Skip to main content

Full text of "Cave vertebrates of America, a study in degenerative evolution"

See other formats


3 1761 06706701 : 


Digitized by the Internet Archive 

in 2007 with funding from 

IVIicrosoft Corporation 




Entrance of Ariguanabo River, Cuba, to its underground channel at 
San Antonio de los Bancs. 

> ■i**^*-* 

— ^ .-■.? >V5->; 

Blind-fish rocks at base of Point Loma, San Diego, California. 









Published by the Carnegie Institution of Washington 

June, 1909 

Publication No. 104 

Copies of this Book 
were first issued 

JUL 91909 

Norinooti QrtH 

J. B. Gusbtng Oo. — Berwick & Smith Co. 

Norwood, Mua., U.S.A. 


Preface vii 

Introductory vii 

Acknowledgments viii 

Conclusions of General Import ix 

A General Consideration of Caves and the Cave Fauna . i 

Caves and the Cave Fauna .............. 3 

Caves in their Relation to the Rest of the Universe 3 

The Nature of the Cave Environment 5 

The Blind Vertebrates and Cave Vertebrates of North America . . _ 9 

The Origin and Dispersal of Cave Animals 12 

The Origin of the Food Supply of Caves 17 

Age of Caves in their Relation to the Variety of Cave Fauna 17 

Divergence in Epigean and Convergence in Subterranean Fishes ...... 18 

Conclusions ................ 21 

Blind and Cave Vertebrates and their Eyes 23 

Mammals 25 

Eyes of the Common Mole 25 

The Cave Rat and its Eyes 26 

The Cave Salamanders ............... 28 

The eyes of Typhlomolge rathbuni 31 

Sclera and Choroid 32 

Pigment Layer Exclusive of the Irideal Parts . 33 

Iris and Ora Serrata 33 

The Retina 35 

The eyes of Typhlotriton spelaus Stejneger 36 

Conclusions as to the eye of Typhlotriton spelaus 40 

Summary in regard to Typhlotriton 41 

The Blind Reptiles 42 

Amphisbana punctata .............. 42 

Methods 42 

General Account of the Eye 42 

Minute Anatomy of the Eye . 43 

Rhineura floridana 48 

Habits ol Rhineura 48 

General Account of the Eye of Rhineura 49 

Minute Anatomy of the Eye of ^A/'«^«ra 50 

Typhlops lumbricalis ............... 54 

General Account of the Eyes of Snakes 54 

Eyes of Typhlops vermicular is . . . . . . . . . . . -55 

Eyes of Typhlops lumbricalis ............ 56 

Conclusions as to the eyes of Blind Reptiles 59 

Amphisbcena 59 

Rhineura 59 

Typhlops . 60 

Eyes of Cyclostome polistotrema stouti 61 

Fishes 62 

General Remarks on the Eyes of Fishes 62 

The eyes of Zygonectes notatus 64 

Typhlogobius : The Point Loma Blind Fish and its Relatives 65 

The Eyes of the Blind Catfish, Ameiurus nigrilabris 69 



Fishes — continued i-age 

The Amblyopsidae 7° 

Relationships of the Amblyopsidae 70 

Distribution of the Amblyopsidae 71 

Amblyopsis spelaits 71 

U'rogluAthys rosa 72 

Typhlkhthys 72 

TyphUchthys subterraneus 73 

Typhlkhthys osborni 74 

Typhlkltthys wyandotte 75 

Chologaster cornutus . 75 

Chologaster papilliferus 75 

Chologaster agassizii 76 

The color of the Amblyopsidae 76 

General habits of Amblyopsis 80 

Respiration 81 

Feeding habits of /4»^/K(?/f« 81 

Habits of Chologaster 85 

Reactions to Light 87 

Breeding habits of Amblyopsis 92 

Rivalry of Males and Secondary Sexual Differences 93 

The Egg and General Development of Amblyopsis 94 

The Migration of the Anus 95 

The Tactile Organs 96 

The 'EdiT oi Amblyopsis 100 

Tioss Amblyopsis ^'■htZT^'i 102 

The Brain of Amblyopsis 106 

Conclusions on the Amblyopsidse 109 

The eyes of the Amblyopsidae no 

Chologaster papilliferus 110 

Chologaster agassizii . . . . ■ ■ 116 

Chologaster cornutus 117 

Typhliththys subterraneus 120 

Troglichthys rosa 126 

Amblyopsis spelaus 134 

Summary of the Eyes of the Amblyopsidic 145 

Development and Later History of the Eye of Amblyopsis 147 

Growth of the Eye from Time of its Appearance 157 

History of the Lens 158 

History of the Scleral Cartilages 158 

History of the Optic Nerve 159 

History of the Development, Maturity, and Degeneration of the Eye 160 

Comparative Rate of Ontogenetic and Phylogenetic Degeneration of the parts of the Eye . 164 

The Future of the Eye 166 

Retardation and Cutting off of Late Stages of the Development of the Eye .... 166 

Causes of Retardation and Cessation in the Development of the Eye 167 

The Eyes of Amblyopsis and the law of Biogenesis 170 

Conclusion 173 

General Summarial Account of the eyes of the Amblyopsidae 175 

Phyletic Degeneration of the eye of the Amblyopsidae 175 

Results of Phyletic Degeneration on the Different Parts of the Eyes of the Amblyopsida 177 

Ontogenetic Degeneration rSo 

Plan and Process of Phyletic Degeneration in the Amblyopsidx 180 

The Cuban Blind Fishes 183 

History of the Work 185 

Zoological position of i'/y^Vf^Ai and Z.««/«fB 187 

Primary and Secondary Sexual Characters 187 

D\sU\hvX\on oi Stygkola 3.nA Luci/uga 188 


The Cuban Blind Fishes — continued f^f^^ 

Nature of the Habitat of ^V^if/foAj and Z.»«'//«^fl 188 

Abundance of .SVXi,'7'c(?/a and /.wf^/^a 1^ 

Origin of the Cuban Blind Fishes 197 

Physical environment of iV^ji.wo/a and yL«£:i/><;fa and their Reactions to it .... 198 

Biological environment of .SV_X^V(»/fi and /.«c//>/;^<» 201 

General habits of Lucifuga and Stygicola 204 

Breeding habits of Lucifuga and Stygicola 204 

The Ovaries of Lucifuga and Stygicola 206 

The Eyes of Z«c//«gif 208 

The Eyes o( Stygicola . 220 

On the Ovary and Ova in Lucifuga and Stygicola 226 

Conclusions in Regard to Lucifuga and Stygicola 232 

The causes of Individual and Phyletic Degeneration 233 




Frontispiece. Entrance to Ariguanabo River, Cuba. Blind-fish rocks at base of Point Loma, San 

Diego, California Title 

Plate A. Twin and Shawnee Caves 6 

1 . Chologaster papilliferus, Spelerpes maculicauda, Spelerpes stejnegeri, and Typhlotriton 

spelceus ............... 12 

2. Spelerpes longicauda zaA Typhlomolge rathbuni 38 

3. Rhineura floridana 48 

4. Eye of Typhlops lumbricalis 54 

J. Atnblyopsis 7d 

6. Chologaster agassizii, TroglicMhys roscr, and Typhlichlhys subterraneus ... 7a 

7. Views of Atnblyopsis, early stages 92 

8. 'XiioSAe. ox^VL% oi Atnblyopsis 2xiA Chologcuter . 98 

9. Heads oi Zygonectes notaius, Chologaster agassizii, Chologaster papilliferus, Typhlichlhys 

subterraneus, Troglichthys ros(x, VinA Atnblyopsis spelceus 110 

10. 'PhologT3iii\\& of \.he ^yts of Atttblyopsis ztiA Troglichihys 132 

11. Carboneria Beach near Matanzas. Cave of the Insurrectos, near the Carboneria . .186 

12. Young of Z«(r//>/^(z in Ashton Cave. Cave Isabella, showing roots . " . . . 190 

13. Stygicola. (Preserved specimens) 196 

14. Living Stygicolas 200 

1 5 . Views of Lucifuga 200 

16. Sections of eye of Z7/c//7/^w ............ 208 

17. Two sections through right eye of Z.«f»/>/^rt . 208 

18. Sections of eye of Lucifuga, showing contents of lens, capsule, and layers of retina . 208 

19. Eyes of Z.7/<://>/frt, showing pigment layer and retina and folding of sclera . . . 208 

20. Eyes of /./^.T/VfX'rt, showing differences in size and structure . ..... 216 

21. Sections through left and right eye-cavities of Z,«<://>/^(i 216 

22. Sections of tyts of Lucifuga, showing pigment layer and cells and oblique and rectus 

muscles 216 

23. Eye of old Z,»(:^«ga, showing pigment mass and fibrillar network about eye . 216 

24. Eye of Lucifuga 222 

25. Eye of Stygicolas znd Luctfugas 222 

26. Eye of Stygicola 222 

27. Ovaries of Lucifuga and Stygicola ■ . 232 

28. Sections of ovaries 232 

29. Sections of ovaries 232 


1. Pigment epithelium. 

pi. Densest pigmented section of the pigment epithe- 
hum, just below the nucleus. 

2. Rods and cones. 

3. Outer nuclear layer. 

4. Outer reticular layer. 

5. Horizontal cells. 

6. Inner nuclear layer. 

7. Spongioblastic layer. 

8. Inner reticular layer. 

9. Ganglionic layer. 
10. Optic-fiber layer. 

a. 0. Ophthalmic artery. 

am. Ameloid bodies of the pigment epithelium. 

*. Brille. 

hac. Rod. 

ci. p. Ciliary process. 

cj. Conjunctiva. 

cj. s. Conjunctival sac. 

chr. or chd. Choroid. 

chr. I. Choroidal lymph. 

chr.f. Choroidal fissure. 

en. Cones. 

en. nl. Cone nuclei. 

eor. or ern. Cornea. 

cps. or cpl. sng. Blood-corpuscles in normal vessels, 

eps. s. Stagnant blood-corpuscles. 

d. Dorsal aspect of eye. 
dr. Dermis. 

e. m. End member of cone. 
/'. cj. Fornix conjunctiva. 
fr. ol. Olfactory pit 

hd. or hyl. Hyaloid membrane. 

H. gl. Harder's gland. 

i. Iris. 

i. I . Outer layer of iris. 

«'. 2. Inner layer of iris. 

i. e. Interpolated cells. 

/. Left side of eye. 

/.*, I?, I? First, second, and third labial scale,| 

Ins. or /. Lens. 

I. e. Lens capsule. 

M. Miillerian nuclei. 

m. m. Middle member of cone. 

mse. or mu. Eye muscle. 

nl. Nucleus. 

nl. con. Cone nuclei. 

ttl. f. Nuclear fragments. 

nl. g. Nuclei of the ganglionic cells. 

nl. I. or »/.' Elongate nuclei of the pars ciliaris. 

nl. Muel. Mullerian nuclei. 

«. op. Optic nerve. 

n. s. Nasal scale. 

oc. Eye. ■ 

0. c. Ocular scale. 

0./. Orbital fat. 

0. s. Ocular scale. 

0t. Otolith. 

/. Pupil. 

p. i. Palpabra inferior. 

/. s. Palpabra superior. 

pi. s. Pigment appearing in optic cavity with senes- 

pi. sph. Pigment spheres. 

p. I. Pigment layer. 

po. s. Preocular scale. 

pr. nl. Processes of the cone nuclei. 

pupl. Pupil. 

r. Right side or retina. 

r, or rt. Retina. 

ro. Rostral. 

scl. Sclera. 

scl. c. Scleral cartilage. 

suio. or s6. orb. Suborbital. 

V. Ventral aspect. 

vit. Vitreous body. 

X. Flattened cells beneath pigmented layer, of doubt- 
ful significance. 

y. Flattened cells beneath inner nuclear layer, of 
doubtful significance. 




A cave is a unit of environment so well circumscribed and of such simplicity 
that we may know its contents, its elements, and its conditions nearly as well as 
the experimental zoologist knows the contents and conditions of his aquarium. 
These contents and conditions are of rare uniformity, changing but little from day 
to night, from season to season, or from decade to decade. The point of chief 
interest in the cave environment is the total absence of light in all parts except 
about its mouth. Probably no animals have a more intimate environmental 
adaptation than those inhabiting caves. This adaptation is largely of color and 
structure of eye, which modifications are surpassed only by the functional adapta- 
tion of the tactile apparatus of the blind forms. 

While no one has followed, and although we may not be able to follow in 
detail, the steps through which the cave animal has acquired this environmental 
adaptation, a knowledge of the present condition of their unchanging environment 
gives us a knowledge of what it has been during their entire period of development. 

We know, or can know, what the present stage of their adaptation is. Not in- 
frequently we know what the condition of the animal was at the start of its cave 
experiences and enough of the steps along its line of evolution (indicated by the 
degrees of adaptation reached by different members of the group) to enable us to 
form so clear a picture of its entire route of evolution that we may conjecture what 
elements of the environment caused the modifications, and by what process they 
were brought about. We have, in other words, a long experiment conducted by 
nature unrolled before us. 

I propose in this work to give an account of the cave as an environment; to 
bring together in a revised form the papers on blind and cave vertebrate animals 
so far published by myself and my students, together with further observations on 
the species previously considered, to consider the habitat, mode of life, and the 
origin of the Cuban blind fishes, and to give an account of their eyes. 

My first experience with blind vertebrates was in 1886, when Superintendent 
Funk sent to Indiana University a living blind fish which had been taken from a 
well at Corydon, Indiana, and which proved to be a new species, Typhlichthys 
Wyandotte, the only representative of the genus so far taken north of the Ohio 
River. Later, when a stay in southern California came in prospect, a study of the 
blind fish, Typhlogobius, living under rocks along the base of Point Loma, was one 
of the first definite plans formed. 

When, in 1890, I returned to Indiana and was once more within reach of the 
caves, the problem again came up. My laboratory is excellently located for the 
study of cave faunas, the series of caves to which Wyandotte, Marengo, Mammoth, 


Colossal, and Nickajack belong, beginning in or about the campus of Indiana 
University. But while seemingly ideally located, and in spite of the fact that 
numerous trips were made to Indiana caves, especially those from which blind 
fishes had been reported, no blind fishes were found till 1896. 

In May, 1896, I was again looking for blind fishes east of Mitchell, Indiana, 
this region being drained by underground streams. East of Mitchell several of 
these find their exit in caves of romantic beauty in the escarpment flanking the valley 
of White River (plate A). The roof over one of the streams has fallen in at two 
places, Dalton's Spring and Twin Caves. At Dalton's Spring the cave-stream 
runs above ground for about 100 yards when it again enters its subterranean course. 
Within sight of the lower opening of the "spring" I saw two bhnd fishes swimming 
in a quiet pool. I secured about 20 specimens and had found the stream which in 
its varying reaches has furnished me with an unlimited supply of specimens which 
have enabled me to give the complete history of the eye of this species, AmUyopsis 
spelcEus De Kay. More material has been obtained from this cave than from all 
others put together. In 1903 the State legislature of Indiana placed the land, 
about 182 acres, on which are the entrances and exits to this stream in the keeping 
of the trustees of Indiana University. While some litigation has arisen as to the 
ownership of the farm, it will probably be permanently preserved as a State park. 


Through grants from the Elizabeth Thompson Science Fund and from the 
American Association for the Advancement of Science I have been able to visit 
the cave regions of southwestern Missouri, about San Marcos, Texas, Corydon, 
Indiana, and Mammoth Cave, Kentucky. In 1902, through a grant from the 
American Association for the Advancement of Science and assistance from various 
other sources, I was able to visit the blind-fish caves of Cuba. Subsequently the 
Carnegie Institution of Washington aided me in making additional investiga- 
tions in Cuba. The part of the present volume dealing with Stygicola and Lucifuga 
is my final report on the work carried on with this aid, and in it a detailed account 
of the Cuban work is given. 

Prof. S. A. Forbes kindly lent the drawing for figure A, plate i. The draw- 
ings of sections of eyes were made under my direction by Mrs. E. R. BieHng in the 
laboratory of Prof. R. Wiedersheim, in the University of Freiburg, Germany, and 
I am indebted to Professor Wiedersheim for placing his laboratory at my disposal. 

I am under many obhgations to various friends, both at home and in Cuba. 
Mr. Oscar Riddle, Dr. John Beede, Mr. John Haseman, Mr. Norman Mclndoo, 
and Mr. T. L. Hankinson acted as volunteer assistants on various Cuban trips, 
always working without remuneration and in part paying their own expenses. 
The late Prof. Jose T. Torralbas, Prof. Carlos de la Torre, Mr. Pascual Ferreiro, 
Dr. Felix Garcia, and the Director of the Cuban Agricultural Station, Prof. F. S. 
Earl, assisted me materially in various ways. 

The assistance of my friend, Mr. Francesco Martinez, has been invaluable. 
His finca, the "Isabella," is at the margin of the cave region of Cuba, and in the 
interval between our trips he ferreted out unsuspected caves, determined their rich- 
ness in blind fishes, and put himself at our disposal in guiding us to his various finds. 

Prof. D. W. Dennis of Earlham College, Richmond, Indiana, made the micro- 
photographs in a manner to leave nothing desired (plates 9, 10, 16-23). 


Mr. Lewis H. Wild, under the direction of Prof. J. Rcighard, made a series of 
photographs of entire eggs and embryos (plate 7). 

Mr. Samuel Garman sent me my first specimens of the blind fish, Troglichthys. 

Dr. B. W. Evermann of the Bureau of Fisheries and the late Prof. W. Norman 
secured me specimens of Typhlomolge. 

Prof. Wm. Roux, Dr. F. R. Lillie, and others kindly consented to the repub- 
lication of articles issued in the journals under their editorship. 

I desire also to express my high appreciation of the interest taken by the authori- 
ties of Indiana University, especially by President William Lowe Bryan, in the 
various trips and plans necessary to bring this work to a successful conclusion. 

The present work forms No. 97 of the Contributions from the Zoological Lab- 
oratory of the Indiana University. 

Finally, I wish to express my indebtedness to her who as Rosa Smith guided 
me to the blind-fish rocks at the base of Point Loma, and who as Mrs. R. S. Eigen- 
mann collected for me at the same place, has acted as editor of the various papers 
that have appeared, and through the twelve years during which my leisure has 
largely gone to the blind vertebrates has ever been ready vnth advice, encour- 
agement, and assistance. 


(i) The bleached condition of animals living in the dark, an individual envi- 
ronmental adaptation, is transmissible and finally becomes hereditarily fixed. 
(See page 80.) 

(2) Ornamental secondary sexual characters not being found in blind fishes 
are, when present, probably due to visual selection. (See page 94.) 

(3) Individual degeneration of the eye may begin in even earlier stages of 
development until nearly the entire development becomes affected, that is, func- 
tional adaptations are transmissible. (See pages 172 and 235.) 




The environment favorable to animal life is limited to a thin layer of water, 
earth, and air. P'rom its deepest to its most elevated point this layer does not 
much exceed lo miles ' in thickness. At no particular point does it exceed much 
more than half this thickness ; and usually the layer is but a few feet thick. About 
half the total thickness is below sea-level and the other half above it. The places 
where the ocean has a depth of 5 miles are few, but in these places the greatest 
depth of possible environment is found. The favorableness of the environment 
diminishes rapidly with the depth. The depth of the possible environment at any 
point on land above the surface is very limited, and beneath the surface it depends 
on conditions ; solid rocks may limit it to the surface and soil may permit mam- 
mals, and especially insects, to burrow several feet beneath the surface. Under- 
ground watercourses, which are caves in the formation, may enable animals to live 
several hundred feet beneath the surface of the ground. The animals thrown out 
by artesian wells attest this. Typhlomolge is occasionally thrown out of the 
artesian well 190 feet deep at San Marcos, Texas. The plant environment stops 
at the surface of the ground ; ^ animal life diminishes rapidly within a few feet of 
the surface unless trees cover the ground. Animal environment definitely stops 
at the tops of trees, though the air above them may be temporarily visited. 

While the depth of the environment at any point is only a few feet on land, 
because the surface of the land itself rises to a few miles above sea-level, the total 
depth of the environment above sea-level is considerable. The fauna rapidly 
diminishes in either direction from sea-level, and were it not that the extreme limits 
of the environment, above and below, furnish rare, sometimes peculiarly adapted 
forms, sometimes relicts, the numbers of individuals and types found would not 
repay the exploration of the ocean depths and mountain heights. 

Since the environment varies vdthin the limits of the possible existence of living 
matter, from the extreme of wetness and dryness, of heat and cold, of depth and 
height, of light and dark, etc., we may divide the environment into many distinct 
units within which the conditions are similar or alike. It is profitable at present to 
call attention only to discontinuous and continuous units of environment. Similar 
or identical conditions may stretch uninterruptedly in one or more directions indefi- 
nitely, permitting the free movement of its inhabitants from one part to another. 
The continuous unit of environment of greatest extent is furnished by the ocean at 
considerable depths. Light and temperature conditions and seasonal fluctuations 
are reduced to the minimum and are nearly uniform under the whole surface of 
the ocean, furnishing an ideal of the type of the continuous environment. This 
particular environment is continuous not only as to space, but also as to time. 

The surface of the ocean forms an equally continuous area, but because tem- 
perature and light conditions differ greatly in different parts of the globe we must 
here deal not with a single but with several distinct units of environment, each large 
in extent. If we assume the conditions in the north polar sea to be identical with 

' Highest mountain, deepest ocean. ' Some fungi are found in caves. 


those of the south polar sea, these form a discontinuous unit of environment, a unit 
whose parts do not form a portion of a continuous area and whose inhabitants 
can not migrate from one part to the other. 

If we assume the conditions in the equatorial Atlantic to be the same as those of 
the equatorial Pacific, we are again dealing with a discontinuous unit — discontinuous 
because the inhabitants of one part can not migrate to the other. If we examine 
these two units more closely, it becomes evident that the Arctic and Antarctic 
oceans have always formed a discontinuous unit. Arctic conditions have never 
prevailed between the two. On the other hand, the equatorial Atlantic and the 
equatorial Pacific were formerly connected in Colombia and formed one continu- 
ous environment. The land area and the fresh waters near the equator from 
Para to the Andes form a continuous unit of environment, and the Galapagos 
Islands to the west of it form a discontinuous unit, each separate island forming a 
continuous unit of a smaller order. It is evident that there are degrees of discon- 
tinuity, depending in part on the length of time the discontinuity has existed, and 
in part on the space separating the nearest parts of the unit. 

Caves are discontinuous units of environment whose elements have always 
been separate. It is possible that in some areas a large complex of different under- 
ground channels exists. An east to west fault has lowered the southern part of 
Texas, or has raised the northern part, many feet. The dividing line is an abrupt 
escarpment across the State. This fault has favored the formation of underground 
watercourses, and inasmuch as river valleys do not cut down to the underground 
channels, it is possible that they form a network of channels or a continuous unit 
which permits the ready migration of its inhabitants from one part to another. 

The lower area on the southern slope of Cuba, between Canas on the west and 
an undetermined point east of Union, is drained by underground rivers. No 
valleys cut down to these rivers, and since this part of Cuba has sunk in recent 
times, the land being only a few feet above sea-level, it is possible that we again 
have a complex of underground channels permitting the migration of its inhabit- 
ants. However, it is also possible that the streams run in separate courses. The 
absence of Lucifuga from the eastern caves favors this hypothesis. At best we 
have here several degrees of continuity. 

The large streams cut the cave region of Kentucky, Indiana, and Missouri into 
sections, their beds lying deeper than the caves. These caves are, therefore, part 
of a discontinuous environment. These facts must be constantly borne in mind 
in considering the origin and dispersal of cave faunas. 

It is quite out of the question in this connection to give even a partial list of 
North American caves, or an account of the North American cave regions. The 
region to which Mammoth Cave belongs reaches from near Bloomington, Indiana, 
through Kentucky into Tennessee and embraces many thousand square miles of 
territory. Only the larger streams whose rapidly deepening channels have made 
the caves possible flow on the surface. "One may travel on horseback all day, 
through certain parts of Kentucky, without crossing a single running stream ; all 
the rain water that falls being carried down through the sink holes into caverns 
below where are the gathering beds that feed the few large open streams of the 
region, of which Green River is an example. It is reported that there are 4,000 
sink holes and 500 known caverns in Edmondson County (Kentucky) alone." * 

' For an account of the principal caves of North America see Hovey, Celebrated American Caverns, Cin- 
cinnati, 1882 and 1896; and Packard, The Cave Fauna of North America, Memoirs of the Nat. Acad. Sci. 
vol. 4, 1888. 



Each cave is a distinct unit of environment and needs special consideration. 
In the present work, we can deal only with the general features of this environment. 
The chief element for consideration is the absence or reduction of the amount of 
light and the relative constancy of other physical conditions. On this basis a 
cave may Ixj divided into three regions: (i) the twilight region just within the 
cave, bounded by the distance to which light penetrates from without — this 
part shades generally from epigean conditions to the real cave conditions ; (2) the 
region of fluctuating temperatures ; (3) the inner cave region. 

These different sections occupy greatly variable parts of different caves. In 
Mammoth Cave the twilight region is large enough to contain a tennis court and 
reaches some distance beyond the "iron door." Some Cuban caves are entirely of 
the twilight character, usually containing an abundant fauna, consisting largely of 
occasional, regular, or accidental visitors from the outside. The second region in 
Mammoth Cave reaches to the Mammoth Dome. On a cold winter day I found 
ice stalagmites on the floor of the entrance gallery just before it enters the dome. 
In certain of the ice caves the entire portion beyond the twilight area may belong to 
this section. In caves of the tropics, on the other hand, it may not exist at all. 
The third part is the cave par excellence — the inner section, but little influenced 
by external conditions. Here there is absolute darkness at all times, both day and 
night, summer and winter following each other without very decided change in 

The temperature differs in the various parts of the same cave and also changes 
slightly with the seasons. In the center of the Shawnee Cave at Mitchell the fluc- 
tuations in temperature during a week do not equal the error of the recording ther- 
mograph arising from unequal trimming of the paper, the absorption of water, 
etc. The total fluctuation during a year is 2.2° C. It is remarkable that this 
record of cave temperature is taken in a cave open at both ends with a current of 
air flowing through it at times. The instrument is placed where it would be least 
affected by these currents, that is, in a large room near the center of the cave about 
15 feet above water-level. 

Glaciferes, or ice caves, are found in various places. They exist wherever the 
prevailing direction of the vnnds and nature of the cave causes a strong inflow of 
air during the winter, reducing the temperature to below the freezing point. The 
summer winds do not blow in the same direction, and convection currents are pre- 
vented by the nature of the cave.* 

Between June, 1906, and February, 1908, the fluctuations in the temperature in 
the water where it leaves Shawnee Cave ranged from a maximum of 17.3° C. to 
7.4°, or through about 10° C. 


' A very extensive list and excellent account of glacieres is given by Batch in his Glaciferes or Freezing Caverns, 
1900. Concerning the cause of glacieres, he says, on page 148: "The cold air of winter sinks into and permeates 
the cave, and in course of time freezes up all the water which, in the shape of melting snow or cold winter rain 
or spring water, finds its way in; and once ice is formed it remains long after ice in the surrounding open country 
has melted away, because heat penetrates with difficulty into the cave." 

^ This range becomes interesting when compared with the range of temperatures in a lake. Professor Birge 
gives the ranges of the water at the surface and at the depth of 18 m. for Lake Mendota: 
Surface, 1895 ... 0° to 24° Bottom, 1895 . . . 1.5° to 17. i" 

Surface, 1896 ... 0° to 26° Bottom, 1896 . . . 3° to 16° 



Conditions of moisture, while practically untform in some parts of caves, fluc- 
tuate in others more than any other element of environment. The maximum degree 
of moisture is naturally found in the pools and streams. On the other hand, in 
the upper parts of Mammoth and Wyandotte Caves the dust lies undisturbed for 
years. In Mammoth Cave the tracks of oxen made in i860 are now shown to 
visitors, and I am told that in Wyandotte the still older tracks of the moccasined 
Indians are perceptible to-day. There are, however, parts of caves where the 
moisture dripping through from above is considerably increased after a rain, and 
the River Styx in Mammoth Cave rises 60 feet above low-water mark. The creek 
in Shawnee Cave sometimes fills parts of the cave to the ceiling. 

The conditions of the water also change very greatly. At ordinary times it 
may be very clear ; after rain it may carry a large amount of sediment. In its low 
condition it may flow very quietly, in its high condition be a torrent. The water, 
then, fluctuates in amount, clearness, and swiftness, with meteoric conditions. 

Charts of simultaneous records on two self-registering barometers show the close 
agreement in changing barometric pressures inside a cave and outside it. One of 
the instruments was placed about 90 feet above the exit of the cave, the other near 
the middle of Shawnee Cave. Records chosen on account of peculiarities in the 
rise and fall of the pressure at certain times leave no room for doubt that baro- 
metric changes similar to those of the outside take place in the caves. 

The following table shows the temperatures for air and water in Donaldson 
and Shawnee Caves in 1906 and 1907 : 

Temperatures for air 

and water in Donaldson and Shawnee Caves. 










ture of air 

ture of 

ture of air 

ture of 

ture at 

ture of air 

ture of 


in center of 

water at its 


in center of 

water at its 


exit from 


exit from 


exit- from 























August . . . 







August . . . 







March . . 













12. 1 









May .... 











June — 






12. 1 


* Tile iiiglier temperatures are caused by rains and last only a few hours after a heavy rain. During the first lo days in September, 1 907 , 
the temperatiu^ of the water was 14.5, 15.6, 17.3, 16, 14.9, 14.6. 13.0, and 15.3 on successive days. During the last 10 days of the month it 
ranged from 150 to 15.5". 

■ From the ist to the 15th the temperature was between 10,6 and 11.6. 

Currents in water and air differ materially in different caves and at times in 
the same cave. In the Cuban blind-fish caves there is neither appreciable air- 
current nor water-current, so that the evaporation from the quiet surface of the 
water forms a covering crust of carbonate of lime and magnesium. In the blind-fish 
caves at Mitchell, Indiana, a small current of water flows during normal conditions. 
The stream becomes a raging torrent in high water. 

Currents in the air may be caused, (i) by the flow of water; (2) by the epigean 
air-currents ; (3) by changes in the atmospheric pressure ; and (4) by differences 
in temperature.* 

' A detailed study of the currents of air and temperature of the water in the Mitchell Caves will be published 
within a year. 




O ft 

? S 

o f^ 











^ o 
e = 

& Q- 

^ 3 

n n 


C §■ 

n "^ 


















5- 3 




s P 

< I 










cr ■ 


In Mammoth Cave a very perceptible air-current flows into the top of the dome 
from Little Bat Avenue. It probably descends to the bottom of the dome and then 
ascends at the side to flow out at Sparks Avenue. This current was flowing at 
the rate of 8,640 feet per hour on November 30, 1902. It is probably caused by 
a thin fall of water which descends from the roof of the dome to the bottom. 

By far the most violent air-current may be caused by a change in the atmos- 
pheric pressure in the air without. These currents are perceptible only in caves 
of considerable extent, and become violent when the opening is insignificant com- 
pared with the size of the cave. 

When the weight of superincumbent air is lightened, the compressed air in the 
cave expands and there is an outrush of air through the opening. If, on the other 
hand, the barometric pressure increases when the superincumbent air column 
gains in weight, there is an inrush of air. I have been at the entrance of Mammoth 
Cave when the internal and external pressures were so equalized that the anemom- 
eter would show ingoing and outgoing currents alternating irregularly every few 
minutes. In 1902 I was also at the entrance ' when the anemometer showed the 
following rates per hour for air going in : November 29, 9 a. m., 46,350 feet ; 6 p. m., 
39,840 feet; November 30, 7 a.m., 50,290 feet; 9"" 40" a.m., 55,830 feet; and 
12'' 30" p. m., 7,800 feet. 

Mr. A. M. Banta reports from Mammoth Cave that on January 31, 1903, "At 
the gate the air-currents were surprisingly fitful. The current was running in 
40 seconds, stopped 15 seconds, flowed out 8 seconds, stopped 10 seconds, and then 
ran in for 2 minutes, when we left." His records give the following rates per hour 
of air going in during February, 1903 : February 18, 12 m., 76,464 feet ; 5'' 30" p. m., 
77,396 feet; 6'' 20"° p.m., 79,896 feet; February 19, 10 a.m., 76,692 feet; 12 m., 
68,904 feet; and February 21, 9 a. m., 56,556 feet. 

I know of no direct record of currents due to changing temperature on the 
outside. Until direct observation with an anemometer had been made the general 
impression among the guides at Mammoth Cave was that air rushed in during one 
part of the year and out during the other. On cold winter days at Mitchell frost on 
the bushes showed that a gentle current of the damp cave air was flowing out from 
the upper part of the cave. The strength of the convection currents is undoubtedly 
dependent in large measure upon the shape of the cave and the nature of the open- 
ing. But the influence of water-currents or winds might at any time be sufl&cient 
to change the direction of the convection currents. 

Nothing very definite can be said about the size of the environment afforded 
by a cave.^ While it is known that some caves are much larger than others, it is 
never certain how large the unexplored or unexplorable part of a cave may be, how 
far the smaller cracks lead, and in how far they may establish intercommunica- 
tion between neighboring caves. 

' A wall partially closes the entrance avenue so that the air passes in and out through a narrow gate where 
the currents were measured. 

' Hovey (The Mammoth Cave of Kentucky 1897, p. 64) makes the longest course in Mammoth Cave 
from the entrance to Grogham Hall about 4.5 miles; the total length of all the known channels is several times 
that. The width and height mav vary greatlv from the many cracks where one has to crawl to Chief City between 
450 feet (Hovey) to 541 feet (Call) long, and'an average width of 175 feet (Hovey) to 190 (Call), with a maximum 
width of 287 feet. 

Blatchley says of Marengo (p. 157), "Marengo Cave has been advertised far and near as containing 7 miles 
of underground passages. Our measurements showed its total length to be 3,850 feet, or 0.7 of one mile. The 
main channels of Wyandotte Cave we determined to be 4.21 miles long." Very many of the caves are but a few 
inches in diameter and too small to be entered. 


The Mitchell Caves can be traced for over 2 miles. Given that they are 3 
kilometers long, their average width is perhaps 8.3 meters. This would give an area 
of 25,000 square meters. As a stream flows their entire length a direct comparison 
can be made with epigean conditions by taking a stream of similar size and length 
above ground, with territory equaling the width of the cave. The fauna of the 
epigean area of equal size is incomparably richer than the subterranean one.* 

The biological environment of cave animals is comparatively simple. While 
much has been written on them, the only account of the interrelation of the animals 
of any cave has recently appeared in a publication by one of my students, Mr. 
A. M. Banta (publication No. 67 of the Carnegie Institution of Washington). 

' For a discussion of the age of caves see page 1 7. 



The blind vertebrates do not belong to one class nor do those within one class 
belong to one family. The blind fauna is very diverse in character and origin, 
but not all families of vertebrates are represented. A certain predisposition in 
habit and structure must be present to enable a species to dispense with light and 
to live in caves. A large blind epigean animal might secure its food and meet its 
mate, but it could not escape its enemies. Large blind forms are therefore impos- 
sible. While the size of a sun-fish (Lepomis) might not preclude it from entering 
caves, the fact that it detects its prey by sight excludes it entirely from the possibly 
blind. There is, on the other hand, no reason why members of the nocturnal 
Silurida;, for instance, should not become blind. 

No large mammals are blind, nor have large mammals permanently taken up 
their abode in caves. Bears visit caves, and raccoons, minks, and ground hogs also 
enter them. The latter two confine their underground wanderings mostly to small 
caves or to caves of their own making. None of these animals permanently live in 
caves ; they are all twilight animals and depend on light for their continued exist- 
ence ; they have normal eyes and are not otherwise modified for life in caves. 

Blatchley reports that a number of cats have established themselves in Wyan- 
dotte Cave, where they bring forth and rear their young. Nothing is known about 
their adaptations. They have exterminated the cave rats and are said to place 
themselves in a narrow passage of the cave and capture bats passing through. 

Neotoma pennsylvanica, a wood rat widely distributed in eastern America, has 
entered caves. It was formerly found in Wyandotte Cave, but has been extermi- 
nated there. In various caves white-footed mice are found, but they are not blind. 

The common mole {Scalops aquaticus), the long-tailed mole (Parascahps 
brewert), and the star- nosed mole (Condylura cristata) burrow in the ground and 
are partly or entirely blind. They are not found in caves. 

Bats, which are twilight animals, but have minute eyes, do not depend on their 
eyes to secure food; they fly at night because their food is then abroad. There 
are in North America and the West Indies a large number of bats partly or totally 
blind. Many, if not all of those of the temperate region, winter in caves; a smaller 
number spend only the day there. They do not secure much, if any, of their 
food in caves and simply use them as shelters in a more systematic manner than 
bears do. 

There are no blind birds, and no birds, as far as I know, permanently live in 
caves. The phoebe utilizes the entrances as it uses all other similarly sheltered 
places to nest. In Cuba a small owl is sometimes found in caves, but I know of 
none that makes it a permanent home. Many owls are adjusted to existence in 
twilight, but that they are dependent on their eyes is shown by the increase in size 
of their eyes. Other animals, depending on their eyes but living in the dusk, 
have similarly enlarged eyes. This is especially well shown by marine fishes liv- 
ing at twilight depth. 

There arc no cave reptiles, nor do reptiles temporarily enter caves for shelter, 
as do mammals. One turtle found a little distance inside of one cave was evidently 
accidental. I have never seen a snake in a cave, but once secured a copperhead 
at the entrance to one. But there are numerous blind lizards and snakes that 


burrow in the ground. Aniella, a small, legless, burrowing lizard of California, 
probably indicates their origin. This lizard has well-developed eyes. It burrows 
in sand and gravel. I have frequently seen it cautiously thrust its head out of the 
ground for an instant as if to take a survey of the field. It evidently still uses its 

Amphisbaenians,* which are vddely distributed over the warm parts of the globe, 
burrow in the ground or live in ant hills, and are partially or totally blind. The 
blind snakes, members of the Typhlopidae, have similar habits.^ 

Many salamanders Hve in damp earth under logs or rocks. It is but natural, 
therefore, that they should be found in or about the entrances to caves, where 
sheltering rocks are not infrequent. Others are true cave animals. Two of the 
salamanders in North America that habitually live in caves have apparently quite 
normal eyes. They are Spelerpes macuUcauda found from Indiana and Kentucky 
to Missouri, and Spelerpes stejnegeri from southwestern Missouri. Two others 
living in caves have quite degenerate eyes, Typhlotriton spelceus from caves in 
southwestern Missouri, and Typhlomolge rathhuni from the caves of Texas. Pro- 
teus, the nearest relative of the latter, lives in the caves of Carniola. There are no 
blind epigean salamanders. Of Anura there are no permanent residents in caves, 
nor are there any blind forms. A jumping animal would be sure to meet with dis- 
aster in a cave if it practiced its usual mode of progression. 

The classes of vertebrates furnishing the largest number of blind forms are 
■fife fish and fish-like vertebrates. Excluding the Branchiostoma, the Cyclostomes 
have for the most part degenerated eyes. PoUstotrema stouti is quite blind. 

Benthahatis moresbyi Alcock is a blind Torpedinid Selachian from Travancore, 
from a depth of 430 fathoms. 

Of the lowest teleosts the Siluridae are represented by Gronias nigrilabris Cope, 
which occurs in a cave near Philadelphia.' The eyes of many other cat-fishes are 
not highly organized and but little used in detecting food.* Other cat-fishes are 
occasionally met in caves, but no others are permanent residents. 
_^- The cave fishes of North America, par excellence, are the Amblyopsidas. All 
the members of this family, 8 in number, have degenerate eyes ; 5 have mere ves- 
tiges; 6 permanently live in caves; i is known only from a spring and another 
from open streams. These v/ill be considered in detail later. 

In Cuba 2 fishes belonging to a marine family, the Brotulidae, have become 
adapted to a cave life in fresh water. Both are blind. Many of their marine 
relatives are also blind. 

Along the coast from San Pedro, California, to Encenada, Lower California, 
but more particularly at the foot of Point Loma, a blind goby lives under rocks 
embedded in sand between high and low tide. 

' "All the members of the family are burrowers, and many live in ant nests. They bore narrow galleries 
in the earth, in which they are able to progress backwards as well as forwards. On the ground they progress 
on a straight line, by slight vertical undulations, not by lateral movements, as in other limbless reptiles; the tail 
of many species appears to be more or less prehensile. The food of these lizards consists of small insects and 
worms. * * * As many as 65 species are characterized in this account; 39 are American, of which only a 
(Chirotes and Rhineura) occur north of the Tropic of Cancer, and 4 {Amphisbana) in the West Indies." — Boul- 
enger. Catalogue of Lizards, vol. 11, p. 430, 1885. 

'There are altogether about lOO species reaching, in the Americas, as far north as Cuba: Typhlops lum- 
bricalis, Yucatan; Typhlops microstomus, Mexico; and Typhlops /ewwij, Guatemala and Mexico. 

' Two blind cat-fishes have recently been described from Brazil. 

* Herrick found that the cat-fishes detect their food, not by means of their eyes or olfactory organs, but 
by the touch and taste organs over the body and in the barbels. 


The fishes, blind or partly blind, living in the depth of the ocean bordering the 
American continents are Ipnops murrayi Giinther and Ipnops agassizii Garman. 
The former lives at depths varying from 955 to 2,158 fathoms and is very widely 
distributed. The second one is known from the type specimens from Lat. 2° 34' N., 
92° 6' W., at a depth of 1,360 fathoms. Iptiops stands alone in a family and is the 
only vertebrate in which no eyes have been found. 

The Brotulidae have several members blind, or with very minute eyes, in various 
parts of the globe. With the exception of the fresh-water species of Cuba, the 
only ones found in the neighborhood of America are Aphyonus mollis Goode and 
Bean, 955 fathoms, and Alexeterion parfaiti Vaillant, 2,736 meters. Other deep-sea 
blind fishes are Aphyonus gelalinosus Giinther between Australia and New Guinea, 
1,400 fathoms; Mancalias shiifeldlii Goode and Bean, 372 fathoms; Paroneirodes 
glomerosus Goode and Bean, 1,260 fathoms; Tauredophidittm hexlii Goode and 
Bean, Bay of Bengal, 1,310 fathoms; Typhlonus nasus Giinther, north of Aus- 
traha and Celebes, 2,150 and 2,440 fathoms. 



It has been shown that many cave animals have good eyes. Epigean animals 
with degenerate or no eyes arc not rare, hence the origin of the cave fauna and of 
the blind fauna are two distinct questions. This was first recognized by H. Gar- 
man and indorsed by Eigenmann and by Hamann. Other writers have usually 
confused the two questions, and indeed they may become one when they concern 
an animal that has become blind concomitantly with its cave colonization. A 
consideration of the forms that arc not found in caves, and the reasons why they 
are not found there, is in this connection possibly more illuminating than the 
direct consideration of the cave forms. 

Caves may have become populated by one of the four following processes : 
(i) Animals may by accident have been carried into caves. 

(2) Animals may, step by step, have colonized caves, becoming adapted 

to the environment as successive generations gradually entered 
deeper and deeper recesses of the caves. 

(3) Animals which had elsewhere become adjusted to do without light 

may have gathered voluntarily in caves. 

(4) Animals may have developed along with the development of the caves. 

First process : This process was imagined by Lankester to operate as follows : 

Supposing a number of some species of arthropod or fish to be swept into a cavern or to be 
carried from less to greater depths in the sea, those individuals with perfect eyes would follow the 
glimmer of light and eventually escape to the outer air or the shallower depths, leaving behind those 
with imperfect eyes to breed in the dark place. A natural selection would thus be effected. In 
every succeeding generation this would be the case, and even those with weak but still seeing eyes 
would in the course of time escape, until only a pure race of eyeless or blind animals would be left 
in the cavern or deep sea. 

While this is a possible mode of origin of cave animals, and even of blind ones, 
it is highly improbable that many or even any animals depending, as he supposes, 
on their eyes have thus come to first colonize the cave. Fishes are annually swept 
into caves, but these are not able to permanently establish themselves in them. 
To do this the fish must have peculiar habits, special methods of feeding and 
mating, before an accidental colonization can become successful, and if they are 
so adapted for a cave existence, they would probably voluntarily colonize the caves, 
without waiting for an accident.' The Amblyopsidas are a small family of fishes, 
8 species being known. They form a very small part of the large fish fauna about 
the North American caves. But since 6, possibly 7, of the species of this family 
are cave dwellers, and only one of the numerous other fishes is permanently at 
home in the caves, we must suppose, if the theory under consideration is the correct 
one, that the accident of being carried into caves happened to 6 or 7 out of 8 of the 
Amblyopsidae, and to only i of all the other fishes about the caves. The absurdity 
of this supposition is self-evident. A comparison of the abysmal fauna with the 
pelagic and shore faunas would probably give us similar results. 

• A distinction ought possibly to be made between the aquatic cave animals that will be discussed under the 
"fourth process," and non-aquatic forms. Non-aquatic cave animals are later immigrants of caves. These must 
either be voluntary recruits from the twilight fauna about the entrance of the cave or they must have become other- 
wise adju-sted to live in the dark. There is no difficulty in accounting for the presence of Myriopoda on this score 
nor for the other forms habitually found under bark and under rocks. Myriopods arc everywhere abundant in 
the caves of North America and they (if any animals) may have accidentally been carried into caves with sticks 
of wood or trunks of trees. 




MrsERBielin§ M. 

A. Chologaster papilliferas. 

B, Bb. Spelerpes stejnegeri. 1 1 2 mm. Wilson's Cave, Sarcoxie, Missouri. 

C, Cc. Spelerpes maculicauda. I 30.5 mm. Wilson's Cave, Sarcoxie, Missouri. 
D, Dd. Typhlotriton spelaeus. I 34 mm. Marble Cave, Missouri. 


Second process : The second theory is that of Herbert Spencer : 

The existence of these blind cave animals can be accounted for only by supposing that their 
remote ancestors began making excursions into the cave, and, finding it profitable, extended them, 
generation after generation, farther in, undergoing the required adaptations little by little. — Popu- 
lar Science Monthly, xliii, 487 and 488. 

I can offer no objection to this theory. It presupposes the existence of caves, 
and it is perfectly possible that many cave animals have arisen in this way. The 
abundant twilight fauna in the entrance of caves argues in favor of it. Spelerpes 
macttlicauda and other salamanders, which are so frequently found a short (lis- 
tance within caves and even in remote recesses, seem to be present colonizers that 
bear out Spencer's view, though it is possible that these should be grouped with the 
animals next to be considered. 

Spelerpes maculicauda has not yet been affected, as far as its structure is con- 
cerned, by its habits. It is a nascent cave form that may result in the future in 
a single blind species of wide distribution, or a number of species in the groups of 
caves that are geographically separated from each other. There can be no question 
whatever, in its case, about an accidental carrying into caves, for if it enter caves 
by accident it must be continually meeting accidents through a very wide region. 

Third process: This view was first expressed by Garman (Science, Oct. 28, 1892, 
p. 240): 

The originals of the cave species [non-aquatic, especially] of Kentucky were probably already 
adjusted to a life in the earth before the caves were formed * *. 

The writer ' independently came to the same conclusion. 

This theory makes the cave simply the collecting ground of animals adapted 
to a cave existence, and leaves the origin of this adaptation an open question. Gar- 
man imagined that the animals become adjusted to cave existence in crevices of 
rocks. Since these crevices are but caves on a small scale, his suggestion simply 
tends to account for the aggregation of the animals found in the caves of Kentucky, 
not for their becoming cave animals in the first instance. 

If at this point we might call mutation to our aid, we would have a satisfactory 
explanation. If mutants arose among any species of animals adapted to cave 
existence, they would find their way into caves or cre\dces if such existed. What 
would happen if there were none need not concern us. But while mutation might 
account for the positive adaptive modifications in cave animals, it does not account 
for the negative or degenerative changes, and the more venerable theory of special 
creation is of equal potency. 

Fourth process: It is certain that in some cases cave animals have developed 
concomitantly with the caves. It seems quite possible that in more cases than 
we have thought the adaptation of an animal to a very complex environment can 
only be explained as the result of concomitant development of environment and 

' In answer to the statement made by Eigenmann, Krause [Promethius, No. 457, p. 652, 1898] said: "Nicht 
weil sie in dunklen Hohlen leben, seien ihre Gesichtsorgane verkiimmert, nicht die Entziehung des Lichtes habe 
diese Organe zuriickgehen lassen, sondern umgekehrt, weil sie sich schon in der Obenvelt dem Leben ohne Licht 
angepasst hatten, waren sic wohl vorbereitet gewesen, in den Hohlen, von volliger Dunkelheit umgebcn, so — 
glanzend, konnte man beinahe sagen — zu reiissiren. * * * Nun, wer's glaubt, mag ja auch bei dera Glauben 
selig werdcn konnen, dass die Hohlen gleichsarazum Turamelplatzund Elysium der Blinden aller Thierklassen 
erschaffen seien. Wir haben diese Sirenenklange aus dem mystischen Dunkel der Gegner des Lichtes und der 
Entwickelungslehre schon ofters gehort; sie stehen in Harmonie rait den iramer starker hevortretenden Bestre- 
bungen, dem Lamarckismus, Darwinismus und selbst dem Weismannismus ein Bein ru stellen." This quotation 
is possibly sufficient to indicate the general lenor of the rest of his article. 


animal. Certain parasitic insects are in the habit of boring through the hard mud 
walls of the nests of mud wasps to deposit their eggs. It seems difficult to explain 
the origin of so complex a habit and of the organ sufficient to pierce the hard wall. 
A mutation to account for it seems inconceivable. It is, however, quite possible 
that the hard wall is a partial adaptation against these very enemies, and that the 
habit of building heavier and heavier walls, and the development of more and 
more efficient organs for piercing them were developed as armor plate and armor- 
piercing shells are interrelated developments. 

From the hills about Horse Cave, Kentucky, one sees valleys about 250 feet 
deep stretching out in four directions. Of the river that is responsible for them 
nothing is to be seen. It is 185 feet beneath the bottom of the valley at the town 
of Horse Cave. The hills are capped with over 70 feet of sandstone. The river 
has had a continuous existence from the time it formed the valley in the sand- 
stone capping, through its later history when it continued the process of valley 
formation in the limestone underlying it, and later still when it hollowed out its 
underground channel in the limestone. There is nowhere any indication that 
there has been a cataclysm in the history of the river. It lies south of the glacial 
area. What is true of the river may be true of the inhabitants still within it. There 
is no reason to think that the ancestors of the blind fishes may not have lived in 
the stream when it flowed over the sandstone capping the hills.' Some fishes of 
any stream stay in the light, others in the shade, others under rocks. The ances- 
tors of the blind fishes probably lived in the shade under rocks and became ad- 
justed to the dark or dusk, existing there long before the caves were formed. When 
placed in open pools Amblyopsis still has that habit. What more natural than 
that this fish should descend farther and farther with the river after it began its 
subterranean course — not suddenly, but gradually ? At first only part of its water 
found its way underground ; but when all its water could flow beneath the surface 
under normal conditions, a part flowed above ground after every freshet, just as 
the water of Lost River of Indiana does at present. It could not sink beneath 
the ground at all until Greene River, into which it empties, had cut a considerable 
distance beneath the surface of the limestone, and thus gave the water in the lime- 
stone rifts a chance to flow out and be replaced with fresh water from the river 
above. As the stream sank beneath the surface, naturally those fishes depending 
on light for food and courtship left it, and only those either negatively heliotropic 
or positively stereotropic remained. 

The blind aquatic fauna looked at from this standpoint is not a new acquisi- 
tion of the present cave stream, but a relict of the fauna of the river when it still 
flowed above ground. The cave and its fauna have developed hand in hand. 
The presence of the cave fishes and other aquatic cave dwellers do not so much 
need explanation (they were present long ago) as does the absence of all of the 
other forms that must have been present when the stream flowed in its epigean 
valley. The prime requisite for a candidate for underground existence is a nega- 
tive reaction to light, or positive stereotropism, or both. 

It must also be evident that a fish depending on its sight to procure its food 
can never become a cave form. Sun-fishes, which are annually carried into the 
present fully developed caves, belong to this class of fishes. They are always 

* Shaler, 1875, considers that during the glacial epoch the conditions in the caves of Kentucky were such 
that the present fauna could not have existed there. 


poor when found in the caves, and will never be able to establish themselves in 
them. On the other hand, there are no reasons why fishes detecting their prey 
either by smell or touch should not be capable of colonizing caves. The cat-fishes 
and Amblyopsidae belong to the latter class. It is surprising that more cat-fishes 
have not established themselves in caves. Among the Amblyopsidae, even those 
with functional eyes depend on touch and vibrations for their food. Chologastcr 
has well-developed tactile organs and poor eyes. It is found chiefly at the mouths 
of underground streams, but also in the underground streams themselves. The 
tactile organs are not different in kind from those of other fishes, and their high 
development is not more marked than their development in the barbels of the cat- 
fishes. The characters which distinguish Chologaster as a fish capable of secur- 
ing its food in the dark are emphasized in Typhlichthys, and the tactile organs 
are still more highly developed in Ainblyopsis. The eyes of the last two genera 
are so degenerate that it is needless in this connection to speak of degrees of degen- 
eration. On account of the structure of their eyes and their loss of protective 
pigment, they are incapable of existence in open waters. With the partial and 
total adaptation to underground existence in the Amblyopsida; and their negative 
reaction to light, it is scarcely possible that for this family the idea of accidental 
colonization can be entertained for a moment. Their structure is not as much 
due to their habitat as their habitat is due to their structure and habit. 

Typhlogobius lives in the holes of shrimps, under rocks, on the coast of southern 
California. It is a living example of the origin of blind forms in dark places 
remote from caves. Here again the "accidental" idea is preposterous, since no 
fish could by accident be carried into the devious windings of the burrows they 
inhabit. Moreover, a number of related species of gobies occur in the neighbor- 
hood. They live ordinarily in the open, but always retreat into the burrows of 
crustaceans when disturbed. The origin of the blind species by the gradual 
change from an occasional burrow seeker to a permanent dweller in the dark, 
and the consequent degeneration of the eye, is evident here at once. Among 
insects the same process and the same results are noted. We have everywhere the 
connection of diurnal species with nocturnal, dark-loving, and blind forms, a tran- 
sition, the result of habit entered into with intent, but no evidence of such a con- 
nection as the result of accident; also numerous instances of daylight species 
being swept into caves, but no instance of one establishing itself there. 

Attention has been called to the difference in the time of origin of the aquatic 
and non-aquatic cave-dwellers. The latter are later immigrants. They neces- 
sarily arrived after some channels had been cleared of water through the stream 
burrowing into still lower channels. The non-aquatic forms are derived, in 
part at least, by migrations from the twihght forms that may have developed with 
the twilight region, and in part they are active immigrants of stereotropic 
or negatively phototropic forms like the Spelerpes. Some of them, like the myrio- 
pods, may even be accidentally brought in with their food and habitat,* but even 
here the active voluntary immigration is, at least, as probable as the accidental one. 

Species widely distributed over a continuous environment may have become 
distributed from one center of development. The same may be said of the species 
found in distant, discontinuous environments where it can be shown that the dis- 
continuity is of recent origin. The same can not be said of species distributed in 

' Decaying logs have been carried into and are found in various parts of the Mitchell caves. 


isolated elements of a discontinuous environment that can not, in the nature of the 
case, at any time have formed parts of a continuous environment. 

Amblyopsis is found on both sides of the Ohio River. The caves of the two 
sides have certainly never formed part of the same complex. It is possible, though 
scarcely probable, that the caves south of the Ohio, inhabited by Typhlichthys at 
one time, formed a continuous environment. It seems evident that Amblyopsis 
could not have migrated from the caves south of the Ohio to those north of the 
Ohio. The different colonies probably had similar but independent histories. 
The cave salamander, Spelerpes maculicauda, is widely distributed in the Missis- 
sippi Valley. It enters caves wherever they are found within its area of distribu- 
tion. It is becoming adapted to a cave existence in widely isolated places. What 
is at present taking place with Spelerpes may have taken place with Amblyopsis, 
except that Spelerpes found its caves ready made, while Amblyopsis was present 
during their making. 

The ancestry of the Amblyopsidae we may assume to have had a tendency to 
seek dark places, wherever found, and incipient blind forms would thus arise 
over their entire distribution. Certainly the fearless, conspicuous blind fishes, 
as at present developed, would have no chance of surviving in the open water. 
Their wide dispersal after their present characters had been assumed would be 
out of the question entirely, except through subterranean waters. The same 
would not be true of the incipient cave forms when they had reached the stage 
at present found in Chologasier. This genus has the habit of hiding under- 
neath objects in the darker sides of an aquarium. These dark-seeking crea- 
tures would be especially well fitted to become distributed in caves throughout 
their habitat. S. Carman's able argument for the single origin and dispersal 
of the blind fishes through epigean waters was based on the supposition that 
the cis-Mississippi and trans-Mississippi forms were identical. The differences 
between these species are such as to warrant the inference, not only that they 
have been independently segregated, but that they are descended from different 
genera. The external differences between these species are insignificant, but 
this is to be expected in an environment where all the elements that make for 
external color markings are lacking. The similarity between Typhlichthys and 
Amblyopsis is so great that the former has been considered to be the young of the 
latter. For reasons that will be fully set forth there is every probability that the 
Cuban blind fishes developed with the caves which they inhabit. 

In conclusion it may be said : 

(i) That the cave fauna is in large part the result of the formation of the caves 
themselves, that environment and habitat developed pari passu. 

(2) That to this original fauna have been added and are being added species 
(such as Spelerpes maculicauda) which, because they are negatively heliotropic or 
positively stereotropic, are gradually becoming adapted to the deeper and deeper 
recesses of caves. 

(3) That to the fauna of the larger caves may also have been added animals 
which had become adjusted to cave existence in crevices, under banks or rocks, etc., 
that is, in small caves. 

(4) That accident has played little or no part in developing the cave fauna. 



Cave existence, reduced to its simplest terms, is the securing of food and the 
meeting of mates in absolute darkness. Food is so scarce that no large preda- 
ceous animals have taken up their abode in caves, hence the largest cave animals, 
such as the cave fishes, have no enemies aside from parasites and disease germs. 
Of the cave fishes Chologaster reaches a length of but 62 mm. ; Typhlichthys, 
55 mm.; Troglichlkys, 55 mm.; Amblyopsis, 135 mm.; Lticifuga, 104 mm., and 
Stygicola, 152 mm. All are insignificant in size. 

The density of the population of any cave, other things equal, is inversely pro- 
portional to the size of the cave. No footl is generated in caves by the growth of 
plants. Directly or indirectly all food consumed in a cave must be imported. 
It may come in through various openings ; usually there are only one or two open- 
ings of any consequence: (a) the "entrance" in a dry cave, {b) the entrance and 
point of inflow of the stream in a wet cave. That cave is best supplied with 
food per square yard which has the smallest area over which the limited supply 
must be distributed. There is, of course, a great difference in the amount of food 
carried in through different openings. An entrance sloping upward naturally will 
not admit as much decaying vegetation as one sloping downward. A narrow 
crack through which water may enter a cave will not admit as much as a large 
opening, through which in times of flood the water may carry tree trunks. These 
matters equalized, I may repeat that that cave is best supplied with food per square 
yard which has the smallest area over which the limited supply entering a given 
opening must be distributed. The density of the fauna varies as the amount of 
food, and hence, other things equal, inversely as the size of the cave. 


Desired lines of research are the relation of the abundance of the cave fauna 
to the age of the particular cave and the comparative degree of adjustment of 
the animals to caves of different ages. We have in North America a series of 
caves reaching from Howe's and other northern caves in the glaciated region to 
the Ohio Valley caves near the edge of glaciation, and the caves of Texas and 
Cuba never affected by glaciation.' 

Howe's Cave in central New York is exceedingly poor in animals, the Texas 
caves are as correspondingly rich, but no detailed comparison has been made. It 
is also known that the Ohio Valley cave salamander, Spelcrpes maculicauda, has 
well-developed eyes, that the Missouri salamander, Typhlotriton, has degenerate 
eyes, and that the Texas salamander, Typhlomolge, has very much more degenerate 
eyes. The degree of degeneration seems here coordinate with the age of the cave. 
Also that the Missouri blind fish has more degenerate eyes than those of the Ohio 
Valley. In a general way the older caves appear to have more intimately adapted 
or more profoundly modified forms than the newer. But here again we lack 
entirely a detailed study. 

' Shaler, 1875, estimates the age of the Kentucky caves at between 750,000 and 2,000,000 years. He further 
maintains that, during the glacial epoch Kentucky was populated by an Arctic fauna and that the cave fauna was 
not derived from this, but from the present fauna of Kentucky, "since the glacial period." I agree with him that 
the present cave fauna of Indiana and Kentucky was derived from or developed concomitantly with the present 
epigean fauna, but am in doubt about the nature of the fauna during the glacial period. 



The struggle for existence with the biological environment as the result of the 
geometric rate of increase tends to divergence in habit and form. It does this by 
preserving variants whenever such possess a character diverging sufficiently in 
amount to give the variant a personal advantage over his fellows — always provided 
the divergent character is transmissible. 

Whether we call the diverging individuals variants in the old sense, or mutants 
in the new, it is to the selection of those among them best adapted to utilize the 
foods of various sorts, to occupy localities of various kinds, to escape the enemies 
of various sorts, and to leave others similar to them in their place when they die, 
that we owe the specific divergence in structure, shape, color, food habits and 
breeding habits of a given family — ■ say the American Characins. The entire 
process tends to the divergence and multiplicity of species. 

The Characins are a family of fresh-water fishes that, in America, range from 
the border of the United States to some distance south of Buenos Aires. They 
form about one-third of the entire South American fresh-water fauna, and have 
diverged in adaptation to diverse food, diverse habitat, and diverse enemies to fill 
nearly every niche open to fishes. The ends of three of the lines of adaptation 
to different food give us mud-eating forms, with long intestinal tract and no teeth ; 
flesh eaters, with shear-like teeth, that make bathing dangerous to life and that 
cut their way out of nets ; and conical-toothed forms, vnth. sharp, needle-like teeth 
and comparatively huge fangs. Greater diversity could scarcely be imagined, 
and one is led to suspect that some of the forms are over-adapted. In their 
divergence in form they have reached almost every conceivable shape as we shall 
see in a moment. 

The struggle for existence with any unit of physical environment, whether 
there be geometric rate of increase or not, tends to convergence in habit and forrri. 
There is no more striking instance of this than the acceptance of the annual or 
deciduous habit of most of the plants inhabiting the temperate zones with their 
seasonal changes, nor is there a more striking illustration of the struggle with other 
individuals than the diversity of form and habit of various forest plants for ground 
and light space. Records of the simultaneous and similar changes in the form in 
the mass of species of any area during changing physical conditions are not want- 
ing. For instance, Scott says: 

The steps of modernization which may be observed in following out the history of many dif- 
ferent groups of mammals are seen to keep curiously parallel, as may be noticed, for example, in 
the series of skulls figured by Kowalevsky, where we find similar changes occurring in such families 
as the pigs, deer, antelopes, horses, elephants, etc. Indeed, one may speak with propriety of a 
Puerco, or Wasatch, or White River type of skull, which will be found exemplified in widely separate 

On some riffles of the San Juan River of Cuba I found a small fish that is very 
strikingly like other fishes inhabiting similar localities in the eastern United States. 
The former is a goby, a marine form, Philypnus dormitator, which has become 
adjusted to conditions found about the riffles of streams ; the others are darters, 
Hadropterus, belonging to an entirely different family of fresh-water fishes. The 
similarity of various "darters" which hve on the bottom of our streams to various 


gobies and blennies that occupy a similar position along the marine shores has 
repeatedly been noticed. 

In the tropics live many burrowing lizards and snakes. Rhineura, one of the 
lizards, lives and acts like an earthworm, and so like an earthworm has it become 
that only a close inspection reveals its true nature. Even the chickens following 
the plows in Florida and Cuba are said to be taken in by the similarity of some 
of the burrowing lizards to earthworms. 

The Characins again furnish striking illustrations. Diverging among them- 
selves, as has been noted above, they have approached, or paralleled, many mem- 
bers of the diverse families of North American fresh-water fishes. Our shads and 
fresh-water herrings have their counterparts in Elopomorphus, Potamorhina, and 
Psectrogaster ; our salmon are paralleled by Salminus and Calabasis; our min- 
nows are paralleled by Tetragonopterus and its relatives. It will take but a slight 
flight of the imagination to detect the striking similarity of some of the Hydro- 
cyninae to our garpikes; our mullets are duplicated by Prochilodus; our top- 
minnows are mimicked by Nannostomus ; and even our festive darters are dupli- 
cated by the species of Characidium, members of this most remarkable family. 

In a dark cave, all those differences between related species which would strike 
the eye, such as protective coloration, recognition marks, decorations of any sort, 
etc., are absent, and related species tend to look alike. It was not until after 
a detailed examination of many specimens that I could invariably distinguish 
Lucifuga and Stygicola, the Cuban blind fishes, from each other. 

On the surface the specimens of Troglichthys rosce very closely resemble Typh- 
lichthys subterraneus from Mammoth Cave, differing slightly in the proportion and 
in the pectoral and caudal fins. These fins are longer in rosa. It is, however, 
quite evident from a study of their eyes that we have to deal here with a case of 
convergence of two very distinct forms. They have converged because of the 
similarity of their environment and especially owing to the absence of those ele- 
ments in their environment that lead to external protective adaptations. It would 
be difficult to distinguish specimens of similar size of Amblyopsis from either 
subterraneus or rosce were it not that it possesses ventrals. 

The eye of T. subterraneus is surrounded by a very thin layer of tissue repre- 
senting the sclera and choroid. The two layers are not separable. In this re- 
spect it approaches the condition in the epigean, eyed member of the family, Cholo- 
gaster. For other reasons, that need not be given here, it is quite certain that 
Typhlichthys is the descendant of a Chologaster. The intensity of coloration and 
the structure of the eye are the chief points of difference. The eye of roscz is but 
about one-third the diameter of that of subterraneus, measuring 0.06 mm. or there- 
about. It is the most degenerate, as distinguished from undeveloped, vertebrate 
eye. The point of importance in the present instance is the presence of com- 
paratively enormous scleral cartilages.' These have not degenerated in propor- 
tion to the degeneration of the eye and in some cases are several times as long as 
the eye, projecting far beyond it, or are puckered to make their disproportionate 
size fit the vanishing eye. This species is unquestionably descended from a species 
with well-developed scleral cartilages, for it is not conceivable that the sclera as 
found in Chologaster could, by any freak or chance, give rise during degeneration 

' Kohl mistook the nature of these structures, as he did of every other connected with these eyes, except the 
lens and ganglionic cells. ^ 


to scleral cartilages, and if it did they would not develop several sizes too large 
for the eye. At present no known epigean species of the Amblyopsidae possesses 
scleral cartilages. The ancestry of rosa is hence unknown. Amblyopsis has the 
scleral cartilages, and the eye of rosa passed through a condition similar to that 
possessed by Amblyopsis, but the latter species has ventral fins and is hence ruled 
out as a possible ancestor of roscB. The epigean ancestry of Amblyopsis is also 
unknown. The ancestry of Typhlichthys being quite distinct from that of rosce, 
the latter species is referred to a separate genus, Troglichthys. 

Judging from the degree of degeneration of the eye, Troglichthys has hved in 
caves and has done without the use of its eyes longer than any other known 
vertebrate. {Ipnops, being a deep-sea form, is not considered.) 

The species of Typhlichthys differ from each other in only a few inconspicuous 
respects. (See page 53.) 



(i) The possible physical environment of animals is composed of units, each 
of which is distinguished by a combination of conditions peculiar to it. 

(2) A unit may embrace one continuous area. 

(3) A unit may have extended in the past over a continuous area, but may 
now be broken up into separate, though similar, parts between which the migra- 
tion of animals is not possible. 

(4) A unit may always have existed of separate and distinct parts (units of a 
smaller order) which together form a discontinuous unit. 

(5) An animal distributed over a continuous, or parts of a formerly continuous, 
unit may have arisen at a single center of dispersal. 

(6) An animal distributed over a discontinuous unit must have had separate 
places of origin or have originated at a time when the parts of the unit were con- 

(7) Each cave consists of a twiUght section, a fluctuating temperature section, 
and the cave par excellence. 

(8) The environment in the third section is chiefly characterized by (c) the 
absence of light ; (6) the constancy of meteorological conditions between seasons ; 
(c) the absence of food except such as is imported. 

(9) All classes of vertebrates, except birds, have blind members. 

(10) Some cave animals (aquatic) have developed pari passu with the devel- 
opment of an underground stream and are among the few inhabitants remaining 
to the stream of its inhabitants during its epigean period. 

(11) Some cave animals (non-aquatic) have gradually colonized caves after 
their formation. 

(12) Some cave animals became elsewhere adjusted to live in the dark and 
later migrated into caves. 

(13) Accident had little or nothing to do with the colonization of caves. 

(14) Some widely distributed cave species have independently arisen in dif- 
ferent places from a widely distributed epigean species. 

(15) Directly or indirectly all of the food supply of a cave must be imported. 

(16) Smaller caves have a relatively richer fauna, because the food supply is 
more abundant. 

(17) Older caves have a more varied and richer fauna. 

(18) Cave animals tend to converge in their evolution; epigean animals, to 




Dr. J. R. Slonaker has found that the eye of the mole (Scalops aquaticus 
machrinus) lies embedded in the muscle beneath the skin, where it appears as 
an inconspicuous dark spot. It is situated well forward on the side of the snout. 
The eye is degenerate and is no longer capable of functioning in distinct vision. 
The most noticeable changes which have occurred are: 

1. The great reduction in the size of the eye. 

2. The much crowded condition of the retina as a result of the decrease in size 
of the eye as a whole. 

3. The noticeable reduction in the size, or the complete absence, of the aqueous 
and vitreous chambers. 

4. The varied modification of the shape and size of the lens, also the peculiar 
cell structure of the lens. 

All the structures of the normal mammalian eye are present in some form or 
other, (i) The conditions found in the adult and at birth have been studied. Very 
little difiference is seen in these two stages excepting an increase in size. 

The eye muscles and the optic nerve are easily traced back to the skull. At 
birth the nerve preseijts in its course from the eye to the skull a peculiar arrange- 
ment. The course is marked by numerous cells and few or no fibers. At the eye 
there is a rapid change from this cell condition to the fiber condition of the nerve 
tract. The fibers have not apparently grown much beyond the limits of the eye. 
In the adult the fibers can be traced to the skull. 

The eye cleft is very small and of practically the same diameter in both hori- 
zontal and vertical sections through it. It meets the eye at such an angle that it 
is impossible for rays of light, should any enter, to pass through the eye along the 
axis of vision. 

All the elements of the normal retina are present, but, owing to the much 
crowded condition, the ganglion-cell layer is much increased in thickness. 

The lens, which is found in a great variety of shapes and sizes, is composed 
of peculiar cartilage-like cells with well-defined nuclei. It is therefore incapable 
of functioning as a normal lens. 

It is very doubtful, therefore, whether the eye of the mole functions in any 
sense. At best it can do no more than distinguish between light and darkness. 





The cave rat, Neoloma magisier, ranges eastward to southern New York and 
south to Alabama, and is not confined to caves. It lives in "cliffs, caves, and rock 
ledges of the mountains, descending into the lowlands, where limestone caves afford 
it security." 

In White's Cave, near Mammoth Cave, Kentucky, it has its nests near the 
entrance, in the twilight region. In Mammoth Cave I found it in Mammoth Dome, 
and it occurs also farther in, far removed from the twilight area. 

Rhoads (Jour. Cin. Soc. Nat. Hist., xix, No. 2, 55, 1897) says of it: 

Any suspicion of blindness or deficient eyesight, such as is exemplified in some of the lower 
orders of animal life in the cave, can not attach to this mammal. As in all the more strictly noc- 
turnal rodents, the eyes of this species are greatly developed ; nevertheless, they are able to make 
most intelligent use of them in broad daylight, if need be. 

In his "Origin of Species," sixth edition, vol. i, page 171, Darwin says that 
the eyes of Neotoma of Mammoth Cave are "lustrous and of large size; and these 

Fig. I. (d) Eye of Mammoth Cave Rat. {b) Eye of Common Gray Rat. 

animals, as I am informed by Professor Silliman, after having been exposed for 
about a month to a graduated light, acquired a dim perception of objects." The 
cave rat, Neotoma, is still abundant in Mammoth Cave. Its tracks are numerous, 
and in places little paths have been made by the rats where they run backward 
and forward along ledges of rock. Since, however, a track once made in a cave 
remains unchanged by wind or weather, the abundance of rats, as judged by their 
tracks, may be misleading. A number of traps were set in the rotunda. During 
three days one trap was sprung and one had the bait removed. No rats were 
caught in the traps and none were caught alive. The author discovered one rat 
rolling a mouse trap about which was too small for it to enter. When approached 
with a light, the rat turned about and stared at the light. It then ran to a pile of 
rocks, but did not attempt to hide ; instead, the rat ran to one end of the pile, then 
along the top back to where it had stood, then stopped and again stared at the light. 

• The histology of the eye is condensed from Dr. J. B. Slonaker's account, from which figures i and 2 are 
taken. See Proc. Ind. Acad. Sci. for 1898, p. 255, 1899. 



An attempt to catch the rat sent it running back and forth along the ledges of 
rock at the side of the cave. Finally the rat appeared at the ground again, and 
despairing of catching it alive, it was killed. Its eyes scemal to Ix; large and pro- 
truding very much as in the common rat. Without question the rat noticed the 
light. It had no hesitation in running from place to place. Later four of these 
rats were sent by express. Only one arrived alive; one had been partly eaten 
by the others. The living one was quite gentle. It permitted itself to be stroked. 
Occasionally it pushed an object away with a sideward motion of the forefoot. 
If provoked it snapped at the object. During daylight it sat quietly in a nest it 
formed for itself of cotton batting, which it pulled into a fluffy mass. At night it 
frequently moved about in its cage. Turning on an electric light near its face 
always produced a twitching of the eyelids, so there can be no doubt that the 
light was perceived. An object held some distance from the cage on one side or 
another was always perceived, but just how precise its vision was has not been 
determined. Its hearing was acute. 

Fio. 2. Retinas of Ncotoma and Common Gray Rat Compared, 
(a) Mammoth Cave Rat. (6) Common Gray Rat. 

Its eyes were as prominent as those of the gray rat. If there was any difference, 
its eyes were larger in proportion to the size of the body weight than those of the 
gray rat. The lens in both cases was enormously large in proportion to the eye. 
The pupil was capable of very wide dilation. A microscopic comparison of the 
retinas also showed little difference. Bits of retina from corresponding parts of 
the eye of a cave rat and a gray rat were hardened by the same process, sectioned 
the same thickness, and stained alike. The results are given in figures i and 2. 

There is little difference except in the thickness of the retina, that of the cave rat 
being thicker. However, the difference may be due to the differences in the ages 
of the animals, the cave rat being fully grown, the gray rat only half grown. The 
thickness of the retinas are proportionate to the size of the eye. The increased 
thickness is largely due to the larger size of the cells of corresponding layers of the 
retina. For instance, the rods and cones are decidedly longer and larger in the 
cave rat. But with the exceptions given the two retinas are nearly alike. 




The salamanders, of which there are many species in the United States, 
habitually live under rocks, logs, and the bark of decaying trees. These all shun 
the light except during the breeding season. Others habitually live in the water 
and are principally nocturnal in their habits, hiding under the banks, logs, or rocks 
in the water during daylight. The eyes of the cave salamanders of North America, 
of which there are four species, range in their structure from the perfectly normal 
to the most degenerate known among the Batrachia. 

Spelerpes maculicauda (Cope) (plate i, fig. c) is common in the caves of the 
Mississippi Valley. As far as I have been able to determine, its eyes have not 
undergone any degeneration. It is abundant and so nearly allied to Spelerpes 
longicauda Green, an epigean species of very wide distribution, that formerly the 
two were considered identical (plate 2, fig. a). 

Spelerpes stejnegeri Eigenmann (plate i, fig. b) is found in the tvdlight regions 
of the caves of southwestern Missouri. Its eyes are also normal. Other species 
of Spelerpes ' are sometimes found in caves. 

Typhlotriton spelceus Stejneger (plate i, fig. d) is restricted to the western 
caves of the Mississippi Valley. It has so far been found in Marble Cave and 

Fio. :i. (a) Head of Spelerpes maculicauda, 54 mm. long. (6) Head of Typhlotriton spetaus, 54 mm. long, 
(c) Head of Typhlomolge rathbuni, 47.5 mm. long. 

Rockhouse Cave, and smaller caves in the same neighborhood in southwestern 
Missouri. It is found under rocks in and out of the water. This is the most 
interesting form, inasmuch as it is a much more typical cave animal than Spelerpes, 
but has not yet reached the degenerate condition of Typhlomolge. Its eyes are 
apparently normal in the larva, but in the adult have undergone marked degen- 
eration. The eyelids are disappearing and the rods and cones are no longer 
present in the adult. The eyes of this species will be dealt with below. 

Typhlomolge rathbuni Stejneger (plate 2, fig. b) is found in the underground 
streams near San Marcos, Texas. It has been taken from the artesian well at 
San Marcos and a surface well. It has also been noticed in one of the caves 
near that place, Ezel's, in which the underground water can be reached. It is 
said to have come out of some artesian wells south of San Antonio. It is a peren- 
nibranch and spends all of its time in the water. Its remarkably long and slender 
legs are not able to support its body when out of the water. Figure 3 shows 

* BUineatus is frequently found about the caves of Bloomington, Indiana. 



B HBselfkKl^ 

A, Aa. Spelerpes longicauda. 147.5 mm. Carlisle, Pennsylvania. 
B, Bb. Typhlomolge rathbuni. 88 mm. San Marcos, Texas. 


the heads of three cave salamanders of North America. The heads were sub- 
jected to the same treatment to prepare them for photography, and photographs 
were taken under approximately the same magnification. 

In February, 1896, the first recorded specimens of this species were cast up 
from an artesian well about 190 feet deep, bored by the U. S. Fish Commission. 
Other specimens have since been thrown up at the rate of 30 to 50 a year. 

The following notes on the habit of this cave salamander are by the late Pro- 
fessor Norman of the University of Texas. 

Unless disturbed, the salamanders appear at all times either resting, or very 
slowly and cautiously walking along. They move a few steps at a time, wait awhile, 
and go again. They have no particular pose when quiet except that they always 
rest on their 4 feet, holding themselves up from the bottom of the vessel and fre- 
quently retaining the exact position of the legs at the moment the motion is 
arrested. If the vessel contain, for example, watercress, they crawl in among the 
branches and stop as when walking on firm bottom, with the legs in such a posi- 
tion as fits easiest for gliding in among the twigs. 

They are never seen to move faster than a slow, easy walk, except when dis- 
turbed by external stimuli. Then one of three methods of locomotion may follow: 
(i) the walking speed may pass into a grotesque run by long strides and corre- 
sponding winds of the body; or, (2) this passes into a combined movement of legs 
and tail, the last acting as fin; (3) at its greatest speed the legs are laid length- 
wise against the body, and the tail only is used for locomotion. 

The legs are exceedingly slender and weak. If the animal is placed on a table 
out of water, the body falls to the table, and at best the animal may wriggle a few 
inches ; but in water the weight of the salamander is so little that the legs are 
amply strong for its locomotion. Dr. Stejneger lost sight of this point when he 
guessed that the animal used its tail for locomotion and its legs as feelers. He 
stated as follows: 

Viewed in connection with the well-developed, finned swimming-tail, it can be safely assumed 
I that these extraordinarily slender and elongated legs are not used for locomotion, and the convic- 
l" tion is irresistible that in the inky darkness of the subterranean waters they serve the animal as 

The motion in water is, for the most part, slow and cautious, the movement 
of the long legs being apparently calculated to produce the least commotion in the 
water. The motion suggests that of a cat creeping upon its prey, or the elephan- 
tine progression of the snapping turtle. The feet are lifted high in walking, and 
the body is kept from the bottom by the full length of the fore arm and leg. In 
ordinary progression the body slopes from nose to tail, which drags (plate 2, fig. b). 
The method of moving the limbs is as follows : I^eft hand and when this is nearly 
ready to place, or usually when placed, the right foot. When the right foot is 
placed, then the right hand and then the left foot. As the hand of one side is 
not raised till the foot of the same side is placed, the enormous strides of the long- 
legged creature cause it to step on its hand or even beyond. Its natural gait is 
a deliberate progression by means of its feet with three feet usually on the ground. 
Any attempt at great rapidity by this means of locomotion results in a most un- 
dignified and futile wriggle. When going slowly, the head is held sloping upward. 
When walking rapidly, it is held sloping down, so that the snout is near the ground. 


No definite information has been obtained as to their habits in nature. They 
show no reaction to light, either as a response by motion to the direction of the 
rays or to the quantity of light. If kept in a vessel, half of which is dark and the 
other half light, the animal is found about as often in one as the other, and on 
emerging into hght from the dark it indicates in no way an awareness of the 
difference. If in a tangle of plants, as watercress, they are found about the same 
as in any other part of the vessel. 

If they are headed against a current, the flowing water acts as stimulus, urging 
them on. If the current strikes them from behind, they move more rapidly in the 
direction of flow. 

The sense of touch is highly developed. There is, however, no experimental 
evidence that this is confined to any particular region. If the surface of the body 
is touched anywhere except at the blunt truncated snout, the animal responds at 
once by moving away. If the stimulus causes it to swim away, it may go (say 
12 or i6 inches) till it strikes the side of the vessel, after which it soon comes to a 
standstill. If, however, it is struck, say with the flat side of a scalpel handle, [ 
sufficiently hard to move the entire animal even an inch backward, it may not 
react, and this may often be repeated before it reacts by moving away. A possible 
explanation of this fact is that in normal life it is every day striking itself against 
obstacles, especially the sides of the vessel (when in confinement). 

The animal is exceedingly sensitive to any motion of the water. Where one is 
kept in water about an inch deep, with its head near the surface, waves of water 
set going by a gentle puff of the breath act as a sure stimulus. 

But little evidence thus far shows in favor of a sense of smell. All attempts 
at feeding (except one) have been in vain. No attention was given to meat or 
other articles placed near it. Examination of a dead specimen showed chitinous 
remains of such Crustacea as Cyclops. 

If a glass rod or other object is held a little to one side and in front of the ani- 
mal, it will cautiously turn its head in the direction of the rod. If the latter is then 
made to describe an arc about the side of the salamander, the head will follow it 
with a continuous motion, expressive of the greatest caution, as far as it can be 
followed without moving any of the limbs. A sudden jar, produced by tapping 
the rod on the bottom of the aquarium at such a time, causes the salamander to 
jerk its head back and rear back on its limbs as far as it can. The same effect 
is produced if the rod is introduced too rapidly. 

If a piece of crayfish tail is held by pincers in the fingers a short distance in 
front or on one side of the head of the salamander, there is the same cautious motion 
forward till the snout comes in contact with it. There is then a momentary hesi- 
tation, followed by a sudden snap and seizure. 

The salamander may be pulled from side to side by the meat, after it has once 
secured a hold, without causing it to let go. All of its caution is apparently directed 
in approaching the food without disturbance. After it has secured a hold it will 
struggle to maintain it. 




The U. S. Fish Commission, through Dr. B. W. Evermann, sent me four speci- 
mens of this salamander and a number of its eggs. Of these, one adult had been 
received in Washington, April 8, 1896, and three young, of different sizes, March 
I, 1896. A few eggs were laid about March 15, 1896. The late Professor Nor- 
man, of the University of Texas, and Professor Bray, of the same place, secured mc 
an additional number. Later, I visited the caves and the artesian well at San 
Marcos, and have been able to observe the living specimens. The specimens 
sent by Professor Evermann were preserved in alcohol ; those sent by Profes.sor 
Norman had been killed in Perenyi's fiuid. The sections were stained chiefly in 
Biondi-Ehrlich's tricolor mixture. 

The following gives the dimension of the eyes in a number of individuals. 
Professor Norman sent only the heads, so the length of his specimens sent can 
be given only approximately. The sizes (in millimeters) were obtained by compar- 
ing the distance between the eyes, with the same distance in entire sf>ecimens. 

Dimensions 0/ the Eyes of Typklomolge in Millimeters. 

Lenhth of 

between eyes. 




















The eye of Typklomolge is, in many respects, much more degenerate than that 
of its European caverniculous relative, Proteus. In Proteus the six muscles are 
all present; in Typklomolge they have entirely disappeared. In the former all the 
layers normal to the retina are present; in the latter the conditions are much 
simpler. In Proteus the lens is still present and blood-vessels still enter the eye; 
in Typklomolge no trace of the lens could be found, except in one individual, and 
blood-vessels no longer enter the eye. While some of the asymmetry may have 
been caused by reagents, it is evident that there is a great deal of fluctuation in 
the shape of the eye. The eye is irregular-oval in outline as seen from above, but 
the optic nerve enters it at the posterior half of its inner face. The eye increases 
materially in size from the smallest to the largest of specimens examined. This 
increase is not directly proportional to the increase in the length of the animal, 
so the young have relatively larger eyes (fig. 4). 

The eye lies immediately beneath the skin, to which it is attached by a connective 
tissue mass which is horizontally elongate. The axis of the eye makes an acute 
angle with the surface of the skin, the eye being directed outward and forward. 
The dermis over the eye does not differ from that in the neighboring tissues. The 
epidermis, in the largest individual, is perceptibly thinner over the eye, i.e. from 
the continuation of the axis of the eye to the surface of the epidermis. The measure- 
ment, in the largest individual, of the epidermis at a point over the eye and 320 /x 
above and below this point gives the following: thickness over the eye 73 /x, 
320 /Lt above the middle of the eye 96 /a, 320 /x down from the eye 80 /x. 

* See Trans. Am. Microsc. Soc. xxi. p. 49, 1900. 



The same elements are found over the eye that are evident in other regions. 
There is no indication of a past free orbital rim; the dermis and epidermis are 
directly continuous over the eye. There are no eye muscles and no glandular 
structures connected with the eye. It is surrounded on all sides, except where it 
becomes associated with the skin, by loose connective tissue meshes filled with 
fatty tissue, and is bound to the dermis by many fibers running in various directions, 
and among these a few pigment cells are found. 


(a) Largest specimens: Cartilaginous elements are found in the sclera of but 
two eyes. In one individual, 90 mm. long, the left eye possesses a cartilage, while 
there is none in the right eye. It is in this case placed just above the entrance of 


Fig. 4. Outline Sketch of Part of Section of Head of Specimen of Typhlomolge rathbuni, 
po mm. long, showing Position of Eye. 

the optic nerve and measures 96/1 in thickness, 160 /a vertically, and 204 /a antero- 
posteriorly. In all other cases the sclera is a thin, flocculent layer not distinctly 
separable from the layers beneath it. It is thickest about the entrance of the optic 
nerve. Over the front of the eye there are a few denser strands, which may repre- 
sent the remains of the cornea. Over the sides of the eye of the largest individual 
the sclera measures from 4 /i to nothing. About the entrance of the optic nerve it 
attains a thickness of 14 /x, and contains many flat nuclei with a length up to 17 fi. 

The choroid reaches a thickness of 20 /j. near the entrance of the optic nerve, 
and dwindles regularly from this point to the distal face of the eye. Blood-vessels 
are found in it next to the pigmented epithelium of the eye. Otherwise it is a 
mass of pigment interlarded with streaks of colorless tissue containing nuclei. 
Over the front of the eye, next to the epithelium, there are a number of colorless 
cells with large, granular nuclei. 

(b) Essentially the same conditions exist in younger specimens, but the parts are 
relatively thinner. The ophthalmic artery, extending approximately parallel 
with the optic nerve during its distal course, is sometimes surrounded by pigment. 




The pigment layer is a thin, compact hiycr, densely pigmented. In an indi- 
vidual 30 mm. long it is about S/t in thickness. As there arc no rods and cones, 
the inner surface of this layer is similar to the outer, that is, the cells form a pave- 

-i. 1 

Fig. 5. (a) Right Eye of Specimen of Typhlomolge 30 mm. long. (6) Exit of Optic Nerve of Same, (c) Iris of Left 
Eye of Same Specimen, (rf) Upper Half of Iris of Right Eye of Specimen of Typhlomolge 70 mm. long. 

ment epithelium. In places, however, processes of the cells extend in among the 

cells of the nuclear layers, for a distance of 40/". in some cases (fig. 50), to the 

inner reticular layer. In the individuals 70 to 90 

mm. long, the pigment epithelium reaches 16 /x in 

thickness. The only indication of a lens was found 

in the eye of a specimen 72 mm. long. In this a 

small lenticular group of cells lay in the opening of 

the pupil. It measured 24X 40/A (fig. 6). 

i I-- 


Marked changes take place in the iris from the 
smallest to the largest individuals examined, so that 
these must be dealt with seriatim. 

The smallest individual is 30 mm. long (fig. 5 a 
and c). On the left side the pupil measures 22 /a in 
diameter ; the distance from the margin of the pupil 
to the ora serrata measures appro.ximately 100 fi. 
The epithelial part of this iris consists of an outer 
layer of dense pigment considerably (14/^) thicker 
than the pigment epithelium of the rest of the eye. 
At the pupil this pigment appears rolled into the 
inner surface of the iris, where it is continuous with the inner layer of cells, which 
consists of a layer of ordinary pigmentless epithelium 6 fi thick, with the nuclei 
elongate and placed obliquely, and 24 /x in length. A few of these ordinarily pig- 
mentless cells show pigment. There is a distinct thickening of the iris at the 
margin of the pupil. The pigment cells lying on the inner face of this region are 
much less densely pigmented than those of the outer layer, and their nuclei are quite 
evident. The pupil is closed with colorless cells belonging to the choroid (fig. 7 a). 

Fig. 6. Lens of Specimen 73 mm. long. 



Very marked changes have been brought about in the specimen 70 mm. long. 
The pupil is now an oblique channel and the lower margin of the iris overlaps 
the upper margin. On the left it is more nearly as in the younger stages, but 
wider (48 fi). The free margin of the iris now reaches the enormous thickness of 

Fio. 7. (a) Right Eye of Specimen 70 mm. long. (6) Right Eye of Specimen 90 mm. long. 

56 fi to 80 /x. The pigmented epithelium has rolled in more, so that the elongated 
nuclei, free from pigment, are crowded together in the region of the ora serrata. 
The pupil is filled in part with pigment, evidently of choroidal origin (fig. 7 a). 

In the right eye of the specimen 90 mm. long the choroidal pigment has forced 
its way into the interior of the eye and forms a conical-shaped mass like a plug in 



the iris and extends into the depth of the vitreous cavity. Apparently on the 
external half of the iris the pigmented layer has become rolled in and folded upon 
itself in the interior of the eye, giving rise to a pigment mass over loo fj, thick. No 
such mass is present in the left eye. The pigment on the inner or upper half of 
the iris is as in the younger stages. The choroidal pigment entering the eye is in 
solid, vermiform strands (fig. 7 b). 


The retina of Typhlomolge is much simpler than that of Proteus. In the latter 
all the layers typical of the perfect retina are still distinguishable. In the former 
the outer reticular layer has entirely disappeared, and the layers between the rods 
and cones and the inner reticular layer form a mass of cells that are homogeneous 
as far as ordinary histological methods permit one to determine. There are no- 
where the shghtest evidences of any rods or cones, either in the largest or smallest 
individual. The nuclei of the outer nuclear, the horizontal, and inner nuclear 
layers are alike. Miillerian fiber-nuclei have not been disringuished as such. 
This layer consists of about five series of nuclei and measures 44 fi in thickness in 
the smallest (30 mm.), and 48 fj. in the largest (90 mm.), specimen; it is between 
32 and 48 /A in the specimen 70 mm. long. 

The inner reticular layer is thin, but well defined. It is 6 /a thick in the smallest 
specimen and 16 fi in thespecimen 70 mm. long. In section the ganglionic layer forms 
a U-shaped mass of cells. In the larger specimens it is about 60 n thick and made 
up of from five to seven series of cells. The vitreous cavity is a widely flaring, 
trumpet-shaped structure, with its pointed end reaching to near the center of 
the eye (fig. 7 c). In the older specimens it is filled by fibers and cellular tissue, 
apparently continuous with the choroid ingrowth from the pupil (fig. 7 b). 

The optic nerve is 17 /u, in diameter in 
the 30 mm. specimen. In the largest spec- 
imen it is 24 ft thick without its sheaths. 
At its passage through the pigmented 
layer of the retina it is contracted to a 
width of but 14 /A. Within this layer it 
expands to 28 fi. After passing directly 
through the ganglionic layer it is distrib- 
uted to the cells of this layer, some of the 
fibers being bent at an acute angle to reach 
the cells near the entrance of the nerve 
into this layer. A large number of iso- 
lated pigment granules are found associated 
with the nuclei of the optic nerve within 
the eye from its entrance to the gan- 
glionic layer. There is no sheath of pig- 
ment such as that found in Typhlogobius. Pigment cells are also occasionally 
present in the very center of the eye (fig. 7 a 2), and are presumably associated 
with the optic nerve. The sheath of the optic nerve consists of a direct continua- 
tion of the choroid layer, which is for a shorter distance pigmented, and of a 
continuation of the sclera (fig. 8). 

Blood-vessels do not enter the eye with the nerve, and none were with cer- 
tainty detected except in the largest individual, where they are closely associated 
with the choroidal mass of tissue that has grown into the eye through the pupil. 

Fic. 8. Exit of Optic Nerve of Eye shown in 6«. 7 t. 



A single specimen of a salamander was discovered in Rockhouse Cave, Barrie 
County, Missouri, by Mr. F. A. Sampson in July, 1891. The specimen was 
described by Dr. Stejneger (Proc. U. S. Nat. Mus., vol. xv, p. 115), as TypMotriton 
spelceus. His diagnosis reads as follows : 

Vertebrae opistoccelous ; parasphenoid teeth ; vomerine teeth ; eyes concealed under the con- 
tinuous skin of the head; tongue attached in front and along the median line, free laterally and 
posteriorly; maxillar and mandibular teeth small and numerous; vomerine teeth in 2 strongly 
curved series; parasphenoid patches separate; nostrils very small; toes 5; 16 costal grooves, 
or 18 if counting the axillary and groin grooves; tail slightly compressed, not finned; toes nearly 
half- webbed; vomerine teeth in two V-shaped series with the curvatures directed forward; gular 
fold strong, very concave anteriorly; color uniformly pale. 

He further wrote, before he discovered Typhlomolge in the underground streams 
of Texas: 

Although many of our salamanders are known to inhabit caves, this seems to be the only 
one, so far discovered, which, like some of the other animals exclusively living in caves, has become 
blind or nearly so. 

A preliminary note by Eigenmann and Denny (Proc. Ind. Acad. Sci. for 1898, 
p. 252, 1899) completes the list of papers dealing with this species. 

In the spring of 1897, I visited Rockhouse Cave and secured a number of 
larvae, which Dr. Stejneger pronounced the larvae of Typhlotriion. Later Mr. 
E. A. Schultze informed me that he had seen this salamander in the underground 
passage leading to Blondi's Throne Room in Marble Cave, Stone County, 
Missouri. In September of 1898, I visited this cave and secured 4 adults and 
3 larvae of Typhlotriton. A large number of the larvae were obtained from Rock- 
house Cave a few days later. Those from the latter cave were found under loose 
stones and gravel in the rivulet at the mouth of the cave. They had been exposed 
to the light. It is scarcely supposable that those from Marble Cave had ever 
been subjected to light. In the caves both larvae and adults are found under 
stones, the old ones in and out of the water. Occasionally one is seen lying on the 
bottom of a pool. 

In the aquarium the larvae creep into or under anything available ; a glass tube 
serves as a "hiding" place. The rubber tube admitting water to the aquarium is 
sometimes occupied by several during a temporary cessation of the flow of water. 
A wire screen sloping from the bottom of the aquarium formed the most popular 
collecting place for the larvae. They collected beneath this, though it offered no 
protection from the light. From these observations it seems probable that stereo- 
tropism rather than negative heliotropism accounts for the presence of this species 
in the caves, and that this reaction has been retained after the long stay of the 
species in caves necessary to account for the changes in its eyes. 

The eyes of the larvae when examined from the surface appear perfectly normal, 
but they are little used in distinguishing objects. When hungry they will strike 
at a stick held in the hand as they would at food. A stick lying undisturbed at 
the bottom of the aquarium is not molested. They strike at a worm when 
touched by it, or when it approaches close enough for its motion to be perceived. 

' By Carl H. Eigenmann and Winfield Augustus Denny. See Biol. Bull. II. p. 33, 1900. 



In the larvas up to 90 mm. long the skin passes over the eye without forming a 
free orbital rim and the eye docs not ])rotrude beyond the general contour of the 
head. In the adult from 97 mm. on, the eye forms a beadlike projection. There 
are in the adult distinct lids. These are closed over the eye, covering it entirely, 
the slit being much too small for the eye. The lower lid is free from pigment, but 


Fig. 9. (a) Diagrammatic Representation of Eye of TyphtolriUm drawn to scale. 
(&) Vertical Section through Cornea and Lids of Adult. 

the upper lid, which closes over the lower, is as thickly pigmented as any other 
part of the body. 

Stejneger says of the eyes that they are "small, only slightly raised, and covered 
by the continuous skin of the head, vrith only a shallow groove to indicate the open- 
ing between the lids, the underlying eyes visible as two ill-defined dusky spots." 

In sections the lids are seen to overlap one another some distance, forming an 
obscure, free orbital rim. Figure 9 6 is a median section of the lids and corneal 
epithelium of an eye 0.954 mm. in diameter, taken from an adult specimen 106 mm. 
in length. In this section the upper lid overlaps the lower lid 0.216 mm., or more 
than one-fifth the diameter of the eye. Passing from the median section toward 
the corners of the eye, the lower lid unites with the underlying tissue first. When 
observed from the top, the upper lid covers the eye entirely. The orbital slit is o. 1 7 
mm. in length. The conjunctival pocket extends some distance forward and back- 
ward beyond the slit. The eye increases in size but little from the larval to the 
adult stage and its growth is not proportional to the growth in length of the ani- 
mal. (See comparative measurements of the eyes at the close of the chapter.) 

The following is a series of measurements (in millimeters) on the larvae of 
Typhlotriton : 


I.en^th of 

Sin of pupil. 

Lenctb of eye. 

Length from 

optic nenre 

to front of 



Rockhouse Cave 

Rockhouse Cave 

Marble Cave 



I. SO 



Sections of the adult and larva from Marble Cave were made in the usual 
manner. The six normal eye muscles are present in Typhlotriton. The m. 
recti form a sheath about the optic nerve in its distal part and spread out from it 



near the" eye. In the adult the sclera is a layer of uniform thickness except in the 
region of the entrance of the optic nerve. It is not usually separated from the 
adjoining parts of the eye, but in places is retracted a short distance from the 
choroid coat by the action of reagents. It is for the most part fibrous, with few 
compressed nuclei, and varies from iS/a to 40 /x. in thickness. In the larva a narrow 
cartilaginous band surrounds all but the ventral wall of the eye. In a specimen 
35 mm. long the width of the band is about 30 fi, its thickness 16 /u,. In three adult 
specimens the sclera of only one had any traces of cartilage. In the right eye of 
the adult specimen 103 mm. long a cartilage about 36 /x. thick, 60 /* wide, and not 
more than 40 /* long is found on the upper face of the eye. The absence of this 
cartilage in the adult has probably no connection with the degeneration of the 
eye. Its presence is probably a larval characteristic which disappears as the gills 
disappear during the metamorphosis. 




Fig. 10. (tf) Section of Retina, exclusive of Pigment Cells, of Larva 35 mm. long. (6) Tangential Section through Rods and Cones about on 
Level with Innermost Extent of Pigment (seen on Right) showing Relative Sizes and Abundance of Rods and Cones, (c) Section of Ret- 
ina of Larva 48 mm. long, (d) Section of Retina of Larva go mm. long, (c) Tangential Section showing Rods and Cones at about 
Inner LimitofPigment(seenon Left). (/) Section of Retina of Adult 106 mm. long, (g) Tangential Section at about Inner Limit of 
Pigment. (A) Section of Retina of Adult 97 mm. long. 

The average thickness of the cornea is 40 ju,. In the adult it is covered by a 
layer of stratified epithelium, 25 /* in thickness, consisting of three rows of cells. 
The cells of the inner row are columnar in shape, those of the middle row rounded, 
and those of the outer row very much flattened and elongated (fig. 9 b). 

In the adult the choroid coat is usually separated from the pigment layer, but 
adheres closely to the sclera. In general it is thicker at the back part of the eye, 
and quite decidedly so at the entrance of the optic nerve. The lens is normal. 
Its size is given in the table on page 40. 

The layers of the retina are well developed in the larva. The retina of the 
larva differs from that of an Anihlystoma larva in the greater thickness of its gangli- 
onic layer. This layer is, in the young larva of Typhlotriton, composed of 5 or 
6 layers of cells. This thickness may in part be an artifact, since the retinae 
examined are shrunken away from the pigment epithelium and the ganglionic layer 


is in contact witli the lens. In the larva 90 mm. long this layer has been reduced 
to not more than 3 series of cells. Aside from the differences noted above, the 
eye of the larval Typhlotriton is apparently normal in all of its histological details. 
The relative thickness in the different sizes of the larvas may be gathered from 
figures 10 a tod and from the comparative table at the end of this chapter. 

Figures 10 a to/ are drawn with the same magnification and show the relative 
thickness of the different layers in the retina; of the larvae of different sizes and of 
the adult. The adult retina is reduced in thickness by the absence of the rods and 
cones and the (partial ?) atrophy of the outer reticular layer and by the thinning of 
the ganglionic layer. The ganglionic layer in the adult contains from two to five 
rows of cells. In this respect, the adult approaches the condition found in the 
Amblystoma more than the young does. The inner reticular layer is comparatively 
thick, that of the young being thicker than that of the adult. 

In the adult the inner nuclear layer is continuous with the outer nuclear layer. 
(See fig. 10/.) 

The inner nuclear layer consists of about 7 series of cells in the smallest larva 
and of 4 to 7 in the largest. The cells in the preparations available can not be 
separated into bipolar and spongioblastic layers, nor are the horizontal cell layers 
distinguishable. The outer reticular layer is well differentiated, but quite thin in 
the larva;, and is irregular in outline, adapting itself to the overlying nuclei which 
encroach on its outlines. In the adult this layer is indistinguishable by the same 
methods that make it conspicuous in the larva. In places there appeared an open 
space where the outer reticular layer should be (fig. 10 A 4), but none of its structure 
remains. It is fair to suppose that the fibers forming this layer are resorbed during 
the metamorphosis. This layer seems to be the very first obliterated by the pro- 
cesses of degeneration both ontogenetic and phylogenetic in this as in other verte- 
brates with a degenerating eye. 

The greatest change during and shortly after metamorphosis takes place in the 
layer of the rods and cones. In the larva 35 mm. long, from the mouth of Rock- 
house Cave, the rods reach an extreme length of 50 /x. The relative sizes and 
number of these as compared vrith the much smaller cones may be gathered from 
figure 12. 

In the larva 90 mm. long the outer segments of the rods are much shorter and 
stain less conspicuously than in the younger. The nuclei of the outer nuclear layer 
are distinctly in 2 layers, whereas in the younger 
specimen they are in 3 less regular layers. The cones 1. S 

are correspondingly fainter than in the young. It ^ ^k 
is surprising that whereas in the larva 90 mm. long 

we find the rods and cones well developed, they have p,^ „ <„) o„ivCon. found in Eyes of Adui«. 
greatly degenerated or practically disappeared in the &Vte,of CeTi°?n ou^.e?Nuc'ie«'£T41^' 

adult only a few mm. longer. In an adult specimen 

97 mm. long the rods have retained their normal shape and position, but no 
differentiation into inner and outer segments was detected. In longer ones most 
of the nuclei of the outer series have become rounded at both ends. But one cone 
was found in eyes of the adult over 100 mm. long. It is shown in figure no. In 
an adult specimen 103 mm. long filmy rods are still evident. They appear as 
conical spaces above the nuclei free from pigment rather than as possessing any 



demonstrable structure. Just at the margin of the place where the pigment has 
been torn from the retina one of these is drawn out to a great length. The pigment 
in this individual extends in places down between the nuclei of the cones. This 
latter condition appears in a very exaggerated form in the eye of Typhlotnolge. 
In tangential section this condition and the filmy rods give rise to the appearance 
represented in figure lo g. 

Distinct signs of ontogenetic degeneration are also seen in other parts of the 
retina. For instance, many nuclei of the inner series of the outer nuclear layer are 
shriveled. In some eyes the ganglionic nuclei have for the greater part lost their 
granular structure and show a homogeneous pasty condition, only a few cells 
with granular nuclei being present (fig. lo/). The same is true in large part of 
the inner nuclei of the inner nuclear layer. This condition of the ganglionic nuclei 
is not entirely confined to the adult but is also found in the larva. 

Some of the modifications in the shapes of the outer nuclei in the adult are 
shown in the figures. In figure 1 1 & the upper part of the nucleus is very much elon- 
gated. This form is of frequent occurrence. In figure 1 1 c is shown the common 
form where the nuclei are simple elliptical bodies, which give no evidence what- 
ever of any processes uniting them with the other elements of the retina. The Miil- 
lerian fibers are profusely present and of very large size in both larva and adult. 

In both adult and young the optic nerve enters as a single strand and passes 
entirely through the layers. A heavy mass of pigment is found following the optic 
nerve to within a short distance of the brain. 

Average Measurements of the Eyes of Typhlotriton. 

Length of Specimen. 








Vertical diameter of eye 

From front of lens to back of eye. . . 
Outer nuclear layer with the rods. . . 

Outer reticular layer 

Inner nuclear layer 

Inner reticular layer 

Ganglionic layer 

Pigment layer 

Optic nerve 































(i) The eye lies just beneath the skin. The skin is but httle thinner over the 
eye than elsewhere and shows no structural characters different from those of 
neighboring regions. 

(2) The eye muscles have vanished. 

(3) The lens has vanished and its place has in part become filled by an ingrowth 
of choroidal tissue containing pigment. 

(4) The vitreal body is very small, if present at all. The vitreal cavity is a 
funnel or trumpet-shaped space. 

(5) The pigmented layer of the retina is a pavement epithelium with indistinct 
cell boundaries and with occasional pigmented processes extending into or through 
the nuclear layers. 


(6) Rods and cones are not found. 

(7) The outer reticular layer has disappeared. 

(8) The inner and outer nuclear layers form one layer of cells indistinguishable 
from each other. 

(9) The inner reticular layer, as usual in degenerate eyes, is relatively well 

(10) The ganglionic layer is well represented and connected with the brain by 
the well-developed optic nerve. 

(11) The epithelial part of the iris is at first simple, with an outer pigmented 
and an inner colorless layer. With age the margins of the iris become folded in- 
ward in such a way that the pigmented layer may be thrown into folds in the interior 
of the eye, while the colorless layer is but little affected. 

(12) Pigment granules, and rarely pigmented cells, are associated in the eye 
with the optic nerve. 

(13) The eye is more degenerate than that of the European Proteus. It is less 
degenerate than that of the North American blind fishes, Amblyopsis, Typhlichthys, 
and Troglichthys, but much more so than that of the species of Chologaster. 


(i) Typhlotriton is an incipient blind salamander living in the caves of south- 
western Missouri. 

(2) It detects its food by the sense of touch without the use of its eyes. 

(3) It is stereotropic. 

(4) Its eyes show the early stages in the steps of degeneration from those of 
salamanders living in the open to those of the degenerate Typhlomolge from the 
caves of Texas. The lids are in process of obliteration, the upper overlapping the 
lower so that the eye is always covered in the adult. The sclera possesses a car- 
tilaginous band in the larval stages but not in the adult. The disappearance of 
the cartilage is probably an incident of the metamorphosis, not of the degeneration 
the eye is undergoing. The lens is normal. The retina is normal in the larva with 
a proportionally thicker ganglionic layer than in the related epigean forms. 

(5) Marked ontogenetic degenerations take place during and shortly after the 
metamorphosis, (a) The outer reticular layer disappears, {b) The rods and 
cones lose their complexity of structure, such as differentiation into inner and outer 
segments, and finally are lost altogether. 



amphisbjEna punctata.' 

Amphisbcma punctata (Bell) is a blind, legless lizard which burrows in the ground. 
It is common in Cuba, to which place it is restricted. How deep it burrows can 
not be stated, but it is often turned up by the plow. The specimens obtained 
ranged from 103 to 245 mm. in length. The head is short, hard and pointed, and 
the tip of the upper jaw projects slightly beyond the tip of the lower. In shape, 
arrangement of the dermal plates, and color of the ventral surface of the body it 
closely resembles an earthworm. The dorsal surface is flesh-color with small 
brown spots. The tail is short and flattened dorso-ventrally. In a specimen 245 
mm. in length, there were 225 annuli on the dorsal side, 202 on the ventral, and 15 
on the tail. In this specimen the tail was one -thirteenth and the head one-thirty- 
fifth the length of the body. 


The lizards were put alive into formalin. They were afterwards put into alco- 
hol. For decalcification, the heads were placed in 5 per cent nitric acid from 20 
to 30 days. A shorter period did not give satisfactory results. Some heads were 
embedded in paraffin and others in paraffin and celloidin. In using the latter 
method the head was embedded in celloidin in the usual manner and hardened in 
chloroform. From chloroform the block was transferred to soft paraffin for 24 
hours and thence to hard paraffin for 24 hours, after which it was embedded in 
paraffin. The best results were obtained from those embedded in parafiin and 
celloidin. Several methods of staining were used ; iron haematoxylin with eosin as 
a counter stain gave the best results. The more modern methods of treating the 
retina with silver could not be applied for lack of fresh specimens. On account 
of the extreme toughness of the cuticle it was impossible to get complete series of 
sections. For comparison the eye of Anolis caroUnensis has been examined. 


The eye of Amphishana appears indistinctly as a small 
black spot beneath the ocular plate (fig. 12). In a specimen 
225 mm. in length, the eye is 352 /x beneath the surface, 420 /it 
'•^ in width, and 360 /x in depth. The conjunctival sac is 116 /t 
in diameter. The conjunctiva is very thin over the cornea, 
Fig. ij. Head of Amphisbana but measurcs 4 tt In thickness over the anterior part of the sac. 

punctata (Bell) showinK Loca- . i . , . i , i j_ ^ 

tion and Relative Size of Eye. 1 he dcrmis and epidcrmis havc the same structure over 
the eye as over the regions near by. This corresponds with 
what Eigenmann ("The Eyes of Rhineura floridana," 1902) found in Rhineura, 
although the eye of Rhineura is a much more degenerate organ than the eye of 
Amphisbcma, but to what extent the eye is degenerated from a more elaborate 
structure can not be stated. Few organs are stationary, and this one is probably 
still in process of reduction. The writer has been unable to obtain the young, 
and there is no means of finding out from the adult whether the eye is degenerat- 

' By Fernandus Payne. See Biol. Bull. Xl. 60, 1906. 




ing at present or not. In each specimen examined the eyes appeared in about 
the same state of degeneration. 

The eye measures 1,224 /x, in circumference and the pupil 104 fi in diameter. 
The uveal part of the iris on each side of the pupil measures 250 fi. The pupil 
and iris occupy 49.3 per cent, or very nearly half, of the entire circumference. 

Harder 's gland is very much larger than the eye. In a cross-section through 
the central part of the eye, the antcro-posterior diameter of the gland is approxi- 
mately three times and the medio-lateral diameter four times the medio-lateral 
diameter of the eye. It is divided into two distinct lobes, the anterior being much 
smaller than the posterior. The gland completely surrounds the eye except over 
the anterior face. Its secretion is poured into the conjunctival sac and from 
thence into the mouth cavity. The large size of the gland in Typhlops led 
Duvernoy to the conclusion that its function was not connected with the eye. 
As its secretion, in Amphishcma, is 
poured into the conjunctival sac and 
thence into the mouth cavity, its 
function must have been, primarily at 
least, connected with the eye. No eye 
muscles are present in Amphisbcena. 
The eye is directed outward and for- 
ward and makes an angle of about 60° 
with a line drawn tangent to the 
dermal plate which covers it. 

Whether the eye is still used as a 
sense-organ is not certain, but since 
the parts are so well developed and 
the eye is not buried very deeply 
beneath the surface, it is probable that 
it is at least susceptible to light. 

The Sclera. — The sclera (scl., fig. 
14) has apparently undergone no 
degeneration whatever. It compares 
favorably with that of Anolis. In 
fact, there is but little difference in its 
structure in the two eyes. At the 
proximal part of the eye, the sclera measures 1 2 /x in thickness, while at the same 
place in Anolis it measures 15 /*. It is continuous over the front of the lens as the 
cornea, which together with the thin wall of the conjunctival sac at this place 
measures 7 /*. Scleral cartilages extend from about the middle of the eye back 
almost to the optic nerve. On each side of the sclera, and forming a part of it, 
are thin irregular layers of pigment in patches. 


The Choroid. — If the blood-vessels in the choroid still persist, the preparations 
do not show them. All that can be seen is a number of densely pigmented cells, 
around and between which are filaments of connective tissue {chr., fig. 14). At the 
entrance of the optic nerve, this layer measures 8 /x in thickness, but gradually 
becomes less forward and vanishes entirely a short distance back of the enlarged end 
of the pigment layer. The pecten, present in Anolis, is not seen in Amphisbcena. 

Fig. 13. Diagram of Eye. showing Parts in their Relation and Dis- 
tance of Eye beneath Surface. 
I, pigment layer; 2, cones; 3, outer nuclear la>'er; 4, outer retic- 
ular layer; 6, inner nuclear layer; 8, inner reticular layer; 9, 
ganglion-cell layer; 10, fiber layer; lens, lens; scl.. sclerotic; 
chr., choroid; cor., cornea; scl. c, scleral cartilage; n. op., 
optic ner\-e; W/. vitreous cavity; con. cav., conjunctival cavity; 
C., outer covering of eye; if., MiUlerian fiber; L., membrana 
limitans externa. 



The Lens. — The lens has retained its natural shape and position {lens, fig. 14). 
It is almost spherical and measures 80 /x in diameter. In most of the sections an 
outer layer of cells extends around the anterior surface of the lens. The interior 
in nearly every case stained as a structureless mass, but in a few sections it appeared 
to be made up of large irregularly shaped cells with small nuclei. If any fibrous 
cells stiU persisted, they did not show. No capsule is present. 

Fio. 14. 

Horizontal Section of Eye showing Different Parts. Retina diagrammatic 
For explanation of letters see fig. 13. 

The Vitreous Body. — The vitreous body {vit., fig. 14) occupies the greater part 
of the eyeball and has certainly undergone but little change. The aqueous cavity 
has entirely disappeared. 

The Iris. — Only the uveal part of the iris remains. It is continuous with the 
pigment epithelium of the retina and has the same structure. In the thickest part 
it measures 68 /a. The cells are similar to those of the pigment layer, except that 
their radial diameter is much greater. The ciliary processes are no longer present. 

The Optic Nerve. — The optic nerve can be traced from the eye, through and 
along the side of Harder's gland. While the nerve could be traced no farther on 
account of an incomplete series of sections, there is no doubt that the connection 
with the brain still exists. The nerve fibers enter the eye in a compact mass, pass 
through the layers of the retina until they reach the nerve fiber layer, where they 



spread out and connect with the nerve cells of the ganglionic layer in the usual 

The Retina. — While the retina has undergone considerable change, all of the 
layers are still present (fig. 15 a). It measures 78 /i in thickness. In Anolis 
about half-way between the anterior and posterior parts of the eye it is 179/1 
in thickness. If the macula lutea is still present, the preparations do not show it. 

The Pigment Layer. — The pigment layer (i, fig. 14), which bounds the retina 
externally, consists of a single stratum of rectangular cells separated by a small 
amount of clear intercellular substance. These cells have large oval nuclei free 
from pigment, almost transparent and with small nucleoli. At the back part of 
the eye, where the pigment layer measures 8 fi, the transverse diameter of the cells 

Fig. 15. (a) Horizontal Section of Retina of ^ffi^/m6<ma ^unrAjto, showing Different Layers. 
(6) Horizontal Section of Retina of A noHs. 

is greater than the radial diameter, but toward the anterior part, where the layer 
becomes thicker, the radial diameter becomes much the greater. The greatest 
thickness of this layer is near the lens, where it measures 68 /a. The outer surface 
of the pigment cells — that which lies ne.xt to the choroid — is smooth and slightly 
convex. The inner surface, on the other hand, is very irregular. The cells at this 
place are very densely laden with pigment and prolonged into filamentous pro- 
cesses which extend between and amongst the cones. In fact, the cones may be 
said to be embedded in the pigment cells. This layer differs but little from that of 
Anolis, except at the anterior part of the eye, where it becomes much thicker. 

The Cones. — No rods are present. The cones (2, fig. 15 a) consist of an upper 
and a basal part. The basal part is elliptical in shape and stains uniformly through- 





out, while the outer portion is longer and somewhat triangular in shape, with the 
smaller side of the triangle resting on the inner elliptical part. This layer measures 
lo II in depth, while the same layer in Anolis measures 13 /x. 

The Outer Nuclear Layer. — This layer is made up of a single stratum of nuclei 
with small dark nucleoli (3, fig. 15 a). Some of these nuclei are almost spherical, 
while others are oval in shape. They are connected with the cones by broad pro- 
cesses which stain darkly. These processes may be very short, in which case the 
cone comes in close proximity to the nucleus ; or they may be drawn out into fila- 
ments as long as or longer than the nuclei themselves. From the inner part of the 

nuclei extend processes which broaden toward the base and 
send numerous ramifications into the inner stratum of the 
outer reticular layer. There is a striking difference here 
between this eye and the normal one. The processes from 
the base of the nuclei pass straight through the outer reticu- 
lar layer, while in certain sections of the normal eye they 
pass through at an angle of about 45° (3, fig. 15 b). 

The Outer Reticular Layer. — The outer reticular layer 
(4, fig. 15 o) is penetrated by the processes from the nuclei 
, of the outer nuclear layer and by a few Miillerian fibers. 
If processes from horizontal cells are present, they were 
not brought out by the method of staining which was used. 
Again, there is but little difference in the thickness of this 
layer in the two eyes, as it measures 6 /a in Amphisbcena 
and 7 /A in Anolis. 

The Inner Nuclear Layer. — The inner nuclear layer 
is a compact mass of somewhat irregular spherical nuclei 
and is 24 fi in thickness (6, fig. 15 a). The corresponding 
layer in Anolis is 59 /*. Spongioblast and bipolar cells can 
not be differentiated from each other. All of the nuclei 
appear to be very much alike, except the nucleated en- 
largements of the fibers of Miiller, which have no definite 
shape and which stain very densely. However, some nuclei, 
more especially those of the inner stratum, stain a very deep 
black color, and show no structure whatever. Parts of 
certain other nuclei stain densely, while the rest retains its 
original identity. Some of the nuclei have 4 to 6 nucleoh. 
In Anolis two other kinds of nuclei appear. A few flattened 

Flc. 16. Diagram showing Compara- , . ^ . ■,. , ^, •jjiTaI-I 

live Mcasurcmcnis of Retina in Eyes horizontal nuclci Can De sccu ucar the middle 01 the layer 

oi Anolis and Amphisbetna. , . , . i r i i • i 

and m the inner stratum are a number of large spherical 
nuclei. Penetrating this layer are many fibers of Miiller. Each fiber as it passes 
through is characterized by a nucleated enlargement. 

The Inner Reticular Layer. — The inner reticular layer measures 20 fi in thick- 
ness as against 45 /a in Anolis (8, figs. 15 a and 15 b). The method of staining 
brought out no definite structures. The fibers of Miiller pass through it as fine 
vertical filaments. Occasionally there is a nucleus from the nuclear layer or from 
the ganglionic layer which lies embedded in the edge of this layer. 



The Ganglionic Layer. — The ganglionic layer (9, fig. 15 a) consists of a single 
layer of nuclei 6 /x in diameter, with now and then another nucleus above or below 
the single layer. From the outer side of these nuclei, fibers which run out and 
penetrate the inner reticular layer can be traced for a short distance. On the 
opposite side are also fibers which continue as fibers of the nerve fiber layer. In 
Anolis this layer measures 23 /a and is made up of loosely connected nuclei, some 
of which are large and spherical, others are smaller and irregular, while still others 
stain very densely. 

The iSferve-fiher Layer. — The nerve-fiber layer is 6 /i, in depth, while in Anolis 
it is 26 fi. 

The Fibers of M tiller. — The Miillerian fibers can be traced from the membrana 
limitans interna to the outer nuclear layer. They commence at the inner surface 
of the retina by a broad conical foot which extends into the ganglionic layer. 
Through the inner reticular layer the fibers pass as fine filaments, but in the inner 
nuclear layer each fiber is characterized by an irregularly shaped nucleus, which 
stains densely and shows no structure. The membrana limitans externa is not 
visible. These fibers differ but little from those in Anolis, except that those in 
Anolis can be traced to the membrana limitans externa, which is plainly visible. 




Rhineura floridana Baird is a legless, burrowing, blind Amphisbanian lizard. 
It is abundant in some parts of Florida. The largest individual secured by the 
author measured 340 mm. The tail is very short, flattened dorso-ventrally, and 
the upper surface of its distal half is strongly rugose. Each of the transverse rings 
is here, with numerous tubercles. The mouth is small ; the tip of the lower jaw 
is some distance behind the tip of the upper jaw. In shape, color, and arrange- 
ment of its dermal plates it strikingly resembles an earthworm. This resemblance 
is heightened by its vermiform progression through the rhythmic movements of 
its annular plates. Its forward and backward locomotion in its burrows is entirely 
due to this vermiform movement. It burrows rapidly, and for this its small, 
hard, conical head is well adapted. The point of the snout is turned down and 
the head then thrust upward in a rooting fashion. An individual will readily dis- 
appear in from half a minute to two minutes. By placing it in a glass vessel partly 
filled with earth its burrowing can readily be seen from below. If placed on a 
bare surface, it for a time will wriggle actively from side to side, snake fashion, but 
without much effect as far as locomotion is concerned. The tail, under such cir- 
cumstances, is dragged behind, as if it had no vital connection with the head. 
Rarely there is a suggestion of a bracing with the tip of the tail against the floor. 
In one minute an individual moved 250 mm. In an attempt at rooting, after the 
snout had become wedged under the edge of an immovable object, the whole body 
to the tip of the tail was repeatedly lifted oflf the floor. 

Rhineura is, as far as known, one of the two blind vertebrates that have been 
found in the fossil state. Baur described a species of Rhineura (R. hatcherii) and 
another Amphisbaenian (Hypsorhina antigua) from the Miocene beds of South 
Dakota. Baur says nothing concerning the dermal plates, so that nothing is 
definitely known about the eyes of this fossil Rhineura. Since all the genera of 

the family Amphisbaenidae have rudimentary eyes, 
the eyes were very probably degenerate before the 
genera became separated. It seems quite certain 
that any fossil members of an existing genus all of 
whose living species have degenerate eyes, must have 
had eyes that were to a greater or less extent degen- 
erate. The time suggested by this find of Baur 
during which the eyes of Rhineura have been degen- 
Fio. 17. Side View of Head of Rhineura Crating is surprisingly long, extending as it does 

showing Surface Plates and Position of. ii^^ ^r^if ±* e 

Eye in Relation to them. tMough about 5 to ID per ccnt of thc fomiatiou of 

sedimentary rocks. 
Rhineura is a burrowing animal, and blind animals which burrow in the ground 
are not found in naturally made caves. The latter are largely populated by species 
that tend to hide in crevices or natural cavities under rocks. It would seem from 
this that the cave fauna was incipient before the existence of caves, and that the 
latter were colonized as soon as they were large enough to admit their present 

' See Proc. Wash. Acad. Sci., iv. p. 533, 1902. 



Rhineura floridana. 

A. Side and dorsal views of tail. 

B. Horizontal section of head, showing Harder's gland and position of eye. 

C. Horizontal section through right eye, showing solid strand of cells, extending from 

Harder's gland to near epidermis. 

D. Horizontal section of left eye, showing extent of pigmentation and lens. 

E. Distal part of another section of same eye, showing different layers of retina at their 

highest development. 2 mm. objective. 

F. Proximal part of another eye, showing cyst represented diagrammatically b text- 





The eye of Rhineura floridana is not visible externally, nor is there any indica- 
tion where it formerly came to the surface. The side of the head is continuously 
covered with plates. There are 4 labials (i, 2, 3, and 4, of fig. 17), the posterior 
of which is comparatively large. Above the labials from in front backward lie a 
single nasal (5), a single loreal (6), a single preorbital (7), and a group of temporals 
(8). Above this series of plates lie a supranasal (9), joined to its fellow of the other 
side, a prefrontal (11), and 2 supraciharies (12, 13). In heads cleared with xylol 
the black eye can be seen to lie underneath the angle between the 2 supraciliaries 
and the preorbital. 

The dermis and epidermis over the eye are not different from these structures 
over neighboring regions except that in one instance (plate 3 c, dt) a solid column of 
cells 32 /A thick extends from Harder's gland to near the epidermis, without, how- 
ever, fusing with the latter. Fisher found that in Trogonophis the epidermis is 
reduced to half its thickness and free from pigment over the eye. In Amphisbcena 
strauchi and A. darwini the skin is not thinner and the pigment is little or not at 
all less over the eye. A conjunctival sac has been described for various Amphis- 
bsenians. No such structure is present in Rhineura. 

Harder's gland (plate 3, figs, b, c, is out of all proportion to the size of 
the eye. In a horizontal section it measures about 4 times as long as the eye 
(medio-laterally) and 3 times as wide (antero-posteriorly). Duvernoy found that 
in Typhlops Harder's gland is 10 times as great as the eye. It is divided into 2 
distinct lobes, that over the anterior face of the eye is histologically quite different 
from that over the posterior face. In vertical section the gland is seen to entirely 
surround the eye except sometimes at its lower posterior quarter. The large size 
of Harder's gland has given rise (Duvernoy) to the conclusion that its function is 
not connected with the eye. Its secretion is poured directly into the tear duct and 
through it into the nasal cavity. 

The distance of the eye beneath the outer surface of the epidermis measures 
between 320 and 560 /a in specimens between 280 and 310 mm. long. It is sur- 
rounded by 2 layers of connective tissue. These are thin over the distal half of the 
eye. Over the proximal narrow end of the eye they become thick ; and since they 
are prolonged beyond the eye, stain a different tint, and readily become separated, 
they are easily distinguishable. They probably represent the sclera and choroid. 
If so, the choroid is practically free from pigment except possibly in rare instances 
where a few pigment granules were detected in cells closely applied to the eye. 
There is no indication of any differentiation into a cornea or capsule of any sort. 
The fibrous sheaths are at the proximal end drawn out into a cone. A supposed 
scleral cartilage has been found in one individual. Here a bar about 20 /j. thick 
extends from over the center of the distal face of the eye for 160 /a around its pos- 
terior face. It stains and has the structure of bone rather than of cartilage. No 
traces of any muscles have been found connected with the eye. 

The eye is directed outward and forward. Its axis is horizontal and makes an 
angle of about 60° with the sagittal plane of the body. It does not occupy a defi- 
nitely fixed position on its axis, for in the eye of one side the choroid fissure was found 
directed caudad, in the other eye ventrad. It is irregularly pear-shaped, with its 
anterior face convex, its posterior face flat or even concave. The eyes in 3 speci- 
mens give the following measurements in microns : 


Measurements (in microns) of eyes of Rhineura. 

Length ok 

Meuio-latebal diameter. 

Anteko-posterior diaueter. 

Distance from surface. 

Left eye. 

Right eye. 

Left eye. 

Right eye. 

Left eye. 

Right eye. 













'*|'A11 the structures vary greatly in different eyes so that the terms "sometimes," 
"usually," "frequently," etc., have to be used much more than is desirable. This 
can not be avoided unless each eye is given a distinct description. 


Fig. 18. (a) Saffittal Section through Middle of Left Eye of Rhineura, about 300 mm lont?. 

(6) Vertical Section through Distal Part of Eye of Rhineura, showing Lens with Capsule. 

(c) Lens of Right Eye of Individual, 275 mm. long. Horizontal Section. 

(d) Left Lens of Same Individual. 



(a) The Iris. — In the structure of the irideal region the eye of this species is 
unique among the degenerate vertebrate eyes so far described. In all other eyes, 
with the possible exception of Troglichlhys, elements of an iris are distinctly recog- 
nizable. In Rhineura the fold of double epithelium between the pigmented and un- 
pigmented part of the retina whose margin is the margin of the pupil has been 
obliterated and the pupillary edge forms the extreme outer edge of the blunt end 
of the pear {p, fig. 19 a). The pigmented layer of the retina in other words merges 
directly into the unpigmented layers of the retina. The entire thickness of the 
retina is thus exposed at the distal face of the eye. 


Fio. 10. (a) Hori/ontal Section of Left Eye of Specimen, 380 mm. long. 

(b) Another Section through Same Eye, showing Exit of Optic Nerve, the Pigmentless 

Condition over Anterior Face of Eye, and Invaginatcd Pigment at End of Pear. 
(fi) Outline of Pigment in Proximal End of Uight Eye of same Individual, showing 
Invagination of- Pigment to form a Cyst. 


(b) The Vitreous Body. — The vitreous cavity is represented by a vertical slit 
extending from the axis of the eye downward to the edge. The choroid fissure 
(fig. i8, chr.f.) thus remains permanently open in so far as the edges of the opposite 
sides of the fissure are not united. A space a few microns wide was found in one 
eye. In other cases there is no real cavity and no vitreous body. The hyaloid 
membrane (fig. i8 and fig. 19, hd) is represented by a few cells with elongated nuclei. 
Blood-vessels were not found in it.* 

(c) The Lens. — In two specimens no traces of a lens were found, but in 
two other specimens a lens was present. There being no pupil and no vitreous 
cavity, the lens is situated in a little depression in the distal face of the retina 
(figs. 18 b, c, d). The lenses differ greatly from each other. In the better 
developed instances (fig. 18 b) it is composed of a spherical mass of cells. The 
nuclei are granular and are surrounded by a hyaline cell body. These little cap- 
sules are closely packed in a slightly darker matrix. The whole lens is surrounded 
by a fibrous capsule containing elongated nuclei. Both eyes of one individual are 
provided with lenses as described. In another individual the 2 lenses differ ma- 
terially not only from those described, but from each other both in structure and 
size. The left lens consists of a lenticular nodule containing about 6 dense nuclei 
(fig. 18 d). On the right side (fig. 18 c) the lens is much larger. It consists of 
2 large nucleated capsules surrounded by a matrix containing a few dense elon- 
gated nuclei similar to those of the capsule surrounding it (figs. 18 b, c, and d, are 
drawn to the same scale). The difference exclusive of size between the 2 lenses 
c and d and the lens represented in figure 18 b, may be due to differences in the 
method of preparation. 

(d) The Retina. — The numbers in the following paragraphs are not consecu- 
tive, but are those used to designate the corresponding layers in the figures. 

(i) The pigment epithelium forms a complete outer layer of the eye exclusive 
of its distal face and a narrow strip along the choroid fissure. The extent to which 
this epithelium is pigmented differs greatly in different eyes. A region along either 
side of the choroid fissure is free from pigment, occasionally parts of the anterior 
face of the eye are free from pigment (fig. 19 b), and very frequently the cells of 
this layer around the distal margin of the eye are free from pigment. Over the 
anterior face of the eye this layer is usually composed of a regular layer of cells 
whether these are free from pigment or not (figs. 19 a, b). On the posterior face the 
series of cells is not nearly so regular. The pigmented epithelium is here invaginated 
and folded upon itself in various ways. The infoldings are sometimes solid masses 
of pigment cells, but sometimes they form hollow spheres which contain a mass 
of concentrically arranged unpigmented material, probably of choroidal origin 
(plate 3, F, and text-fig. 19 c.) What the significance of these cysts may be I can 
not conjecture. Indications of similar structures were seen in the eyes of Amblyopsis. 

The narrow stalk of the pear-shaped eye is usually filled with an irregular 
jumble of pigment cells. In favorable sections it is seen that these are also the 
result of an invagination of the pigment epithelium from the pointed end of the eye 
(fig. 19 b). The pigment epithelium has not been reduced at the same rate as 
the rest of the retina; as a consequence it is infolded in various ways. Small 

' The figures were drawn with camera lucida from sections mounted in balsam; 2 mm. objective and 4 eye- 
piece. The horizontal sections were made from above down and are so drawn that the anterior face of the figure 
IS toward the top of the page. 


pigment cells are sometimes found in the inner layers of the retina among the gan- 
glionic cells and along the optic nerve within the eye. Pigment cells were also 
found in the eyes of Typhlomolge (figs. 5 a and 7 a, z). There are rarely any 
pigment cells over the distal face of the eye. 

(i a) X, nuclei. — In the eyes of Rhineura, Typhlichthys, and Troglichthys a 
few cells with elongated, tangentially placed nuclei are present between the pig- 
mented epithelium and the outer nuclear layer. They are distinctly outside of 
the outer limiting membrane (figs. 19 a, 6; plate 3, fig. E, x). The origin of these 
nuclei is difficult to explain. Possibly they are derived from the pigment epithe- 
lium which in some of the unpigmented regions (fig. 19 b, x) is more than one layer 
deep. If the outer layer should become pigmente<l, the inner nuclei, if they 
remained unpigmented, might give rise to these longitudinal cells. 

(2) Rods and cones are not present. There is in some cases a distinct space 
between the pigment epithelium and the outer nuclear layer. This space when 
present is partially filled with filmy, hazy structures, but nothing suggesting defi- 
nitely either a rod or cone was detected (fig. 19 a and plate 3, fig. e). 

(3) The outer nuclear layer consists of about 2 series of elliptical nuclei. They 
form a compact and distinct layer a few microns from the outer limiting mem- 
brane (figs. 18 a, 19 a, &, and plate 3, fig. e). 

(4) The outer reticular layer is represented by a series of distinct but irregular 
gaps between the outer nuclei and the inner nuclei. Horizontal cells are not pres- 
ent (figs. 19 a, b, c, and plate 3, fig. e). 

(6) The inner nuclei are smaller, rounded, and less granular than the outer 
nuclei. They do not form as compact a layer as the outer nuclei. It is impossible to 
distinguish between bipolar and spongioblastic cells (6 in figs. 18, 19, and plate 3). 

(8) The inner reticular layers, as is usual in degenerate eyes, are well developed 
in the eyes of Rhineura. They are frequently crossed by MuUerian fibers. 

(9) The ganglionic layer is represented by a number of nuclei loosely grouped 
about the vitreous slit. The individual nuclei are distinctly larger than those of 
the inner nuclear layer and less oval than those of the outer nuclear layer (9 in the 

(10) A distinct optic fiber layer is not present and the optic nerve is nowhere 
within the eye a compact strand of fibers. A loose flocculent strand of fibers passes 
through the proximal part of the retina. Its path through the pigmented layer is 
difficult to trace. Beyond the eye the optic nerve can be followed by means of the 
fibrous sheaths and pigment cells associated with it rather than by the presence of 
any fibers with a distinctly nervous structure. The optic nerve leaves the eye, not 
at the proximal end or the narrow end of the pear, but anterior to the pigment 
mass in the narrow part of the pear (fig. 19 b, n.op.). 



Typhlops lumbricalis (Linnaeus), a blind snake, is generally distributed in the 
West Indies and Guiana. The specimens examined were obtained in the neigh- 
borhood of Canas, Province Pinar del Rio, Cuba. It is a burrowing form that 
lives just beneath the surface, being thrown out even by the plow. 

The snakes were first placed in formalin and after a few days were transferred 
to alcohol. Only one young specimen was obtained, and it was preserved in 
Zenker's fluid. For decalcification, the heads of some were placed for at least 3 
days in 10 per cent nitric acid and others in Perenyi's fluid from i to 2 weeks. One 
series was stained by the iron haematoxyhn process, the others with haemalum and 
eosin. It was very difficult to obtain satisfactory sections and especially complete 
series from the specimens, since no method was found to decalcify properly and to 
get the integument in condition for sectioning. 

The lengths of the individuals examined were 10, 20, 21, and 21.5 cm. The 
color is brown above, on the ventral side it is yellowish white. The body is cov- 
ered with scales of uniform size, while those of the head are somewhat larger. The 
surface of the entire body is very smooth and shining and rather hard. The tail, 
which is about one-twentieth of the body's length, ends in a short, sharp spine. 
The mouth is small and lies on the ventral side some distance back from the tip of 
the snout. 


Snakes differ from other animals in having the edges of the two eyelids entirely 
grown together. A disk-shaped, conjunctival sac is thus formed and the layers 
over the eye between this sac and the exterior form the "brille." Six weakly 
developed muscles are present. The 4 straight ones arise in the neighborhood of 
the foramen opticus, while the 2 oblique ones arise from the surface of the prefrontal 
which is turned toward the eye socket. 

Closely connected with the eye is Harder's gland, whose function is doubtful. 
Leading from this gland is a single duct, which either empties into the duct from 
Jacobson's gland or directly into the mouth cavity. The secretions of the gland 
are thus not functional in connection with the eye. 

The sclera consists of closely woven fibers. Cihary muscles are not found, but 
next to the iris is a great bundle of equatorial muscle fibers running obliquely, 
which seem to be a continuation of the iris musculature. The ciliary processes are 
weakly developed. 

The retina consists of the usual layers. The nerve-fiber layer is very thin (0.003 
to 0.004 mm.). 

The ganglion-cefl layer consists of a single, rarely two, layer of small cells, each 
with a very large nucleus (0.012 to 0.013 mm.). The inner reticular layer contains, 
at apparently regular intervals, elongated, oval nuclei (0.042 to 0.045 mni.). The 
inner nuclear layer consists of two kinds of cells (0.052 to 0.054 mm.). The outer 
reticular layer is very thin (0.004 to 0.005 miTi-)- 

' By Effa Funk Muhsc. Sec Biol. Bull. vi. p. 261, 1903. 




coals -s, ■ 

Eye of Typhlops lumbricalis. 

A. Horizontal section, from specimen 20 cm. long. A and B two-thirds objective, 

2 inch eyepiece. 

B. Transverse section, from specimen 2 I cm. long. (Scales not shown.) 

C. Diagram of eye of adult. 

D. Diagram of eye of young. 


The sensory epithelium consists of the outer nuclear layer and the cone layer 
which is made up of single and twin cones. There are no rods. A single cone 
consists of two sections, an outer extremely small section, 5 to 6 fi in length and an 
inner much larger section, almost completely filled with a larger, pear-shaped, 
strongly refractive body, the ellipsoid, 14 to 16 fx. in length and 8 to 9 /* across its 
widest part, which is turned toward the limiting membrane. The twin cone con- 
sists of two parts, one similar to a simple cone, the other cylindrical and very slender, 
its structure being otherwise like that of a simple cone. It is probable that the two 
parts of the twin cone are connected with but one nucleus. The nuclei of the cones 
vary greatly in form, and leading from these into the inner layers of the retina are 
relatively very large fibers or processes. Passing between the limiting membranes 
are the radial supporting Miillerian fibers. 


The work thus far on blind snakes has been done by Kohl on Typhlops vermi- 
ctUaris, a species found in Greece and the southwestern part of Asia, and on Typh- 
lops braminus, a species found in the islands of the Indian Ocean and in Africa 
south of the equator, accounts of which are given in his "Rudimentare Wirbelthier- 
augen." ' He found that in depth the eye of Typhlops vermicularis is equal to 
about one-sixth that of Tropidonotus. The brille is thicker in Typhlops than in 
Tropidonotus and compared with the axial diameter of the respective eyes it is seven 
times thicker. In Typhlops the brille is equal in thickness to about half that of 
the ordinary skin of the head. In Tropidonotus it is equal to one-fourth. 

The cornea of Typhlops measures 0.0052 mm., and compared with the relative 
sizes of the eyes is equal to about half that of Tropidonotus, which measures 0.064 
mm. The conjunctiva is thickened at the edge of the disk-shaped sac and consists 
here of gland cells, the fornix conjunctiva. The supporting membranes of the 
eyeball, choroid, and sclera are relatively equal to about half those of Tropi- 

Harder's gland in Typhlops is many times larger than the eyeball. The six 
muscles are present. The lens is elliptical, while that of Tropidanotus is almost 
globular. The ratio of the lens volume of Typhlops to the eye volume is i to 14.04, 
while in Tropidonotus it is i to 3.6. The lens epithelium of the former is relatively 
6 times greater than that of Tropidonotus. 

The retina at the back of the eye of Typhlops, and the retina of Tropidonotus 
bear the actual ratio of 8 to 13, while compared with the eye axis in each case the 
Typhlops retina is 4 times greater. The fovea centralis and area are absent. 

The fiber layer has its greatest thickness near the exit of the nerve and gradually 
becomes thinner until, near the iris, scarcely a fiber is found. The globular gan- 
glion cells are arranged in a single layer except occasionally for short distances, 
when they lie in a double row. The inner nuclear layer seems to be subdivided 
into four layers. 

There are no twin cones. Each cone consists of a cone cell, stalk, middle and 
end members. The cone nuclei lie in two series, but the stalks vary in length so 
that the distal ends of the cone members reach nearly the same level. 

' Kohl, Dr. C, Rudimentare Wirbelthieraugen, Erster Theil, Heft 13, Bibliotheca Zoologica. Vcrlag von 
Theodor Fischer, 1892, Cassel. 




The eye shows through the large ocular scale, which entirely covers it. It appears 
as a black spot surrounded by an unpigmented circle. The preocular, also a large 
scale, overlaps the ocular and reaches just to the edge of the eye (figs. 20 a, b). 

Compared with one of the garter 



po.s. a& 



Fig. 20. (a) Dorsal View of Head of Typhlops, 21 cm. long 
(6) Lateral View of Head of same Specimen. 

snakes and in proportion to the size 
of the head, the eye of Typhlops lum- 
hricalis is situated farther from the 
surface and occupies far less space, 
while Harder's gland, associated with 
the eye in both, is relatively much 
larger in Typhlops. In a specimen of 
Typhlops lumbricalis 21 cm. in length, 
the eye measured 0.306 mm. in width, 
and 0.387 mm. in depth. The greatest width of the gland of the same was 0.711 
mm. and the length was 1.067 mm. The gland completely surrounds the eye up 
to the edges of the conjunctival sac (plate 4, figs, a, b). In proportion to the 
size of the eyes, the gland of the garter snake is much smaller than that of 
Typhlops lumbricalis, but compared with Rhineura fioridana the gland of Typhlops 
lumbricalis is but little more than half as large. 

The eye is covered by layers of epidermis and dermis that differ from these 
same layers on neighboring parts by laeing thinner, more compact, and free from 
pigment and glands. The ocular scale, however, which covers the eye region, 
does not differ in thickness from the other scales of the head (plate 4, fig. a). 

A conjunctival sac is present with a diameter at least as great as the greatest 
width of the eye bulb. The conjunctiva, which forms this sac, is very thin over 
the cornea and next to the brille, where it measures 0.003 '^^^ ^t the edge of the 
sac it is differentiated into glands, the fornix conjunctiva, and measures 0.016 mm. 
(plate 4, figs. B and c, F. cj.). 

In horizontal section, the eye axis is seen to be turned forward about 30° away 
from a line at right angle to the horizontal axis of the body. 

Eye muscles are present, but from the sections used, the exact number could 
not be determined. 

Choroid and Sclera. — The dense pigmentation makes it impossible to dis- 
tinguish between the different coats at every point. Beyond the retina with its 
pigment layer is an open vascular space, and this is followed by another dark layer, 
the two together representing the choroid. The choroidal pigmentary layer seems 
to consist of long fibers circularly arranged. The sclera can be followed by start- 
ing with the outer covering of the optic nerve and tracing its continuation 
about the eye. 

Iris and Ciliary Processes. — Here again the pigmentation makes it difficult 
to determine the structure. Both iris and ciliary processes are present, for the 
black layer extends over the anterior surface of the lens, leaving a pupil equal in 
diameter to about one-fourth of the circumference of the lens. At points near 
the equator of the lens this dark layer is enlarged into the ciliary processes and in 
connection with the capsule helps to hold the lens in place. 



Cornea. — This structure is present and can be traced to the region of the 
ciliary processes. 

Lens. — A large lens is present, its depth being equal to about two-fifths of the 
eye depth. From the sections httle could be determined about its structure. A 
well-developed capsule surrounds it (plate 4, fig. c). 

Retina. — The same layers are present that are found in snakes in general, 
but the comparative thickness of the various layers is different. In the garter 
snakes, for instance, the retina is of a uniformly even thickness even to the ciliary 
process, a single layer of cells continues on over the surface of the processes and 
iris, but in Typhlops lumbricalis the retina at the back of the eye is very thick and 
gradually becomes thinner till it ends a short distance from the ciliary processes 
(plate 4, fig. c). At this point the arrangement could not be definitely determined 
in the sections. At the back the retina, exclusive of the pigment layer, measures 
0.0725 mm. 







Fic. 21. (a) Section of the Retina of an Adult Specimen, 21 cm. long. 
(6) Section of the Retina of a Specimen, lo cm. long. 

Ends of fibers were seen projecting inward from the ganglion-cell layer, but 
no definite fiber layer could be distinguished (10 in fig. 21 b). 

The ganglion-cell layer (9 in the figures) consists of a single row of large 
nucleated cells, somewhat irregularly arranged (0.008 mm.). The inner reticular 
layer (8) consists of a mass of fibers interwoven in a close network. This layer 
measures, at the back of the eye, 0.015 mm. 

The inner nuclear layer (6) consists of at least 3 layers of cells, loosely arranged. 
The course of some of the fibers can be followed among these cells. This layer 
measures 0.016 mm. 

The outer reticular layer (4) is very thin and consists of a few fibers so arranged 
as to leave a great number of spaces between the two nuclear layers. The distance 
between the nuclear layers is about 0.005 "^™- 

Cones. — The sensory epithelium shows two distinct parts, an inner layer of 
nuclei (3) and an outer row of cones (2). In the sections these two were so separated 



that a loose tissue was visible, consisting probably of the limiting membrane and 
ends of the Miillerian fibers. The outer nuclear layer in the adult consists of a single 
row of nuclei, with a mass of quite homogeneous material about them. This part 
of the sensory epithelium measures 0.018 mm. The cones are pear-shaped bodies 
with the smaller end pointing outward, and at intervals of every four or five a 
shorter one occurs. Each element is differentiated into two parts. By the iron 
haematoxylin process of staining, the outer small end is densely stained, while the 
body of the element is a light granular mass (fig. 21 a). 

The pigment layer (i) is a continuous layer of even thickness, similar in every 
respect to that of the garter snake. 

One young specimen, 10 cm. in length, was examined. The eye as a whole, 
as well as the lens, is nearly spherical. The eye measures in width 0.290 mm. 
and 0.322 mm. in depth. All parts are so developed that the vitreous cavity is 
relatively much smaller than that of the adult. The coats are thicker, the ciliary 
processes better developed, the lens capsule thicker, and the retina at the back 
actually measures one and two-thirds the depth of the adult retina. The ele- 
ments of each layer are much more numerous than in the adult, and they are 
packed much more closely together (fig. 216). The ganglion nuclei are apparently 
arranged one against the other. In the inner reticular layer occur the "interpolated 
cells." These were not found in the sections of the adult eye that were examined. 
The cells of the inner nuclear layer are smaller and arranged in five or six rows. 
There is a well-developed outer reticular layer similar in its make-up to the inner 
reticular. Instead of a single row of cone nuclei with its surrounding homogeneous 
mass, as in the adult, this layer in the young consists of five or six rows of small 
closely arranged cells. The cones likewise are smaller and more numerous (fig. 

Comparative Measurement of Retinal Layers in millimeters. 

Fiber layer. 

cell layer. 


























Typhlops lumbricalis (adult) 

Typhlops lumbricalis (young,io cm.) 

Relative Proportions of Eye Parts. 

Tropidonotus natrix. 

Typhlops vermicularis. 

Typhlops lumbricalis (adult). 

Eye depth. 

2.5541 mm. 

0.4399 "im. 

0.4032 mm. 

Brille : 

Eye axis:: i : 77.4 

1 : 10.77 

1 : 12.5 

Cornea : 

Eve axis:: i : 39.9 

1 : 84.6 


Lens depth : 

Eye axis:: i : 1.56 

i: 303 

i: 2 5 


Eve axis:: i : 21.63 

1 : 38.58 


Retina at back : 

Eye axis:: i: 19.19 

i: 5-36 

i: SS 




(i) The eye muscles have entirely disappeared. 

(2) Only the uveal parts of the iris remain. 

(3) The lens has retained its shape and position, but its structure has been 
greatly changed. No capsule is present. 

(4) Harder's gland is many times larger than the eye and pours its secretion 
into the conjunctival cavity and thence into the mouth. 

(5) The sclera, scleral cartilages, cornea, vitreous body, and pigment epithe- 
lium have undergone but little change unless it be in the reduction in size. 

(6) The cuticle passes over the eye unchanged. 

(7) The aqueous cavity is no longer present. 

(8) All the layers of the retina are still present. As shown in figure 6, the 
great reductions in the depth of the layers, in comparison with those of Anolis, have 
taken place in the nerve fiber, ganglion cell, inner reticular and inner nuclear layers. 

(9) If the eye has been reduced from an eye of the average size, all parts have 
certainly undergone considerable change, and this change has been approximately 
equal among the several parts. 

(10) The retina does not show such a profound change as either the iris, muscles, 
or lens. However, it has been greatly changed, as it extends only 50.7 per cent of 
the distance around the eye. 

(11) The eye of AmphishcBna shows that the more active parts of the eye are 
the ones to degenerate first. They are the parts which have been most affected. 


(i) The eye of Rhineura has reached its present stage as the result of a process 
of degeneration that probably began in the early Miocene. 

(2) The dermis and epidermis pass over the eye without any modifications. 
The conjunctival pocket has vanished. 

(3) Harder's gland is many times as large as the eye and pours its secretion 
into the tear duct and thus into the nasal cavity. 

(4) The eye muscles have disappeared. 

(5) A cornea is not diflferentiated. 

(6) The lens is absent in half the eyes examined and varies greatly in those 
in which it is present. 

(7) The vitreous body has practically disappeared. 

(8) The pigment epithelium is variously pigmented. It is of greater extent 
than is sufficient to cover the retina and has been variously invaginated or puckered 
over the proximal and posterior faces of the eye. 

(9) An uveal part of the iris is not present. 

(10) The eye of Rhineura does not represent a phylogenetically primitive 
stage ; it is an end product of evolution as truly as the most highly developed eye. 

(11) The adult eye shows few indications that there has been a cessation of 
development at any definite ontogenetic stage. It does not resemble as a whole 
any ontogenetic stage. 


(12) An arrest in the ontogenetic development has taken place in so far as the 
number of cell multiplications concerned in forming the anlage of the various 
parts of the eye have decreased in number, and in the lack of union of the lips of 
the choroid fissure. 

(13) It is possible that the absence of cones or rods is due to an arrest in the 
histogenesis of the retina, but since these structures are normally formed in the 
young of Typhlotriton and disappear with age, it is possible that their absence in 
the adult eye of Rhineura is also due to ontogenetic degeneration. 

(14) The irregularity in the structure and existence of the lens and the great 
reduction of the vitreous body offer evidence in favor of the idea of the ontogenet- 
icaUy and the phylogeneticaUy earlier disappearance of the ontogenetically and 
phylogenetically newer structures. 

(15) Horizontal nuclei found between the pigment epithelium and the outer 
limiting membrane are probably derived from the proximal layer of the optic cup. 

(16) The different layers of the retina have reached a degree of differentiation 
out of proportion to the great reduction of the dioptric apparatus and general 
structure of the eye. 


(i) The dermis and epidermis over the eye differ from the same over neigh- 
boring parts, by being thinner, more compact, and free from pigment and glands. 

(2) The conjunctival sac is present and has a width at least as great as the eye. 

(3) Harder's gland surrounds all but the distal part covered by the conjunc- 
tival sac. 

(4) Eye muscles present, but their number and structure could not be made out. 

(5) A large lens with capsule is present. 

(6) The various layers of the normal snake retina are present but the com- 
parative thickness is different. 



The eyes of this myxinoid of the Pacific coast were examined by Allen and by 
Stockard. Allen found that they show a very primitive structure, which is in 
reality the result of a complex process of degeneration. The eyeball is found em- 
bedded in a mass of fat about three times its size. In one case, the eye was found 
to lie some distance beneath the outer surface of the mass of fat. Normally, how- 
ever, the corneal surface lies on a level with the surface of the fat and is often 
flattened to form a rather extensive free surface. No eye muscles nor traces of 
such were discovered. No oculomotor nerves were found. No traces of them 
are discoverable in embryonic life (Kupffer). The choroid and sclerotic coats 
are represented by a very thin layer of unpigmented, non-vascular connective 
tissue without any appreciable distinction between corneal and sclerotic portions. 
The retina remains in the early condition of an optic cup, the outer layer (pigment 
layer) not being fused with the remaining layers. All specimens showed the layer 
in question to be widely separated from the bulk of the retina. This pigment 
layer is composed of a single layer of cubical cells devoid of pigment as far as 
could be ascertained. A layer corresponding to that of the rods and cones in higher 
vertebrates is clearly present. The nuclei of these structures (outer nuclear layer) 
are strikingly well developed and regularly arranged. Certain characteristic cells 
of the inner nuclear layer could be readily made out. The ganglionic layer is 
represented by cells scattered irregularly throughout the inner reticular layer. 
Fibers from these last-named cells can be traced in a more or less direct course to 
the optic nerve. The outer rim of the optic cup is in many cases differentiated in 
such a manner as to suggest a rudimentary iris. A structure unmistakably like 
an iris was found in one specimen examined. The celhilar structure of this rudi- 
mentary iris is almost identical with that of the pigment layer. No indications 
of muscle fibers or pigment are to be seen. Certain deeply staining coagula within 
the optic cup give evidence of a vitreous body. Some large, clearly-marked cells, 
probably those of the vitreous body, are found attached to the surface of the retina. 
Evidences of a choroid fissure are to be seen in the fact that the ventral part of 
the retina is thinner than the dorsal in almost all specimens. In one case the 
choroid fissure was found to persist. The most striking feature, however, is the 
extreme variation. The optic nerve enters the eye at various angles. Variation 
occurs in all parts of the eye and is especially notable in the measurements of the 
thickness of the retina and the dimensions of the eye as a whole. 

Stockard found that the lens-bud results from a contact of only a portion of 
the optic cup with the ectoderm. This structure continues to develop for a time 
until, in an embryo considerably more advanced and measuring 15 mm. in length, 
one sees the lens-bud with a slight indication of a constriction about the periphery 
of its area of union with the ectoderm, as if it were preparing to pinch oflF. Here 
the progressive development of the lens ceases and degeneration begins. It soon 
disappears entirely. He considers the cessation of development in the lens due 
to the absence of a durable contact with the optic cup upon which lens formation 
is directly dependent. 




It is not the intention to review the literature on the normal eyes of fishes. A 
list of papers dealing with their macroscopic aspect has been furnished by Ziegen- 
hagen in 1895, while those dealing with minute structure have been enumerated 
by Krause in 1886 and Cajal in 1894. The current literature is discussed periodi- 
cally by Virchow in "Die Ergebnisse der Anatomic und Entwickelungsgeschichte." 

The topographical relationship of the cells of the retina obtained an entirely 
new light by the application of the methylene-blue method chiefly on the part of 
Dogiel, and the Golgi method principally through Ramon y Cajal. The layers 
of the retina of fishes as made out by Ramon y Cajal are as follows, beginning at 
the periphery and going toward the center of the eye: 

1. Epithelial-pigment layer. 5. Horizontal cells. 8. Inner molecular layer. 

2. Rods and cones. 6. Bipolar cells. 9. Ganglionic layer. 

3. Outer nuclear layer. 7. Spongioblasts. 10. Optic fiber layer. 

4. Outer molecular layer. 

Throughout this work the layers are designated on the figures and frequently in 
the text by these numbers. The literature bearing on the eyes of the blind species 

will be given under the different species. 

(~y^ (~^C\^ C) r ^ The horizontal relations, especially the 

^-^^^ /^^^t\ /^ y^ mosaic of the single and twin cones in the 

retinas of fishes, has been dealt with by 
Hanover, Miiller, Krause, Friis, Ryder, 

/^\~\ ^S^/''N^ /*K V^ Beer, Eigenmann, and Shafer. 

v_A^ LJvJ()() It was found that in many fishes the 

single and twin cones form a regular mo- 
saic. The number of parts entering into 
each unit of the retinal mosaic is remark- 
ably constant for any species, but differs 
considerably in different species of fishes. 
The "shape" of the unit differs in different 
parts of the retina. The pattern may be 
made up of twin cones only.* The axes or 
hues joining the centers of the components 
of each twin if continued may be at right 
angles to each other and form a square 
(fig. 22 a), or they may be approximately 
parallel (Sebastodes, c), or they may be 

F.o.«. Types of Single and Double Cones in Retinas varlously inclined to each Other and form 
of Various Fishes. rhombs {ScorpcBHa, b). 

In other genera (Perca, Micropterus, Etheostoma, and Pimephales) a single 
cone is placed in the center of each of the units of 4 twin cones (d). In still others 
{Blennius, e) a single cone is added at each angle of the unit, and in still others 

' Krause fou|id only single cones in the eel. 


(Salmo, Coregonus, /), a single cone is found both at each angle and in the center 
of the unit. The most complicated unit (Esox, g) is composed of 5 twins, 4 form- 
ing the sides and i a diagonal, and of 4 single cones, i in each corner. These 
patterns are all regular, but not mathematically so. 

In some families (Silurida; and Catostomida;) no regularity could be made out. 
In general the number of rods is inversely proportional to the number of single cones. 

Fig. 23. I to 6 show Section of Eye of Bass 6 cm. long. The Eye measured 3.8 mm. in Diameter from 
Cornea to Back, and 4.7 mm. from Anterior to Posterior Edges. 7, 8, and q show Sections of Eye of 
Bass 33.5 cm. long. The Eye was 10 mm. in Diameter from Cornea to Back, and 13 mm. from 
Anterior to Posterior Edges. All figures drawn to the same magnitication. C, part turned toward 
cornea; By part pointing from cornea. 

In the black bass, the only species in which the pattern was examined over the 
entire eye, the number of components in each unit of the mosaic is the same, but 
the shape of the pattern varies regularly from a rectus at the anterior and posterior 
faces of the eye, to a rhomb above and below. The elements of the unit and the 
entire unit increase in size with the growth of the eye. New elements are not 
added after the pattern has been established. 




Of the eyes of a number of species of normal fishes, namely those of Cyma- 
togaster aggregatus, Carassius auratus, Ameiurus sp., Coregonus sp., and Zygonectes 
no/a/M5 examined, I shall briefly describe the eyes of but one. 

Zygonectes notatus (Rafinesque) was selected for comparison, since it is a 
member of the Cyprinodontidae, a family closely related to the Amblyopsidae. I 
am not aware that this species has any advantage over 
other species of the family. It has large, well-developed 
eyes, that we may assume to be fully and normally devel- 
oped. The material examined was alcoholic. It had been 
preserved by simply placing in alcohol without any intention 
of future histological examination, but the structures were all 
well preserved for making out the horizontal relations of the 
single and twin cones. The protoplasmic and nervous 
processes of the cells were of course not brought out as 
with Golgi's method. 

A specimen 38 mm. long had the eye 2.24 mm. in 
length, 2 mm. in vertical diameter, 1.12 mm. from axis of 
optic nerve to front of iris, 1.6 mm. from axis of optic nerve 
to front of cornea; lens 0.96 mm. in diameter; pigment 
layer measures 56 /a; outer nuclear layer, 36 fi; outer reticu- 
lar, 4/i,; tangential cell layer, 9 /u,; inner nuclear, 40 /a; inner 
reticular, 52 /u-; ganglionic layer, 12 /u,; optic-fiber layer, 28 /a; 
total thickness of retina, 237 /*. 

The regularity of arrangement of single and twin cones 
is very striking. The basal part of the single cones con- 
tains refractive granules increasing in size outward where 
the series ends in a lenticular vacuolated body separating the 
granular from the distal part of the rod. The twin cones are 
all without granulation. This marked difference between 
the two enables one to distinguish between them at a glance 
in tangential sections. The twin cones are arranged in 
series in such a manner that the axes joining the cones in any 
neighboring series are at right angles to each other, while in 
every alternate series they extend in approximately the same 
or parallel directions. The single cones alternate in all 
directions with twin cones (fig. 24 b). 
Fio. 24. (a) Section through Retina Thc outer nuclei are Irregular, comprcssed , and elongate, 
(i) conc"a"tem oTsame formlug two distinct kycrs. The outer molecular layer has 
an irregular outer boundary produced by the process extend- 
ing toward the outer cells. The inner nuclear layer is divided into an outer layer 
of small bipolar cells and an inner layer of larger, more coarsely granulated 
spongioblastic cells. When any breaks occur in the retina, owing to mechanical 
or chemical causes, they usually occur between these outer bi-polar and inner 
spongiose cells of the inner nuclear layer. 




San Diego Bay is in part surrounded by mud flats which are covered by water 
at high tide. Sand beaches take the place of the mud flats where the channel 
approaches the shores. On the ocean shores a sandy beach stretches several miles 
to the southeast from the mouth of the bay, while on the west rises the point of 
land called Point Loma. The entire ocean beach at the base of this promontory 
is rocky. In many places all the earth has been removed by the action of the 
waves, leaving the bare rock ; in other places, and more especially between the outer 
point and Ballast Point, large bowlders lie embedded in the sand (frontispiece). 
These are all covered at high tide, while but a few small pools remain about the 

Fig. as. 

(a) Young GUlichthys mirabilis (lirard. 
(6) Larva of Clci'dandia or Lepidogobius. 

From mud flats of San Diego Bay. 
From surface of San Diego Bay. 
(c) Clevetandia ios Jordan and Gilbert. From San Diego Bay. 
(rf) Quielula y-cauda Jenkins and Evermann. From San Diego Bay. 

rocks at low tide. Many of these rocks are covered with seaweeds, actineans, and 
especially large chitons. All these localities are inhabited by relatives of the Point 
Loma blind fish. The sloughs traversing the mud flats of the bay are inhabited 
by GUlichthys mirahilis Cooper, the young of which is represented in figure 25 a. 

In the mud flats every tide pool as large as a man's hand contains Clevelandia 
ios (fig. 25 c) ; nearer low-water mark in similar localities Quietula y-cauda are 
found, but less abundant than Clevelandia ios. On digging in the sandy beaches 
of the bay specimens of another species of this group, Ilypnus gilberti, are some- 
times found buried in the sand. In the crab holes under the rocks about Point 
Loma occurs the most remarkable of this family, the Point Loma blind fish, Typhlo- 
gobius californiensis (fig. 26 a). In deep water off Point Loma lives still another 
goby, Gobius nicholsoni. 



It is thus seen that almost every nook available has been taken possession of 
by these diminutive fishes. All of them have the two ventrals united along the 
median line and a thin membrane stretched across their bases to form a pouch. 
By appressing the ventrals and then raising them, a partial vacuum is formed in 
this pouch and the fish is enabled to cling to any substance with which its ventral 
happens to be in contact. In confinement the blind fish frequently utihzes the 
surface of the water of an aquarium for a surface of attachment. 

All the species in the bay have the habit, if disturbed, of hiding in crab or 
clam holes. Clevelandia will sit on its tail and pectorals until the hand is near it ; 
then with a quickness which would do honor to a Johnnie Darter, with a flirt of 
the tail and a stroke of the pectoral, it disappears into its hole, from which, how- 
ever, it at once thrusts its head to await developments. Several of them frequently 
take refuge in the same hole. 

Gillichthys is the largest of these gobies. About San Diego the young are 
abundant throughout the year. The adult can be caught with hook and line in 
quantities, especially just at the return of tide during summer. Toward their 
spawning season they retire to their respective crab holes, and no morsel, how- 
ever tempting, will liu-e them forth. At San Diego they begin to spawn about the 

Fio. »6. (o) Typklagobiu! califomiensis Steind. From base of Point Loma. 
(6) TypUagobius about 25 mm. long. 

end of March. The young, when first observed, have but few color cells. They 
are very active, jumping several times their own length if left dry in a watch crystal. 
The young of this species but little resemble the adult. The maxillary does not 
reach beyond the eye, the color is in more or less well-defined crossbars, and the 
scales, which in the adult are cycloid, have several large teeth. 

Clevelandia is by far the most abundant of the gobies, and in fact the most 
abundant of any fish in the bay of San Diego. They are found everywhere between 
high and low water mark, and doubtless form an important item of the food of 
the larger fishes. They spawn in the early part of May. The young rise to the 
surface at night, and are then sometimes taken in the surface dredge. They can, 
however, be procured more abundantly in the latter part of May in the pools left 
at low tide about the piles of wharves. 

The most remarkable of the gobies is undoubtedly the blind one inhabiting 
the crab holes under rocks at Point Loma. In its pink color and general appear- 
ance it much resembles the blind fishes inhabiting the caves of southern Indiana. 
Its peculiarities are doubtless due to its habits. The entire bay region is inhabited 
by a carideoid crustacean which burrows in the mud, which, like the Wind fish, is 
pink in color. Its holes in the bay are frequented by Clevelandia, etc., while at 


the base of Point Loma, where the waves sometimes dash with great force, the 
blind fish is its associate. 

On rough days few fishes are seen, though ever so many stones are overturned. 
On mild days, on the contrary, at very low tides quantities are found almost invari- 
ably in company with one of the crustaceans mentioned above. Sometimes the 
fishes live quite out of water on the damp gravel and sand under a rock, but more 
frequently small pools of water fill all the depressions under the rocks, and the 
fishes swim rapidly away to hide in the crab holes, several of which always branch 
from the cavity in which the rock has lain. Very rarely are the fishes found 
swimming in rocky tide pools. 

In the bay the gobies habitually live outside of the holes, descending into them 
only when frightened ; but at Point Loma they rarely leave their subterranean 
abodes, and to this fact we must attribute their present condition. How long 
these fishes have lived after their present fashion it would be hard to conjecture. 
The period which would produce such decided structural changes can not be a 
brief one. The scales have entirely disappeared, the color has been reduced, the 
spinous dorsal has been greatly reduced, the eyes have become stunted, and the 
whole frontal region of the skull and the optic nerves have been profoundly changed. 

The skin, especially that of the head, has become highly sensitized. The skin 
of the snout is variously folded and puckered. The nares are situated at the end 
of a fleshy protuberance which projects well forward, just over the mouth. At 
the chin are various short tentacles, and a row of papillas (which probably bear 
sensory hairs) extends along each ramus of the lower jaw and along the margin 
of the lower limb of the preopercle. The eye is, however, the part most seriously 
affected. It is quite evident and apparently functional in the young (fig. 26 b). 
Objects thrust in front of the fish are always perceived, but the field of vision is 
quite limited. With age the skin over the eyes thickens and they are scarcely 
evident externally. As far as I could determine they do not see at this time, and 
certainly detect their food chiefly, if not altogether, by the sense of touch. A 
hungry individual will swim over meats, a fish, or a mussel, etc., intended for its 
food without perceiving it by sight or smell, but as it comes in contact with any 
part of the skin, especially that of the head region, the sluggish movements are 
instantaneously transformed, and a stroke of the fins brings the mouth immediately 
in position for operations. 

Ritter's experiments showed that it would not choose between light and dark, 
but, "On the whole, both from these observations on the living fish, and from the 
structural conditions, ... I am of the opinion that the power of perceiving 
light is not wholly lost even in the adult." 

The optic nerve is very slender and the lens proportionately very large. 

In the youngest individual caught (fig. 26 b), the membranes of the fins were 
thin, the color cells well formed and arranged not unlike those of the young Gil- 
lichthys. The movements were similar to those of the other gobies, and not at 
all sluggish like those of the adult. Their favorite position is standing or sitting 
with the broad pectorals extending out at right angles to the body. In this posi- 
tion the fish can, with a sudden stroke of its pectorals, move quickly and rapidly. 
In the old fish the fins are thick and smaller in proportion, and all the vivacity 
seems to have disappeared. The color has degenerated, or at least not developed 
in proportion to the growth of the fish. 



All these gobies are tenacious of life, especially the blind ones. Several of 
the latter have been kept in a half-gallon jar of water for several weeks without 
change of water, and others have been kept several months in confinement in my 
laboratory. When the water becomes somewhat stale, they frequently rise to the 
surface and use the water as a plane to which they attach themselves by means 
of their ventrals. The earhest date at which I procured young was October 25. 
The smallest caught at that time is represented in figure 26 b. 

The covering of the ovarian egg consists first of a finely striate membrane, the 
zona radiata of all teleostean eggs. Exterior to this is a network of threads with 
the meshes coarsest at the entodermic pole and forming almost a continuous mem- 
brane at the ectodermic pole. When the eggs are deposited, the meshwork of 
threads is stripped off the egg and remains attached to the zona radiata around 

Fig. 27. Larval Typhlogobius in its membrane. 

the micropyle. In the eggs deposited naturally by the females in confinement 
the threads were wound together to form a cord at the micropylar end of the egg. 
The cords of many of these eggs were attached to each other, and the eggs thus came 
to be laid in bunches like those of grapes. In their natural habitat the eggs are 
fastened by the threads to the lower surfaces of the rocks under which they live, 
and the membranes are expanded into long club-shaped bags. The yellow of 
the blind-fish egg is entirely confined to the yolk, which contains many oil globules. 
The granular protoplasm is opaque. In females with ripe eggs they are frequently 
to be seen forming a yellow band along the flanks. 

The eye in the larvae just about to be hatched (fig. 27) is apparently normal. 

The histology of the adult eye was studied by Ritter, who comes to the follow- 
ing conclusion : 

1. In the smallest examples studied the eyes, though very small, are distinctly visible even in 
preserved specimens — so distinctly that the lens is plainly seen. In the largest examples, on the 
other hand, they are so deeply buried in the tissue as to appear even in the living animals as mere 
black specks, while in preserved ones they are in many cases wholly invisible. 

2. Neither in small nor in large specimens does the epidermis over the eye differ in thickness or 
structure from that of adjacent regions. In the large individuals the much greater thickness of the 
tissue here is brought about by an increase in the sulvepidermal connective tissue, the growth of 
which can be seen taking place in the embryonal connective-tissue cells that are found here. 

3. As is the case with rudimentary organs generally, the eye is subject to great individual varia- 
tion in size, form, and degree of differentiation. 

4. The only parts of the normal teleostean eye, no traces of which have been found, are the 
argentea, the lamina suprachoroidea, the processus falciformis, the cones of the retina, the vitreous 
body proper, the lens capsule, and in one specimen the lens itself. 

5. In the parts present the rudimentary condition of the organ is seen in the very slight develop- 
ment of the choroid, no cellular elements being present in this excepting in the chorio-capillaris, 
and here to a quite limited extent, the rest of that layer being composed exclusively of pigment; 


in the fact that the choroid gland is composed entirely of pigment; in the fact that the iris, though 
of fully the normal thickness, is almost entirely of pigment, there being on its outer surface in some 
specimens a small amount of cellular material, which probably represents the ligamentum annulare; 
in the great proportional thickness of the pigment layer of the retina and the entire absence in it of 
anything excepting pigment; in the incomplete differentiation of the layers of the retina, there being 
in some individuals scarcely more than a trace of the external reticular layer separating the two 
nuclear layers, and there being in no specimen studied a retina sufficiently developed to enable one 
to homologize with certainty the layers marked out; in the minute size of the optic nerve, and the 
fact that it is ensheathed in a thick layer of pigment for nearly its entire course through the retina; 
and, finally, in the small size of the motores oculi. 

6. The surest evidences of actual degeneration are found, first, in the greatly augmented quantity 
of pigment in all the parts that are at all pigmented in the normal eye ; and, secondly, in the presence 
of pigment in regions where none is found in the normal eye, as in the hyaloid membrane. 

No undoubted instances of degeneration through the breaking down and dissolution of the 
tissue without the formation of pigment, such as have been described particularly by Looss, have 
been found, though in a single specimen (the one in which no lens is present) a process of this nature 
may be taking place. 


All that is known of this fish is contained in the following extract from Cope's 
paper (Proc. Acad. Nat. Sci., Phila., 1864, p. 231) : 

For a knowledge of the first genus of blind SUurid from our country, I am indebted to my friend 
Jacob Stauffer, secretary of the Linnsean Society of Lancaster, an ardent explorer of the zoology 
and botany of southern Pennsylvania, and who has furnished me with many valuable notes and 
specimens. This fish, of which specimens have been taken in the Conestoga Creek, a tributary of the 
Susquehanna, is simply a blind representative of the ordinary type of Silurids, characteristic of 
North America, and is not to be arranged with the exotic groups. * * * The color of the upper 
surfaces, tail, fins, barbels, and under jaw is black ; sides varied with dirty yellow, abdomen and 
thorax yellowish white. * * * A specimen died in 20 minutes after capture, when put in water, 
though uninjured ; the Ameiiiri, like other cat-fishes, will live for many hours after complete removal 
from their element. It is occasionally caught by fishermen, and is supposed to issue from a 
subterranean stream, said to traverse the Silurian limestone in that part of Lancaster County and 
discharge into the Conestoga. 

Two specimens of this fish present an interesting condition of the rudimental eyes. On the left 
side of both a small perforation exists in the corium, which is closed by the epidermis, representing 
a rudimental cornea; on the other the corium is complete. Here the eyeball exists as a very small 
cartilaginous sphere with thick walls, concealed by the muscles and fibrous tissue, and filled by a 
minute nucleus of pigment. On the other the sphere is larger and thinner walled, the thinnest 
portion adherent to the corneal spot above mentioned ; there is a lining of pigment. It is scarcely 
collapsed in one, in the other so closely as to give a tripodal section. Here we have an interesting 
transitional condition in one and the same animal, with regard to a peculiarity which has at the same 
time physiological and systematic significance, and is one of the comparatively few cases where the 
physiological appropriateness of a generic modification can be demonstrated. It is therefore not 
subject to the difficulty under which the advocates of natural selection labor, when necessitated to 
explain a structure as being a step in the advance toward, or in the recession from, any unknown 
modification needful to the existence of the species. In the present case observation on the species 
in a state of nature may furnish interesting results. In no specimen has a trace of anything rep- 
resenting the lens been found. 



The Amblyopsidae are a small family of fashes, first brought to the notice of 
naturalists by W. T. Craige, who presented a specimen to the Philadelphia Academy 
in 1842. De Kay, "Natural History of New York" (Reptiles and Fishes, p. 187, 
1842), gives a brief description of Amblyopsis spelcEUs. It was followed at once by 
articles by Wyman (1843 and later, 1850, 1854 a and b) and other articles by Thomp- 
son (1844) and by Telkampf (1844). Renewed interest in the Amblyopsidae was 
aroused by Agassiz's discovery of an epigean relative, Ckologaster cornutus, in the 
ditches of rice fields in South Carolina. 

Typhlichthys subterraneus was described by Girard in 1859 from a well near 
Bowling Green, Kentucky ; Ckologaster agassizii, by Putnam from a well at Leba- 
non, Tennessee, in 1872 ; Ckologaster papilliferus, by Forbes in 1882. In 1898 the 
present author described Typklicktkys rosce, and a short time afterwards he 
demonstrated that this species is generically distinct from Typklicktkys, naming 
it Troglicktkys. More recently (1905) he described Typklicktkys osborni and 
Typklicktkys wyandotte. 


The Amblyopsidae are members of the order Haplomi, first characterized by 
Cope.* They have recently been defined by Boulenger, as follows : 

Air-bladder, if present, communicating with the digestive tract by a duct. Opercle well devel- 
oped. Pectoral arch suspended from the skull ; no mesochorochoid. Fins usually without, rarely 
with a few spines; ventrals abdominal, if present. Anterior vertebrae distinct, without Weberian 

The order consists of a number of families of which the Galaxiidse and Aplochi- 
tonidae are found in the fresh waters and occasionally in the oceans of the south 
temperate zone; the Scopelidae are found pelagic and abysmal in the ocean, the 
Kneriidae in Africa, the Dalliidae in Alaska and Siberia, the Poeciliidae in fresh 
water and along the shores of the tropical and temperate zones, and the Esocidae in 
fresh waters of the north temperate zone. 

The Amblyopsidae are distinguished from the other families by the doubling 
forward of the alimentary tract, the opening of the oviduct and anus being placed 
close behind the throat, in front of the pectorals. 

The genera of the Amblyopsidae may be distinguished by the following char- 
acters : 

a. Ventral fins present; pyloric cceca 2 or 3 Amblyopsis 

aa. Ventral fins absent 

6. Eye a vestige ; pyloric cceca 2 

c. Sclera with cartilages Troglichthys 

cc. Sclera without cartilages Typhlichthys 

bb. Eye well developed ; body pigmented ; pyloric cceca 4 Ckologaster 

' Proc. Amer. Assoc. Adv. Science, Indianapolis, 1872, 328 and ;}^3. 



(A) side, (B) dorsal, and (C) ventral views. 



Amblyopsis spelseus De Kay. Plate 5. 

AmUyopsis spelmis, De Kay, Nat. Hist. N.Y., Reptiles and Fishes, 1842, p. 187, Mammoth Cave, Ky. — Wyman, 
Ann. and Mag. Nat. Hist., xii, 1843, p. 298; Amer. Jour. Sci. and Arts, xlv, 1843, PP- 94 to 96, Kentucky. — 
Thompson, Ann. and Mag. Nat. Hist., xiii, 1844, p. 112. — Telkampf, Muller's Arch., 1844, pp. 381 to 
394, taf. 9. — Wyman, Proc. Bost. Soc. Nat. Hist., iii, 1850, pp. 349 to 357. — ACASSiz, Amer. Jour. Sci. 
and Arts, xl, 1851, p. 127. — Wyman, Proc. Bost. Soc. Nat. Hist., iv, 1854, p. 39s, v, p. 18; Amer. Jour. Sci. and 
Arts, xvii, 1854, p. 258. — Poey, Mem. Cuba, ii, 1853, p. 104. — Gunther, Cat. Fishes Brit. Mus., vii, 1868, 
p. 2, Mammoth Cave, Ky. — Putnam, 1872, Amer. Nat., p. 30, fig., Lansing, Mich. (p. 20], well near Lost 
River, Ind. — Cox, Report Geol. Res. of Ind., Rhodes Cave, near Corydon; Gulf of Lost River. — Cope, 
Report Geol. Res. of Ind., iii and iv, 1871 and 1872 (1872), p. 161, Little Wyandotte Cave, Ind.; Ann. and 
Mag. Nat. Hist., 1872, Little Wyandotte Cave, Ind. — Jordan, Rept. Geol. Nat. Res. of Ind., vi, 1874 (1875), 
p. 218, Mammoth Cave. — Cope, Rept. Geol. Nat. Res. of Ind., viii, ix, x, 1876, 1877, 1878 (1878), p. 
483, Little Wyandotte Cave, Ind. — Jordan and Gilbert, Synopsis, 1883, p. 324. — Packard, Cave Fauna 
of N. A., Mem. Nat. Ac. Sci., 1886, p. 14, Hamer's and Donnelson's caves, Lawrence Co., Ind.; 
Clifty cave; Elrod's cave (p. 127), 4 miles west of Orleans, Ind. ; Mammoth Cave, Ky. — Hay, Rept. Geol. 
and Nat. Res. of Ind., xix, 1894, p. 234. — Jordan and Evermann, Fishes N. A., 1896, i, p. 706. — Blatch- 
LEY, Rept. Geol. Nat. Hist. Res. of Ind., xxi, 1896, p. 183, Sibert's well cave, a part of Little Wyandotte 
Cave, and in caves near Mitchell, Ind. — Eigenmann, Proc. Ind. Ac. Sci., 1897(1898), p. 230; Degeneration of 
the Eyes of the AmUyopsidcB, its Plans, Processes, and Causes, Proc. Ind. Ac. Sci., 1899, p. 239 (summary). — 
Eigenmann and Voder, Ear and Hearing of the Blind Fishes, Proc. Ind. Ac. Sci., 1898 (1859), P- ^4*- 
Eigenmann, Eyes of the Blind Vertebrates of N. A., Archiv f. Entwickelungsmech., viii, 1899, p. 545; Pop. 
Sci. Mo., Ivi, 1900, p. 485; Marine Biological Lectures, 1900, for 1899, p. 113. — Cox, Report Bureau of 
Fisheries, 1904, p. 392, issued 1905. 

Most of the Amblyopsidae are confined to the caves of the Mississippi drainage 
basin. Amblyopsis spelceus has the widest distribution. It is recorded from the 
following places: Mammoth Cave, Kentucky; Rhode's Cave, near Corydon; 
Lost River and one of its "Gulfs"; Elrod's Cave, Orange County; Little Wyan- 
dotte, near the southern boundary of Indiana; Hamer's and Shawnee Caves in 
Lawrence County, Indiana ; Clifty Caves, near Campellsburg, Washington County. 
Vague reports of blind fishes have come from near Milford in northern Indiana; 
from Lansing, Michigan ; and from Hiram, Ohio. None of the alleged specimens 
from the north had been preserved and none could be secured until recently, when 
I received a specimen of Amblyopsis from near Hiram, Ohio, with a letter to Prof. 
H. H. Lane, in substance as follows: 

Hiram, Ohio, July 7, 1906. 

The fish was brought by a student who resided near the place where it was found. The state- 
ment made was as follows : The township of Shalersville built a roadway of logs and earth across 
a swamp, known locally as the Podunk Swamp. The next spring the roadway sank out of sight 
and in its place there was a canal of reddish brown water. This fish was said to have been caught 
out of this water. The swamp I have occasionally visited, but have never seen any fish in the 
water. After the sinking of the road referred to the county rebuilt it at considerable expense only to 
have it sink out of sight again as before. It has not been touched since and the same stretch of water 
across it is there to-day. The swamp is one of the kind common to the glacial area and is surrounded 
by morainic hills. It was no doubt originally a lake and has been converted into a swamp by the 
growth of vegetable matter. 

This specimen makes the other northern records also probable. 

The specimens from Milford, Indiana, were reported to have been caught 
under circumstances identical with those reported for the Hiram specimen. 

This species is thus known to be distributed east of the Mississippi, both north 
and south of the Ohio River, which divides the cave region, and also far north in 
northern caves or even in glacial swamps. It is probable that it has a very wide 
distribution in the ground water. It has become quite rare in and about Mam- 
moth Cave. I have visited this cave several times, also Colossal Cavern, Cedar 
Sinks, and other caves in Kentucky, but so far have not succeeded in capturing or 
seeing any specimens south of the Ohio River. 

• This cave, plate A, has been variously called Shawnee cave, Donnelson's and Donaldson's cave. 


I have visited many caves in the Lost River region of Indiana and others have 
visited different caves without finding this species. 

Amblyopsis has been pumped out of a well at Mitchell, Indiana. I have taken 
it in only three caves; one specimen in Clifty Cave and one in Hamer's Cave. 
The only place where this species is known to be at all abundant is in the caves of 
the Donaldson farm of Indiana University. 

Troglichthys rosx Eigenmann. Plate 6, Figs. A, B, c. 

Typhlichthys suhterraneus, Garman, Bull. Mus. Comp. Zool., xvii, 1889, p. 232, wells and caves, Jasper County, 
Mo.; not of Girard. — Kohl, Rudimentare Wirbelthieraugen, 1892, p. 59. 

Typhlichthys rosce, Eigenmann, Proc. Ind. Acad. Sci., 1897 (1898), p. 231, Sarcoxie, Mo. 

Troglichthys rosce, Eigenmann, Science, N. S. ix, 1899, p. 280, Day's Cave, Sarcoxie, Mo. ; Degeneration in the 
Eyes of the AmUyopsida, its Plans, Processes and Causes, Proc. Ind. .Acad. Sci., 1898 (1899), p. 239 (sum- 
mary); Eyes of the Blind Vertebrates of N. A., Archiv f. Entwickelungsmech., viii, 1899, p. 573; A Case 
of Convergence, Proc. Ind. Acad. Sci., 1898 (1899), p. 247. — Cox, Report U. S. Bureau of Fisheries, 1904, 
p. 391 ; issued 1905. 

This species has thus far been collected by Miss R. Hoppin and by myself 
at Sarcoxie, Missouri. Miss Hoppin found it in Wilson's Cave, Day's Cave, 
Center Creek, and wells. Her reports were published in full by Mr. S. Garman. 
I found the fish in the fall of 1898, in a pool just within the mouth of Day's 
Cave. Judging from the localities where it is said they occur either in wells or 
in caves, the species is distributed over an area 300 miles long by 100 miles broad. 

It has been reported to me as occurring in wells at Cassville, Marionville, 
and Springfield, Missouri, and somewhere in Arkansas, in a spring in Newtonia, 
from a cave at Joplin, Missouri, and another near Springfield, Missouri, and from 
Turnback Cave near Marionville. A specimen from Arkansas is said to be in the 
United States National Museum. 

It is said that 7 miles southeast of Lead Hill, in the left hollow off Cane Sugar 
Orchard Creek, a half mile below an old mill, there is a cave where blind fishes 
have been found. These were described in such a way as to leave no doubt of the 
authenticity of the locality. Mr. C. H. Thompson, of the Shaw Botanic Garden 
in St. Louis, gave the following account of a cave reported to him: 

In a cave about 13 or 14 miles north of Frederickstown, St. Franjois County, Missouri, there 
is a stream of water averaging 4 to 6 feet wide and i to 3 or 4 feet deep. In these deeper "pools " 
by feeling under the rocks one will find fish which are blind. The stream does not flow out at the 
mouth of the cave, but a few rods down the slope of the hill, directly below the cave entrance, a large 
spring breaks out. This is probably the same stream as that found in the cave. The spring forms 
the source of Coldwater Creek. By consulting the map the source of Coldwater Creek, as there 
indicated, is northeast of Frederickstown. Coldwater runs in a northeast direction through .St. 
Genevieve County into the Mississippi. From the map the location of the cave is in all probability 
the extreme southeast corner of St. Franjois County. 

Typhlichthys Girard. 

The characters of the three known species of Typhlichthys are purely technical 
and may be summarized as follows : 

a. Width of head more than 6 in length to base of caudal; length of head 3§; first anal ray 

nearer base of middle caudal ray than to anus wyandotte 

aa. Width of head 5 in length to base of caudal; length of head 3 to 3.4; orbital fat-mass elongate, 
inconspicuous in life, not projecting; cheeks little swollen; eye on an average 0.16 

mm. in diameter, the smallest 0.14 mm suhterraneus 

aaa. Width of head 4.5 in length to base of caudal; length of head 3J; orbital fat-mass round and 
very conspicuous in life, projecting dome-shaped beyond contour of surrounding parts; 
cheeks much swollen; eye less than mm. in diameter osborni 



Chologaster agassizii. Dorsal, side, and ventral views. 

Troglichthys rosae. (a) dorsal ; (A) side ; (c) ventral views. 
Typhlichthys subterraneus. (J) side and (e) dorsal views. 



Typhlichthys subterraneus Girard. Plate 6, p-igs. D, e; Tejrt, fig. 28. 

Typhlichthys subterraneus, Girard, Proc. Ac. Nat. Sci. Phila., 1859, p. 62, well near Bowling Green, Ky. — 
Putnam, Amcr. Nat., vi, 1872, 17, Mammoth Cave, Ky.; Lebanon, Tenn.; Moulton, Ala. — Jordan, Kept. 
Geol. and Nat. Res. of Ind., 1874 (1875), vi, p. 218, Mammoth Cave, Ky. — Jordan and Gilbert, Synopsis 
Fishes N. A., 1883, p. 325. — Hay, Geol. and Nat. Res. of Ind., xix, 1894, p. 234. — Jordan and Everuann, 
Fishes North and Mid. Amer., i, 1896, p. 704. — Eigenmann, Eyes of the Blind Vertebrates of N. A., Archiv 
f. Entwickelungsmech., 1899, p. 545; Proc. Ind. Acad. Sci. 1898 (1899), p. 239 (summary). — Cox, 
Report Bureau of Fisheries for 1904, p. 389, 1905. 



Fio. a8. (a) Side and (b) Dorsal View of Head o( TypUichlkys iublarantus. 

Typhlichthys subterraneus Girard was discovered at Bowling Green, Kentucky, 
and later found in Mammoth Cave. For a time specimens of this species and of 
Amblyopsis found a ready market at Mammoth Cave, and this probably has had 
much to do with its later scarcity in this place. It was subsequently caught in 
other caves, to be sold at Mammoth Cave. The author has taken it in Roaring 
River of Mammoth Cave, where it was occasionally found swimming freely, but 
more often under large rocks to be brought out only by tapping the rocks with the 
net handle or one's foot. The difficulties in collecting this species (as well as other 
material) in Mammoth Cave arise from the great extent of the cave and the in- 
convenience of transporting collecting apparatus to the remote places where alone 
these fishes are to be found. 



The author has also taken this species in a small cave at the edge of the town 
of Glasgow, Kentucky, where it is moderately abundant and easily accessible, but 
on account of the limited extent of the environment very few were caught on any 
one trip. One was found under a floating board in this cave. One other speci- 
men was secured after an extensive exploration on foot and on hands and knees 
in a cave at Cave City, Kentucky. 

Typhlichthys osborni Eigenmann. 
Typhlichthys osborni, Eigenmann, Biol. Bull., \iii, p. 63, Horse Cave, Ky. 

Typhlichthys osborni is known only from Horse Cave, Kentucky. The town 
of Horse Cave is situated at the junction of two intersecting valleys. Their streams 

Fig. ag. (a) Side and (6) Dorsal View <n li 


have long ago disappeared from the surface and now flow 185 feet beneath the city. 
In the heart of the town is a sink or depression with vertical walls which was prob- 
ably caused by the falling of the roof of a large cavern. At one end of the sink an 
inclined plane leads into the underground stream, which supplies the city with 
water. The stream also furnishes the power to light the city. A dam across the 
cave furnishes the head for the power, but so modifies the conditions above it as 
to make collecting practically impossible. A convenient break in the dam made 


it possible, on one of three visits, to ascend the stream to a pile of fallen rocks from 
under which the water flows and which makes further progress impossible. This 
stretch is not great. It was noted for the abundance of blind crawfish ; no blind fishes 
were found here. On the right side of this stream, near the entrance, an older, dry 
channel leads off. At the end of the gallery a small rivulet runs to the left through 
a series of small pools separated by thin vertical partitions ; to the right it expands 
into a broad stream, quite shallow, but with such a depth of soft mud at the bot- 
tom that progress was impossible without a boat. In this expanse Typhlichlhys 
osborni was very abundant. In the fall of 1907 this cave was visited again, but no 
fishes were found where previously they had been abundant. 

Below the dam in the main cave the stream is swift and the floor so rock-strewn 
that progress is difficult and dangerous and fishing unprofitable. 

Typhlichthys wyandotte Eigenmann. 

Typhlichlhys sublerraneus, Eigenmann, Proc. Ind. Acad, Sci. 1897 (1898), p. 23c, Corydon, Ind. ; not of Girard. 
Typhlichlhys wyandotle, Eigenmann, Biol. Bull., viii, Jan., 1905, p. 63. 

Typhlichthys wyandotte is known from a single specimen from Corydon, Indi- 
ana, sent in 1886 by Superintendent Funk of the schools of Corydon to Indiana 
University. This is the only record of the genus north of the Ohio River. Repeated 
efforts to secure additional specimens have failed. 

Key to ike Chohgasters. 
a. Eye large, contained 5.5 times in the length of the head. 

b. Eye over i mm. in diameter; tactile papillae very small; sides with 3 well-defined longitu- 
dinal lines cornulus 

bb. Eye less than i mm. in diameter; tactile papillae large papilli/erus 

aa. Eye contained 10 times in the length of the head; color very faint agassizii 

Chologaster cornutus Agassiz. 

Chologaster cornutus, Agassiz, Amer. Jour. Sci. and Arts, xvi, 1853, p. 135, Ditches of rice fields at Waccama in 
S. C. — GuNTHER, Cat. Fishes Brit. Mus., vii, 1868, p. 2. — Putnam, Amer. Nat., vi, 1872, p. 30. — Jordan 
AND Gilbert, Synopsis Fishes of N. A., 1883, p. 325. — Gilbert, Bull. U. S. Fish Comm., viii, 1888, p. 22, 
Okefinokee Swamp, Millen, Ga. — Jordan and Evermann, Fishes North and Mid. Amer., i, 1896, p. 
703. — Eigenmann, Degeneration of Eyes of Amblyopsidse, its Plans, Processes, and Causes, Proc. Ind. 
.Acad. Sci., 189S, p. 239 (summary); Eyes of the Blind Vertebrates of N. Amer., Archiv f. Entwickelungsmech., 
viii, 18S9, p. 543; Marine Biological Lectures, 1899 (1900), p. 113. 

Chologasler avilus, Jordan and Jenkins, in Jordan Proc. U. S. Nat. Mus., viii, 1888, p. 356, pi. 44, fig. 8, Outlet 
of Lake Drummond, Dismal Swamp, near Suffolk, Va. — Cox, RefKjrt Bureau of Fisheries for 1904, p. 
386 (issued 1905). 

The Chologasters have a wide and discontinuous distribution. Chologaster 
cornutus Agassiz has been found in the ditches of rice fields in South Carolina ; in 
the Okefinokee Swamp at Millen, Georgia ; and in the Jericho Canal near Suffolk, 
Virginia, in an outlet of Lake Drummond. Its range is entirely east of the Alle- 
ghany Mountains, and it is found in lowland streams only. I visited the locality 
near Suffolk, but found no specimens. 

Chologaster papillifenis Forbes. Plate i, Fig. A. 

Chologasler papilli/erus, Forbes, Amer. Nat., March, 1881, and Jan., 1882, Cave spring in southern Illinois. — 
Jordan and Gilbert, Synopsis Fishes N. A., 1883, pp. 325, 890. — Jordan and Evermann, Fishes 
North and Mid. Amer., i, 1896, p. 704. — Eigenmann, Proc. Ind. Acad. Sci., 1897 (1898), p. 231; 
Degeneration in the Eyes of the Amblyopsidae, its Plans, Processes, and Causes, Proc. Ind. Acad. Sci., 
1898, p. 239 (summary); Eyes of the Blind Vertebrates of N. A., Archiv f. Entwickelungsmech., 1899, 
p. S4S; Marine Biological Lectures, 1899 (1900), p. 113. 

Chologaster papilli/erus Forbes is known only from cave springs in Clinton and 
Jackson Counties, Illinois. Most of the specimens have come from a spring in 
Jackson County, Illinois. 


Chologaster agassizii Putnam. Plate 6. 

Chologaster agassizii, Putnam, Amer. Nat., vi, 1872, p. 22, well at Lebanon, Tenn. ; Mammoth Cave, Ky. — 
Jordan, Kept. Geol. Nat. Res. of Ind., vi, 1874 (1875), p. 218 (reference to Putnam's specimens). — Hay, 
Geol. and Nat. Res. of Ind., xix, 1894, p. 234. — Jordan and Evermann, Fishes North and Mid. Amer., 1896, 
I, p. 704. — Eigenm.\nn, Proc. Ind. Acad. Sci., 1897 (1898), p. 230; Eyes of the Blind Vertebrates of N. A., 
Archiv f. Entvvickelungsmech., VIII, 1899, P- 54^! Proc. Ind. Acad. Sci., 1898 (1899), pp. 239, 251; Marine 
Biological Lectures, 1899 (1900), p. 113. 

Chologaster agassizii Putnam is known only from Lebanon, Tennessee, and 
caves about Mammoth Cave. I have taken it in the Styx in Mammoth Cave and 
in Cedar Sinks, near Mammoth. I found the Chologaster in only one locality in 
Mammoth Cave. A short distance after descending the corkscrew the Styx appears 
on the right. On one visit Chologaster was abundant in and around the remains 
of an old boat, but I secured only a few small specimens on account of their agility 
and the easily roiled water. They were much more alert than the blind members of 
the family and made quickly for the lower edge of the wall of the cave, below 
which many of them escaped. On a subsequent visit the locality had been quite 
modified, and I secured even fewer specimens than before. 

Cedar Sinks, the other locality from which I secured Chologaster agassizii, is a 
highly interesting region. It lies several miles from Mammoth Cave and is 
reached over a rough road leading, without modifications from the engineer, up and 
down the steep slopes of the interminable sink holes of the region. Cedar Sinks 
was formed by the caving in of the roof of an enormous cave room. The vertical 
walls of the room are still standing. I have been told the bottom of the sink 
embraces about 4 acres. In the bottom are 2 funnel-shaped depressions holding 
water. The walls of the funnels are so steep that it is just possible to climb out if 
one has been foolish enough to slide down. At the base of the highest rock bound- 
ing the sink are two openings. One leads to an extensive underground stream 
which can be followed a very restricted distance ; the other, to a stream and cave 
which must be quite extensive, judging from the inflow of air at the time of one of 
my visits. Small pools or streams in one of the entrance galleries yielded a few 
specimens of Chologaster. 


The three species of Chologaster are colored with varying intensity from C. 
cornutus, which is darkest, to C. agassizii in Mammoth Cave, in which the 
color is faintest. The color cells are in all cases arranged in a definite pattern. 
This is determined by the underlying muscles. The pattern consists of three lon- 
gitudinal bands on the sides following the line where the muscle segments are 
angularly bent and cross stripes along the line separating successive segments 
(plate 6, upper figures). 

The lower side of the head and the abdomen of Chologaster papilliferus are 
sparingly pigmented and translucent. The underlying liver and gills give the 
parts a rosy tinge. On the sides and top of the head pigment is abundant. There 
is a more densely pigmented area extending along the middle of the back, l^egin- 
ning as a narrow stripe at the nape and widening to the dorsal fin behind, where it 
occupies the entire back. On the sides are 3 narrow stripes, which, owing to the 
accumulation of pigment in 2 layers, are quite dark. Each stripe has a lighter 
central band, widest at the middle of the sides. A light band, without the con- 
spicuous bordering dark stripes, runs along the middle of the belly. The sides are 
thickly covered with a layer of pigment, leaving usually colorless lines where con- 


nective tissue separates successive myotomes. On the sides of the tail the pigment 
is dense on either side of these colorless lines. A dark band extends along the sides 
of the head through the eye. The top of the head is dark (plate i, fig. a). 

The pattern of Chologaster cornutus agrees with that of C. papilli/erus. The 
longitudinal bands are much darker and wider and without the light central 
streak. The middle band is much wider than the others and is continued forward 
to the tip of the snout. The amount of color present varies very greatly with the 
locality from which the specimens come. 

The general color of C. agassizii is light gray (plate 6, upper figures). The 
scales are lighter than the area surrounding them. The color pattern is more striking 
than in the other species of the genus. Each somite is bordered by a dark line. The 
lines of successive somites are separated by an almost imperceptible colorless line. 
A broad, not sharply defined, band extends along the sides. The middle of this is 
lighter than the margin. Another one extends between the somites and the ven- 
tral musculature, another from the nape between the lateral somites and the dorsal 
muscles, and a diverging one from near the nape to either side of the dorsal fin. 
Dark areas are caused by the accumulation of pigment along the borders of the 
small muscles of the fins. Still another dark area is found about the caudal. The 
ventral surface is white, except the accumulation of pigment along the lines sepa- 
rating the muscles. The fins are uniformly light gray. A light area is found on 
both the upper and lower part of the caudal peduncle, just within the first short 
rays of the caudal. 

The general color of Typhlichthys is cream and pink. It is abundantly pig- 
mented. In younger specimens the pigment is arranged in more definite areas 
about the head. In the old it is more uniformly distributed, being, however, spe- 
cially abundant about the brain. The pigment pattern of the body is precisely as 
in Chologaster except that the individual pigment cells are minute and their aggre- 
gate not evident except under the lens. 

The retention of the color pattern of Chologaster in Typhlichthys is not less 
interesting than the retention of similar habits. It is perhaps due to diff'erent 
causes. The color pattern in Chologaster is determined by the underlying mus- 
cular structure and the retention of a similar pattern in Typhlichthys is due to the 
same underlying structure rather than to the direct hereditary repetition of the 
color pattern. In Amhlyopsis the color is much less marked than in Typhlichthys. 
Amblyopsis is flesh-colored, ranging to purple in the gill-region, where the blood of 
the gills can be seen through the overlying structures, and over the liver, which can 
he seen through the translucent sides and ventral wall. About the head and bases 
of the fins the color is yellowish, resembling diluted blood. The surface of the body 
is slightly iridescent and that of the head has a velvety, peach bloom appearance. 

The general pink color of Amblyopsis is due to the blood, not to any abnormal 
development of blood-vessels in the dermis. In the fins where the blood-vessels 
are near the surface, the general eff'ect is a yellowish color. The surface vessels of 
the dermis also appear yellowish. It is only on account of the translucent condi- 
tion of all the ti.ssues, permitting the deeper vessels to show through a certain thick- 
ness, that the pink effect is produced. Amblyopsis has always been spoken of as 
white. The term "white aquatic ghosts" of Cope is very apt, for they do appear 
white in the caves and their gliding motion has an uncanny effect. All alcoholic 
specimens are white. 


The chromatophores in Amhlyopsis are dififerentiated and contain color before 
the yolk is absorbed. The black chromatophores are minute granules, few (15 
or thereabout) to the segment. In an older larva the pigment was much more 
abundant. The eyes are pigmented early, shortly before hatching, and, owing to 
their pigment, they soon become conspicuous and remain so till the fish has reached 
50 mm. in length, when the overlying tissues have become thick. The pigment 
of the body is lost, or, what amounts to the same thing, does not increase much with 
age. There is an abundance of pigment cells in the adult, but they are very poor 
in pigment, and, being in the dermis and covered by the thick layer of epidermis 
rich in glands, are not apparent. Pigment cells are also abundant in deeper tis- 
sues in the adult, so that, while no pigment is visible on the surface, an abundance 
of chromatophores is present in deeper tissues. 

The pigment cells can not be made to show themselves, i.e. become greatly 
pigmented, even by a prolonged stay in the light. The old, if kept in the light, will 
not become darker ; and a young one reared in the light until ten months old not 
only showed no increase in the pigmentation, but lost its pigment, taking on 
the exact pigmentless coloration of the adult. Pigment cells appear late in 
Amblyopsis. When the young are two months old pigment is abundant. This 
pigmented condition is evidently a hereditarily transmitted condition. It disappears 
with age. In the first instance this disappearance was probably individual. But as 
in the flounder, the depigmentation has also become hereditarily transmitted, for 
even those individuals reared in the light lose the color. 

Numerous facts and experiments show that, while pigment may be and is devel- 
oped in total darkness, the amount of color in an individual animal depends, other 
things equal, directly on the amount of light to which it is habitually exposed. 

A number of apparently contradictory observations may be noted : 

(a) The absence of pigment in pelagic animals or their larvae, which depend on 
their colorless condition for their existence, is evidently due to causes entirely dif- 
ferent from those preventing the formation of pigment in cave animals. Natural 
selection has, in pelagic animals, eliminated the color. 

{b) The migration of pigment granules due to temperature or light and the 
expanding of chromatophores, when an animal is over a dark background or in 
the dark, and the contracting over a light background, which may take place at 
once or at the expiration of several days, is evidently also a different question. 
The observations of Cunningham, Agassiz, and Semper along this line are of interest. 

(c) Fischel (A. M. Anat., vol. xlvii, pp. 719-734, plate xxxvi, 1893) has 
noticed that larvae of salamanders reared in water at 6° to 7° are dark, while others 
kept in water from 15° to 58° are light. 

{d) Flemming (A. M. Anat., vol. xlviii, pp. 369-374, 1896) found that with 
uniform temperature in two vessels side by side, the one dark, the other light, the 
salamander larvae in the dark vessel develop pigment cells rich in color granules ; 
the larvae in the white vessels become pale, although the number and character of 
the pigment cells is not otherwise changed. The difference is entirely due to the 
character of the vessels, for if the larvae are taken from the dark to the light vessel, 
they become light-colored in a few days. 

(e) Semper ("Animal Life," p. 89) records that " * * * in the tadpoles of our 
common toads and frogs the pigment is equally weU developed in yellow, blue, or 


red light, and in absolute darkness." This was to be expected, for even in the 
young of cave animals pigment is, as a rule, well developed. 

(/) Pouchct (Arch, dc Physiol, et d'Anat., 1876, and Rev. Scient., vol. xiii, 1897) 
has demonstrated that change in color cells, such as are mentioned under (b) and (d), 
is brought about by the reflex control of the eye. The section of the great sympa- 
thetic nerve puts an end to the changes of color under the influence of light. 

The lower and upper surfaces of the flounder, the one protected and the other 
exposed to the light, give the most striking example, and the argument is clinched 
here by the fact, noted by Cunningham and McMann, that a flounder whose lower 
side is for long periods exposed to the light takes on color. Loeb has shown that in 
the yolk sacs of Fundulus embryos more pigment cells are developed if the embryos 
are kept in the light than when they are kept in the dark. However, in the body, 
and especially in the eye, the pigmentation was not affected by the absence of light. 

The general absence of color in cave animals is conceded. Packard states "as 
regards change of color, we do not recall an exception to the general rule that all 
cave animals are either colorless or nearly white, or as in the case of Arachnida and 
insects, much paler than their out-of-door relatives." Chilton has made the same 
observation on the underground animals of New Zealand. Similar observations 
have been recorded by Lonnberg, Carpenter, Schmeil, and Vire. 

Hamann enumerates a number of species living both in caves and above ground. 
In such cases the underground individuals are paler than the others. This confirms 
similar observations of Packard. 

Poulton has mentioned that Proteus becomes darker when exposed to the light. 
This has been verified by others. In Typhlotriton, larvae living at the entrance of 
a cave are dark, while the adult living farther in are much lighter, but with many 
chromatophorcs containing a small amount of color. Epigean fishes found in 
caves are always lighter in color than their confreres outside. 

We have thus numerous examples of colored epigean animals bleaching in 
caves, and also bleached cave animals turning dark when exposed to the light. We 
have also animals in which the side habitually turned to the dark is colorless, while 
the side habitually turned to the light is colored. Finally we have cave animals 
that are permanently bleached. 

Natural selection can not have affected the (foloration of the cave forms, for it 
can be of no consequence whether a cave species is white or black. It could only 
affect the coloration indirectly in one of two ways : first, as a matter of economy, but 
since the individual is in part bleached by the direct effect of the darkness, there is 
no reason why natural selection should come into play at all in reducing the pig- 
ment as a matter of economy; second, Romanes has supposed that the color 
disappeared through the selection of correlated structures, a supposition he found 
scarcely conceivable when the variety of animals showing the bleached condition 
was considered. 

Panmixia can not account for the discharge of the color, since it returns in some 
species when they are exposed to the light and disappears to a certain extent in 
others when kept in the dark. Panmixia, Romanes thinks, may have helped to 
discharge the color. In many instances the coloration is a protective adaptation, 
and therefore maintained by selection. Panmixia might in such instances lower 
the general average to what has been termed the " birth mean." Proteus is perhaps 


such an instance. But in this species the bleached condition has not yet been 
hereditarily estabhshed, and since each individual is independently affected, "the 
main cause of change must have been of that direct order which we understand by 
the term climatic." 

Since, however, the bleached condition, which in the first instance is an individual 
reaction to the absence of light, has become hereditarily established in Amblyopsis 
so that the bleaching goes on even when the young are reared in the light, it is evi- 
dent that in Amblyopsis we have the direct effect of the environment on the individual 
hereditarily established. 


The general impression given by Amblyopsis is that of a skinned cat-fish swim- 
ming on its back. The largest individual secured by mc measured 135 mm. in 
total length. Individuals as large as this are rare. The usual length of an adult 
is about 90 mm. At Mammoth Cave I was told of an individual having a length 
of 200 mm. 

Amblyopsis is found in pools in the cave streams it inhabits. I have secured 
as many as 12 from a pool perhaps 10 by 50 feet in size. Very rarely they are to 
be found in the riffles connecting the pools. I have seen them lying at the bottom, 
or swimming, rather gliding, through the water like "white aquatic ghosts." In 
the aquarium they lie at the bottom or at various depths in the water, their axes 
making various angles with the horizontal, their pectorals folded to their sides. 
When swimming slowly, it is chiefly by the use of the pectorals. The strokes of 
the pectoral are lazily given, and the fish glides on after a stroke till its impetus is 
exhausted, when another stroke is delivered. The fishes frequently roll slightly from 
side to side at the exhaustion of the result of a stroke. When swimming rapidly, 
the pectorals are folded to the sides and the locomotion is then similar to that of 
a salamander, by the motion of the tail. They readily adjust themselves to differ- 
ent depths and are usually perfect philosophers, quiet, dignified, unconcerned, and 
unperturbed, entirely different from such eyed species as minnows and sun-fishes, 
which are sometimes found in caves, and which are much more readily disturljed by 
any motion in the water, making it almost impossible to capture them. The pec- 
torals are also almost exclusively used when quietly rising in the water. At such 
times the pectorals are extended laterally and then pressed to the sides, beginning 
with the upper rays. A downward stroke is delivered in this way, not quickly, but 
with apparent lazy deliberation. In swimming forward the pectorals are brought 
forward, upper edge foremost. The center of gravity seems to be so placed in 
regard to their various axes that the fish does not lose its balance whatever its posi- 
tion. It floats horizontally in the water without any apparent effort to maintain 
its position. It floats with the main axis inclined upward, with the snout some- 
times touching the surface of the water, apparently lifeless. Once one was seen 
resting on its tail in a nearly vertical position, and one while quietly swimming 
leisurely turned a somersault and swam on undisturbed. At another time the 
same individual rolled completely over. When one is kept out of water for a short 
time, it frequently goes in a corkscrew-shaped path through the water, continually 
spinning around its long axis. In their quiet floating position it is difficult to deter- 
mine whether they are alive or not. 




The number of respiratory movements of Amblyopsis averaged 19 a minute in 5 
observations, reaching a maximum of 30 in a small individual and a minimum of 
14 in a large one. This is in strong contrast to Chologaster, the number of whose 
respiratory motions reached an average of 80 per minute in 5 observations, with a 
minimum of 56 and a maximum of 108 in a small specimen. Loeb has called my 
attention to the more rapid absorption of oxygen in the light than in the dark ; this 
extended would probably mean the more rapid absorption of oxygen through the 
skin of light-colored animals, a matter of doubtful value, however, to species living 
in the dark. 

The gill filaments are small as compared with the gill-cavity. In addition to 
the oxygenation through the gills, oxygenation probably takes place through the 
skin. Ritter has suggested the same for Typhlogobius. 

Cutaneous respiration is not unique in Typhlogobius and the Amblyopsidae. In 
the viviparous fishes of California the general surface, and especially the fins which 
have become enormously enlarged, serve as respiratory organs during the middle 
and later periods of gestation. The fins are a mass of blood-vessels with merely 
sufficient cellular substance to knit them together. There is, however, no pink 

Skin respiration would account for the extreme resistance to asphyxiation in 
Amblyopsis and Typhlogobius. About 45 examples of Amblyopsis were carried in 
a pail of water 400 miles by rail with only a partial change of water 3 times during 
24 hours. A smaller number may be kept for days or weeks — probably indefi- 
nitely — in a pail of water without change. The characteristics of Typhlogobius 
along this line have been set forth elsewhere. 


The first speculations on the feeding habit of Amblyopsis are those of Cope. 
He remarks : 

The projecting lower jaw and upward direction of the mouth render it easy for the fish to feed 
at the surface of the water, where it must obtain much of its food. This structure also probably 
explains the fact of its being the sole representative of the fishes in subterranean waters. No doubt 
many other forms were carried into the caverns since the waters first found their way there, but 
most of them were like those of our present rivers, deep-water or bottom feeders. Such fishes would 
starve in a cave river, where much of the food is carried to them on the surface of the stream. 

The speculations of Cope are entirely erroneous as pointed out by Putnam, and 
we shall see that the deductions based on them naturally fall to the ground. 
Dr. Sloan recorded one Amblyopsis which he kept 20 months without food. 

Some of them would strike eagerly at any small body thrown in the water near them, rarely 
missed it, and in a very short time ejected it from their mouths with considerable force. I tried 
to feed them often with bits of meat and fish worms, but they retained nothing. On one occasion 
I missed a small one and found his tail projecting from the mouth of a larger one. 

Wyman also found a small-eyed fish in the stomach of an Amblyopsis. 

Hoppin was also struck by the fact that if not capable of long fasts, Typhlich- 
thys (Troglichthys) must live on very small organisms that the unaided eye can not 
discern. Garman found in the stomachs of Troglichthys, collected by Hoppin in 
Missouri, species of Asellus, Cambarus, C^iUhophUus, and Crangonyx. 


All the specimens of Amblyopsis from the Mitchell Caves so far examined by 
me contained very large fatty bodies, a condition suggesting abundance of food. 
The stomachs, as far as examined, always contained the debris of Gammarus. 
One young Amblyopsis disappeared on the way home from the caves and had evi 
dently been swallowed by one of the larger fish. 

The young Amblyopsis reared in the aquarium seemed to feed on the minute 
forms found in the mud at the bottom of its aquarium. Some Ccecidotea placed in 
its aquarium soon disappeared, and the capture of one of these was noted under a 
reading glass. The fish was quietly swimming along the side of its aquarium; 
when it came within about an inch of the crustacean it became alert, and with the 
next move of the Cacidotea it was captured with a very quick, well-aimed dart on 
the part of the young fish. Others were captured while crawling along the floor of 
the aquarium. 

Mr. Fernandus Payne has made extensive observations and experiments on the 
feeding habits of this fish, and his notes follow : 

The following experiments have been made to determine what the blind fishes eat and more 
especially how they detect the presence of their food. Incidentally some correlated reactions have 
been observed. 

In the laboratory, after the fishes had become accustomed to their new conditions, I had no 
trouble in getting them to eat isopods, amphipods, young crawfish, and diptera and salamander 
larvae. They will also take meat from the end of a thread when it is moved about in front of them 
or brought in contact with the body. The meat, in nearly all cases, is ejected either before or after 
it has been swallowed. From these observations it seems that they will eat any small animal which 
moves about in the water. According to my experiments they prefer amphipods. This may be due 
to the fact that they are more active than the other animals, and hence their presence is more easily 
detected. If isopods and amphipods are placed in the same aquarium, the amphipods are eaten 
first. The young of 25 cm. in length readily eat cyclops, daphnids, and aquatic fly larvae. I have 
seen them eat fly larvae until the stomach was so full that they had difficulty in keeping the larvae 
from wriggling out again. 

In the caves both variety and quantity of food are limited. Crangonyx and Ccscidotea appear in 
considerable numbers, but most of them seem to stay under rocks in running water while the fishes 
are confined to the quiet pools. Young crawfish are certainly not plentiful, for the adults are not 
very numerous. Whether Cyclops or any other small Crustacea are present in any abundance, I 
am unable to say. I know of nothing else on which the young blind fishes could feed. In July, 
1906, 1 took a number of young from the gill cavity of the mother, put them into a box made of cheese- 
cloth and sunk the box in a quiet pool of water in the cave. They remained in this place for about a 
month and were growing nicely. I have no doubt they would have lived here much longer had not 
the cloth become full of holes and freed them. I put no food into the box, so they must have eaten 
the small organisms in the water. 

Fishes must find their food either by the sense of sight, taste, touch, or smell, or by a combination 
of two or more of these. In most fishes sight undoubtedly plays the predominant part in locat- 
ing and seizing food. This factor is excluded in Amblyopsis. Herrick has given some excellent 
experiments bearing upon this question. He finds that practically the whole cutaneous surface of 
Atneiurns is sensitive to both tactile and gustatory stimuli, but that the gustatory stimuli are of the 
greatest value to the cat-fish in procuring food. The hake (Urophycis tenuis) catches its food by 
sight, only when the food is in motion. Bits of meat, etc., lying on the bottom are usually found by 
the aid of the free ventral fins. From these and other experiments, Herrick concludes that the hake 
receives both tactile and gustatory stimuU by means of the free fin-rays and to some extent, doubtless, 
by other parts of the outer body surface. He was unable to determine whether or not smell played 
any part. When food is thrown into the aquarium the tomcod {Microgadus lomcod) catches it while 
it is falling through the water. The ventral fins are used in locating sapid substances lying on the 
bottom. He cut the olfactory nerves to see whether smell played any part in the detection of food 
and found that the fishes with the nerves cut acted in every respect like normal fishes. The sea 



robin {Prionotus carolinus), according to Morrill and Herrick, finds its food largely by sight and by 
the use of the free pectoral fin-rays which are tactile in function. The king-fish {Menticirrhus 
saxatUis) uses sight somewhat, but in the main the tactile organs are used as most of the food was 
taken by contact, and non-nutritious substances were generally taken. The toad-fish (Opsanus 
an) did not find concealed bait and seemed to get its food wholly by the visual and tactile senses. 

Herrick concludes from his experiments that fishes which possess terminal buds in the outer 
skin taste by means of these organs and habitually find their food by their use. Fishes which lack 
these organs in the skin have the sense of taste confined to the mouth. The delicacy of the sense of 
taste in different parts is directly proportional to the number of terminal buds in these areas. 

A mhlyopsis h;is terminal buds scattered over the entire head. They are most numerous on the 
lips and the tip of the snout. I did not determine whether or not they were present on other parts 
of the body. My experiments indicate that, if they are present on parts other than the head, they 
are but few in number. While these fishes are without doubt able to taste with the buds on the 
lips and snout, practically all of their food is found by means of the tactile sense. I am unable to 
say how the terminal buds compare in number with those of other fishes. The young fishes up to 
20 mm. in length do not have terminal buds developed. Since this is the case they have only the 
tactile sense for finding food, for smell plays only a minor part, if any. 

Ritter says that in Typhlogobius the tactile sense has not only not increased, but has actually 
diminished pari passu with the diminution of the power of sight. Such is certainly not the case in 
AmUyopsis. Eigenmann says: 

"(i) The eyes were degenerating and the tactile organs developing beyond the normal before 
the permanent underground existence began. 

" (2) The eyes continued to degenerate and the tactile organs to increase after the permanent 
entrance to underground waters." 

The tactile organs are arranged in rows or ridges. An examination of the number of individual 
tactile organs in the same 3 ridges in each of 8 fishes gives the following counts: 

Length of specimen. 

Number of organs in ridges. 

Length of specimen. 

Number of organs in ridges. 


I a 3 

I I I 


14 14 10 

16 16 10 





I a 3 
16 16 10 
21 26 18 
47 42 22 
42 40 28 

Whether the individual tactile organ is more highly developed in the adults than in the young 
would be difficult to say. At any rate the above figures show that in the adults tactile organs are 
much more numerous than they are in the young. 

Since, in the blind fishes, the factor of sight is entirely eliminated, we have left the senses of 
taste, touch, and smell by which they may find their food. In testing which of these is concerned 
I used about 50 individuals. My best results were obtained by placing the fishes in battery jars 
(one in each jar) 7 inches in diameter and 8 inches high. The water in these jars was from 4 to 
5 inches in depth. This enabled me to eliminate all factors except those which I introduced. 

During the summer of 1906 I kept a number of fishes in an aquarium in the cave. I tried to get 
them to eat meat, but had no success. In September of the same year, I transferred other fishes 
to the laboratory at Bloomington, where my experiments were made. Some individuals begin to 
eat in a few days, others not till several weeks after they are confined. The young, from 25 to 40 
mm. in length, and the adults seem to become adjusted to their new conditions much more readily 
than those about half-grown. Those from 60 to 70 mm. in length are much more sensitive to me- 
chanical stimuli than either the young or adults, and further their sense of fear seems more highly 
developed at this time. 

When first brought into the laboratory, I kept the fishes in a dark room so as to have the con- 
ditions as nearly normal as possible. As this necessitated a light while making the observations, 
I abandoned it for a lighted room, but several observations were made in the dark room, where 
I often tried to feed them meat from the end of a thread. After 4 weeks, I got some of the larger 
ones to take a few pieces, and one large fish took 5 pieces in as many minutes. The next day I 
found the meat lying on the bottom of the aquarium. In no case did I get the fishes to take meat 


before it came in contact with the lips. Once when I was feeding them meat, the thread touched 
the lips of one of the fishes. It immediately snapped at the thread. Before I could bring the meat 
in contact with the lips, it snapped at the thread a second time. This seemed to indicate that they 
do not readily distinguish between edible and non-edible substances. Again, I lowered a pair of 
forceps into an aquarium where there were 15 fishes. All of them were attracted by the disturbance 
in the water. Fishes 18 inches away turned and swam in the direction of the forceps. I kept the 
forceps moving just enough to create slight vibrations in the water. Every fish came up and snapped 
at the forceps and some of them snapped 2 and 3 times. At the least disturbance on the surface 
of the water these fishes would swim upward as if expecting something to eat. They are able to 
follow the disturbance anywhere about the aquarium and do it quickly and accurately, turning at 
any sort of an angle. This experiment was made with non-edible objects, so taste and smell could 
have played no part whatever. The tactile organs remain as the only means by which these vibra- 
tions were detected, located, and followed. 

I suspended pieces of fresh beef in the aquarium to see whether the fishes were able to locate 
it while it remained stationary, but in no case did they pay any attention to the beef. After I had 
placed the fishes in the light in individual jars, I had no trouble to get the larger ones to take meat 
from the end of a thread. I also fastened bits of absorbent cotton to the thread as I had done with 
the meat, and at first they took the cotton just as readily as they had the meat. The cotton was not 
swallowed, but ejected as soon as taken into the mouth. The fishes turned if any part of the body 
was touched, but never snapped until the cotton or meat came in contact with the lips. How- 
ever, after a few trials with the cotton the snap was not so vigorous, and if continued, the cotton was 
refused altogether. In the course of 4 hours, I got one fish to take the cotton 11 times, but after 
that it seemed to be able to perceive the difference and though I kept this individual several months 
no amount of persuasion could induce it to take another piece of cotton. After this it acted toward 
the cotton just as it did toward the beef until the cotton came in contact with the lips, when it would 
refuse it. I did get it to take cotton soaked in beef juice. I tried 2 fishes by placing bits of meat on 
the bottom of the aquarium. In swimming close to the bottom, the meat touched the ventral sur- 
face of the body or the pectoral fins. In each case, the fishes stopped, backed up a little until the 
lips touched the meat, and then snapped at it. This seems to indicate that this species might, in 
some cases, take food which was not in motion and that it might locate its food partly by taste. 
I tried these fishes with cotton in the same manner as I had done with the meat, and they reacted in 
exactly the same way until the cotton touched the lips, when they refused to take it. One fish did 
snap up one piece of cotton. 

I also tested their ability to taste by squirting beef juice on various parts of the body. I got no 
reaction that I could not get with pure water. I dropped beef juice, a drop at a time, on the surface 
of the water. The fishes were attracted by the vibrations, came to the surface, and snapped at the 
drop, but they also reacted in the same manner toward drops of water. They are not able to locate 
the center of disturbance as readily when the drop falls behind them as when it falls on the side or 
in front. This experiment again shows how sensitive these fishes are to vibrations in the water and 
how accurately they are able to locate them. 

I might add that slight disturbances, such as the dropping of amphipods into the water, often 
cause the fishes to sink gradually to the bottom and remain quiet for several seconds, after which 
they begin to swim slowly about. At this time the swimming is accomplished mostly by the use of 
the pectoral fins. By a backward stroke, the fins are brought against the body, and then, as the fish 
glides forward, they are allowed to float out at right angles to the body, the filamentous edge dragging 
on the bottom. We might term this the "seeking reaction." Amphipods which touch the fins 
or any other part of the body at this time are snapped up immediately. 

I mentioned before that the fishes were confined to the quiet pools. It seems to me that their 
manner of getting food accounts, in part, for their habitat. They eat living animals, and these ani- 
mals are found by the vibrations which they make in swimming. In running water the fishes could 
not detect these vibrations. 

A few observations on the memory of Amblyopsis may be placed on record in this connection. 
When the fishes are first brought into the laboratory, they are very sensitive to mechanical stimuli. 
If kept in a place where they are constantly subjected to stimuli, they soon pay much less attention 
to them. I kept some fishes in battery jars on my table. At first, when I struck the table lightly, 
they always responded by springing upward. After a few weeks they responded much less often, 
and after several months they paid very little attention to jarring of any kind. 


It was mentioned before that one fish, in the course of 4 hours, took 1 1 pieces of cotton from the 
end of a thread and after that refused to take it again, although the fish was kept for several months. 
In this case, then, it learned to discriminate within a very short time, and remembered the difference 
between the cotton and the meat. It took the meat, if brought in contact with the lips, after it 
refused the cotton. 

Another fish was tested by dropping water on the surface of the aquarium. The fish came to 
the surface and grabbed at the drop. I tested the fish once everyday for 12 days, and on the twelfth 
day it refused to grab, but came up near the surface, poised as if ready to grab, and then sank 
slowly toward the bottom. The thirteenth day it responded, but not very readily. For the next 
8 days I tested it every day and got no attempt at grabbing, although it came near the surface every 
day. I did not test it again for 3 days, when it again snapped at the drop. It came up to the sur- 
face at the first few drops, but sank gradually toward the bottom. Upon continuation of the drop- 
ping, it came up again and grabbed. I then left it undisturbed for 5 days before testing and again 
it grabbed. This was the twenty-ninth day of the experiment. I then started with an interval of 
I day and increased it by i day each time, thus making the intervals i, 2, 3, 4, and 5 days. It 
did not snap at the drop until after the interval of 5 days. This was the forty-fourth day of the 
experiment. I again waited 5 days before testing the fish and got no response further than that 
the fish came near the surface. On account of the lack of time the experiments were discontinued. 
Whether the fish would eventually have learned not to snap at the drop, I can not say, but that 
memory plays some part in its reactions is evident from my observations. 

The conclusions reached are as follows: 

(i) Sight is as a matter of course excluded from food seeking. 

(2) The olfactory sense, if any, plays a very minor part in detecting food. 

(3) The sense of taste enables them to discriminate between things in contact with the snout. 

(4) The tactile sense is the one by which they find and precisely locate their food. 


The following extract, from a letter from Mr. E. B. Forbes, is of interest: 

Doubtless you have received the little Chologaster which I sent you yesterday. The spring in 
which they are found is in an almost inaccessible part of Jackson County and I drove 1 7 miles from 
Cobden, Illinois, in a wagon to this place. The spring is a very large one, flowing from the bottom 
of a 250-foot cliff of flint and limestone. The little fishes were found under stones at the edges of 
the spring, very close to the bluffs, and when disturbed they swam back under the cliff. After the 
rough drive home they were still alive and seemed vigorous when handed over to the expressman. 
I found this species in other springs than the large one mentioned and have no doubt that it is rather 
widely distributed. None were found at any considerable distance from the face of the cliff. 

I found that Chologaster agassizii acts similarly in the River Styx in Mammoth 
Cave. As soon as my net touched the vi^ater they darted in under the ledge of 
rock at the side of the little pool in which I found them. 

The Chologaster in general make-up is like Amblyopsis, but somewhat more 
elongate. It sits with its pectorals extended. When it moves horizontally for some 
distance the pectorals are usually pressed to the sides, the propelling being done 
largely by the tail, very much after the manner of a salamander, which it resembles. 
In swimming toward the surface it uses its pectoral fins chiefly, and the fish usually 
sinks to the bottom as soon as its efforts to raise itself are stopped. 

Individuals kept in aquaria with one end darkened either collected in the dark- 
ened area, floating about, or under leaves or sticks in any part of the aquarium. 
They are frequently found under a floating board where they float with the tops 
of their heads in contact with the board, their bodies slanting downward. 

Typhlichthys, living in total darkness, has retained the habit of staying under 
floating boards and sticks and under stones. Miss Hoppin noticed that Trog- 
lichthys swims with its back to the aquarium, and I have repeatedly noted the 
same in the young of Amblyopsis up to 50 mm., and the still younger Amblyopsis 
frequently hides under rocks. 


Chologaster papilliferus detects its food entirely by the sense of touch. Two 
which were kept in an aquarium for over a year were starved for a few days. They 
became very nervous, continually swimming along the sides of the aquarium. Some 
individuals of Asellus were introduced. These, though quite near, produced no 
effect if moving in front of Chologaster. The moment one came in close proximity 
to a fish from any direction, by a flashlike motion it was seized. None of them 
were swallowed. The fishes became very alert after the introduction of the sowbugs 
and when swimming forward would strike at a part of a leaf if it came in contact 
with the head of a fish. It seemed evident that the eye gave no information of 
the character of the object. As the Asellus was not altogether to their taste, Gam- 
marus was introduced. One of these, swimming rapidly toward the chin of the 
Chologaster from behind and below, was instantly seized when it came in contact 
with the fish. The eye could not have located the Gammarus at all. The action 
is in very strong contrast to the action of such a fish as Lepomis, which detects its 
food by sight. It is undoubtedly this peculiar method of locating and securing 
food which has enabled the Amblyopsidaj to establish themselves in caves. 

On March 20 the eyes were removed from 7 living specimens of Chologaster 
papilliferus with the following results: 

Within half an hour after removing the eyes, examples of Asellus were intro- 
duced into the aquarium, which were readily detected and captured. In captur- 
ing them the chologasters were not as accurate as fishes might be expected to be 
that do not ordinarily depend on their eyes to help in locating prey. It may be 
borne in mind, however, that the eyes were removed from the surface and that in 
addition to the removal of the eyes some of the tactile organs were probably dis- 
turbed or destroyed. 

A rod held in the hand was readily perceived by the blinded fishes, who avoided 
it with as much dexterity as an Amblyopsis would, except that their actions in 
avoiding the rod were very much quicker than the action of an Amblyopsis. The 
latter, if approached from in front, will back water with its pectorals and then, if 
the rod comes nearer, it will turn to one side or another, frequently with lazy delib- 
eration. Chologasters, on the other hand, would turn tail with a flashlike motion 
when the stick was approaching them. They could be approached from the back 
more readily than from other regions. 

The action of the blinded fishes was in this respect precisely like that of an 
unblinded one in the same aquarium. Removing the eyes makes no appreciable 
difference in the appearance of the fish, and a number of colleagues were asked 
whether the fishes were detecting the rod by sight (with the eyes) or by tactile 
sensation. Not knowing that the eyes had been removed, the verdict, in the major- 
ity of cases, was in favor of the eyes; in the other cases it was doubtful. There 
was no general disturbance of the fishes in the aquarium when the rod was intro- 
duced. Only the ones immediately concerned responded. 

On April 4 I was able to touch each of 5 blinded chologasters on the snout 
with a glass rod before it made any attempt to get away. The same is true of some 
which had not been blinded. 

The blinded chologasters readily swim about in the aquarium, regardless of 
protection or of contact with the sides of the aquarium. They not infrequently 
lie at the bottom, but the general tendency is to swim about freely. One of them 
lived for 2 years after the operation. 


At lo a. m. of one day the blinded fishes were removed from the large aquarium 
and replaced by a number with eyes. These at first remained at the bottom, but on 
the following morning they were swimming about as the blinded ones had been. 
The general conclusion from these experiments is that the Chologaster papilliferus 
with comparatively well-developed eyes can get along without them just as well as 
with them. 


A long series of observations and experiments was made to determine the reaction 
of Chologaster and Amblyopsis to white and monochromatic light. Incidentally 
other characteristics were brought out. 

Some previous experiments on blind or blinded vertebrates may be recalled. 
Dubois (Compt Rend., t. ex, pp. 358-360) and Semper (p. 79) record that Proteus, the 
blind salamander of Europe, is sensitive to diffuse light. Graber records that blinded 
salamanders prefer dark chambers to light ones. Korang (Centralblatt f. Physiol. 
VI, pp. 3-6) notes that concentrated light deprived of heat rays thrown upon the leg 
of a frog whose brain had been laid bare and covered with extract of beef, caused 
it to respond each time with reflex movements. 

That Amblyopsis avoids the light, even the diffuse daylight of a room, is 
without question. An aquarium was divided in the center by a black partition; 
one end of the aquarium was covered and the sides painted black, and a small 
opening was left in one of the lower corners of the partition to enable the fishes to 
move readily from one chamber to the other. The fishes had no difficulty in find- 
ing this opening, and at the beginning of the experiment, before the fishes had 
quieted down from the excitement incident to moving them, they swam back and 
forth from one chamber to the other as rapidly as it was possible to note the changes. 
The following are some of the results obtained at separate times: 

Experiment I : Observation on 6 individuals placed in the above aquarium, 
May 12, 1906, gave, between 9.43 a. m. and 10.20 a. m., 104 events in the dark, and 
220 in the light. 

This would indicate that the fishes have a preference for the diffuse daylight of 
the room over that of the dark chamber. But these specimens had been in the 
light several days, so the light-perceiving or light-reacting organs may have been 
fatigued. Subsequent events and tables indicate the opposite in such a striking 
way that the evidence is conclusive. A rapid moving of different individuals 
from one chamber to another was due to the excitement caused by preparing the 
aquarium, and the preference shown for one or the other conditions of illumina- 
tion was entirely overcome by the excitement produced. 

Experiment II : Conditions as in the first experiment with the same 6 individuals 
in the afternoon of the same day, the aquarium placed so that sunlight entered 
the lighted end of the aquarium. Result, 114 events in the light, 204 in the dark. 

The second experiment shows that there is an inclination to seek the dark 
rather than sunlight. That the fishes had not gotten into a normal condition is 
evidenced by the rapid changes of different individuals from light to dark and 
vice versa. Toward evening as the direct light was excluded the fishes began to 
go over to the lighted compartment. 

' For further studies see Payne, Biol. Bull, xili, pp. 317-3J3. 



Experiment III: On May 13 the same 6 specimens were used under the same 
conditions as in experiment I. The aquarium had been quiet since 5 p. m. the 
evening before. 


In the 

In the 


In the 

In the 


In the 

Id the 


In the 

In the 

A. U. 

A. H. 

A. U. 

A. u. 

h. m. 








h. m. s. 

9 21 













9 45 3 


9 21 














9 45 15 



9 21 













9 46 15 



9 21 













9 48 



'9 22 












9 48 40 



'9 22 














9 49 



9 24 














*9 50 15 



i' 'k 













9 50 30 



'9 36 
9 26 
















'9 51 












» Two exchanged. 

" One other came out, went back. 

3 Two exchanged, the one last out returning. 
* One came out, but went back at once. 

Experiment IV: On May 13, from 2 to 3 p. m., during the period correspond- 
ing to the time when records were made the day before, the fishes stayed in the dark 
chamber except occasionally when one would come into the Ught only to quickly 
turn and swim back into the dark. 

Experiment V: On May 15 the fishes remained in their dark chamber nearly 
all day except during the excitement caused by changing the water, when they 
swam freely into the light. It is evident that the incessant changing during the 
first observations recorded was due to excitement caused by the change of water 
and aquarium. 

A small opening was made in the front of the dark chamber, through which 
observations were made. A few individuals on this occasion came out. 

Experiment VI: On May 17 no blind fishes were in the light chamber between 
8.30 a. m. and 9.20 a. m. Through an opening in the top of the dark chamber several 
were observed to come to the opening between the two chambers but quickly to 
withdraw. The sides of the light chamber were painted with a wedge-shaped dark 
area the better to protect the dark chamber from oblique rays. 

Effect of jarring. — The aquarium was moved slightly in order to note the 
effect of jarring. While no fishes had been in the light chamber during the morn- 
ing, 4 were now out in a few moments; these returned and during 7 to 10 minutes 
the changing to and from the dark chamber was kept up. 

At 9'^ 30"' 17 approach opening of dark chamber without going out. i6 approach corner above the opening. 
At 9'' 37"' drew off 2.5 inches of water to 0.5 inch from level of top of opening. 

At g*" 45™ one came to opening and returned; 7 went through opening. Evidently still some disturbance. 
Left the observations at 9'' 48™. 

After the fishes had become quiet it was seen that while they were constantly 
moving past the opening it was rare that one passed out into the light chamber, 
and then they invariably showed signs of uneasiness, frequently turning sharply 
round and reentering the dark chamber, at other times making a complete circuit ; 
this at a time when there was no direct .sunlight. 

At 12 m. a dark tunnel was constructed by leaning a black pane of glass against 
the dark partition, leaving an opening at the side of the aquarium opposite to that 
in the opening of the first partition. For some time after this was done the fishes 
stayed in the light chamber in which they had been put, without being able appar- 


ently to find their way out. After a day, however, all had collected in the dark 
chamber and it was rare that any of them came out into the light chamber. They 
remained in the dark chamber for days without coming out, except occasionally 
at night. On May 24 the blind fishes remained in the dark compartment until 
night, when all collected in the light compartment, only to be found back again 
in the dark the next morning. 

Everything indicates that they readily perceive light, even the diffuse light of 
a room, and that they individually react negatively to light. 

Four Amblyopsis which had been kept for a day in a vessel painted black and 
covered to exclude the light were experimented upon as follows: a ray of light 
from a microscope mirror about 2 inches in diameter was thrown on each success- 
ively. After from i to 5 seconds the fishes became uneasy, the uneasiness giving 
place to discomfort, the fishes making vigorous efl"orts to get out of the ray. 

Another jar, not painted, containing both Amblyopsis and blind Cambarus, 
was placed where light could be reflected upon them from the mirror of a micro- 
scope. The Cambarus, if in motion, came suddenly to a halt ; if quiet, it backed or 
moved off at once. The fishes also responded to the light but it took several times 
as long for them to do so. 

Bright sunlight appears to be irritating ; if exposed to it, the fishes swim about 
uneasily. A shadow passed suddenly across them when in the diffuse light of a room 
does not affect them, nor do they, when swimming, seem disturbed by a ray of light 
entering the dark chamber through a small hole in the paint made for the experiment. 

Two examples kept in a pail in my ceUarwere quietly floating, but when a lighted 
match was held above them the fishes at once darted to the bottom and sides of 
the pail. The heat could not have been a factor in this case ; the reaction to the 
light of the match was quick and violent. 

A similar observation was made on 40 individuals in two aquaria. They were 
captured one morning, and the observation made the second night after. They 
had been kept in the dark during most of the intervening time. A lighted match, 
held near the aquaria, produced a very general and active movement among all the 

Even more striking than this was the action of a colony of Amblyopsis in an 
open pool. During the bright part of the day the fishes remained under the rocks 
at the bottom. Occasionally a nose could be seen poking out from under a rock ; 
perhaps one of the fishes came out at times during the day. In the morning and 
evening and at night, they could be seen swimming in various parts of the pool. 

The following experiments make it evident that the direction of the light does 
not influence the actions of these fishes, but that their behavior is due to a per- 
ception of difference in the intensity of light. A large box, covered at its southern 
end, was sunk into the ground where the water of a spring flowed through it. 
Throughout the lighter parts of the day the fishes stayed in the shade of the south- 
ern part of the aquarium. It was only in the evening, in the morning, and at night 
that the fishes ventured forth. A similar box 2x4x8 feet, divided in the middle 
by a partition running to near the bottom, had lids hinged so that either or both 
compartments could be covered and darkened. Within a short time after one of 
the compartments was darkened all of the individuals would be found in the dark- 
ened compartment, irrespective of the direction of the sun's rays. 


Mr. F. Payne has made further studies and found that their negative heliotro- 
pism is sufficient to overcome their positive geotropism if an 800 candle-power arc 
lamp is used 16 inches from the aquarium. He also found that the young fish 
to an inch in length react more strongly to light than older ones, even if their eyes 
are destroyed, and that one part of the body is as sensitive as another to a pencil of 
strong light. 

The 7 blinded chologasters mentioned previously were placed at 9 a.m. in an 
aquarium which was dark at one end and light at the other, but with no partition 
between. In the bottom of this aquarium, extending from the lighted into the 
darkened area, was placed a plate of glass propped up at one edge so as to enable 
the fishes to get under it. The conditions in the two parts of the aquarium were 
as nearly alike as possible except as to light. The blinded chologasters collected 
in the darkened half of the aquarium and remained there. The reaction was quite 
positive. No sunlight entered the aquarium — only the diffuse light of the room. 
The same reaction took place when sunlight entered the aquarium. 

Later, the pane of glass was taken from the bottom of the aquarium and placed 
against its sides, and the fishes collected behind it in the dark end. A number of 
normal chologasters in another aquarium had the same habit of squeezing them- 
selves in between the sides of the box in which they were and the small glass 
aquarium placed in it. It is evident that Chologaster is also negatively helio- 
tropic and positively stereotropic. 

A series of observations was made to determine to what rays, if any, Amblyopsis 
reacts most vigorously. For this experiment a glass jar 3 feet long and 8 inches in 
diameter was divided into 6 compartments by 5 partitions. Each partition had a ver- 
tical slit extending half-way up from the bottom to enable the fishes to swim freely 
from one compartment to another. The compartments were thus all connected. A 
cap was screwed tightly over the end of the jar, which was placed horizontally in a 
window-sill where each compartment would have an equal amount of light.' The 
jar was surrounded with bands of tissue paper in several layers of violet, blue, green, 
orange, and pink so that each compartment was lighted by one series of rays. 
Three Amblyopsis were used for these observations; they were selected for their 
size and named, A, the smallest, b, the middle-sized, c, the largest. These fishes 
had been in confinement some time, but had been transferred from the cave, with 
as little exposure to light as possible, to a dark room where they were very seldom 
exposed to the light. Observations were made as opportunity presented itself. 

It was found that some compartments were visited by a certain fish without any 
definite regard for color. During January, for instance, fish c moved out of the pink 
and orange compartments but once ; fish A remained almost exclusively in yellow, 
visiting pink once, orange once, and green 4 times. Fish B, on the other hand, 
remained mostly in the violet, visiting blue 7 times and green 3 times. From this 
we must conclude either that different individuals react differently or that one color 
does not produce a stronger reaction than another, and the latter seems the more 
reasonable conclusion. (See table on page 91.) 

To determine whether the apparatus had anything to do with the distribution 
and also whether widely separated elements of the spectrum would cause the fishes 

' For over a month these fishes were sealed in this jar without change of water. 



Series of Observations to determine to which Rays Amblyopsis react 

most Vigorously. 











h. m. 

Dec. 16 



A, B 

B, C 









A, C 


B, c 



A, B 






B, C 


A, C 





A, C 














Jan. 4 












































































































A, B 









A, C 









Fish No. c 































to react positively or negatively, they were put into a rectangular aquarium im- 
pervious to light, except at the ends, and divided by a median partition. The ends 
were covered with translucent celluloid film, care being taken, of course, to have 
each end equally light. Random observations taken through 20 days show: 
A, once in the blue compartment and 34 times in the red ; b, 6 times in blue and 39 
times in red; c, 27 times in blue and 18 times in red; a total of 34 times in the 
blue and 91 times in the red. 

If only A and b had been used, we would have been justified in concluding that 
Amblyopsis is positively tropic toward the red end of the spectrum as against the 
blue. If only c had been used, we would have been justified to draw the opposite 
conclusion. The fishes in the red compartment had become nervous and were 
swimming near the red window, that is, on the side opposite the opening between 
the compartments. Their proneness to remain in the same compartment may 
have been partly due to this nervousness, the cause for which was not apparent. 

Four specimens of Chologaster were placed in the apparatus having 6 different 
colored compartments. Between January 26 and February 4 rather irregular 
observations were made. 

The number of specimens for each compartment on a purely chance distribu- 
tion would have been 12.6, leaving out of consideration the element that the end 


compartments contained but one opening, only one compartment bordering each. 
A strong positive reaction toward violet is indicated, and a strong negative reaction 
toward pink and blue. The totals were : violet, 25 ; blue, 6 ; green, 1 1 ; yellow, 13 ; 
orange, 14; pink, 7. 

To test these results, the second aquarium, with but two compartments and three 
specimens, was used. The specimens are marked A smallest, B medium sized, c 
largest. In series I, 20 out of 24 events occurred in the red. The windows were 
interchanged, transposing colors, when, in series II, out of 24 events 13 occurred 
in the red. This indicates a decided positive reaction toward the red. In series 
III, 16 out of 20 events occurred in the red. A new aquarium was substituted, with 
the windows side by side, looking toward a west window. Out of 17 events (series 
IV) 13 occurred in the red. The colors were then interchanged, so the fishes would 
be compelled to change compartments in order to be in the same light. In this 
series 27 out of 29 events occurred in the red. These series give conclusive evi- 
dence that the affinity of these fishes is strongly in favor of the red. It may also be 
noted that the smallest specimen was most frequently found in the blue. 


The eggs are laid by the female, to the number of about 70, into her gill cham- 
ber. Here they remain for perhaps 2 months, till the yolk is nearly all absorbed 
and the young fish has attained a length of about 10 mm. If at any time a female 
with young in her gill pouches is handled, some of them are sure to escape. This 
was observed and gave rise to the idea that this species is viviparous. 

We owe the first observations on the breeding habits of Amblyopsis to Thomp- 
son, who states that a fish "was put in water as soon as captured, where it gave 
birth to nearly 20 young, which swam about for some time, but soon died * * * they 
were each 4 lines in length." It is unfortunate that the highly interesting suppo- 
sition of Thompson that they were viviparous has gained common currency. 

Putnam adds to the above, judging from some data in his possession, that the 
young are born in September and October, and further along remarks that they 
are "undoubtedly" viviparous. 

The first young I obtained were secured on May 9, 1896. The little fishes 
could move actively for a few moments, but as they were encumbered with much 
yolk, they soon settled to the bottom and remained quiet. A large number of old 
ones were in the water in which the young were found, and the mother of this lot 
was not identified with certainty. Another lot of young obtained on September 5 
of the same year were much farther along in their development. Some were pre- 
served and others placed in various aquaria, where one lived to be 10 months old. 
As before, the parent was not with certainty determined, simply because it was 
taken for granted that they were viviparous and the ovaries only were examined. 
Two other lots of young were obtained on June 5, 1897. One of these lots was in 
the stage of the first lot obtained, with a large amount of yolk still present, while 
in the other lot the yolk had almost entirely disappeared. These had been carried 
in the gill cavity of the mother, and it became evident either that the fishes were 
not viviparous at all, or that their viviparity was not nearly of the pronounced 
character hitherto supposed. 

On March 11, 1898, 29 individuals were captured. Four were females with eggs 
in their gill cavities. The youngest stage among these was at the end of scgmen- 




Views of Amblyopsis, in the early stages. 

A and B. Embryos on egg, (A) younger stage, (B) older stage. 

C. Larva at time of hatching. 

D. Older larva. 


tation, the oldest was a gastrula covering but one-third of the yolk. The eggs had 
not been developing more than 5 days, probably not more than 2 at the utmost, 
and decided beyond a doubt that these fishes are oviparous and not viviparous. 
In one individual 6i eggs v^^ere found, in another 70. The exact number in the 
other two, I can not give, but the number does not differ greatly from the above. 
From one side of one I took 35 eggs, from another individual an uncertain number. 
The remaining eggs were left in the gills to develop, but all that were not subse- 
quently preserved finally died. 

The female with eggs can readily be distinguished by her distended gills, and 
since dead eggs become opaque, such can readily be distinguished through the 
translucent opercles and branchiostegal membrane. Dead eggs are retained in the 
gill cavity till they disintegrate. 

I have never secured as many young from any female as the eggs enumerated 
above. This may have been either on account of the dying of many eggs or the 
liberation of the young during the struggle of capture. 

Emphasis need be laid on the fact that Amblyopsis is not viviparous and that 
its breeding period extends at least from the first of March to November and 
probably throughout the year. A female with nearly ripe eggs was secured on 
September 9, and since these would have been carried either as eggs or young for 
about 2 months longer, November is a safe limit. During March the spawning 
season is evidently at its beginning, and it is during this month and April and May 
that the early stages may be looked for with the greatest confidence. 

No eggs were deposited in the laboratory. Females with eggs in the gill 
cavities had to be sought for in the caves. To secure embryological material when 
a female containing favorable stages was captured, she was isolated in a small 
aquarium and the number of eggs needed freed from the gill cavity by gently 
raising the edge of the operculum. The rest of the eggs were permitted to remain 
in their natural surroundings until another lot was wanted. During the early 
stages of development the edges of the operculum are closely pressed to the neck 
and there is no danger of freeing more eggs than are wanted unless the fish is 
roughly handled. During the later stages of development the tension of the oper- 
culum is relaxed and eggs or larvae can be much more easily removed, but there 
is a correspondingly greater danger of liberating more young than are wanted. 
If the female is disturbed or confined during the latest stages of brooding, some 
or all of the young will escape. The eggs freed from the gill cavity will continue 
their development uninterruptedly, but the gill cavity of the female offers such a 
unique and self -regulated hatchery that they were usually left in it. 


In an aquarium containing six specimens of Amblyopsis, two took a great 
antipathy to each other and engaged in vigorous contests whenever they came in 
contact. Frequently they came to have a position with broadside to broadside, 
their heads pointing in opposite directions. The fight consists in quick lateral 
thrusts toward the antagonist to seize him with the mouth. The motion is in- 
stantly parried by a similar move by the antagonist. This blind punching may 
be kept up for a few seconds, when, by their vigorous motions, they lose each 
other and jerk themselves through the water from side to side, apparently hunt- 
ing for each other. At this time they are very agile and move with precision. 


When the beUigerents meet, one above the other, the snapping and punching is 
of a diflferent order. While jerking through the water, just after a round, if one 
of the belligerents touches one of the neutrals in the aquarium, it frequently gives 
it a punch, but does not follow it up, and the unoffending fellow makes haste to 
get out of the road, the smaller ones most quickly. If, after an interval of a few 
seconds, a belHgerent meets a neutral, they quietly pass each other without paying 
any further attention ; whereas if the two beUigerents meet again, there is an im- 
mediate response. Whether they recognize each other by touch or by their mutual 
excitability, I do not know. In another aquarium I saw one belligerent capture 
the other by the pectorals. After holding on for a short time it let go, and all dif- 
ferences were forgotten. The thrust is delivered by a single vigorous flip of the 
tail. These fights were frequently noticed, and, as far as determined, always 
occurred between males. 

The absence of secondary sexual differences in the cave fishes is a forcible 
argument in favor of sexual selection as the factor producing high coloration in 
the males. The absence of secondary sexual differences in caves opposes the idea 
of Geddes and Thomson, that the differences are the external expression of maleness 
and femaleness. 


The eggs are large, measuring 2.3 mm. in diameter. The yolk is translucent, 
of various tints of amber. The yolk measures 2 mm. in diameter and contains a 
large protruding oil-sphere i to 1.2 mm. in diameter. When the egg is deposited, 
the yolk is flabby and composed of yolk-spheres of various sizes loosely put 
together. After the egg has been in water for some time, the yolk forms a tense 
rounded mass. The egg is heavier than water. The oil-sphere lies uppermost 
in the egg, and the germinal disk forms at the side of the egg. Attempts at artificial 
fertilization have not been successful beyond obtaining well-developed germinal disks. 

The rate of development will probably be found to vary considerably with the 
temperature of the water. In a series of eggs in which the gastrula covered half 
the yolk when observations began, the blastopore was reduced to the size of the 
oil-sphere in 9 hours, when the embryo encircled about a third of the yolk; 16 
hours later the blastopore was closing. The rate of development of the series of 
eggs taken in May was as follows, the mother containing the eggs having been 
kept in a small aquarium without change of water and at the temperature of an 
ordinary living room. The temperature of the water in the cave is 12° C, that 
in the room was 22° C. 

On May 4, 9 p. iti., the gastrula covered approximately half the yolk. It lies eccentric, neither 
below nor at the side, the germ being evidently heavier, the oil-sphere at the top. 

May 5, 6 a. m., the embryo surrounds about a third of the egg, the blastopore is about as wide 
as the oil-sphere, i.2mm.,andthelatter seems to fully fill it. At 2.30p.m. the embryo is i. 6mm. long 
and has 4 protovertebrae. At 6 p. m. the blastopore has narrowed considerably and invariably lies 
at one side of the oil-sphere, the embryo lying oblique to the vertical axis of the egg. This eccentric 
position becomes more and more evident as the blastopore closes toward lop. m. The embryo is 1.76 
mm. long, with 6 protovertebrae. At 10 p. m. the eyes and brain are shaped like the ace of spades, 
the eye lobes evidently not yet narrowly separated from the brain by a narrow stalk, the blastopore 
closing, the embryo 1.92 mm. long, and with 10 protovertebra;. On May 6, at 6 p. m., the embryo 
lies horizontal around the margin of the yolk ; the cavity of the central nervous system has appeared ; 
a large piece has been eaten out of the yolk; the lens is just beginning to develop. There are 12 
or 13 proto vertebrae. At 8 a. m. the embryo is 2.4 mm. long; at 11 a.m. no marked change is seen; 
at 6 p. m. tail is beginning to bud out; embryo, 3 mm. long, encircles half the yolk; 17 proto- 
vertebrae present. 



There is a regular change in the position of the embryo with development. 
The blastoderm is formed at the side of the yolk. When the gastrula covers half 
the yolk, the egg has rotated so that the gastrula covers more of the lower than 
of the upper surface of the yolk. Still later, some hours before the closing of the 
blastopore the latter structure lies to one side of the yolk-sphere, which always 
occupies the upper pole of the egg ; the embryo extends from this region obliquely 
over the yolk. After the formation of the tail the embryo is always found coiled 
about the upper half of the yolk. The period spent in the egg lasts about a month. 
In the laboratory some embryos hatched in about 28 days, but in the cold cave 
streams this period would probably be several days longer. The yolk has been 
but little affected at the time of hatching, measuring 1.8 mm., the oil-sphere about 
I mm. ; and since the yolk is all absorbed before the young are freed from the 
giil membrane, probably another month is spent under the gill membrane. 

Fig. 30. (a) Internal \nAtomy oi Amblyopsis spelaus. i, anus; 2, opening of oviduct; 
3, oviduct; 4, ovary, which is single; 5, liver; 6, duodenum; 7. gall 
sac; 8, pectoral fin; 0, one of pyloric caeca; 10, ca;cum; 11, stomach; 
12, spleen; 13. air bladder; 14 and 16, intestine; i^, pancreas; /., liver. 

(b) Alimentary Canal of Chologaster cornutus, ^..pyloric cxca; s., stomach; i'.,vent. 
(cS Alimentary Canal of Chologasler papilli/erus, 

id) Alimentary Canal of CItoUtgasler agassizii. 

(c) Alimentary Canal of Typhlichtkys subterraneus. 

The young, on hatching, are about 5 mm. long and lie on their sides. The 
motion of the tail produces no effect other than to cause them to spin around 
with the yolk for a pivot. The metamorphosis of the larva into the definitive fish 
is completed before it leaves the gill cavity of the mother. The longest individ- 
uals I have secured from the gill cavity measure about 10 mm. 


Certain structures gain an entirely new significance in the light of the breed- 
ing habits. These are the enlarged gill cavities with the small gills, the closely 
applied branchiostegal membrane, and the position of the anus and sexual orifices. 

The anus in all of the species has undergone a curious translocation. The 
primary cause of the transposition probably lies in the ovary and oviduct, and not 


in the alimentary canal. The opening of the oviduct has moved forward until 
it lies in front of the pectorals and it has carried the anus forward with it. In 
newly hatched individuals the anus has its normal position behind the ventrals. 
When the fish has reached a length of 25 mm., the anus has reached a point in 
front of the ventrals, but it is still nearer the ventrals than the pectorals; with 
a length of 35 mm. the anus has moved forward to just below the insertion of 
the pectorals. In mature specimens it lies considerably in advance of the pec- 
torals (plate 5, fig. c). The forward movement of the sexual orifice takes place in 
both sexes. 

Nothing is definitely known of the advantages of the location of the opening of 
the oviduct. They can be inferred from the habit of Amblyopsis in carrying its 
eggs in the gill cavity. Located as it is, the oviduct may be covered by the gill 
membranes of the 2 sides alternately, or, if the fish takes an oblique position in 
the water with the head down, the eggs may flow directly into the gill cavities, 
being carried downward by gravity and held in the groove in front of the anus 
by adhesion. 

It is difficult to imagine even a formal explanation of the origin of the position 
of the sexual orifice in the Amblyopsidae. The anus was probably carried forward 
as the result of the forward movement of the sexual orifice, and it is this that 
demands explanation. Very probably the habit of carrying the young in the gill 
pouches antedates the present position of the anus. The eggs may have been 
allowed to flow into the gill openings, the female occupying a position vdth head 
downward during oviposition. If this were the case, then, while the individual 
skill would count for much in transferring the ova, a variation or mutation which 
lessens the distance between the sexual orifice and the gills would be of distinct 
advantage and would probably be transmitted by natural selection. The actual 
transfer of the ova into the gill cavity has not been observed. 


The tactile organs are among the most important in the consideration of the 
blind forms. Their minute structure will form the basis of a separate paper. 
The prominent tactile organs about the head of Amblyopsis have been mentioned 
by nearly every writer, and they have been figured by Putnam-Wyman and Leidig ; 
but the figures of the distribution of the ridges are worthless. The description 

by Professor Forbes of Chologaster papilli- 
fcrus is the only systematic enumeration of 
the ridges that has appeared. The accom- 
panying figures (32 and 33), drawn by me 
with the camera lucida, verified and copied 

Flc. 31. Tacttte Orgun in Hea.d ot Larvsc Amblyopsis. See l^ -\/r tt f) P„y friim fVlf> PYOrt pvtpnt 
also Piate 70 just above Yolk. Larva was placed Oy iVir. U. \J. \^0X, glVB inC CXdCl CXCLni 
in weak osmic acid which brought out outlines j „ ... „ r .i "j ;„ /I ..„ 7^7^. «j. .,.,*„ 

of structure. 8 mm. + 4 ocular. and positiou of the ridgcs in Amblyopsis, 

Typhlichthys, and Chologaster papilliferus. 
It will be seen that in the number and distribution of the tactile area the three 
forms agree very closely, the eyed form having the same number and dis- 
tribution of ridges or rows that the blind forms have. In C. papilliferus most 
of the ridges are much less prominent than in the blind species, being sunk 
into the skin. About the nose and chin, however, the ridges are as prominent 



as in the other species. In Chologaster cornutus there are no distinct ridges at all, 
the tactile organs being arranged as in other species of fishes. In specimens of 
the same size the papillae are not more prominent in papilli/erus than in cornutus. 
It is only in the oldest of papilliferus that the papillae become prominent. The 
number of individual papillae in each tactile ridge differs considerably with age 
(size), so that an exact comparison between the large Amblyopsis and the much 
smaller species of Chologaster and Typhlichthys can not be made. From a num- 
ber of counts. Professor Cox found that ridge No. 6 contains in Chologaster papilli- 
ferus, 6 organs; in Typhlichthys, ii ; in two specimens of Amblyopsis, respectively 
3.33 and 4.25 inches long, 12 and 20. The tactile ridges in the head of Amblyopsis 

(rf, e) Distribution of Tactile Ridges in Typhlichthys iubtcrraneus ; dorsal and side views. 

Fig. 32. (a, 6, c) Distribution of Tactile Ridges in ^mWyo^m; lateral, dorsal, and ventral views. 

are shown in plate 8, figures A and b. The outermost layer of skin has been re- 
moved from a small area over the right eye of A, showing the numerous taste buds. 
Figures c and d show head of Chologaster papilliferus under slightly greater mag- 
nification. Figure d shows especially the tactile organs about the mouth. The 
skin passes over the eye without a free orbital rim, and the eye does not show well. 

Aside from the tactile organs in ridges there are many solitary ones not evi- 
dent from the surface in Amblyopsis. When the epidermis is removed by macera- 
tion, the dermal papillae on which they rest give the whole head a velvety appearance. 

In the young, at least, of Amblyopsis, each of the tactile organs of the ridges is 
provided with a club-shaped filament abruptly pointed at the end (fig. 31). They 
wave about with the slightest motion in water and are so numerous as to give the 
whole head a woolly appearance. 



Tellkampf has remarked : 

The blind fish is found solitary and is very difficult to be caught, since it requires the greatest 
caution to bring the net beneath them without driving them away. At the slightest motion of the 
water they dart off a short distance and usually stop. * * * During my stay at Mammoth 
Cave I observed that the Amblyopsis * * * remained motionless while I moved a burning lamp 
around them, but they were disturbed by a slight motion of the water, proving that the light 
made no impression upon their optic nerve, while their sense of touch was acute. 

Fio.*33. (a, b, e) Distribution of Tactile Ridges in Troglichthys. Side view of entire fisli. dorsal and ventral views of head. 
id,e,f} Distribution of Tactile Ridges in Chohgasler papiUi/erus. Side view of entire tish, dorsal and ventral 
views of anterior part of body. 

Dr. John Sloan in Packard, 1887, wrote: 

We carried our lighted candles within a few inches of them when near the surface, but they 
seemed wholly insensible to their existence; but if a drop of tallow fell in the water near them, 
they would swim rapidly away. I brought home 12, as many as could live in my bucket. Of these 
12 caught in September none died until next June, when the water became warmed to near 70°, 
when several of them died with tetanic convulsions (?). I put the remainder in my cellar, where the 
temperaturerangedfrom45° to6o°, whereone, "Blind Tom," lived 11 months, making 20 months 
of existence without having taken any visible food. While in my aquarium they manifested total 
indifference to light and sound. * * * They manifest great sensibility on the back and sides to 
any approaching body, but do not notice an attack from below. It is not possible to capture one 
by a side sweep of the net, but by passing it under him a considerable distance below and bringing 
it up slowly there is no difficulty in taking them. In their native pools and in the aquarium when 
disturbed they do not strike the bottom or sides of their surroundings, but seem to have a sense of 
resistance (if the term is pardonable) which protects them. 

Miss Hoppin in Garman remarked : 

I am very sure they [cray-fishes], as well as the white-fish [Troglichthys] have the tactile sense 
developed in an unusual degree. At the least touch upon the water they dart away. * * * Nu- 
merous tests convince me that it is through the sense of touch, and not through hearing, that 
the fish is disturbed. * * * If I strike the vessel so that the water is set in motion, he darts 
away from that side through the mass of water, instead of around in his usual way. If I stir 
the water or touch the fish, no matter how lightly, his actions are the same. 











Photographs of ihe tactile organs of Amblyopsis and Chologaster. 

A. Head of Amblyopsis from above, showing tactile ridges. 

B. Same head from side. Tactile organs especially numerous about mouth. 

C. Head of Chologaster papiliiferus, from above, under slightly greater magnification 

than A. 

D. Same head, from side, especially showing tactile organs about mouth. 


Blatchley states: 

* * * the least movement of the water frightened them, and they darted rapidly away, usually 
at right angles to the course they were pursuing. The sense of touch, rather than that of hearing, 
is, in my opinion, the one which has been intensified by long residence in the dark and silent recesses 
of the caves. 

I have not found the slightest difficulty in capturing Amhlyopsis with a small 
dip net, either from a boat or while wading through the subterranean stream, and 
I have caught one in the hollow of my hand. At such a time any amount of noise 
I was capable of making did not affect the fishes found swimming in the water. 
Frequently they were taken in the dip net without apparently taking any note 
of the vibrations produced in the water until they were lifted out of it ; very rarely 
a fish became noticeably scared. Such a one would dart off a few feet or a few 
inches and remain on the qtii vive. If not pursued, it soon swam off quietly; if 
pursued, it not infrequently escaped by rapidly darting this way and that; when 
jumping out of the water, often an abrupt turn in the opposite direction from which 
it started would land it in the net, showing that their sense of direction was not 
very acute. At other times, if disturbed by the waves produced by wading, one 
or another individual would follow a ledge of rock to the bottom of the stream, 
where it would hide in a crevice. But very frequently, much more frequently 
than not, no attention was paid either to the commotion produced by the wading 
or by the boat and dip net. In general it may be said that the fishes in their natural 
habitat are oblivious to disturbances of the water until frightened by some very 
unusual jar or motion, probably a touch with the net, when they become tensely 
alert. The fact that they are not easily frightened suggests the absence of many 
enemies, while their frantic behavior if once scared gives evidence that occasional 
enemies are present and that they are very dangerous, or that the transmission of 
the instinct of fear is as tenacious as the transmission of physical characters. 

Contrary to Sloan's observation, that they detect the presence of a solid sub- 
stance in their path, I have never noticed that the fishes in confinement became 
aware of the proximity of the walls of the aquarium when swimming toward them. 
Instead, they constantly use the padded, projecting lower jaw as bumpers. Even 
an extremely rapid dart through the water seems to be stopped by the projecting 
jaw without serious inconvenience. 

Sticks, straws, etc. are never avoided by the fishes, even when the fishes had 
not been disturbed for hours. By this I mean that they are never seen to avoid 
such an object when it is in their path. They swim against it and then turn. An 
object falling through the water does not disturb them even if it falls on them. 
Gently moving a pencil in front of them does not disturb them much, but if the 
pencil is held firmly in the hand it is always perceived and the fish comes to a 
dead halt half an inch before it comes in contact with it. On the other hand, they 
may be touched on the back or tail before they start away. They glide by each 
other, leisurely and dignified, and if they collide, as they sometimes do, they 
usually display no more emotion than when they run against a stick. But this in- 
difference is not always displayed, as was noted under the head of breeding habits. 

A number kept in an aquarium having a median partition in which there was 
a small opening were readily able to perceive the opening, swimming directly for 
it when opposite it. This observation is in direct contrast to their inability to 
perceive solid substances in their path. A sharp tap on the sides of an aquarium 


in which 6 blind fishes were swimming, where they had been for a number of days 
undisturbed in a dark room, caused nearly all of them to dart rapidly forward. 
A second tap produced a less unanimous reaction. This repeated on successive 
days always brought responses from some of the inmates of the aquarium. Those 
responding were not necessarily the nearest to the center of disturbance, but some- 
times at the opposite side of the aquarium or variously distributed through it. 
After a few days the fishes took no notice of the tapping by any action observable 
in the artificially lighted room. 

Such tapping on a well-lighted aquarium containing both Chologaster and 
Amblyopsis was always perceived by the Amblyopsis, but the only response from 
these imperturbable philosophers was a slight motion of the pectorals, a motion 
that suggested that their balance had been disturbed and that the motion was a 
rebalancing. The Chologaster, on the other hand, invariably darted about in a 
frantic manner. One individual of Amblyopsis floating on the water was repeat- 
edly pushed down by the finger without being disturbed ; but if touched on the 
side, they always rapidly dart away. 

From everything observed it is quite evident that Amblyopsis is not keener in 
perceiving objects or vibrations than other fishes, and ordinarily pays much less 
attention to them. Mr. Payne's observation on the feeding habits leads one to 
conclude that they possess greater power of discrimination between vibrations. 
Some observations on young Amblyopsis are of interest in this connection. 

The young with a large amount of yolk still attached show a well-developed 
sense of direction. A needle thrust into the water near their heads and in front 
of them causes a quick reaction, the young fishes turning and swimming in the 
opposite direction. They will do this two or three times, then, becoming exhausted, 
will remain at rest. Sometimes an individual will not move until it is actually 
touched by the needle. The needle must come within about an eighth of an inch 
of the fish before it is noticed. Then, if the needle produces any result, it causes the 
fish to quickly turn and swim some distance, when the fish falls to the bottom again 
and remains at rest. If the needle be placed behind the fish, it will swim directly 
forward ; if at the side or about the middle, it swims directly forward or turns and 
swims in the direction opposite the origin of the disturbance. Younger specimens 
have no power over the direction of their progress — the wiggling of the tail simply 
produces a gyration, with the yolk as pivot. 

A young blind fish, 6 months old, swims about in a jerky manner, chiefly by 
use of its pectoral fins. It keeps close to the side of the vessel, usually with its 
back to the glass. (The aquarium was a cylindrical jar 300 mm. in diameter 
and 300 mm. high.) From whatever direction it may be approached it perceives 
a stick thrust toward it as readily as a seeing fish can, and will invariably dart 
away a short distance, sometimes making sharp turns to avoid the stick and always 
successfully. It can be approached from the top nearer than from the sides or from 
in front. It does not avoid the sides of the aquarium, which it frequently strikes. 


Anatomically considered, the ear of Amblyopsis is normal. Numbers of ears 
together with the brains have been dissected out. These were treated with Flem- 
ming's strong solution or with Hermann's fluid, either of which stained the nerve 
matter black. In the first place, the three semicircular canals are present and 


each has its ampulla fully developed. The three ampullae and the sinus utriculus 
superior communicate with the utriculus in front, behind, and above. Below^, the 
utriculus communicates with the sacculus, which terminates posteriorly in an 
appendage, the lagena. The three ear bones are present, one in the recessus 
utriculi, one (the largest) in the sacculus, and the other in the lagena. 

The auditory nerve divides into two branches, the ramus anterior and the 
ramus posterior. The ramus anterior divides into three branches — the ramulus 
ampullae anterioris, which extends to the anterior ampulla; the ramulus ampullae 
externae, which extends to the external ampulla; the ramulus recessus utriculi, 
which extends to the recessus utriculi. The ramus posterior gives off a heavy 
branch, the ramulus sacculi, which extends to the sacculus. The rest of the ramus 
posterior divides into the ramulus lagenae, which extends to the lagena; and the 
ramulus ampulla posterioris, which extends to the posterior ampulla. Another 
branch, the ramulus neglectus, which is normally given off where the ramus pos- 
terior divides into the ramulus ampulla posterioris and ramulus lagenae, has not 
been identified. 

The normal fish ear has seven auditory spots — the macula acusticus recessus 
utriculi, three cristae acusticus ampullarum, macula acusticus sacculi, papilla 
acusticus, and the macula acusticus neglecta. In Amblyopsis all of these auditory 
spots are present. 


(By Fernandus Payne.) 

Until the time of Bateson and Kreidl, it was generally taken for granted that 
fishes could hear because they had ears. Bateson concluded from his observations 
on congers, flatfishes, pouting, etc., that fishes perceive the sound of sudden shocks, 
but do not seem to hear the sounds of bodies moving in the water. Kreidl was 
the first to make experiments to test the hearing of fishes. He experimented on 
the gold-fish (Carassius auratus) and concluded that gold-fishes do not hear with 
the ear, but that they do react to sound waves by means of sense-organs in the 
skin. Lee's observations supported Kreidl's results, and he further concluded 
that the sole function of the ear in fishes is equihbration. Parker was the first to 
get positive evidence against the conclusions of Kreidl and Lee. His experiments 
were based on Fundulus heterodUus. He used three classes of fishes; first, nor- 
mal, that is, unmaimed, ones ; second, fishes with the auditory nerves cut ; and third, 
fishes with the skin rendered non-sensitive but vdth the ears intact. 

His apparatus consisted of a heavy aquarium with a slate bottom, two glass 
sides, and two slate ends, one of which he replaced by a piece of deal board to 
serve as a sounding board. To the middle of one edge of the sounding board he 
attached a stout beam of wood so that it stood out horizontally about i m. in 
the plane of that end. He stretched a bass-viol string from the free end of the 
beam over a bridge in the center of the sounding board to its opposite side. When 
the string was plucked or bowed, it produced about 40 vibrations per second. 
The fishes to be experimented upon were placed in a small cage suspended from 
a cord attached at its ends to the walls of the room. The end toward the sound- 
ing board was covered vdth coarse cloth. 

He subjected 10 normal fishes each to 10 tests, and from the 100 tests he got 
96 pectoral-fin responses. Fishes with auditory nerves cut responded only 18 
times in a total of 100 trials, and Parker thought these 18 times were in part acci- 
dental occurrences and in part due to the slight movements of the aquarium caused 
by the vibrating string. Instead of the vibrating string he substituted an electric 
tuning-fork which vibrated 128 times per second. With the tuning-fork, where 
the vibrations of the aquarium could be ehminated, he got no responses with the 
earless fishes. Fishes in which the skin was made insensitive, but with the ears 
intact, responded to sound 96 times in a total of 100. These fishes reacted almost 
exactly as the normal ones did. From these results Parker concludes that the 
killifish hears. Although his conclusion, that a fish hears, is contrary to Kreidl 
and Lee, he does not say that the observations of these men are entirely wrong, 
for the ears in different fishes may function differently. In fact, Parker found no 
evidence of hearing in the smooth dog-fish (Mustelus canis) when he subjected it 
to the same experiments as the killifish. Bigelow used Parker's methods of experi- 
menting and reexamined the gold-fish. He concludes that the gold-fish hears.' 

' Since writing the above Korner in Lucae's Festschrift, 1905, reviewed the evidence advanced to show that 
fishes can hear, and concludes that while they react to rapidly repeated tone-vibrations such as are produced by 
a tuning-fork or an electric bell, it is not proven that they perceive this with their ears. He used 25 species of 
fishes and found that in no case did any of these 25 species react in any way to a single sharp click. He con- 
cludes from these experiments that fishes do not hear. 


From the evidence at hand it is very probable that some fishes hear and that 
others do not. The following experiments have been made on the blind fish 
Amblyopsis spelceus. Various opinions have been expressed about the hearing of this 

Wyman states: 

It is said that the blind fishes are acutely sensitive to sounds as well as to undulations produced 
by other causes in the water. In the only instance in which I have dissected the organ of hearing 
(which I believe has not before been noticed), all its parts were largely developed. 

The following words of Professor Cope are frequently quoted : 

If these Amblyopsis be not alarmed, they come to the surface to feed and swim in full sight, 
like white, aquatic ghosts. They are then easily taken by the hand or net, if perfect silence is pre- 
served, for they are unconscious of the presence of an enemy except through the medium of hearing. 
This sense is, however, evidently very acute; at any noise they turn suddenly downward and hide 
beneath stones, etc., on the bottom. 

Subsequent writers have generally disagreed with Cope. Dr. Sloan (in Packard, 
1884) wrote: 

I tested their hearing by hallooing, clapping my hands, and striking my tin bucket when they 
were in easy reach and near the surface. In no instance did they change their course or notice the 

Miss Hoppin (Garman) failed to get any response from Troglichthys as long 
as noises only were resorted to. She says : 

I may scream or strike metal bodies together over him, as near as possible, yet he seems to take 
no notice whatever. 

Blatchley states that noises do not attract them. 

Eigenmann's observations (Proc. Brit. Ass. A. Science, Toronto Meeting) on 
Amblyopsis confirm those of Miss Hoppin on Troglichthys. No ordinary noises 
produced had any effect on Amblyopsis. Whistles, tuning-forks, clapping of hands, 
shouting in the reverberating caves, were alike disregarded. 

Amblyopsis, since it is blind, does not require precautionary methods to exclude 
sight as a possible disturbing element. 

If there are sounds in the water of the caves that concern the blind fishes and 
the ears are sound-perceiving organs, we might expect the ear to be better de- 
veloped along with the tactile organs as a compensation for the loss of sight. But 
if there are no sounds, we might expect them to degenerate along with the eye 
unless the function is something else than sound perception. Amblyopsis has few, 
if any, enemies in the caves. There are certainly none that make sounds, so the 
ears of the fishes would not be kept on the alert for them. There is less variety 
of sounds in the air of the caves than on the outside. This may make but little 
difference, as sound generated in the air does not penetrate readily into the water. 
Rippling of the water is certainly perceived more readily by the tactile organs than 
by the ear. Besides, the fishes are confined to the quiet pools. 

My methods of experimenting were practically the same as those of Parker 
and Bigelow. I used a heavy slate-bottomed aquarium, 24 inches long, 14.5 
inches high, and 12.5 inches wide. I removed the glass from one end and substi- 
tuted a board 2 inches thick. This served as a sounding board. The fishes were 
confined in a smaller aquarium (4x5x8 inch) suspended in the larger. The end 
of the smaller aquarium was covered with cheese-cloth toward the sounding board. 


The large aquarium rested on a masonry pedestal, which eliminated all vibra- 
tions of the floor. The small one was suspended by cords running from one side 
of the room to the other. 

After various trials with small tuning-forks which produced vibrations between 
icx) and 512 per second, which gave negative results, I used a large fork 12.5 inches 
in length vibrating 100 times per second and which produced a large volume of 

I used (a) unmaimed blind lishes and (b) others whose auditory nerves had 
been cut. I also attempted work with fishes whose lateral line nerves and nerves 
to the skin had been cut, but the specimens either died or did not revive suffi- 
ciently to get normal reactions. 

(a) Unmaimed blind fishes when placed in the aquarium invariably dart to the 
bottom and remain there for a short time, after which they begin to swim about 
rather rapidly. They soon become more quiet if nothing further disturbs them, 
but continue swimming about in a leisurely way, stopping now and then for a few 
seconds at a time. After they have been in the aquarium for 12 or 24 hours, these 
stops are more frequent and longer. The fishes strike various attitudes during 
these stops, but they seldom rest upon the bottom. Instead they are usually poised 
as if ready to dart away. The body seems so well balanced that they have no trouble 
in maintaining any position they may happen to take. During these stops the tail 
always projects straight backward and the pectoral fins stand at right angles to the 
body. If at this time the sounding board is caused to vibrate, the fish responds 
either with a quick movement of both the tail and pectoral fins or by the pectoral 
fins alone. Twenty fishes were each subjected to 5 tests, and out of the 100 trials 
there were 97 responses and 3 failures. 

(b) Fundulus, with the auditory nerves cut, acts as normal blind fishes do in 
swimming slowly or in resting, but when stimulated, loses entire control of its 
equilibrium. Parker suggests that in resting or swimming slowly the fish depends 
upon the eye for orientation, but in quick movements the ear comes into play. The 
reactions of Amblyopsis seem to support this suggestion, for with both auditory 
nerves cut they have no control of their orientation. When resting, they lie on the 
side or back, either at the surface or on the bottom. In swimming slowly they 
sometimes move forward in irregular lines, but when they attempt rapid locomotion, 
they move in irregular spirals about the long axis of the body and make no progress 
one way or the other. With only one auditory nerve cut the movements are quite 
different. The fish is able to move forward, but it goes in a corkscrew-like path, 
turning over on its axis as it swims along. The same result was obtained by 
Eigenmann by thrusting a pin into one of the auditory organs. 

The operation of eliminating the ear is a comparatively easy one to perform. 
Of those operated on, more than half recovered. They generally lived for 2 or 3 
weeks, and some even longer. The observations were made from i to 2 days after 
the operation. With these fishes three kinds of responses were obtained. If they 
were perfectly quiet when the sounding board was caused to vibrate, they either 
responded by a slight movement of the pectoral fins or by a movement of both 
caudal and pectoral fins. If, on the other hand, they lay with the body quiet and 
with the pectoral fins moving slowly when the sounding board was caused to 
vibrate, they responded by stopping the fin movements. Ten fishes were each 
subjected to 10 tests, and out of 100 tests there were 96 responses. This result 


differs very little from the reaction of fishes not operated upon. Since the ears 
have been eliminated, there is only one conclusion to reach and that is, that 
blind fishes detect vibrations with a frequency of loo per second by means of sense- 
organs in the skin. As stated, I have not been able to eliminate the skin and 
lateral-line organs, and so can not say definitely whether or not the ears play any 
part in the reactions of normal blind fishes. Since the reactions are the same, ear 
or no ear, the part the ear plays in sound-wave perception, if any, is certainly small. 

Using the word "hearing" in the sense in which Kreidl and Parker used it, that 
is, if we define hearing to be the sensation received through the ear and caused by 
vibrations either in the air or water, the experiments cited do not enable one to 
conclude definitely whether the blind fishes hear or not. If they do hear, their 
power in this direction is very limited. 

The rssults show conclusively that they detect waves of loo vibrations per 
second by means of sense-organs in the skin. 



(By E. E. Ramsey.) 

A comparison of the microscopic appearances of the brain of a normal fish 
and that of the blind fish, Amblyopsis spelceus De Kay, discloses a number of inter- 
esting conditions. The optic lobes and the optic tracts are measurably degenerate. 
The hemispheres are larger in Amblyopsis than in the average of normal brains. 
The brains of Campostoma anomalum, Percina caprodes, Eupomotis gibbosus, and 
Amblyopsis were measured with regard to the comparative widths of the optic 
lobes and the hemispheres. Five fishes of the same length were taken of each 
species. The averages obtained are as follows : 


Optic Lobes. 


Comparative widths. 

Campostoma anomalum . . . 

Eupomotis gibbosus 

Percina caprodes 

Amblyopsis spelaeus 






p. ct. 


It is thus seen that the hemispheres are relatively larger in the blind fish than in 
the more normal forms, and that the optic lobes are relatively much smaller in the 

There is no noticeable variation in the cerebellum. In length there is a marked 
shrinkage, chiefly in the optic lobes, as shown by the position of the cerebellum 
which lies directly on the lobes. In the normal brain the cerebellum is situated 
well back, hardly reaching the lobes. The following table gives an idea of the 
length of the brain, as compared with the length of the fish. The brain length is 
measured from the tip of the olfactory lobes to the posterior part of the cerebellum : 





Length of body. 

Length of brain. 

Per cent. 

Length of body. 

Length of brain. 

Per cent. 



























9.8 av. 

6.3 av. 

The result shows the brain of Amblyopsis to be only two-thirds as long as that 
of Campostoma. This shrinkage in width and length is great enough to show itself 
in the extent to which the cranial cavity is filled. A great depth of fatty tissue cov- 
ers the dorsal surface of the brain. The only other external modification of any 
note is the absence of either optic nerves or optic chiasma. 

The optic lobes are normally composed of 7 layers, which from outside to inside 
are as follows: 

(i) A peripheral zone. * 

(2) An optic fiber layer from the optic nerve. 


(3) An optic cell layer. 

(4) A deep cell layer. According to Krause this layer contains in its outer part the cells 
which serve as terminal stations for the optic nerve, and in its inner sublayer the end 
stations for the fifth layer (Marklager). 

(5) A deep fiber layer. 

(6) A granular layer. 

(7) The ependyma and its epithelium, which lies next to the ventricle of the lobes. 

The optic lobes of Amblyopsis show a marked degeneration. The dorsal walls 
are not more than half or two-thirds as thick as those in the normal brain. Its 
contour is so flattened that the ventricle is almost obliterated (i6 in fig. 34 b). The 
torus longitudinalis, which in the normal brain is suspended in the ventricle in the 
median line entirely below the layers of the lobes, is between the lobes and on nearly 
the same level with them. The torus thus forms a commissure connecting the 
lobes. The band of fibers connecting them dips downward in the normal brain 
and crosses to the opposite side through the torus; in the degenerate lobe they 
cross from one side to the other in almost a straight line (15 in fig. 34 b). The 
shrinkage in length is shown in the fact that the hypophysis is crowded forward to 
the anterior level of the lobes- 

Fio. 34. (a) Cross-section of Brain. Amblyopsis spdaus near Anterior part of Optic Lobes. Specimen 77 mm. long. 
(6) Cross-section ttirough Middle of Optic Lobes of Amblyopsis spdcms. Specimen 77 mm. long. I, first 
layer of optic lobe; 2, degenerate optic liber layer; 3, optic celt layer; 4, deep cell layer; s, deep 
fiber layer; 5a, diagonal iibers of deep fiber layer; 6, granulated layer; 7a, optic tract region; 
13, ependyma; 15, torus longitudinalis; 16, ventricle. 

The optic nerve of the normal brain is derived from the second and fourth layers 
of the lobes. The fibers of the second layer pass downward on both sides of the 
lobes, and the inner ones cross over at the ventral surface, where they join the fibers 
of the same layer from the other side. They then continue forward and downward 
to the optic chiasma as the optic tracts. The fifth layer is composed of diagonal 
fibers and descending fibers. These latter nerves pass downward and become a 
part of the optic tract. 

As has been said, the wall of the optic lobes of Amblyopsis has undergone con- 
siderable shrinkage in thickness. The outer layer is not changed. The second 
layer, which is derived from the optic nerve, is entirely wanting. The optic nerve 
is represented by a small bundle of tissue, which is probably the remnant of the 
neurilemma. In the brain where the second layer should be, there is a narrow 
space containing practically no tissue. The third layer is unchanged. The fourth 


layer consists normally of two sublayers ; the outer one has both nerve fibers and 
nerve cells — the latter according to Krause being the terminal stations of the optic 
nerve — and the inner sublayer has the terminal stations of the fifth layer in it. 
The outer sublayer is entirely atrophied in the lobes of the blind fish ; and the inner 
one, if at all present, is indistinguishable from the third layer (3 and 4 in fig. 34 b). 

The fifth layer is reduced to diagonal fibers. The descending fibers which join 
the optic tracts are atrophied. The diagonal fibers are more apparent than in the 
normal brain. These fibers form a broad commissure in the torus longitudinalis, 
which runs laterally to the outer edge of the lobes, where it turns back into the 
substance of the brain just beneath the ventricle and becomes diagonal. Cross- 
sections of fibers arising from various levels of the lobes are shown (5 in fig. 34 b). 

The sixth layer is a granular layer. Its thickness is less than in the normal 
brain. No other change is noticeable. The thickness of the seventh layer, epen- 
dyma, is not more than half that of a normal brain. The cells show some shrinkage. 

The differences in the lobes thus appear to be: first, in the atrophy of the 
second layer; second, the outer sublayer of the fourth layer is entirely gone; 
third, the descending fibers of the fifth layer are wholly wanting ; fourth, the granu- 
lar layer is not so thick and the ependymais not only thinner but reduced in the 
numebr of its cells. 

The optic tracts, that part of the nervous tissue which lies between the optic 
lobes and the optic chiasma, are entirely wanting. The space occupied by these 
tracts in the normal brain is in this brain partially occupied by tissue in which I 
have not been able to make out any structure. All the stains that have been tried 
have failed to reveal any cells. These tracts do not take the stains with the same 
readiness and in the same degree that those in normal brains do when subjected to 
exactly the same treatment. Three fishes, Amblyopsis, Campostoma, and Eupomoiis, 
were killed and the heads placed in Fohl's mixture for the same duration of time. 
The brains were removed from the skull as soon as they were sufficiently hardened 
and were placed in the same bottle in order that the conditions might be alike. 
The three were embedded in the same block and sectioned side by side. The 
tissue of the tracts of the brains of Campostoma and Eupom^tis differentiated very 
well — but the degenerate brain showed no structure. 

In the dissections of the head of the blind fish, I have been unable to find any 
indications of optic nerves leaving the lobes. In both the dissections and the 
sections which have been made of the entire head and brain, there seems to be no 
break in the enveloping membranes on the anterior ventral surface of the lobes 
where the optic nerves originate. The vestiges of the optic nerve can be followed 
backward from the eye for a short distance. The only tracts leading away from 
the lobes are those which connect them with cerebral hemispheres and cerebellum. 
Those which pass forward to the hemispheres are from the diagonal fibers of the 
fifth layer. These pass laterally, but before reaching the lateral aspect of the 
lobes, turn downward through the granular and epithelial layers, and then course 
forward toward the ventral surface of the hemispheres. 



1. Amblyopsis spelaus is found from Mammoth Cave north to Michigan. It is the only 

blind species occurring on both sides of the Ohio. 

2. No direct comparison of specimens from south and north of the Ohio has been made. 

3. There are 3 species of TypMichthys occurring in 3 different localities, one of them north 

of the Ohio. 

4. Troglichthys is confined to the caves of southwestern Missouri. 

5. The 3 species of Chologaster are found in 3 disconnected areas. 

6. The color pattern of Clwlogaster is controlled by the underlying musculature. 

7. Amblyopsis has been permanently bleached so that even individuals reared in the light 

do not acquire color. Its colorless condition is due to the transmission of the environ- 
mental adaptation in past generations of cave-dwellers. 

8. Respiration is probably in part carried on through the skin. 

9. Amblyopsis is a bottom and pelagic (ubiquitous) feeder on living, moving animals. 

10. Chologaster does not depend upon its eyes for detecting and securing prey, or for avoiding 

a rod held in the hand. 

11. Amblyopsis is negatively phototactic. It seeks the dark regardless of the direction or 

wave length of the rays of light. 

12. In well-lighted, open pools Amblyofsis hides under rocks during daylight. 

13. Chologaster when deprived of its eyes is negatively phototactic, and positively stereotropic. 

They are positively tropic to red as against other rays of the spectrum. 

14. Amblyopsis probably breeds during the entire year, but more individuals carry developing 

eggs between March and May. 

15. Amblyopsis is not viviparous, but the eggs to the number of about 70 are carried in the gill 

chamber of the female from fertilization till the larvae are about 10 mm. long. The 
eggs hatch in about a month, having a length of about 5 mm. 

16. There are few, if any, secondary sexual characters which argues in favor of the origin of 

these through sexual selection as against Geddes and Thompson's explanation that they 
are the result of maleness. 

17. In newly hatched Amblyopsis the anus is in the normal position, behind the ventrals. 

When the fish reaches a length of 25 mm., the anus has reached a point in front of the 
ventrals; when 10 mm. longer, the anus has moved forward to between the bases of the 
pectorals. In mature specimens it lies anterior to this point. 

18. The heads of the Amblyopsidae are provided with tactile ridges, rows of tactile organs 

regularly and definitely arranged. 

19. These fishes are not keener in perceiving vibrations than other fishes. They may have 

greater power of discrimination between vibrations. 

20. The ear of the Amblyopsis is normally developed. These fishes do not " hear " in the 

ordinary sense of the word. 

21. The external peculiarities of the brain of Amblyopsis are the absence of optic nerve and 

chiasma; the hemispheres are relatively larger than in other fishes and the optic 
lobes are much smaller. 

22. The dorsal walls of the optic lobes have only half the normal thickness, the differences 

being due to (a) the atrophy of the second layer; (b) the outer part of the fourth 
layer has disappeared ; (c) the descending fibers of the fifth layer are wholly wanting ; 
(d) the granular layer is thinner than normal and the ependyma is thinner and has 
fewer cells; (e) the optic tracts are wanting. 



The Amblyopsidse offer exceptional facilities for the study of the degeneration 
of eyes. They furnish gradations in habits from permanent epigean species to 
species that have for ages been established in caves. The eyes of the following are 
considered : 

1. Chologaster cornutus Agassiz. Locally abundant in the lowland streams and swamps in 

the South Atlantic states from Virginia to Florida. Maximum length about 55 mm. 

2. Chologaster agassizii Putnam. Found in the underground streams of Kentucky and 

Tennessee. It is rare. Maximum length 62 mm. 

3. Chologaster papilliferus Forbes. Found under stones in the springs of Southwestern 

Illinois, in Union and Jackson counties. Maximum length 55 mm. 

4. Amblyopsis spelaus De Kay. Widely distributed in the caves east of the Mississippi both 

north and south of the Ohio River. Maximum length 135 mm. 

5. TypUichthys subterraneus Girard. Found with the latter species in the caves east of the 

Mississippi, but confined as far as known to the south side of the Ohio River. 

6. Troglichthys rosa Eigenmann. Found in the caves west of the Mississippi River. Maxi- 

mum length 55 mm. 

The first tw^o species mentioned live, as far as known, altogether in terranean 
streams; the others, altogether in subterranean streams. Chologaster has well- 
developed eyes, the others mere vestiges. We have thus two epigean species with 
well-developed eyes, one subterranean species with well-developed eyes, and three 
subterranean species with greatly degenerate eyes. The three latter species are 
descended from three distinct terranean ancestors. Amblyopsis is the only member 
of the family possessing ventral fins, and Troglichthys has scleral cartilages which 
are not found in the other members except Amblyopsis. 

It must be apparent that an experiment on a vast scale has been conducted by 
nature, leaving us but to read the results. Moreover the experiment is one in 
evolution without the assistance or intervention of natural selection. 


The only account of the eyes of Chologaster papilliferus Forbes, aside from the 
measurements in the description of the species, is a note by Wright. Professor 
Wright obtained his specimen from Prof. S. A. Forbes, and therefore had C. papil- 
liferus. He announced that the pigment is absent in the pigmentary layer of the 
retina of this species. But this condition was unquestionably either accidental or 
due to the reagents employed. Chromic acid partly or wholly removes the pig- 
ment, leaving the cells in good condition. 

The vertical diameter of the eye in a specimen 39 mm. long is 640 fj-; in a 
specimen 55 mm. long (the largest secured), 960 fi. The distance from the point 
of entrance of the optic nerve to the front of the cornea is 560 /* and 900 fi, respec- 
tively, in the two specimens. The distance from the point of entrance of the optic 
nerve to the front of the epidermis over the eye is 600 /x in the smaller specimens, 
the lens about 360 /a in diameter. For further measurements see the table, page 120. 

The eye is small when compared with that of other fishes of the same size, and 
especially so when compared with the eyes of Zygonectes. It is located high up 
on the side of the head, its upper surface being nearly on a level with the top of the 
head. It is directed outward and forward. In a specimen 35 mm. long it is 1.44 
mm. from the tip of the snout and 0.88 mm. long. The distance between the 
eyes is 1.60 mm. 




Heads seen from above and showing the relative sizes of the eyes of : 

A. Zygonectes notatus; D. Typhlichthys subterraneus, about 35 mm. long; 

B. Chologasfer agassizii, 4 1 mm. long; E. Troglichthys rosas, 38 mm. long; 

C. Chologaster papilliferus, 35 mm. long; F. Amblyopsis spelaeus, 35 mm. long. 



The dermis over the eye is thinner than elsewhere and devoid of pigment. The 
epidermis passes directly over the eye without any free orbital rim. It is much 
thinner, 24 /a in specimen 39 mm. long, than elsewhere about the side of the head 
(50 to 60 fi) and consists solely of epithelial cells ; those at the base are columnar, 
those at the free end of the epidermis are fiat. All the other elements of the 
epidermis — goblet cells and mucous cells, very abundant all about the eye — are 
totally absent over it (fig. 35 a). 


!• 10. 35. (o) Section through Lower Left Half of Iris of Chologasler papUli/erus, seen from in front. 
i, iris; c. cornea; ep, epidermis; d, dermis; sub. o., sutiorbital. 
(i) Section of RiRht Half of Head of Chologasler papilli/trus. 
(c) Section through Retina at Entrance of Optjc Nerve, 
(rf) Inner Surface of Retina nearly tangential at Entrance of Ontic Nerve, 
(e) Vertical Section of PiKnicnt Cells of Retina, depiRmented witK Chromic Acid. 
(/) Tangential Section through figment Cells. Upper part of figure passes through nucleated 
part of cells, middle through processes of cells, and lower through cones oilly. 


The 6 normal eye muscles are present in Chologaster. The 4 rectus muscles 
arise near a common point just behind the point of exit of the optic nerve from the 
skull. The M. rectus superior passes from this point outward, upward, and for- 
ward. The M. rectus inferior passes nearly horizontally outward and forward. 
The M. rectus externus passes nearly straight out at right angles to the axis of the 
body to the posterior face of the bulb. The M. rectus internus is probably the 
longest, passing outward and forward to the anterior face of the eye. 

The two oblique muscles originate near a common point well in front of the 
exit of the optic nerve and are inserted near the insertion of the M. rectus superior 
and inferior. There is nothing remarkable about any of these muscles and they 
are mentioned solely as a basis of comparison with the condition found in Ambly- 
opsis. The space from the wall of the brain case outward about the eye muscles 
and eye is bounded by a connective tissue capsule. Within this capsule, the space 
between the muscles and the posterior part of the optic pit and the eye is filled with 
fat. Above this capsule lies another mass of fat and below it still another (fig. 35 b). 
The supraorbital does not help to protect the eye, which lies entirely lateral from 
it and extends above it. The suborbital bones are thin, hollowed sheets of bone 
backing the suborbital mucous canal. Their number, etc., has not been deter- 
mined, but their location is of importance in view of a statement made by Kohl 
concerning their absence in Troglichthys. 

The sclera is represented by a thin fibrous capsule which is sometimes widely 
separated from the eye by reagents. In the largest specimen it is but 4 /^ thick. It 
is continued over the front of the eye in contact with the dermis as a thin cornea 
(fig. 35 a). This is much more compact than the rest of the sclera. It readily 
separates from the dermis. The sclera is never at any place cartilaginous. I 
was at some trouble to demonstrate the absence of cartilage, even in the largest 
specimen, in order to detect if possible the homologues of the cartilages in Ambly- 
opsis and Troglichthys rosa, and can state positively that no cartilage is found 
associated with the eye of Chologaster papilliferus or in fact with the eye of any of 
the species of Chologaster. 

The choroid is very thin. Just within the sclera is a homogeneous, sometimes 
excessively thin, layer containing a few nuclei, the suprachoroidal lymph space. 
If the eye contracts through reagents, the choroid which clings to the eyeball is 
separated from the sclera by the widening of this space. Pigment is not abundant 
except over the iris and below and at the sides of the entrance of the optic nerve. 
About the entrance of the optic nerve a mass of pigment is prominent, being espe- 
cially conspicuous inlhe largest specimen (fig. 35 c). A mass of pigment which 
may be homologous with this has been described by Ritter in Typhlogobius, who 
found no cellular structure in the pigmented mass in Typhlogobius and identified 
this pigmented mass as the choroid gland. A choroid gland or the rete mirabile is 
not found. A processus falciformis is not present. Blood-vessels are not numer- 
ous and it was impossible to separate a distinct vascular layer of the choroid. In 
the largest specimen the choroid is much richer in blood-vessels ventral of the 
pigmented mass at the entrance of the optic nerve than elsewhere. The capillary 
layer reaches here a total of 9 /a in thickness. A layer of excessively thin pigment 
cells lies close to the pigmented layer of the retina. It is so thin and so closely 
applied to the pigmented layer of the retina that it is only in a few tangential sections 
that this part of the choroid becomes evident. 


The optic nerve enters the retina as a single strand. It spreads out in all direc- 
tions as soon as it has passed the pigmented part of the retina (fig. 35 c). Some of 
the fibers pass behind the ganglionic cells just within the entrance of the optic 
nerve, a condition of importance in the interpretation of the distribution of the optic 
nerve in the blind members of the family. The diameter of the nervous opticus at the 
entrance into the pigment layer is 32 fi in the largest specimen. The nerve is not 
spread out over the ganglionic layer, but is distributed in well-defined tracts between 
the nuclei. There is no nerve fiber layer proper (figs. 35 c, d). These strands of 
fiber not only entirely displace the ganglionic cells along their track, but also plow 
into the granular layer. 

The pigment layer of the retina is very thick, as compared with the other layers, 
a condition recalling that described by Ritter for Typhlogobius and usually to be 
found in degenerate eyes. 

For a comparative statement of the thickness of the various layers of the retina, 
see table on page 120. 

The pigmentary layer is half the total thickness of the retina in the smaller 
specimen, while in the largest it is still thicker, measuring 104 of the 168 /ot of the 
retinal thickness. 

About one-eighth of the outer part of this layer usually appears as a solid mass 
of pigment where the margins of the cells touch. Just within this is a region where 
the cells are contracted, there being large, open, pigmentless spaces; at the inner- 
most part there is again an accumulation of granular or rod-shaped pigment granules 
which obscure almost everything else in the ordinary sections. (See fig. 35 c.) Speci- 
mens preserved in chromic acid lose most or all of their pigment, which becomes 
brownish or disappears. The nuclei of the pigment cells are very irregular in outline 
(fig. 35/), appearing to have no more definite shape than those of white blood 
corpuscles. Hollow processes extend from the cell body downward to near the 
external limiting membrane (figs. 35 e and/). 

About the bodies of the cones the pigment is in thin strands, of which there are 
8 to 12 to each twin cone; farther out it forms a complete ring about them. The 
cones are twins, rarely triplets. The twins are nearly all arranged in such a man- 
ner that the line, which may be termed the axis, connecting the centers of the com- 
ponents of a twin are nearly parallel and form approximately part of an equatorial 
circumference of the eye (fig. 36 b). There is, therefore, no resemblance to the 
condition found in Coregonus and Zygonecies even if we omit for the present the 
consideration of the rods (or single cones). The cones consist of an outer segment 
(80 fi long in the largest specimen) with a tendency to become oblique near their 
outer ends. In chromic preparations these readily split into disks. They stain 
faintly but evenly. They are joined by a translucent interval to the body of the 
cone, an ellipsoid iDody 5 /x -f 10 /a taking on a deep stain (fig. 36 a). These rest 
apparently on a membrane cylinder extending from their base to near the external 
limiting membrane, a distance of 10 yx. Here they rest on a deeply staining cone- 
shaped cell body which pierces the external limiting membrane and is extended 
as a less deeply staining, nodulated process to the outer reticular layer, where it 
spreads out into a cone-shaped base. 

The rods or single cones are very much fewer in number and not regularly 
arranged. They are much fewer than the number of nuclei in the outer unclear 




Beee c 

layer exclusive of the twin cone nuclei. But extending just with- 
out the external limiting membrane a large number of short 
processes are seen between the cone nuclei (fig. 35 e). Whether 
these are degenerate rods, I am unable to say. 

The outer nuclear layer differs materially in the younger, i.e., 
smaller (29 mm.), specimens and in the largest specimens. In the 
younger specimens it consists of several layers of cells exclusive of 
the cone cells, which in this case can be counted with the layer of 
rods and cones. In the larger specimens this is reduced to a single 
layer of nuclei less densely packed, with occasionally a horizontal 
nucleus near the base which is less granular, staining a more uni- 
form color. 

In the largest specimen the outer nuclear layer makes up about 
7 per cent of the total thickness of the retina, in the smaller 
specimens it is slightly thicker, forming 10 per cent of the total 

The outer granular layer differs also materially in the largest 
and smallest specimens. In the largest it forms a thin layer entirely 
free from nuclei and with a total thickness of but 2 or 3 /x. 

In a specimen 39 mm. long this layer is 5 /i thick, distinctly 
granular, contains a few round nuclei — not differing from those 
of the inner nuclear layer. (See fig. 35 c.) These are probably 
members of the layer of fulcrum cells. The latter are not separable 
from the underlying bipolar cells in other regions. In tangential 
sections they appear in groups of two, but are much fewer in 
number than the twin cone cell. The inner layer of the inner 
nuclear layer is composed of distinctly larger cells in the largest 
specimen and separated from the rest by a slight interval. The 

5-7 ^3 00 




Fio. 36. (a) Vertical Section through Retioa of Chologaster pa^U/erus. (6) Section from Outer Margiu of 
Retina to Base of Cone Bodies, (c) Section throuKh Cone Nuclei, (rf) Section throuRh Basal Segments 
of Cones. («) Section through Outer Limiting Membrane, Outer Nuclear Layer with Cone Cell Processes. 
outer Reticular Layer and Outermost Layer of Outer Nuclear Layer. Nuclei in last layer are frequently 
in pairs, but do not correspond to and are much less numerous than twin cones. (/) Section through Lower 
Half of Left Eye seen from behind. The eye was depigmented. Oval nuclei ni. I. at ora serrata. scl,, sclera; 
ciu, choroid; *., iris; {g) Lower part of Iris of Zygonecies noiaius. 


cells here form a distinct row in section, while the rest for the most part are 
irregularly placed. The difference in size is especially noticeable near the entrance 
of the optic nerve. The nuclei are mostly spherical. A few nuclei are found 
more elongate and with their longer axis at right angles to the retina (Mliller's 
fiber nuclei). The largest spherical nuclei measure 5 /i in diameter. The inner 
granular layer varies in thickness and contains few cells. 

The ganglionic layer consists of a single layer of nuclei, rather irregularly placed. 
The nuclei measure 6 /i in diameter. For reasons explained in a previous para- 
graph a distinct nerve-fiber layer is not present. A thin nucleate membrane, the 
hyaloid membrane, containing the blood-vessels, lies directly on the ganglionic layer 

(fig- 35 0- 

It is quite evident from the foregoing that the retina is very much simplified as 

compared with that of Zygonectes. The point of greatest degeneration lies between 
the outer nuclear and inner reticular layers. The horizontal nuclei are all but 
entirely eliminated. The bipolar cells are, in the adult, reduced to two layers of 
nuclei, and the spongioblasts are reduced to a single layer of cells. Even this dis- 
tinction and differentiation is only seen in the largest individuals. Twin cones are 
abundant and apparently not lacking in number and structure, but are arranged 
in a different manner. Rods are much fewer in number than in either Coregonus 
or Zygonectes. 

The chief difference between the youngest and oldest specimens of papilliferus 
examined lies in the thickness of the pigmented layer and the outer nucleated and 
the outer granular layers. The relative thickness of the pigmented layer increases 
very much with age. 

The irideal region needs a few words since its structure helps to explain certain 
conditions in the blind fishes. The epithelial part is composed of two layers of 
cubical cells, of which the outer are the larger. The 
outer cells are normally filled with pigment to such 0'-^r\ 

an extent that their outlines can not be made out, the <^ 00 (] O 

inner cells are free from pigment. The outer layer .0„OO d Q '^Q 
passes directly over into the pigmented layer of the ^ CA '^0n'Su 
retina. Wheretheinnerlayerofthe iris merges into the c9 '^'^OQ^^ ^ 

inner layers of the retina it is composed of a group of (->, Qm ^rT^O (v f) 
cells with elongated nuclei (fig. 36/, n/./.). The uveal ^^"^Q (J 
part of the iris is composed of a thin layer of cells a 

with irregular nuclei, and the pigment cells of this vxc. 37- Nuclei of Epuheiiai Layer oc 

,,. ,.,,.,. Lens o' Chologastcr. 

layer are much thmner than the epithelial pigment. 

The ligamentum ciliary does not contain many muscle fibers, but is abundantly 
supplied with granular nuclei. The things of greatest importance are these granu- 
lar nuclei, the epithelial pigment and the oval nuclei at the ora serrata. As com- 
pared with the same region in other fishes the shortness of the section of the iris is 
at once striking (fig. 36 g). The absence of ciliary muscles and the insignificance 
of Decemet's membrane are also notable. 

The lens offers no peculiarities. The shape of its epithelial nuclei may be seen 
in figure 37. 




Only a single specimen of this species appears to have been put on record. 
Putnam described it from Lebanon, Tenn. The present account is based on live 
specimens secured by me in the river Styx in Mammoth Cave and in Cedar 

The eye of Chologaster agassizii Putnam is much smaller than that of C 
papilliferus . In a specimen 41 mm. long it is placed 2.08 mm. from tip of the 
snout, the eye measuring 0.72 mm. in diameter. The distance from eye to eye 
is 2.72 mm. It is elliptical in outline, with the lateral face depressed. It is directed 
outward. The optic nerve, which, at its origin, is surrounded by pigment for a 
distance of 2.4 mm., extends almost straight inward. The dermis over the eye is 
essentially as in papilliferus. The epidermis is less simplified. It is thinner than 
in the surrounding tissue, but goblet cells are found in it, although they are much 


Fig. 38. Chologaster agassizii. 

(a) Cross-section of Part of Head. 

(6) Vertical Section through Retina of a Specimen 38 mm. long. 

(c) Vertical Section through Retina of a Specimen 62 mm. long. Rods, Cones, and Pigment Layers omitted. 

(d) Lower part of Iris of the Same Specimen, 62 mm. long. 

smaller and much less numerous than elsewhere. The sclera and choroid are as 
in papilliferus, including a pigment mass below the exit of the optic nerve just 
within the sclera. The optic nerve measures 24 /x at its point of entrance into the 
pigment layer of the retina, and it is thus one-fourth smaller in diameter than in 

The proportionate thickness of the retinal layers as compared with the 
layers of papilliferus is seen in the table. The maximum thickness in the largest 
specimen is but 130 /a as compared with 166 fi in papilliferus. This differ- 
ence is almost entirely due to the thickness of the pigment layer, which is 74 fi 
in the largest agassizii and 104 /i. in the largest papilliferus, leaving a difference 
of but 6 ft in the other layers. The pigmented layer is, on an average, much thinner 


than in papilliferus. Yet the per cent of the total thickness of the retina in pig- 
ment is larger than in normal fishes. The nuclei of the pigmented epithelium are 
irregular in outline. The part of the pigment layer about the nuclei forms a mass 
of pigment in which cell boundaries can not always be made out. The pigment 
alx)ut the nucleus is in granules ; farther in, about the cone bodies, it is in prisms. 
I have not been able to make out rods. The cones are irregularly elongate so that 
the cone bodies are at various heights. The pattern of the twin cones has, there- 
fore, not been made out. 

The outer nuclear layer consists of nuclei conical in shape, partly outside the 
outer limiting membrane as in papilliferus, and a number of oval nuclei form- 
ing a double series within these in the younger, a single series in the older 

The outer reticular layer is distinct to the iris. Horizontal cells could not, with 
certainty, be identified. Some of the cells lie without the inner nuclear layer in 
the outer reticular layer and may be fulcrum cells. The inner nuclear layer is 
three to four series of cells deep. Miillerian nuclei are present. If artificial 
splitting should take place, the innermost series of nuclei separates from the outer 
layers ; these probably correspond to the spongioblast cells of other retinas. The 
inner reticular layer is well defined and contains very few cells. The ganglionic 
layer consists of a single series of nuclei. A distinct optic fiber layer is not 

The iris is much as in Chologaster papilliferus, much shorter in section than 
in Chologaster cornutus. The inner cells of the retinal part are pigmented around 
the margins of the pupil, while in papilliferus only the outer cells carry pigment. 


The eye of Chologaster cornutus Agassiz is much larger than that of the other 
species of the genus. The retina on the other hand is simpler. The details of 
the measurements are given at the end of the account of this eye. But two 
specimens were available for examination; they were preserved in alcohol and 
respectively 27 and 43 mm. long. The very remarkable retina deserves much 
fuller treatment than is possible with the limited material available. 

Leaving out of consideration the accessory structures of the eye as choroid, 
sclera, muscles, etc., which are scarcely if at all different from the same structures 
in papilliferus, the retinal characters may be briefly described. 

The pigment layer is very thick as compared with the rest of the retina, form- 
ing over 60 per cent of the total thickness. The pigment cells form a sheath com- 
mon to any pair of the twin cones. 

Connections between the cones and the outer nuclei could not be made out. 
There are apparently fewer cones than nuclei. For the relation of the cones to 
the underlying cells and of the latter to the nuclei of the inner nuclear layer, see 
figures 39 c and d. 

The outer nuclear layer consists of a series of nuclei closely packed together 
with their longer axes vertical. Occasionally a fainter staining nucleus is 
found among the bases of these cells with its longer axis horizontal (figs. 39 a 
and 40 6). 



The outer reticular layer is well developed. Its boundary is irregular on 
the side of the inner nuclear layer, but more regular on the side of the outer 
nuclear layer. 

Horizontal cells are very few and widely separated, if, indeed, this layer is repre- 
sented at all. A few cells horizontally placed are present on the inner face of the 
outer reticular layer (fig. 39 a). 

The inner nuclear layer is represented in the smaller specimen by two series 
of small rounded nuclei (fig. 39 a, 5-7). In the larger specimen a single irregular 
series represents this layer (figs. 40 6, c, 5-7). Besides the rounded nuclei there are 
a few irregular-shaped ones and other elongated ones. Some of the latter lie in 
the plane of this layer, others at right angles to it. The latter are probably Miil- 
lerian nuclei. 

OMOg^^ C "0 ^o. 

"S o 


Fig. 39. Chologaster cornulus from a Specimen 27 mm. long. 

(a) Entrance to Optic Nerve and Part of Retina, 2 mm. and 6. 

(b) Oblique Section through Pigment Layer to near Outer Nuclear Layer, 2 mm. and 4. 

(c) Bases of Cones and Underlying Nuclei of Outer Layer. Nuclei, in black, are in deeper focus, a mm. and 8. 

(d) Nuclei of Outer Nuclear Layer and Deeper-lying Nuclei of Inner Nuclear Layer, 2 ram. and 8. 

The inner reticular layer is well developed and contains a few round nuclei, as 
in papilliferus. In addition, it contains some vertically elongated nuclei at times 
reaching through half the thickness of the layer. These are also evidently Miil- 
lerian nuclei. Some of them extend from the ganglionic layer outward, others 
from the inner nuclear layer inward (fig. 40 b). 

The ganglionic layer is very imperfect, being represented by scattered nuclei 
embedded in the inner layer of the reticular layer. In this layer we have a decided 
degeneration by a reduction of the number of elements (fig. 40 a, 9). 

A nerve-fiber layer is not evident in cross-section. 

The pigmented layer has not been decreased nor have the reticular layers 
degenerated materially beyond Chologaster papilliferus. The nuclear layers, on 
the other hand, have been very materially affected. The outer layer has been 
much reduced. But this need not necessarily imply degeneration. The inner 



nuclear layer has been reduced one-third and more from the lowest point in papilli- 
ferus. There is no longer any definite difference between the inner spongiose and 
outer bipolar cells of this layer, a difference that is usually well marked and is still 
evident in papillifcrus. An equally marked change has unquestionably occurred 
in the ganglionic layer where a layer of cells, continuous but for the strands of the 
n. opticus passing between them, has dwindled to irregularly scattered cells. 

Oe t • •• V 










8- -. 




8— . 









.^ "^ 

Fio. 40, (a) Section Tangential to Ganglionic Layer, showing Distribution of Ganglionic Nuclei, 9. 

On Left, 4-7, Row of Nuclei of Inner Nuclear Layer, 2 mni. and 4. 
(6) Section of Retina through Old Individual (47 mm. long). Pigmented Layer left 

Blank. All Nuclei as seen in one Focus except Vertical Miillerian Nucleus, which 

is from Another Section. 2 mm. and 4. 
(c) Fragment of Same Retina at Another Point. 
((/) Cells of Lens Epithelium, Surface and Tangential, 2 mm. and 4. 
(«) Cells and Blood Cells from Hyaloid Membrane. 

The position of the Miillerian fiber nuclei is also unique in this retina. 

The eye is in some respects more degenerate than that of Typhlichthys sub- 
terraneus. The dioptric arrangements in this eye and the cones are better developed 
and the layers in general are better differentiated than in T. subterraneus, 
but the nuclear layers are in the latter species composed of more series of cells. 
A section of the iris is much longer than in either of the other species of this genus. 
Since the differences in the eye and retina of the species of Chologasier are largely 
a matter of measurements, the following tables are added : 



Measurements of the Eyes of Chologaster in Groups. 

[Treated alike.] 



Medio-lateral diameter 





with the lens. 

a. Eve dissected out and measured directly : 






C. cornutus Agassiz, 32 mm. long 





45 mm 




C. papillifcrus Forbes, 32 mm. long 



(720 without lens) 



C. agassizii Putnam, 39 mm. long 






b. Head mounted in balsam, the eyes measured 


from above : 

without the lens 

C. papilliferus Forbes, 35 mm. long 


(612 without lens) 
(688 with cornea) 

C. agassizii Putnam, 41 mm. long 


(576 with cornea) 

c. Heads sectioned in paraffine : 


C. cornutus Agassiz, 27 mm. long 




C. papilliferus Forbes, 39 mm. long 



560 with cornea 


C. agassizii Putnam, 38 mm. long 


738 about 

520 with cornea 


Measurements of the Retina of Species of Chologaster. 

[Oaly averages £rom two to nine measurements are given in each case.] 

C. cornutus. 

C. papilliferus. 

C. agassizii. 

43 mm. 

39-39 mm. 

55 mni. 

38 mm. 

63 mm. 


Outer nuclear 






















Outer reticular 

Inner reticular 








It is seen that the retina of agassizii differs from that of papilliferus almost alto- 
gether in the decrease of the thickness of the pigment epithelium. The retina of 
cornutus differs from that of agassizii in the reduction of the layers inside of the 
pigment epithelium. 


The eye of Typhlichthys subterraneus has not heretofore been made the sub- 
ject of study. The following account is based on 3 specimens, 20, 25, and 45 mm. 
long respectively, from a small cave in the town of Glasgow, Kentucky, and a 
number of specimens of various sizes, the largest 54 mm., from Mammoth Cave, 
Kentucky. These were all collected by myself in the early part of September, 1897. 

The eye of this species is in general less degenerate than that of Amblyopsis. 
The accessory structures are, on the other hand, much more degenerate than in 
Amblyopsis. The eye can not be seen from the surface. The region of the eye 
is, however, more conspicuously apparent than in Amblyopsis on account of the 
thinner tissues of this smaller species through which the orbital fat-mass can be 
seen. The 'eye can not be seen even in heads cleared vnth oil on account of 
the almost total absence of pigment about the eye and its total absence in the 
eye itself. 


The eye is surrounded by a large mass of fat through which connective tissue 
cells are scattered. A distinct separation of the orbital fat from the other fatty 
tissues in this neighborhood by connective tissue membranes such as are found in 
Amblyopsis is not noticeable in this species. A few pigment cells are found scattered 
through the fat-mass. They are nowhere massed together so as to become evi- 
dent to the naked eye. In one eye not a single pigment cell is found about its 
surface, in another three are found on the surface of the connective tissue sur- 
rounding the eye. In no case is the pigment about the eye of any significance, 
for it is as abundantly found throughout the fatty tissue surrounding it. 

No trace of eye muscles are present. Scleral cartilages are entirely absent, a 
condition in striking contrast to that found in Troglichthys rosm, with which this 
species has been confounded. 

Sclera and Choroid. — The sclera and choroid coats are not separable in this 
species. In specimens up to 40 mm. in length the eye is surrounded by a very 
thin membrane containing here and there a nucleus, and in the region of the choroid 
fissure and near the exit of the optic nerve a few capillaries. In the oldest speci- 
men, 54 mm. long, the tissues about the eye are distinctly more fibrous, but even 
here I have not been able to separate the layers. From the front of the eye a 
strand of tissue similar to that surrounding the eye extends outward. A blood- 
vessel reaches the eye with the optic nerve, and a few capillaries are found on the 
surface of the eye and in the hyaloid membrane, but the details of their distribu- 
tion I have not made out. This primitive condition of the outer layers of the eye 
is not so striking as at first appears when the conditions in Chologaster are taken 
into consideration, for even in Chologaster the choroid and sclera are insignificant. 

The Eyeball. — The eye is on an average 1.68 mm. in diameter and has reached 
this size when the individual has reached 25 mm. in length. In specimens of this 
length the cells of the retina are still undergoing division. In a specimen 20 mm. 
long it has a diameter of 1.42 mm. Its maximum differentiation is not reached 
at the time it first reaches its maximum diameter. The eye is probably potentially 
functional throughout life as a light-perceiving organ. A minute vitreal cavity, 
remnants of the hyaloid with its blood-vessels, outer and inner nuclear as well as 
inner, and usually also the outer reticular layers are well differentiated, and the 
optic nerve is certainly still connected with the brain at a time when the fish has 
reached a length of 40 mm. 

The position of the eye is not fixed, so that in different series of sections, pre- 
sumably cutting the head in the same planes, the choroid fissure occupies various 
positions and the eyes are cut in various directions. With this general sketch the 
various layers may be taken up in detail. 

Pigment Layer (i in figs. 41 a, 43 c). — No pigment granules are present in the 
eye, a condition in great contrast to that in either Amblyopsis or Chologaster, where 
the pigment is least affected by the degeneration processes. The absence of 
pigment in this eye is indeed unique among vertebrates. Whether pigment 
is developed in earlier stages and disappears I have not been able to determine. 
In the specimens 40 mm. and less in length the pigment layer consists of a series 
of cells, but little separated from the underlying outer nucleated layer. The sepa- 
ration between the layers is greatest near the exit of the nerve and at the iris. In 
older individuals a considerable space is formed between the pigment layer and 



the outer nucleated layer on the dorsal and proximal parts of the eye, but since in 
all of the cases under consideration a good share of this space is attributable to 
reagents, a more detailed description is useless. However, in these regions delicate 
protoplasmic processes extend inward to the nucleated layer. The nuclei of the 
pigmented layer stain much more faintly than those of the rest of the retina with 
Biondi-Ehrlich, but just as deeply as the others with haemalum. The cells of the 
pigment layer are in one series, but occasionally a cell is found below the level 
of the rest. A few cells very elongate in section may be mentioned here. They 
were found (fig. 41 c) on the inner face of the pigment layer. These are important 
in the interpretation of the structure of the eye of Troglichlhys rosce, where they 
are also found. Their origin and significance are not known.' 

Fig. 41. (a) Sagittal Section through Right Eye of Typhlichthys subtcrraneus, 25 mm. long. 
(6) Miilierian Nuclei (?) from Retina of Individual 25 mm. long, 
(c) Horizontal Section of Eye of Individual 40 mm. long. 

Rods and Cones with their Nuclei. — While the outer nuclear layer is very well 
developed indeed, the rods and cones are not definite. In the most highly 
developed eye there is a distinct outer limiting membrane. Without this are filmy 
processes continuous with those from the pigment cells. Very rarely one sees an 
elliptical, slightly granular body which may or may not be a cone body. The 
outer nuclear layer is in some cases quite distinct, consisting of a compact series 
of outer (cone?) nuclei, irregularly elliptical in outline, below which are a few 
cells of a second series (rod nuclei ?) sometimes with their longer axes parallel with 
those of the outer layer, sometimes horizontally disposed. 

' See also Rhineura. 


Cells of bizarre appearance were noted near the iris in one of the younger indi- 
viduals (fig. 41 b). Some of these are long, club-shaped, with rounded end turned 
inward, others the reverse, still others with long, elliptical outer segments and smaller 
inner segments. 

The cones are certainly less developed than in Amblyopsis, while the reverse 
is the case with the nuclei belonging to them. 

The outer nuclear layer seems but little more degenerate than in Chologaster as 
far as differentiation is concerned, being of course very much more limited in extent. 

Outer Reticular Layer (4 in the figures). — A distinct break between the outer 
and inner nuclear layers, of varying thickness, where nuclei are absent or few 
and far between, is present. A distinct boundary line for this layer does not exist, 
and a reticulate appearance is only to be seen in short stretches, otherwise the 
layer is only distinguished in the preparations by the absence of nuclei. In the 
younger specimens examined this layer is not differentiated, the nuclear layers 
forming one continuous structure. It is quite evident from this that tissue differen- 
tiation is not completed in the eyes of this species till very late. 

I'he Inner Nuclear Layer. — The nuclei of the inner layer are of two sorts, larger, 
granular, more faintly staining ones, and smaller, more homogeneous, deeper 
staining ones. In one individual they are seen to be surrounded by a compara- 
tively large cell body whose outlines are made distinct by the branches of the 
Miillerian fibers. In thickness this layer exceeds both the nuclear layers in 
Amblyopsis. It was not possible to identify nuclei belonging to the Miillerian 
fibers as such. Supporting fibers can be followed in some individuals from the 
ganglionic layer through the inner reticular and the inner nuclear layers, in which 
they branch to send processes between the regular cells (fig. 41 c). Once peculiar 
horizontal nuclei were noticed on the inner face of this layer. They are marked 
y in figure 41 a. 

The Inner Reticular Layer. — Horizontal cells are not present in the inner reticu- 
lar layer. Otherwise the layer offers no peculiarities. Owing to the persistence 
of the union of the lips of the choroid fissure and the consequent merging of the 
ganglionic into the outer layers at this point, the inner reticular layer appears 
horseshoe-shaped in a vertical longitudinal section (fig. 41 a, 8). In a section going 
through the plane of the choroid fissure (fig. 42 a, 8, and plate 3, fig. D, of Rhineura) 
it appears as a central area in the eye, free from nuclei. This condition, which is 
seen in all but the eyes of the oldest individuals, is of importance in interpreting 
the conditions seen in Troglichthys rosce. In the older individuals the nuclear 
layers become thin on either side of the choroid lips and the reticular layer ap- 
proaches the pigment layer (fig. 41 c). The layer is well developed. Its relative 
thickness may be gathered from the comparative table. 

The Ganglionic Layer. — There is no distinct optic fiber layer. The ganglionic 
layer consists of a single layer of cells irregularly disposed about the vitreal cavity 
where this is present and forming a solid core of cells behind the vitreal region 
inclosing blood-vessels and hyaloid nuclei. Some of the cells appear to send fibers 
into the inner nuclear layer in the older retinas. These may be Mullerian nuclei, 
since in Chologaster cornutus such are found in this layer. The total number of 
nuclei counted in one example as belonging to this layer is 100, not very greatly 
different from the number noticed in specimens of Amblyopsis. In specimens up 



to 40 mm. in length the choroid fissure is a well-marked structure. The pigment 
layer and inner layers merge into each other here, and the ganglionic layer is con- 
tinuous with the pigment layer. As stated above, the inner reticular layer does not 
surround the ganglionic layer at this point. A vertical longitudinal section of the 
eye has the general appearance of a section through a Graafian follicle (fig. 41 a). 
The ovum would correspond in position to a cell in the ganglionic layer, the stalk 

Fig. 42. (a) Vertical Section through Left Eye of Individual 25 mm. long. 

(b) Vertical Section through Left Eye of Specimen 40 mm. long. 

Uvea shows well as a Series of Elongated Nuclei, »/. 7. 

through Ganglionic Layer, does not pass through Pupil. 

Inner Layer of Cells of 
Section, while passing 

of the ovum to the lips of the fused choroidal fissure, the outer follicular cells to 
the nuclear layers, and the interior cavity of the follicle to the inner reticular layer 
of the eye. 

Optic Nerve. — The optic nerve is not as distinct at its exit from the ganglionic 
layer as in Amblyopsis, but in specimens even 40 mm. long there is no difficulty 
in tracing it to the brain. In specimens of the latter size it has a diameter of 9 /t. 
It contains many elongated nuclei, some of which are also seen with the optic fibers 
within the eye (fig. 42 b). The covering of the optic nerve partakes of the same 



indefinite nature as that of the eye itself, with which it is continuous. No pigment 
accompanies the nerve as a distinct layer, but here and there, as in the covering of 
the eye, a pigment cell may be seen, while about its entrance into the brain cavity 
some pigment cells are also found. 

Epithelial Part of the Iris. — The pigment cells, as in Amblyopsis, decrease in 
height toward the irideal portion of the retina, where they become a series of pave- 
ment cells with rounded nuclei directly continuous with a layer of cells with elon- 
gate elliptically nucleated cells forming the inner layer of the iris. The homologues 
of the elliptically nucleated cells are found in the iris of Chologaster in the region 
of the ora serrata. At the junction of the outer and inner layers of the iris the 
cells are sometimes heaped up, making the irideal margin quite thick (fig. 43 b). 
There is in some cases a distinct free pupil (fig. 43,) while frequently the opening 
is directly continuous with the choroid fissure which may remain open in this 
region (fig. 41 c). 

Fig. 43. (a> Iris of Eye shown in 43 a. 

\o) Section througli Iris and Lens of Right Eye of Typhlichthys 4a mm. long, 
(c) Median Vertical Section of Left Eye of Same Individual. 

Lens. — The lens was not found in all eyes ; when present it is situated at the 
anterior end of the choroid fissure or behind the iris. It consists of but very few 
cells. These cells are undifferentiated. No fibers or other signs of differentiation 
are at all evident. The lens cells are not distinguishable from the neighboring 
cells, and only the faint lines seen to surround the group serve to distinguish 

Vitreous Body and Hyaloid. — ■ The choroid fissure is distinctly evident in speci- 
mens at least 42 mm. long, not as a distinct fissure, except in front, but as a line 
along which the various nucleated layers of the retina are merged. In the distal 
part of the retina the fissure is not entirely closed, and it here leaves an opening 
into the vitreous cavity which is more distinct and larger in the large specimens 
than in the smaller ones (fig. 41 c). The vitreous cavity, when present at all, is 
confined to a very narrow region just behind the lens. Here a few oval nuclei 
and an abundant supply of blood-vessels are to be found (figs. 41 c, 43 a, b), 
the latter communicating with the exterior through the open part of the choroid 
fissure. The vitreal body or cavity does not extend far into the eye, and in the core 
of ganglionic nuclei, where the vitreal cavity does not extend, the hyaloid mem- 



brane is represented by blood corpuscles and by a few cells with elongated nuclei 
whose longer diameters are parallel with the optic nerve. 

Measurements of the Eyes of Typhlichthys sublerraneus. 

Length of fish. 

Diameter of eye, 

Diameter of eye, 
































































In December, 1889, Carman published an account of cave animals collected 
by Miss Ruth Hoppin in Jasper County, Missouri. Among them were a num- 
ber of what were supposed to be Typhlichthys sublerraneus Cirard. A compari- 
son of the eyes of two of the specimens collected by Miss Hoppin with the eyes of 
specimens of Typhlichthys sublerraneus from Mammoth Cave showed that the 
western specimens represented a distinct species, and that Kohl must have based 
his account of the eye of Typhlichthys on specimens from Missouri. 

In the spring of 1897, I visited the caves examined by Miss Hoppin, at Sarcoxie, 
Missouri, but as my stay was limited and the caves were full of water I did not 
succeed in getting any additional material. In September, 1898, through a grant 
from the Elizabeth Thompson Science Fund I was enabled to make another and 
this time successful effort to secure this highly interesting material. 

Kohl described the eyes of Typhlichthys, basing his account on two specimens 
respectively 36 and 38 mm. long. Dr. Mark informed me that at least one of 
these specimens came from Missouri, and Kohl's account was certainly drawn 
from Missouri specimens only. 

He found that the bulbus is nearly spherical, with a diameter of 0.04 mm. 
The orbit is a very flat cavity that offers little protection to the eye. Suborbitals 
are totally wanting and in their place is a cartilaginous protecting capsule, placed 
over the bulbus dorsally and laterally, and made up of several cartilaginous plates 
0.02 mm. thick. Between the plates the connective tissue frequently contains 
thick and large nuclei which are sometimes united into groups. One such mass 
he thinks has been taken for the lens by Wyman (Putnam, fig. 5). It lies 0.195 
mm. from the outer surface of the epidermis. All tissues covering the eye show 
absolutely no difference from neighboring parts. Eye muscles are not found, but 
sometimes there are stiff connective tissue strands connecting the cartilaginous 
bands with the tissues immediately surrounding the eye. The eye in the speci- 
mens examined he considers in the stage of the formation of the secondary eye 
vesicle. There is still a large cavity present representing the primitive eye cavity 
which is only being encroached upon by the invaginating outer cells, which in 
part are precociously ganglionic, sending each a process to the optic stalk. The 
optic stalk no longer shows a cavity, which he assumes became obliterated by the 
direct ingrowth of nerve fibrils and not in the usual way. The invagination of 


the inner layer may have progressed farther in one eye than in the other, but there 
is always a considerable space still left between the inner and the outer layers of 
the primitive eye vesicle. The elements of the inner layer, the ganglionic cells, 
he found to send their processes directly inward. They must have gradually re- 
volved, since in the normal eye the nerve processes are directed outward. Some 
of the fibers cross each other on their way to the outlet for the nerve. Not all of 
the invaginated cells send processes. Among those that do there are smaller, round 
cells without a trace of fibers. From these the rest of the nervous parts of the 
retina, including of course other ganglionic cells, would probably have arisen. 
The outer layer of the secondary eye vesicle is also single-layered. The cells are 
elongate, with oval nuclei, and without a definite arrangement. They are con- 
nected with the few cells of the optic stalk that still remain. Connective tissue 
cells are found in the nervus opticus. They are probably mechanically active in de- 
generation by separating the elements. He found no sheath to the optic nerve, 
as described by Wyman. The lens he found to be a spherical cell heap o.oi mm. 
in diameter in the distal pole of the eye. It lies just within the sclera and the cup 
of invagination. The sclera is made up of several layers of very fine fibrillae. 
Nuclei are not found in it, but nuclei are found on its outer surface. No vessels 
are found in the choroid, which consists of connective tissue cells more numerous 
on the dorsal than on the ventral surface. The Typhlichthys eye is "absolut 
pigmentlos." The surrounding tissues are rich in pigment, which, however, is not 
related to the eye. There are pigment masses found here and there, but especially 
between the bulb and the cartilaginous capsule. 

It is hard to arrive at the proper explanation of the structure of this highly 
degenerate eye even with an abundance of material, and it is probably not to be 
wondered at that Kohl in the work outlined above did not see the eye muscles, 
mistook the sclera for suborbitals, parts of the retina for the choroid, interpreted 
the pigmented epithelium of the eye as an extra optic pigment mass, mistook the 
inner reticular layer for the primary optic cavity, the nuclear layers for the pig- 
ment epithelium, etc., and arrived at a thoroughly erroneous idea of the general 
structure of the eye and based his theories on the degeneration of eyes in general 
on his conception of the structure of this eye. The invaginating cells of the 
primary optic vesicle are supposed to have been directly converted into the gang- 
lionic cells, which are usually among the very last products of the histogenesis of 
the retina.* 

By supposing that the eye was arrested at the beginning of the invagination, 
and that the invaginating cells rotated on their axes and were converted directly 
into ganglionic cells, Kohl derived the nucleated layers from the outer pigment- 
producing layer of the primary vesicle, at the same time ruling the pigment layer 
out of the eye. 

The eye is very small and situated so deep that it is impossible to see it from 
the surface (fig. 44 a). In the upper half of a head cleared in xylol it is just evi- 
dent to the naked eye as a minute black dot (figs. 44 b, c). As in Typhlichthys 
and in Amblyopsis, it is surrounded by a fat-mass filling the orbit. It is not at all 

' The mistakes of Kohl, esperially as far as they are the result of criticising work done on Amblyopsis while 
he was working on another species, seem to me to point a moral. A certain species must not be too readily taken 
as an exponent of a family, order, or class, and a knowledge of related species and geographical distribution is not 
altogether to be neglected. 



uniform in shape in different individuals or even the two sides of the same indi- 
vidual. It can be located and seen in cleared heads solely on account of the pig- 
ment which is always abundant over the distal face of the eye. It is located so far 
beneath the surface as to occasionally lie in contact with the brain case nearly 
opposite the posterior end of the olfactory lobe. It has thus been withdrawn 
much farther than in the other blind species. 

It is very much smaller than the eye of either T. subterraneus or Amhlyopsis. 
Its size is, however, quite variable, measuring 40, 49, 56, 64, 54 by 96, 56 by 120 /* 
in different instances, exclusive of choroid and sclera. 

6ra in 

Fig, 44. (o) Cross-section of Part of Head of Troglichlhys, 25 mm. long, showing Position and 
Proportions of Eye. 
(6) Head of Troglickthys from above, showing Relative Positions of Tactile Organs and Eyes, 
(c) Part of Same Head, showing Eyes with their Peculiar Pigmentation and E>istribution 
of Pigment Cells in Surrounding Tissues, 

The muscles of the eye were in no case normal. I have not found more than 
two rectus or more than one oblique muscle belonging to any one eye. They can 
best be made out from horizontal sections. In cross-sections it is very difficult to 
identify or follow them. 

The best-developed rectus was found in a specimen 35 mm. long. It is com- 
posed of a number of normal fibers forming a bundle 20 /u, in thickness, and from 
its origin to its insertion it is 256 fi long. The remarkable peculiarity of this muscle 
is that 100 /A of this is a tendon 4 /* in thickness (fig. 46 b, msc. r.). The tendon spreads 
into a cone-shaped mass of fibers attached to the pro.ximal face of the eye. Traces 
of two muscles were made out connected with the right eye of another individual. 



The oblique muscle is attached by a tendon to the face of the eye opposite 
that of the attachment of the rectus (fig. 46 a, msc). In the best-developed condi- 
tion it was found to be but 9 /t in diameter, taking its origin at a point on the level 
of the lower surface of the olfactory nerve where the latter pierces the ethmoid 
and 160 fi laterad from it. The muscle itself is in this instance about 200 /x in 
length and is attached to the eye by a tendon of equal length. The rectus in the 
same individual is 208 /a long. 

In all the cases enumerated above the muscles of the opposite side were not 
nearly so well developed. In the one with the weU-developed rectus the oblique 
was indistinct, while in the one with the well-developed oblique the rectus is also 
well developed, but the striations are not distinct. 

The scleral cartilages form one of the striking features of this eye. They are 
quite variable, forming a more or less complete covering for the eye. In some they 
are several times as long as the eye and in such cases extend much beyond the 
eye. In one eye 49 /* in diameter the length of one of the cartilages reaches 160 /x 
(fig. 45 a). They have not kept pace in their reduction with the reduction of the 

Fig. 45. (a and b) Two Cross-sections of Eye of Specimen preserved in Alcoiiol, 38 mm. long. Sec- 
tions show Variable Extent of Pigment, Ciioroidal (ck,) Pigment, and Scleral 
Cartilages. Extent of latter represented by dotted lines in figure a. 

eye in size. As a consequence individual cartilages either extend beyond the eye 
or are bent at acute angles in their endeavor to apply themselves to the shrunken 
eye (fig. 46 a, scl.c). These cartilages were mistaken for the suborbital bones by 
Kohl. There is absolutely no ground for this supposition. The suborbitals are 
present (fig. 44 a, subo.) and widely separated from these cartilages. Further, the 
eye muscles are attached to the cartilages and to similar ones in Amblyopsis. 

The presence of these large cartilages is the more remarkable when we con- 
sider that none are found in Typhlichthys subterraneus, and in the species of 
Chologaster, which in other respects resemble Typhlichthys in all but the develop- 
ment of the eye and the color. It is quite evident that Troglichthys and Typh- 
lichthys are not derived from a common ancestor (except, of course, remotely). 
Their present superficial resemblances are the result of converging development 
under similar environments. A species similar to Chologaster agassizii gave rise 
to Typhlichthys subterraneus. What the ancestry is of Amblyopsis and of Tro- 
glichthys is not known. The cartilages are bound together by an abundant fibrous 
connective tissue containing a few corpuscles. (These I have found nowhere as 
abundantly as represented by Kohl.) 



The choroid, in so far as this layer can be distinguished from the sclera, consists 
of a dense layer of fibers closely applied to the eye. Over the distal surface it is 
split into two layers between which there are a greater or smaller number of pig- 
ment masses (fig. 45 b, ch.). These would prove effective to prevent the performance 
of the natural function of the eye were it functional. Pigment cells are much more 
sparingly found in other parts of the choroid. Blood-vessels are very few in number, 
a condition to be expected in such a minute organ. This layer was mistaken for 
the sclera by Kohl. 

The eye proper of Amblyopsis differs very greatly in different individuals, 
but in general it maintains a certain degree of development from which the many 
individual variations radiate. The eye of TrogUchthys rosa has similarly a 
general type of structure which is maintained, but with many variations. This 
type is more degenerate than that of either Amblyopsis or Typhlichthys subterraneus. 

Fio. 46. Two Horizontal Sections through Eye, showing Extent of Scleral (scl.) Cartilages and Tendons of 
Oblique (a, msc.) and Rectus Muscles (6, msc.r). Fig. a represents section just above Fig. b, from an 
individual 34 mm. long. Drawn under magnification of 560 diameters. 

The eye of TrogUchthys has been derived from an eye like that of Amblyopsis by 
the disappearance of pigment from the posterior part of the retina and the reduc- 
tion of the central mass of ganglionic cells to the vanishing point. In the most 
highly developed eye of T. rosce (9, fig. 47) I found but three of these cells. Both 
in size and in structure the eye of T. roscR is the most rudimentary of vertebrate 
eyes so far known, except that of Ipnops which is said to have vanished. 

The vitreous cavity and the hyaloid membrane have vanished. The eye has 
collapsed, the margins of the iris have probably fused, and the pigmented and 
inner layers of the iris separated from each other. With this general sketch the 
elements of the eye may be taken up in detail. 

The pigment layer is variously developed (i in figs. 46 and 47) and may be quite 
different on the two sides of the head. One peculiarity is practically always present 
and very striking. The layer forms a covering over the distal face of the eye where, a 
priori, there ought to be no pigment, and is thinnest or absent over the proximal face 



where it ought to be most highly developed. Kohl has cut the Gordian knot by 
excluding this pigment from the eye entirely by the choroid (sclera), but there is 
certainly no such membrane intervening between this pigment and the rest of the 
eye as Kohl has figured. On the contrary the choroid very clearly surrounds it, 
and from its own epithelial structure there is no room for doubt as to its nature. 
As said, its extension over the sides and back part of the eye differs materially in 
different eyes. In a number of instances no pigment cells are present either on 
the sides or at the proximal surface; in others the sides are well covered. If by 
any means the tissues of the eye are separated from each other, the space is always 
formed between the pigmented layer and the rest of the eye. Processes are at such 
times seen to extend down from the pigment cells toward the rest of the retina. The 
cell boundaries and nuclei of the pigment cells are for the most part distinct. The 
cells are deepest over the distal pole of the eye and from this point they decrease 
in size to the proximal pole. Toward the upper face, where the pigment epithelium 
approaches the lens, the densely pigmented cells are transformed into much thinner 


chr. - 

nl. I— 

Fig. 47. Horizontal Section through only Eye with Central Ganglionic Cells. 
From an Individual 34 mm. long. 

pigmentless cells. These are probably the homologues of the pigmentless cells 
over the distal face of the eye of Amblyopsis, and, if so, are all that is left of the 
outer layer of the iris. 

The explanation of the condition of the pigment epithelium in this eye presents 
more difficulties than any other structure. In the eye of T. suhterraneiis no pig- 
ment is developed, but the pigment epithelium is normally developed. In this 
eye pigment is formed in the cells that are present, but the epithelium has any- 
thing but a normal structure. The pigment cells in the proximal face of the eye 
have either disappeared or been displaced. The only other alternative, that they 
are present but without pigment and indistinguishable from the cells of the outer 
nuclear layer, while possible, is scarcely probable, for in many eyes there is but a 
single layer of cells representing all of these structures, and in other cases even these 
have vanished. The objection to the idea that the cells have vanished is to be 
found in the fact that they are so well developed over the distal face. This point 
can only be settled by a study of the development of the eye, but one other sug- 
gestion may not be out of place. A comparison of this eye with that of Amblyopsis 



will suggest the homology of the anterior cell mass in the latter case, with the pig- 
ment cells always present between the retina and the irideal pigment layer in the 
former species. This correspondence is further strengthened by the fact that 
frequently the pigment in T. rosce over the front of the eye is in more than 
one layer of cells. Since, however, I was unable to arrive at an entirely satis- 
factory explanation of the origin of this pigment mass in Amblyopsis, it will not 
help us much, should the two structures be homologous. 

Attention may be called here to the fact that both in Amblyopsis and in the pres- 
ent species the lens — and therefore the lost pupil — are not situated at the distal 
pole of the eye, but above this point, and that both in regard to the pupil and the 
eye in general the location of the pigment masses in the two species is the same. 

The pigment is granular, not prismatic. 

scl. c. 

sd. e, 

Fio. 48. Cross-sections through Ri^ht and Left Eye of an Individual 35 mm. long. Sections be pass through Lens. 
Fig. a is a Composite from ^ Sections. Fig. b represents one Secliont but the " Lens " is from the Next Section. 

The lens is the only structure of the eye concerning which Kohl has not made 
any mistake.* It is a small group of cells closely crowded together and about 
ID or 12 /x in diameter (figs. 48 a' and b, I). There are no signs of fibrilation or the 
result of any other histogenic process; it appears as an aggregation of indifferent 
cells. On its surface there are at times cells that are evidently of an epithelial 
nature, being flattened so that their sections appear much longer than deep. It 
lies at the upper outer face of the eye at the margin of the pigment mass described 
in the last section. It is not covered by pigment or other retinal substance. 
Kohl considered this condition a primary one. The lens, however, does not lie in 
an incipient secondary optic cavity, the vitreal cavity, as Kohl supposed, but in 
the remains of such a structure. Under the circumstances it is doubtful whether 
the uncovered condition is primary. It seems more probable, considenng the 
condition in Amblyopsis, that the lens was inclosed by the closing of the pupil 
over the eye, and that the present naked condition is the result of the subsequent 
degeneration of the iris over it. That the latter is the phylogenetic origin of its 
present condition there is no doubt. 

' Considering the history of the lens in Amblyopsis, I am not sure now whether Kohl was or was not mistaken 
about these cells. 



Cl. C 

A to G, Photographs of the eyes of Amblyopsis ; H, eye of Troglichthys. 

A. Horizontal section of right eye of fish 9.5 mm. long. 

B. Dorsal face of horizontal section of left eye of fish 25 mm. long. Optic nerve 

directed forward and inward. 

C. Cross-section of left eye of fish 1 00 mm. long. 

D. Anterior face of transverse section of left eye of fish 123 mm. long. 

E. Transverse section of left eye of fish 1 30 mm. long. No definite structures are 

distinguishable aside from scleral cartilage. 

F. Transverse section of right eye of fish from v^hich E was taken. 

G. Cross-section of right eye of fish 1 05 mm. long, showing large vesicle formed by 

pigment epithelium and remainder of retina as small nodule on its distal face. 
H. Eye of Troglichthys rosae showing large scleral cartilages and different layers of 
the eye. 


The Retina : The elements of the retina proper, i.e. the ganglionic, nuclear, and 
reticular layers, form a vesicle arranged so that the cellular elements surround a 
central (the inner) reticular layer. These may be taken up seriatim. The cellular 
elements are of three sorts. 

(i) Behind the lens and behind the pigment layer, sometimes also over the side 
of the retina, lie a few cells with elongated nuclei {nl,l, in figs. 45 a, 6,46 b, 49 a) 
and so arranged as to suggest an epithelial covering for the underlying structures. 
Some of these cells were supposed by Kohl to represent the choroid, with which they 
have absolutely no connection. It is possible that some of these lateral cells are 
modified pigment cells, but even this seems doubtful. I am unable to refer the 
cells of this nature situated laterally over the retina to any structure in the normal 
retina. Such cells are, however, found in the eyes of T. subterraneus between 
the pigment epithelium and the nuclear layers (fig. 41 a), and whatever their origin 
the two structures are unquestionably homologous in the two eyes. It is probable 
that the cells with elongated nuclei to be found behind the lens are of different 
origin and significance. They may be the remains of the elongated cells found 
in the inner surface of the iris of Chologaster, cells which are still present in both 
Amblyopsis and T. subterraneus. It is also possible that they are the remains of 
the hyaloid nuclei. 

(2) The ganglionic cells, which in Typhlichthys are arranged around the vestige 
of the vitreal cavity and in Amblyopsis form a central core and are distributed over 
the front of the retina, are in this species practically confined to the latter location. 
All there is left of the central core of ganglionic cells in Troglichthys rosa is three 
cells in the most highly developed eye found (fig. 47 and plate 10, fig. n). In the 
other eyes no indication of these cells was detected. If these cells come to be formed 
at all in the present eye, they migrate forward, where they form the anterior wall of 
cells surrounding the inner reticular layer. The fibers of the ganglionic cells extend 
directly from the ganglionic cells through the reticular layer to the exit of the optic 
nerve. The cells must, as Kohl has suggested, have undergone a rotation on their 
axes to send their fibers directly to the optic nerve, unless only the lineal descendants 
of those ganglionic cells immediately surrounding the entrance of the optic nerve 
in Chologaster are here represented, a supposition not to be entertained. The 
ganglionic nuclei are occasionally notably larger than the nuclei of the rest of the 
retina, but they are by no means always so. 

(3) The cells of the nuclear layers join those of the ganglionic layer. The 
cells of the inner and outer nuclear layer and the horizontal cells are indistinguish- 
able from each other. They form, in the most highly developed condition, figure 47, 
3-7, a layer three cells deep covering the sides and the proximal surface of the 
inner reticular layer. In some cases the layer is reduced to a single series of cells, 
and even these are occasionally absent. There is no sharp distinction between the 
nuclei of this layer and those of the ganglionic layer, so that the boundary between 
these cells and the ganglionic cells is not marked. In some instances these cells 
appear to be directly continuous with the cells surrounding the origin in the optic 
nerve. This condition led Kohl to imagine that the primary optic stalk had 
become filled with nerve fibrils. 

Of the reticular layers the outer (8, in fig. 47) is not developed. The inner 
reticular layer forms, with the optic fibers traversing it, the spherical or pear-shaped 


central mass of the retina. No cells are developed in the reticular layer. The 
optic fibers appear to pass directly through the reticular layer. This condition is 
probably apparent rather than real. First the vitreous cavity disappeared, bring- 
ing the ganglionic cells and the optic fiber layers together in the center of the 
eye. This condition has just been reached by T. subterraneus and Amblyopsis. 
In the present species the ganglionic cells have disappeared from the center, and 
only the optic fiber layer remains. This is represented by the individual fibers 
passing from the ganglionic cells to the exit of the optic nerve. They do not form 
a compact nerve, but the fibers pass individually to the exit in the most direct route 
from their respective cells. 

I have been unable to trace the optic nerve for any distance beyond the eye. 
In one case it leaves the eye as a loose bundle 12 /a in diameter; in another case 
it is more compact, being but 4 /* in diameter. It is surrounded by a sheath of 
varying thickness and complexity. In one case there are a few cells about the nerve, 
and these are covered by the tendon of the rectus muscle, which forms a complete 

Measurements in ft: The scleral cartilages vary from 18 to 40 in thickness. 
The distance from the distal face of the retinal pigment to the ganglionic cells varies 
from 30 to 40. The pigment cells have a maximum depth of 14, dwindling from 
this to 2 or 3 on the sides. The nuclear layers reach a maximum thickness of but 
10. The inner reticular layer, including the optic fiber layers, is about 40 in all 
directions, reaching a proximo-distal length of 70. The lens measures from 10 to 15. 


The eyes of Amblyopsis have been described by Tellkampf, Wyman, and Put- 
nam. These authors gave general accounts of the eyes as far as this could be done 
without serial sections, and their accounts are far from satisfactory. It is therefore 
unfortunate that Kohl, who had less material of a supposed Typhlichthys from 
Missouri, should have based a criticism of the facts observed by Wyman in Amblyop- 
sis on what he saw, especially since scarcely a statement made by Kohl corresponds 
to a condition found in Amblyopsis, or even the Typhlichthys subterraneus from 
Mammoth Cave. An abstract of Kohl's result are given under Troglichthys. 

Tellkampf first pointed out the presence of rudimentary eyes and states that 
these can be seen in some specimens as black spots under the skin by means of 
a powerful lens. The statement that the eyes are externally visible in some speci- 
mens, which was afterwards thrown in doubt by Kohl, is perfectly correct. The 
eye of Amblyopsis can be seen as a black spot with the unaided eye in specimens up 
to 50 mm. in length. 

Wyman, in Putnam, figured the optic nerve, a lens, and muscular bands attached 
to the exterior of the globe, but did not recognize them as homologues of the muscles 
of the normal eyes of fishes. In a four-inch fish Wyman found the eye to be one- 
sixteenth of an inch in its long diameter. A nerve filament was traced to the cranial 
wall, but farther it could not be followed. The eye is made up of (i) a thin mem- 
brane, the sclera; (2) a layer of pigment cells, the choroid, which were most abundant 
about the anterior part of the eye ; (3) a single layer of colorless cells larger than 
the pigment cells, the retina; (4) just in front of the globe, a lenticular-shaped 
transparent body, the lens; (5) the whole surrounded by loose areolar tissue. 



Wyman was mistaken in his identification of Nos. 2 and 4, and part of 3. 

Of this species I have had an unlimited supply of fresh material from the 
Shawnee Caves in Lawrence County, Indiana. I shall first give the histology of 
the eyes of fishes from 25 mm. long to their maximum size, 135 mm. The details of 
the development of the eye will follow. 

In well-fed adult specimens of Amblyopsis there is no external indication of 
an eye. In poor individuals the large amount of fat surrounding the eye and 
collected in a ball-shaped mass becomes apparent through the translucent skin. 
In young specimens, before they have reached a length of 50 mm., the eyes are 
perfectly evident from the surface. By this I do not mean that they are conspicuous, 
for the minute eyes would not be conspicuous were they situated just beneath the 
skin. The skin is not modified in the region over the eyes, but has the same structure 
it possesses in the neighboring regions. This condition is in strong contrast to 
the conditions described for Chologaster papilliferus. The position of the eye can 
be determined from the surface in older individuals by certain tactile ridges, being 

'- s«6o. 

Fic. 49. (a) Section of Right Half of Head of Chologaster, through Eye. 
(p) Section of Right Half of Head of Amblyopsis, through Eye. 

between a long longitudinal ridge (supraorbital) situated caudad of the posterior 
nares and two vertical (suborbital) ridges. They can also be approximately located 
by the mucous canals, being situated above the middle of the suborbital canal 
forward from the fork of the suborbital and rostral canals. The exact location in 
relation to these ridges differs, however, to some extent in different specimens. 

The skull is surprisingly little modified, there being deep orbital notches, large 
enough to accommodate a large eye. The maintenance of this skull structure 
long after the eye has dwindled is significant in the consideration of the causes of 
degeneration and will be referred to again. 

The change in the relation of the eye to surrounding tissues as well as the relative 
size can best be gathered from the accompanying figures or cross-sections of Cholo- 
gaster and of Amblyopsis, drawn with the same magnification, but from different 
sized individuals (figs. 49 a, b). 

Beneath the dermis (black in the figures) a thick layer of connective tissue has 
developed in Amblyopsis. The large fibrous capsule occupied by the eye, eye 


muscles, and orbital fat in Chologaster has in Amhlyopsis become largely filled with 
fat. There is no indication of fatty degeneration; it is simply the accumulation 
of fatty cells in the eye cavity. The eye is very small and lies on the floor of the 
optic capsule. The infraorbital and supraorbital fat-masses described for Cholo- 
gaster papilli/erus are also large in Amhlyopsis and form especially large masses in 
front and behind the optic capsule. In Chologaster the brain extends forward beyond 
the front of the eye, while in Amhlyopsis the brain does not extend as far forward, the 
anterior portion of the brain cavity being filled with fat. Attention may also be 
called here to the presence and position of the suborbital bones which Kohl says 
are represented in Troglichthys by the cartilaginous masses forming a hood over 
the front of the eye. These cartilages {scl., fig. 49) are present in front of the 
Amhlyopsis eye, and it can readily be seen that they have nothing to do with the 
suborbital bones {sub. 0). 

The adult eye of Amhlyopsis with its appurtenances may now be taken up 
seriatim. The eye occupies the lower part of the eye cavity. It is surrounded by 
loose connective tissue, which is so associated with the eye that if contractions 
occur through reagents, as frequently happens, a space is left between the eye 
with its connective tissue and the septum forming the lower floor of the eye cavity. 
Above the eye with its connective tissue is the large accumulation of fat mentioned 
previously. From the eye to the inner wall of the orbit extends a continuation of 
the connective tissue surrounding the eye. In this continuation of the connective 
tissue the optic nerve and eye muscles extend. In the longest individual, 135 mm. 
long, the eyes were 5 mm. from the surface of the epidermis. 

The shape of the eye together with the pigment variously scattered in the con- 
nective tissue associated with it is very variable, differing from subspherical in the 
smaller individuals to long spindle-shaped in the old. Considerable difference 
is found in the shape of the eye itself. See table of measurements, page 144. 

Pigment is found in very variable quantity and variously scattered in the con- 
nective tissue surrounding the eye. The amount of this pigment seems to vary 
inversely with the amount of pigment in the eye itself and to increase with age. 

As Wyman has stated and figured, eye muscles are present in Amhlyopsis, 
but, contrary to his statement, they are the homologues of the normal eye muscles. 
Not all preparations are equally good for tracing the muscles. They are best 
demonstrated in heads treated entire by Golgi's method and sectioned in celloidin. 
While the muscles have been noted in a variety of preparations the description will 
be drawn from those treated by Golgi's silver method and stained at times with 
haemalum or Biondi-Ehrlich's 3-color stain.^ 

In one individual the upper rectus and upper oblique muscles are inserted 
together on the upper median surface of the eye, or more exactly on the upper 
posterior angle of the upper scleral cartilage. The lower oblique is inserted opposite 
this place. From these places the oblique muscles extend inward and forward. 
The origin of the lower oblique is 0.72 mm. in front of its insertion, while the larger 
upper oblique extends a little farther forward, being inserted 0.85 mm. behind its 
origin. It takes its origin in the projecting angle of a cartilage above and in advance 
of the origin of the lower oblique. In the inner part of the orbit a small muscle 

' Golgi's method did not give the desired results for nervcus structures, but by staining with the above 
methods the material was found excellent for general purposes. 


extends from the inferior oblique horizontally backward, taking its origin with the 
rectus muscles. This muscle in its posterior extent has the characteristics of the 
inner rectus. But whether or not its fibers reach the eye, I was unable to determine. 
If they do, they reach it with the fibers of the lower oblique. 

The rectus muscles arise from the lateral margin of the bone forming the brain 
case, just behind the anterior end of the brain, the upper rectus taking its origin 
behind the others. They extend as four bundles forward in a connective tissue 
tube. Before leaving this tube they are reduced to three bundles by the union 
of a small bundle situated above the others in the tube with the largest bundle 
situated nearest the outer margin. One of these is the lower rectus. The largest 
one is the upper rectus and the one joining it, in all probability, the external rectus. 
The external rectus, if I am correct in the identification, is not distinguishable from 
the latter during the rest of its course nor in its insertion in the sclera. The entrance 
of the rectus muscles into the connective tissue sheath occurs 0.5 mm. behind their 
insertion in the eye. In this eye we have the two oblique muscles, the upper rectus, 
the lower rectus, a small bundle of fibers following for the most part the course 
of the upper rectus, the external rectus, and a small bundle of fibers extending from 
the origin of the rectus muscles forward to the lower oblique which may be the 
inner rectus. We have at least five, probably all six, of the muscles normal to fish 
eyes. But that this is not always the case is very strikingly emphasized by the fact 
that the eye of the opposite side of the same individual lacks the upper oblique. 

In another individual the superior rectus and superior oblique are the only 
muscles present on the left, while on the right the upper rectus is the only muscle 
present. The preparations of this individual are particularly favorable for tracing 
the muscles. They are stained with Mayer's hasmalum and indigo carmine. The 
muscles are stained an intense blue, while the connective tissue through which they 
pass is light purple. 

In still another specimen both the oblique muscles are present on the left and 
three of the rectus muscles, one of which, the interior, extends forward in the inner 
part of the orbit and joins the lower oblique as in the first individual described. 
No fibers of this muscle reach the eye. On the right side of the same individual 
the upper rectus and but one oblique muscle are present. In still other individuals 
not suitable for tracing the muscles, their fluctuating number has been noted, and 
their varying method of attachment to the eye is also a matter frequently noticed. 

Inside of the loose connective tissue surrounding the eyes there is a more compact 
sheath. This is thickest in front of the eye where it contains spherical nuclei and 
holds one to three compact cartilages which usually are disposed to form a hood 
over the front of the eye. These cartilages, described by Wyman in this species 
and by Kohl in rosa, and taken by the latter as the remnants of suborbital bones, 
have nothing whatever to do with the latter structures. Their nature can be under- 
stood from their close association with the eye, by the fact that they are closely 
bound together by the scleral connective tissue, and by the fact that some, at least, 
of the eye muscles are attached to their outer surfaces. They are unquestionably 
scleral elements {scl.c. in figs. 49 to 52). There may be some hesitation in accepting 
this view of the nature of these cartilages since no cartilage whatever is found in the 
sclera of Chologaster. Their position, usually in front of the eye, is also anomalous 
if they are scleral cartilages. It may be stated, however, that the eye of Amblyopsis 



is not simply a miniature normal eye. The whole eye has collapsed with the dis- 
appearance of the vitreous humor, and looked at in this light there is no difficulty 
in the position of the cartilages which have fallen together over the front of the eye. 
The presence of granular nuclei in front of the eye over the region of the iris has 
been noted by Kohl in Troglichthys and observed by me. These nuclei are probably 
the homologues of the nuclei found in the ligamentum pectinatum of Chologaster. 
In shape, number, and size the scleral cartilages differ very much. In one 
instance cartilages extend continuously from the exit of the optic nerve more than 
half-way over the side and around the front of the eye. In another a single cartilage 
lies directly in front of the eye, and on the opposite side of the same individual 
a single cartilage lies behind the eye. The sclera is much more developed than in 
Cfwlogaster, consisting, aside from the cartilages, of an abundant fibrous tissue. 

Fig. 50. Section through the Eye of Amblyopsis spelaus t $ mm. long, killed with Chromic Acid and stained 
with fiiondi-Ehrlich's three-color mixture. This is the most highly developed eye seen, 3 mm. and 4. 

The choroid is a thin membrane closely applied to the eye. It contains a few 
oval nuclei parallel with the surface of the eye. Pigment cells are few, irregularly 
scattered, and not at all uniform in different eyes. The pigment cells are rounded 
masses usually much thicker than the whole of the choroid in regions devoid of 
pigment. About the entrance of the optic nerve is frequently a large accumulation 
of pigment corresponding with the increase in the amount of choroidal pigment 
in Chologaster at the same place. Even this mass is not uniformly present. Some- 
times granular masses interspersed with pigment are found here, which give one 
the impression of a degenerating mass. An especially large accumulation of pig- 
ment is found in the eye represented by figure 53. Blood-vessels are present in the 
choroid. They are apparently as great in relative capacity as in Chologaster. In 
an individual with the vascular system injected, a vessel, 0.0 1 mm. in diameter, 
approaches the eye with the optic nerve, but it does not enter the ball with the latter. 



It breaks up into smaller vessels distributed in the choroid. A vessel is usually found 
in a groove of the pigment layer of the retina. This groove extends along the dorsal 
wall of the eye — otherwise it might be taken for the choroid fissure (fig. 50, cps.). 
A somewhat larger vessel than at other points is found near the iris, where this 
structure appears to be continuous with a deep choroidal groove (fig. 51 a, cps.). In 
the young a blood-vessel enters the hyaloid cavity at this point. 

- scl. c. 

The eye itself, exclusive of choroid 
and sclera, differs greatly both in size and 
inner structure although the general ar- 
rangement of the retinal cells remains the 
same in all cases. In some cases the pig- 
ment layer of the retina forms a large mem- 
branous bag many times too large for the 
inclosed structures which lie as a small 
ball in this comparatively vast cavity. In 
such eyes found in old individuals the 
wall in many places is free from pigment. 
In general the pigmentation of the retina 
varies inversely as the pigmentation of 
the choroid. In other individuals the eye 
forms a compact mass of cells (fig. 53). 
To anticipate somewhat, the vitreous 
cavity with the hyaloid membrane and 
its blood-vessels have entirely disappeared, 
the ganglionic cells have in large part been brought together into a solid mass, 
and the irideal opening has usually become closed. 

Pigmented Layer and Cones. — The pigment cells as they appear in the best 
preparations of the better-developed eyes may be described first (fig. 50). The 
cells are longest near the entrance of the optic nerve. They possess an outer seg- 
ment, not determinable in all cells, which is free from pigment. They have a homo- 

FlG. 51. 

From AmbiyoPsis ps mm. long killed in Picric 
* :a with a Mixture of Hicmalum and 

Acid and stainei 

Indigo Carmine. Figures made with Bausch and Lomb 

iS Immersion and 4 Eyepiece. 
(a) Section of Right Eye. Choroidal Groove with one of 

Scleral Cartilages in front of Eye. Nuclear Layers 

thinner than usual. Densely Pigmented Segments of 

Pigment Cells form a Conspicuous Layer just below 

Pigment Nuclei. 
(6) Next Section after 51 a, showing Group of Elongate 

Uveal Cells. 



geneous, vesicular, ellipsoidal nucleus situated near the outer end of the cell. This 
nucleus is strikingly different in shape and constitution from the same structure in 
Chologaster. It stains but faintly and then homogeneously. Just within the 
nucleus there is a well-defined mass of dense pigment forming a cap over the inner 
side of the nucleus and at times encroaching on the rotundity of its inner outline. 
This pigment mass evidently has its counterpart in Chologaster where a solid band 
of pigment is found just within the nucleus. In depigmented cells this pigment cap 
is seen as a deeper-staining, more dense protoplasm than the rest of the cell. From 
this pigment segment a prolongation, much poorer in pigment and containing a 
central uniformly staining core, extends toward the interior of the eye. This core, 
which in reality extends also into the pigmented section, occupies the position of 
the cones in Chologaster. In no case have I been able to trace any connection 
between these bodies and the outer nuclear layer. They are sometimes in several 
esgments or in a number of spherical bodies, and occasionally two are seen side by 
side in the same cell in tangential section. In position they certainly suggest cones, 

Fig. 52. Section near Posterior Face of Left Eye of Small Individual, showinR particularly Position 
of one of Scleral Cartilages behind Eye and Thick Choroid filled with more or less Angular Mass 
of Granular Pigment. This Eye shows one of the Largest Accumulations of Pigment noticed. 

and this suggestion is heightened by the presence in the inner end of some of the 
cells of a vesicular structure very similar to the nucleus, but frequently with an 
angular indentation on the surface. These occupy the relative position of the cone 
bodies, they are by no means found in all eyes. The evidence seems to point most 
strongly in favor of the supposition that they are cones. One of the cells measures 
as follows : diameter of cell, 0.007 rnni- '> nucleus, 0.007 by 0.007 mm. ; deeply pig- 
mented mass, 0.007 mm. ; total length of cell, 0.036 mm. No rods have been found. 

In many individuals, and strikingly so in two specimens 25 mm. (fig. 53, am.) and 
35 mm. long (fig. 54 h) respectively, deeply staining spherical bodies, much smaller 
than the nucleus and staining much deeper, are present in the pigment cells. Those 
stained with haemalum are quite dark and give the appearance of a large centro- 
some. These I take to be myeloid bodies noted in the pigment cells of the frog and 
other forms. 

In most individuals the high development of the pigmented region, above 
described, is not found. In some individuals the pigmented layer is composed of 
flat pavement cells, forming a large vesicle (plate 10, figs, d and g). In others the 



pigment is either entirely absent or very sparingly developed. As mentioned 
above, the pigmentation of the eye seems to vary inversely with the pigmentation 
of the surrounding structures. 

The pigment is in all cases granular and differs in this respect from the pris- 
matic pigment of the eye of Chologaster. 

Iris. — The pigment cells decrease in height toward the irideal part of the eye, 
where they are replaced by a layer of pigmentless cells forming a thin membrane 
(fig. 50). The nuclei of these cells stain darker than the bodies of the cells, 
which is the reverse of the conditions seen in the pigmented cells. In individuals 
up to 35 mm. long similar cells ex- 
tend along the line of the vanishing 
choroid fissure (figs. 54, a and /). 

The pigmentless membrane is ap- 
parently the relic of the outer pig- 
mented layer of the iris. If so it has 
undergone greater changes than the 
rest of the pigmented layer, for it is 
well pigmented in all the species of 

The inner layer of the iris is fre- 
quently entirely separated from the 
outer layer and not infrequently is 
entirely obliterated. (A few rounded 
pigment masses are always found 
within the eye at this point.) In 
other individuals a minute opening 
is still present and the outer layer of 
the iris is continuous with the inner, 
which contains some of the elongate 
nucleated cells found in the region of 
the ora serrata in Chologaster. These 
are much more regularly present in 
Typhlichthys subterraneus. These 
nuclei are variously grouped in different eyes, as is represented by the figures 50, 
51 6, 54 b, d, e. The exact significance of the various structures about this region 
in the eye can not always be determined owing to their presence or absence 
in dififerent individuals and their great variability when they are present. In this 
region are sometimes a few cells with elongate nuclei that can not be identified with 
any of the structures considered. These may represent all that is left of the hya- 
loid. Blood-vessels are usually not found in the eye of the adult. 

Between this pigmentless membrane and the rest of the retinal structures, i.e. 
within the pigment epithelium, there is in the majority of the adult eyes an irregular 
mass of pigmented cells. I am entirely at a loss to account for this mass unless 
with the shrinking of the eye as the result of the loss of the vitreous body and lens 
and the consequent closing of the pupil, the margin of the iris is rolled inward 
and some of the pigmented cells of the outer layer of the iris come to lie within the 
eye after the closing of the pupil. The iris is seen to be rolled in the way imagined 

Fig. 55. Horizontal Section through RiKht Eye of Specimen, 25 mm. 
long from above. A Large Branch of Optic Nerve is seen to pass 
in front of Cone of Ganglionic Ceils. This is not Constant, and 
in Left Eye of Same Individual the Largest Strand passes 
behind Number of Ganglionic Cells lying in front of Inner 
Reticular Layer and the Central Ganglionic Mass. 



in many sections of Chologaster and the method of the closing of the pupil in Typh- 
lomolge is as I have suggested. 

The Nuclear Layers. — Within the pigment and cone layer lies a nuclear layer 
made up of about four series of cells (3 to 7 in figs. 50, 54 e). The nuclei reach 
from 2.5 to 3.5 fjL in diameter. Rarely I have succeeded in staining the smaller nuclei 
different from the larger. They are, in such cases, more refringent, the large nuclei 
being granular. The larger nuclei may be the spongioblasts. In a young indi- 
vidual this difference was well marked. Here the smaller cells were confined to 
the proximal part of the eye (fig. 53). A separation of the nuclear layer into an 


Fio. 54. From an Individual 35 mm. long tilled in Perenyi's Fluid and stained with Mayer's Hsemalum. 

(o) Outer Nuclear Layer in Center, Choroidal infolding on Left. Lower Part of figure passes through Choroidal Fissure Area 

and Pigment Cells are here undifferentiated, quite different from those of the Dorsal Part of Same Section. 
(6) Further Forward and shows Strands of Optic Ncrie (». of.) and Elongated Nuclei of Inner Layer of Iris irregularly arranged 

(W. ;.). Choroid and Sclera can not be separated from each other e^ept where Latter is differential as Cartilage, in front of Eye. 
(c,iO Surface and Deeper Focus of Section passing through Iris and Central Ganglionic Cells. In fig. d Irideal Structure with 

Elongated Inner Nuclei is well shown, 
(«) Passes near Center of Eye. Choroidal Fissure Epithelium seen below and Irregular Mass of Section through Elongated Irideal 

cells (t.;. I.). 
(/) Passes through Optic Nerve and Pupil of Same Eye as fig. e. 

Figs, a to d arc from Left Eye, e and / from Right Eye. All under Lenses • mm, and 4. 

inner and outer with an intervening outer reticular layer I have noticed but once. 
In this eye a slight separating space was found on one side, and here there were 
one or two cells that may be fulcrum cells. If so, it is the only indication of this 
layer in all the preparations made. The suppression, partial or total, of the sepa- 
ration into an outer and inner layer, has also been noted by Ritter in Typhlogohius. 
The Inner Reticular Layer. — This layer is always well developed ; occasionally 
a few nuclei extend partially in from the outer nuclear layer. It is frequently thicker 
on the dorsal half of the eye (fig. 54/) than on the ventral half, but sometimes the 
reverse. In figure 50 the ventral half is but 0.012 mm. Nuclei have but once been 


found in this layer, and I have not been able to identify Mullerian nuclei as such 
either in this or the nuclear layers. The ganglionic layer forms a compact mass 
of nuclei, somewhat funnel-shaped, with the narrow end toward the exit of the nerve 
(9 in figs. 50-54 e). I have found from 60 to 125 nuclei in this mass. At the 
wide end of the funnel this mass of cells is directly continuous with the cells of the 
nuclear layers. The cells in this intermediate layer are of the large type, and as 
they give off fibers to the optic nerve, they may be classed as ganglionic or possibly 
as cells belonging to the spongioblasts. 

Optic Nerve and Lens. — The optic nerve is always evident in the eye itself 
except in very old individuals. It passes as a compact thread through the pigmented 
layer into the ganglionic layer. Here it breaks up into smaller bundles, the fibers 
of which pass in part to the cells within the ganglionic core, while the greater part 
pass to the large cells at the outer rim where the ganglionic cells pass over into the 
cells of the granular layers. The fact that these large cells give off the greater part 
of the optic fibers suggests whether or not these cells are really the ganglionic cells, 
while the cells forming the core are such cells as are seen at the entrance of the optic 
nerve in Chologaster (z in fig. 35 c) and there form a plug around which the optic 
fibers pass directly to the ganglionic cells. The bundles of fibers passing to the 
anterior cells never pass through the mass of core cells but at one side of this mass. 
In the right eye of an individual 25 mm. long they pass out in front of the mass; 
in the left eye of the same individual, behind them. 

Outside the eye itself the matter of following the optic nerve becomes a much 
varying task. In very young, and up to 25 mm., there is no difficulty in tracing 
the optic nerve to the brain. In newly freed individuals (about two months old) 
the optic nerve passes nearly obliquely down and in, while in an individual 25 
mm. long it passes horizontally back and in toward the foramen for the optic 
nerve. In the latter individual the nerve leaves the eye, not as might be expected 
at the posterior inner face, but at the anterior inner, making a sharp turn as it 
leaves the eye. Its compact nature is entirely lost after leaving the eye, forming 
a loose bundle several times as thick as the optic nerve within the eye. It is here 
surrounded by a very thin film of pigment, which in its turn is surrounded by 
layers of fibrous tissue. 

In individuals much more than 25 mm. long it is usually no longer possible to 
follow the nerve to the brain. It can be followed some distance, but usually dis- 
appears before reaching the optic foramen. In but one instance did I succeed in 
following it into the brain cavity in an adult specimen. The structures surround- 
ing the optic nerve are as variable as those surrounding the eye. In one case it is 
surrounded by various layers of pigment, while in others scarcely any pigment is 
found with it. 

The most highly differentiated lens * was found in an individual 130 mm. long, 
i. e., a very old one. The lens in this case consists of a few nuclei about which 
there are concentric layers of a homogeneous tissue (fig. 54/). In other individuals 
structures approaching this condition were found (fig. 55 a), in one a large cell, 
in another a cell with concentrically arranged lamellae. The lens, in an individ- 
ual 25 mm. long, could not be found at all, and in another 35 mm. long could 

' It is certain that this is not the lens. The name " secondary lens " may be applied to it. Similar structures 
are found occasionally in Rhineura and Lucifuga. 



not be determined with certainty. The relative development of the lens is not 
dependent on age. The lens described by Wyman w^as undoubtedly one of the 
scleral cartilages, for these cartilages are frequently nodular in this species and 
one usually lies in front of the eye. 

The supposition of Wyman that one of the scleral cartilages is the lens need not 
be criticised too severely. The structures described above as the lens are con- 
sidered such, more because they could not be identified as anything else, and be- 
cause nothing else that could with certainty be considered a lens could be found 

Fig. 55. (a) Two Successive Sections through Right Eye of very Old Individual 130 mm. long, showing a Lenslilce Body. 
(bj Outline Section of Left Eye of Individual 108 mm. long, showing highly developed Lenslike Body. 

aside from these structures, rather than on any direct evidence. The development 
of the eye would indeed lead one to suppose that the lens is actually placed entirely 
outside the optic cup, and in that case none of the structures here described can be 
the lens. With as much variation as is found in all the structures it is not improb- 
able that the lens may, in some individuals, be found within the optic cup, and in 
others outside of it. 

The progressive ontogenetic degeneration of the eye after maturity will be given 
in the section dealing with its ontogenetic history. 

Measurements of the Eye of Amhlyopsis in fi 

Ixtnrth of 

Diameter of 
Eye, Axial. 

Diameter of 
Eye, Vertical. 

Pigment Layer 

















































1. There are at least 8 species of "blind fishes," Amblyopsida;, inhabiting 
North America; 3 with well-developed eyes and 5 with mere vestiges. 

2. The 5 species with vestigial eyes are descended from 3 generically distinct 
ancestors with well-developed eyes. 

3. The genera can be more readily distinguished by the structure of their eyes 
than by any other characteristic. 

4. The most highly developed eye is much smaller and simpler than the eye of 
normal -eyed fishes. 

5. The structure of their eyes may be represented by the following key to the 
genera and species of Chologaster : 

a. Vitreous body and lens normal, the eye functional. No scleral cartilages. 
Eye permanently connected with the brain by the optic nerve. Eye mus- 
cles normal. No optic-fiber layer. Minimum diameter of the eye 700 fj.. Chologaster 
b. Eye in adult more than i mm. in longitudinal diameter. Lens over 0.5 
mm. in diameter. Retina very simple, its maximum thickness 83.5 
/x in the old; the outer and inner nuclear layers consisting of a 
single series of cells each; the ganglionic layer of isolated cells. 
Maximum thickness of the outer nuclear layer 5 /i; of the inner 
layer 8 /i. . . . . . . . . . . cornutus 

bb. Eye in adult less than i mm. in longitudinal diameter. Lens less than 

0.4 mm. Outer nuclear layer composed of at least 3 layers of 

cells ; the inner nuclear layer of at least 3 layers of cells, the former 

at least 10 fj. thick, the latter at least 18 /i. 

c. Pigment epithelium 65 fx thick in the middle-aged, 102 in the old . papUliferus 

cc. Pigment 49 (i thick in the middle-aged, 74 in the old ; 24-30 per 

cent thinner than in papUliferus. Eye smaller . . . agassizii 

aa. The eye a vestige, not functional ; vitreous body and lens mere vestiges ; the 
eye collapsed, the inner faces of the retina in contact; maximum diame-* 
ter of eye about 200 /u.. 

d. No scleral cartilages; no pigment in the pigment epithelium; 
a minute vitreal cavity ; hyaloid membrane with blood- 
vessels. Pupil not closed. Outer nuclear, outer reticu- 
lar, inner nuclear, inner reticular, ganglionic, and pig- 
ment epithelial layers differentiated. Cones probably none. 
No eye muscles. Maximum diameter of eye 180 fi. Eye 
probably connected with brain throughout life . . Typhlichthys 
dd. Scleral cartilages; pigment in the pigment epithelium ; vitreal 
cavity obliterated; no hyaloid membrane. Pupil closed. 
Some of the eye muscles developed. No outer reticular 
layer. Outer and inner nuclear layers merged into one. 
Eye in adult not connected with the brain. 
e. Pigment epithelium well developed; cones well developed; 
ganglionic cells forming a funnel-shaped mass through 
the center of the eye. Pigment epitheUum over the 
front of the eye without pigment. Maximum diameter 
of eye about 200 ix ..... . Amhlyopsis 

ee. Pigment epithelium developed on distal face of the eye, 
rarely over the sides and back. No cones. Nuclear 
layers mere vestiges ; the ganglionic layer restricted to 
the anterior face of the eye just within the pigmented 
epithelium. Maximum diameter of eye about 85 /^ . Troglichthys 


6. The steps in degeneration are seen in figure 66, page 176. 

7. The structure of the vestigial eyes differs much in different individuals. 

8. The eye of Chologaster is an eye symmetrically reduced from a larger, normal 
fish eye. 

9. The retina in Chologaster is the first structure that was simplified. 

10. Later the lens, and especially the vitreous body, degenerated more rapidly 
than the retina. 

11. The eye of Typhlichthys has degenerated along a different line from that 
of Amblyopsis, its pigmented epithelium having been most profoundly affected. 

12. The eye muscles have disappeared in Typhlichthys. 

13. Troglichthys shows that the steps in the degeneration of the muscles were in 
the direction of lengthening their attaching tendons, finally replacing the muscles 
with strands of connective fibers. 

14. The scleral cartilages have not kept pace in their degeneration with the 
active structures of the eye. 

15. The lens in the blind species, if present, is, for the most part, a small group 
of cells without fibers; in Amblyopsis it disappears early. 

16. The proportional degeneration of the layers of the retina is shown in figure 
67, page 179. 

17. With advancing age the eye of Amblyopsis undergoes a distinct ontogenetic 
degeneration from the mature structure. 

18. The phyletic degeneration does not follow the reverse order of development. 
None of the adult degenerate eyes resemble stages of past (phyletic) adult condi- 

19. The degenerate eyes do not owe their structure to a cessation of develop- 
ment at any past ontogenetic stage, i.e., at any stage passed through in developing 
a normal eye. 

20. Cessation in development occurs in the reduction of the number of cell 
generations produced to form the eye and in histogenesis, not in cessation of mor- 
phogenic processes. 

21. In some cases (Typhlichthys) there is a retardation in the rate of develop- 
ment, the permanent condition being reached later in life than is usual in fishes. 
(It is possible that the pigment of the pigment epithelium never comes to develop 
at all. It is, however, impossible to assert this until the embryos of this species are 
examined. It is possible that the pigment degenerates before the stages that I have 
examined are reached.) 




The present chapter describes the developmental stages of the eye of the blind 
fish Amblyopsis spelaus and gives the history of the eye during growth, maturity, 
and old age. Questions of special interest in the history of this very degenerate 
organ are: 

1. Do the rudiments of the eye appear as early as usual or later? 

2. How much does the eye grow from the time of its appearance? 

3. When does each part of the eye reach its maximum (a) in size, (h) in mor- 

phogenic development, (c) in histogenic development ? 

4. When does the eye as a whole reach its maximum development ? 

5. Are there evidences of a slowing down of the rate of the developmental 

processes : (c) cell division, {b) cell arrangement, (c) cell differentiation ? 

6. Are there evidences of a cutting ofT of late developmental stages, that is, 

are there any parts of the normal eye that are not developed ? 

7. Does the eye develop directly toward the condition of the adult or does it 

follow palingenetic paths and then retrograde to the condition found 
in the adult? 

8. What parts of the eye degen- 

erate first ? 

9. What is the comparative rate 

of the ontogenetic degenera- 
tive modifications of the vari- 
ous parts of the eye, and how 
does their rate compare with 
the rate of phylogenetic de- 
generation implied by the 
structure of the adult eye 
of Amblyopsis and the dif- 
ferent stages of degeneration 
reached by other members 
of the family? 
10. Is there any evidence for or 
against the dictum of Sedg- 
wick that structures which 
have disappeared from the 
adult organization are re- 
tained in the embryo only if 
the organ was of use to the larva after it had ceased to be of use to 
the adult? 


The development of the eye has been followed in several series of living embryos 
and in sections of these embryos. The earlier stages of the eye as they were 
observed in the series obtained on May 4, 1901, will be described.' Where advisable 
other series will be described also. The first indications of the eye are seen in living 
specimens when the embryo is about 1.5 mm. long, at about the time of the forma- 
tion of the first protovertebra. This size was reached in the present series in 2.5 to 

' For an account of the general development of this series see p. 95. 

Fig. s6 

(a)^ Outline of Head of Embryo between 1.3 and 1.5 mm. long. 
(6) Outline of Brain and Optic Thickening in Mounted Embryo r.6 
mm. long, with 4 Protovertebrae (3.30 p.m.. May 5). 

(c) Outline of Brain and Optic Thickening in Living Embryo 1.92 
mm. long with 10 Protovertebrie (12 p.m.. May 5). 

(d) Outline of Brain and Optic Vesicle of Living Embryo 2.4 mm. 
long with 10 Protovertebrs (12 p.m., May 5). 



3 days from fertilization. The degree of development when the eye begins to form 
is exactly as in fishes with normal eyes. 

At 1 1 a. m., May 5, 1901, the head was slightly raised so that its outlines appeared 
definite and clear, while the remaining outlines of the embryo were hazy. It was 
not possible at that time to distinguish eyes (fig. 56 a). At 2^ 30" p. m., when the 
embryo has reached a length of 1.6 mm., the eyes form prominent lobes on either 
side of the brain. The lobes are distinguishable in living embryos, but stand out 
much more prominently in embryos mounted entire. In an embryo prepared in 
this way, a camera outline of which is reproduced (fig. 56 b), the eye protuberance 
(oc.) has a length of 80 /u, and projects 36 /u. beyond the lateral margin of the brain. 
Sections of embryos at this stage of development show the brain to be still joined 
with the ectoderm. There is no indication of any cavity in the central nervous 
system at this time and the eye lobes are solid, symmetrical, lateral protuberances 
with their anterior margins but 48 fi from the tip of the brain. At 6 p. m. 
the embryo had reached a length of 1.76 mm. and 6 protovertebrae had been 
formed. The eye was no longer a symmetrical swelling on the side of the 
brain, but its outer, posterior angle was now distinctly farther back than the pos- 
terior inner angle. In other words, the lobes had grown laterad and were bent 
backward. The lateral projection of the eye beyond the contour of the brain 
amounts to 48 fi and has a longitudinal extent of 100 fi (fig. 56 c). The greatest 
diameter — measured from the anterior inner angle of the eye to the posterior 
outer — was 116 /a. Sections show the nervous system, including the eye, to be 
still a solid mass of cells, which anteriorly is still continuous with the ectoderm. 

Histologically there is no differ- 
ence between the cells com- 
posing the optic lobes and 
those composing the brain. 
There is a slight indication in 
the arrangement between the 
two optic lobes suggesting a 
lateral traction of the cells. At 
9 p. m. the characters of the 
eye shown at 6 p. m. had be- 
come intensified without other 
material change. The embryo 
had reached a length of 1.92 
mm. and 10 protovertebrae had 
been formed. The optic lobe 
was still broadly united with 
the brain, but its lateral growth 
was largely represented in the 
lobe extending back. There 
was no cavity as yet in the 
nervous system. A little later 
the canal of the central nervous 
system made its appearance, for at 1 2 p. m. it was well formed. There was probably 
some fluctuation as to the rate of growth in length and the degree of differentiation 

Fig. 57. 
(a) Outline of Brain and Optic Vesicle of Livinj; Embryo between sizes 

of those shown in tigs. 56 d and 57 b (5.50 a.m., May 6). 
(6) Outline of Brain and Optic Vesicle of Living Embryo 2.4 mm. long, 

with 12 or 1.^ Protovertebra; (S a.m.. May 6). 
(c) Horizontal Section through Left Eye of Embryo about 3.44 mm. 

long, 3 Sections Ventrad of one represented in fig. 56 d. 
(,d) Horizontal Section through Head of Same Individual, showing Optic 

Vesicle (11 a.m., May 6). 
(«) Outline of Brain and Optic Vesicle of Embryo 1.68 mm. long, with 5 

Protovertcbra: from Livmg Specimen. 



the tissues reach, for, in embryos of another series, some individuals had a well- 
developed canal, while others of the same size did not. At 12 p. m. the embryos 
had reached a length of 2.4 mm. (fig. 56 d). At s*" 30"' a. m.. May 6, the eyes had 
become a pair of fiaps lying along the sides of the brain or diverging from near 
its anterior end and connected only in front by the contracted optic stalk (fig. 
57a). The split in the optic lobe which separates it into an outer and an inner 
layer had developed to such an extent that it could readily be made out in living 
embryos. At 8 a. m. some of the embryos were still only 2.4 mm. long and 12 to 
13 protovertebra; had been formed (fig. 57 b). The changes in the eye from 12 
p. m., May 5, to 12 noon. May 6, were not very great, and consisted chiefly in the 
constriction of the optic stalk and the consequent gradual separation of the optic 
lobe from the brain. The skin had not yet begun to thicken to form the lens (figs. 

The changes from noon till 6 p. m., May 6, when the last embryo of this 
series was preserved, consisted largely in the shifting of the optic vesicles as the 
result of the development of the olfactory pits. Seventeen protovertcbr?e had de- 
veloped and the embryo was about 3 mm. long. 

Fig. s8. Horizontal Sections through Optic Stalk (fig. a) and Optic Vesicle 
(fig. b) of Embryo of Second Series. 

For later stages I am compelled to draw on another series of embryos which I 
also observed through the earlier stages described above. They were taken from 
a female that was captured March 11, 1898, and that contained eggs in the early 
stages of gastrulation. 

The eyes had reached a stage seen at about 2.5 to 3 days from the beginning of 
development. An outline of the development may be given to connect this series 
with that just described. The rate of development was considerably slower than 
in the preceding series. Figure 57 e (March 13, 10 a. m.) was taken from a living 
specimen, showing 5 protovertebra;. Sections demonstrated that at the stage repre- 
sented by figure 576 the neural tube was still a solid structure. The distance from 
edge of eye to edge of eye measured 164 /*. 

About a day later the larva; were 2 mm. long. The neural canal had been 
formed and extended out into the now well-formed vesicle through a distinct optic 
stalk. Sections showed that the epidermis was still unmodified over the eye, with 
no indication of a thickening to form the lens. 



Figures 58 a and b show horizontal sections through the base of the optic stalk 
and through the middle of the optic vesicle respectively. The embryo is 2 mm. long 
and in about the same stage of development as those 2.8 mm. long of first series. 

During the next 24 hours the embryo grew to a length of 2.4 mm. At this 
stage the tail was free for 0.4 mm. of its length. Embryos 24 hours older than the 
last were found to be 2.5 to 2.8 mm. in length. The latter, while not longer than 
the oldest embryos of the first series described, are evidently farther along in the 
development of the eyes. In all of these specimens (figs. 59 a, c) the eyes have 
become greatly modified. The secondary optic vesicle has been formed by the thick- 
ening of the skin to form the lens. The retinal wall of the vesicle is three series of 
cells deep, while the wall destined to form the pigment epithelium has become 

("IG. 59. (a) Horizontal Section of Head of Embryo 3.5 mm. long, two Sides at Different Levels. 

(6) Left Eye of Same Embryo as that from which fig. 59 a was taken, showing First Indication of Leu 

(c) Transverse Section through Dorsal Part of Optic Stalk of Embryo 2.7 mm. long. 

id) Optic Vesicle and beginning of Lens in another Specimen 2.7 mm. long. 

(e) Transverse Section of Optic Vesicle and beginning of Eye of a Cymalogaster larva, 1.5 mm. long. 

thin and is composed of a single series of cells. The eye, at this stage, does not 
differ materially from that of a Cymalogaster ' larva about half as long. (Com- 
pare figs. 59 c, d.) 

There is no indication of a differentiation of an iris. The secondary cup is a 
shallow, bowl-shaped structure, the depression being entirely filled by the thicken- 
ing of the skin which is giving rise to the lens (figs. 59 b and d). 


In specimens 4.4 mm. long the eye had become a deeper cup than it was during 
the 3-mm. stage. The lens, which no longer fills the entire cavity, has become 

' Cymalogaster is a telcost with large and well-developed eyes, 
compared with figures 60 d, e (Amblyopsis). 

Figures 60 a, b {Cymalogaster) should be 



a spherical mass of cells, solid in some cases (fig. 60 d) but with a cavity in 
others. It is still connected with the skin. In one case the lens was a vesicle with 
a distinct epithelium bounding the cavity (fig. 60 e). In the other cases there 
seemed to be no regularity in the arrangement of the lens cells. 

The pigmented layer has become very thin compared with the thickness of the 
rest of the retina. Its thickness increases toward the margin of the cup. The 
retina is very thick, with about 5 layers of nuclei ; these are crowded except at the 
free margin of the retina, which is free from nuclei. There is no histological differ- 
ence between the different cells of the retina unless there is an appreciable elonga- 
tion in the cells at the margin of the cup. 

Optic fibers are not yet developed. 




Fio. 60, 

(a) Transverse Section of Eye of Cymalogasler larva, 3.2 mm. long. 

, . Transverse Section of Eye of Cymalogasler larva, 4.5 mm. long. 

(c) Transverse Section of Eye of AmblyopHs embryo, 4.4 mm. long. 

(rf) Section of Right Eve of Larva, 4.4 mm. long. Nuclei all drawn without a change of focus. 

(c) Vertical Section of Eye of another Larva, 4.4 mm. long. 


The embryo is hatched at the beginning of this period. The least differentiated 
eye of this stage is represented in vertical section in figures 6i a and b. The second- 
ary vesicle has become more definitely formed. The vitreous cavity is reduced in 
size and the retina has become distinctly thicker, but shows as yet no differentiation 
into different layers. 



In a larva 5 mm. long the eye is still in contact with the epidermis on one side 
and the incipient dura mater on the other. The epidermis is distinctly thinner 
over the eye, reaching an extreme thinness of 16 /x as compared with a thickness 
of 40 /x at a distance of 100 fi below the eye and of 0.24 fi at 100 fi above the eye. 

The lens lies directly beneath the skin. In this particular eye (fig. 62 a) it is an 
ellipsoid, 30 /* by 38 fi (36 by 28 in another eye). It is entirely separated from the 
skin and takes on a deeper stain. The cells of the lens are not very regularly grouped, 
but apparently they are arranged about a median point or space. The lens lies 
entirely outside of the eye in contact with the outer face of the dorsal part of the 
iris. The eye proper is a subspherical solid mass with only a shallow depression 
below the lens representing the vitreous cavity and choroid fissure. In the eye 
more particularly described here the depression is filled largely with blood corpuscles 
(fig. 62 a, cpl.sng.). 

The pigmented layer is not more than 4 /a thick, and is very sparingly pigmented 
over the posterior face of the eye. At the iris and the lower margin of the choroid 
fissure it is continuous with the inner layers of the retina through cells whose nuclei 
are distinctly elongate. The retina proper, from the pigmented layer to the vit- 
reous cavity, is 64 fi thick. 





fiichd- cplsng 


Fig. 61. Two Vertical Sections of Eye of Individual about 5 mm. long. Fig. a taken through Lens, Vitreous 
Cavity, and Choroid Fissure. Fig. 6, Second Section Proximal to that from which fig. 61 a was drawn and 
passes through Innermost Part of Vitreous Body. Layers of Retina have not yet begtin to be differentiated. 

It is differentiated into a nuclear layer (the outer and inner together) and the 
ganglionic layer, separated by the incomplete inner reticular layer. The ganglionic 
layer is composed of two sorts of cells. Those nearer the vitreous cavity have much 
more distinct nucleoli than those nearer the reticular layer. Cell multiplication 
is still going on. 

The optic nerve is well developed, forming a solid strand of fibers, 12 /u. in diam- 
eter, readily traceable to the brain. 

The muscles are represented by strands of cells closely crowded. No striation 
is evident. 


In embryos 6 mm. long the cells giving rise to the oblique muscles and those 
for at least 2 of the recti can be distinguished. Scleral cartilages are not yet formed. 

In 3 of the specimens sectioned there was no indication of a lens. In others it 
was well developed. Cell division was still going on in the retina. 

The optic vesicle was very shallow. The rim of the vesicle was wide and still 
continuous with the choroid fissure, which showed as a shallow groove along the 
ventral surface. The choroid fissure, instead of leading into a central secondary 



optic cavity, led to the mass of ganglionic cells (fig. 62 c). This condition of the 
choroid fissure and its relation to the interior of the eye leads me at this point to say 
a few words concerning the general structure of the eye. In the description of the 
eye of the adult I considered that the central ganglionic mass was the result of the 
collapsing of the eye with the disappearance of the vitreous body and cavity. I 
was justified in this conclusion by the process of degeneration going on in the eye 
of Typhlomolge, Typhlichthys, and Typhlogobius. Whatever may have been the 

e-ih.^. ^^,,^_^^^ 


Fig. 62. (a) Anterior Face of Transverse Section of Left Eye of Larva 5 mm. long. Sections run obliquely 
in such a way that Ki^lit Eye is cut first, series beginninK in front. Divergence from Spherical 
Outline is due to Pressure of Brain on Proximal Face and Epidermis on Distal Face. 

(6) Anterior Face of Transverse Section of Left Eye of Larva 6 mm. long. No Lens in connection 
with this Retina. 

(c) Parasagittal Section of Eye of Larva 6 mm. long, showing Ventrally Choroid Fissure represented 
by space between Pigmented Layers and Vitreous Cavity represented by Shallow Depression on 
Ventral Face. Retina differentiated into (iangiionic. Inner Reticular, and Nuclear Layers. 

(rf) Anterior Face of Transverse Section through Right Eye of Larva, 7.5 mm. long. 

(e) Horizontal Section through Middle of Eye of Larva, 7 mm. long, showing Choroid Groove. 

phylogenetic process in Amblyopsis, it is evident that ontogenetically the mass of 
cells does not arise as imagined. It appears from the embryos that the condition of 
the adults arises more as the result of a contracting of the retinal area without a cor- 
responding decrease in the size of the eye as a whole than as the result of the col- 
lapsing of a vesicle followed by the coalescence of the walls brought together by the 
collapse. Sagittal sections of the eye (fig. 62 c) show the lips of the choroid fissure 
drawn apart with the contraction of the retina, only the dorsal two thirds of the 
eye reaching full development. From a study of the embryos of this size the point 


of exit of the optic nerve which marks the proximal end of the choroid slit alone 
gives evidence that potentially, at least, we have to do with an eye from which a 
central cavity has disappeared, i. e., in which it does not develop. 

The optic nerve is well developed, arising apparently from the ventral cells of 
the ganglionic mass, that is, those immediately lining the potential optic cavity. 

The pigment cells are well developed and have a varying depth in different parts 
of the eye. They are low and without pigment over the front of the eye and the 
ventral surface near the choroid slit. 

The retinal layers proper are differentiated into the ganglionic layer or mass 
which occupies the central and lower part of the interior of the eye. Apparently 
only the more centrally placed cells of this mass give rise to fibers. The inner 
reticular layer surrounds the ganglionic mass above and partly on the side, not at 
all below. The nuclear layers are well developed, without a differentiation into 
outer and inner layers or any indication of an outer reticular layer. The latter 
structure is apparently never formed at all. 


The variability in the rate of development of the eye is well seen in a series of 
specimens about 7 mm. long and whose eyes are little if any beyond the stage of 
development reached in other specimens only 5 mm. long taken from another female. 
In the former the eye is in contact with the dura proximally, but is withdrawn from 
the epidermis by 36 fi or more. A strand of cells extends from the eye upward 
and outward to the thinnest part of the epidermis. The epidermis is distinctly 
thinner over the eye than in neighboring regions. 

The eyeball is subspherical, with a shallow groove along its ventral surface 
representing the choroid slit (fig. 62 e). 

In half of the specimens of this size examined no lens could be detected. In 
one the lens was a comparatively large pear-shaped structure whose cells were 
undergoing degeneration, if the numerous dark granules in them were indicative 
of degeneration. In one individual in which no lens could be found on one side, 
that of the other side was probably represented by a small group of cells lying 
between the eye and the skin (Ins. 63 c). The cells were breaking apart and the 
outline of the structure as a whole was irregular. In all cases the lens lies out- 
side the iris, and in fact the entire vitreous space is not large enough to hold the 
lens in such eyes as still show this structure. 

The pigment layer is pigmented over the dorsal part of the eye. In vertical 
sections no pigment appears below the entrance of the optic nerve. The iridian 
part of the layer is, as usual, without pigment. The ganglionic cells, as in the 
last stages described, are exposed to the exterior through the choroid fissure, or 
where this is not evident there is no differentiation into different layers along the 
line of the choroid fissure. The ganglionic cells placed at the distal face of the eye 
give off fibers to the optic nerve. Fibers have not been definitely traced to the cells 
of the same series occupying the proximal or middle position. The optic nerve 
reaches a thickness of 20 ft and breaks up into bundles a short distance within the 
eye. These bundles radiate, forming an incomplete funnel-shaped structure. The 
incomplete inner reticular layer only partially separates the ganglionic and the 



nuclear layers. The relative development of the pigment layer and the inner 
reticular layer both show a less degree of differentiation than the same layers 
in the eyes of another series of larvae only 6 mm. long. This is due to the 
individual variation in the rate of development, not to degeneration since the 
last stage. 

Dividing cells are found in the nucleated layer. In the nuclear layers some 
nuclei elongated in a vertical direction are probably the nuclei of the Mullerian 


— kW 

'-i^'^Q stretin. 


sLni.ex.•itt,fi'^'^l'd■ ^ ' 


KiG. 63. 



Horizontal Section lo y. Dorsal to that given in fig. 60 ^, and showing Iris and Vitreous Cavity. 

Outline of Lens of Same Eye as that shown in figs. 60 e and 61 a but at a Level Dorsal of fig. 6r 

Region between Eye and Epidermis of Larva 7.5 mm. long, showing Degenerating Lens. 

Lens of Larva about 7 mm. lung. 

Vertical Section near Center of Right Eye of Fish 9.5 mm. long. 

Anterior Face of Transverse Section through Eye of Fish g.5 mm. long. 

Horizontal Section through Left Eye of Fish o-S mm. long. 


In larvae 9 to 10 mm. long the eyes lie from 60 to 100 /a removed from the epi- 
dermis and in contact with the brain capsule or but little separated from it. Their 
average measurements are: longitudinal diameter, 114 /a; antero-posterior, 98 /li.; 
vertical, 106 /a (figs. 63 e, 64 V). 

The epidermis over the eye has assumed the thickness found over neighboring 
regions, and from now on till death by old age there are no external modifications 
to indicate the former position of the cornea. 

The pupil is still open, and also the choroid fissure in the region of the pupil 
(figs. 63 e, g). In the proximal portions the choroid fissure is indicated by the 
absence of pigment along the ventral line (fig. 63 /). The vitreous cavity is a 


shallow depression in the distal face of the eye with a very narrow slit, sometimes 
a line, separating the iris from the solid mass of cells representing the retina. The 
vitreous cavity formed by the ventral invagination, that is, proximal of the iris, is 
obliterated in some individuals except in so far as the absence of pigment along a 
median line and in the union of the ganglionic layer with the pigmented layer along 
this line indicates its presence. The choroid fissure has been noted in an individual 
over IOC mm. long, so that evidently in some cases it may not close. Blood-vessels 
are still present in the vitreous cavity as far as it is developed. The distance from 
the exit of the optic nerve to the ventral margin of the pupil is considerably less 
than the distance between the exit of the optic nerve and the dorsal margin of the 

A few nuclei, probably the remnants of the hyaloid membrane, lie over the 
distal face of the retina. 

In ID specimens sectioned, all of them from 9.5 to 10 mm. long, the lens has 
disappeared without leaving any trace. 

The pigmented layer increases in thickness from the iris to the exit of the optic 
nerve. Its pigmentation also increases from the iris to the optic nerve. Within 
any one cell the pigment is uniformly distributed. In the dorsal part of the eye the 
pigment reaches to the iris, while in the ventral it does not reach so far, and in fact 
in a line from the optic nerve to the iris very few (only about 3) cells are pigmented. 
The maximum thickness of this layer is 12 /a. 

The inner cells of the iris have taken on their elongate shape which distin- 
guishes them in the eye of the adult, where the region of the iris and pupil can not 
otherwise be distinguished. 

The layers of the retina are now well developed except that the ganglionic mass 
of cells occupying the center of the eye is continuous with the outer nuclear and the 
pigmented layers along the ventral line. The outer and inner nuclear layers are 
represented by about 4 rows of nuclei immediately within the pigmented layer. 
The cells represented by these nuclei are not separable into an outer and an inner 
layer histologically, nor is there any break indicating the presence of any outer 
reticular layer. The cells form a compact layer of approximately uniform 
thickness. There are no indications of cones in any of the eyes examined. 

The inner reticular layer is well developed except along the region of the cho- 
roid fissure, where, as has been said above, the nucleated layers of the retina meet. 
There is possibly one exception to this in one of the eyes, in which the reticular 
layer surrounded the optic nerve at its entrance to the eye (fig. 63/). 

The space ventral to the central axis of the eye is occupied by the mass of 
ganglionic cells. This mass is irregularly trumpet-shaped, with the narrow end 
of the trumpet at the entrance of the optic nerve and the wide end at the distal 
part of the retina, where its cells are continuous with those of the nuclear layers. 
In the distal face of the trumpet, in what would be its hollow end, there is a dis- 
tinct conical area free from cells and abundantly supplied with fibers (fig. 63 g). 
It is possible that this represents the optic-fiber layer. The optic nerve is well 
developed, but its fibers seem to go to their respective cells directly without first 
going to this apparent optic-fiber layer. The outer nuclear layers measure about 
20 fi, the inner reticular about 8 fi, and the ganglionic layer about 32 /a in thickness. 
The changes taking place between 10 and 25 mm. are insignificant. 




The eyes of adult individuals from 25 to 75 mm. long were fully described in 
a previous chapter, and the eyes of very old individuals were mentioned briefly. 
The most highly developed eye found was that of an individual 75 mm. long. 
This eye is much above the average in the development of its pigmented layer, 
etc. Perhaps 25 mm. represents the stage at which the eye as a whole reaches its 
maximum development. 


The question of the rate and amount of growth of the eye from the time it 
appears can best be answered by the following table of measurements of the eyes 
of successive sizes of embryos. Attention should be called to the great varia- 
bility of the size of the eye in any one stage or in successive stages of development. 

It is seen from this table that the eye reaches the full vertical and longitudinal 
diameter of the adult when the embryo is only 2 mm. in length. Since the eye 
does not make its appearance till the embryo has reached a length of 1.5 mm. 
and the lens does not begin to develop until i mm. has been added to the length 
attained by the embryo after the eye has reached its full size, that is, not until it 
has reached a length of 2.5 mm., it is apparent that from the beginning the eye 
is in longitudinal and vertical diameter equal to the full adult eye. 

Table of Measurements of the Eye from the Time of First Appearance to Maturity. 

[Alt measurements are given in micra, except lengths of embryos, wliich are in millimeters.] 

Condition of embryo; living, 

or if preserved, 

direction of the sections. 











Mounted entire . 








Mounted entire. 




Length of 










6.5 to 7 

5-5 to 7 

6.5 to 7 

9 to 9.5 

9 to 9-5 







60 toicS 

Number of 
















1 60 




Axial diame- 
ter from cor- 
nea to optic 





80 and 108 







16 to 48 

16 to 36 or none 
18 to 5oor none 


of optic 



* The following gives the individual measurements of the eyes of the seven specimens whose average is here noted : 

No. 1. 

No. i. 

No. 3- 

No. 4- 

No. 5. 

No. 6. 

No. 7. 

LonRilmlinai diameter . . . 

Verlical diameter 













The lens begins to develop when the embryo is about 2.5 mm. long (fig. 59 b). 
It forms as a thickening of the skin where the optic vesicle is in contact with it. 
It is still connected with the skin when the embryo has reached a length of 4.5 mm. 
(Compare figs. 59 b, 59 c, 60 d, 60 e, 63 c with figs. 60 a, 60 b, the latter repre- 
senting the development of a normal lens.) The history of the lens after this stage 
is somewhat uncertain. It is well established that the cells composing it never 
lose their embryonic condition, that they are never differentiated into fibers. In 
iTiany eyes, certainly in all in which a lens could be detected in later stages, the 
lens becomes separated from the skin (fig. 60 e). The separation is completed 
when the larva has reached a length of 5 mm. (fig. 62 a). From this stage on, 
the lens begins to be resorbed ; in some 6-mm. larvae it could no longer be found 
(fig. 62 b). In 7-mm. larvae exactly half the eyes were without a lens (figs. 63 b, 
c, d), and in 9 to lo-mm. larvae no trace of a lens could be detected. The his- 
tory of the lens is completed. Judging from this rapid and universal disappear- 
ance of the lens in the young I am inclined to the opinion that the structure 
described in the adult eye as a lens is not a lens. 

The lens is the first organ to stop developing, the first to begin to degenerate, 
and the first to disappear. 


Attention was called to the variation of the scleral cartilages. A study of the 
development of the cartilages has enabled me to detect perhaps a greater degree 

of uniformity of plan, if not of structure, in these carti- 

^c lages than I was able to make out from a study of the 

"" " adult alone. It would seem that there are normally two 

/S] cartilaginous bars of variable shape developed. One or 

W ' ' both of them may be replaced by two or more smaller 

* cartilages. One of the cartilages is found over the distal 

Fio. 64. ca) |cieraj ^^^j.^'^F^^"'^^'**!;! f^^e of thc eye and the other on the posterior face caudad 

tak"n. ■ of the optic nerve. The earliest stages at which carti- 

(J) Scleral Cartilage of Left '^ . ° 

Same Si ^°'^" ^'''' °' l^gcs Were noticed were 9.5 to 10 mm. (figs. 63 g, 64 a, 0) 
long. In one fish 10 mm. long there were in the right 
eye about 10 cartilage cells, all directly over the pupil and iris. In the left eye 
there were about 22 cells, all over the dorsal part of the iris, none of them in 
front of the pupil. There were no traces in these eyes of scleral cartilages 
elsewhere. The cartilage cells were still for the most part isolated, not bound 
together into a definite cartilage. 

In another fish 10 mm. long the cells were definitely bound together into a 
small cartilage in each eye, that of one side encroaching on the pupil, that of the 
other side not. 

In a fish 25 mm. long there were two cartilaginous masses in each ej'e. One 
of these was over the distal face of the eye, the other over the caudal face of the 
eye caudad of the exit of the optic nerve (plate 10, fig. b). The one over the distal 
face curved ventro-caudad. 

In a fish 30 mm. long the cartilages were confined to the caudal half of the 
eye and were developed in such proportions that they encroached on the eye. 


The development of these cartilages to such unexpected size indicates that these 
cartilages are self-determining and not conditioned by the stimulus to growth by 
the eye with which they are in contact. In the right eye of this fish there were 
two cartilages in close contact with each other over the distal face. A third car- 
tilage lay on the dorsal, surface of the proximal part of the eye. The larger one 
of the two distal cartilages measures 63 by 32 by 65 /*, with a maximum diameter 
of the eye of 12 /x. 

In a fish 22 mm. long there were no cartilages on the proximal faces of the 
eye. In the right eye there was a cartilage 128 ju. long by 40 /u, thick, curved along 
the ventral part of the distal face. In the left eye there were two much smaller 
cartilages on the distal face of the eye. 

In a fish 35 mm. long there were two cartilages in the left eye placed as in 
the fish 25 mm. long, but they were larger. In the right eye the distal cartilage 
was represented by two cartilages in contact with each other. 

From the above it is seen that the distal cartilage arises first (lo mm. stage), 
the proximal ones not till much later (25 to 30 mm. stage). The cartilages do 
not reach their maximum size till later.' 

The distal cartilage in older fishes is frequently nodular and lies in front of 
the eye, where it was taken to be the lens by one of the earliest observers. In a 
specimen 90 mm. in length a globular cartilage 62 /a in diameter lay just over 
the pupil of the eye, which had a total diameter of 84 /a. One or the other car- 
tilage not infrequently encroached on the general outline of the eye. 

In the left eye of an individual 105 mm. long there were no traces of a scleral 
cartilage; the right eye was not examined. In the right eye of an individual 
108 mm. long there was a single large cartilage, 134 /u. by 208 fi, lying at one side 
of the center of the distal face of the eye. In the right eye of an individual 123 mm. 
long a minute cartilage was found on the proximal face of the eye. It was not 
determined whether one occurred over the distal face. In the left eye of the same 
fish a large cartilage lay over the distal face (plate 10, fig. d). 

In the left eye of the largest fish a single large cartilage 64 /x by 96 /a in sec- 
tion occupies the region to one side of the distal face (plate 10, fig. d). In the right 
eye (plate 10, fig. f) the distal cartilage measured 48 /a by 160 /x in section, and two 
smaller proximal ones were also present, one of them 24 /x by 32 /a in section. 

The scleral cartilages are the last structure to appear in the development of 
the eye; they grow during the greater part of life and retain their structure to the 


The details of the formation of the optic nerve have not been followed. No 
indications of it were seen in the eyes of the embryos 4.4 mm. long. In the eyes 
of embryos 5 mm. long it is well developed, forming a solid strand of fibers 12 /a 
in diameter which is readily traceable to the brain. The optic nerve increases 
but little, if any, after its formation. Its development is rapid. In subsequent 
stages it is not always traceable from all the cells forming the ganglionic mass. 
In the 6-mm. larvae its fibers were distinctly traceable from the cells nearest the 
choroid fissure, while in later stages they were more distinctly traceable from the 

' In the original the words " and there is no evidence of degeneration in them even in the oldest fish " 
completed the sentence. This is not strictly true and is omitted. 


distal cells of the ganglionic group. The optic nerve can be followed to the brain 
in all the larval stages and in the young fish up to 25 mm. in length (plate 10, 
fig. b). The optic nerve is evident within the eye in older stages up to about 
100 mm. ; in the very oldest ones it could not be found. In individuals much 
more than 25 mm. long it was not possible to follow the nerve to the brain, though 
it could usually be followed for some distance from the eye. The fibers are never 
medullated, and so far I have not been able to give them a differential stain. 


The history of the eye may be divided into four periods : 

The first period extends from the appearance of the eye till the embryo reaches 
4.5 mm. in length. This period is characterized by a normal palingenetic devel- 
opment except that cell division is retarded and there is very little growth. 

The second period extends from the first till the fish is 10 mm. long. It is 
characterized by the direct development of the eye from the normal embryonic 
stage reached in the first period to the highest stage reached by the Amhlyopsis 
eye ; its latter half is further characterized by the entire obliteration of the lens. 

The third period extends from the second period to the beginning of senescent 
degeneration, from a length of 10 mm. to about 80 or 100 mm. It is character- 
ized by a number of changes, which, while not improving the eye as an organ of 
vision, are positive as contrasted with degenerative. There are also distinct degen- 
erative processes taking place during this period. 

The fourth period begins with the beginning of senescent degeneration and 
ends with death. It is characterized by degenerative processes only, which tend 
to gradually disintegrate and eliminate the eye entirely. It is questionable 
whether these changes should be called senescent. It may be urged that they are 
the result of disuse in the individual, or that the end product of these degenerative 
changes is the typical structure of the eye of Amblyopsis. 

First Period. — During the first period the eye arises as a solid outgrowth 
from tljB solid central nervous system when the embryo is about 1.5 mm. long. 
The outgrowth increases rapidly in size during the next 0.5 mm. of growth in 
length. The solid lateral outgrowth is bent back along the side of the brain, and 
its connection with the brain becomes constricted into the optic stalk. A cavity 
approximately central arises in the optic lobe at the same time that a cavity ap- 
pears in the central nervous system, which occurs when the embryo is about 2 mm. 
in length. The two layers of the optic vesicle formed by the appearance of the 
cavity are of about equal thickness. A little later the secondary optic vesicle is 
formed by the thickening of the skin over the eye and the consequent cupping of 
the distal face of the eye. The process reaches its culmination when the embryo 
has a length of 4.4 mm. The lens is still connected with the skin, and the two 
layers of the secondary vesicle have become very different, the proximal one being 
one-layered, the distal one several-layered. The details of the changes of this 
period have been given in the preceding pages. 

At any time up to this length the eye might, as far as its structure is concerned, 
give rise to a perfect eye in the adult. The eye so far foUows phylogenetic paths 
with the reservation that no adult ancestor is supposed to have had eyes like these 
embryonic stages. 


The Second Period. — The development during the second period is direct and 
leads to the condition obtaining at the end of that period. Some of the processes 
are palingenetic, some are of purely ontogenetic significance, while still others (if 
I may make the distinction) are degenerative. 

The optic nerve develops at the beginning of the period in an undoubted phylo- 
genetic way. As in the case of the eye as a whole, the nerve develops directly into 
its full size. The details of its history are given under the head of the optic nerve. 
The latter half of the history of the lens belongs entirely to this period. Its his- 
tory is also given under another head. The changes the lens undergoes during 
this period are all katagenic, and some time before this period closes the lens has 

The direct development of the optic vesicle of the beginning of this period into 
the eye as found at the end of this period is very difficult to interpret satisfactorily. 

A comparatively very narrow marginal part of the secondary optic vesicle is 
converted into the epithelial part of the iris. The lens is almost always entirely 
excluded from the optic cup when the iris develops. The extreme shallowness of 
the optic cup and the comparative thickness of the retina would lead one to expect 
the choroid fissure proper to be a very short structure. The shallow cup develops 
into the adult eye by processes like those that take place in normal eyes. These 
purely palingenetic processes operating on so deficient material give rise to condi- 
tions that are not palingenetic. In the closing of the choroid fissure of the normal 
eye the thing of chief concern is the union of the infolded margins of the optic cup 
from the margin of the pupil to the point of exit of the optic nerve and the closing 
in of the retina around the optic nerve at its exit from the eye. In Amblyopsis 
the former process has become insignificant, and the latter the prominent process. 
This is further complicated by the fact that the vitreous cavity has ontogenetically 
disappeared nearly as much as phylogenetically, so that, while the processes of 
changing the optic cup into the eye are palingenetic, the material operated upon 
being quite different from that normally obtaining in fish embryos, the resulting 
stages of the eye are not palingenetic. 

The choroid fissure, which is distally a distinct slit leading into what remains 
of the optic cavity, becomes proximally a groove in a solid mass of cells. The 
closing of this groove takes place at various times, or it may remain permanently 
open. This condition has undoubtedly been brought about by a contraction of 
the area of the retina and the consequent heaping up of cells, either concomitantly 
with, or as the result of, the obliteration of the optic cavity. The funnel-shaped 
mass of cells in the center of the Amblyopsis eye is thus the result of the phylogenetic 
rather than the ontogenetic disappearance of the optic cavity. 

I must confess that an easier way of explaining the developmental stages would 
be reached by assuming that the central mass of cells, through which the optic 
nerve passes, is not really ganglionic — that only the distal cells of the mass are 
ganglionic — and that the proximal ones are the homologues of the cells found 
at the point of entrance into the eye of Chologaster (fig. 65, z). This would imply 
that a cavity has not disappeared from the center of these cells (because there 
never was one), and that the entire vitreous cavity has been reduced to that now 
found in the embryo, and that no part of the cavity has disappeared in toto. This 
interpretation is especially suggested by figure 62, c. This would account for the 



.111. ex, 

Fig. 65. Exit 
showing GangI 

fact that the optic nerve does not form a central strand through the funnel of 
ganglionic cells, but passes through it in several strands as it does through the 
mass of cells at the entrance of the optic nerve (fig. 65). The objection is that 

it would not account for the position of 
the exit of the optic nerve, vi'hich should, 
according to this view, be at the proxi- 
mal end of the choroid fissure. The 
second objection is found in the phylo- 
genetic stages of degeneration indicated 
in different eyes, notably that of Typh- 
lomolge. Furthermore, it would not 
account for the groove that is un- 
°'. o.p'i£?""'^'/?"„^y=2,' choiogaster papmiferus, doubtcdlv fouttd along the ventral side 

glioDic Cells and (s) cells at Entrance 01 Optic Nerve. •' ^ 

of the larval eye, nor would it account 
for the presence of the inner reticular layer around the optic nerve. It would, 
moreover, make it necessary to assume that the cells found about the entrance of 
the optic nerve in Chologaster have been retained in Amblyofsis out of all pro- 
portion to the other structures of the eye. These objections seem to me fatal to 
this second supposition. 

During this period the differentiation of the several layers of the retina also 
takes place. At the beginning of the period the pigmented layer is represented 
by a layer of thin cells without pigment. At the end of the period it is composed 
of cylindrical cells 12/1, high which are markedly pigmented. Pigment granules 
first make their appearance when the larva is about 5 mm. long. The remainder 
of the retina is at the beginning of the period several cells deep without any dif- 
ferentiation into layers. The inner reticular layer first appears as a number of 
irregular spaces separating the ganglionic from the nuclear layer when the em- 
bryos are 5 mm. long. These spaces soon unite into a single layer, but this does 
not occur till the very latest stages of the period when the choroid fissure has been 
closed for some time, and in fact they may never form a layer entirely around the 
central ganglionic cells. In earlier stages the layer extends between the dorsal and 
lateral parts of the ganglionic and nuclear layers. The nuclear layers never 
becomt separated into outer and inner ones, nor is an outer reticular layer ever 
formed. There is no indication of cones such as are seen in some adult eyes. 
Miillerian fibers are well formed in older individuals at this period. 

The development of the scleral cartilages described under another head also 
takes place toward the close of this period. No dividing cells have been found in 
the eyes of specimens more than 7 mm. long. The nuclei of the retina in the 
lo-mm. stage are all granular and measure 4 to 5 /* in diameter. 

The Third Period. — This extends from the time the fish has reached a length 
of 10 mm. till marked senescent changes begin, which take place when the fish 
approaches 100 mm. in length. 

The nuclei of the retina, when the fish has reached a length of 25 mm., are 
no longer alike. There are two types of cells in all layers : cells with larger granu- 
lar nuclei, and cells with smaller compact or dense nuclei. The difference is per- 
haps due less to histogenesis than to the process of degeneration which has already 


set in. The cells with smaller nuclei are probably degenerate. In the oldest 
fish only cells of the second type are found. 

A number of changes take place during the third period, some of which can 
be classed neither as progressive nor as retrogressive. As the fish grows, the eyes 
are farther and farther removed from the surface. In the fish 25 mm. long they 
are nearly i mm. below the skin, and in the largest specimen examined they are 
as much as 5 mm. beneath the surface of the skin. The scleral cartilages develop 
progressively probably during the entire period, in some cases encroaching on the 
regular outline of the eye. Other processes which are progressive nevertheless 
do not tend to make the eye a more perfect organ of vision. The pupil, for in- 
stance, becomes closed in many cases, or reduced to a very minute opening. The 
vitreous cavity, which was still evident, becomes, concomitantly with the closing 
of the pupil, entirely obliterated. The pigmented layer becomes a variable struc- 
ture, the pigment granules being in many cases entirely absent. Rarely the pig- 
ment layer changes to a high columnar epithelium. The stages of this period have 
not been successively observed as in the younger period, and the genetic relation- 
ship of different stages is not always apparent. 

The Fourth Period. — This extends from the time the fish has reached a length 
of about 100 mm. to the end of its life. There are distinct features that charac- 
terize the eye of this stage (plate 10, figs. c-g). 

The fibrous capsule enveloping the eye is distinctly thicker than in younger 
stages. The scleral cartilages are as well developed as at any time.* The eye- 
muscles, as far as present, show no indication of degeneration and their striation 
can readily be made out in aU individuals. 

The most marked changes take place in the size of the eye itself. The pig- 
mented layer becomes distended to form a thin-walled vesicle of two or three times 
the diameter of the eye in previous stages (plate 10, figs. F and g). This develop- 
ment of the pigmented layer beyond the requirements of the retina has also been 
seen in the eyes of Rhineura and other blind vertebrates. The cells of this layer 
become spherical or attenuated and the columnar epithelium converted into a 
thin epithelium thickened in places. Within this vesicle, whose sides may be 
compressed, as in figure f, the rest of the retina forms an insignificant little ball 
of tissue. In an eye of an individual 105 mm. long whose pigmented epithelium 
forms a vesicle 320 /x in diameter, the rest of the eye forms a small sphere 60 /n in 
diameter in contact with the iridian part of the pigment (plate 10, fig. g). The 
elements composing this little ball and representing the retina have also under- 
gone a marked senescent modification. The optic nerve is no longer evident.^ 
The ganglionic cells no longer form a compact mass, but are either unidentifiable 
or irregularly scattered. The cells of the outer nuclear layer are also less regular. 
While in the second period and up to 95 mm. in length two sorts of nuclei are 
distinguishable, some of them small and dense, others larger and granular. In these 
later stages they are all small and dense, no granular ones being present, and their 
outlines are less well defined than in the young. 

' In the left eye of a specimen 105 ram. long no cartilages were found. It is not possible to say whether they 
had disappeared or were never developed. Because of the irregularity in the development of these cartilages and 
their large size in other individuals of this period, I am inclined to think cartilages never appeared in this specimen. 

' The optic nerve can be traced as a very delicate filament through the pigment layer in an individual 123 
mm. long. In this eye the choroid fissure was still open. 



In a fish 25 mm. long the smaller nuclei measure 2.5 /x, the larger ones measure 
3-5 to 5 /lA. In the specimen 123 mm. in length the nuclei measure but 2 to 3 /u.. 
Evidence that the smaller nuclei in the younger specimen are degenerate is fur- 
nished by the fact that optic fibers can not be traced to the smaller ganglionic 
nuclei in a 25-mm. specimen. 

The most disorganized eye found is the left one of the largest fish examined, 
130 mm. long (plate 10, fig. e). The fibrous sheath (sclera) is thick; the cartilage 
is large, 64 by 96 fi in section. The eye itself is a disintegrated mass abundantly 
provided with granular pigment and without well-defined outline or structure. 
The right eye of the same specimen is less degenerate (plate 10, fig. r). It is an 
elongated vesicle 60 by 256 fi in section, with a large cartilage to one side of its 
distal half, 48 by 160 /a in section, and two smaller proximal ones, one of which 
measures 24 by 32 fi in section. Associated with the retina of this eye is a struc- 
ture that I described as a possible lens in my first paper. It consists of a few 
nuclei about which there are concentric layers of a homogeneous tissue. Consider- 
ing the fate of the lens in all the young fishes examined, it seems very doubtful, 
if not impossible, that this structure should be a lens. 

That the eyes of these largest individuals belong to the fourth period is seen in 
the fact that they become distended vesicles whose parts are finally resorbed after 
undergoing degenerative changes. The scleral cartilages offer an exception to the 
general fate. 

Summary of the Origin, Development, and Degeneration of the Eye and its Parts. 

Earliest appear- 
ance or differ- 

End of 


End of 

End of 

Beginning of 





Rarely and then 
after lo 














Before 25 








100 or before 
Before 2S 

Before 25 
Before 25 



Beyond 130 

Beyond 130 

Beyond 130 

130 mm. and beyond 
130 mm. and beyond 



Choroid fissure 

Pigmented layer. .. 

Outer nuclear 

Outer reticular.. . . 
Horizontal cells. . . 

Inner nuclear 


Optic fiber layer 

or nerve 

Scleral cartilages . . 

Corneal epithelium 

? I do not know. — Does not take place. 



On pages 134 et seq. an outline of the probable phylcgenetic history of the eye 
of Amhlyopsis is given. In the preceding chapter the rate of ontogenetic degen- 
eration and its extent has been found to vary in different parts of the eye. It has 
also been found that certain parts begin to degenerate earlier than others. We 
shall now attempt to discuss briefly the ratio between the rates and extent of onto- 
genetic degeneration and the rate and degree of phylogenetic degeneration implied 
by the structure of the eye. The discussion is somewhat intangible, but certain 
definite results can be obtained by it. 


In order to compare the ratio between the ontogenetic and phylogenctic rates 
of degeneration, it is necessary to use some stage in the development of the eye as 
the point which phylcgenetic degeneration has reached. For such a point we 
shall use the optimum reached by various parts of the eye during their develop- 
ment. It is certain that the phylogenctic stage is below this optimum, that some 
of the degeneration in individual eyes is due to phylogeny, but since we do not 
know how much of the descent from the optimum is due to heredity and how 
much to the peculiarities of the environment and the resulting functionless life of 
the parts during the life of the individual, it will be best to take the optimum as 
above indicated. 

All phylogenctic time is taken as a unit, although some parts of the eye have 
been degenerating longer than others. The ontogenetic degeneration leads from 
the optimum to the vanishing point for most parts of the eye. 

Ontogenetically the lens degenerates very rapidly, reaching its vanishing point 
from its optimum during the period in which the fish grows not more than 5 mm. in 
length. The rate of its phylogenctic degeneration must have been proportionately 
rapid, for at its optimum in Amblyopsis it is minute and its cells are undifferentiated. 
In the epigean relatives of Amblyopsis the lens is one of the parts least affected, so 
that it must have degenerated very rapidly in its later phylogenctic history, after 
the fish had entered the caves. 

At its best the vitreous body is so inappreciable in amount that I have not been 
able to consider its ontogenetic degeneration. Its phylogeny has approached the 
vanishing point toward which most parts of the eye are heading. 

The retina may be considered in its extent and in the degree of the histogenic 
differentiation of its parts. In the matter of its extent or size there is little change 
from its optimum until its disintegration in old age. Its ontogenetic changes are 
slight. Its optimum is comparable with that of the lens and indicates a rapid and 
great reduction from the lowest retina of epigean relatives. The ontogenetic and 
phylogenctic rates of degeneration in the extent of the retina differ greatly, the 
former having come practically to a standstill. 

In its histogenic differentiation the retina is not comparable with the lens, for it 
rises above the embryonic phases. In fact, in its histogenic differentiation the retina 
rises far above the requirements of the case, and the most highly developed eye of 
Amblyopsis approaches the lowest of its epigean relatives. Over any given area it 
is doubtful whether the ganglionic and inner reticular layers are more degenerate or 
as degenerate as the same parts in the eyes of Chologaster cornutus. It is certain 
that in their highest development the parts between the inner reticular and the pig- 
mented layers are below the lowest point reached by the corresponding parts in the 
epigean species mentioned. The same is true of the pigmented epithelium. 

The simplification of the structure of the retina from its maximum to its mini- 
mum in ontogeny is of greater extent than its simplification from the lowest differ- 
entiated retina found in epigean species to the maximum found in Amblyopsis. 

From the foregoing we may conclude that there is no constant ratio between 
the extent and degree of ontogenetic and phylogenctic degeneration, and that the 
observed rate of ontogenetic degeneration is not necessarily proportionate to the 
rate of phylogenctic degeneration inferred from the degree of degeneration of the eye 
at its optimum. 



There can be no doubt that the phylogenetic fate of the eye, exclusive of con- 
nective tissue, sheaths, sclera, etc., is total disappearance. The most degenerate 
ontogenetic eye indicates as much. There are no relatives of Amblyopsis that have 
reached this condition, but Troglichihys has an eye distinctly more degenerate than 
that of Amblyopsis. It may offer a clew as to whether any of the ontogenetically 
degenerate eyes, such as are found in old specimens of Amblyopsis, are prophetic 
of the condition through which the eye will pass in its route to the vanishing point. 
The most highly developed eye found in any specimens of Troglichihys (plate lo, 
fig. h) is comparable in a general way with the eyes of the old of Amblyopsis. The 
pigmented epithelium is larger than the requirements of the eye in both cases, 
and the scleral cartilages are disproportionately developed in both cases. The 
ganglionic cells extending through the center of the eye of the younger Amblyopsis 
are absent in both cases. Only 3 cells have been found in this region in all the eyes 
of Troglichihys examined. When we attempt a closer comparison, our efforts fail. 

We may conclude that if Troglichihys indicates one of the steps through which 
the eye of Amblyopsis will pass to its annihilation, the degenerative phases seen in 
the oldest specimens of Amblyopsis indicate only in a general way the phylogenetic 
path over which the eye will pass in the future. 



In my first paper on the Eyes of Blind Vertebrates (Roux' Arch. viii. p. 596, 
1899) I said : 

Cessation of development takes place only in so far as the number of cells are concerned. 
The number of cell generations produced being continually smaller results in an organ as a con- 
sequence also smaller. In this sense we have a cessation of development (cell division, not 
morphogenic development) in ever earlier stages. That there is an actual retardation of devel- 
opment is evident from Amblyopsis and Typhlichihys in which the eye has not reached its final 
form when the fish are 35 mm. long. 

I am convinced now that this statement did not go far enough. There is, indeed, 
a gradual retardation in all processes of development which frequently terminates in 
a complete arrest of development before the final stages of normal eyes are reached. 
This is especially true of the lens. In discussing the changes it will be best to keep 
separate the three groups of processes concerned in development. 

The proof of the limiting of the number of cell divisions mentioned has been 
brought out in the chapters on the development. It has also been seen that the rate 
of division is very much retarded. In the retina it stops altogether at the time the 
fish has reached a length of 5 to 7 mm., and very rarely more than two dividing 
cells are found in any eye. In its first stages the eye is thus about equal in size to 
the adult eye. Cell division stops earlier in the lens, where no new cells are formed 
after it is cut off from the skin. The lens is at this time relatively as well developed 
as the retina. In both the retina and the lens cell division ceases in late stages, and 
the total number of cell generations is very much limited. The lens is looked upon 
as phylogenetically a new structure, and we have, by the stopping of its later stages 
of cell division, a step in the elimination of a phylogenetically new structure. This 
is, however, of no consequence because it is not differential, for the retina, a phylo- 
genetically older structure, suffers a similar stoppage. There is no evidence, then, 


that phylogenetically younger structures lose their power of cell division earlier 
than phylogenetically older ones. 

The retardation of the morphogenic processes, cell arrangement, movement, 
union and separation, etc., is conspicuous in the delay of the closing of the choroid 
fissure and all that this implies. There is no conspicuous stopping of this process 
except in the occasional failure of the choroid fissure to close at all. 

Histogenic processes are also distinctly retarded, and in conspicuous instances 
suffer an entire stoppage. While the eyes of 3-mm. specimens of Cyniatogaster 
or Carassius and Amblyopsis are nearly alike, in the former two the tissue differ- 
entiation has progressed vastly farther by the time the fishes have reached a length 
of 10 mm. Histogenesis is carried surprisingly far in many degenerate eyes. In 
Rhineura, for instance, the layers of the retina are differentiated far beyond the 
requirements of the case. In Amblyopsis the process, as far as it can be made out 
with the methods available, falls short of the normal development.' The cells of 
the lens never lose their embryonic characters ; they are never transformed into lens 
fibers. Cones are rarely if ever developed in the retina, and an outer reticular layer 
never. In normal development the cones and the outer reticular layers are the last 
to differentiate, so that we have certainly a cutting off of late ontogenetic stages. 
The question whether these are also phylogenetically young may be passed over. 

The total evidence from the three processes is that none of them proceed with the 
push and rapidity found in normal structures, and though they are normal, they 
grow weaker with development and frequently give out altogether. But with all 
this lack of vigor, while there is more variation in each structure developed than has 
been noted in normal eyes, the point to which cell division, cell arrangement, and 
histogenesis are carried, in different individuals, is about the same. The causes 
leading to the changed development are of approximately equal value in different 
specimens from the same locality. 


The retardation and arrest in the ontogenetic development of the eye of Amblyop- 
sis may be due to one of several possible causes. They are either conditioned by 
something outside the cells composing the eye, or they are inherent or predeter- 
mined in the egg cell from which the eye is ultimately derived. The conditioning 
factor, if it lie outside the eye, may be a peculiarity in the physical and chemical 
environment in which the fish lives, or a lack of stimulation or an inhibition exer- 
cised by some other part of the body. Unless we assume that the eye of Ambly- 
opsis has reacted and does now react differently to the physical and chemical 
environment from that of some of the relatives of Amblyopsis, physical and 
chemical factors may readily be eliminated as contributing directly to the retarda- 
tion and cessation. 

Although, in discussing the phylogenetic degeneration of the eyes of cold- 
blooded vertebrates in general, I have insisted that cross-country conclusions must 
be guarded against, I then saw no objection, and now see none, to considering the 
different members of the Amblyopsidae as homogeneous material within the bounds 
of which we may expect similar causes to effect similar results. The different stages 

' The difEculties, for instance, of differentiating with Golgi methods the bipolar cells of an eye whose total 
diameter falls short of 0.2 mm. can readily be imagined. 


(phyletic) of development found in the eyes of the different members of the Ambly- 
opsidae are all referable to the difference in time in which they have been subjected 
to their present environment. The only environmental condition surrounding the 
developing eggs of Amblyopsis to which the peculiarities of development might 
be attributed is the total absence of light. 

Temperature, oxygen pressure, chemical composition, etc., of the surrounding 
medium may be entirely excluded from the possible agents affecting the eye, inas- 
much as normal eyes are developed by other fishes in the same water and under 
all possible fluctuations of the above conditions within the limits of the possibility 
of fish life. But the same objection holds in attributing the lack of development to 
the absence of light. Chologaster agassizii, a member of the Amblyopsidae, which 
always lives in caves in exactly the same conditions under which Amblyopsis lives, 
has nevertheless normally developed, though small, eyes. 

While guarding against the possibility of attributing too much weight to the 
results obtained in other families of animals, it still may be mentioned that many 
fishes living perpetually in total darkness develop normal eyes. This is also true of 
the young of all viviparous animals which develop in more or less complete darkness. 

If, then, so closely related fishes as Chologaster and Amblyopsis are subjected 
to the same environment which is minus a certain element and both develop their 
normal parental structure, one developing a normal eye, the other a very abnormal 
degenerate one, it is scarcely warrantable to say that the abnormal structure in one 
of them is due to the absence of the one element (light) from the environment. 
Moreover, if the development is controlled by the absence of light, there is no reason 
why development should be normal, even to the extent of forming a normal start and 
should then be arrested or retarded. The fact that the presence or absence of light 
is not the controlling factor in the retarded development of the eye of Amblyopsis 
does not vitiate the supposition that a certain amount of change may not be pro- 
duced on the eyes of an individual by rearing it in the light. Such change would, 
however, stand on a par with the ontogenetic degeneration of the eye with age in 
the absence of light ; that is, it would be a functional adaptation due to use. 

Experiments have been in progress to test the effect of light. So far only nega- 
tive results have been obtained. One young has been reared till it was 6 months 
old. It was obtained from the caves at a time when it was ready to swim about 
freely; that is, when the eye was already fully formed. There was no difference in 
the gross anatomy of the eye of this individual as compared with that of others. The 
minute anatomy, as the result of an accident, was not available for study. The 
others examined in earlier stages have not been reared beyond a length of a few 
millimeters, and the effect of the light, if any, was not appreciable. From the 
observations on the development of the eyes — which show that some processes are 
arrested very early — it would seem that the only rational way to determine the 
effect of light on the total development is to colonize the adults in an outdoor pool 
where the young can be reared, from the fertilization on, in normally lighted waters. 

The lack of development of the eye not being chargeable to any factor in the 
environment, is there any factor within the fish that inhibits its development, or 
whose absence fails to furnish the stimulus necessary to the development ? If so, 
this factor must be present or absent at the time the retardation begins or some time 


The inhibition, if any, might operate through a mechanical crowding on the part 
of a neighboring organ or the greater selective power in eliminating the food requisite 
for the development of the eye. The first cause may be eliminated, for there is no 
evidence whatever of crowding other than that found in normal eyes ; in fact, in 
all stages beyond the earliest, the eye is much smaller than the optic sockets can 
easily accommodate. 

The question of selective food elimination is not so readily disposed of. The 
ophthalmic artery provides the eyes abundantly with blood, so it is not an absence 
of this that causes the supposed starving. Indeed if the retardation were due to a 
lack of blood supply we would be removing the problem one step from the eye with- 
out solving it. Besides, Loeb's experiments have shown that the action of the 
heart may be greatly diminished without affecting the rate of growth of the larval 
fish. The blood supply being abundant, is there any other organ that may drain it 
of the nutriment necessary for the proper growth of the eye? Leaving aside the 
question whether an organ can be starved by having the nutriment requisite for 
development withdrawn from the blood by another organ, I can think of no organ 
or set of organs that attain an unusual growth aside from the tactile organs of the 
skin. This system of organs is undoubtedly very highly developed in the adult and 
has also attained a remarkable degree of development at the time the fish is lo mm. 
long. It is, however, not unusually developed in the earlier stages before hatching 
and shortly thereafter when the cessation of cell division, the most important 
element of the stunted optic development, takes place. Besides this, the tactile 
organs of Chologaster, which possesses normal eyes, are very highly, if not so elab- 
orately, developed as in Amhlyopsis. I have experimentally determined by eliminat- 
ing the eyes altogether that the tactile organs in Chologaster papilliferus are amply 
developed to enable the fish to live indefinitely without the use of its eyes. The 
same must also be true of Chologaster agassizii, which lives permanently in caves. 
While not impossible, it seems, therefore, very improbable that the tactile organs 
affect the development of the eyes in Amhlyopsis and not in Chologaster} 

I know of no other organs in Amhlyopsis whose development differs from that 
of Chologaster in a degree sufficient to make it a successful contestant for a food 
supply in Amhlyopsis and not in Chologaster. 

What has been said concerning organs whose presence might affect the develop- 
ment of the eyes is equally true concerning organs whose absence might deprive the 
eye of the necessary stimulus to reach normal development. I know of no organ, 
either in Amhlyopsis or Chologaster, whose absence in the one and presence in the 
other might account for the difference in the degree of development reached by the 
eyes in the two fishes. 

The conclusion is forced upon us by the above considerations that neither in the 
environment nor in the fish itself is there a factor sufficient to account for the early 
arrest in cell division, the retardation of the morphogenic processes, and the stopping 
of the histogenic processes. We are therefore entirely justified in assuming that 
the determining cause of the method of development lies in the cells themselves and 
is inherited. The great development of the scleral cartilages beyond the needs of 

' As an example bearing on this subject attention may be called to the tactile apparatus of the Silurida?, 
which is certainly in many instances more elaborate than that of Amhlyopsis, and yet the eyes are normal, though 


the eye also tend to locate the formative or hereditary power in the cartilages them- 
selves rather than in the stimuli to their development that they receive from their 
contact with the developing eyes, for they develop entirely beyond the needs of these 

The causes operating in ontogeny and phylogeny that have led to the limited 
power of development and differentiation I have fully considered in the concluding 
chapter, which was also published in the Popular Science Monthly.^ The conclu- 
sion is reached that the phylogenetic degeneration, which is equivalent to saying the 
limited power of development found in the cells entering into the eye of the indi- 
vidual, is the result of functional adaptation during the lifetime of past individuals 
to the total disuse of the eye. This adaptation, it was concluded, was transmitted 
to a certain extent to the succeeding generation through the usual vehicles of trans- 
mission. There has always been and is yet a serious objection to this conclusion, 
because the method of the transmission of functional adaptations to the organiza- 
tion of the egg so as to limit or extend its powers is not known. 

Recently, while admitting that functionally adaptive structures arise develop- 
mentally without reference to function, Driesch has maintained that: "Wer hier 
von 'Vererbung' friiher einmal functionell 'erworbener' Eigenschaften reden will 
verlasst den wissenschaftlichen Boden, denn wir wissen von solcher Art der Verer- 
bung gar nichts." 

Possibly we might find a warrant for the assumption of the transmission of func- 
tional adaptation to the germ cells in the writings of Driesch himself, though he 
might not thank us for it. He maintains that certain developmental results whose 
proximal cause he is not able to determine may be produced by factors working in a 
distant part of the embryo. Without entering into a discussion of the validity of 
these factors working at a distance, if they are really factors and capable of acting, 
as Driesch imagines, why may not functional modifications effect changes in the 
hereditary cells in a similar manner ? 

I conclude that retardation and cessation in development are not due to onto- 
genetically operating causes, but they are inherent in the fertilized ovum — they 
are inherited. 


During recent years the law variously termed von Baer's law, Agassiz's law, 
Haeckel's law, or the law of biogenesis, has been frequently called into question. Its 
general tenets are : (i) every individual in its development repeats in brief the devel- 
opment of the race ; (2) closely related forms have a similar ontogeny, and the nearer 
two animals are related the longer their embryos are alike ; (3) the embryos of high 
animals pass through stages resembling the adult stages of lower animals ; and 
(4) in every ontogeny there are, among the truly ancestral stages, stages which are 
adaptive and have been acquired during ontogenetic development. 

No objection has been raised to the fourth tenet in so far as its acceptance does 
not commit to the acceptance of the first. In objection to the first of these proposi- 
tions Hurst writes : 

I do not deny that a rough parallelism exists in some cases between ontogeny and phylogeny. 
I do deny that the phylogeny can so control the ontogeny as to make the latter into a record of the 

' The same conditions are foui>d in Lucifuga. ' See the next chapter. 


former — even into an imperfect record of it. * * * Vestiges, and these only, can give any em- 
bryological clew to past history which could not be equally well made out from comparative anatomy. 

Zittel finds cases in paleontology both in support of and against this first propo- 
sition : 

All know that it (development of Antedon) does not in the remotest manner agree with the facts 
of paleontology. * * * No observations of embryology would warrant our imagining the former 
existence of graptolites or stromatophores. No stage in the development of any living brachiopod 
informs us that numerous spine-bearing genera lived in Pala-ozoic and Mesozoic times. * * * 
The beautiful researches of Hyatt, Wiirtemberger, and Branco have shown that all ammonites and 
ceratites pass through a goniatite stage, and that the inner whorls of an ammonite constantly re- 
semble, in form, ornament, and suture-line, the adult condition of some previously existing genus 
or other. 

Smith finds that "the development of Placenticeras shows that it is possible to 
decipher the race history of an animal in its individual ontogeny." 

But it is not the intention to review the numerous expressions of opinion pro 
and con which have appeared on this subject in recent years. A full discussion of 
the literature to 1897 has been given by Keibel. 

The eye of Amblyopsis presents, however, such an excellent opportunity to test 
an opinion vaguely expressed by Balfour in his " Embryology," and carefully and 
clearly stated by Sedgwick and reiterated by Cunningham in his " Sexual Dimor- 
phism " and in other places, that the facts presented in the foregoing pages may 
be reexamined in their relation to this point. 

Balfour says : 

Abbreviations take place because direct development is always simpler, and therefore more 
advantageous ; and, owing to the fact of the foetus not being required to lead an independent exist- 
ence till birth, and of its being in the mean time nourished by food-yolk, or directly by the parent, 
there are no physiological causes to prevent the characters of any stage of the development which 
are of functional importance during a free, but not during a foetal, existence from disappearing from 
the developmental history. * * * In spite of the liability of larvae to acquire secondary characters, 
there is a powerful counterbalancing influence tending toward the preservation of ancestral char- 
acters in that larvae are necessarily compelled at all stages of their growth to retain in a. functional 
state such systems of organs, at any rate, as are essential for a free and independent existence. It 
thus comes about that, in spite of the many causes tending to produce secondary changes in larvae, 
there is always a better chance of larvae repeating, in an unabbreviated form, their ancestral history 
than is the case with embryos which undergo their development within the egg. 

The most concrete critique of the law of biogenesis has been offered by Sedgwick. 
After rejecting the second proposition by showing that, while in many cases the 
adults differ more from each other than the young, in other cases the embryos differ 
more from each other than the adults, he takes up the main question stated in the 
first proposition by a consideration of "The Significance of Ancestral Rudiments in 
Embryonic Development." It is, indeed, around this phase of the subject that 
the discussion has centered. His views are best given by a series of excerpts from 
his paper. Thus Sedgwick states that 

* * * The tendency in embryonic development is to directness and abbreviation and to the 
omission of ancestral stages of structure, and that variations do not merely affect the not-early 
period of life where they are of immediate functional importance to the animal, but, on the contrary, 
that they are inherent in the germ and affect more or less profoundly the whole development. 

The evidence is of this kind : i. Organs which we know have only recently disappeared are not 
developed at all in the embryo. For instance, the teeth of birds, the fore limbs of snakes, reduced 


toes of bird's foot (and probably of horse's foot), the reduced fingers of a bird's hand. * * * 
2. Organs which have (presumably) recently become reduced or enlarged in the adult are also 
reduced or enlarged in the embryo. * * * 3. Organs which have been recently acquired may 
appear at the very earliest possible stage. * * * The latter arrangement ["ancestral organs 
have disappeared without leaving a trace"] seems to be the rule, the former the exception. 

I think it can be shown that the retention of ancestral organs by the larvae [embryos?] after 
they have been lost by the adult is due to the absorption of a larval or immature free stage into em- 
bryonic life. A larval character thus absorbed into the embryonic life, its disappearance is no 
longer a matter of importance to the organism, because, the embryo being protected from the struggle 
for existence, the pressure of rudimentary functionless organs is unimportant to it. Characters 
which disappear during free life disappear also in the embryo, but characters which, though lost 
by the adult, are retained in the larva may ultimately be absorbed into the embryonic phase and 
leave their traces in embryonic development. 

To put the matter in another and more general way. The only functionless ancestral structures 
which are preserved in development are those which at some time or another have been of use to 
the organism during its development after they have ceased to be so in the adult. * * * But another 
explanation is possible, which is that organs which are becoming functionless, and disappearing 
at all stages, may in some case disappear unevenly, that is to say, they may remain at one stage after 
they have totally disappeared at another. 

The question seems to me not quite so simple as imagined by Sedgwick. De- 
generate organs may or may not be better developed in the young than in the adult. 

(i) They are better developed in the young if they are still functional in the 
young after they have become functionless in the adult. 

(2) They may be better developed in the young, if they were of use to the 
young, after they ceased to be of use to the adult. 

(3) They may be well developed in the young after complete disappearance in 
the adult if the material is used for other purposes in later life. 

(4) They are better developed in the young if their presence is essential to pro- 
vide the necessary stimulus to bring about or to inhibit cell movements or cell dif- 
ferentiation in the development of other organs. 

(5) They are supposed to be no better developed in the young than in the 
adult, if they ceased to be of use to the young when they lost their use in the adult. 

The material entering into the formation of the eyes is not used for the building 
up of other organs, and it is uncertain whether the eyes positively or negatively influ- 
ence the development of other organs, so that a discussion of numbers 3 and 4 of the 
above possibilities is not profitable. Inasmuch as both young and adult live perma- 
nently in total darkness, and the eye of the young can not be functional under the 
present mode of existence, the first possibility is also eliminated from the discussion. 

In Amblyopsis, which carries its young in its gill cavity, we are undoubtedly 
dealing with an animal in which the eyes are useless in the young as well as in the 
adult and in which they became totally useless in the young at the same time that they 
became totally useless in the adult, that is, at the time the species took up permanent 
quarters in the caves. Do the eyes in this case repeat the phylogenetic history of the 
eye, or have the eyes in the embryo degenerated in proportion to their degeneration 
in the adult ? In this form the question is whether a perfect or better eye is produced 
to be finally metamorphosed into the condition found in the adult, or whether 
development of the eye is direct. 

We have seen in the preceding pages that the foundations of the eye are nor- 
mally laid, but that the superstructure, instead of continuing the plan with new 


material, completes it out of the material provided for the foundations, and that in 
fact not even all of this (lens) material enters into the structure of the adult eye. 
The development of the foundations of the eye are phylogenetic, the stages beyond 
the foundations are direct to the present adult condition of the eyes from which 
they are now ontogenetically degenerating to the vanishing point. 


The study of the development and its related questions shows: 

1. The eye of Amblyopsis appears at the same stage of growth as in fishes 
developing normal eyes. 

2. The eye grows but little after its appearance. 

3. All the developmental processes are retarded and some give out prematurely. 
The most important of the latter is the cell division and the accompanying growth 
that provides the material for the eye. 

4. The lens appears at the normal time and in the normal way, but its cells 
never divide and never lose their embryonic character. 

5. The lens is the first part of the eye to show degenerative steps and it disappears 
entirely before the fish has reached a length of i mm. 

6. The optic nerve appears shortly before the fish reaches 5 mm. in length. It 
does not increase in size with the growth of the fish and possibly never develops 
normal nerve fibers. 

7. The nerve does not increase in size with growth of the fish. 

8. The optic nerve gradually loses its compact form, becomes flocculent, 
dvrindles, and can not be followed by the time the fish has reached 50 mm. in 
length. In the eye it retains its compact form for a much longer time, but disap- 
pears here also in old age. 

9. The scleral cartilages appear when the fish is 10 mm. long ; they grow very 
slowly — possibly till old age. They do not degenerate at the same rate as other 
parts of the eye, if they degenerate at all. 

10. The history of the eye may be divided into four periods : 

(a) The first period extends from the appearance of the eye till the embryo 
reaches 4.5 mm. in length. This period is characterized by a normal 
palingenetic development except that cell division is retarded and there 
is very little growth. 

ip) The second period extends from the first till the fish is 10 mm. long. 
It is characterized by the direct development of the eye from the nor- 
mal embryonic stage reached in the first period to the highest stage 
reached by the Amblyopsis eye. 

(c) The third period extends from the second period to the beginning of 
senescent degeneration, from a length of 10 mm. to about 80 or 100 
mm. It is characterized by a number of changes which, while not 
improving the eye as an organ of vision, are positive as contrasted 
with degenerative. There are also distinct degenerative processes 
taking place during this period. 

{d) The fourth period begins with the beginning of senescent degeneration 
and ends with death. It is characterized by degenerative processes only 
which tend to gradually disintegrate and eliminate the eye entirely. 


11. For a summary of the origin, development, and degeneration of the eye and 
its parts see table, page 164. 

12. There is no constant ratio between the extent and degree of ontogenetic 
and phylogenetic degeneration. The observed rate of ontogenetic degeneration 
is not necessarily proportionate to the rate of phylogenetic degeneration inferred 
from the degree of degeneration of the eye at its optimum. 

13. If TrogUchlhys indicates one of the steps through which the eye oi Ambly- 
opsis will pass to annihilation, the degenerative phases seen in the oldest specimens 
of Amhlyopsis indicate only in a general way the phylogenetic path over which 
the eye will pass in the future. 

14. Some late stages of development are omitted by the giving out of develop- 
mental processes. Some of the processes giving out are cell division, resulting in 
the minuteness of the eye and the histogenic changes which differentiate the cones 
and the outer reticular layer. 

15. There being no causes operative or inhibitive either within the fish or in 
the environment that are not also operative or inhibitive in Chologasler agassizii, 
which lives in caves and develops well-formed eyes, it is evident that the causes 
controlling the development are hereditarily established in the egg by an accumu- 
lation of such degenerative changes as are still notable in the later history of the 
eye of the adult. 

16. The foundations of the eye are normally laid, but the superstructure, 
instead of continuing the plan with additional material, completes it out of the 
material provided for the foundations. The development of the foundation of the 
eye is phylogenetic, the stages beyond the foundations are direct. 



As in all organs no longer of use or hindrance, and therefore no longer under 
the control of selection, the individual variations in the structure of the eye of 
Amblyopsis, Troglichthys, and Typhlichthys are very great.' There is also 
a marked change in the eye w^ith age. It is therefore necessary to distinguish be- 
tween individual variations and stages in ontogenetic and phylogenetic degeneration. 
The eye of each species has a general structure which is typical for the species. 
The individual variations have been sufficiently described under the respective 


The steps in degeneration in the Amblyopsidae are indicated in figure 66. The 
most highly developed eye is that of Chologaster papilliferus. The parts of this 
eye are well proportioned, but the eye as a whole is small, measuring less than 
I mm. in a specimen 55 mm. long. The proportions of this eye are symmetrically 
reduced if it has been derived from a fish eye of the average size. The retina is 
much simpler than in Zygonectes. The simplifications in the retina have taken 
place between the outer nuclear and the ganglionic layers. The pigment layer 
has not been materially affected. These facts are exactly opposed to the 
supposition of Kohl that the retina and the optic nerve are the last to be affected, 
and that the vitreous body and the lens cease to develop early. In Chologaster 
papilliferus (b) the latter parts are normal, while the retina is simplified. That 
the retina is affected first is proved beyond cavil by cornutus (a). The vitreous 
body and the lens are here larger than in papilliferus, but the retina is very greatly 
simplified. Cornutus, it must be borne in mind, lives in the open. The eye of 
Chologaster agassizii (c) differs from that of papilliferus largely in size. There 
is little difference in the retinas except the pigmented layer, which is about 26 
per cent thinner in agassizii than in papilliferus. 

If we bear in mind that no two of the eyes represented here are members of 
a phyletic series, we may be permitted to state that from an eye like that of cornutus, 
but possessing scleral cartilages, both the eyes of Amblyopsis and Troglichthys 
have been derived, and that the eye of Amblyopsis represents one of the stages 
through which the eye of Troglichthys passed. The eye of Amblyopsis (h) is the 
eye of Chologaster cornutus minus a vitreous body with the pupil closed and with 
a minute lens or more probably none at all. The nuclear layers have gone a step 
farther in their degeneration than in cornutus, but the greatest modification has taken 
place in the dioptric apparatus. 

In Troglichthys (i) even the mass of ganglionic cells present in the center of 
the eye as the result of the collapsing after the removal of the vitreous body has 
vanished. The pigmented epithelium, and in fact all the other layers, are repre- 
sented by mere fragments. 

The eye of Typhlichthys (g) has degenerated along a different line. There is 
an almost total loss of the lens and vitreous body in an eye like that of papilliferus 

' This is also true of the eye of Luci/uga and Stygicola. 



without an intervening stage like that of cornutus, and the pigment layer has lost 
its pigment, whereas in Amblyopsis it was retained. 

The reduction in size from the normal fish eye went hand in hand with the 

Flc. 66. DiafH'ams of Eyes of all Species of Amblyopsida" and Typhlomolge, d,e,g.h.aad i drawn under same 
magnification, (a) Clwhga-ti cornutui, (6) Ciwlogasler papiiltjerus^ (c) Chologaster agassizii, drawn 
to scale; (rf) Retina of Ckohgasler cornulus; (e) Retina of Chotogaster papHlijerus; (/^ Eye of Typhlo- 
molge under lower magnification than d-f; (s) Eye of Typhlichthys sublerrantus ; (A) Eye of Amblyopsis 
spelaus; (t) Eye ol Troglichthys rosa, 

simpHfication of the retina. There was at first chiefly a reduction in the number 
of many times duplicated parts. Even after the condition in Chologaster papillif- 
erus was reached the degeneration in the histological condition of the elements did 
not keep pace with the reduction in number (vide the eye of cornutus). The 


dioptric apparatus disappeared rather suddenly, and the eye, as a consequence, 
collapsed with equal suddenness in those members which, long ago, took up their 
abode in total darkness. The eye not only collapsed, but the number of elements 
decreased very much. The reduction was in the horizontally repeated elements. 
The vertical complexity, on which the function of the retina really depends, was not 
greatly modified at first. 

In those species which took up their abode in total darkness the degeneration 
in the dioptric apparatus was out of proportion to the degeneration of the retina, 
while in those remaining above ground the retinal structures degenerated out of 
proportion to the changes in the dioptric apparatus, which, according to this view, 
degenerates only under conditions of total disuse or total darkness which would 
necessitate total disuse. This view is upheld by the conditions found in Typhlo- 
gobius, as Ritter's drawings and my own preparations show. In Typhlogohius 
the eye is functional in the young and remains a light-perceiving organ throughout 
life. The fish live under rocks between high and low tide. We have here an 
eye in a condition of partial use and the lens is not affected. The retina has, 
on the other hand, been horizontally reduced much more than in the Amblyopsidae, 
so that, should the lens disappear, and Ritter found one specimen in which it was 
gone, the type of eye found in Troglichthys would be reached without passing 
through a stage found \n Amblyopsis ; it would be simply a horizontal contracting 
of the retina, not a collapsing of the entire eye. 

The question may with propriety be asked here : Do the most degenerate eyes 
approach the conditions of the pineal eye? It must be answered negatively. 



The different structures of the eye may now be taken up in detail. 

(a) The eye muscles are normally developed in Chologaster. They are present 
to a greater or less extent in Amblyopsis. They have been reduced in number in 
Troglichthys, where the half nearest the eye has been replaced by bundles of fibrous 
tissue. In Typhlichthys they have vanished. 

(b) The scleras of the different members are not comparable on account of the 
presence of cartilage in some species and not in others. Both this layer and the 
choroid are insignificant in Chologaster and Typhlichthys. In Amblyopsis cartilages 
different in size and number are found anywhere about the eye, being frequently 
present in shape and position to suggest a displaced lens. In thickness the cartilages 
are disproportionate to the size of the eye. In Troglichthys we have a still more 
evident misfit, for the scleral cartilages are both too long and too thick. Evidently 
the scleral cartilages have not decreased in size in the same ratio as the eye, or, 
what amounts to the same thing, they develop beyond the present needs of the 
eye. (See also Luci/uga.) 

(c) The choroid is thin in all cases except where pigment cells are situated. 
These are frequently several times as thick as the rest of the choroid. In Ambly- 
opsis the pigmentation of the choroid is inversely proportional to the pigmentation 
of the retina. 

(d) The lens has already received sufficient attention. It is merely necessary to 
insist again that, as long as an eye is functional to any extent, the lens — in fact the 
dioptric apparatus in general — does not degenerate and that when absolute disuse 


comes, the lens, both phylogenetically and ontogenetically, disappears rapidly. In 
Typhlogobius Ritter found the lens absent in one very old individual, and Cope 
found that in Gronias the lens is sometimes present on one side, while not on the 
other. In Amblyopsis and Typhlichthys it has degenerated to a mere vestige, or 
is gone altogether. Ritter, after considering the structure of degenerate eyes as 
far as known at the time, came to the conclusion "that the lens disappears before 
the retina; and that, where degeneration takes place at all in ontogeny, the lens is 
affected first and most profoundly." With the first part of this statement the more 
recent observations are in full accord. It is, however, doubtful whether the lens 
is ever the first part affected ; in fact the retina always leads, but certainly the lens, 
if affected at all, is affected profoundly. 

(e) There is more variety in the degree of development of the pigment epithelium 
than in any other structure of the eye. Ritter has found that in Typhlogobius 
this " layer has actually increased in thickness concomitantly with the retardation 
in the development of the eye, or it is quite possible with the degeneration of this 
particular part of it. An increase of pigment is an incident to the gradual diminu- 
tion in functional importance and structural completeness." There is so much 
variation in the thickness of this layer in various fishes that not much stress can be 
laid on the absolute or relative thickness of the pigment in any one species as an 
index of degeneration. While the pigment layer is, relative to the rest of the retina, 
very thick in the species of Chologaster, it is found that the pigment layer 
of Chologaster is not much if any thicker than that of Zygonectes, but exception 
must be made for specimens of the extreme size in papilliferus and agassizii. 
In other words, primarily the pigment layer has retained its normal condition, 
while the rest of the retina has been simplified, and there may even be an increase 
in the thickness of the layer as one of its ontogenetic modifications. Whether 
the greater thickness of the pigment in the old Chologaster is due to degeneration 
or the greater length of the cones in a twilight species I am unable to say. 

In Typhlichthys, which is undoubtedly derived from a Chologaster-like an- 
cestor, no pigment is developed, the layer retains its epithelial nature and remains 
apparently in its embryonic condition. It may be well to call attention here to the 
fact that the cones are very sparingly developed, if at all, in this species. In 
Amblyopsis, in which the degeneration of the retina has gone farther, but in which 
the cones are still well developed, the pigment layer is very highly developed, but 
not by any means uniformly so in different individuals. The pigment layer reaches 
its greatest point of reduction in rasa where pigment is still developed, but the layer 
is fragmentary except over the distal part of the eye. We thus find a development 
of pigment with an imperfect layer in one case, Troglichthys, and a full-developed 
layer without pigment in another, Typhlichthys. In the chologasters the pigment 
is prismatic ; in the other species granular. 

,(/) In the outer nuclear layer a complete series of steps is observable from the 
two-layered condition in papilliferus to the one-layered in cornutus, to the undefined 
layer in Typhlichthys and the merging of the nuclear layers in Amblyopsis, and 
their occasional total absence in rosa. The single cones disappear first, the cones 
long before their nuclei. 

(g) The outer reticular layer naturally meets with the same fate as the outer 
nuclear layer. It is well developed in papilliferus and agassizii, evident in Cholo- 



gaster cornutus, developed in spots in Typhlichthys, and no longer distinguishable 
in the other species. 

Qi) The layers of horizontal cells are represented in papilliferus by occasional 
cells; they are rarer in cornutus and beyond these have not been determined with 

(j) The inner nuclear layer of bipolar and spongioblastic cells is well developed 
in C. papilliferus and C agassizii. In cornutus it is better developed in the young 
than in the older stages, where it forms but a single layer of cells. There is evi- 
dently in this species an ontogenetic simplification. In the remaining species it is, 
as mentioned above, merged with the other nuclear layer into one layer which is 
occasionally absent in Troglichthys. 

(j) The inner reticular layer is relatively better developed than any of the other 
layers, and the conclusion naturally forces itself upon one that it must contain 
other elements besides fibers of the bipolar and ganglionic cells, for, in Amblyopsis 
and Troglichthys, where the latter are very limited 
or absent, this layer is still well developed. Hori- 
zontal cells have only been found in the species of 

(k) In the ganglionic layer we find again a com- 
plete series of steps from the most perfect eye to the 
condition found in Troglichthys. In papilliferus 
and agassizii the cells form a complete layer one 
cell deep exxept where they have given way to the 
optic fiber tracts which pass in among the cells 
instead of over them. In cornutus the cells have 
been so reduced in number that they are widely 
separated from each other. With the loss of the „ , ^. ^ . ..t-,^,. 

^ Fig. 67. DiaRram showing per cent of Total Thick- 

vitreous cavity the cells have been brought together ZZ:hrld7:cCl!^^%Z:;/s!iT^':'. 
again into a continuous layer in Typhlichthys, Y^.,:^^ ,t\T^;„'"Z^X^tl:%ZJ't^-, 
although there are much fewer cells than in cornutus tlii'^l^^.TVASp^srt fetfc' • 
even. The next step is the formation of a solid core '°' ^'■"«'''*""'*- 
of ganglionic cells, and the final step the elimination of this central core in Troglich- 
thys, leaving but a few cells over the anterior face of the retina. 

(/) Miillerian nuclei are found in all but Amblyopsis and Troglichthys. In 
C. cornutus they lie in part in the inner reticular and the ganglionic layer. Cells of 
this sort are probably also found among the ganglionic cells of Typhlichthys. 

We thus see that the simplification or reduction in the eye is not a horizontal 
process. The purely supporting structures like the scleral cartilages have been 
retained out of all proportion to the rest of the eye. The pigment layer has been 
both quantitatively and qualitatively differently affected in different species. There 
was primarily an increase in the thickness of this layer, and later a tendency to 
total loss of pigment. The degeneration has been more uniformly progressive in 
all the layers within the pigment layer. The only possible exception being the 
inner reticular layer, which probably owes its retention more to its supporting 
than to its nervous elements. Another exception is found in the cones, but their 
degree of development is evidently associated with the degree of development of 
the pigmented layer. As long as the cones are developed, the pigmented layer is 
well developed, or vice versa. 



The simplification of the eye in cornutus has been mentioned in the foregoing 
paragraphs. It may be recalled that the nuclear layers are thinner in the old than 
in the young. There is here not so much an elimination or destruction of element 
as a simplification of the arrangements of parts, comparatively few being present 
to start with. 

The steps in ontogenetic degeneration can not be given with any degree of finality 
for Amblyopsis on account of the great variability of the eye in the adult. While 
the eyes of the very old have unquestionably degenerated, there is no means of 
determining what the exact condition of a given eye was at its prime. In the largest 
individual examined the eye was on one side a mere jumble of scarcely distinguish- 
able cells, the pigment cells and scleral cartilages being the only things that would 
permit its recognition as an eye. On the other side the degree of development 
was better. 

The fact that the eyes are undergoing ontogenetic degeneration may be taken, 
as suggested by Kohl, that these eyes have not yet reached a condition of equilibrium 
with their environment or the demands made upon them by use. Furthermore, 
the end result of the ontogenetic degeneration is a type of structure below an)1:hing 
found in the phlyogeny of the vertebrate eye. It is not so much a reduction of the 
individual parts as it is a wiping out of all parts. 


Does degeneration follow the reverse order of development or does it follow 
new lines, and if so, what determines these lines ? Since the ontogenetic development 
of the eye is supposed to follow in general lines its phyletic development, the above 
question resolves itself into whether or not the eye is arrested at a certain stage of 
its morphogenic development, and whether this causes certain organs to be cut off 
from the development altogether. In this sense the question has been answered 
in the affirmative by Kohl. Ritter, while unable to come to a definite conclusion, 
notes the fact that in one individual of Typhlogobius the lens which is phyletically 
a new structure had disappeared. This lens had probably been removed as the 
result of degeneration rather than through the lack of development. Kohl supposes 
that in animals placed in a condition where light was shut off more or less, every 
succeeding generation developed its eye less. Total absence of light must finally 
prevent the entire anlage of the eye. Time has not been long enough to accomplish 
this in any vertebrate. Phyletic degeneration is looked upon as the result of a long 
series of " H mmungen " which in successive generations appeared in ever earlier 
time of ontogenetic development in always lower stages of the development of the 
individual eyes. The eye develops after the vertebrate type. At certain stages 
the rate of progress is diminished and in most cases finally completely ceases. A 
retardation has developed which after a shorter or longer period ends in the cessa- 
tion of all development. The first appearance of the retardation falls in a time 
of embryonic or post-embryonic development that in the phylogeny corresponds 
to the moment when the lack of light became operative. The period in ontogeny 
which lies between the first disturbance in development and its cessation corre- 
sponds to the phyletic time during which the development of the eye is checked at 
a continually lower stage of development. The point of cessation in ontogeny cor- 
responds to the time when the eye reached its equilibrium. If in ontogeny there 


is undoubted degeneration, it is always an indication that the eye has not yet 
reached the point where it is in equilibrium with its functional requirements. 

Cessation of development does not take place at the same time in all parts of 
the eye. Those not essential to the perception of light are disturbed first. The 
retina and the optic nerve are the last affected, the iris coming next in the series. 
Because the cornea, aqueous and vitreous bodies, and the lens are not essential 
for the performance of the function of the eye, these structures cease to develop 
early. The processes of degeneration follow the same rate. Degeneration is 
brought about by the falling apart of the elements as the result of the introduction 
of connective tissue cells that act as wedges. Abnormal degeneration sometimes 
becomes manifest through the cessation of the reduction of parts that normally 
decrease in size so that these parts in the degenerate organ are unusually large. 

Kohl's theoretical explanation here given somewhat at length is based on the 
study of an extensive series of degenerate eyes. He has not been able to test the 
theory in a series of animals living actually in the condition he supposes for them, 
and has permitted his erroneous interpretation of the highly degenerate eye of 
TrogUchthys to lead him to this theory of the arresting of the eye in ever earlier 
stages of ontogeny. It has been shown in previous pages that this most degenerate 
eye is in an entirely different condition from that supposed by him. The mere 
checking of the normal morphogenic development has done absolutely nothing 
to bring about this condition, and it could not have been produced by the checking 
of development in ever earlier and earlier stages of ontogeny, for there is no stage 
in normal ontogeny resembling in the remotest degree the eye of TrogUchthys. 
The process of degeneration as seen in the Amblyopsidae is in the first instance 
one of growing smaller and simpler — not a cutting off of late stages in the develop- 
ment. The simplified condition, it is true, appears earlier and earlier in ontogeny 
till it appears almost along the entire line of development, even in the earliest stages. 
But the tendency for characters added at the end of ontogeny to appear earlier and 
earlier in the ontogeny is well known, and there is no inherent reason why an organ 
disappearing in the adult should not eventually disappear entirely from ontogeny. 
The fact that organs which have disappeared in the adult have in many instances 
not also disappeared in the ontogeny and remain as so-called rudimentary organs 
has received an explanation from Sedgwick. For a discussion of this see the chap- 
ter on the Law of Biogenesis. 

In Amblyopsis, where the eye has not been functional at any period of ontogeny 
for many generations, where degeneration begins at an early period and continues 
till death, the degenerate condition has reached the early stages of the embryo. 
It is only during the first hour or so that the eye gives promise of becoming any- 
thing more than it eventually does become. The degree of degeneration of an organ 
can be measured as readily by the stage of ontogeny when the degeneration becomes 
noticeable as by the structure in the adult. The greater the degeneration, the farther 
back in the ontogeny the degenerate condition becomes apparent, unless, as stated 
above, the organ is of use at some time in ontogeny. It is evident that an organ 
in the process of being perfected by selection may be crowded into the early stages 
of ontogeny by post-selection. Evidently the degenerate condition is not crowded 
back for the same reason. How it is crowded back, I am unable to say. A satis- 
factory explanation of this will also be a satisfactory explanation of the process 


by which individually acquired characteristics are enabled to appear in the next 
generation. The facts, which are patent, have been formulated by Hyatt in his 
law of tachygenesis. Histogenic developipent is a prolonged process, and onto- 
genetic degeneration is still operative, at least, in Amblyopsis. 

Degeneration is not the result of the ingrowth of connective tissue cells as far 
as I can determine. It is rather a process of starving, of shriveling, or resorption 
of parts. 

From the foregoing it is evident that degeneration has not proceeded in the 
reverse order of development, rather the older normal stages of ontogenetic develop- 
ment have been modified into the more recent phyletic stages through which the 
eye has passed. The adult degenerate eye is not an arrested ontogenetic stage of 
development, but a new adaptation, and there is an attempt, in later ontogeny at 
least, to reach the degenerate adult condition in the most direct way possible. 




The Cuban blind fishes were discovered by the surveyor D. Tranquilino San- 
dalio de Noda. They were described as Luci/uga subterraneus and Lucifuga 
dentatus by Poey, in his "Memorias sobre la Historia Natural de la Isle de Cuba," 
tomo 2, pp. 95-114, 1856. Poey recorded them from the cave Cajio, near La Guira 
de Helena, La Industria, half-way between Alquizar and Guanimar, the Cave of 
Ashton, the Cave of the Dragon, on the cattle farm San Isidro, near Las Mangas, 
La Concordia, a cave near the bee house of the coffee plantation La Paz, and a well 
near the tavern Frias. 

Poey stated that Lucifuga dentatus from some of the caves had vestiges of eyes, 
while those from others were without the least vestige of eyes. Poey later added 
some notes on their distribution in his "Enumeratio Piscium Cubensium." In 
1863 Gill (Proc. Acad. Nat. Sci. Phila., 1863, p. 252) recognized Lucifuga dentatus 
as the ijpQ of a distinct genus, which he called Stygicola. 

No additions were made to the knowledge of these fishes until March, 1902, 
when I visited Cuba with Mr. Oscar Riddle expressly to secure material for the 
study of their eyes. We visited several of the caves mentioned by Poey and many 
others, securing 119 specimens of both species. One of the specimens contained 
four young, making in all 123 specimens. The discovery that the blind fishes are 
viviparous, and that the young have fairly well developed eyes, made it seem very 
desirable to secure a full series of embryos and also if possible to rear some of them 
in the light. 

The expenses of this trip were defrayed in part by a grant from the American 
Association for the Advancement of Science and in part from subsidiary work on 
the fresh-water fishes of the western end of the island. The results as far as pub- 
lished are included in an article on the "Fresh-water Fishes of Western Cuba" 
(Bull. U. S. Fish Com., 1902, pp. 211-236, plates 19-21, 1903). 

Grant No. 64 of the Carnegie Institution made additional work in the field 
possible. It was planned to spend the entire breeding season near the caves and 
rear young in the light, but for reasons that will appear the grant was exhausted in 
apparently determining that these fishes do not breed in the places visited. 

My trip to Cuba in March, 1902, made it seem probable that the blind fishes 
give birth to their young in February. Many recently born young of Lucifuga 
were obtained at that time, and one of the females caught contained young nearly 
ready to be born. The California viviparous fishes, with which I had extensive 
experience^ and which give birth to young in a similar degree of maturity, carry their 
young about 5 months. On these premises I concluded that early stages of the 
young of the blind fishes should be found during the middle of September. Allow- 
ing a month for the probably more rapid development in the tropics, I visited the 
caves the latter part of October and first part of November. 

' The specimens were numbered as they were collected, i, 2, etc., and when referred to are given by their serial 

' On the viviparous fishes of the Pacific coast of North America, Bull. U. S. Fish Com., 189a, pp. 381-478, 
2^ plates, 1894. 



Aside from obtaining young it was planned to build cages in a well-lighted cave 
in which the adult would be compelled to carry and give birth to their young in the 
light. The body walls in the majority of individuals would offer little or no obstacle 
to the penetration of light to the embryos. 

Dr. J. W. Beede, of the Geological Department of Indiana University, acted as 
volunteer assistant and rendered very valuable aid in collecting fishes, making the 
cages, and taking the traverse to the various caves in the chief cave region about 
Caiias. Only a single individual with young was obtained and one other with nearly 
mature eggs. Two cages were built and fishes were confined in them and the cages 
sunk in the Modesta, a well-lighted cave in which fishes were naturally abundant. 

On December i a few fishes were collected and sent me by Mr. F. Martinez, of 
Canas. Although these promised little better success than the ones collected in 
October and early November, I started for Cuba again on December i8, 1903, 
accompanied on this trip by Mr. John Haseman, as volunteer assistant. It was 
again found that this was not the breeding season, as no fishes with young were 
found at all. The cages were found intact and received a new supply of fishes. 

On May i a number of fishes were sent me by Mr. F. Martinez, and as these 
promised no young the trip planned for May was abandoned. On June i, when 
Mr. Martinez was again to send me samples, he was unable to obtain any fishes on 
account of high water. 

Between June and August I could not get away from my routine work, but this 
period was later covered by Mr. Haseman. On August 15, 1904, I started again 
for Cuba, accompanied by Mr. Hankinson as volunteer assistant. I returned Sep- 
tember 7. On this trip, which was more extensive than the former, I obtained two 
females with young, one a Lucifuga containing 10 young, and one a Stygicola con- 
taining I young. On this occasion I visited two new locaHties. At one of these, 
Jovellanos, from which Poey reported Stygicola, I obtained nothing. At the other, 
the Carboneria farm, on the north coast near Matanzas, I obtained my first speci- 
mens from the northern slope of Cuba. I am under many obligations to Dr. Felix 
Garcia, the harbor health officer of Matanzas for the opportunity to visit the 

At this time the cages in the Modesta were found to be entirely spoiled, the wire 
screening having corroded in large pieces. I succeeded in bringing living fishes to 
Indiana, but it was not possible to bring large numbers. There was great mor- 
tality en route on account of the extreme sensitiveness to cool water, which rules 
entirely out of court the idea of colonizing them in some of our northern caves. 

In June, 1905, two of my students, Mr. J. Haseman, who had accompanied me 
on one of the trips, and Mr. Norman Mclndoo, made another tour of the caves, 
but with no better success as far as embryos were concerned. They secured but 
one female with young. 

The following papers have appeared on material gathered during the various 
Cuban trips : 

1. The Blind Fish of Cuba. Science, N. S., xvi, p. 347. 

2. Eigenmann, C. H. The fresh-water fishes of western Cuba. Bull. U. S. Coram. Fish and Fisheries, 

1902, pp. 211-236, plates 19-21. 

3. The water supply of Havana. Science, N. S., x^'ni, pp. 281-282. Aug. 28, 1903. 

4. In search of Blind Fish in Cuba. World To-day, V, pp. 11 29-1 136. 

5. Auf dcr Suche nach blindcn Fischcn in Cuba. Die Umschau, vil, pp. 365-367. 

6. Hay, W. P. On a small collection of crustaceans from the island of Cuba. Proc. U. S. Nat. Mus., xxvi, 

pp. 429-435. Feb. 2, 1903. 

7. Lane, H. H. The ovarian structures of the viviparous blind fishes, Lucifuga and Stygicola. Biological 

Bulletin, vi, pp. 38-54, 1903. 


Carboneria beach near Matanzas. 

Dividing line between naked beach (on right) and sand-filled area (on left). 
Rift separates the two zones. Bushes on extreme left mark line of older beach. 

^mf<^_ ^ 

^ft {■ 

Cave of the Insurrectos, new the Carboneria, from entrance. Pool of water 
showing at bottom of cave. 


8. Muhse, E. F. The eyes of 7"ji/>A/o/ii /«m6r«co/«j (L.), a blind snake from Cuba. Biol. Bull., v, pp. 261- 

270, Oct. 1903. 

9. Pike, F. H. The degenerate eyes in the Cuban cave shrimp, Palcemonetes eigenmanni Hay. Biological 

Bulletin, xi, pp. 267-276, 1906. 

10. Payne, F. The eyes of A mphisbiena punctata (Bell), a blind lizard from Cuba. Biol. Bull., xi, pp. 60- 

70, plates I and 11, July 1906. 

11. Weckel, A. L. The fresh-water Amphipoda of North America. Proc. U. S. Nat. Mus., xxxn. De- 

scribing a new Amphipod, Gammarus ccecus, from the Modesta Cave, Cuba. pp. 47-49, 1907. 

12. Haseman, J. D., and Mclndoo, Norman N. On some fishes of Western Cuba. Proc. Acad. Nat. Sci. 

Phil., 1906. 


Lucifuga and Stygicola are members of the Brotulidae, of which Jordan and 
Evermann say: "These fishes are closely related to the Zoarcids. In spite of 
various external resemblances to the Gadidae, their affinities are rather with the 
blennioid forms than with the latter." 

They are most closely related to the genera Brosmophycis and Ogilbia, with 
which they have a distinct caudal peduncle in contradistinction to the numerous 
other American genera of the family. Brosmophycis marginatus (Ayres) occurs 
on the coast of California in moderate depth. Ogilbia ventralis (Gill) occurs in 
rocky pools about the Gulf of California and at La Paz. The other member of the 
genus, Brosmophycis cayorum, was taken on a shoal covered with algas at Key West. 

Other members of the family are found at great depths in various parts of the 
world ; one, Brotula barbata, occurs about Cuba in water of moderate depth. 

The genera Lucifuga and Stygicola differ from each other in their dentition. 
Stygicola has teeth on the palatines; Lucifuga has none. In Stygicola the nape 
is more strongly arched than in Lucifuga. The maximum recorded size of Stygi- 
cola is 152 mm. ; of Lucifuga, 104 mm. 


The male organ of Stygicola consists of a conical papilla, two-lobed at the tip 
and surrounded by a dermal pouch. It reaches to the second or third anal ray, 
being turned either to one side or the other of the anal. It is pigmentless, but is 
covered from in front by a pigmented dermal flap. 

In color, Lucifuga varies from a faint pink to lilac-pink and lilac. There is, 
in general, an increase of pigment with age. Stygicola varies from pinkish lilac to 
steel-blue, with transparent edges to the fins. There is no regular increase of color 
with age in this species nor is there any distinction in the sexes. Both black and 
light-colored individuals are found side by side in caves. It is possible that light- 
colored individuals have lived in the remote recesses of the cave and that the black 
ones have remained in the lighted chambers, but there is no direct evidence on this 

The males of Stygicola are distinctly larger than the females. The average 
length of 137 females caught is 98.2 mm., the largest one being 140 mm. The 
average size of the 82 males is 107 mm., the largest one being 152 mm. long. In 
the first lot secured the males were in excess of the females. in the ratio of 100 
females to 115 males. In all I have 137 females to 82 males. Counting the first 
43 specimens secured, there is but an appreciable difference in the average of the 
fins as far as these could be counted, the average formula for the female being, D. 
91.4; A. 74; and for the males D. 91. i; A. 73.3; or the average for the two, 
D. 91.2; A. 73.6. 


Of Lucifuga * 74 males have an average length of 63.5 with a maximum of 104, 
and 82 females have an average length of 58 mm., with a maximum of 95. Only 
specimens over 50 mm. in length were considered. 

While the average number of rays diflfers considerably in the two species, the 
number in each varies so much that the numbers in individual cases overlap, the 
individuals of Lucifuga reaching as high as 88 dorsal rays, and the individuals of 
Stygicola as low as 87. The same is true with the anal. 


Stygicola is known to be distributed from Jovellanos and Alacranes on the east 
to Canas. Lucifuga is confined to the region from Guira de Helena westward to 
Canas. The entire region between Alacranes and Canas on the southern slope is 
drained by underground rivers. In the Canas region, the two species live side by 
side with apparently no choice, except that while the young of Lucifuga are abun- 
dant in shallow water among the roots of trees I have not been able to see or secure 
Stygicola shorter than 60 mm. except as larvae from the mother. Stygicolas are 
perhaps more abundant in the deeper, darker caves, though they are also found 
in the shallowest, while lucifugas are more abundant in shallower, more open caves, 
they in turn being found in the deeper caverns. Blind fishes resembling Stygicola 
or Lucifuga have been reported to me from well-like caves at Merida, Mexico. 
None have been captured. Other blind fishes which may be related to them are 
said to occur in Jamaica. 


Within the area over which they are distributed the blind fishes of Cuba live, 
as far as known, in well-like caves in coralline limestone. The character of the 
region in which they live can best be understood from an examination of the Finca 
Carboneria, just outside of the Bay of Matanzas. There is here a coral strand 
about on a level with high water. 

At the point of contact between ocean and land there is an abrupt wall, 5 to 10 
feet high, profusely covered with seaweed, the nearly tideless water coming to the 
top of the wall where there are shallow, panlike pools replenished by waves and 
spray. Immediately on top of the wall follows a low,- naked, jagged mass of rock 
resembling a huge sponge with its numerous pits and points. This area is in- 
habited by innumerable moUusks. This low area is separated by a cleft (plate 11, 
fig. a) forming a sharp line of demarcation from a second zone similar to the first, 
but in which the pits and depressions in the rock have become filled with sand which 
gives foothold to tufts of plants. Over this lizards scamper from rock to rock. 
Following this there is an abruptly sloping beach, the outer half of which is rocky 
and sandy, partly covered with cactus and other low-growing plants, the inner or 
land half being covered with shrubs and trees. All of these zones occupy per- 
haps 100 yards. They are followed by the level, practically treeless, meadow, 

' The following account was published of the first 53 specimens of Lucifuga secured: The females are dis- 
tinctly larger than the males. In making the average for the size of the sexes, individuals less than a year old 
were not considered, because differences in the sexes, if present, could be but very slight, and because in such young 
the sex could not always be determined with certainty. An examination of all specimens makes it probable that 
at the end of a year after birth the young are about 50 mm. long. In obtaining the average size of the sexes only 
those specimens over 50 mm. were considered. The males above this size measure 59.7 mm. on an average, with 
a maximum of 94 mm. ; the females measure 71. i mm. on an average, with a maximum of 93 mm. Of the speci- 
mens over 50 mm. long, 23 were males and 22 females, or 100 females for every 104.5 males. Counting the fin 
rays of the first 43 specimens over 50 mm. long, we get males, D. 82.1, A. 67.4; females, D. 81.9, A. 68. The average 
formula for those less than 50 mm. long is D. 83; A. 67.2, or for all together, D. 8a.6; A. 67.5. 



perhaps 0.75 mile wide and less than 10 feet above sea-level. It is such a beach 
as is shown in figure A, plate 11, raised to a Httle higher elevation. There is here 
but little sandy soil, the underlying rock coming near the surface. The slope of 
the hill behind this level stretch is composed of bare rocks very similar to those 
of plate II, figure a, except that the gnarled roots of the densely growing stunted 
shrubs and trees twist about the rocks and into the crevices. The character of this 
area was very well described by my host, who dryly remarked, when I asked him 
whether I should go on horse to the caves on top of the hills, "No, you will go on 
your hands and knees." The disagreeable impression that these hills make on 
one traversing them on foot in the heat of the day is heightened by the innumerable 
hermit-crabs that lurk in every cranny and scamper over the rocks. At an eleva- 
tion of about 100 feet is another level stretch of rocks with a thin layer of sandy soil. 
Within less than a quarter of a mile from the ocean is a natural well, improved 
somewhat with the chisel. It is circular, with a diameter of about 6 feet and is 
less than 10 feet deep. It is evidently situated along the line of an original fissure 
in the coralline rock such as is shown in plate ii, figure A, for there are openings 
in opposite sides of the deeper part of the well that have an indeterminable extent. 
The surface of the water in this well is near sea-level, about 4 feet below the level of 
the land. The water, over 5 feet deep, is perfectly fresh and blind fishes were 
more abundant in this well than in any other area of the same extent. 

Flc. 6S. DiaRrams of Cave of the Insurreclos iind the Carboneria Well (fig. R, plate ii) taken from X. 3. Depression 
about Mouth of Cave; a. Dry Cave; 1, The Pool of Water near Sea-level, S.L, and with Submerged Stalactites 
and Stalagmites; f, Side Rifts in Carboneria Well. 

There are a number of caves on the plateau over the hills and I visited two of 
these. They are within 4 or 5 miles of the seashore. Their mouths lie at an ele- 
vation of about 100 feet (87 and 93 by barometer). In general character these caves 
are like others visited in Matanzas province, i. e., at Matanzas and at Alacranes, or 
Alfonso XII. They occur in a level area and from a distance there is nothing to 
indicate their presence. There is first a slight depression in the level country 
(fig. 68 (3) ). From one side of this depression a fissure, whose upper and lower 
surfaces are approximately parallel, extends down at an angle of about 45° or 
more (plate 11, fig. b). The slope is in all cases very steep, though not always 
regular. In horizontal section the walls appear to form sections of a circle so that 
these caves all suggest fragments of hollow cones. At a depth of about 80 feet 


water is encountered in a crescent-shaped pool. The caves extend down for an 
indeterminable distance below the water-level. The surface of the water in the 
caves is near sea-level. Light penetrates to all the recesses of these caves, one of 
which is called Cueva dos Insurrectos from the fact that a company of Cubans 
was quartered in it during the Revolution. Figure B, plate ii, gives a glimpse 
down the Cave of the Insurrectos from the entrance X in figure 68 to the pool of 
water at the bottom, at a vertical distance of 83 feet. These caves are inhabited 
by Stygicola, but in very much fewer numbers than the well near the .seashore. 
One specimen was secured. 

A cave in the side of the hill at the edge of Matanzas shows essentially the same 
character. The slope is very much steeper and the cave is much smaller. There 
is the same sort of pool at the bottom as in the Cave of the Insurrectos. I secured 
no fishes in the Matanzas Cave, though it probably contains them. We were told 
that into this cave the Cubans, shot during the Revolution, were thrown by the 
guardians of Matanzas. 

On the southern slope of the island, both at Alacranes and westward about 
Canas, are formations very much like each other and very much like the condition 
represented in figure 68, with these exceptions: the territory is farther from the 
sea ; the pockets corroded in the surface rocks are much deeper and larger, and are 
filled with a stiff red clay. 

Bananas are grown in the pockets of soil about the caves at Alacranes. About 
Canas most of the territory is still in its primitive condition, covered with manigua, 
a straight-stemmed, smooth-barked, but irregular-surfaced, sapling that grows in 
such abundance mingled with other bushes and vines that it obscures the nature of 
the ground and makes progress through it impossible without the machete. 

Frequent clearings made to convert the manigua into charcoal and prepare the 
soil for seed tobacco reveals the nature of ground to be a series of jagged rocks with 
pits and depressions filled with the aforesaid red clay. The roads through this 
region are simply trails along which the manigua has been removed. The rocks 
are in the natural condition or worn a little by the two-wheeled vehicles which 
alone are usable here. The wheels of these are so large that they bridge most of 
the pits between rocks. Traveling over the roads in the manigua in one of the two- 
wheelers is quite a serious performance. Where the soil is a little thicker, tobacco, 
casava, and other things are grown. I do not know whether the formation is con- 
tinuous from Canas to Alacranes, but it seems quite certain that we have to deal 
with the same sort of structure in both places. It is a raised coral beach somewhat 
shattered and with a thin, in many cases interrupted, layer of soil. 

The entire southern slope of the area from Alacranes and Union to Canas is 
drained by underground streams which, for the most part, are inaccessible. The 
underground drainage begins further north than the northern edge of the manigua. 
At San Antonia de los Banos ' a stream is seen to enter the ground, and a few yards 
from this place, where the thin limestone roof of the underground channel has 
given way, the stream can be seen. (See frontispiece.) 

For reasons to be mentioned at once the streams are inaccessible. In August 
of 1904 a very heavy rain caused a small torrent to run in the road leading south 
from Canas for a distance of about a mile to the Finca Rosa, where the water 
spread out over a depression of several acres, so shallow that the depression was 

' The elevation of the railroad track is 62.92 m. 



Root breaking up into rootlets in Ashton Cave. Young of Lucifuga are 
found among these rootlets. 

Cave Isabella, showing group of roots coming through crack in roof. Taken 

vkdth artificial light. 


not perceptible to a casual observer. Mr. Francesco Martinez, who lives within a 
mile of the place and has been my guide about the caves of Caiias, informed me 
that the water would all disappear in a day, but that there was no distinct opening 
to any stream below the surface. Though I have been able to get to the ground- 
water in many caves about the neighborhood, none of the caves had any intimate 
connection with an underground stream, for, while the surface water was extremely 
muddy and abundant and all of it was carried oflf as rapidly as it would have been 
in a surface-drained area, the water in the caves to the south, in which direction the 
drainage flows, remained limpid and showed no appreciable rise. I was told, 
however, that during an unusual freshet in 1886 the entire region about Modesta 
Cave became flooded and, naturally, the cave was overflowing.' The underground 
streams come to the surface in a series of "ojos de agua." I visited two of these. 
One of them is in the Cienaga near the Playa of Guanimar. The water simply 
rises here in a pool 20 feet across in a swamp and is conducted in an artificial canal 
by the side of a road to the sea. I did not make extensive observations in this neigh- 
borhood, for the Cienaga has a great number of soft places with unknown depth, 
from which even the highway with a ditch on either side was not altogether free. 
One of the ditches showed the ground to Ije permeated with canals uj) to a foot in 
diameter. In this Cienaga many of the southward -flowing subterranean streams 
find their exit, doubtless others have a subaqueous exit in the ocean ; two others are 
found at Batabano on the coast just south of Havana. On the northern slope the 
most famous of the exits of the underground rivers is the Vento Springs, which 
supply the city of Havana with water. I have described these in Science (N. S. 
xviii, pp. 281-282, 1903). I should say that this spring does not yield half as 
much water as that at Guanimar. Underground streams and tunneled mountains 
are not rare in other parts of Cuba, though I have not connected them directly with 
the blind fishes. 

I was told a cave passes through a hill west of Matanzas, over which the United 
Havana Railroad runs. I was also told that at Cardenas, only lo or 15 miles from 
the Cave of the Insurrectos, there are underground streams with blind fishes, but 
this information reached me too late to make a personal inspection. 

The most famous of the underground streams and tunneled mountains in all of 
Cuba is the Sumidero which I visited. This region is half a day's travel by horse 
from Pinar del Rio. I found no blind fishes here, and it is extremely doubtful 
whether any occur in the main stream which twice pierces mountains in the course 
of a mile amid the most impressive cave scenery I have seen. 

In the blind-fish area drained by underground streams the surface water reaches 
the underground streams through sink-holes, fissures, and "caves." 

The sink holes are shallow and imperceptible. One at Finca Rosa, I have 
described above ; another is at Aguada on the United Havana Railroad, where, in 
extreme cases the water rises to stand several feet over the railroad track and then 
gradually disappears entirely.^ The difference in the nature of. the sink holes of 

' Mr. Martinez gave me the following facts: Rain unless protracted makes no impression on the water in 
the caves — as measured by visual standards. Aher a rain of 3 days and nights it rises 6 or 8 inches. In 1886, 
after a long rain of 5 days and nights the water in the well at Isabella rose to within 5 feet of the surface. Ordi- 
narily it is almut 50 feet from the surface. In the Modesta Cave in which the water is normally 15 feet from the 
surface the water rose to the top and over, till it stood i foot in the house of Modesta, and between the houses 
at Isabella and Modesta the water was in places 5 to 6 feel deep. The rain water does not run off in surface 
streams, but all of it sinks into the ground. At the time of the high water the water disappeared from the surface 
at Modesta in 2 days, while in the deeper places it did not disappear for 5 or 6 days. 

'The lowest part of the land at Aguada del Cura is 45.77 m. above the Nueva R. R. station in Havana. 
The railroad track is 3.82 m. higher. 



Cuba and of Indiana seems due to the difference in the thickness of the soil, which, 
as stated, is extremely thin in this part of Cuba. In the manigua frequent fissures 
or narrow wells lead down to the groundwater. 

There are, finally, the so-called "caves" which also lead down to groundwater. 

As stated above, the caves at Alacranes are of essentially the same character 
as those of the Carboneria. There are several of these. I have visited three, but 
obtained fishes from only two, the "M" and Donkey. 

Into the deeper parts of one of the caves visited, the Pedregales, light does not 
penetrate ; stalactites and stalagmites are clear, tinted rosy, and pure in tone when 
struck. The usual pool of water did not contain any fish at the time of our visit. 
An amusing incident occurred at this place. Our guide evidently thought our chief 
object was to view the marvels of cave formations. When we asked whether there 
were any caves in the neighborhood with fishes in them, he remarked, "Yes, but 
the fish don't amount to anything, they haven't any eyes." 

The "M" cave consists, first, of the slight depression in the general surface, 
and second, of the opening at one side of the depression leading down to the water. 
The slope is here gentle enough for a zigzag path in the shape of the letter " M " and 
enables cattle to get to the water at a vertical depth of 83 feet. Light penetrates this 
cave, and indeed the part directly down from the opening is well lighted. The pool 

Fig. 6Sa. Diagram of the Kentucky Cave Region, after Slialer. A. Sandstone and limestone stiowing ordinary topography. 
B, Sinic holes. C. Domes below large sink holes. D. Upper line of caverns first formed. £. Lower line of caverns. 
F. Cavern filled with stalactite. G. Lowest line of caverns filled with water. H, Masses of pebbles. 

of water leads off to the left, so that the remote part of the pool is in perpetual dark- 
ness. This condition makes this cave an ideal place to observe the reaction of the 
blind fishes to light. As in the Cave of the Insurrectos the caves extend down for 
an undetermined distance below the surface of the water and blind fish could fre- 
quently be observed here far below the reach of our lo-foot dip net. 

The Donkey Cave is similar to the "M" Cave, but the descent is steeper and 
there is a large shallow expanse of water on the left of the shaft of light from the 
opening. The depression at the mouth of the cave is here 9 feet below the general 
surface and the water is reached at 64 feet below the surface. Water was formerly 
pumped from this cave for purposes of irrigation. 

The caves about Canas differ from those of the Carboneria and Alacranes. 
They are cistern-shaped sink-holes rather than caves in the ordinary sense of the 
word, but on account of the absence of soil there are no funnel-like depressions on 
the surface to indicate their presence. There is absolutely no general surface indi- 
cation that one is in a cave country in traveling through it, and it is not until 
standing at the very brink of one that the presence of a "cave" may be suspected. 
All of the caves in the Canas region are modifications of the Modesta type. They 
are dome-shaped rooms (fig. 69) whose roofs are in different stages of dilapidation 
and collapse. They have a circular doughnut or crescent-shaped pool of water 



at the bottom. In most cases 
the roof is very thin ; that is, 
the dome is just beneath the 
surface, the room being high. 
More rarely the roof is thick 
and the cave correspondingly 
low. In one case the roof is 
intact and a narrow tunnel 
slopes down to the cave from 
the side. In several cases a 
vertical shaft leads down at 
the edge of the cave, in other 
cases a smaller or larger open- 
ing or openings occur near 
the middle of the dome, while 
not infrequently more than 
half of the roof has fallen, 
forming a slope down one 
side, while at the opposite 
side the overhanging walls 
still stand (fag. 70). The 
latter is the Ashton type 
found in several of the caves 
on the Finca Ashton. In all ^'o- *«• d^k"™ °f Modrau caw. 

the caves visited there was a pool of water. (There are said to be dry caves, but 
we had no time to visit them.) In one case the pool forms a simple sheet of water ; 

Fiu. 70. Diagram of Ashton. Hypothetical Outline of Cave before Fall of Right Part of Roof is indicated by Dotted Lines. 



very frequently there is an island in the water beneath the opening in the roof, 
and in the Ashton type the water has become restricted to a crescent at the base of 
the wall still standing. It is possible that the Carboneria and Alacranes caves 
belong to the latter type of caves. 

Almost invariably one or more trees (Ficus) stand over the cave and send long 
roots down through the cave to the water below, where they break up into number- 
less rootlets (plate 12, fig. a). The roots were very useful in gaining access to the 
bottom of some of the caves. During my earlier trips, access was gained to most 
of the dome-shaped caves by climbing down the roots or a bamboo pole. In the 
later trips the roots were still the most effective ladders to some of the caves, but I 
substituted a portable rope ladder for the slippery bamboo pole. 

Finca Isabella 
a. Well 


o ^ 

I mile 





Modes ta 

Modesta, I. 



Frias X 

^ Well® nnca\Frlas 


Fig. 71. Partial North and South Section through Cave Region about Canap, Cuba. Entire area has subterranean drainage. 
Road from Caiias becomes a stream in heavy rains and sinks within the area inclosed in circle. Caves marked with small 
circles were located by traverse readings, those marked x were located by guess. There are caves south of area mapped, 
but land slopes to ocean so that water is found very near surface. There are many others in area covered that are not 
indicated on this map. 

The density of the caves may be gathered from the accompanying sketch of a sec- 
tion extending south from the station Canas on the Western Railroad but not quite 
to the southern edge of the cave region (fig. 71). The caves marked with a cipher 
(o) were located by traverse readings by Dr. Joshua William Beede, of the Geo- 
logical Department of Indiana University, who volunteered his services on one of the 
trips. The caves marked with a star (x) were "discovered" on a subsequent visit 
and located by estimate. Numerous "wells" and other caves are not indicated, 
but from the number located an idea of the abundance of the caves can be formed. 
They are about as numerous as sink holes in the cave regions of Indiana and 
Kentucky. There are caves south of the area mapped, but the land slopes to the 
ocean 15 miles to the south, so that water is found very near the surface. 



An attempt was made to determine the relation of the water in the various caves 
to a general level of groundwater and to ocean-level. An aneroid barometer was 
used for this purpose, but although it was of latest pattern and its vernier read to 
I foot, I am afraid that the readings arc approximations only, because allowance 
for barometric changes could not readily be made. 

Barometer readings along the line of the Western Railroad compared with the 
elevations determined by the engineers of the line may give us an approximation to 
the dependence that may be placed on the respective readings. 


Barometer readings. 



Barometer readings. 



Aug. 22, 


Aug. 26, 


Aug. 22, 

Aug. 26, 





Arroyo Narranjo. . 


Rancho Boyeros . . 
Santiago de las 

















181. 2 






House of Finca 













' Accepted engineer's determination. 

The engineer of the United Havana railroads furnished the following eleva- 
tions of stations in the cave region. The elevations given are above the Villa 
Nueva station at Havana, not above sea-level. As the line crosses the Western 
Railroad at Rincon and the elevation of its rails above sea-level at Rincon is 
252.13 feet, I estimate Villa Nueva to be 23 feet above sea-level. 


Elevation above 
Villa Nueva. 


Elevation above 
Villa Nueva. 









Rio Almendares water-level 

Palenque station 

Rio Almendares water-level to the face 

of the superior rail . . 

Vento station 

Guines station 

Rio Seco station 

San Nicolas station . . ... 

Aguada del Cura station 

Palos station 

Depression of land at Aguada del Cura, 

Bermeja station 

Height above this point to the face of 

Rincon station 

Goven station 

Buenaventura station 

Duivican station 

Saladriyas station 

Duran station 

At Canas there is a well in the yard of a store about 100 yards from the railroad 
station. On August 26, 1903, the surface of the water in this well stood very near 
sea-level, i. e., exactly 100 feet below the surface of the ground. 

Mr. A. P. Livesey, general manager of the Western Railroad, kindly furnished 
me with the depths of 3 wells.' Tabulating these and the depths obtained in the 

'He wrote: " Regarding the depths of wells along our line, I may say that these vary very considerably, not 
only in the different localities, but also during the two seasons, viz., wet and dry, but for your information and 
guidance I give below the average depths of 3 of our company's wells, which are used to obtain water for our 
locomotives. They are as follows: Salud, 100 feet; Guira, 50 feet; Artemisa, 80 feet," 



various caves, together with the elevations of the mouths of the caves, we get the 
following results: 


of station. 

Elevation of suriace 
of groundwater 
above sea-level. 


of station. 

Elevation of surface 
of groundwater 
above sea-level. 





Isabella Finca caves : 

1. Modesta 

2. Miserid 

3. Hawey (new) . . 

4. Hawey No. i . . 













Isabella Finca caves : 

5. Open pool at 


6. Isabella 

7. Drago 

8. Frias 

9. Ashton No. i . . . 

10. Ashton (new) .. . 

11. Banos 

12. San Pedro 






The elevations of the water of the caves together with the well at the house 
Isabella fall into two groups : first, those from numbers i to 8 in which the elevation 
of the water does not vary more than 11 feet. This amount may easily be due to 
change in barometric pressure during the various readings and to the personal equa- 
tion. It seems probable that the water in these caves, most of which are south of 
Finca Isabella, is at a level, and that this level is between 8 and 19 feet above sea- 
level. The Finca Isabella is about 15 miles north from the coast, or 10 miles from 
the Cienaga, in which some of the underground rivers rise to the surface. 

The second group, from 9 to 12, are east of Isabella; 9 and 11 are near each 
other; 12 is 2 miles or more east of 9 and 11, and I am not certain about the loca- 
tion of 10. These readings were taken August 25, 1903, in the order: 10, 12, 11, 9. 
The first reading at the house was at 6*' 30"" a. m., when the barometer stood at 
1,114 feet. The trip consumed all of the morning. About 3 p. m. the barometer 
stood at 1,179 at the house, so it is very probable that the high elevations may in 
part be due to the change in barometric pressure. 

For comparison we have the data for the caves, Adolfino and Insurrectos, at the 
Carboneria, near the north coast. 

The Cave of the Insurrectos is about 93 feet above sea-level according to barom- 
eter. The water is 83 feet below the surface and according to that 10 feet above sea- 

The top of Adolfino is 87 feet above sea-level, the water is 80 feet below, or 7 
feet above sea-level. The surface of the Carboneria well is about 4 feet above 
sea-level, the surface of the water is at sea-level and more than 5 feet deep. 

There is every indication that the water has risen about 10 feet in the caves 
in very recent geologic times. In all the caves stalagmites are seen to rise out of 
the water, in some cases from a depth of at least 10 feet. As these could only have 
been formed on ground free from water it is evident that the water must have risen 
in the caves. As the water is now near sea-level, this rise is probably due to the 
subsidence of the western end of the Island of Cuba. This subsidence is general, as 
stalagmites are found submerged on the northern and southern sides of the island. 

' The water in the Donkey and " M " caves, according to barometric readings from the railroad station at Union, 
is below sea-level. It is not at all probable that this reading is correct, but it indicates that the groundwater 
level is here again very near sea-level. At " M," according to barometric reading, it is 83 feet below the gen- 
eral level of the surface and at the Donkey it is 73. 





A. Drawing of black individual of Stygicola. 

B, C, D. Stygicola. Pfiotograpfis of preserved specimens. 




The number of fishes in any cave differs very greatly. They are rare in caves 
entirely inclosed ; in those entirely open and not connected with hidden recesses 
they are also very rare or absent. They are most abundant in caves with both well- 
lighted and dark portions and those that are continued subterraneously. The den- 
sity of tic distribution of the fishes evidently varies greatly, directly with the food 
supply. The food supply itself varies with the openness of the cave to the external 
world. The question arises whether the caves visited are independent pockets or 
form part of a continuous underground system of channels, and whether the fauna 
of the caves visited may be easily exhausted or continuously replenished from the 
extensive subterranean channels and reservoirs. Collections made in the same 
caves indicate that there is an undoubted decrease in the numbers and that the 
decrease is not usually compensated by immigration from the underground reser- 
voirs. It has rarely proved worth while to visit the same cave twice on any of the 
stays in the cave region. The results of three visits to the "M" and Donkey 
Caves on October 25, November 2, and December 23 illustrate the point. In 1904, 
I secured 15 fishes in the Donkey Cave on October 25 ; 5 on November 2 ; and 3 
on December 23. In the "M" Cave I secured 20 in March, 1902 ; 19 on October 
25 ; 14 on November 2 ; and 9 on December 23, 1904. Equal efforts were made 
on each occasion and an equal amount of time was given to the caves. 

On June 24, 1905, Mr. Haseman secured 4 fishes in the Donkey Cave and 7 in 
the "M." The Donkey thus yielded 15, 5, 3, and 4 fishes respectively, on succes- 
sive visits; the "M," 20, 17, 14, 7, 7. Both of these caves are with deep recesses in 
which fishes could be seen but not secured. 


Without doubt the remote ancestors of the Cuban blind fishes lived in the ocean 
and were adjusted to live in the light and to make use of it in detecting their food, 
their enemies, and their mates. Equally without doubt, their less remote ancestors 
became adjusted to do without light and lived in total darkness, either at a depth in 
the ocean or more probably in the crevices in Cuban coral reefs. If in the former, 
they entered the subaqueous exits of Cuban rivers ; if in the latter, they are older 
than the rivers themselves, having remained in their original habitat in the crev- 
ices of the coral reefs as these were elevated to their present and even greater 

The latter seems to me the more plausible theory. The fresh-water blind fishes 
of Cuba are as old as the parts of Cuba they inhabit. They are part of the result 
of the formation of the island. 

The deeper recesses of the crevices and rents in the naked reef at the Carboneria 
already described are probably now inhabited by fishes of some sort, possibly by 
Ogilvia among others. Attention has been called to the fact that within less than 
a quarter of a mile from them, in a coral reef raised only 4 feet above the ocean-level, 
there is a rift essentially like those found in the naked reef skirting the ocean. This 
rift contains fresh water, and blind fishes are abundant at a place where a circular 
opening has been cut to form a well. 

It is entirely within the range of probability that the ancestors of these fishes 
lived in this rift when it was 5 feet lower and contained salt water and that they 


gradually, as the reef was raised, became adapted to fresh-water conditions. But 
if this rift with its well contain descendants of its original marine inhabitants, there 
is no reason why the same should not be true of the wells and caves and rifts of the 
more elevated coral reefs of Cuba. In other words, there is no reason why the 
blind fishes should not have developed over the entire area and risen with the entire 
area over which they are now known to be distributed. Stygicola is found from 
Canas at least as far east as Jovellanos ; Lucifuga only west of Guira and at least 
as far as Canas. There is, furthermore, no special reason why the blind fishes 
which have been reported from the natural wells at Merida in Mexico and from 
Jamaica should not be identical or related to the Cuban species, why they should not 
have been independently derived in different places from one or more species 
widely distributed in cracks and crevices of coral reefs. 



Whatever conditions may have been in the past, at the present light is entirely 
absent from some of the places inhabited by the blind fishes while others are as well 
lighted as any stream. In the only cave I entered which light does not penetrate, 
the pools of water, in every respect similar to those in other caves, contained no blind 
fishes. On the other hand, in Ashton Cave, parts of which are as well lighted as 
any stream, blind fishes live side by side with eyed fishes. In a few of the best- 
lighted caves no blind fishes were found, but in Modesta, where an ii-foot opening 
in the ceiling lights a space 35 by 45 feet so that pebbles and fishes can be seen vnth 
perfect distinctness at a depth of water of 10 feet and more, bhnd fishes are abundant. 
The same is true of similar caves, well or partially lighted. 

Blind fishes were abundant in Tranquilidad, a dark cave into which light pene- 
trates through a narrow shaft over 20 feet deep and then only illuminates the margin. 
They were also abundant in the open well at the Carboneria, about 6 feet in 
diameter and with a total depth of about 10 feet. 

It is to be emphasized that blind fishes are abundant in well-lighted caves only 
when these are connected with underground channels that extend into the dark. 
Such caves contain many more fishes than caves that are totally dark. The reason 
for this lies entirely in the much greater abundance of the food supply in caves 
open to the surface; the lighting of the cave is incidental. 

The reaction of the blind fishes to light can be as well studied in the " M " Cave 
as in any aquarium ideally constructed for the experiment. The pool of water in this 
cave varies from 5 to about 20 feet across, and from a few inches in depth to many 
feet — certainly over 10 feet and possibly 50. The pool is probably between 150 
and 300 feet long. A direct shaft of light reaches the pool near one end so that 
the water is well illuminated within this shaft. The right end, near which the shaft 
of light reaches, is shaded by rocks and is so dark that a lamp is of distinct assistance 
in exploring its 2 to 3 feet of depth. The other end of the cave is in total and per- 
petual darkness. Fishes are abundant in this cave. I have seen very few within 
the shaft of light and most of those were driven there by my movements. In the 
shade of the rocks to the right, on the contrary, they are abundant, and in the larger 
dark parts of the cave to the left they are also abundant though relatively less so 
than on the right. Here we have a very distinct reaction to the light — all the fishes 


avoiding it. Cattle come down to drink in this cave within the shaft of light. The 
indirect result of this is a great abundance of blind-fish food. In the movements 
and distribution of the fishes in this cave we have a clear balance struck between 
the positive attraction to the food and the negative response to the light. 

The same reactions demonstrating perception and tropic relations to light 
are seen in the Donkey Cave near by. In this cave I have never seen a blind 
fish within the shaft of light, but have seen and caught them in numbers in the 
expanse of shallow water in the shadow and total darkness to the left of the shaft 
of light. While fishing in Ashton in December, 1903, I caught 3 specimens in the 
lighted part of the cave and about a dozen in the dark recesses to the right of the 

Unfortunately, on account of the difficulty of getting about over the jagged 
country, I have been able to visit but few caves at night, but the observations in 
the Carboneria well were exceedingly instructive. 

A few bushes growing over the well shade it to a certain degree. As stated 
elsewhere, poles and fence rails were placed slanting into the water crossing each 
other and in sufl&cient number to form a teetering foothold that enabled me to stand 
waist deep in water. From this position every part of the well was within reach of 
my net, except pockets in the sides too small for the net and the indefinitely extend- 
ing side rifts I have mentioned. 

On visiting the well about 9 a. m. perhaps as many as 10 stygicolas were seen 
swimming about or resting on the wood or sides of the well. I entered the well 
but succeeded in catching only one fish ; the others readily escaped either by making 
for the dark side rifts or by hugging the walls of the well and entering the small 
pockets where I could not get them. There seemed to be no hesitation in their 
actions. I again entered this well the same night. Liberally discounting the 
result for the experience already gained in entering the well and knowledge of the 
location, the result alone is evidence of a distinct difference in the actions of 
the fishes at night and in the day — I caught twelve. 

Their actions were quite different. While in the daytime they seemed able to 
locate the dark recesses and make for them with precision, their action at night 
gave distinct evidence of confusion and lack of ability to readily escape. They 
could be easily followed with the pencil of light from the lamp and picked up with 
the net. 


The fluctuations in the air temperature of caves with small openings are, in a 
climate like that of Indiana, reduced to a few degrees Fahrenheit, and must be re- 
duced to a minimum in a climate like that of Cuba. The temperature of the water 
will also fluctuate but little. The air of caves that are open like that of Ashton 
will, on the contrary, fluctuate to nearly the same extent as that of the epigean 
neighborhood. The nights of the Cuban winter are cool and the temperature of 
the water in the open pools of these caves may be reduced a few degrees. No direct 
observations are at hand on this point. 

The temperature of the water in 18 caves containing fishes, taken in June, 1905, 
showed a total range from 74° to 76.5° Fahrenheit. Only 2 caves had a temperature 
as low as 74°; 3 of 75°; 5 over 75.5°; 6 of 76°; 2 of 76.5°. 


Observations between August 22 and 25 showed slightly higher temperature for 
open caves, thus: in June, 1905, the temperature at Banos was 75.8°, at Ashton, 
75.6°, and on August 25, 1904, it was 77° at Banos and the same at Ashton. In 
the "M" Cave, a closed one, the temperature was the same, 75° Fahrenheit. 

The blind fishes are adjusted to withstand slight fluctuations in temperature. 
Some were kept in aquaria and the water became distinctly chilled over night and 
warmed during the day. While they lived for several days in these aquaria, they 
were always sluggish or numb in the morning. A more distinct reaction of the 
same sort was noticed in the only fish I succeeded in bringing home alive. It could 
scarcely move after an early September night in Indiana. A still greater reaction 
was noticed in several I succeeded in bringing alive to Louisville and which suc- 
cumbed to the frosty weather on the way from the depot to the hotel. 


In all caves in which collections were made the water is clear as crystal. It 
will easily rank with the water of Lake Tahoe and of the limestone springs of Florida, 
as among the most transparent natural water in the world. Fishes can readily 
be seen at depths of 15 and 20 feet or more, with the aid of a bicycle acetylene lamp. 
The water at the Vento Spring is of the same nature, but I was informed that it 
becomes slightly roiled after heavy rains. 



The water is everywhere highly charged with salts of lime and magnesium 
In all cases where the surface of the water is not disturbed by breezes, a crust of 
these salts forms like a thin ice over the surface of the water. When one disturbs 
the water, the crust breaks up into small fragments which fall through the water 
like snow through the air. Occasionally a larger flake, a foot square, may fall to 
the bottom; sooner or later they are dissolved again. With falling of the level 
of the water some of the crust is left on shore and gives an index of the amount 
of rise and fall in the water during a year. 


The ordinary fluctuation in the amount of water in the caves is very small ^ 
about one foot during a year — judging by the flakes of lime left on the banks. 
I have mentioned elsewhere that, after long-continued rains, water flooded the 
entire region about Modesta, the cave was full to the top, and the water stood 
several feet over the ground. All of this retreated in a few days. Such fluctua- 
tions are very rare. 


Concerning the size of the environment little can be said. The pools accessible 
are easily measured, none of them exceeding a few square meters in surface, 
but the size of the underground connections is naturally unknown. The rapid 
disappearance of the water after heavy rains indicates extensive underground 



Living Stygicolas. 
Position of body and fins in swimming and differences in color of different individuals. 







P 03 



s^ ? i 


0) fi) 4-1 


03 =- o 

3 '^ OQ 



• 1-1 e 


™l ^ 

^o Q- 








n o^ 








« 3 



<= 3- 


3 "< 


C^ ^ 

W = 





S- ^ 


8 ? 














During March small frogs are found abundantly at the margins of the pools 
in some of the caves. I do not know that these affect the lives of the blind fishes 
in any way. Tadpoles were found in the Carboneria well. It is possible that 
these may form some part of the food of the fishes during some seasons of the 
year. They are but casual associates of the blind fishes in some of the caves. 

Fishes other than the blind ones were found in Ashton and some of the small 
open caves about Modesta. They were all Girardinus metallicus Poey, a species 
very abundant all the way to Pinar del Rio. The female reaches a maximum 
length of 79 mm., but is usually much smaller; the maximum length of the male 
is 45 mm. The largest specimens taken in Ashton are 41 mm. and 38 mm. These 
fishes are active swimmers, living near the banks, and while a few may be cap- 
tured by the blind fishes, they are themselves too small to attack even the young 
of the blind fish. 


The blind fishes are carnivorous, securing living prey. Their food consists 
largely of 4 species of crustaceans, 3 of which are blind cave forms. Probably 
every living animal of the proper size is used by the blind fishes for food. 

Cirolana cubensis Hay. 

This species was described by Hay in Proceedings of the National Museum, 
VI, page 430, as follows: 

Body oval, a little more than twice as long as broad, widest a little behind the middle, rather 
strongly convex, and perfectly smooth. Head a little broader than long, slightly produced in front. 
Mesosome broader, with its greatest width at the fifth segment; coxal plates of the second, third, 
fourth, fifth, and sixth segments successively more enlarged and more strongly produced backward 
at an acute angle. The plate of the seventh segment is about the same size as the one preceding it. 
Metasome narrower than mesosome, of five segments, each of which, except the last, has the lateral 
angles strongly produced posteriorly; telson as long as the metasome, its margins gently curved 
and convergent for about two-thirds of its length, and then rather abruptly strongly convergent to 
form a short, obtuse tip. The eyes are altogether wanting. First antenna with three basal seg- 
ments and a short flagellum which, when extended backward, reaches slightly beyond the posterior 
margin of the first thoracic segment. Second antenna with five basal segments, and a long, slender 
flagellum which may extend slightly beyond the middle of the body and is composed of about 
twenty-nine segments. The mandible, maxillae, and maxillipeds do not present specific characters 
of importance, being of the type usual in the genus. The appendages of the mesosome are of mod- 
erate strength, and are armed with a few rather stout spines and stiff setje. The branchial append- 
ages of the metasome are membranaceous and small ; the uropoda are well developed, the outer branch 
lanceolate in outline, the inner much broader and very slightly longer, and with the tip somewhat 
accuminate; both branches and the margins of the telson as well bear a rather dense fringe of hairs. 
Color in alcohol, white, with no markings of any kind. Length, 5 mm. 

Of the species of Cirolana known to inhabit American waters, C. mayana, which occurs on the 
coast of Yucatan and Colombia, is the nearest relative of the present species. Between the two, 
however, there are several important structural differences. The physiological differences between 
this species and all the others of the genus must be very great to admit of its living in the subterranean 
streams of fresh water. It may be added that Cirolana cubensis is very distinct from Cirolanides 
texensis Benedict, which occurs in the waters which flow from the large artesian well at San Marcos, 


This species is everywhere abundant and may attack the fishes if it succeed 
in attaching itself to them. I have not caught any fish with them attached, but 
in small aquaria in which many of them were placed as food for the fishes they 
soon turned the tables and fastened themselves upon the fishes. In some of the 
caves cirolanas exist in vast numbers. At the base of the shaft of Tranquilidad 
they were so numerous and voracious that it was impossible to stand in water 
long enough to light our lamp. They fastened themselves in numbers on the feet 
and went to work with such a will that it was impossible to stand still. 

Palsemonetes eigenmanni Hay. 

This extremely slender and graceful shrimp is abundant in all the caves. It is 
essentially pelagic in habit, though it is frequently seen resting on various objects 
on the bottom. Its eyes have been described by Pike. The species was described 
as follows in the Proceedings of the U. S. National Museum: 

Carapace thin, very delicate and transparent, in form slightly compressed near the middle of 
the body but rather broad anteriorly ; the anterior border, below the eye, is produced as a broad, 
obtuse angle, which bears, near its lower margin, an acute, forwardly directed spine ; this spine is 
the anterior end of an obscurely marked ridge, which extends obliquely downward and backward 
along the sides of the carapace. The rostrum is long, slender, compressed, and rather markedly 
upcurved; on its superior margin it bears a row of 6 or 8 slender, acute teeth, which begins well 
back on the carapace and extends forward to the rostrum ; these teeth are directed obliquely for- 
ward ; the inferior margin is unarmed ; the tip of the rostrum is acute and reaches forward to a point 
opposite the distal extremities of the antennal scales. The eyes are much reduced in size, are with- 
out pigment, and the corneal surface comes to an obtuse point in front. The first antenna has the 
basal segment well excavated above and provided with a small, acute spine at the outer distal angle ; 
there are two long and one short flagella, the short one slightly exceeding the rostrum, the long ones 
somewhat longer than the body. The second antenna has the basal segment provided with a small 
spine near the distal end ; the antennal scale is broad and with subparallel margins ; the tip is slightly 
rounded, and there is a small, obtuse spine at the outer distal angle ; the flagellum is slender and 
about twice as long as the body. The mandible has an incisor portion with three or four sharp teeth, 
a small molar surface with several obtuse teeth, but is without a palpus. The third maxilliped is 
not strongly developed and presents no characters of importance. The first pair of pereiopods is 
chelate, and except for its much smaller size is exactly like the second ; the chela is slender and weak ; 
the carpal segment is long and slender; the meros is of about the same length, but stouter; the 
remaining segments short and rather thick. The remaining pereiopods are very long and slender. 
The abdomen is of the form usual in this genus, but the sixth segment is neither elongate nor com- 
pressed ; the telson narrows gradually from the base to the obtusely angulate tip ; on the upper 
surface there is on each side at about the middle and again about one-fourth the distance from the 
tip a small, appressed spine; at the tip there is on each side one minute and one long, slender 
spine, and in the middle a fringe of setae. Color in alcohol, white. Length, 23 mm. 

They differ very markedly from Palamonetes antrorum Benedict, hitherto our only known blind 
Palmmoneles, in the shape of the rostrum and the character of the chelse. The shape of the eye is 
rather remarkable, even in a group, where through atrophy the eye tends toward the conical form. 
I know of no other in which it is produced into a blunt point. So far as I have been able to ascer- 
tain, this is the first record for this genus in Cuba. In the material from San Isidro there is one 
specimen which agrees in every way with the types, but the other two differ in such a manner as to 
lead me to believe that a second species may be found to inhabit the subterranean waters of Cuba. 
The two specimens just mentioned have the sixth segment of the abdomen 2.5 times as long as deep, 
and the antennal scale is more slender and acute. Unfortunately, the rostrum of one is entirely 
gone, while of the other only the abdomen remains. 


Epilobocera cubensis Stimpson. 

This crab, which reaches a width of several inches, was observed in many of the 
caves. It is probably found in all of them though not in great abundance. If the 
adult affects the blind fishes at all, it is to feed on them. I have found the young 
of this species in the stomach of Stygicola. 

Gammarus csecus Weckel. 

The following technical description will be found in Proc. U. S. Nat. Mus., 
XXXII, page 47. 

Eyes absent. First antennae more than half as long as the body ; second segment of the pe- 
duncle slightly longer than the first and about three times as long as the third; flagellum composed 
of twenty to thirty elongated segments, each bearing a few short hairs at the distal end ; secondary 
flagellum reaching slightly beyond the third segment of the primary flagellum, composed of four 
segments, the distal one short and furnished with long hairs. Second antennae are about two-thirds 
as long as the first pair with the peduncle extending far beyond that of the first pair; ultimate seg- 
ment of the peduncle only slightly longer than penultimate which is greatly elongated and about equal 
in length to the antepenultimate ; flagellum composed of about twelve segments, which are shorter 
than those of the first antennae and furnished with more hairs. 

The carpus of the first gnathopods of the male is triangular and elongated, with the anterior 
margin furnished with a few long hairs and numerous short ones; propodus narrower than the car- 
pus, twice as long as broad, with the anterior margin concave, armed sometimes with a fascicle 
of hairs, the posterior margin convex, and the palm almost straight, slightly convex, and armed with 
four or five spines and a few short hairs ; dactyl as long as the palm and fitting it closely. Second 
gnathopods with a carpus broader than in the first pair but similarly armed ; propodus not so broad 
as the carpus, about twice as long as broad and larger than in the first gnathopods; posterior mar- 
gin almost straight; anterior margin slightly convex and usually furnished with one or two fascicles 
of hairs ; palm very oblique, slightly concave at the center, armed with five or six spines at the tip 
of the closed dactyl, and one or two spines and a few short hairs on the margin; dactyl strongly 
curved, as long as the palm. 

Both margins of the coxal plates of the third, fourth, and fifth peraeopods are serrate and fur- 
nished with spines, those on the anterior margin being smaller than those on the posterior. Postero- 
lateral angles of the third and fourth abdominal segments are produced backward and end in a blunt 
tooth. The last two or three abdominal segments are furnished dorsally with a few short spines. 
The first uropods project slightly beyond the second pair. In both pairs the rami are about equal 
in length and slightly longer than the peduncle. The third uropods were broken ofif in the few 
specimens which I had for examination. Telson cleft to the base, armed distally with a few short 

I found this blind amphipod in Modesta in the roots of trees. It was not 
abundant and was not observed in any of the other caves in which no special 
search was made for it. It was hidden among the rootlets of Ficus in a way in 
which it would not be noticed unless special care was taken to look for it. It is 
quite probable that it may be found in many of the caves. 

In addition to the above mentioned species dragon-fly larvae were found in the 
stomachs of some of the fishes. 


In parts of Ashton a green alga forms a dense mass over many square feet of 
bottom. Young lucifugas are abundant in the alga, but this is the only instance 
of its occurrence in association with blind fishes and it scarcely deserves considera- 
tion as part of their normal environment. 

The only plant worth considering as forming part of the biological environ- 
ment of the blind fishes is the tree sending roots to the water. The roots break 


up into innumerable rootlets harboring numberless cirolanas and many young 
and small lucifugas. These trees are found in all the caves of the Canas region. 
The roots sometimes extend vertically as much as 40 feet before striking water. 
At other times roots run along the ground down the slope of the cave as in Ashton, 
finally breaking up into rootlets (fig. 70 and plate 12). 


The position in the water and action of body and fins in swimming of Stygicola 
are amply indicated in plate 14, which is from instantaneous exposures on fishes 
confined in a 5-gallon aquarium. It is seen that the posterior part of the body 
moves from side to side, eel-fashion. The long dorsal and anal fins move in the 
same way, waves of motion passing from in front back. These fins, on account 
of this motion, are not well shown in the photographs. The pectorals move in- 
dependently of each other. One may be forward, the other back. They are used 
in guiding largely. When the fish is swimming very slowly, the wave-move- 
ments passing along the dorsal and anal fins are the chief means of locomotion. 
In swimming rapidly the motion of the body comes chiefly into play. The fishes 
svwm indifferently up or down, with the back up or lying on their sides. The ac- 
tions of Lucifuga are essentially like those of Stygicola. 

These fishes are much more readily disturbed than Amblyopsis of the Indiana 
caves, and when disturbed they swim swiftly in a less distracted way. On the 
whole they are much harder to catch than the Amblyopsis. 

The action of the stygicolas in the Carboneria well in daytime and at night 
has been detailed. Two instances that seem to indicate that fishes "remember" 
localities must be put on record for what they are worth. One of these is of a 
fish at the right end of the "M" Cave, and the other in the left, dark part of the 
Donkey. In the "M" Cave the same fish, three times within an hour and a half, 
apparently made straight for an opening under the wall of the cave and escaped. 
In the Donkey Cave the same thing happened about a big stalagmite that rises 
out of the water. Several times within half an hour the fish came out, but each 
time it darted back among the nooks in the stalagmites with apparently as much 
decision as a mouse in seeking its hole. Perhaps in both cases the action was a 
reaction merely to the vibrations set up by my net. Perhaps the location of the 
solid stalagmite and the wall were perceived by the approaching fish and the escapes 
into nooks below the wall were simply necessary sequences in following along the 
solid wall until an opening was reached. Whatever it was, the repeated escape of 
the two fishes was as interesting as it was aggravating. Very frequently when dis- 
turbed they descend in the water and escape into depths beyond the reach of 
the net. 

The character of food has been detailed under the head of Biological Environ- 
ment. I am unable to give any direct observations on the securing of this food. 


In March of 1902, on my first trip, Mr. Riddle secured a female lucifuga con- 
taining 4 young, lacking but 3 or 4 mm. of being as long as the smallest lucifugas 
caught in the caves (plate 15, fig. c). This was the first intimation we had that 
these fishes are viviparous. No other embryos were obtained at that time. An 
examination of the ovaries of all the females caught and the size of the young led 
me to suppose that March was the close of the breeding season. With the grant 


from the Carnegie Institution I expected to remain in Cuba during the entire 
breeding season to secure a full series of embryos and to rear young in the light. 
Unfortunately for this plan the fish seem to have no general breeding season, and 
the appropriation was exhausted in determining that fact. I visited Cuba late in 
October, which was supposed to be the beginning of the breeding season if March 
was the end, but there was no indication that this time was near the breeding 
season. I had collections made early in December and again visited Cuba late 
in that month. But while, as before, there were indications that some individuals 
were ready to breed, there was no indication of the approach of a general breed- 
ing season. I next had collections made the first week in May without results. 
I revisited the caves late in August and early in September and finally, near the 
end of June, sent two of my students, Mr. John Haseman and Norman Mclndoo, 
to the caves. The former had accompanied me on one of the trips, and both were 
in every way thoroughly competent to get everything possible. 

To summarize: The caves were examined by myself and Mr. Riddle early 
and late in March, 1902 ; by Mr. Martinez early in May, 1903 ; by Mr. Haseman 
and Mr. Mclndoo late in June, 1905 ; by myself and Mr. Hankinson late in August, 
1904; by myself and Dr. Beede late in October, 1904; by Mr. Martinez in Decem- 
ber I, 1903; and by myself and Mr. Haseman late in December, 1903. 

The net results of these numerous trips for Lucifuga are: Late in March I 
secured one female with young about 20 mm. long, or nearly ready to be born; 
the ovaries in most of the other females were minute, the largest eggs measuring 
356 fi ; in two ovaries there were eggs 560 ft and 850 /a in diameter, both of these 
containing spermatozoa. Late in June a female with 15 young, 12 mm. long, was 
obtained ; the ovaries of the remaining fishes were small. On August 23 a female 
with 10 nearly grown young was obtained. The ovaries of all the others were min- 
ute. Late in October and December the ovaries of all females secured were minute. 

The young from the female in March were at least 3 months old. This would 
give a breeding period whose outside limits would extend from December to the 
end of August. The examination of numerous ovaries does not indicate a general 
breeding season, though a larger per cent contained large eggs in March than in 
other seasons. The best season to get material is probably March to May. 

The net results for Stygicola are : 

In March the ovaries of Stygicola are mostly small, with eggs not exceeding 200 
/i. One female taken at this time contained eggs 600 to 700 /* in diameter and 
her ovary was abundantly supplied with spermatozoa. In May no mating females 
were secured. In June the ovaries were mostly minute. Two of those secured 
contained turgid ovaries in which the structures were distinctly lobulated. 

On September i, I obtained a female with one young from the Carboneria. 
Other females had large ovaries, probably recently freed from young. Most females 
had small ovaries. One contained large eggs. The rest contained small eggs. 

On October 30, I obtained a Stygicola from Alacranes containing two young. 
The mother was 92 mm. long and her ovary contained eggs 880 /* long, which 
were evidently mature} At the same time I obtained 47 other females from 77 to 

' In an ovary containing spermatozoa in abundance, days if not months before the ripening of the eggs, an 
occasional early ripening should naturally result in the development of the embryo. The present case is probably 
one of this sort. Two eggs evidently started to develop long before the others were mature. The ripening of the 
eggs at different times may lead to different sized larvae in the same ovary unless the earlier larvx digest the sper- 
matozoa present before the other eggs become ripe. 


115 mm. from Alacranes and Canas, in all of which the ovary was empty and in 
most cases at its minimum. 

In December all the ovaries but two were minute. In one ovary a single large 
egg 720 /I was found, in the other the ovary was large and the eggs reached a 
maximum of 640 /a. Thus, nearly mature eggs were found in December and 
March, and young in September and October. 

If the species breed annually and irregularly throughout the year and the 
young are carried but 3 months, at least one-fourth of all the females caught at any 
season of the year should be with young. If the young are carried but 2 months, 
one-sixth of all the females should be with young. If the species breed at some 
definite season of the year and this period is not more than 3 months long, all of 
the females should be with young near the middle of the breeding season. 

The results are wide of any of these marks ; and the only conclusion possible 
is that either there is no definite breeding season, but individuals breed at any 
time during the year, or the fishes breed only at longer intervals than a year, and 
in either case while breeding they migrate to undetermined regions. That these 
regions are not far away is shown by the fact that occasionally breeding females 
reach the upper accessible parts of the cave. Between breeding times they are 
found in the upper, readily accessible parts of the cave. 

I found that while Amblyopsis probably breeds throughout the year a larger 
per cent breed in March than in other seasons. A similar condition may exist in 
the Cuban Wind fishes. 


The minute structure of the ovary of Lticifuga is elsewhere described. The 
ovary consists of a pair of delicate walled sacks united behind and with the ovif- 
erous tissues attached along the middle of its dorsal and ventral wall except for 
a short distance behind. It is placed in the mesentery between the dorsal wall 
of the body cavity and the rectum and stomach. In enlarged ovaries the oviferous 
tissue is seen to be lobulated, the lobules being attached anteriorly and free pos- 
teriorly. These lobules are arranged like shingles, the anterior ones overlapping 
the posterior ones. When the ovaries contain no larvae or ripe eggs, they extend 
far forward, the posterior oviferous tissues reaching but little behind the stomach. 
When eggs mature, the ovary becomes turgid and the oviduct apparently shortens, 
so that the posterior part of the stomach comes to lie in the fork near the anterior 
end of the ovary. 

The spermatozoa are evidently, as in Cymatogaster, which is another vivi- 
parous fish, transferred to the female long before the eggs are mature. When 
mature the eggs are probably 850 fi in diameter, or even larger. Spermatozoa 
were found in an ovary containing eggs but 560 /x in diameter. 

The number of young found in Lucifuga were 4, 15, and 10 respectively. The 
young were nearly all turned with their heads toward the front of the ovary, a 
condition duplicated in the ovary of Cymatogaster with nearly mature young. The 
condition of the young in the ovary with 4 young is well shown by the photograph 
(plate 15, fig. c). There were 2 young on each side. The largest eggs in this 
ovary were 200 fi in diameter. 


The condition in a female 90 mm. long containing 15 young, about 12 mm. 
long, was as follows: there were 11 on the left side, one of which had an ovarian 
lobe in its mouth, and several had the gill covers hooked over ovarial lobes, the rest 
being free in the cavity (plate 15, fig. d). There were 4 on the right side, one of 
which had the head turned to the rear, and one was so firmly attached to the 
ovarian lobe by the gills that it was practically impossible to get it loose without 

One ovary of Stygicola containing 1 1 large eggs, at least one of which is free in 
the ovary, is distended much more than the few eggs would warrant, being 16 mm. 
long and 12 mm. wide. The outer tunic is quite thin. The eggs are nearly of 
the same size and measure 848 ju. in diameter. The general features of the ovary 
of this species are given in plate 27, fig. A. The details of the structure are given 
in another chapter by Lane. 




The snout of Lucifuga is broad and depressed to the posterior edge of the max- 
illaries — duck-bill shaped. The eye is distinguished without difficulty in the trans- 
lucent living individuals, and even in specimens preserved in formalin or alcohol 
it is readily distinguished up to very old individuals. 

In the older specimens the skin over the eye readily discloses the location of 
the organ. There is over the eye in these specimens a hemiovate elevation sepa- 
rated from the rest of the skin of the head by a distinct groove. The skin in this 
ovate arch is not any less abundantly supplied vi^ith pigment than any other part 
of the head, and there are no other distinguishing features to indicate that it is 
better adapted to admit light than any other part of the skin of the head. In 
some cases it is even more densely pigmented than neighboring regions. The 

region is proportionately larger 
in young individuals than in old, 
but is more conspicuously de- 
marked in the older than in the 

Removing the skin shows 
that beneath the ovate arch lies 
a mass of orbital fat, approx- 
imately in the center of which 
the eye lies embedded. The 
orbital fat-mass seen from above 
has an oval shape, considerably 
longer in the axis of the head 
than transversely. Behind, the 
mass touches the orbital process 
of the frontal bone. The eye is 
placed approximately over the 
middle of the maxillary. 

The proportion of the or- 
bital space or socket occupied by the eye differs greatly in individuals of 
different sizes. In younger individuals, just about to be born, the eye fills a 
large part of the socket (plate i6, fig. b), while in the old it forms an insignificant 
dot in a mass of fat and connective tissue, hundreds of times larger than the eye 
(plate 2i). The relation of the eye to the surface is similarly conditioned with age. 
In the young it lies near the surface, while with age it becomes farther and farther 
removed, retaining however its relative position in the orbital fat-mass until old 
age, when possibly it may move nearer to the skull. 

Seen from the surface, that is without sectioning, the eye presents great fluctu- 
ations in size. These are in part conditioned by the size of the individual, but in 
part are independent of size. Other things being equal, the eye decreases in size 
progressively from birth to its disappearance in extreme old age. This process is 
accompanied by, if it is not responsible for, the appearance of pigment masses. 
These are either intimately associated with the eye, as in the development of great 

Fio. 72. 

(A) Outline Camera Drawings of Eye of 4 Young of Female shown in plate 
isC, from Sides, Left Eye on Left, Right Eye on Right, so that Middle of 
Pairs is Anterior, o. Fish 18 mm. long; 6, 18.5 mm.; c, 19 mm.; d, 20 
mm. For details of these Eyes see figs, plates 16 to 18, 16 mm. and 4. 

(B) Eyes of Mother of 4 Young, shown in A, drawn to Same Scale; a, from 
above; 6, from sides. For sections, see plate 21. 

(C) Outlines of Eyes of No. os. a Fish 53 mm. For sections see plate 20 
c and plate 24 A, 16 ram. and 6. 



■< o 
o> «. 

^ 3" 

» o 

ft OQ 

3 a- 


• o 

(» -^ 





OQ ^ 

n' ^ 

<* o 

o -^ 

ft 3 

>< n 

„ o 

o ■" 

S g 

2. y 







3 sr 

ft -" 
f -. 


» « 
cr _ 

g ^ 













8 i 

a> n 

^ a, 

- o 
n "O 











1: 2; 



IT 2 

a. ? 







31 » 

• lA 

X S 


■ o_^ 





2 -^ 

(J 2,£. 

"J CA 

fj. 1 


01 o 
- 3 



o n 

< •< 































Eyes of Lucifuga, 25 mm. long. 

A. Right eye, showing vesicular arrangement of pigment layer and retina and 

folding of sclera. From above. X 1 00. 

B. Left eye, shriveled and sclera similarly folded. 



pigment cushions on the eye, or in extra ocular regions at times in contact with 
the sclera, at other times in the orbital fat some distance removed from the eye. 

While, other things being equal, we find a progressive decrease in the size of the 
eye with age, we do not find that individuals of the same size have eyes of the 
same size. On the contrary, the eyes of individuals of approximately the same 
length may be very different in size and, as we shall see later, in structure also. For 
instance, of 4 young taken from the ovary of one mother and differing from each 
other by not more than 2 mm. in total length, we have the eyes of two individ- 
uals without a lens and the eyes of the other two with large lenses. The eyes 
measure 272, 320, 384, and 416 /* respectively, or, after clearing in xylol, which 
permitted a more minute measurement, 260, 280, 375, and 425 (fig. 72, a). De- 
tailed measurements of these eyes will be found in the following table: 

Measurements in /x 0/ Eye 0/ Female Lucifuga and of Four Young contained in her Ovary. 

[x, as they 

were talc en 

from the ovary ; y, cleared in xylol 

; z, sectioned.] 




Left Kye 








in eye. 






76 > 




























































Right Eye 









in eye. 






76 • 























































' Lower embryo of right ovary, 
■ Upper embryo of left ovary. 
* Lower embryo of left ovary. 

* Upper embryo of right ovary. 

s Vertical distance between imier margios of scleral cartilage. 


Still more striking is the variability in the size of opposite eyes in the same 
individual whatever its length. There are minute differences in the size of the eyes 
of the two sides at all times, the individual with two eyes exactly alike is probably 
not to be found, but the differences in mind are of a much larger order. For in- 
stance, in the mother of the 4 young mentioned above, the left eye had a longitudinal 
diameter of 170, the right eye 225 ; that is, the right eye was a third longer than the 
left. Instances of this sort are by no means rare, there being a marked difference in 
a number of the individuals secured. In one of the oldest secured, the eye of one 
side is all but gone, that of the other still well defined (plate 23 and plate 24, d). 
In a much younger one, 43 mm. long, I have found no eye in one side. In another 
the left eye bears the ratio of i to 3 to the right eye, which is therefore almost 
nine times as large as the left (plate 24, figs. A and b). 

Such big differences between the eyes of the two sides, fluctuating in amount 
in different individuals, but readily seen in living specimens, are found in about 10 
per cent of individuals. Sections usually showed that such differences whenever 
they existed were largely to be found in the pigment layer which in the large eyes 
was vesicular and the retina shriveled and retracted to the pupil, leaving a large 
space between the pigment epithelium and the rest of the retina (plate 21, fig. b; 
plate 22, fig. a; plate 24, figs. A and c). 

Note. — One element of error is present in the exposition'of'the eye of Luci/uga. Lucifuga and 
Slygicola live together in the same caves. There is no difficulty in distinguishing these after they 
reach acertain size. What that size is I can not say, but at 60 mm. they are conspicuously different. 
The smallest specimen of Slygicola unquestionably determined is 60 mm. in length. Possibly 
the two species are superficially indistinguishable when young, and some of the young specimens 
mentioned below 60 mm. and used in preparing the following account may in reality be stygicolas. 
All specimens below 60 mm. secured had the characters of Lucifuga. The probability of this pos- 
sible error is not as great as it may appear at first sight, as an analysis of the origin of the specimens 
less than 60 mm. will show. Seven of the specimens less than 60 mm. sectioned are from the cave 
of Jaiguan. From this cave 23 fishes were taken, 5 of which were stygicolas. The smallest of the 
stygicolas was 81 mm. and considerably larger than the smallest undoubtedly distinguishable spec- 
imens. If there were no specimens of Slygicola between 60 mm. and 81 mm. long when they 
could have been readily distinguished, it is probable that there were none smaller. The 3 smallest 
specimens of Slygicola measured 81, 90, and 97 ram. respectively. From Hawey I secured only 
Lucifuga, at least 3 of them being larger than the smallest specimens, permitting an unquestioned 
determination. From La Fria the only 2 over 60 mm. long were Slygicola,vi\n\e those below 57 mm. 
were apparently all lucifugas. Two of those sectioned, 54 and 57 mm. long, may be considered 
lucifugas without a doubt. This leaves one 27 mm. and one 28 mm. in doubt. In Los Baflos we 
secured no large specimens ; all the small ones were referred to Lucifuga. In Ashton large and small 
were all referred to Lucifuga, the smallest one sectioned from this place being 53 mm. ; it is undoubt- 
edly a Lucifuga. 

The proportion of stygicolas to lucifugas among individuals over 60 mm. is: stygicolas, 43; 
lucifugas, 36. Lucifuga does not reach a size over 104 mm., and comparing the ratios of lucifugas to 
stygicolas, between the smallest determined Slygicola 60 mm. and the largest Lucifuga 104 mm., we 
get Slygicola 32 mm., Lucifuga 36 mm., or a ratio of i to 1.25. But of the 32 stygicolas between 60 
mm. and 94 mm., 10 came from the "M " Cave which is remote from the region where lucifugas were 
found. Eliminating these, we would get a ratio of 22 to 36, or i to i Vi for the region where both 
are found. This, other things being equal, would give us the probability that any of the 
younger specimens found in the region where both species were found was a Slygicola or a Luci- 
fuga. More than this, in the "M" Cave, about 60 miles removed from any cave in which Luci- 
fuga was found, Slygicola is very abundant, but we secured no specimens less than 60 mm. long in 
five trips, nor were any small ones found in the Donkey and Carboneria, where only Slygicola 



occurs and where it is abundant. This makes it seem probable that the young of Stygicola live in 
deep water and are not found in the open sink holes or that their habits otherwise prevent them 
from being found. 

One more element tends to show that all the young are Lucifuga. Lucifuga differs from Stygicola 
in the shape of the nape, the scales of the head, the teeth, and the number of fin-rays. These char- 
acters in the young were always those of Lucifuga as far as could be made out. 

With these preHminary remarks the details of the structure of the eyes of differ- 
ent individuals may be given. The different parts of this account may be begun 
with a description of the conditions obtaining in the 4 young and their mother 
(referred to as 76) since there can never be any question concerning the genetic 
relationship of the eyes. 


The six normal eye muscles are all present in the young of 76, both in those with 
a large eye and those with a small eye. The muscles in one of the large-eyed 
specimens and one of the small-eyed specimens have the following maximum 
diameter : 





Dorsal oblique. . . 
Ventral oblique.. 
Dorsal rectus .... 



Anterior rectus . . . 
Ventral rectus .... 
Posterior rectus . . 





These muscles can all be traced quite readily from their origins to their inser- 
tion (plates 16 to 18, msc.) and are apparently quite normal. 

In the mother of these young the oblique muscles can be very readily traced 
in the socket in front of the eye, but their insertion in the eye is by fibers bent 
nearly at right angles. 

The dorsal, ventral, and posterior rectus of the left eye can be traced from 
their origin to their insertion. The posterior rectus is an exceedingly slender 
thread, and with the ventral rectus diverges from their origin, they converge again 
at their insertion. The dorsal and ventral recti are merged with the oblique 
muscles so that they appear as continuous strands, with the fibers mentioned 
above diverging from their union. (See also plate 22,^fig. c.) 

In the right eye the posterior rectus is attached to^the^eye independently, the 
ventral rectus and oblique are much more remote from the eye than in the left eye 
at the point where the connecting fibers are given oflf to the eye. 

In one of the largest individuals, 93 mm., the oblique muscles can be seen in the 
socket, but I have not been able to connect them with the eyes. The dorsal and 
ventral recti are present and possibly the posterior rectus. The muscles are, in 
other words, not so very different from those in 76. 


The sclera is most highly developed in the eyes of unborn young about 20 mm. 
long. It is well developed, with its cartilages, in 12 mm. young. Its most striking 
feature is the large scleral cartilage. This in the young 20 mm. long is a segment 
of a hollow sphere with a large opening for the iris. The edges of the proximal 
opening are at times curved in. It resembles a convex shield with an opening in 

* The four young specimens bear the serial numbers 76, a, b, c, and d. 


the center. In these early stages the cartilage is usually in contact with the iris in 
front, but diverges widely from the eye proximally and not infrequently extends 
beyond the eye. The inevitable conclusion is reached by an examination of such 
figures as scl. c. of plate i6, figure a; plate 17, figures A, b; plate 18, figure b, 
that the sclera was built for an ontogenetically or phylogenetically much larger 
eye than the largest found, and that the sclera has not been reduced at the same 
ratio as the eye itself. There is here no possibility of an artificial shrinking causing 
the space between the sclera and the eye, because this space is filled with undis- 
turbed tissue, and the only indication of a shrinking is sometimes noticeable prox- 
imal of the eye, between it and the fibrous part of the sclera. 

The ratio of the largest eye found in the young (76 a) to the eye suggested by 
the sclera is about as 45 to 85 ; in the smallest eye among the young of 76 (i. e. 76 b), 
it is about as 20 to 49. The eye, however, even if as large as suggested by the 
scleral cartilage, would still be a very small eye, unless the scleral cartilage formed 
but a rim over the front of the eye. 

The cartilage is only about one cell deep, except near the outer rim where it is 
occasionally thickened. Over the back of the eye stretching from the proximal 
edge of the scleral cartilage there extends a slack membrane very much thinner than 
the cartilage and apparently continuous over the surface of the cartilage as an ex- 
ceedingly thin membrane. Near the scleral cartilage this proximal membrane has 
a definite outline which is at times lost toward the optic nerve, the membrane 
becoming flocculent and its substance less readily distinguishable from the con- 
nective tissue filling the socket. A similar membrane more uniform in outline 
and consistency over the front of the eye represents the cornea. 

The scleral cartilages degenerate shortly after birth. In the eyes of recently born 
individuals they differ from those in the eyes of the unborn by fitting close to the 
bulb. They have apparently been drawn to the bulb and in this process lost their 
symmetrical shield shape and are at times bent in acute angles, at other times their 
free margins project considerably beyond the eye. In one case, an individual 
(No. 203) 25 mm. long (scl.c. plate 19), the cartilage in shrinking to the eye was 
thrown into a fold extending some distance from the eye. The pockets formed 
between the layers of cartilage in this fold are filled with pigment apparently belong- 
ing to the retina. This peculiarity is found in both eyes of this individual. The 
cartilages in free living individuals are much more variable than in the unborn young, 
and even in one individual only 28 mm. the cartilage of one eye has entirely dis- 
appeared, while that in the other is a minute bar folded upon itself. In only a 
single case, to be described shortly, were there any traces of cartilage in specimens 
over 40 mm. long. The fibrous part of the sclera differs greatly in thickness in 
different eyes of older fishes or even in the same eye. 

The greatest amount of difference between the sclera of the mother and the 
unborn young described above (76) is undoubtedly found in the cartilage. In the 
right eye of the mother there is no definite cartilage at all; there is a nodule of 
substance at the lower margin of the iris that may be the remnant of the cartilage, 
but otherwise there is nothing in this eye to indicate that there ever was any car- 
tilage associated with it at any time. In the left eye of the same individual are two 
nodules of cartilage, one tangent to the dorsal surface of the eye (plate 21, fig. a), 
the other in a vertical section through the middle of the eye somewhat below the 


level of the optic nerve. The former retains its distinct cartilaginous nature while 
the latter has lost it to such an extent that it is only by inference that it can be 
considered of cartilaginous origin. These are the only cartilages seen in eyes of 
individuals over 40 mm. long. The fibrous part of the sclera is as well developed 
as in the younger eyes, and indeed near the nodules of cartilage in the left eye it 
is distinctly thicker than in the younger stages. The sclera as a whole no longer 
forms a capsule much larger than the eye ; it fits snugly against the eyeball, except 
in the cornea of the right eye, where it forms an arch over the iris and pupil in the 
normal way. Where the cornea joins the sclera proper in the right eye, there is 
again a material thickening of tissues. 

The cornea in older individuals undergoes many modifications. It retains 
its shape for but a short time after birth. In 16 individuals over 24 mm. long it 
retained its original outline in only 4 eyes in 4 different individuals, one 28 mm. 
(plate 20, fig. a), one 38 mm., one 53 mm., and one 65 mm. long (plate 21, fig. a), 
the mother mentioned above. In the other eyes the aqueous space is obliterated, 
and the cornea more or less disintegrated. In cases where the vitreous cavity had 
disappeared, and the pupil had become closed, the cornea was at times replaced 
by a lenticular mass, cellular rather than fibrous (plate 20, fig. c). 

The points of interest are that the sclera develops early and on a scale much 
beyond the present needs of the eye, i.e., it preserves a past phylogenetic stage far 
better than the other parts of the eye, and yet ontogenetically it degenerates much 
more rapidly than any other part, with the possible exception of the lens. 


In unborn young about 20 mm. long there is considerable space between the 
sclera and choroid. At first sight this may be taken as the result of shrinkage on 
the application of reagents, but a closer inspection shows the space to be filled with 
an undisturbed gelatinous substance interspersed with nuclei. It represents the 
suprachoroid al lymph space. Immediately in contact with the eye, the gelatinous 
matrix is replaced by fibers. The normal condition of the gelatinous layer is 
further testified to by the dendritic choroidal pigment cells that are scattered through 
it and occasionally are arranged into a thin layer, dividing the mass into approxi- 
mately two equal layers. Still further evidence is given by the occasional blood- 
vessels passing through it. 

In 76 a there is a fine capillary meshwork in the choroid. In the meridian of 
the optic nerve an artery approaches the entrance of the optic nerve from below 
and a vein much thicker leaves it above. The vein is made up of two branches in 
the choroid near the entrance of the optic nerve, one branch coming from above, 
the other from below. The artery enters the retina along the lower edge of the 
optic nerve. The vein leaves the retina in this eye over the lower margin of the 
iris. The meshwork of blood-vessels over the inner surface of the retina contains 
many far beyond capillary size, closely approaching in thickness the retina itself. 
There is a median vessel extending from the lower edge of the pupil along the sur- 
face of the retina up to a level with the upper surface of the lens. 

In the left eye of 76 the ophthalmic vein measures 50 fi in diameter, while the eye 
itself measures but 170 fi. In the eyes of this individual I have not been able 
to make out any blood corpuscles, nor have I been able to identify the ophthalmic 


In a specimen 94 mm. long the ophthalmic vein can readily be traced. In a 
specimen 93 mm. the ophthalmic vein of the right eye is seen to measure 40 fx as 
compared with a diameter of the eye of about 100 fi. A few blood cells are seen in 
this eye. A considerable mass of pigment is developed in the choroid, in places 
15 /A thick. It is not possible to make out any vascular network either in the cho- 
roid or in the eye. Very few blood-vessels are seen about the eye itself, although 
the vessels leading to and from the eye are very large and filled with blood corpuscles. 

In the eyes of older individuals there is a great diminution in blood in and about 
the eye. The capillary meshwork in the choroid and the vitreous vessels are no 
longer readily distinguishable, their reduced size being further indicated by the 
absence or inconspicuousnessof the large choroidal veins seen in 76 a. The ophthal- 
mic vein is, however, very large and well filled with corpuscles in even the oldest 
individuals. It has here the appearance of a sinus rather than a vessel. Certainly 
the necessity of the eye does not require a vessel equal to nearly half of the total 
diameter of the eye as in the case of 42. 

The entire vascular arrangement gives the impression of being abnormal. A 
key to the large blood-vessels or sinuses is probably found in several of the eyes of 
Stygicola to be described later. In them it was definitely determined that blood 
lakes had formed in and about the eye that were entirely cut off from the circu- 


Near the eyes of all specimens above a certain size there are found masses of 
pigment. They are probably cells gorged with pigment which are aggregated in 
one or several masses. For instance, near the left eye of the largest fish examined, 
there is a large (80 x 128 /* in section) pigment mass 144 /* from the eye. It is oval 
in its proximal end ; truncate in its distal. Some of the denser fibers of the capsule 
surrounding the eye extend out to it. Another less distinct pigment mass is found 
in contact with the eye in a manner to make it difficult to determine its relation to 
the eye. It may be part of the retinal pigment (plate 23, fig. b). On the right 
side there are several pigment masses located in the orbital fat near the eye : one, 
80 X 96 /a; another circular mass, 32 ju,; another, 80 /i* in diameter near the eye ; and 
still another, 32 x 48 /a. Some of these are evidently composed of lobes or distinct 
subsidiary masses. In very thin sections it can be seen that the cells composing 
the masses are filled to distention with granules about 0.7 /x in diameter, just such 
as are found in the pigment of the retina and in the subepithelial pigment of the 
skin. The cells measure 9 to 14 /u, in diameter. They are rounded, sometimes 
flattened where they are in contact. When fully pigmented their well-defined 
outlines and the occasional undoubted relation of nuclei to them are the only 
indications that they are cells. 

Remote from the densely filled cells, a number of cells can be made out in one 
individual in which the nucleus is located at one margin and the cytoplasm con- 
tains a few, or even but one, granule, while in others no granules are found. The 
nucleus is always kidney-shaped with the concave side toward the cytoplasm. There 
is, for instance, one nucleus near one of the large masses, similar to the nuclei in the 
mass flattened on one side and associated on that side with a hyaline bag of definite 
outline and containing a number of the pigment granules ; near it is another with 


more pigment in a more elongate mass. In the youngest individual (38 mm.) 
with whose eyes pigment was found associated, it is close to the optic nerve on one 
side of the body and along a fibrous strand on the other. 

The cells are fully charged with pigment, and no cells could be found with but 
a few granules. In the next largest (43 mm.) there is a large pigment cushion on 
the posterior face of the left eye. There are also a few fully pigmented cells scat- 
tered distad from the eye. 

In individuals 44 mm. and 54 mm. long the pigment is also associated directly 
with the eye, but the parts can not be readily distinguished. 

In an individual 53 mm. long there is a mass distad from the right eye over the 
pupil, and another proximal to the left eye. These are the beginning of the masses 
seen near the eye in older individuals. 

In an individual 57 mm. long there are small masses of pigment cells some 
distance removed from the eye. On the left side the mass exceeds the size of the eye. 

In the left eye of an individual 63 mm. long there is a large amount of pigment 
immediately around the eye and also masses removed some distance from the eye. 
The same is true of the right eye, which is large and vesicular. 

In an individual 65 mm. long there is a small pigment mass remote from the 
eye and a larger amount directly associated with it. 

In an individual 69 mm. long (plate 22, fig. a) there are masses of pigment 
near the eye which is vesicular. In an individual 80 mm. long small masses are 
found near the eye and there is much pigment in the eye. 

In the right eye of an individual 84 mm. long there is a very thick (30 /*) mass 
of quadrate pigment cells in the choroid along the lower surface of the eye. The 
pigment layer of the retina is but 4 /* thick and there is a lenticular mass of pigment 
cells, 46 X 34 ju, in section, in the pupil. The vitreous cavity is obliterated. 

In one of the largest fishes, 93 mm., there are large masses near the eye as well 
as a cushion of pigment affixed to the eye (plate 22, fig. b, pi. s.). 

From the above it is seen that the pigment masses make their appearance at 
about the time the eye begins to actively degenerate, a short time after birth, and 
that they reach their maximum development when the eye has reached the vanish- 
ing point. The masses are first seen in a fish 38 mm. long in association with the 
optic nerve and the muscles near the eye. In slightly older individuals the pig- 
ment masses appear as lenticular cushions applied to the sclera, and in still older, 
when the fish has reached 50 mm., other masses are seen more or less remote 
from the eye, although pigment cushions may still be seen in some of the larger 
specimens. In the very largest there are several masses in the neighborhood of the 
eye or where it has disappeared. 

While it is practically impossible to make out the structure of the pigment 
masses in their most intense development, it is evident that they are made up of 
rounded bodies densely pigmented, several of which are bound by fibrous tissues 
into subsidiary masses many of which together form the larger masses described. 

No doubt the smaller rounded bodies are cells. In their most intensely pig- 
mented condition it is impossible to demonstrate this. In certain favorable cases 
the individual pigment granules can be made out, as well as their arrangement in 
the cell. In the very largest individuals some cells were found that contained but 
one or very few pigment granules. 


The appearance and gradual increase of these pigment cells and masses with 
the beginning and progressive degeneration of the eye makes an intimate depend- 
ence of the one phenomenon on the other very plausible. That pigment cells may 
sometimes appear and become pigmented at some distance from the degenerating 
eye is seen in the optic cavity of the largest individuals, where cells with but few pig- 
ment granules were seen remote from the eye. Furthermore no phagocytes or 
pigment cells in the process of gorging were seen in the eye. But in one case at 
least there were found a number of fully pigmented cells between the pigment 
layer and the rest of the retina. There seems to be little doubt, therefore, that 
there is direct association of at least some pigmented cells with the degenerating 
eye. Other indications as to the possible origin of the pigment masses are given 
under the head of the lens. In some of the degenerating lenses cells containing 
pigment granules were found. These cells are 6 fi to 9 /x in diameter. They are 
most numerous in the lens of an individual 25 mm. long before accumulation 
of pigment cells into masses has taken place. 

I have noticed similar pigment accumulations in the eye of Amblyopsis. 

Pigment is found in very variable quantity and variously scattered in the con- 
nective tissue surrounding the eye. The amount of this pigment seems to vary 
inversely with the amount of pigment in the eye itself and to increase with age. 


The variation in the lens is not equaled even by the variation in the sclera. 
Here, as in the sclera, we may begin the account with a description of the conditions 
in the 4 unborn young taken from the ovary of a single mother. In 2 of these, 
a and d, the lens is still present; in the other 2 there is no indicatio-n of it. In a and 
d it consists of a sphere (plate 17, fig. b, plate 18, fig. a) incased in a fibrous mem- 
brane of varying thickness, flocculent peripherally, becoming dense and firm and 
containing nuclei proximally. The contents of this membrane are evidently under- 
going histolysis. It is an amorphous, granular substance with partially dissolved 
masses, some of them still showing nuclei. At other places the nuclei have degen- 
erated into black chromatin lumps. There is absolutely no indication of lens 
fibers. The cortical layer of the mass is at times compact over the distal surface 
and this is the only indication of an epithelium covering this part. In the lens of 
a stained with iron haematoxylon, the center which chiefly contains the masses men- 
tioned above is in part quite black. In h of these young the only indication of the 
lens is a small vacuity in the connective tissue between the edges of the iris (plate 
18, fig. b). There is nothing about this space except its position to indicate that 
it was ever in the remotest way connected with a lens or its capsule. The lens 
in stiU younger ones (12 mm.) is much as described in a and d. It consists of a 
fibrous capsule filled with a mass of undifferentiated cells. 

In small individuals, ranging from birth with a length of about 24 mm. to 38 
mm. in length, the lens is usually present in a more or less advanced degree of 
degeneration. In the degeneration the solid contents of the lens capsule largely 
disappears, the capsule collapsing or not. 

In a young 25 mm. long the lens capsule of the right is very much shriveled, like 
a collapsed balloon and contains only about a dozen small cells, some of them 
nucleated, others in part filled with dark brown pigment granules. These look not 




Terences in 

Eyes of Lucifuga. 

A and B. Eyes of opposite sides of young, 28 nun. long, showing great diffe 

size of eyes and also of general structure. In A lens capsule shows well. 

C. Eye of individual 53 mm. long. X 200. Eye of opposite side very large 
and vesicular (represented in plate 24 A.) 



y^ ■•■: 

: /•-' 

Sections through left and right eye-cavities of Lucifuga, specimen No. 76 (see plate ISC). 

These figures have same magnification as plate 1 6B (of head of one of her young) 
with which they should be specially compared. Eye-cavities very large as compared 
with those of the larva (plate I 6B), while eyes are much smaller. X 60. 






B C 

A. Right eye of a Lucifuga, 69 mm. long. The pigment layer with choroid and sclera 

( / , chr. & scl ) is vesicular, walls of vesicle having shriveled somewhat. Section 
is along optic nerve (n. op.) and shows large, round pigment mass (pi. sph.) 
between pigment layer and lower part of retina. X 60. 

B. Left eye of adult, 93 mm. long. Eye has become very small, pigment layer 

incomplete. Very large mass of pigmented cells has accumulated over and in 
front of eye ; strand of connective tissue (with nerve fibers) extends from front of 
eye out toward surface. X 60. 

C. Oblique (lower) and rectus muscles sending common tendon to eye. From 

fish 44 mm. long. 





Eye of an old Lucifuga, 94 mm. long. 

A. Left eye-socket witfi contained eye and pigment-mass at its left. X 60. 

B. Part of same section, X 375, sJiowing fibrillar network about eye. Eye a nodule 

of cells in which distinction can only be made between pigment and retina. 
Part of pigment mass at extreme left. (For right eye see plate 24 D). 


unlike white corpuscles that have been abundantly fed with pigment granules. 
Whether they carried these in with them or whether the remnant of the lens had 
undergone a pigment degeneration, I am not able to say. 

The eye of the other side is much damaged in sectioning, but is essentially the 

In a young 24 mm. long, evidently just born, the lens capsule of both eyes is a 
large balloon, little wrinkled, and containing but little stainable material, all of it 
of the same nature as that described above. 

In a young 27 mm. there is no remnant of a lens in the left eye, while in the 
right there is the filmiest suggestion of the lens capsule, but nothing more. 

In an individual 28 mm. long the lens of the right eye is represented by a nearly 
empty capsule, that of the left is shriveled, contains pigment, and is entirely in the 
vitreous space, the pupil having closed. 

In an individual 38 mm. long the left lens is represented by a large empty col- 
lapsed capsule, that of the right being small and collapsed. 

The lens capsule is the last part of the lens to disappear. In specimens over 
40 mm. long, it was observed in only two doubtful cases ; in all others there was no 
trace of it left. 

It is quite evident from the structure of the lens displayed in the unborn young 
18 mm. long that it had passed its point of highest organization and was obviously 
far along on the route of degeneration. Indeed the lenses of the young (12 mm.) 
show no signs of fiber formation and also show indications that they have begun to 

Conspicuous and remarkable are the fibrous lens capsule which persists after its 
contents have disappeared, the irregularity of the contained cells in their highest 
development and their irregular distribution, and finally the pigment-fed phagocytes 
in the capsule. 


On account of the fluctuation in the size of the eye it is difficult to determine 
whether the end of its development is reached with a length of 12 mm. or not until 
a length of 20 mm. In the 4 embryos, 76 a, b, c, and d, about 20 mm. long, the eyes 
fluctuate from a maximum 425 yx in longitudinal diameter in the longest, to 260 /* 
in the shortest. If the embryo with the smaller eye had been of smaller size, it 
would have been but natural to come to the erroneous conclusion that the eye 
increases with age till the fish reaches a length of 20 mm. The same is true in 
respect to the differentiation of the retina. One can not say in general that the 
retina progresses in any respect between the length of 12 mm. and 20 mm. I can 
only say that the most highly developed retina was found in an unborn individual 
20 mm. long (plate 18, fig. a, and plate 24, fig. e). 

In the retina of the youngest individuals (12 mm.) there is a distinct differentia- 
tion into a ganglionic layer occupying 0.24 of the total thickness, an inner fibrous 
layer of the same thickness, a nucleolar layer 0.32 of the total, and a pigment layer 
occupying 0.20 of the entire retina. The boundaries of the different layers are not 
equally regular at all places, and the nuclear or ganglionic layer sends a connect- 
ing series of nuclei in an irregular manner through the reticular layer in different 
places. The pigment layer is well pigmented. The inner cell layer of the uvea is 


not pigmented and forms a distinct ciliary process. Between the latter and the rest 
of the retina there is an accumulation of elongate nuclei. 

This retina has reached a stage in an irregular process of histogenesis, or it 
has earlier stopped at such a stage of differentiation, or finally, it has reached its 
present condition as a degeneration from an earlier, more highly differentiated 
stage. From the material at hand it is impossible to determine when the ret- 
ina reaches its highest stage of development and when it begins to degenerate. 

A slightly higher stage of differentiation is found in one of the eyes of one of the 
unborn young of 76. In this eye, the retina has about the same total thickness. 
There is found in places a very distinct separation of the outer layer of nuclei into 
an inner layer, a reticular layer, and an epitheUal layer. To one of the epithelial 
nuclei a cone is found attached (plate 24, fig. e). A ciliary process is not seen in 
this eye nor in the group of elongate nuclei so conspicuous in the younger stage. 
The inner layer of the uvea, as well as the outer, is pigmented. 

Beyond birth only general processes can be described without entering into a 
minute description of each eye. The retina degenerates progressively and it 
seems to do this accompanied by one of two modifications in the general structure 
of the eye. The eye may shrivel (plate 20, figs, b, c; plate 21, fig. a), the pig- 
ment layer lying close against the rest of the retina ; or the pigment layer may sepa- 
rate itself from the rest of the retina and become very greatly distended, the retina 
itself forming but a small segment of the eye vesicle (plate 22, fig. a; plate 24, figs. 
A, c). Plate 22, figure a, represents such an eye, in which the retina is well con- 
tracted and the pigment layer shriveled. The optic nerve passes through the 
vesicle. The beginning of such a modification is probably to be seen in plate 21, 
figure B. In other cases the retina is drawn out laterally (plate 24, figs. A, c). 
Such vesicular eyes were also found in old individuals of Amblyopsis. There does 
not seem to be any increase in the amount of pigment, and, since it is scattered 
over a larger area, the pigmented layer of these vesicular eyes is less densely pig- 
mented than that of the shriveled eyes. In one eye conditions normal to a fish eye 
are more nearly retained. 

I am not able to say that one part of the retina undergoes a more rapid 
degeneration than another. They all reach the vanishing point with extreme 
old age. 

In an old individual (94 mm.) the eye of one side consisted of a few vacuoles 
surrounded by nucleated fibrous tissue (plate 24, fig. d). It is impossible to deter- 
mine to what these parts of the eye belonged. There are also scattered pigment 
granules and cells, while near this eye are a few pigment masses. The eye of the 
other side is better preserved and represented in plate 23. In one eye, which is 
shriveled to very small dimensions, a peculiar lenslike structure occupies most of 
the interior. Such lenslike structures I found in Amblyopsis and erroneously 
considered them the lens. In Rhineura it is distinctly seen that the structure fills 
an invaginated pocket of the pigment layer. 

A census of a series of eyes of individuals from the time of birth to old age gives 
us the following statistics concerning the lens, the vitreous space (that is, between 
retina and iris), and the aqueous space (between iris and cornea) : 



Statistics of Lens, showing Vitreous Space {between Retina and Iris) and Aqueous Space 

{between Iris and Cornea). 







25 (103) 
24 67 











Left eye. 

Empty capsule 

Large, entirely filling 

vitreous cavity 
Large capsule 



Capsule with pig- 


Large, empty 













Right eye. 

Collapsed empty 


Large, empty . . 


Filmy capsule . . 
Nearly empty . . 

Capsule with pig- 
Small, collapsed 








Empty vesicle . . . 




VriRiODS Spacb. 

Left eye. 


Small, filled by lens 






Left very large . . . 












Right eye. 

Very large 



Very small 


? .... 
Very small 

o .. .. 
Very small 

o .. .. 


Large . . . . 

o .. .. 

o .. .. 



Agraous Spacc 

Left eye. Right eye. 



Very large 

? . 




* lo the left eye the lens is not distinguishable, but is probably represented by a collapsed capsule in part filling the vitreous cavity. 




The account of the eyes of Stygicola is based (i) on two young born October 
20, each about 20 mm. long ; (2) on the mother of the above, 92 mm. long ; (3) on 
various other older fishes, from 60 to 135 mm. long.* The early stages of the 
development and the history of the eye between 20 and 60 mm. is not known. 

On October 30 I obtained a Stygicola at Alacranes. She gave birth to two 
young on the evening of October 31, at Canas. They were born tail foremost. 
The ovary of this specimen contained eggs 0.88 mm. in diameter, or nearly ripe. 
The 2 young are referred to as 125 a and 125 b. 

The head of 125 b, seen from above, is represented in plate 25, figure B, and 
the eyes are represented by plate 25, figure c. The eyes of the one born at Canas 
(125 b) were symmetrical, nearly of the same size. The eyes moved, and as far as 
I could judge were as readily movable as the eyes of other young fishes. 

The eyes were silvery, the argentea being apparently well developed. The iris 
was well distinguished, the pupil too large for the lens, having a downward directed 
notch continuous with the choroid fissure which is still visible as a pigmentless 
streak. While small, there was nothing in the general appearance of the eye that 
would lead one to conclude it might not be functional. 

The eyes are so placed in relation to the brain that a line tangent to their pos- 
terior faces would be tangent to the anterior face of the optic lobes. This condi- 
tion corresponds very well to the position in Lucifuga of equal size. 

Table of Measurements. 

Curmit No. 

135 ' 







90 mm. 

60 mm. 

88 mm. 

93 mm. 

135 mm. 














Vertical diameter . . . 





Outer nuclear 

















































Blood-vessel in eye . . 
Optic nerve 

« From outer margin of scleral cartilage, unless otherwise stated. 

■ From outside of pigment to outside of j>i(inient. 

* Total thickness of retina 67 m. as compared with 237 ix in Zygonectes. 

* These eyes lie 0.5 mm. below the surface. 
s This eye lies 0.3 ram. below the surface. 

About the left eye of the second young (125 a) there was a large accumulation 
of blood, which in section is seen to be in the choroid layer and mixed with the 
orbital fat. Measurements of the eyes of the young, as well as of the mother, 
are shown in the above table, and see also plate 25, G. 

' A single larva obtained on September i, between lo and 12 mm. long, is not well enough preserved to be 


In these young the eyes are in contact with the skin and fill a large part of the 
fibrous orbit. With age the eyes come to be farther and farther removed from the 
skin, and lie in the orbital fat, which may be many times the size of the eye. For 
instance, in the mother of the young (125 a and b) the eyes are approximately in the 
middle of the large eye cavity, which is over a thousand times as large as the eye, 
having on the left side a vertical diameter of 1.8 mm. and a lateral diameter of 3 mm., 
whereas the eye has an average diameter of but 0.2 mm. The eyes are about 0.13 
mm. removed from the surface. The eye cavity is filled with cavernous connective 
tissue mesh work holding fat. About the eye the meshes are stronger and very 
rich in blood-vessels. About the eye in this individual, as in all old ones, there are 
also large accumulations of pigment. 

Parts of the eye have certainly begun to degenerate before birth. The lens 
leads in this respect. After birth there is a rapid general degeneration of the eye. 
This is not directly proportional to the increase in size of the fish. For instance 
(see table), in a specimen 60 mm. long the eyes are distinctly farther reduced than 
in one of 88 mm. The left eye (plate 25, fig. g) in life was surrounded by stagnant 
blood. The choroidal blood-vessels were distended and the vessels of the vitreous 
body were also abnormally large. The entire eye was compact, and the retina, 
slightly withdrawn from pigment layer by reagents, shows a drawn-out process indi- 
cating an intimate relation between two layers. Figure f (right eye) shows eye nor- 
mal to this stage. The retina has shrunken away from the pigment layer somewhat 
and an artificial space has also been formed in places between sclera and choroid. 

As in Lucifuga, the eyes of opposite sides have at times undergone different 
modifications ; the eye on one side may be contracted into a compact ball, while on 
the other it is distended into a hollow sphere, eight or ten times as great in cubic con- 
tents. The left eye of 125 a and the left of 126 (plate 25, fig. o, and plate 26, fig. a) 
indicate that in these two eyes at least, the compression is associated with an accu- 
mulation of blood in the choroid vessels and in the orbital fat. While this blood 
does not have the appearance of a clot, the corpuscles have a very different staining 
reaction from those in the vessels. In 126 / there is a small vessel in front of the 
iris which contains normal blood (plate 26, fig. A, cps.), otherwise this eye is shut 
off from the circulation. The left eye of 125 a was certainly cut off from the 
circulation by the formation of a large blood lake about the eye. There is evi- 
dence in the right eye of 125 that extra limital blood has accumulated about this 
eye also. It would seem from these examples that one of the principal causes of 
degeneration is a disturbance in the circulation. 

Figure A of plate 26 shows the left eye of No. 126, 88 mm. long. The choroid 
blood-vessels are distended vnth blood corpuscles which stain differently from 
those in one of the choroid vessels. Other spaces or vessels filled with blood were 
found in tracts passing through the orbital fat-mass, past the eye. The iris was 
infolded and the pupil closed with a fibrous tissue containing blood-vessels. The 
lens was a flaccid membranous bag containing pigment granules and a few nuclear 
remains. The pigment layer variously pigmented (i) appears in two layers in 
places, and within it are found large rounded masses of pigment. The retina con- 
sists of ganglionic cells, and an outer layer of cells and a reticular layer, approxi- 
mately divided in the middle by an irregular cellular layer. The optic nerve in 
the figure is supplied from neighboring sections. 



The eye muscles of 125 b are well developed, with some anomalies in one of 
the recti of each side. 

The two oblique muscles arise just below the point of exit of the olfactory nerve 
from the brain cavity, downward and medial of the middle of the olfactory pit. 
They are attached to the membrane connecting the ethmoid with the vomerine 
cartilage. They extend backward in a canal bounded above by the ethmoid, 
below by the vomer, and laterally by another cartilage. The upper oblique is 
regularly horizontal-oval, measuring 34 fi by 48 fi. The lower oblique is slightly 
crescent-shaped in section with a diameter of 25/* by 83 /u,. These muscles are 
attached on the sclera so that the tips of their insertion are just in contact with the 
posterior rim of the scleral cartilage. The superior and inferior recti have their 
points of insertion on the cartilage just outside the insertions of the oblique. 

The anterior rectus of the left side is inserted on the anterior face of the scleral 
cartilage. It has a diameter of 20 fi near its insertion. It has its origin just in 
front of the exit of the optic nerve. On the right side the muscle arises just below 
the exit of the optic nerve, extends out and then curves down and joins the fibers 
of the inferior rectus, following the fibers of this muscle and becoming indistin- 
guishable from them. 

The posterior rectus arises far back, just below the origin of the ear capsule. 
It extends out and forward, with a diameter of about 30 /i and attaches to the pos- 
terior face of the eye. 

The superior and inferior rectus muscles are much stronger than the others; 
they arise much farther forward than the posterior rectus, about on a plane con- 
necting the posterior faces of the eyes. The upper rectus has a broad point of 
origin, the inferior rectus a narrower one below it. The upper rectus curves upward, 
forward, and out ; the lower runs in a nearly straight line obliquely down, out, and 
forward. On both sides the upper rectus gives off fibers to the lower. The method 
of the two sides is different ; on the right a compact bundle of fibers branches off 
from the root of the muscle, passes toward the lower rectus to whose inner face 
they become joined. The fibers pass from the origin of the superior to the in- 
sertion of the inferior rectus. While some fibers seem to have a similar course on 
the left, the conspicuous thing here is that fibers form an arch between the upper 
and lower recti, their origin and insertion being both on the eye. The important 
point is that in the eyes of the young the muscles, while varying to a degree on the 
two sides, are all well developed. The muscles are still conspicuous in a specimen 
97 mm. long, but in the mother of the young described and in older fishes, I have 
not been able to find any muscles (plate 25, figs, e, f, g, msc). 


The scleral cartilage is well developed at birth. Whereas, in Lucifuga, it formed 
a partial shield over the distal face of the eye, its pupilary diameter being much less 
than the diameter of its proximal rim, it here forms a ring about the equator of the 
eye the diameter of whose proximal rim is less than that of the distal opening. 
The walls of the ring are thickest in front, where they reach 30 /i, tapering back- 
ward. The ring in some cases fits the eye and does not, as in Lucifuga, suggest 



Mrs.ERBiehn^ del. 

B Meisellilh. 

Eye of Lucifuga. 

A. Righleye of Lucifuga, 53 mm. long. Left eye compact; sfiown in plate 23 A. Retina drawn 
from one section; optic supplied from several sections. X 440. 
B and C. Left and right eye of Lucifuga. X 270. 

D. Right eye of Lucifuga, 94 mm. long. (For left eye see plate 23.) 

E. Part of retina of eye of Lucifuga shown in plate 18 A. 




Eyes of Stygicola and Lucifuga. 

A. Lucifugas 20 mm. long and ready to be liberated. See plate 1 3 D (or others from the same ovary. 

B. Dorsal view of head of a young Stygicola, about 20 mm. long. 

C. (a) right and (A) left eyes of a young Stygicola 20 mm. long (No. I25A). In b details of marking of iris and ball are shown. 
D ,o) and (i) right and left eyes of the only other individual from the same ovary (No. 125a). 

E. Section of the left eye shown in C b. 
F and G. Vertical sections of right and left eyes shown in D a and A. Seen from in front. X 390. 
H. Section through middle of lens of left eye. 



/ '"^ •»■ ^X. A, '?♦ ■* 

- ti.op. 

MrsERBtelmg ilel 

Eye of Stygicola. 

A. Leh eye of Stygicola, 88 mm. long. From behind. X 390. 

B. Fragment ot pigment-layer, poor in pigment. From a neighboring section. 

C. Pigment-cell from choroid of same eye; few sections removed. 

D. Right eye of same fish, with same magnification. 

E. Left eye ot Stygicola, No. 1 25 (mother of 1 25 a and i). 92 mm. long. From in front. X 390. Section 

_LI: I.. ,1 L __»: 

B M: 

CI>C1 lIU"., 


a larger eye, but in others it is considerably larger. The cartilage degenerates rap- 
idly. In the eye of the mother of 125 a and b (92 mm.) only a few cells are left 
(plate 26, fig. E, scl.c). 

This history of the scleral cartilage in Ltici/uga and Stygicola is in distinct con- 
trast to its history in Amblyopsis. In the latter it appears as the last of the eye 
structures and remains after everything else has disappeared. The early history 
in Luci/tiga and Stygicola is not known, but it disappears even more rapidly than the 
lens, only a few cells sometimes remaining longer. Aside from the cartilage the 
sclera consists of a thin, fibrous, nucleated membrane over the proximal face of the 
eye and a similar membrane, and the cornea over the distal face. The cornea may 
remain for a long time after birth or it may, especially if the eye becomes compact, 
disappear and be replaced by an accumulation of cells such as have been seen in 
Amblyopsis and Lucifuga. With age the sclera becomes a fibrous capsule of varying 
thickness (plate 26, figs, e and f). 


The choroid in the eyes of the young consists of a thin membrane containing 
blood-vessels and pigment cells. Its structure can best be seen where it has acci- 
dentally become removed from the pigment layer by reagents. The blood-vessels 
may become so distended with blood that the thickness of this layer becomes several 
times its normal thickness. Between the choroid and the sclera in the young is a 
well-developed suprachoroidal lymph space. In contrast with Lucifuga, where 
this space is largest between the choroid and scleral cartilage, it is usually thin or 
absent in this region but comparatively well developed on the proximal face of the eye. 
In the old it is not evident (plate 25, figs, e and f, chr. I.). As in Lucifuga there 
appears, concomitant with the degeneration of the eye, an accumulation of pigment 
in the orbital fat or in the choroid. The outlines of such a mass in contact with the 
eye are shown in plate 26, figure f, representing the eye of a fish 135 mm. long. 
The accumulations of pigment in both eyes are very large — larger than the eye. 


'^" ' As in the eyes of Lucifuga, the lens degenerates and disappears more rapidly 
than other parts. The methods of degeneration are seen in the lenses of 125 a 
(plate 25, fig. h). The nuclei become distended, the chromatin accumulating in a 
few nucleoli-like granules. The membranes of the nuclei next dissolve and there 
results a mushy mass containing lumps of chromatin. The contents of the lens 
capsules are next removed in a manner not clear. Toward the end of this process 
the lens may be found to consist of a shrunken, fibrous membrane containing 
pigment granules and accumulations of pigment — possibly cells. There is not a 
vestige of the lens left in old individuals. 


f The iris in the young appears much darker from the surface than the rest of 
the eye. In sections it is found to be not more densely pigmented than the pigment 
layer. The epithelial part may be entirely pigmented, or the inner cells may be par- 
tially free from pigment. At birth the pupil is larger than the lens and it remains 


SO, becoming even larger in those eyes which become distended. In the eyes which 
contract, on the other hand, it becomes closed and the opening finally is entirely 
obliterated. The iridcal parts are then indicated by a layer of pigment much thicker 
than elsewhere about the eye (plate 26, figs, e and f). This pigmentation over the 
front of the eye is not unique in the Cuban blind fishes. It is well marked in 
Typhlomolge and so striking in Troglichthys that Kohl in this species ruled the 
irideal pigment entirely out of the eye. Where the iris joins the retina, in the ciliary 
region, there are cells with elongated nuclei as usual. 


The vitreous space and its history differ greatly in different eyes. For instance, 
in 125 a, right, it is very large, while on the other side it is almost entirely filled with 
the enlarged hyaloid blood-vessels (plate 25, figs. F and g). The factors that con- 
dition its structure in these two eyes also control its later history. In the eyes that 
contract the space becomes rapidly reduced and finally becomes obliterated (plate 
26, figs. A and r). In the eyes that become vesicular it remains, unless the vesicu- 
lation is so pronounced that the eye becomes a hollow sphere with the pigment 
layer forming the larger part of the circumference, and the retina is literally turned 
out to form the front part of the sphere. In the latter case the vitreous space, or 
body, naturally is turned entirely out of the eye. 

Between the vitreous body and the hyaloid membrane, or in the latter, pigment 
cells are sometimes found in old individuals. 


The pigment layer in the young 20 mm. long is a thin epithelial layer, well pig- 
mented. In places where it is artificially separated from the rest of the retina 
processes extend from the pigment layer down, and from the nucleated layer up, 
indicating more than mechanical contact between the layers (plate 25, fig. g). 
Such conditions argue that there is a beginning, at least, of the differentiation of 
cones. The pigment layer never becomes more highly differentiated than in 
these young. In eyes that become distended the amount of pigment being scat- 
tered over a wider area is much less dense at any one point. (Compare right and 
left eyes of 126, plate 26, figs, a and d.) 

Sometimes pigment cells, or simply accumulations of pigment, separate from 
the layer and come to lie between the layer and the rest of the retina. These are 
also found in Lucifuga (plate 26, fig. a). Whether this is a case of active migration 
of pigment cells or simply a result of mechanical crowding, I am unable to say. 

The remaining layers of the retina may best be considered together. In young 
20 mm. long (plate 25, figs, e, f, g) these consist of a ganglionic layer consisting of 
several series of cells. A nuclear layer, also consisting of several series of cells, lies 
immediately beneath the pigment layer. The nuclei are similar, there being little 
differentiation. Between the nuclear layer and the ganglionic layer is a sharply 
defined, broad, reticular layer. This is differentiated into a wider outer, a nar- 
rower inner, and a very narrow, more densely staining intermediate layer. In 
favorable sections stratifications can be made out in this entire layer. 

Miillerian fibers are seen, but I have not identified the nuclei belonging to them. 


I am not sure whether the conditions seen in older eyes are to be taken as the 
result of retrogressive changes, or of an abortive differentiation of an outer reticular 
layer with a separation of the nuclear into an outer and inner nuclear layer. 

The definiteness of the reticular and outer nuclear layers is no longer found in 
older fishes. Instead, there is found an irregular series of cells bounding the epi- 
thelial face of the retina (plate 26, fig. a). The other nuclei of the outer layer, seen 
in younger fishes, are scattered irregularly through the retina, leaving, however, a 
distinct, inner reticular layer. The outer reticular layer (4) so formed is much 
thicker than in normal retinas, and it is otherwise so irregular that it can scarcely 
be considered the homologue of the outer reticular layer of normal retinas. 

In the oldest eye examined even this degree of regularity is gone (plate 26, fig. f). 
There is an undoubted reduction in the number of nuclei, remains of which are seen 
as dark granules among the nuclei. The character of the nuclei in the older indi- 
viduals differs ; some still show granules, while others stain uniformly. 

The optic nerve is very well developed in the young 22 mm. long, and can be as 
readily followed to the brain as in normal eyes. It becomes proportionately more 
slender with age. It is observable even in the oldest eyes, but in them it has been 
impossible to trace the optic nerves outside the bulb. 


(By Henry H. Lane.) 

In the Biological Bulletin, vol. 6, No, i, December, 1903, the ovarian 
structures of Cuban cave fishes, Luci/uga and Stygicola, were described as 
minutely as the few specimens then at hand would allow. A much larger series 
of the ovaries of these fishes has since been put at my disposal. A study of these 
44 ovaries (21 of Lucifuga and 23 of Stygicola) enables me to correct some minor 
errors and to make some observations additional to those already recorded, and these 
are submitted as follows : 

A few terms may be defined for the sake of clearness. These are : 

oviduct: the unpaired duct leading from the ovary to the urogenital pore. 
It is not in Teleosts generally the homologue of the Mullerian duct of 
other vertebrates. 
ovisac: the anterior enlargement and continuation of the oviduct, covering 

the ovary proper. 
ovary: the organ containing the ova. It is, however, sometimes convenient 
to speak of the ovisac and ovary proper together simply as the "ovary." 
In such cases the context prevents ambiguity. 
stroma: the tissues of the ovary proper other than the ova and their follicular 


Externally the ovary is a Y-shaped, subcylindrical organ (plate 27) with a bilateral 
arrangement of the stroma. Its greatest diameter is usually immediately posterior 
to the point where the two horns begin. These horns of the ovary are right and left 
in position and may be long enough to inclose between them the posterior part of 
the stomach, though there is much variation in their length (plate 27), Within the 
ovisac the stem of the Y is divided by a median partition with which the ovarian 
structures proper are associated, in some ovaries more distinct than in others. This 
median sagittal partition extends posteriorly to the region I have chosen to consider 
the beginning of the oviduct, where in most cases only the part attached to the 
ventral wall persists ; in others the part attached to the dorsal wall is also present, 
though separated from the ventral part by a fissure. From the tips of the ovarian 
horns slender but comparatively strong threads of connective tissue, inclosing 
blood-vessels, run cephalad and fasten to the peritoneal walls, thus assisting in 
securely holding the ovary in position. Dorsally, a mesovarium suspends the organ 
from the peritoneal lining of the body cavity, while ventrally there is a correspond- 
ing attachment, the mesorectum. Each horn of the ovary is supported by its own 
fold of peritoneum and these two become united at or near the point of division of 
the horns and are continued posteriorly as the single, thicker mesovarium sup- 
porting the body of the ovary and the oviduct. The mesorectum is not always 
complete in the region of the ovarian horns. 

The oviduct, which has its external orifice at the urogenital pore, increases 
gradually in size as it extends forward toward the ovary and finally becomes the 
ovisac surrounding the ovary proper. 


The size of the ovary varies with the age and size of the females as well as the 
state of development of the embryo or ova within it. One specimen, a Lucifuga 
65 mm. long, had an ovary only 16 mm. long and 8 by 9 mm. in largest diameters, 
altho it contained 4 nearly ripe young, each 18 to 20 mm. in length. One non- 
pregnant Lucifuga,?>2, mm. long, had an ovary but 12 mm. in length. These meas- 
urements were made on preserved specimens. 

The point of division into the two horns is usually a little less than halfway from 
the anterior tip of the ovary to its posterior end. The two horns themselves are 
rarely equal in size, though there is no evidence of any tendency toward an unpaired 
condition through the " phylogenetic resorption" of one side such as Ryder found in 
Gambusia patruelis. 

The space between the ovisac and the inclosed ovary varies in size and is con- 
tinuous with the lumen of the oviduct. The growth of the young results in a grad- 
ual stretching of the ovisac, and to a certain extent of the oviduct also, so that near 
the close of gestation these structures are so extremely thin that their cellular nature 
can not be satisfactorily made out (plate 29, figs. a,b,c). Apparendy within a short 
time after the birth of the young, the ovisac and oviduct contract and reassume the 
form and appearance found in the ovaries of mature non-pregnant females. The wall 
of the ovisac is then quite thick (ov.s., plate 29, fig. d) and the lumen very small. 

The stroma of the non-pregnant ovaries is a bilateral mass which occupies most 
of the space within the ovisac. Its general shape resembles that of the ovisac, 
being fusiform in its main part, with its greatest diameter just posterior to the 
division into the two horns. The horns of the stroma are attached to the horns of 
the ovisac along their median surfaces (plate 29, fig. a), the whole stroma 
forming a V. In the stem of the V the stroma forms a median dorso-ventral 
partition within the ovisac, which is to be looked upon as representing the area of 
fusion of the originally distinct right and left ovisacs (fig. b). Near its posterior 
end this partition is cut across laterally by a fissure and the two prongs thus formed, 
one dorsal, the other ventral, gradually disappear toward the oviduct (figs, c and 
d) . Sometimes the dorsal one disappears first, sometimes both are equally extensive. 

The stroma has many somewhat pointed and comparatively large lobes, which 
are usually connected with the main mass by a slender "neck" of tissue. Dr. 
Eigenmann observed that these lobes are sometimes held in the mouth of the young 
fish during a part at least of its later development. Whether the young derives 
any nourishment from the lobes can not be stated with certainty. The whole stroma 
in the non-pregnant ovaries is distended by a large amount of lymph and adipose 
tissue contained in the sinuses described below, especially when approaching the 
reproductive period as shown by the maturity of the ova. 

The largest ova can be seen through the ovisac by the unaided eye as opaque 
white dots in the preserved specimens (plate 27). The follicles surrounding these 
ova usually lie some distance beneath the surface of the stroma and a tubular inden- 
tation of the epithelial covering of the latter extends down to the follicle (plate 29, 
fig. e). The blind end of this pit is so closely applied to the follicular membrane 
that it usually requires a very close inspection to discover its independence. It is 
then found that the follicular membrane at this place is only a single cell layer in 
thickness. Stuhlmann describes a similar indentation of the epithelium over the 
ova in the ovary of the viviparous blenny, Zoarces viviparus Cuvier. 


By the time the young are well advanced, i.e., i8 to 20 mm. long, the lymph 
sinuses of the stroma have mostly lost their contents and the stroma itself has 
become very greatly reduced and compressed into a narrow median wall (plate 29, 
fig. b). There are no "pockets" in which the young are carried as in Cymatogaster 
and other EmbiotocidcB (Eigenmann), or as in Anableps (Wyman), but instead the 
young attach themselves by the mouth to the ovarian lobes, or lie free within the 
lumen of the ovisac. 

The single oviduct, as well as the ovisac, is widely distended in pregnant females 
when the young are well advanced. In the non-pregnant females the duct is a 
cylindrical, thick-walled, muscular tube with numerous folds on its inner surface, 
which is covered with a layer of columnar epithelium similar in all respects to that 
of the ovisac. 


The Ovisac. — The ovisac, as noted, varies greatly in appearance, depending 
on the length of the pregnancy or the time since the close of that period. In nor- 
mal, non-pregnant ovaries it varies from 100 to 150 /n in thickness. Structurally 
it is composed of at least 4 cell layers. The outermost is the ordinary peritoneal 
layer continuous with the lining of the body cavity; second, a thicker layer of 
longitudinal muscle fibers which lie immediately below the peritoneal covering; 
third, a somewhat thicker band of transverse muscle fibers ; fourth, the inner lining 
of epitheUum, which, in some instances at least, contains numerous blood capillaries. 
In the case of pregnant females, the ovisac is more or less thinned through 
stretching, until, when the young are well advanced, the cell layers can scarcely be 
distinguished. (See plate 29, figs. A, B, c.) 

The Ovary. — The ovarian structures proper are highly vascular and much lobed. 
In some instances the egg foUicles are surrounded by a network of blood capillaries. 
The greater part of the stroma is split up into numerous sinuses (st., plate 28), 
many of which are larger than any of its blood-vessels. The epithelial layer cover- 
ing the stroma frequently contains numerous capillaries each with a diameter of 

5 to 8 /i (plate 29, fig. r). In some instances these capillaries are very numerous ; 
in others, they are scarcely, if at all, perceptible. This difference is due to the 
diflferent degrees of distention of the stroma by the lymph in its sinuses. The 
stroma itself consists of a mass of connective tissue and non-striated muscle fibers 
in which are embedded the ova in various stages of development. 

The Follicle. — The smallest ova (5 to 10 fi in diameter) have no trace of a follic- 
ular membrane around them individually. Somewhat larger ova (100 to nearly 
400 fi in diameter) are surrounded by a single layer of elongate cells, quite similar 
to the stroma cells. In the case of more mature ova (over 400 /a in diameter) there is 
a distinct follicle consisting of a single layer of appressed quadrangular cells about 

6 /i in depth ; outside of this is a layer of somewhat irregular cells, in many cases 
surrounding blood capillaries 6 to 10 /x in diameter. The thecal wall outside the 
capillary layer consists of from i to 3 cell layers of long, spindle-shaped cells 
resembling those composing the stroma itself. The medium-sized ova (about 
400 /x in diameter) He close beneath the surface of the ovary (0., plate 29, fig. d), but 
the largest ova (600 fi and over in diameter) arc usually found rather deeply em- 
bedded within the stroma, except for the tubular indentation from the surface of 


the latter which reaches down to the follicle and possibly aflfords later a means of 
entrance of the spermatozoa to the mature egg. 

The Ova. — In the smaller ova (under 400 /u, in diameter), the nucleus is usually 
quite distinct and has approximately one-third the diameter of the whole ovum. 
The cytoplasm, not yet deeply laden with deutoplasm, has usually a reticulated or 
alveolar appearance. In the larger ova (over 400 /x in diameter), the cytoplasm 
becomes more and more heavily laden with deutoplasm, until in the largest it is 
almost wholly obscured by the latter. The nucleus at the same time becomes cor- 
respondingly more difficult to find, not increasing much in size as the ovum develops. 
No traces of maturation were detected in even the largest ova found (750 /* and over 
in diameter). 

More or less deeply within the stroma the ova arise in masses of several hundred 
ova each. In size the smallest discernible ova measure from 5 to 10 /x in diameter 
and have well-defined nuclei {s.o., plate 28, fig. a). As development proceeds a 
number of ova in each "nest" may increase more than the others; at a later stage 
it can be seen that a few of these are gaining on their fellows ; and still later one is 
seen to be outstripping all the others in that "nest." Sometimes one {l.o., plate 28, 
fig. a) gains the ascendency so early that the remainder {s. 0., plate 28, fig. a) never 
show any marked increase or difference in size among themselves. In any case, 
in the final stage of development, a single ovum is left in the ''nest," and this now 
seems to migrate till it rests just beneath the surface of the ovary itself (0., plate 28, 
fig. c). Where several large ova seem to have been developed in one "nest," close 
scrutiny, at least in the case of the less fully developed ova, invariably reveals a 
separating layer of very thin, elongated stromal cells (st.c, plate 28, fig. f) such as 
originally surrounded the whole "nest" (compare with st.c, plate 28, figs. A or 
d), thus showing that these larger, closely adjacent ova are derivatives each 
from an originally distinct "nest." 

The fate of the other ova which at first lay in the same "nest" vrith the larger 
ovum is a question of interest. Two possibilities suggest themselves: either the 
growing ovum absorbs the neighboring ova into its own substance, or they disinte- 
grate in situ without becoming a part of the larger ovum. 

Certain of the larger ova very strongly suggest the first possibility. In these 
the smaller ova are grouped together at the side of the much larger one, or may even 
surround it, and are apparently undergoing a greater or less amount of disintegra- 
tion. In the case shown (a. 0., plate 28, fig. b) this disintegration has gone so far that 
the outlines of the small ova are quite indistinct and in some cases apparently only 
the nucleus remains, and this, too, is no more than an irregular mass of chromatin. 

The atrophy of the small ova is evidently a rapid one, for there is no sign of any 
pigmentation or other mark of a gradual degeneration of the cells. Moreover, wher- 
ever a "nest" contains a larger ovum and smaller atrophying ova around it, the 
cj^oplasm of the ova is confluent, at least in the case of those most advanced in dis- 
integration. This in itself is good evidence of the assimilation of the degenerating 
ova by the larger one. In short, there is here a struggle for existence among the ova 
of each ' ' nest.'' ' The successful ovum either produces a rapid degeneration of the sur- 
rounding ova, or taking advantage of such a condition produced in them by some 
unknown factor, assimilates them into its own substance. It is hard to deter- 
mine what is the all-important cause of the initiation of the more rapid growth of 


the superior ovum, but one possibility is that of a more fortunate situation in regard 
to the source of nutrition. In most cases, if not in all, the ova which gain in size 
over the others lie in such a position in the "nest" as to be more nearly in contact with 
the stromal blood capillaries than the others, and this very likely furnishes the expla- 
nation of the phenomenon noted. 

In case, for any reason, no ovum in a "nest" develops, all the ova in that "nest" 
undergo a slow, pigmented degeneration, or atrophy. The evidence for this lies in 
the presence within the stroma of masses of yellowish-brown cells which do not stain 
with haematoxylon. For convenience I shall speak of them as the "yellow cells." 
Their nuclei may be evident only as small, deeply staining masses of chromatin ; 
or the chromatin may have the form of a more or less definite spireme ; but in many 
cases the nuclear material shows signs of karyorrhexis, being decidedly broken up 
and apparently migrating into the cytoplasm of the cell. A reference to plate 28, 
D and E, will show that these cells can not be red blood corpuscles, for, not only do 
they not lie within a blood-vessel, but they are also more than twice as large as 
undoubted red blood corpuscles in the same section. They have none of the 
characters of leucocytes or phagocytes, but they do exhibit the typical brown or 
yellowish pigmentation of degenerating epithelial cells. 

As shown {st.c, plate 28, fig. d), the yellow cells lie within a space surrounded 
by exactly the same sort and arrangement of stromal cells as those which inclose 
undoubted "nests" of ova. Hence they must be regarded as either degenerating 
ova or cells which have taken the place of ova. As already stated, they have none 
of the characters of phagocytes. One other possibility is suggested by the fact of 
the viviparity of these fishes ; that is, the possibility that these yellow cells may rep- 
resent a sort of corpus luteum. Aside from the structure of these cells, which do not 
have more than a very faint resemblance to the lutein cells of mammalian corpora 
lutea, one consideration very effectually disposes of this possibility ; namely, the fact 
that these yellow cells do not occur in the larger ovaries, i.e., in those of the more 
mature females, but on the contrary they occur in the smaller and even the smallest 
ovaries at hand. They certainly can not therefore be of the nature of corpora lutea 
cells. That they are degenerating ova seems to me the most probable conclusion, 
for the following reasons: 

(a) The "yellow cells" occur only in masses, exactly similar, in point of 

number and size of cells as well as in position in the stroma, to the 

"nests" of young ova. 
{b) The masses of "yellow cells" are surrounded by the same sort and 

arrangement of stromal cells as surround the "nests" of ova. 
(c) The "yellow cells" exhibit the typical brown pigmentation of slowly 

atrophying epithelial cells. 
{d) There are no cells of sufficient size in these ovaries which could have 

these characters except degenerating ova in "nests." 
These fishes are undoubtedly descended from oviparous forms, and viviparity 
is probably a comparatively recent acquirement, though most probably attained 
before the change of habitat from the sea to the underground streams of Cuba. 
Some at least of the marine members of the Brotulidae are also viviparous. 
The production of the many "nests," each with its hundreds or even thousands 
of young ova, is a reminiscence of the oviparous condition, when it was necessary for 


the preservation of the species that a multitude of young be produced, as in the 
case of the oviparous fishes. The condition of viviparity, providing as it does 
for the greater safety of the young during the most critical period of their 
development, and their habitat in caves where the number of enemies is prob- 
ably greatly less than in the sea enable these species to maintain themselves by 
the production of fewer offspring. 


1. The ovary in Lucifuga and Stygicola consists of a mass of stroma containing 
the ova and covered with epithelium ; the whole structure is V-shaped and is con- 
tained within the ovisac; the latter is continued to the urogenital pore as the oviduct. 

2. The epithelium, lining the ovisac and covering the ovary proper, is unique 
in that it frequently contains numerous blood capillaries. 

3. The sinuses within the stroma are filled with lymph and adipose tissue. 

4. Lucifuga and Stygicola are viviparous blind fishes which give birth to but 
few young, 2 to 15 so far as yet observed. 

5. The young are not developed in separate sacs, but lie within the lumen of 
the ovisac, gradually compressing the ovarian stroma as they develop. 

6. The ova arise in "nests" or masses of several hundred each. The smallest 
observed have a diameter of 5 to 10 /x. 

7. One ovum from each "nest" is developed to maturity; the other ova of the 
"nest" undergo rapid degeneration and are ultimately absorbed into the substance 
of the large ovum. 

8. In those "nests" in which none attains maturity, all the ova undergo a slow, 
pigmented degeneration in situ. 

9. The destruction of so many ova at an early stage is an adaptation to the vivipa- 
rous habit. 

ID. Viviparity is probably a comparatively recent acquirement of these fishes, 
though attained before these genera left the sea for the fresh-water cave streams. 



1. Lucifuga and Stygicola are two marine fishes that have remained in the 
cracks and caves of the coral beaches which they inhabited, as these caves were 
elevated and became filled with and enlarged by fresh water. They have become 
entirely adjusted to a fresh-water environment. 

2. Stygicola is known from both the north and south slopes from Alacranes to 
Matanzas and Alfonso XII. Lucifuga is known only from the south slope west of 

3. The caves in which the fishes were found are all well lighted, but are always 
connected with dark underground channels. Each cave has only a limited supply 
of fishes that may be replenished from the underground channels. 

4. Lucifuga and Stygicola are negatively heliotropic. They are adjusted to 
withstand but slight temperature changes. They feed on crustaceans and odonata 

5. Both species are viviparous, giving birth to 2 to 15 young about 25 mm. long. 
Both probably breed throughout the year. Spermatozoa are transferred long before 
the ripening of the eggs. Lucifuga breeds probably most abundantly through 
March and May in shallow places. Its young are abundant near the surface. 
Stygicola breeds in unknown places and its young are not seen near the 

6. The eye decreases in size progressively from birth to extreme old age con- 
comitantly with the appearance of masses of pigment cells in the orbital fat. 

7. The eye varies greatly in diflferent individuals of the same size — from 260 
to 425 /A in length, in brothers and sisters in the same ovary. 

8. The ontogenetic degeneration results either in the shriveling of the entire 
structure or the great distention of the pigmented layer. One process may be found 
on one side, the other on the other side of the same individual. 

9. The eye muscles are all present in the young, but undergo a variable amount 
of degeneration with age, disappearing entirely in very old of Stygicola. 

10. The sclera is self-determining in both Lucifuga and Stygicola. In Luci- 
fuga the cartilages at the time of birth are too large for the eye, forming a shield 
over the face of the eye. In Stygicola it forms a ring about the middle of the eye. 
After birth they very rapidly degenerate and disappear entirely by the time Lucifuga 
has reached less than half its maximum length. In Stygicola it remains longer. 

11. There is evidence that there is an early disturbance of the vascular system 
of the eye resulting in the formation of large blood lakes about the eye. 

12. The lens has begun to degenerate before birth. Its contents liquefy, the cap- 
sule shrivels, and finally disappears at a length of about 40 mm. 

13. It has not been determined when the histogenesis of the retina ends and its 
degeneration begins. The most highly developed retina was found in an unborn 
young of Lucifuga 20 mm. long. In this retina the outer nuclear, outer reticular, 
inner nuclear, inner reticular, and ganglionic layers are more or less distinctly 



Ovaries of Lucifuga and Stygicola. 

A. Dorsal aspect. Round, opaque, while dots are larger ova seen through 

stroma and ovisac. 

B. Ventral aspect. X 2 diameters. 




'-^ — ■ — a»=^ 



v/. c 

A. "Nest ' of small ova (s. c), each about 10 /^ in diameter, one larger ovum (/.o.) 

SOa* in diameter. Whole nest contained within special arrangement of stromal 
cells (st. c). X about 300 diameters. 

B. Developing ovum (m. o.). surrounded by rapidly atrophying small ova {a.o.); 

I., lumen of ovisac; n., " germinative spot"; st., stroma. Diameter of large ovum, 
120 yit. X 500 diameters. 

C. Cross-section of one horn of ovary. /., lumen of ovisac ; o., ova ; ov. s., ovisac ; 

St., stroma. Guide line to st. crosses place of attachment of stroma to median wall 
of horn of ovisac. X about 50 diameters. 

D. "Nest of yellow cells." Diameter of individual "yellow cells" (j). c.) about 15 fJ-sl. c, 

arrangement of stromal cells around yellow cells, as around " nest " of small ova 
(st. c, in fig. A.) X 200 diameters. 

E. Few "yellow cells" more highly magnified to show pigment-granules and general 

appearance of slow degeneration. Nuclei can not be distinguished in photo- 
graph, though distinct enough in section. X about 800 diameters. 

F. 3 adjacent " nests " of ova (r?.), each with developing ovum ; st. c , stromal cells which 

separate " nests " and likewise developing ova from one another. X 2 1 diameters. 




Sections of Ovaries. 

A. Cross-section through horns of pregnant ovary. 

B. Cross-section through middle part of pregnant ovary. Ovisac collapsed when fetuses were removed. 

C. Cross-section through posterior part of pregnant ovary. 

D. Ooss-section of non-pregnant ovary with stroma in two lobes, one dorsal, other ventral. 

E. Part of cross-section of non-pregnant ovary. 

F. Part of epithelial covering of non-pregnant ovary showing capillaries (c;m). 





It may now be profitable to take up the causes leading to the small degree of 
degeneration found in Chologaster, the degeneration of the eye in Amhlyopsis, 
Typhlichthys, and Troglichihys to a mere vestige, together with the total disap- 
pearance of some of the accessory structures of the eye, as the muscles, in some of 
the species. In the outset of this consideration we must guard against the almost 
universal supposition that animals depending on their eyes for food are or have been 
colonizing caves, or that the blind forms are the results of catastrophes that have 
happened to eyed forms depending on their eyesight for their existence. This idea, 
so prevalent, vitiates neariy -everything that has been written on the degeneration 
of the eyes of cave animals. 

The degeneration of organs ontogenetically and phylogenetically has received 
a variety of explanations. 

(i) The organ diminishes with disuse (ontogenetic degeneration, Lamarck, 
Roux, Packard) and the effect of this disuse appears to some extent in the next gen- 
eration (phylogenetic degeneration, Lamarck, Roux, Packard). 

(2) Through a condition of panmixia the general average maintained by selection 
is reduced to the birth mean in one generation (ontogenetic, Romanes, Lankester, 
Lloyd Morgan, Weismann) to the greatest possible degeneration in succeeding 
generations (phylogenetic, Weismann), or but little below the birth average of.'the 
first generation (Weismann's later view, Romanes, Morgan, Lankester). 

(3) Through natural selection (reversed) (the struggle of persons) the organ may 
be caused to degenerate either (a) by the migration of persons with highly developed 
eyes from the colony living in the dark (Lankester), or (b) through economy of 
weight and nutriment or liability to injury (phylogenetic purely, Darwin, Romanes). 

(4) Through the struggle of parts (a) for room an unused organ in the individual 
may be crowded (ontogenetic, Roux), (b) for food, this may lead to the development 
of the used organ as against the disused through a compensation of growth (Goethe, 
St. Hilair, Roux) ; this ontogenetic result becomes phylogenetic through transmis- 
sion of the acquired character (Roux), or is in its very nature phyloblastic (Kohl). 

(5) Through the struggle between soma and germ to produce the maximum 
efficiency of the former with the minimum expenditure of the latter (ontogenetic 
and phylogenetic, Lendenfeld). 

(6) Through germinal selection, the struggle of the representatives of organs 
in the germ (ontogenetic and phylogenetic, Weismann). 

(7) To these special considerations should be added the recently suggested gen- 
eral process of mutation. 

The idea of ontogenetic degeneration is intimately bound up with the idea of 
phylogenetic degeneration. Logically we ought to consider first the causes of indi- 
vidual degeneration and then the processes or causes that led to the transmission 



of this. Practically it is impossible to do so, because many of the explanations are 
general. Only number (4) of the above may be taken in the ontogenetic sense purely, 
though it was certainly also meant to explain phylogenetic degeneration. In many 
of the explanations of particular cases of degeneration more than one of the above 
principles are invoked, though only one was meant to be used. In most cases, how- 
ever, the discussions of degeneration have been in general terms, without direct 
bearing on any specific instance of degeneration in all its details. It must be evident 
that such discussions can only by accident lead to right results. 

By the Lamarckian ontogenetic degeneration is considered the result of lack of 
use and consequent diminished blood supply. The results of the diminution caused 
by the lack of use during one generation are transmitted in some degree to the next 
generation, which thus starts at a lower level. A continuation of the same con- 
ditions leads finally to the great reduction and ultimate disappearance of an organ. 

No one, so far as the author knows, has attempted, or, perhaps better, suc- 
ceeded, in accounting with this factor in detail for the degeneration of the eye. 
Packard's explanations are evidently a mixture of Lamarckism and Darwinism. 

Packard says, "When a number, few or many, of normal seeing animals enter 
a totally dark cave or stream, some may become blind sooner than others, some hav- 
ing the eyes slightly modified by disuse, while others" may have in addition physi- 
cal or functional defects, especially in the optic nerves and ganglia. " The result of 
the union of such individuals and of adaptation to their stygian life would be broods 
of young, some with vision unimpaired, others with a tendency to blindness, while 
in others there would be noticed the first steps in degeneration of nervous power 
and nervous tissue." Packard evidently had invertebrates in mind. He clearly 
admits the cessation of selection or panmixia which is implied by his supposition 
that those born with defects may breed with the others. He supposes that the 
blind fauna may have arisen in but few or several generations, a supposition 
that may be applicable to invertebrates, but certainly may not be applied to the 
vertebrates. At first those becoming so modified that they can do without the 
use of their eyes would greatly preponderate over those " congenitally blind." 
" So all the while, the process of adaptation going on, the antennae and other tactile 
organs increasing in length and in the delicacy of structures, while the eyes were 
meanwhile diminishing in strength of vision and their nervous force giving out, 
after a few generations, perhaps only two or three, the number of congenitally 
blind would increase, and eventually they would, in their turn, preponderate in 
numbers." Packard seems here to admit the principle of degeneration as the 
result of compensation of growth, the nervous force of the eye giving out with the 
increase of the tactile and olfactory organs. It is somewhat doubtful in what sense 
the term " congenitally blind " is used, but it probably means born blind as the 
result of transmitted disuse rather than blind as the result of fortuitous variation. 
The effects of disuse are thus supposed through their transmission to have given 
rise to generations of blind animals. The continued degeneration is not discussed. 

In 1873, 1874, and 1890, Romanes, in a series of articles in "Nature" and later 
in "Darwin and after Darwin," n, page 291 et seq., maintained that the beginning 
of degeneration is due to cessation of selection, and continued degeneration to the 
reversal of selection and final failing of the power of heredity. Selection he supposed 
to be reversed because the organ no longer of use "is absorbing nutriment, causing 


weight, occupying space, and so on, uselessly. Hence, even if it be not also a source 
of actual danger, economy of growth will determine a reversal of selection against 
an organ which is now not only useless, but deleterious." This process will con- 
tinue until the organ has reached " so minute a size that its presence is no longer 
a source of detriment to the organism, the cessation of selection will carry the reduc- 
tion a small degree further; and then the organ will remain as a 'rudiment.'" 
Since, however, we can not consider that the force of heredity is everlasting, it will 
eventually fail and the organ dwindle still further and disappear. This failure 
of heredity, Morgan (" Animal Life," page 793) is unable to distinguish from the 
effect of disuse without which " the reduction of organs is difficult to explain." 

The principles involved in this explanation are panmixia natural selection, and, 
according to Morgan, disuse transmission. 

Weismann ("Nature," 1886, and " Essays," vol. 11, i) contended that cessation 
alone, or panmixia as he terms it, is sufficient to account for all degeneration. He 
later gave up this view for his theory of germinal selection, of which more later. 

Roux, starting with the then generally accepted view that acquired characters 
are transmitted, attempted chiefly to explain degeneration in the individual. 
Degeneration is looked upon as the result of a struggle among the parts for 
(a) room and (b) food. He emphasizes the fact that a reduced functional activity 
continued for a long period reduces the functional possibility of an organ (page 
176). The diminished use not only brings about this simple atrophy, but also 
the reduction, by stronger neighbors, to such a volume as is still of advantage to the 
animal. Disused organs that are not in the struggle for room may maintain them- 
selves a long time. The struggle among parts for food, which implies the principle 
of compensation of growth of Goethe, need not take place through the withdrawal 
of blood, but may take place through the more active osmotic selection by the 
stronger organ of food that would otherwise go to the weaker. 

Without doubting that both these principles are active agents in degeneration, it 
may be seriously doubted whether they were effective in the degeneration of the 
eyes in question. Certainly there can be no question of a struggle for room, for 
the position and room formerly occupied by the eye is now filled with fat which 
can not have been operative against the eye. The presence of this large fat-mass 
in the former location of the eye, the large reserve fat-mass in the body, the uni- 
formly good condition of the fish, and the low vitality which enables them to live 
for months without visible food, all argue against the possibility that the struggle for 
food between parts was an active agent in the degeneration of the eyes. 

Kohl considers that "Der Grund, und direkter oder indirekter Anlass zum 
Eintreten der Entwickelungshemmung ist Lichtmangel." The method of the direct 
operation of the lack of light he conceived to be as follows : The ancestry of blind 
animals lived where the light was uninterrupted and they had developed eyes. 
They got into an environment where the light was shut ofi" more or less. The first 
generations retained their fully developed eyes without, however, being able to 
put them to full use. In consequence during phylogeny other organs became 
highly developed to compensate for the disuse of the eye. (Through natural selec- 
tion?) Thus touch organs (Myxine, Siphonops) or the auditory organs (Talpa 
and possibly Typhlichthys) became more highly developed. The eye was unneces- 
sarily highly developed. A process of degeneration (Riickbildung) began, which 


was never very extensive. Much more potent in placing the eye in harmony with 
its environment was the fact that every succeeding generation developed its eye less. 
This process of Hemmung of the eye did not begin until the developmental force 
began to go to the development of the compensating organs. On account of the 
loss of this developmental force the eye was unable to reach, in successive genera- 
tions, the former grade. The degeneration is thus explained as the result of a 
struggle of parts, although this term is nowhere used, acting through the princi- 
ple of compensation. The same objections may be offered to this explanation 
of Kohl as to all his theoretical discussions ; they are based on the assumption 
of conditions and processes that have no existence. The high development of 
" compensating " organs is not primarily the result of the loss of the eye, but the 
high development of the former organs permitted the disuse and later degeneration 
of the later. His whole process is a phylogenetic one without a preceding onto- 
genetic one, though on this point he does not seem to be very clear himself, for on 
one page we are told that degeneration leads to retardation, and on another that 
degeneration is a consequence of retardation. 

Lendenfeld endeavors to apply Roux's Kampf der Theile with reversed selec- 
tion to explain the conclusions reached by Kohl on the processes and causes of 
degeneration. The struggle is represented to take place between the germ and 
soma, the former endeavoring to keep the latter at the lowest efficient point as 
weapon for the germ. If a series of individuals gets into the dark, the organs of 
vision are of no advantage, and reversed selection will bring about their degenera- 
tion. The saving in ontogeny appears first as a retardation and then a cessation 
of development. 

Weismann later accepted the view of Romanes, Morgan, and Lankester of 
the inadequacy of panmixia to explain the whole phenomena of degeneration, 
and in his " Germinal Selection " rejects the idea of reversed selection and suggests 
a new explanation for what Romanes attributed to the failure of heredity and the 
Lamarckians to disuse transmission. The struggle of the parts, of Roux, has 
been crowded back by him to the representatives of these parts in the germ. 

"The phenomena observed in the stunting, or degeneration, of parts rendered 
useless show distinctly that ordinary selection, which operates by the removal of 
entire persons, personal selection, as I prefer to call it, can not be the only cause 
of degeneration ; for in most cases of degeneration it can not be assumed that slight 
individual vacillations in the size of the organ in question has possessed selective 
value. On the contrary, we see such retrogressions affected apparently in the 
shape of a continuous evolutionary process determined by internal causes, in the 
case of which there can be no question whatever of selection of persons or of a 
survival of the fittest, that is of individuals with the smallest rudiments." The 
gradual diminution, continuing for thousands and thousands of years and cul- 
minating in its final and absolute effacement, can only be accomplished by ger- 
minal selection. Germinal selection as applied to degeneration is the formal 
explanation of Romanes' failure of the hereditary force and the establishment of 
disuse effects in the heredity through the struggle of parts for food. "Powerful 
determinants will absorb nutriment more rapidly than weaker determinants. The 
latter, accordingly, will grow more slowly and will produce weaker determinants 
than the former." If an organ is rendered useless, the size of this organ is no longer 


an element in personal selection. This alone would result in a slight degeneration. 
Minus variations are, however, supposed to rest "on the weaker determinants of 
the germ, such as absorb nutriment less powerfully than the rest. This will enable 
the stronger determinants to deprive them even of the full quantum of food cor- 
responding to their weakened capacity of assimilation and their descendants will 
be weakened still more. Inasmuch now as no weeding out of the weaker deter- 
minants of the hind leg (eye) by personal selection takes place on our h}'pothesis, 
inevitably the average strength of this determinant must slowly but constantly 
diminish, that is, the hind leg (eye) must grow smaller and smaller imtil it finally 
disappears altogether." "Panmixia is the indispensable precondition of the whole 
process; for owing to the fact that persons with weak determinants are just as 
capable of life as those with strong, solely by this means is a further weakening 
effected in the following generations." 

This theory presupposes the complex structure of the germplasm formulated 
by Weismann. But granting Weismann the necessary structure of the germplasm, 
can germinal selection accomplish what is claimed for it? I think not. Grant- 
ing that variation occurs about a mean, would not all the effects claimed for minus 
variations be counteracted by positive variations? Eye determinants, that on 
account of their strength secure more than their fair share of food and thereby 
produce eyes that are as far above the mean as the others are below, may leave 
descendent determinants that are still stronger than their ancestry. It is evident 
that a large, really extravagant development of the eye in such a fish as Chologaster 
would not effect the removal of the individual by personal selection, still less so in 
Amblyopsis, which not only lives in comparative abundance, but has lived for 20 
months in confinement without visible food. It seems that all the admitted objec- 
tions to degeneration by panmixia apply with equal force to germinal selection. 
This, however, would be changed were the effect of disuse admitted to affect the 
determinants, and this it seems Weismann has unconsciously admitted. So far 
we have considered germinal selection in the abstract only. In the concrete we 
find that degeneration is not a horizontal process affecting all the parts of an organ 
alike as Weismann presupposes, not even a process in the reverse order of phyletic 
development, but the more vital, most worked parts degenerate first with disuse 
and panmixia, the passive structures remain longest. The rate of degeneration 
is proportional to the past activity of the parts and the statement that "passively 
functioning parts, that is, parts which are not alterable during the individual life 
by function, by the same laws also degenerate when they become useless" is not 
applicable to the eyes. As one example of the unequal degeneration we need only 
call attention to the scleral cartilages and the rest of the eye of Troglichthys roscB.^ 

All are agreed that natural selection alone is insufficient to explain all, if any, 
of the processes of degeneration. All either consciously or not admit the principle 
of panmixia, and all are now agreed that this process alone can not produce exten- 
sive degeneration. All are agreed that the important point is degeneration beyond 
the point reached by panmixia, the establishment of the degenerating process, what- 
ever it may be, in the germ, or in other words, breaking of the power of heredity. 
It is in the explanation of the latter that important differences of opinion exist. 

* I must again guard against cross-counter conclusions. In the Brotulidx the passive cartilages are among the 
first things to go. 


Weismann attempts to explain the degeneration beyond the point which pan- 
mixia can reach by a process which not only is insufficient, if all his premises are 
granted, to produce the desired result without the help of use transmission, but 
has as its result a horizontal degeneration which does not occur in the eyes. 

Romanes supposed degeneration, beyond the point which may be reached by 
panmixia, to be the result of personal selection and the failure of the hereditary 
force. The former is not applicable to the species in question and is denied by 
such an ardent Darwinian as Weismann to be applicable at all in accounting for 
degeneration. Moreover the process as explained by Romanes would result in a 
horizontal degeneration which has no existence in fact. The second assumption, 
the failure of hereditary force, is not distinguishable, as Morgan has pointed out, 
from the effect of use transmission. 

The struggle of parts in the organism has not affected the eye through the lack 
of room, since the space formerly occupied by the eye is now filled by fat and not 
by an actively functioning organ. It is not affected by the struggle for food, for 
stored food occupies the former eye space. It could only be affected by the more 
active selection of specific parts of food by some actively functioning organ. It is 
possible that this has in fact affected the degeneration of the eye. The theory 
explains degeneration in the individual and implies that the effect in the individual 
should be transmitted to the next generation. This second fact seems but the 
explanation of the working of the Lamarckian factor. 

Mutation can produce definitely directed evolution such as we find in the 
degenerating eye only when each step, each successive mutation, has an advantage 
over the mother or sister lines. I do not think that any one after familiarizing 
himself with the variation of the eye and its insignificance will maintain that this 
minute organ is now or has been for many generations of selective value. If it is 
not of selective value, mutation is as powerless to account for its condition as is 
natural selection of favorable variations. 

The eyes of the two sides vary so much, independent of each other, that we 
are forced to conclude that there has been no check on their variation for a long 

The only answer to the objection that the eyes are not the result of personal 
selection is that they may be so correlated with another organ inversely propor- 
tionate to it, that the selection of individuals with this other organ in favorable 
condition carries with it the selection of individuals with the eye in decreasingly 
imperfect condition. No such organ is available. 

The Lamarckian view, that through disuse the organ is diminished during the 
life of the individual, in part at least on account of the diminution of the amount 
of blood going to a resting organ, and that this effect is transmitted to succeeding 
generations, not only would theoretically account for unlimited progressive degenera- 
tion, but is the only view so far examined that does not on the face of it present 
serious objections. Is this theory applicable in detail to the conditions found in 
the Amblyopsidae ? Before going farther, objections may be raised against the 
universal assumption that the cessation of use and the consequent panmixia was a 
sudden process. This assumes that the caves were peopled by a catastrophe. But 
it is absolutely certain that the caves were not so peopled, that the cessation of 
use was gradual and the cessation of selection must also have been a gradual pro- 


cess. There must have been ever widening bounds within which the variation of 
the eye would not subject the possessor to elimination. 

Chologaster is in a stage of panmixia as far as the eye is concerned. It is true 
the eye is still functional, but that the fish can do without its use is evident by its 
general habit and by the fact that it sometimes lives in caves. 

The present conditions have apparently existed for many generations, as long 
as the present habits have existed, and yet the eye still maintains a higher degree 
of structure than reversed selection, if operative, would lead us to expect, and a 
lower degree than the birth mean of fishes depending on their eyes — the condi- 
tion that the state of panmixia alone would lead us to expect. There is a staying 
qualify about the eye with the degeneration, and this can only be explained by the 
degree of use to which the eye is subjected. 

The results in Chologaster are due to panmixia and the limited degree of use 
to which the eye is put. Chologaster agassizii shows the rapid diminution of the 
eye with total disuse. 

The difference in the conditions between Chologaster and Amblyopsis, Typh- 
lichthys and TrogUchthys is that in the former the eyes are still in use, except 
when living in caves ; in the latter they have not been in a position to be used for 
hundreds of generations. The transition between conditions of possible use and 
absolute disuse may have been rapid with each individual after permanently enter- 
ing a cave. Panmixia, as regards the minute eye, continued. Reversed selection 
was inoperative, for economy can not have affected the eye for reasons already 
stated. Simply the loss of the force of heredity, unless this was caused by disuse 
or the process of germinal selection, can not have brought about the conditions, 
because some parts have been affected more than others. 

Considering the parts most affected and the parts least affected, the degree of 
use is the only cause capable of explaining the conditions. Those parts most 
active during use are the ones reduced most, viz., the muscles, the retina, optic 
nerve, and dioptric appliances, the lens and vitreous parts. Those organs occupy- 
ing a more passive position, the scleral cartilages, have been much less affected 
and the bony orbit least. The lens is one of the latest organs affected, and not 
at all during use, possibly because during use it would continually be in use. It 
disappears most rapidly after the beginning of absolute disuse both ontogenetically 
and phylogenetically. All indications point to use and disuse as the effective agent 
in molding the eye. The process does not, however, give results with mathe- 
matical precision. In Typhlichthys suhterraneus the pigmented layer is affected 
differently from that of Amblyopsis . The variable development of the eye muscles 
in different species would offer another objection if we did not know of the variable 
condition of these structures in different individuals. Chilton has objected to the 
application of the Lamarckian factor to explain degeneration on account of the 
variable effects of degeneration in various invertebrates. But such differences in 
the reaction are still less explicable by any of the other theories. 

University of Toronto 

Acme Library Card Pocket 
Under Pat. "Rel. Index FUe"