The Biological bulletin

BIOLOGICAL BULLETIN

OF THE

flDarine Biological Xaborator\>

WOODS ROLL, MASS.

Editorial Staff.

E. G. CONKLIN — The University of Pennsylvania,
JACQUES LOEB — The University of California.
T. H. MORGAN — Bryn Mawr College.

W. M. WHEELER — American Museum of Natural
History.

C. O. WHITMAN — The University of Chicago.
E. B. WILSON — Columbia University,

] •

/

Efcitor.

FRANK R. LILLIE — The University of Chicago,

VOLUME V

WOODS HOLL, MASS.
JUNE, 1903, TO NOVEMBER, 1903,

?

PRESS Of

7nE NEW ERA PRIMIHC Conp*«i'
LAKCASIER. PA

CONTENTS OF VOL V.

No. T. JUNE, 1903

PA<;K

AXEL LEONARD MELANDER AND CHARLES THOMAS BRUES :

Gi/esfs and Parasites of the Burrowing Bee Halictus i

J. B. JOHNSTON: The Origin of the Heart Endotheliu m in Amphibia 28^

J. W. SCOTT : Periods of Susceptibility in the Differentiation of

Unfertilized Eggs of Amphitrite t 35

ARTHUR W. GREELEY : Further Studies on the Effect of Varia-
tions in the Temperature on A nimal Tissues 42

BENNET M. ALLEN: The Embryonic Development of the Ovary

and Testis of the Mammalia (Preliminary Account} 55

No. 2. JULY, 1903

HENRY LESLIE OSBORN : Bunodera corniita sp, nov. : A New
Parasite from tlie Crayfish and Certain Fishes of Lake
Chautauqua, N. Y. 63

]. B. JOHNSTON : On the Blood Vessels, their Valves, and the

Course of the Blood in Lumbricits 74 '

VERNON L. KELLOGG: Two New Genera of MaJlophaga 85

FRANK R. LILLIE : Experimental Studies on tlie Development of

the Organs in the Embryo of the Foivl ( Gallus domesticus} 92 >

W. C. CURTIS : Cross obothrium laciniatuin and Developmental

Stimuli in the Cestoda 125 ^

No. 3. AUGUST, 1903

S. J. HUNTER : On the Conditions Governing the Production of

Artificial Parthenogenesis in Arbacia 143

ESTHEK F. BYRNES : Heterogony and Variation in some of the Copc-

poda of Long Island. 152

CASWELL GRAVE : On the Occurrence among Echinoderms of Larvce
with Cilia Arranged in Transverse Rings, with a Sugges-
tion as to their Significance 169

iii

\

IV CONTENTS.

No. 4. SEPTEMBER, 1903

MAULSBY W . BLACKMAX : The Spermatogenesis of the Myriapods.
II. On the Chromatin in the Spermatocytes of Scolopendra
heros ................................................................. 187

HELEN DEAN KING : The Effects of Heat on the Development of .

the Toad' 's Egg .................................................... 218

THOMAS H. MONTGOMERY, JR.: On Floscularia conklini, nov.
spec., with a Key for the Identification of the Known
Species of the Genus ................................. ............ 233

No. 5. OCTOBER, 1903

C. M. CHILD: Form Regulation in Cerianthiis ...................... 239

EFFA FUNK MUHSE : The Eyes of the Blind Vertebrates of North
America. VI. The Eyes of T^phlops Lumbricalis (Lin-
me/is'), a Blind Snake from Cuba ............................ 261

ANNIE E. PRITCHETT: Some Experiments in Feeding Lizards witJi

Protectively Colored Insects ...................................... 271

S. J. HOLMES : Sex Recognition Among Amphipods .................. 288

T. H. MORGAN : Regeneration of the Leg of Amphiiima means... 293

No. 6. NOVEMBER, 1903

H. F. THACHER : Absorption of the Hydra nth in Hydroid Polyps.. 297

C. M. CHILD : Form Regulation in Cerianthiis ....................... 304

ADELE M. FIELDE : Artificial Mixed Nests of Ants ................ 320

ADELE M. FIELDE : A Cause of Feud between Ants of the Same

Species Living in Different Communities ...................... 326

|. F. GARBER : Dimorphism in Blissus leucoptcrus ................ 330

RAYMOND PEARL : On two Cases of Muscular Abnormality in f/ie

Cat ................................................................. 336

Vol. V. June, ipoj. No. i

BIOLOGICAL BULLETIN.

GUESTS AND PARASITES OF THE BURROWING

BEE HALICTUS.

AXEL LEONARD MELANDER AND CHARLES THOMAS BRUES.

During the months of summer every roadside presents a field
of busy insect-activity, as varied and interesting as it is unseen
and unheeded. Those insects, however, that we do notice are
seen during their idling moments and hence we are generally ac-
customed to stigmatize all as idlers with no aim beyond song or
frolic. But insects have a busy life --another phase of their
existence which many of us overlook. If we inspect some road-
side more attentively we shall be surprised to see many of the
self-same idlers working with diligence. Spurred by parental
anxiety these insects excavate their nests and store them with
food, doing for their young what their parents have done for
them.

Out of this multiplicity of insect-life we shall select as an
example one of the burrowing bees of the genus Halictus, and
endeavor to tell what may be seen on any summer day. Halictns
(Chloralictus] pruinosns Robertson is a brilliant greenish bee,
measuring about one third of an inch in length, which lives over
an extended range, occurring from New Mexico, through Illinois,
to Massachusetts. It is the commonest Halictine at Woods
Hole, in the last-mentioned state, where the following observa-
tions were made. During the early part of summer these bees
commence their excavations along the roadsides wherever a
sandy slope presents a favorable situation, ami continue their
activities until early autumn, the colonies increasing in size, and
becoming more closely settled as the season advances. They
seem to be in the height of their vigor during the early part of
September in this region. Although their social instincts are
not so highly developed as those of Apis or Bombns, these bees

2 MELANDER AND BRUES.

depart in their habits from the strictly solitary bees in that a
male and two or three females are generally necessary for the
successful direction of a single menage. Moreover, a large
number of nests are usually associated as a colony which may
be scattered over a considerable distance or so populous that the
tunnels almost intersect by their irregularities. The openings to
the nests, however, are always separated by a distance of two or
three inches or more. It can thus be readily seen that Halictus
lives under conditions more or less similar to those of their more
gregarious relatives, the ants, and hence it is not surprising that
they are forced to harbor the same class of guests, and to be
exposed to the same vicissitudes as are their cousins.

In constructing their nests the bees dig by means of their
mandibles in the sandy clay, forming a hole of a diameter only
slightly greater than will admit the largest female. The wall is
then banked up with a plaster formed by the aid of saliva. Im-
mediately behind the entrance is a short blind passageway, only
large enough to allow a bee to turn on itself within.

This niche, which is always less than an inch from the entrance,
serves simply to allow the bees to pass one another in the in-
terior of the nest. From this point the gallery extends nearly
straight back into the hill side, for a distance of a few inches and
then slopes downward to the end — a total length of a foot or
so. Near the further end jut a number of small diverticula
radially extending from the main tunnel.

These are the nurseries of the young bees, where are stored the
pollen and honey which is destined to serve as food for the bee
larvae of the coming generation. The excavation of the tunnels
is a matter of considerable toil, requiring many days for its com-
pletion, but so industriously do the little bees work that at the
close of day a miniature mound of sand has accumulated on the
hill-slope below the opening. During the warm portions of the
day the site of each colony of nests is a scene of inspiring activ-
ity. The air is filled with an ever-changing swarm of bees, each
bent on its own task of excavation or of collecting honey and
pollen, while from the openings of completed nests others can be
seen peering about and eying everything that comes within their
range of perception. At night everything is quiet, the trespas-

GUESTS AND PARASITES OF HALICTUS. 3

sers and robbers, too, have ceased their work, and the colony
slumbers in peace.

The structure of the nest was ascertained by the ingenious
plaster-cast method advocated by Prof. J. B. Smith. By this
means the galleries of Halictits are seen to depart but little from
those of the other burrowing bees. A passage-way for exit and
entrance in addition to the regular one opening on the dumping
ground, such as is constructed by Augochlora humeralis Patton,1
was never noticed in the case of H. pniiuosns, the vigilance re-
quired to guard two openings having probably prevented such an
extravagance. All the burrows which we dug out, a dozen or so
in number, extended in a nearly horizontal direction, and were
always built on the very steep slopes along the roadsides. By

Fig. I. Diagram of Halictus nest, a, plan; b, elevation.

this means none of the excavated dirt accumulated about the
doorway, which was even cleared of all debris with but little effort
on the part of the bee. The relatives of pruinosns in Texas,
morphologically of the same species, select a level spot for their
nesting-site, dig vertical burrows, and place the accumulated dirt
in an irregular cone about the opening. A photograph of these
nests is given for comparison.

During the latter part of nest-construction when the pollen has
been gathered and the eggs laid, their home is continually threat-
ened by thieves and kidnappers against whom a guarded watch-
fulness must be maintained. The sentinels are generally the

JJ. B. Smith. Proc. Am. Ass. Adv. Sci.t 1898, p. 368.

4 MELANDER AND BRUES.

males, who sit at the doorway, their rounded heads neatly filling
out the entrance. When the female returns pollen-laden, the
little guard slips into the first side passage while she enters, and
then as quickly returns to his post. The incomers are perceived
at a distance of half a foot, probably announced by the buzzing
of their wings. Even when the little watchers can not see the
female coming they dart half way out of their retreat at her ap-
proach. With antennae vibrating and mandibles spread the males
either manifest a joyful greeting for their nest-mates or show an

FIG. 2. Nest of Halictus near Austin, Texas.

equal degree of hostility towards any stranger that may venture
too near.

The most dreaded of the enemies of the Halicti is perhaps the
little velvet ant, Mutilla canadensis Blake, which is common
nearly everywhere in North America, running about on the nests
of these bees, its distribution practically coinciding with that of
this species. Perhaps it is the stridulation produced by the ab-
domen of these intruders that arouses the ire of the guard at the
door, for no sooner does one approach a nest than the watcher,
if it be a female, rushes out and pounces upon the JMutilla, en-
deavoring to sting it to death. Down the hill-slope they roll,
heedless of everything but an inborn desire to annihilate each

GUESTS AND PARASITES OF HALICTUS.

5

other. The Rlutilla, too, is armed with a powerful sting, half the
length of her abdomen, but the sagacious Halictns grasps her
enemy about the waist and most successfully evades the sharp
thrusts. These combats continue for many minutes, concluded
either by the invulnerable Mutilla slipping from the bee's grasp,
for her body is hard and sleek, or by the death of the more
plucky Halictns. Each colony, where everything seemingly is
peace and content, is thus turned into a field of carnage, with the
bodies of one or more females ruthlessly tumbled to the bottom
of the hill. If the bee escapes unscathed, which happily is the
more usual outcome of these struggles, she spends a few moments
in preening her body, and then returns to her nest. But no

"';:-\'"- ' '-i'

. rr»-v. . *•',' .

FIG. 3. Nest of Halictiu at Woods lioll, Mass.

greeting awaits her after her loyal struggle. When she hurriedly
left the nest the male waiting his turn in the tunnel below quickly
took her place as guard at the door, and now he blocks the
entrance as obstinately as though it were a stranger begging
admittance. The taint of Mntilla is still to be recognized on the
body of the female and probably overpowers her family smell.
For quite a minute she must remain at the door parleying with
her mate before he is convinced of her identity.

This observation is of interest when considering the organic
dependence of instinct. Fear of Mntilla has been cultivated

O MELANDER AND BRUES.

through natural selection and heredity till it manifests itself in the
actions just recorded. But the conduct of the male towards his
nest-mate, an inhospitable act which a gleam of reasoning intelli-
gence would not permit under the circumstances, lends itself
rather to the theory of a mechanical instinct, actuated in this
case by the chemical nature of Mntilla s poison. If this be so it
will be questioned why the bee does not behave as when Mntilla
itself approaches. Does the mixture of Mntilla -influence and

compel an impassive head-on greeting while

FIG. 4. Combat between Mutilla and Halictus. " Down the hill they roll heed-
less of everything but an inborn desire to annihilate each other."

Mntilla alone induces the male-watcher to turn tail in the manner
described on the next page ?

One little bee once displayed an originality not noticed again.
For fully twenty minutes she had waited at the entrance of her
home, gently urging admission by advancing to the nest -opening
once each minute. The male would retreat a short distance each
time but not sufficiently far to admit the female, who would then
retire, resting with her antennae almost touching those of the

GUESTS AND PARASITES OF HALICTUS. /

stubborn gate-keeper. Finally she turned about and crept back-
ward to the male, resting a moment with her sting before his
face. When she now turned, the male seemed convinced, and
the wearied female entered in the usual way. In this case did
the female flaunt her own poison to overcome that of Mutilla as
a -passport to her home ? It might seem so ; but the simplicity
of such a physiological action is quite equalled by the complex-
ity of the intelligence displayed.

When a male bee guards the opening the approach of Mutilla
produces a far different effect upon the watcher. Instead of
rushing out on the marauder, the defenseless male adopts the less
foolhardy measure of "turning tail," but still keeps at the en-
trance of the nest. Now the convex abdomen neatly fits the
opening, forming a parasitic-proof shield, and Mutilla must needs
leave. When no other bee is behind a female watcher, she never
rushes out, leaving the nest unguarded, but adopts a manoeuvre
similar to the male's, but instead of inflexibly curving her abdo-
men over the opening, she reaches afar with her sting.

Canadensis, however, is not the only Mutillid that worries the
Halictines. On numerous occasions Myrmosa unicolor Say1 and
Mutilla infensa sp. nov. were found crawling about, but these species
do not appear to have become nearly so annoying. From one
square meter of Halictus -colony fully fifty specimens of canadcnsis
were taken during the summer, whereas in all but ten specimens
of the Myrmosa were observed. Mutilla fcrntgata Fabr. and
vesta Cresson were also found prowling over the nests, though
these species are doubtless parasitic on the larger burrowing in-
sects which associate with Halictus, for the large size of their
bodies would not permit entrance into the Halictus nests. More-
over, they may crawl quite close to the doorkeeper and elicit no
attention ; possibly their stridulation is pitched to an unrespon-
sive key and their odor stimulates no reaction.

Almost as ardent a persecutor of the bees is to be found in a

1 It is time to abandon superfluous names. Myrmosa jtnicolor Say, described as a
male, and M. thoracica Blake, described as a female, have paraded in collections
quite long enough as distinct species. Inasmuch as Mr. H. L. Viereck has recently
taken the initiative (Ent. ATtius, 1902, p. 72) in consolidating some of the species of
Mutillidse, we shall follow him in the nomenclature of this paper. The males of this
species fly abundantly among the roadside flowers, in company with males of cana-
lensis &nA.ferriiguta ( = castor Blake = Lepeleterii Fox \_fenestrata Lepeletier] ).

8 MELANDER AND BRUES.

new species of Plwra} This little fly takes a stand near an open-
ing and patiently awaits an unguarded moment. Then she quickly
slips in to deposit an egg in the pollen so industriously stored.
One Plwra persisted in her attempts to enter for several hours.
Driven back a half inch by the doorkeeper she gradually and
slowly returned until she nearly touched his face. Then a sudden
lunge half way out of the nest on the part of the bee would drive
her back again. This was repeated over and over, the dogged-
ness of the parasite and her slow approach seeming to exasperate
the little watcher. By turning his head he tried to follow her
movements, but from their very slowness was unable to discern
her position. Only when his palpi were touched would he make
a sudden dart. PJiora depends on her agility as well as on her
deliberateness. On each return of the female bee, after a fifteen-
minute foraging trip, the parasite would jump about excitedly
and possibly would get a chance to oviposit on the pollen mass
during a dart at the bee. A moment's rest on the threshold
would grant the nervous little fly ample time to infect the unsus-
pecting bee. The behavior of the bees towards Phora is quite
different from the action of ants towards these guests. Unless
irritated by the persistence of the parasite, Halictns is passive
and does not notice its presence. Even the incoming females dc
not'see the fly at a distance of half an inch. On the other hand,
ants are put in a state of fright by the proximity of these flies.
During the attacks of the ant-decapitating phorid, Apocephalus
Pergandci Coq. upon the species of Camponotustferruginea in
the north, and niacttlatus var. sansabeanns in the south, the ants
rush in the wildest excitement with wide-spread mandibles at the
agile fly. Can this difference result from the bees never seeing
their offspring and being consequently unaware of their fate,
whereas the ants have a personal acquaintance with the ravages
of these parasites ? It might seem so, but we must remember
that in the case of Pachycondyla Jiarpax, at least, a phorid larva
is not only tolerated in the nest, but is also fed by its host.2 In
this case, however, no harm is done to the species by the pres-
ence of the fly, whereas with Halictns it must mean the death ot
the brood.

1 P. halictoriini, described in the sequel.

2 Wheeler, W. M., Am. Nat., 1901, p. iocf]e( seq.

GUESTS AND PARASITES OF HALICTUS.

The most conspicuous of the smaller Hymenoptera that fre-
quent these grounds is a little species of Loxotropa. Time and
again this insect was observed crawling stealthily over the nest-
colony, tapping its antennae on the ground as it moved. During
this deliberate progress it covers an inch in four seconds, but as
soon as its nears a selected opening its movement slows down to
an almost imperceptible advance. Still holding its long and
clubbed antennae extended straight forward, their tapping now
reduced to a slight nervous vibration, it gradually insinuates itself
into the nest, even beneath the very jaws of the gatekeeper.
Often after crawling so far into the nest that only the tip of its
abdomen is visible, it finds the nest unsuitable. Then it deliber-
ates no longer, but makes a hasty exit, leaving the astonished

FIG. 5- Loxotropa ruficornis Ashm. Halictits, $ . Pliora cata, sp. nov.

sentinel to reach in vain with questioning antennae for its bold
and impudent disturber.

As interested an observer of the incoming bees as is the Pliora,
is a tachinid fly. This species hovers over the breeding ground
and suddenly circles over a particular hole. Is it attracted to
the nest by the hollowness of the sound of its vibrating wings as
it flies over an opening, or does it discern the state of advance-
ment of the household below by an instinct less mechanical ?
Like its relatives, this species chooses the moment when the
incoming bee pauses at her threshold quickly and quietly to
oviposit on her pollen mass and thus infect her offspring.

IO MELANDER AND BRUES.

A number of ants, foragers from near-by nests, are always to
be found on the nesting-ground. These belong to harmless
species which do not molest the bees. When an ant and a bee
meet on the nest there is no encounter, each retreating good-na-
turedly to go her own way. The Stenammas, especially, have a
stridulatory note as plaintive as that of Mutilla, yet this is unno-
ticed by the bees; even the watchers rest unaroused in their
doorways while the ants pass them by. The little red thief ant
is also found nesting in the midst of the bee-colony. Evidently
it is here to ply its vocation of tunnelling into the chambers of
the bees to steal from them their honey.

The little beetle, Btfoccra concolor, seems quite at home with
the bees. Although it belongs to a family of fungus-beetles, it,
nevertheless, must have some intimate connection with the bees,
as it was repeatedly observed running familiarly in and out of the
nests. It is quite possible that it may live upon the pollen in
deserted nests which has become mouldy by the growth of fungus
hyphae. The mixture of pollen and honey is thus readily turned
into a mass of fungus under certain conditions.

The woes of the Halicti are not yet at an end. Another insect
is as persevering in its depredations as its colleagues, and accom-
plishes by boldness what the others try by stealth. This is a
larger foe, PJiilantJius punctatits by name, which audaciously
builds its nest in the center of the Halictns colony, and when
ready swoops down on a bee, stings it to death, and carries it
home. Not one but many bees meet this death at the sting of
their unsuspected neighbor, who plans her murders so that they
take place at the flowers where the bees are at work.

When we consider the persistence of the Mutillas we can ap-
preciate the extent to which specialization in keeping the nest
parasite-proof has been carried by this bee. Seldom are the
entrances left unguarded, and never is a stranger bee granted ad-
mission. In this respect Halictns is far more conservative than
the wasp Trypoxylon. Although mistakes in selecting their own
domicile from a cluster of fifty similar nests were frequently
made, the watchers always recognized these visitors as strangers
and were instantly ready to show fight. Trypoxylon, a wasp
which also guards its doorways, on the contrary, makes no ob-

GUESTS AND PARASITES OF HALICTUS.

I I

jection to the free entrance of visitors of the other sex, as has
been shown by the Peckhams.1 Mutilla canadcnsis appears to
be the most dreaded enemy, as it alone is noticed by the
bees. With a little reasoning ability many of the other parasites
could be readily annihilated, whereas no move is made for pro-
tection against these foes except by the guard at the door. But
how are the bees to know, even in the case of Mittilla, that their
guests mean harm to their progeny ? Probably they do not in a
strict sense. It is evident, however, that the instinct of guarding
the entrance to the nest could have been developed through the

FIG. 6. Stethopathus occidentalis, sp. nov., lateral view.

action of natural selection of favorable variations in habit, while it
would be difficult to derive a number of specific reactions towards
the different guests in the same manner. The very commonness
of Mutilla and its conspicuous size are probably the reason that
a specific reaction has been developed in this single case.
Halictus is far less sensitive to its surroundings than many of
the fossorial wasps are, coming and going even though we dis-

1 " Instincts and Habits of the Solitary Wasps," p. 79, 1897.

12 MELANDER AND BRUES.

turbed the nest and remained close by. Its one fear is centered
in Mutilla. With thief-ants to rob its nests, parasites to prey on
its offspring, and in constant danger of being carried away bodily
by a wasp, itself numerous in individuals, it is remarkable that
Ha/ictus should have become a dominating type throughout such
a wide territory.

This ends the list of the enemies of the bees as we have ob-
served them. Many other insects abound on the nesting-site,
but most of these, at least, are accidental visitors which neither
harm nor are harmed. Several beetles, spiders, flies and other
insects are included in this list which we give for reference in
conclusion. The smaller species live near the Halictus as they
would do anywhere, and not through preference, and the larger
ones in part are attracted to our observation ground to prey on
the smaller. These transients are such as a careful observation
of any limited field would bring to notice. They are the partici-
pants in life's continual struggle, each seriously and unwittingly
playing its part.

PART TWO.

A LIST OF THE INSECTS, INCLUDING THE ACCIDENTAL VISITORS,

FOUND ABOUT THE COLONIES OF HALICTUS PRUINOSUS,

ROBERTSON, AT WOODS HOLE, MASS.

JULY- AUGUST, 1902.

Class ARACHNIDA.
Epeirid sp.

A minute larval spider was several times seen. It has no
connection with the Halictus.

Bathyphantes formica Emerton.

Quite a number of specimens of this strange spider were ob-
served running in their zigzag course over the ground. Like
the last it is an accidental visitor, occurring on the colony during
its search for food. We are indebted to Mr. Nathan Banks for
the determination of this species.

GUESTS AND PARASITES OF HALICTUS. I 3

Acarina spp.

Two species of mites were obtained, one of which (Biyobia
prate )i sis Garm. ?) occurred in numbers.

Class MYRIAPODA, DIPLOPODA.

Polyxenes fasciculatus Say.
Numerous specimens found crawling about on the sand.

Class INSECTA.
Order Thysanura.

The genus Podnra, represented by many specimens, was found
associated with the former.

Order Hemiptera.
Aleurodes sp.

The larval form of an Aleyrodid was discovered on the nest.
Probably it is that of A. corni Hald., the commonest form of the
Atlantic States.

Order Diptera.

Family CHIRONOMID^:.
Ceratopogon hollensis sp. nov.

Third vein in part confluent with the first, ending much beyond the
middle of the wing, wings in large part hairy, not uniform in coloration,
but not spotted ; eyes well separated ; tarsal claws simple, of an equal
length ; legs not spinose beneath ; metatarsus much longer than the second
tarsal joint.

Female. — Head fuscous, proboscis black. Antennas fuscous, the joints
uniformly moniliform, slightly longer than broad, the last twtr joints longer.
Eyes widely separated, the front yellowish. Mesonotum pruinose, sparsely
and uniformly covered with short black bristles. Abdomen dark fuscous,
lightly gray pruinose, apically hairy. Pleurae paler fuscous, smooth. Hal-
teres dark fuscous, the stems paler. Legs slender, uniformly yellowish,
except that the knees, and the tips of the femoral and tarsal joints are very
narrowly black ; tibiae provided with several simple long but slender hairs
on the outer edge ; no bristles below, tarsi somewhat hairy, claws small,
uniform, simple, empodium small. Wings sparsely covered with short
bristle-like hairs, more or less serially arranged. These become obsolete
at the very base, cinerascent with a pale brown tinge becoming stronger
along the basal part of the course of the anterior heavy veins, gradually
interrupted in front of the anterior cross-vein, then gradually recommencing

14 MELANDER AND BRUES.

to end abruptly before the tip of the first vein. The crotch of the furcation
of the light vein crossing the anterior cross-vein is darkened by an accumu-
lation of pigment and by an increase in the number of hairs.

Length, 0.85 mm.

Woods Hole, Massachusetts, August, 1902.

The nearest relative of this species is C. variipcnnis Coq.

It is not unlikely that the species is an halictophile, as it was
several times seen upon the nests, thus suggesting its myrme-
cophilous relatives. It may also be the cause of the presence of
some of the proctotrypidse here listed, as some of them are
known to prey on the larvae of various species of the genus.
This is the case with Adcliopria Ashm., a Diapriid, which is
parasitic on a Texan species of Ccratopogon)

Family MYCETOPHILID.E.

Sciara sp.

A Sciara would frequently fly over the nesting- site and alight
on the open ground. It is an accidental visitor more at home in

the nearby grass.

Family PHORID.E.

Phora halictorum sp. nov.

Female. — Length, 1.5-2.25 mm. Head black, subshining, antennae
black ; palpi dull yellow, with stiff black bristles below ; proboscis not ex-
serted ; front long, flattened, punctured, shining, its bristles reduced in size,
and those of the middle row placed high up. Anterior four proclinate
bristles small, the remaining ones placed normally.

Thorax black, subshining, the dorsum finely pubescent, the pleurae
lightly pruinose, ten bristles present on the hind edge of the mesonotum,
dorsum with coie pair of dorsocentral and four marginal scutellar bristles.

Abdomen black, shining though not brilliant, not bristly, lightly pruinose
basally along the sides ; ovipositor short, retractile, piceous.

Legs piceous, front legs somewhat lighter, front coxae dull yellowish,
middle and hind coxae piceous, hind coxae with the usual ridge on the pos-
terior side ; hind femora stoutest, twice as thick as the front ones, middle
femora intermediate, all the tibiae with short bristles, biserially arranged
on their outer side, those of the front tibiae ten to twelve in number and
approximated into one line towards the inner forward edge, those of the
other tibiae in two separated series, for the middle tibiae four in the outer
and six in the inner row, and for the hind tibiae seven in the outer and ten
in the inner rows ; front tibiae without terminal spurs, middle tibiae with one

1 See Wm. H. Ashmead, BIOL. BULL., 1902, p. 15.

GUESTS AND PARASITES OF HALICTUS. I 5

long spur three fourths the length of the metatarsus, hind tibiae with two
moderately long spurs, the outer one two-thirds as long as the inner, which
is nearly as long as that of the middle tibia.

Wings hyaline with faint cinereous tinge, not brilliantly iridescent, the
heavy veins nearly black, reaching very nearly to the middle of the wing.
First vein but slightly bowed, third vein nearly straight, furcate, costal
bristles fine and short, thickly placed, distributed as follows : four proximal
to the humeral cross vein, twenty-two (double series) bordering the costal
cell, ten (double series) bordering the marginal cell, and six (in double
series) along the submarginal, /. e., the furcation of the second heavy vein.
Thin veins dark, the fourth longitudinal slightly flexed only at its extreme
base, so that the cell in front is slightly wider than the one behind, ending
a little closer to the wing-tip than the second light vein does, seventh vein
evident, extending into the wing-margin. Halteres whitish, their stem
dusky.

Ma'e. — Length, 1.75111111. Differs as follows : frontal bristles stouter,
abdomen smaller, genitalia not distinct, small, the central filament fleshy,
short, directed backward. Tibial setulse and the inner spur of the hind
tibiie reduced in size, ridge of the hind coxae large ; costal bristles not uni-
form, disposed thus : prehumeral four, twenty along costal cell, four along
marginal cell and four along the submarginal. The inner bristles are
minute, becoming larger at the third pair of the costal cell, and from thence
are much stronger than in the female.

Described from several specimens, collected as described in the
previous account at Woods Hole, Mass., July-August.

This species is related to agarici Lintner l but differs by the
longer bristles on the tibiee, longer front, four scutellar bristles,
etc. The habits also are quite different, as agarici feeds upon
decaying mushrooms.

Phora rostrata sp. nov.

Female. — Length, 1.5-1.75 mm. Black, shining, legs more or less yel-
low, lower frontal bristles proclinate, third vein forked.

Head shining black, especially smooth and polished on the front and
vertex. Front with the normal chaetotaxy except that there are only two
proclinate bristles at the lower edge. The front is also sparsely hairy
besides the large bristles. Median longitudinal groove and ocellar tubercle
unusually well-marked. Antennas black, with a distinctly plumose arista.
Proboscis piceous, very large and strongly exserted, as long as the head-
height. It is slender at the base where the rather small bristly spindle-
shaped black palpi are inserted, then much enlarged, swollen and bifurcated
at the extremity. The bifurcation is produced by a splitting of the apex by
a horizontal slit in the proboscis. Thoracic dorsum shining, hairy as usual,

1 loth N. Y. Kept., pp. 399-406.

1 6 MELANDER AND BRUES.

with one pair of dorsocentral and two scutellar bristles. Abdomen black,
nowhere bristly. Legs pale yellow, the tarsi sometimes brownish ; hind
tibiae very indistinctly ciliated and with a single weak spur, as have also
the middle pair. On the inner side at the apex the posterior pair have
several transverse rows of short black bristles. Wings yellowish hyaline,
the costal vein reaching distinctly beyond the middle of the wing and with
very short cilia. First vein ending a little closer to the tip of the second
than to the humeral cross vein. Fourth vein evenly arcuate, fifth vein
sinuate as is also the sixth ; seventh vein present. Halteres yellowish,
blackened at the tips.

Described from two female specimens collected at Woods
Hole, Mass., July 15, 1902, about the burrows of Halictus pnt-
inosiis.

This species is readily recognizable on account of the excessive
development of the proboscis, which is evident!}' adapted to some
peculiar method of food-getting. It is also characterized espe-
cially by the very shining front, which seems to place it near to
the European P. minor Zett., with which it agrees is some other
characters.

Phora cata sp. nov.

Male and Female. — 0.8-1.2 mm. Black, legs and palpi yellowish or
brown, antennas of male enlarged. Anterior frontal bristles proclinate.
Head black, front short, about as wide as long, subshining, faintly gray
pollinose in the male, two anterior bristles proclinate, the others all present
and arranged as usual. Antenna; wholly black in the male, in which sex
the third joint is enlarged and ovate so as to be very conspicuous, in the
female they are of the usual size and slightly yellowish at the base ; arista
pubescent. Palpi light yellow, strongly bristly. Proboscis of female pro-
jecting, stout and horny. Thorax shining, black, hairy, with one pair of
dorsocentrals and two marginal scutellar bristles. Abdomen black. Legs
yellowish-brown, the anterior pair lighter. Posterior femora ciliated below
on apical half, their tibiae without any rows of small bristles on the outer
side ; four posterior tibiae each with a delicate apical spur. Wings hyaline,
the costal vein not quite reaching to the middle of the wing, its cilia short
and closely placed. Third vein far from the costa at its base, and forked
very near the apex. Tip of first vein twice as far from the humeral vein
as from the tip of the second. Fourth vein slightly but evenly curved, re-
curved at the extreme tip. Fifth slightly diverging from the fourth to its
tip, which is as far behind the wing tip as the fourth is before it. Seventh
vein faint but distinct. Halteres yellowish in the female, piceous in the male-
Described from a single pair from Woods Hole, Mass. The
lighter color of the female is most likely due to her apparently

GUESTS AND PARASITES OF HALICTUS. I/

immature condition. They were taken on the sand in the midst
of a colony of Halictns.

This species can readily be recognized in the male sex by the
enlarged third joint of the antenna. The female is not so charac-
teristic, but can be distinguished by the combination of structural
characters given in the description. It resembles most closely
P. agarici Lintner, but has very short costal bristles.

Stethopathus Wand.

Among the insects frequenting the ground immediately about
the //rt//V///^-burrows was one extremely small form, which from
its quick motions we immediately suspected to be a wingless
phorid fly. Such it indeed proved to be, but of quite a different
sort from any of our previously discovered North American
species. Its occurrence in New England is quite unexpected and
considerably extends the range of such forms, as none have
hitherto been seen in America north of central Texas.

Its associations with the Halictus may be doubtful, although
no specimens could be found elsewhere whereas three females
were captured where the burrows of the bees were abundant.
Nests of Lasins niger and of Stenamma fnlvmn, variety piceum
also abound in such locations, but close scrutiny of the ant nests
revealed no specimens of the Phoridae. The fact that species of
Phora occur as parasites of these bees would make it seem not
improbable that the StetlwpatJins has similar habits. We have
also a single winged male phorid, captured at the same time, but
which is probably the male of some other undescribed form on
account of its larger size and the different chaetotaxy of the head.
The description of this interesting little wingless fly, one of the
smallest known of all the Diptera, is given herewith.

Stethopathus occidentalis sp. nov.

Female. — Head rounded triangular, much rounded on the sides and at
the hind angles and obtusely pointed in front, about two thirds as long as
wide above, vertex descending rather steeply and evenly. Eyes small,
about one and one third times as large as the second antennal joint,
coarsely facetted with hemispherical ommatidiaas usual. Antennas placed
at the bottom of the deep frontal cavities. Proboscis long and stout,
equal to the head-height ; palpi small and slender, thickest near the tips,
with stout macrochaetaeonthe inner side. Ocelli present, placed in a small

IS

MELANDER AND BRUES.

triangle on the vertex. Head with four closely approximated macrochaetae
at the middle of the front margin, two widely-separated ones near the
anterior corner of the eye directed inwards and two outwardly directed
ones near the posterior angles ; a series of small macrochaetae below and in
front of the eye.

Thorax small, twice as wide as long, truncate before and behind ; sinu-
ate on the sides and narrowed behind, so that the pleurae are slightly visi-
ble from above. Thorax rather sharply arched above, and much narrowed
below on the sides. Dorsal surface with a pair of long macrochaetae just

FIG. 7- Stethopathus occidentalis sp. nov., dorsal view.

behind the anterior angles and four smaller marginal ones along the pos-
terior edge.

Abdomen considerably swollen, but with large and strongly chitinized
dorsal plates. The first is only a narrow band, contiguous with the second
which is very large and contiguous with the third. The fourth and fifth
are separated by a white membrane such as covers the abdomen else-
where. Seen from above the abdomen is twice as wide as the thorax and

GUESTS AND PARASITES OF HALICTUS. 19

flattened, oval in cross-section. No ventral sclerites are present. Each
segment is margined behind with small bristles and is hairy elsewhere as
is the entire body. Glandular opening of the fifth segment l in the shape of
an arcuate slit. External genital organs of the usual form. Legs rather
stout, the tibiae with two apical spurs.

Length, 0.75 mm. Testaceous, head and thorax darker above, espe-
cially directly about the ocelli. Abdominal plates dark fuscous, the mem-
branous parts almost white, with a small fuscous spot at the insertion of
each hair.

This form is a typical representative of the Stethopathinae and
strange to say, it approaches more nearly to the East Indian
Stethopathus occllatns Wand, than to any of the species that have
hitherto been discovered in America. Indeed, it is here regarded
as congeneric with the former, although the two species are from
such widely separated regions of the earth.2 It may be neces-
sary later to separate these two forms, but at the present state of
our knowledge of this group it does not seem advisable. The
American species resembles 5. ocellatits Wand, in possessing
ocelli, being utterly destitute of wings and halteres, and in having
a similarly shaped head and abdomen. But differences in form
are also evident : the thorax is only twice as wide as long, in-
stead of three times as in ocellatits, the palpi are clavate, not
spindle-shaped, and the chaetotaxy is somewhat different, although
conforming to the same general type. Although its habitus seems
to be quite different from that of the genus dEnigmatias Meinert
(which it may be recalled, has just been discovered in Arizona,3
a locality quite distant from its home in Denmark), yet this spe-
cies may possibly prove to be a close relative.

A point perhaps of minor importance, but nevertheless inter-
esting as bearing upon its systematic position, is the fact that the

1 In previous papers the gland opening has been referred to the fourth segment of
the abdomen, but the very short first segment in the present species leads us to believe
that this sclerite is concealed in the other American species and that there, too, the
gland really opens on the fifth segment.

2 Many cases might be mentioned of monotypical or very small genera of insects,
which have an inexplicably wide discontinuous distribution. Amphizoa with two-
species, one in western North America and another in Tibet ; Syntelia, which is rep-
resented by two species, occurs in Mexico and eastern Asia ; and the water-beetle,
Pelobius, occurring in western Europe, Tibet and Australia. For further references
to the close approximation in certain details of the faunre of eastern North America
and Asia, see C. C. Adams, " The Southeastern United States as a Center of Geo-
graphical Distribution of Flora and Fauna," BIOL. BULL., Vol. III., pp. 115, et seq.

3 1). W. Coquillett, Can. Ent., 1903, p. 20.

2O MELANDER AND BRUES.

American species has, like the East Indian form, bare, non-pubes-
cent macrochaetse, while the other American species of this sub-
family have them pubescent.

Family TACHINID/E.
Metopia leucocephala Rossi.

The interested observer of the Halictus mentioned in the first
part was captured for identification, and proves to belong to
this widely distributed species.

Order Hymenoptera.

Family BRACONID.E.

Subfamily CHELONIN.E.

Chelonus brevipennis sp. nov.

Female. — Length, 2 mm. Ferruginous, head piceous black, wings
reaching just beyond the base of the abdomen.1 Antennae 21 -jointed,
tapering as usual, and almost as long as the body, ferruginous at the base,
black at the tip, the third joint four times as long as thick, the apical joints
more or less quadrate-moniliform. Eyes smaller and less densely hairy
than usual. Head almost smooth above, shining, thinly pale pubescent,
piceous black above, ferruginous below, palpi yellow. Thorax ferruginous,
pronotum coarsely rugose reticulate above, mesonotum less distinctly so,
metanotum small, quadrate, not toothed at the posterior angles, rugoso-
reticulate ; pleurae not roughly sculptured, somewhat shining. Abdomen
with no traces of sutures above, dark ferruginous and sparsely white hairy ;
gradually broadened from the base and rounded at apex ; finely and irregu-
larly reticulately striate longitudinally, especially at the base. The incurved
margin is emarginate at the apex of the abdomen. Ovipositor stout, black.
Legs long and slender, yellow, the femora clavate.

Described from a single female specimen collected at Woods
Hole, Mass., in a burrow of Halictus pJ-uinosns.

The present species seems best referable to Chclomis because
of its pubescent eyes. The apex of the abdomen however is
emarginate, somewhat as in Gastrotheca Guerin. Unfortunately
as the wings are rudimentary they can not be used to determine
its affinities. The only other apterous species belonging to this
subfamily are included in Acampsis Wesmael, from which the
present form differs by its unsegmented. abdomen.
1 For neuration, see BIOL. BULL., 1903, p. 189, Fig. 5.

GUESTS AND PARASITES OF HALICTUS. 21.

Family CHALCiniD/E.1

Eupelmus rhizophelus Ashmead.2

This remarkable chalcidid with vestigial wings in the female
was seen rather commonly about the Halictus burrows. As it has
been previously bred from cynipid root galls by Mr. Ashmead, it
is no doubt an accidental visitor to the bee nests.

Eupelmus Ashmeadii sp. nov.

Female. -- Length, 3.5-4 mm., ovipositor 0.5 mm. Shining green
varied with ferruginous on the thorax and with luteous and black on the
abdomen. Head shining green, with a sparse white pubescence. Man-
dibles brown, black at the tips, palpi black. Antennae long, the scape
yellow, reaching to the ocelli, flagellum black, about once and one half
the head-height, last joint acutely pointed. Head less than twice as wide
as long, the space between the eyes above narrow, so that the lateral
ocelli are close to the eye-margin. Face rugoso-punctate with a median
carina extending from the clypeus to the insertion of the antennae. Pro-
thorax shining brown. Mesonotum very closely punctate, not at all shining,
brown in front and green behind, concave medially behind, on each side
of the depression it is raised and almost carinate, then slopes down to the
reflexed margin ; anteriorly it is raised to form a broad triangular tubercle.
Pleurae ferruginous except in front where they are green. Metanotum
golden, closely punctate, bilobed, sharply declivous, forming a right angle
with the mesonotum. Wings deeply infuscated, paler at base and slightly
so at apex, with a narrow cross band of white just before the stigmal vein.
Marginal vein equal to one third the length of the wing, stigmal vein
moderate, one half the length of the post-marginal. Abdomen shining
black, pale luteous on the basal third. Sheaths of the ovipositor bright
ferruginous, almost as long as the abdomen. Legs brown, darker on the
front and hind femora, tarsi yellowish except the tips.

Described from three female specimens collected at Woods
Hole, Mass., July and August, 1902.

This pretty species was associated with the much smaller
brachypterous species, Eitpclinns rliizopliclus Ashm., on the bur-
rows of Halictus pruinosus. It is named in honor of Mr. Wm.
H. Ashmead, who determined it as an undescribed species.

Henicopygus subapterus Ashmead.

We have seen' this species running actively about on the ground
among Halictus burrows at Austin, Texas. Like the species of
Eupelmus, it may be an accidental visitor.

1 We are indebted to Mr. Wm. H. Ashmead for his kindness in determining the
species of Chalcididaj.

2 For wing-neuration, see BIOL. BULL., 1903, p. 189, Fig. 7.

22 MELANDER AND BRUES.

Encyrtinae gen. et sp. indesc.

Among the Chalcididae there is a single specimen which Mr.
Ashmead, who has kindly examined it, informs us represents
an undescribed genus of Encyrtinae. Unfortunately it is too
poorly preserved to permit of an accurate characterization in the
large and difficult group to which it belongs.

Cirrospiloideus (Miotropis) platynotae Howard.
A single female of this species was captured.

Superfamily PROCTOTRYPOIDEA.
Family SCELIONID.E.

Telenomus sp.

There is a single pair representing an apparently undescribed
species in this large and difficult genus.

Caloteleia Marlattii Ashmead.
This active little species is a regular visitor about the nests.

Caloteleia parvipennis sp. nov.

Female. — Length, 2.5 mm. Yellow, varied with darker. Head
black, very smooth and polished above the antenna;, finely punctured on
the vertex and with larger punctures intermixed. Mandibles yellow at the
base, black at the tip. Antennal scape pale yellow, reaching a little above
the vertex, the pedicel small and rounded, yellow, the flagellum about one
and one half times the length on the scape, black, the first flagellar joint
twice as long as the pedicel, then the joints decrease in size to the fourth,
the following six forming a thick oval club with closely articulated joints.
Thorax entirely yellow, except the tegulae which are black, mesonotum
finely punctulate, with two rather faintly marked furrows, scutellum large,
semicircular, smooth. Metathorax very short, emarginate in the middle,
smooth on the sides. Abdomen polished and perfectly smooth, except for
coarse longitudinal striae on the first and at the base of the second seg-
ments. The petiole is short quadrate, and bears a quite distinct polished
black tubercular horn at its base ; basal half of abdomen otherwise yellow-
ish varied with brown, apical half black ; third segment longest, second
nearly as long, others much shorter. Legs including the coxae yellow.
Wings short, reaching only to the middle of the abdomen. Marginal vein
short and swollen, stigmal about one third as long as the lengthened post-
marginal, costal margin sparsely ciliated.

Described from one female specimen taken at Woods Hole,
Mass., on a slope that was thickly riddled with the burrows of

GUESTS AND PARASITES OF HALICTUS. 23

Halictus. When captured it made no attempt to fly, the wings
evidently being too much atrophied to be of functional use.

This form can be readily recognized by its short wings. It
does not seem to be very closely related to any of the other
North American species.

Scelio ovivorus Riley.

This large and coarsely sculptured Scelionid was originally
bred by Scudder from the eggs of the common New England
grasshopper (Dissosteira Carolina] so that its occurrence is evi-
dently not connected with the presence of the Halictiis colony.
Nevertheless it was often seen intermingling with the bees.

Family DIAPRIID.E.
Loxotropa ruficornis Ashmead.

This is a common species always to be found on the breeding
ground of these bees. Its habits have already been noted in the
preceding part of this paper.

Family BETHYLID.E.
Empyris subapterus sp. nov.

Female. — Length, 3.25 mm. Black, head and thorax subopaque, ab-
domen shining ; antenna;, mandibles at tips, palpi, tegulae and extreme
tip of abdomen rufous ; sparsely pale pubescent. Head about one third
longer than wide, closely and finely punctate with fewer larger punctures
intermixed. Antennas reaching about to the tegulae, scape stout and
curved, three times as long as its thickness at the tip ; following joints of
about equal length, except the first flagellar, which is shorter ; pedicel more
slender, the other joints slightly wider than long. Eyes hairy, ocelli
present. Prothorax sculptured like the head, with a transverse impressed
line anteriorly. Mesonotum very short, less than half as long as wide,
without grooves or furrows. Tegulae rufous. Scutellum basally with a
deep transverse linear fovea. Metanotum about one and one half times
as long as wide, with a median longitudinal carina and a fainter one close
on each side of it anteriorly, also a lateral and an apical carina present ;
surface elsewhere finely transversely rugulose ; posterior face sharply
declivous, shining and punctulate. Wings abbreviated,1 just attaining the
apex of the metanotum ; with a small stigma near the apex, a narrow,
submarginal cell and an equally long but wider basal cell ; costal margin
fringed. Legs, including the coxae, dull rufous. Abdomen polished black,
the margin of the penultimate segment and the apical half of the last seg-
ment ferruginous.

'For figure see BIOL. BULL., 1903, p. 189, Fig. 2.

24 MELANDER AND BRUES.

Described from several female specimens collected at Woods
Roll, Mass., running about among the burrows of a colony of
Halictus pruinosus Robts.

This species greatly resembles Mesitius in habitus, but has a
transverse furrow at the base of the scutellum instead of two
foveae. It can hardly be the undescribed female of E. carbonarius
Ashmead, on account of the difference in the sculpture of the
metanotum. It is apparently the first subapterous form to be
described in this genus.

Family FORMICIDJE.

Lasius niger Linneus.

Stenamma fulvum var. piceum Buckley.

Solenopsis molesta Say.

This last named species is the only one that derives any direct
benefit from the presence of the bees.

Family MUTILLID.E.
Mutilla canadensis Blake.

This is the most conspicuous of the enemies of the bees. It
has been fully noticed in the preceding part.

Mutilla infensa sp. nov.

Female. — Clothed with sparse appressed white pubescence becoming
densef apically, and with scattered long erect hairs. The hairs are black
on the vertex, dorsulum and second abdominal segment and become
whitish on the under side of the body and beyond the second segment of
the abdomen. Coarsely sculptured species ; head finely and closely punc-
tate, thorax and petiole of the abdomen coarsely reticulate, abdomen much
less deeply and more distantly punctured than the head, the apical segments
with finer punctures, meso and metapleurse shining, not or but little stri-
gose, nearly smooth, pygidium longitudinally closely but irregularly striated,
the striae very weak and vanishing apically. Head quadrate, concave
behind, in profile also rounded ; eyes prominent, round, subshining, their
facets distinct ; mandibles straight, strong, pointed, untoothed ; scape stout,
as long as the three basal joints of the flagellum, basal flagellar joints sub-
equal. Thorax elongate-oval, nearly as broad as the head, the front mar-
gin and angles well defined, posterior surface of the metanotum not sharply
declivous, somewhat flattened and rounded above. Petiole of the abdomen
flattened above, constricted from the second segment, one fourth broader
than long, its front angles sharp and prominent, its ventral carina weak,

GUESTS AND PARASITES OF HALICTUS. 25

very obtusely angulate at the middle and minutely toothed in front. Legs
slender, provided like the body with silvery erect hairs, four or more strong
spines on the outer edge of the hind tibiae, the tibial spurs and spines black.

Ferruginous or somewhat darker, the mandibles, the flagellum except
its basal joint, /. e., the third antennal joint, more or less of the second ab-
dominal segment, and all of the other segments of the abdomen, from the
third apically, both ventrally and dorsally black. Legs including the coxae
piceous or black. Second segment of the abdomen with a varying extent
of the front margin, a diffused median vitta and the hind margin more
strongly black or blackish. On each side of the median stripe is a pair of
conspicuous rounded testaceous spots. Last ventral segment sometimes
reddish.

Length, 4.75 mm.

Woods Hole, Massachusetts. Parasitic on Halictus prui-
nosus (?).

The edentate mandibles, the facetted eyes and the nodose
petiole of the abdomen would lead one in placing this species in
the small group scrupea, where it is obviously distinct from the
only other known female by its rugose thorax, etc. Notwith-
standing this, we shall have to disregard the well-marked om-
matidia and place the species in the group occidentalis , intermedi-
ate.between cariniceps Fox and mgitlosa Fox, differing from each
by the structure of the pygidium, etc., but related by its general
habitus, sculpture and chfetotaxy.

Mutilla vesta Cresson.

*

Mutilla ferrugata Fabricius.

Like the former species this too is doubtless parasitic on the
larger Hymenoptera such as PJiilanthus or the Pompilidae that
nest near by.1

Myrmosa unicolor Say.

The males of this species fly about the roadside flowers while
the females are frequently found about the bee nests. Their
presence is undoubtedly due to the bees.

Family PHILANTHID.E.
Philantus punctatus Say.

This species was observed nesting in the very midst of several
of the colonies of Halictus.

1 In Europe Sichel records J\I. incompleta Lep. as parasitic on Halictus (cf. Horae,
Soc. ent. Ross., VI., p. n) and M. coronata as a parasite of Larra anathema
(ibid., p. 12).

26 MELANDER AND BRUES.

Sphex ichnenmonea Linn., and a species of Pompilidae were
also seen digging their nests in the compact sand of the road in
the vicinity of the bee colony. They have no connection with
the presence of the bees, but associate with them as the same
condition of soil and surroundings are suitable for each.

Order Coleoptera.

Family COCCINELLID.E.

Microweisea misella Leconte.1

The species of this genus are reported to be of great economic
importance as they greedily prey on scale insects. The presence
of the Aleyrodes may have had an influence in bringing this spe-
cies to our notice.

Family ENDOMYCHID/E.
Aphorista vittata Fabricius.

Family PTINID.E.
Casnocara scymnoides Leconte.

Family SCAPHIDID.E.
Baeocera concolor Fabricius.

The last three species are fungus-eating beetles, which may
come to the Halictits nests to feed on the fungus overgrowing
the stores of abandoned or damp nests. It is certain that during
the course of the season numerous nests are left unfinished, either
deserted voluntarily by the bees for some whimsical reason or
not completed by the death of the bees.

Family RHIPIPHORID.E.
Myodites fasciatus Say.

Inasmuch as Fabre and others have found the larvse and pupse
of a member of this family in the cells of an European species of
Halictus, it is quite interesting to note the occurrence of M. fasci-
atus about the colonies of the American form. Several specimens
were taken while sweeping with the net among the swarming
bees as they entered and left their nests.

Several other beetles were found crawling over the nest but
were visitants too accidental to record.

1 This is the species known in our lists under the generic name Swilia or Pentilia.
The present name was proposed by Cockerell ( Can. Ent., 1903, p. 38).

GUESTS AND PARASITES OF HALICTUS.

2/

In conclusion we may present the following diagram showing
the interrelationships of the most important of the insects we
have observed. For Halictns it is indeed a "whirlpool of life"
with only too many vortices centered upon its unfortunate self.

Mutilla

Myodites
Philanthus s
Metopia •

/
n

Hi
Mutilla

Phora

HALICTUS -*-*s. Stethopathus
A K ^~% Ceratopogon

* f

Baeocera Solenopsis j||

Proctotrypidse

A f

y

Pompilus
Sphex

Spiders

Chrysis

Polyxenes < zz Smilia

Aleyrodes
Thysanura

Sciara

THE ORIGIN OF THE HEART ENDOTHELIUM IN

AMPHIBIA.1

J. B. JOHNSTON.

The origin of the heart endothelium in Amphibia has been the
subject of several special investigations and of a considerable
volume of discussion. The question of fact may now be regarded
as settled. The work of Brachet has given definite and conclu-
sive evidence that the endothelium is derived directly from the
entoblast, as had been shown to be very probable by the earlier
work of Rabl and Schwink. The question now of interest is,
how is the derivation of the heart endothelium from the entoblast
in amphibia to be harmonized with its known origin from the
mesoblast in all other vertebrates ? The problem is that of the
homology of the heart endothelium of amphibia. Granted that,
as Ziegler contends, the condition in amphibia is to be regarded
as the result of ccenogenetic modification, exactly what is the
modification that has taken place ? What is the definite explan-
ation of the striking difference between amphibia and other ver-
tebrates ? As Brachet has pointed out, the term " ccenogenesis "
can not be invoked as a magic symbol to dispense with the whole
matter. It is not enough to say that in amphibia the endothelial
cells remain connected with the entoblast until a late period and
become separated after the mesoblast sheet has split off. This
offers no escape from the difficulty pointed out by Morgan ('97,
p. 151) that the heart endothelium must be considered to have a
different origin from the rest of the heart.

The work upon which the present paper is based has been
done upon the eggs of a salamander which have been used for
class study for the past two years. The species has not been
identified because no adults have yet been taken. I hope at some '
later time to give a description of these eggs and to deal with
some other features of the embryology of the species. The eggs

1 Studies from the Zoological Laboratory of West Virginia University, No. 7
February 27, 1903.

28

HEART ENDOTHELIUM IN AMPHIBIA. 29

have proved very favorable for study and the facts are so clearly
made out that they are thought to offer a solution of the problem.

The earliest indication of the formation of the heart endothe-
lium is found in the rapid multiplication of the cells of the ento-
blast just behind the mouth anlage, at a period when the head is
slightly turned downward and before the gill slits have begun to
appear. As shown in Fig. i, the nuclei in this part of the ento-
blast are small, rounded, very numerous and closely crowded,
and many of them are in some stage of mitosis. The nuclei in
the remainder of the entoblast are larger and irregular, being
much distorted by pressure of the yolk grains, and mitotic fig-
ures are rare. The area described extends for a considerable
distance backward from the mouth, and the same conditions pre-
vail on the cephalic surface and the sides of the pharynx close to
the mouth anlage. Rapid growth in these latter regions con-
tinues later than behind the mouth and is connected with the
formation of head mesenchyme. The region of growth behind
the mouth is noticeable in both transverse and sagittal sections,
but it is of short duration and in slightly later stages the cells are
relatively larger and the nuclei have the appearance of resting
nuclei. At the point nearest the mouth the cell divisions con-
tinue until the time of separation of the heart endothelium.

The formation of the mesoblast and its early differentiation
furnish the facts of greatest significance for our problem. In the
head and anterior part of the trunk the mesoblast is split off from
the entoblast to a point some distance from the mid-ventral line,
where the delamination appears to stop. That this is a definite
limit beyond which delamination does not go is evidenced by the
distinct separation between entoblast and mesoblast which often
occurs even in very early stages (Fig. 2, a and b\ by the total
absence of nuclei in the outer half of the entoblast ventral to the
limit mentioned, and by the future history of the ventral portion
of the mesoblast. Mitotic figures often appear very early in the
ventral edge of the mesoblast sheet (Fig. i), and although they
do not appear in the sections drawn in Fig. 2, they are usually
more numerous there than elsewhere in the mesoblast. The
result of rapid growth here is to cause a decided thickening of
the ventral edge of the mesoblast, and in this thickening the body

J. B. JOHNSTON.

3a

b.

FIGS, i, 2 ANI> 3.

HEART ENDOTHELIUM IN AMPHIBIA. 3!

cavity early makes its appearance (Fig. 3, a). With further
growth the body cavity enlarges and the entoblast is laterally
compressed between the cavities of the two sides. As a result,
the growing entoblast behind the mouth, above described, takes
on the form of a keel. Later the body cavity (pericardial cavity)
spreads ventrally and mesially, and the mesoblast insinuates
itself between the heart endothelial cells and the ectoblast and
later between these cells and the entoblast. This movement is
due entirely to the growth and spreading of the mesoblast earlier
split off and not to a further delamination from the entoblast.
There is no sign of any further delamination of mesoblast after
the stage shown in Fig. 2, but on the contrary the mesoblast
grows continually more and more sharply distinct from the
entoblast after that period. The pushing down of the mesoblast
in the region of the heart, which accompanies the enlargement
of the pericardial cavities, is well advanced while the thickened
ventral edge of the mesoblast farther caudally has not shifted its
position (Fig. 3, a, b, <r). The region in which the delamination
of mesoblast does not reach the mid-ventral line extends caudally
to a point a little behind the middle of the embryo and this region
probably includes the blood island described by Brachet. The
writer has not yet fully investigated this region, but if the surmise
here made is correct, the reasoning applied to the question of the
heart endothelium will apply equally well to the blood island.
To recapitulate, there is a mid-ventral area or keel of entoblast
extending backward from the mouth anlage, from which no
mesoblast is split off in the species studied. From this area the
heart endothelium (and perhaps the blood) are formed.

A second fact of some interest for us is that the mesoblast
shows a tendency to split off late, so that it is already divided
into regions when it first separates from the entoblast. This is
seen especially in the formation of the mandibular arch. As
shown by Fig. 2. e, the mandibular arch mesoblast, at its first
appearance is separated from the rest of the mesoblast by the
first gill slit, and it never has any connection with the mesoblast
bounding the pericardial cavity. Indications of the second gill
slit also appear very early, so that in some cases the hyoid arch,
which is continuous with the pericardial cavity, seems to be split

32 J. B. JOHNSTON.

off from the entoblast separately from the rest of the meso-
blast. Finally, single cells wander off from the cephalic surface
of the entoblast and go immediately to the formation of head
mesenchyme.

The mode of formation of the heart endothelium from the ven-
tral keel of entoblast differs in details in different forms. In the
Urodeles studied by Brachet, the keel of entoblast extending from
the mouth anlage to the region of the liver splits off as a contin-
uous rod, the cells of which later arrange themselves into a tube.
In the species studied by the writer the cells of this keel do not
remain in a continuous rod but split off singly or in groups of a
few cells and form a loose mass which remains connected with
the entoblast longest at the end nearest the mouth. At this
point there is continued growth and there is probably a migra-

FIG. 4.

tion of cells from this point backward, and also upward into the
several branchial arches, to form the aortic arches. The splitting
off of this keel of entoblast is taking place simultaneously with
the spreading ventrally and mesially of the pericardial mesoblast.
My preparations leave no doubt whatever that the heart endo-
thelium is formed from the most superficial portion of the ento-
blast in the mid-ventral region and that the lateral sheets of
mesoblast are formed wholly outside of this area. Brachet's de-
scription makes it clear that the same thing is true of the Urodeles
which he studied, but this important relation seems not to have
attracted his attention and the fact is not mentioned by him.

We are now able to state definitely the nature of the cceno-'
genetic modifications connected with the formation of the heart
endothelium in amphibia. According to the earlier accounts

HEART ENDOTHELIUM IN AMPHIBIA. 33

which have recognized the derivation of the endothelium directly
from the entoblast, the mesoblast sheets were split off first, and
later — consequently from deeper layers of entoblast — the cells
destined to form the heart endothelium were split off. Since in
other vertebrates only one layer of cells is split off and the heart
endothelium is differentiated from a part of this mesoblast, the
conclusion that the endothelium of amphibia has a different
origin from that of other vertebrates was unanswerable. In the
species studied by the writer (and also, apparently, in those
described by Brachet) the mesoblast sheets are split off earlier
from the entoblast except in the region in which the heart
endothelium will appear, and later the endothelium is split off
from a part of the entoblast which has not given rise to any
(other) mesoblast. Therefore, in these forms, the heart endo-
thelium is derived from the same source as the mesoblast sheets,
namely from the superficial layer of entoblast, and the difference
between these amphibia and other classes of vertebrates consists
only in a somewhat general tendency for the mesoblast to split
off relatively late and to be marked out into definite organs, or
organ-anlages, at the moment of splitting off. This is seen not
only in the splitting off of the heart endothelium at a little later
time and separately from the rest of the mesoblast, but also in
the same mode of formation of the mandibular and hyoid arches
and of a part of the head mesenchyme. The writer believes that
a reexamination of other amphibia, at least of Urodeles, at the
proper stages of development will show the process here described
to be characteristic for amphibia. In brief, then, the heart
endothelium of amphibia is strictly mesoblastic, although it is not
at any stage identified with the undifferentiated mesoblast, being
split off from the entoblast in the same manner as the rest of the
mesoblast, but somewhat later and separately.

DESCRIPTION OF FIGURES. ABBREVIATIONS.

a., aortic arch cells; arch., archenteron ; br.i, first branchial arch; c., ccelome ;
fff., ectoblast; en/., entoblast; g.i, g.2, g. J, first, second, and third gill slits; /;.,
heart region ; h.e., heart endothelium ; ky., hyoid arch ; /., liver region ; m., site of
mouth ; m.a. , mandibular arch ; rues. , mesoblast \p.c. , pericardial cavity ;ph., pharynx.
Small crosses indicate the position of mitotic figures. In Figs. I and 4 resting
nuclei are shown as black spots.

34 J- B. JOHNSTON.

FIG. I. Transverse section through the ventral part of a young embryo immedi-
ately behind the site of the mouth, to show the area of growth in the entoblast pre-
paratory to the formation of the heart endothelium. The dotted circle indicates the
position in which the foregut appears in the next section forward.

FIG. 2, a, b, f, e. Transverse sections nos. 212, 224, 235, 251 of an embryo in
which the first gill slit has just made its appearance. Sections IO microns thick.

FIG. 3, a, b, c, d, e. Transverse sections nos. 273, 292, 306, 318, 325 of a later
embryo in which the separation of the endothelial cells from the entoblast is nearly
completed. Sections IO microns thick. The sections shown in Fig. 3, a, b, c, t are
approximately at the same levels as those shown in Fig. 2,a,b,c,e, respectively.

FIG. 4. Median sagittal section of the region between the mouth and liver of an
embryo of the same age as that shown in Fig. 3. Cell boundaries are shown
wherever they can be seen. The heart endothelial cells are evidently continuous
with the entoblast behind the mouth, but independent at all other points.

All figures were drawn with Zeiss apochromatic lenses and camera. Figures I, 2,
and 3 were drawn with 16 mm. objective and no. 4 ocular ; Fig. 4 with 8 mm.
objective and no. 4 ocular, and all have been reduced to one third in reproducing.

BIBLIOGRAPHY.
Brachet.

'98 Developpment du coeur chez les Amphibiens urodeles. Archives d'Anatomie

micr., T. 2, p. 251-304, 1896.
Morgan.

'97 The Development of the Frog's Egg, an Introduction to Experimental
Embryology. New York, 1897.

Ralb.

'86 Ueber die Bildung des Herzens der Amphibien. Morph. Jahrb., Bd. 12, pp.

252-273. 1886.
Schwink.

'91 Untersuchungen iber die Entwickelung des Endothels und der Blutkorperchen

der Amphibien. Morph. Jahrb., Bd. 17, pp. 288-333. l89I-
Ziegler.

'02 Lehrbuch der vergleichenden Entwickelungsgeschichte der niederen Wirbel-
tiere. Jena, 1902.

PERIODS OF SUSCEPTIBILITY IN THE DIFFEREN-
TIATION OF UNFERTILIZED EGGS
OF AMPHITRITE.

J. W. SCOTT.

While studying the unfertilized egg of Amphitrite at the Ma-
rine Biological Laboratory, Wood's Holl, Mass., I verified Fisch-
er's 1 result that the eggs could be caused to develop cilia by
squirting them from a pipette, by transferring them from one dish
to another, or by some other sort of mechanical agitation. I be-
lieve however that it is inadmissible to speak of this development
as parthenogenesis, meaning the production of a normal embryo
from an unfertilized egg. Ciliated, swimming structures result,
but their differentiation takes place with only partial or abnormal
and usually without any definite segmentation. I will discuss
the morphology of these processes in another paper. Lillie2 has
shown clearly a similar differentiation in the Ch&topterus egg.

In addition to Fischer's results, I found : (i) At least two criti-
cal periods in which the egg is highly susceptible to mechanical
stimulation, one period thirty to forty-five minutes, the other
eighty to one hundred minutes after they are removed from the
body and placed in sea-water ; (2) slight agitation is more effec-
tive in the second period than in the first ; rougher handling is
better in the first than in the second in which the eggs are more
easily broken into fragments. (3) Frequent and • moderate
squirting after thirty to fifty minutes seems more effective than
one hard squirting after the same time.

In my early experiments with certain salt solutions, the results
were sometimes discrepant, and there was great variableness in
the number of swimming eggs obtained under apparently identi-
cal conditions. About this time Fischer's paper came into my
hands. He had shown that " parthenogenetic development can
be produced by adding a small amount of Ca-salt to sea-water "

1 Fischer, Martin H., Am. Jour. Phys., 1902, III., p. 301.

2 Lillie, F. R., Archiv fur Entwickelungstriechanik der Organismen, 1902, XIV.,

P- 377-

35

36 J. W. SCOTT.

and by "mechanical agitation." "The unfertilized eggs of
Amphitrite" he says, " develop to the trochophore stage if, after
residence in sea-water from one half to one hour, they be squirted
from a small nozzled pipette into another dish of sea-water."
"The method is an uncertain one," depending upon "state of
ripeness," and a " previous residence in sea-water or in one of
the sea-water-salt solution mixtures is essential." He had
noticed that some eggs are very sensitive to " mechanical manip-
ulation," but rarely develop when treated " immediately after
they are cut out of the body of the animal."

Already convinced that there was a time-factor to be consid-
ered, I planned the following series of experiments. In each
series a set of eggs, removed from a single female at the same
time, was used. Due precautions were taken to prevent fertiliza-
tion by previously washing the female thoroughly in fresh water.
The hands of the operator, the dishes and pipettes used, were
carefully sterilized in the same way. For the same reason, sea-
water was used which had been raised to a temperature of 60° C.,
cooled and aerated. After washing in fresh water, the Amphi-
trite was placed in a dish of sterilized sea-water until the eggs
were removed. In the following four experiments the eggs were
removed from the female at 2.10 P. M. July 30, and were at once
transferred very carefully to fresh sterilized sea-water.

Experiment I. — The object of this experiment was to test the
effect of transferring from one dish to another. In order to get
a standard amount of agitation, the eggs were allowed to fall, one
drop at a time, from the mouth of a pipette held one inch above
the water. The different lots of eggs and the time each was trans-
ferred are given below :

1 control, transferred 2:IO P. M.

2 transferred 2:27 P. M.

3 " 2:43 P. M.

4 " 2:58 P. M.

5 " 3:13?. M.

6 " 3:43 P. M.

7 4:13 IJ- M.

8 4:43 P. M.

The dishes containing the transferred eggs were left undis-
turbed until 10 P. M., when eggs were taken from i, 2, 4, 6, 8

SUSCEPTIBILITY IN EGGS OF AMPHITRITE. 37

and examined. Care was taken to avoid disturbing those left in
the dishes.

The control showed nearly all eggs unchanged ; in a few the
germinal vesicle had broken down and they were darker (more
opaque) in color ; a few had started to segment.

Lot 2. The germinal vesicle had broken down in nearly all ;
a " perivitelline space" found in about 20 per cent., but was
rather small in most of this number. Most of the eggs were
light (translucent) in color.

Lot 4. An irregular "perivitelline space" in 40-50 per cent.
The germinal vesicle was broken down in practically all, the
light-colored as well as the dark. There were a few extra-ovates.

Lot 6. The germinal vesicle broken down in nearly all ; a
"perivitelline space" in 40—50 per cent., irregular in some; a
smaller number are blackened.

Lot 8. A prominent "perivitelline space" and contracted pro-
toplasm in 50—60 per cent.; the rest have the germinal vesi-
cle intact.

Anipliitrite eggs frequently begin differentiation if left in sea-
water entirely undisturbed. This is shown in the above control.
The experiment so far disclosed no marked phenomena, and I
give the above descriptions to indicate the comparatively uniform
development at this time. No evidence of normal segmentation
was found at any time in this and the three following experiments.

All the dishes were again examined at 9:30 the next morning,
as the advanced stages afford a better means of testing the effects
of transference. Below is given the estimated number of ciliated
eggs found in 2,OOO of each lot. Aside from the swimming eggs,
the different lots were in practically the same condition as on the
previous evening. No further description is then necessary.

Lot Time Transferred Number Ciliated

Number. from Beginning. in 2,000 Eggs.

1 O mill. O

2 I? " O

3 33 " i°

4 48 " 10

5 63 " 4

6 93 " 60

7 123 " 10

8 153 " 10

38 J. w. SCOTT.

Experiment 2. — The object of the experiment was to test the
effect of a more violent method of transferring. The eggs were
taken up in a pipette and squirted with moderate pressure into
the dish of sterilized sea-water from a distance of two or three
inches ; then water in the dish was taken up three times and
squirted at the surface. The control was simply transferred.

An examination of these eggs was made at 10:15 P. M., when
their condition was not much different from those in Experiment I,
except that more showed effects of the agitation. The next morn-
ing, 10:15 A. M., the following results were obtained :

Time Transferred
from Beginning.

Number Ciliated
in 2,000 Eggs.

O mill.

0

7 "

2

23 "

IO

38 •«

40

55 "

o

70 »

3

100 "

20

130 "

4

1 60 "

0

Lot
Number.

I control.

2

3

4

5

6

7

8

Experiment j. — Eggs were transferred in the following lots
and squirted moderately as in Experiment 2. Thereafter they
were squirted again moderately at frequent (10-15 min.) inter-
vals, up to 4:40 P. M. Examined at 9 A. M. July 3 i.

Lot Time Tr msferred Number Ciliated

Number. from Beginning. in 2,000 Eggs.

1 control. » O min. o

2 32 " 60

3 60 " o

4 i 20 " 4

Experiment 4. — The eggs were squirted violently, the water
vigorously agitated by squirting with a pipette, and then left
undisturbed.

Condition at 8:45 A. M., July 31. Many fragments present.

Lot Number. Time Transferred Number Ciliated

from Beginning. in 2,000 Eggs.

1 control. o min. o

2 33 " 40

3 60 " 2

4 i 20 " o

SUSCEPTIBILITY IN EGGS OF AMPHITRITE.

39

So

60

40

20

I have taken the above experiments as typical examples. I
have occasionally obtained a much larger per cent, of swimming
eggs, frequently a smaller number, and sometimes none. Accept-
ing the number of swimming structures as a fair test of develop-
ment of this kind, we may make again the following statements :

I. In the differentiation of unfertilized eggs of Amfhitrite, pro-
duced by transference, squirting or other methods of agitation,
there are at least two periods in which they are highly suscep-
tible, one thirty to forty-five minutes, the other eighty to one
hundred minutes after being put into sea-water.

I have attempted to depict this idea on ordinate paper, shown
in the accompanying figures. Abscissas represent time from
the beginning of an experiment, ordinates the relative number of

20 40 60 80 100 120 140

FIG. I. i. Gently transferred, experiment I. 2. Very moderately squirted,
experiment 26. 3. Moderately squirted, experiment 2. 4. Violently squirted,
experiment 4.

swimming eggs in 2,000 of each lot. Fig. i gives the results of
four experiments produced by different degrees of agitation.
Fig. 2 shows all the observations of these four experiments com-
bined in a single line ; where two observations were made at the
same time their average is taken (in one case only). The dotted
line gives my idea of the curve of susceptibility, as brought
about by a moderate degree of shaking.

2. By comparing experiments i, 2 and 3 (Figs, i, i, 2, 3) we
find slight agitation is more effective in the second period than in
the first ; rougher treatment causes more to develop in the first
period, but injures some in the second.

1 60

J. W. SCOTT.

80

60

40

20

3. Frequent and moderate squirting after thirty to forty-five
minutes seems more effective than one hard squirting, after the
same time. Compare experiments 3 and 4.

A comparison with fertilized eggs is of interest. The normal
egg throws off the first polar body in less than thirty minutes
after fertilization, and the first cleavage appears about thirty
minutes later. According to Loeb's l view, the sperm in the
case of parthenogenetic eggs acts simply to hasten, or accelerate,
processes which are already present in the egg. It hasf requently
been noticed that the unfertilized egg of Ampliitritc, if left undis-
turbed in sea-water, will often show some phenomena of differ-
entiation. Assuming Loeb's theory as a working hypothesis, we
should expect artificial means to be slower than fertilization.
This proves to be the case ; fertilized eggs develop cilia sooner

\

20

40

60 80

FIG. 2.

IOO

I2O

than those squirted. Sometimes there is not much difference in
AmpJiitrite. Presumably, then the two critical periods mentioned
correspond to processes in the normal egg that are active about
the time for the appearance of the first polar body and the first
cleavage ; there is the same relative time between them. Further
work is needed to prove this. Delage 2 states that the starfish
egg is highly susceptible to " artificial fertilization " between the
breaking down of the germinal vesicle and the appearance of the
first polar body.

However this may be, it is certain that there arc processes of
differentiation going on in the unfertilized eggs of Amphitrite which

'Loeb, J., Am. Jour, of P/iys., 1901, Vol. IV., No. IX.

2 Delage, Y., Archiv d. Zvol. Exper. et. Gen., 1901, T. IX., Nos. 2-3.

SUSCEPTIBILITY OF EGGS OF AMPHITRITE. 4!

may be started into activity at definite intervals by mechanical agi-
tation. These processes are, for the most part at least, indepen-
dent of the processes that cause segmentation. I have noticed,
as a rule, that the riper the eggs are the more cleavage is found,
but I am convinced that it is never normal beyond the first few
segmentations, if at all. It would seem, then, that the sperm
introduces the active cause of this process.

It has been shown by Lyon L in the fertilized Arbacia egg, that
there are recurring periods of susceptibility to KNC poisoning,
and to lack of oxygen. Each period of susceptibility is followed
by a period of resistance. On the other hand, in the unfertilized
eggs of Amphitrite, there are at certain times unstable conditions,
during which a small amount of agitation will set these unstable
forces free, and lead to some definite characteristics of more ad-
vanced development (/. e., production of cilia, etc.).
HULL ZOOLOGICAL LABORATORY,
UNIVERSITY OF CHICAGO.

, E. P., Am. Jour. ofP/iys., 1902, Vol. VII., No. I.

FURTHER STUDIES ON THE EFFECT OF VARIA-
TIONS IN THE TEMPERATURE ON
ANIMAL TISSUES.

ARTHUR W. GREELEY.

This paper contains an account of a series of experiments
which are the outcome of others of a similar nature described in
two earlier papers by the same author.1 This previous work has
called attention to the fact, noted by other observers, that the
fluidity of the protoplasm of any of the Protozoa studied, varies
directly with the temperature up to a certain critical point (28°
35° C), above which the protoplasm suddenly goes into heat
rigor, or coagulates. My own work has shown that as the tem-
perature is lowered below the normal, a similar coagulation sets
in which causes the cell to lose water.2 This loss of water and
coagulation is accompanied by a gradual cessation of the vital
activities of the cell, and brings about certain very definite mor-
phological changes that result in the formation of resting cells,
which consist only of an undifferentiated mass of protoplasm.
In the case of Monas, these changes were carried further by ex-
posing the cells to a still greater reduction in the temperature,
and the resting cells were finally broken up into many small
spores, each of which reproduced the motile organism when re-
turned to the normal temperature. As the temperature is raised
above the normal, the protoplasm takes up water and all its vital
activities are accelerated, until coagulation suddenly ensues at
the critical point.

1 Greeley, American Journal of Physiology, 1901, VI., p. 122. BIOLOGICAL
BULLETIN, 1902, III., p. 165.

2 This fact that lowering the temperature to 1° to 5° C. and raising the temperature
above the critical point has the same effect upon protoplasm (i. e., coagulation and
loss of water), has received an interesting verification in the recent work of Fischer
on Lepidoptera (All. Zdt. Jilr Kntomologic, October 15, 1901). In experiments on
the artificial production of seasonal varieties of Vanessa anteopa by exposing the larvje
to different degrees of temperature, Fischer discovered that precisely the same varia-
tions in the adult forms are produced by lowering the temperature to 1° C. or raising
it to 40° C. , while modifications in the temperature within those limits gave strikingly
different results.

42

EFFECT OF TEMPERATURE ON ANIMAL TISSUES. 43

In order to determine whether similar structural changes, as
have already been described in the cases of Stcntor and Monas,
could be produced on other forms as well, the low temperature
experiments have been continued on many other Protozoa, both
Infusoiia and Rhizopoda, and in all of them changes identical with
those described above have been obtained. Monas is the only
form in which it has been found possible to control the formation
of spores, but in all the others resting cells were formed at the
low temperature, which reverted to the motile condition when
restored to the normal temperature.

I. THE REVERSAL OF VITAL PHENOMENA BY A REDUCTION

OF THE TEMPERATURE.

The results of these low-temperature experiments on the Pro-
tozoa suggested an interesting comparison to the experiment of
Loeb's,1 in which the tentacles and polyps of a Campanularian
Hydroid were reduced to the undifferentiated protoplasm of the
stolon by bringing them in contact with some foreign substance.
It appeared that for the Protozoa a lowering of the temperature
as well as a contact stimulus brings about just such a reversal of
the vital phenomena until the undifferentiated resting cell is
formed, while a small increase in temperature accelerates the met-
abolic processes. To see if a lowering of the temperature brought
about similar changes in the more complex multicellular animals
a series of experiments was begun on the fresh-water Hydra.

It was at once observed that whenever a Hydra is exposed to
a temperature of 4° to 6° C., the tentacles gradually become
shorter and thicker, and are finally completely absorbed into the
body. As the absorption goes on, the ectoderm and entoderm
cells of the tentacles lose their individuality and form an undif-
ferentiated mass of protoplasm, which is slowly taken into the
body of the Hydra (see Fig. 4). The tentacleless body of the
Hydra becomes slowly resolved into a dense spherical mass of
coagulated protoplasm, in which no distinction between the in-
dividual cells can be made out, and remains in this condition as
long as it is kept at a low temperature (see Fig. 3), but quickly
forms tentacles and a double layer of cells again when it is re-

1 Loeb : American Journal of Ph ysiology, 1900. //". , p. 178.

44

ARTHUR W. GREELY.

turned to the temperature of the room. Thus a lowering of the
temperature seems to produce essentially the same effect on
Hydra as the contact stimulus on the Campanularian Hydroid in
Loeb's experiment. Likewise the structural changes appear to
be identical with those produced by the low temperature upon
the Protozoa.

Hydra react to variations in the temperature in another way
which is interesting when compared to the reactions of Protozoa

FIG. i. A budding Hydra after an exposure of twenty-four hours to a tempera
ture of 6°C. The body is slightly shortened and thickened, and the absorption of
the ectoderm and entoderm cells has begun in the tips of the tentacles.

under the same conditions. It has been a fact of common obser-
vation that the rate of cell division varies directly with the tem-
perature for all temperatures below the critical point. In my
experiments on Stcntor I showed that a lowering of the temper-
ature not only inhibits cell division but brings about the reverse
process. If a Stentor in the process of division be placed at a

EFFECT OF TEMPERATURE ON ANIMAL TISSUES.

45

temperature of 4°C. a fusion of the partially divided halves takes
place. Among Hydra the formation of buds, which finally
become distinct individuals, may be considered analogous to the
process of cell division among Protozoa. It was found that if a
Hydra in the earlier stages of the process of budding be placed

FIG. 2. The same Hydra as in Fig. I, after an exposure of six days to a tem-
perature of 6°C. The absorption of the tentacles and bud is nearly complete.

at a temperature of 4°C., not only does the growth of the bud
stop instantly, but an absorption of the bud into the body of the
parent commences, and continues until all traces of the bud have
disappeared. (See Figs. I and 2.) In order to demonstrate
this absorption of the bud, great care is needed in lowering the

FIG. 4. FIG. 3.

FIG. 3. The final resting stage of Hydra, formed after an exposure of four to
seven days to a temperature of 6°C. The body consists of an undifferentiated mass of
protoplasm.

FIG. 4. The tip of a tentacle of a Hydra that has been exposed to a tempera-
ture of 6°C. for twenty-four hours, showing the dissolution of the octoderm and
endoderm cells.

temperature. If the temperature is quickly reduced to i°C. the
Hydra go to pieces, but if the temperature be maintained at from
4° to 6°C., and is not suddenly varied in either direction, the
process of absorption can be easily seen. Six or seven days are
required for the complete disappearance of the bud. These two
experiments seem to show that among the Ccelenterates as well

46 ARTHUR W. GREELY.

as the Protozoa, a lowering of the temperature brings about a
reversal of the vital phenomena and the formation of simple rest-
ing stages.

II. THE EFFECT OF VARIATIONS IN THE TEMPERATURE UPON
DEVELOPMENTAL PROCESSES.

It has been frequently observed that eggs, spores, cysts, or
other resting stages of the motile organism which are formed to
tide over some unfavorable conditions in the life-history of the
animal, will not develop into the motile form unless they are ex-
posed to the very conditions that brought about their formation,
and normally intervene before development commences. Thus
Braem l found that the statoblasts of Bryozoa, and the winter
eggs of Apus would develop only after they had been exposed
to a certain degree of low temperature. In this case the phys-
ical change produced in the protoplasm of the egg or statoblast
by the low temperature seems to be necessary before the devel-
opmental processes can originate.

Dr. Loeb suggested to me the possibility that the same thing
might be true for the development and metamorphosis of the
chrysalids of a common moth, Cecropia, that are formed in the
Autumn, but do not complete the metamorphosis until the follow-
ing Spring. To test this hypothesis and see if other means be-
sides that of low temperature would suffice to start development,
the following series of experiments was performed :

On October 15, 1901, before they had been exposed to any
frosts, a large number of cocoons were brought into the labora-
tory. Many of these chrysalids were found to be parasitized by
an ichneumon fly, and only a small number were available for
the experiments. The cocoons were kept in the laboratory at
a temperature of 20° C. until November 27. They were then
divided into two lots. One lot was kept constantly at a tempera-
ture of 20° C., as a control series, and the other was placed out-
doors for six days, at a time when the temperature fell below o°
C. each night. At the end of the six days, these cocoons were
brought back into the laboratory, and kept with the others at a
temperature of 20° C. On January 27, the chrysalids that had

1 Braem, Jahrber, Settles. Ges. f. nat. Cult. (7,ool. Bot. Sec.), 1895, p. 2.

EFFECT OF TEMPERATURE ON ANIMAL TISSUES. 47

been exposed to the low temperature began to produce moths,
and all of them had completed the metamorphosis by February
3. None of the chrysalids that had been kept at a temperature
of 20° C. showed any signs of development. Several of the
cocoons were opened, but the chrysalids were in the same con-
dition, as far as could be seen, as when they were collected.

This result indicated that a lowering of the temperature at
least accelerates the metamorphosis of the chrysalids. To de-
termine whether the effect of the low temperature on the larva
consisted in an extraction of water from the protoplasms, as was
the case in the low temperature experiments on the Protozoa,
the experiment was now varied as follows : The cocoons, that
had been kept constantly at a temperature of 20° C., were now,
on February 3, divided into three lots. One lot was retained at
the room temperature, 2O°C., another lot was exposed outdoors
to a temperature of about — 10° C. for seven days, and the third
lot of four cocoons was placed in a desiccator over sulphuric
acid for two days. These four cocoons, while in the desiccator,
lost water as is shown by the following record of weights :

Weight when Placed in Weight when Removed from

Desiccator, Feb. 3. Desiccator, Feb. 5.

1. I7-0555 g- 17-031 g-

2. 15.184 g. 15.173 g.

3. 16.630 g. 16.603 g.

4- i5-"5 g- 15-095 g-

These four cocoons produced moths on March 4, 10, 13 and
14. On March 24, moths emerged from the cocoons of the
second lot that had been exposed to the low temperature, but
on March 26 the cocoons of the control series that had been
kept continuously at the room temperature produced moths
also, showing that this last exposure to a low temperature was
too late to have any effect. The desiccation hastened the devel-
opment by about two weeks. We see from this experiment that
the original exposure to a low temperature in November, soon
after the cocoons were first brought into the laboratory, hastened
the development by two months, and that the desiccation within
two months before all the cocoons produced moths sufficed to
accelerate the development materially. These experiments are
far from satisfactory because of lack of material, but they furnish

48 ARTHUR W. GREELY.

testimony to the conclusion already reached by Braem and others,
that in resting stages of this sort, development can commence
only after some physical change has occurred in the proto-
plasm through the action of a low temperature or other changes
in the external conditions. These experiments further seem to
show that the changes produced in the protoplasm by lowering
the temperature are identical with those produced by an extrac-
tion of water, as has already been indicated in the experiments
on Protozoa.

It is interesting to note that the same forms of stimuli (/. e., a
lowering of the temperature and an extraction of water), which
hasten the development of the moth, also produce artificial par-
thenogenesis "of the starfish egg. This fact lends weight to the
idea, expressed by Loeb, that artificial parthenogenesis consists
merely in the acceleration of developmental processes already
present in the egg.

III. EFFECT OF TEMPERATURE ON THE ABSORPTION
OF WATER BY MUSCLE.

If the conclusion drawn from these earlier experiments, that a
reduction of the temperature produces changes in the protoplasm
that cause it to lose water is true, then variations in the tem-
perature ought to have a decided effect on the absorption of
water by muscle or other animal tissue. The experiments of
Loeb1 and Webster2 on the gastrocnemius of the frog have
demonstrated that this muscle always behaves in a very constant
way, as far as can be determined by its change in weight, toward
each salt solution in which it is immersed. In some salt solu-
tions the muscle always absorbs a definite amount of water at the
normal temperature, in others of the same osmotic pressure it
always loses a definite amount. The only variation in this be-
havior of the muscle toward salt solutions occurs with the change
of seasons, the muscle of winter frogs differing widely from
summer frogs in this respect. This fact had been the only indi-
cation that temperature in any way affected the absorption of
water by the frog's muscle, and the influence of the temperature

1 Loeb, Archiv. f. d. ges. Physiol., 1899, LXXV., p. 303.

2 Webster, Univ. of Chicago Decennial Publications, 1902, X., p. 105.

EFFECT OF TEMPERATURE ON ANIMAL TISSUES. 49

alone was not clear in this case. In order to ascertain the influ-
ence of temperature upon this process and to obtain, if possible,
some quantitative estimate of its action, I started a series of
experiments to test the absorption of water by frog's muscle in
the same solution at different temperatures.

All the salt solutions were used at dilutions isotonic with ;«/8
NaCl which is supposed to represent as nearly as possible the
average osmotic pressure of the muscle. When tested at the
normal temperature (2O°C.), the solutions of all the salts experi-
mented with, fall into one of three classes : first, those solutions
which cause the muscle to absorb water as is shown by its in-
crease in weight, for example, the univalent salts, KC1 and
NH4C1, and salts with a bivalent anion and two univalent kations
as Na2SO4 ; second, those solutions which cause the muscle to
lose water as shown by its decrease in weight, for example, salts
with a bivalent kation and two univalent anions like CaCl2 or
or SrCl2 ; third, those solutions which leave the water content of
the muscle unaltered. LiCl is the best example of this third
class. ;;//8 NaCl usually falls in this group, although in my ex-
periments, I found that w/8 NaCl caused a slight increase in
weight, and that ;;//6 NaCl was more nearly isotonic with the
muscle.

The method used in the experiments was the same one that
has been elaborated so successfully by Webster.1 A large
amount of each solution was made up isotonic with ;///8 NaCl,
and was then divided among dishes which were kept constantly
at the following temperatures : i °C., 2O°C., 25°C., 27°C., 29°C.,
3 i °C, 36°C, 38°C., 45 °C. and 55°C. The gastrocnemius muscle
of the frog was used in the experiment. The muscles were care-
fully weighed and then distributed among the dishes containing
the solution to be tested at the temperatures named above. The
muscles were weighed after remaining in the solutions for three
hours, and again after twenty-four hours, and the gain or loss in
the water content calculated in percentages of the original weight
of the muscle.

The results of the experiments are given in Table I., in which
are given curves showing the effect of temperature on the absorp-

1 Webster, loc. cit.

ARTHUR W. GREELY.

tion or extraction of water after the twenty-four-hour exposure
to each of the solutions. The curves for the three-hour exposure
to the solutions are not given, because the twenty-four-hour

TABLE I. Curves showing the effect of temperature on the absorption of wat £
by the gastroc-nemius muscle of the frog in various solutions isotonic with m/8 NaCl.
The period of exposure to the solutions was twenty-four hours.

EFFECT OF TEMPERATURE ON ANIMAL TISSUES. 5 I

curves are entirely typical of the results obtained in both cases.

In considering the temperature effects we may classify them as
regards their bearing on the absorption phenomena in the three
classes of solutions mentioned above.

First, of those salts whose solutions, isotonic with m/S NaCl,
cause an absorption of water at the normal temperature (20° C),
the following were used : NaCl, KC1, NH4C1, Na2SO4, K2SO4
and K2C0O4. In all of them, as is shown by the curves, the ab-
sorption of water varies directly with the temperature up to a
certain critical point, in the neighborhood of 25°C., at which a
sudden loss of water commences which increases rapidly with a
further elevation of temperature. The form of the curve is the
same for all the solutions, regardless of whether the initial ab-
sorption is great or small, and at about 5o°C. the loss of water
becomes practically the same in all the solutions. The form of
these curves is strikingly like that showing the direct effect of
temperature upon the amount of water in protoplasm, indepen-
dently of the specific action of any salt solution, as is shown by
the curve for ;«/8 NaCl which approximates as nearly as possible
the fluid which normally bathes the muscle during life, as far as
its chemical composition is concerned. For these reasons it
seems probable that the effect of temperature upon the absorp-
tion of water in these solutions is due to the physical changes
induced by the variations in the temperature in the protoplasm
of the muscle. The rise in temperature may also accelerate the
specific chemical action of the solution upon the muscle proteids,
but in any event this only increases the result produced by the
temperature alone. The amount of water in the protoplasm of a
Protozoan varies directly with the temperature up to the critical
point which marks the beginning of heat rigor, and it is interest-
ing to find that the same thing occurs in muscle when immersed
in solutions which are isotonic with its own substance. Above
the critical point the heat rigor causes the same loss of water in
all the solutions regardless of their chemical composition.

The same temperature effects are still better shown in curves
of the absorption in solutions of the second class, /. t\, those
which cause neither a gain or loss of water in the muscle at the
normal temperature. Of the solutions of these salts, three were

52 ARTHUR W. GREELY.

used: LiCl, MgCl2 and ;«/6 NaCl. Usually MgCl2 has been
described to act like CaCl2, BaCl2 and SrCl2, in whose class it
would naturally fall, in causing a loss of water. But although I
tested its action many times, in all my experiments it had practi-
cally no effect on the weight of the muscle at the room tempera-
ture. It will be seen by examining the curves for LiCl and
MgCl,, that- at a temperature of i° C, the muscle loses a small
amount of water. This loss of water decreases steadily as the
temperature is raised until just above 20° C., an absorption of
water commences, which increases until the critical point is
reached. Above this point the muscles lose water very rapidly
just as in the other solutions. Thus in these solutions, which
appear to have no effect on the muscle at the normal tem-
perature, there is a loss of water at low temperatures, a gain at
temperatures between the normal and the critical point, and a
very rapid loss above the critical point, which is exactly the
effect that changes in temperature have been shown to have
on the protoplasm of the cells of Protozoa, when only the
physical condition of the protoplasm is modified by the varia-
tions in the temperature. ;// / 8 NaCl also should have no effect
on the amount of water in muscle at the room temperature, but
in many cases, especially with the muscles of winter frogs, this
solution appears to be hypotonic to the muscle substance. As
Webster has shown, the osmotic pressure of the muscle varies
with the season of the year, being higher during the winter,
which is the condition we should expect from the observed effect
of low temperatures on protoplasm. In my experiments (with
winter frogs), ;;//8 NaCl invariably caused a slight gain in weight,
but w/6 NaCl was found to be isotonic with the muscle, and the
curve for this solution corresponds exactly with that for LiCl.
Thus in all these solutions which appear to have no chemical
effect on the muscle, as far as can be determined by the changes
in weight, the amount of water in the muscle varies directly with
the temperature up to the critical point, and inversely with
the temperature above that point, and it is reasonable to suppose
that the same thing occurs in the muscle in its normal surround-
ings within the body.

The curves for CaCl2, BaCl2 and SrCl2 are very different from

EFFECT OF TEMPERATURE ON ANIMAL TISSUES. 53

those of the first two classes of salts. Solutions of these salts,
isotonic with ;«/8 NaCl, cause a loss in weight of about 20 per
cent, at a temperature of 20° C. It will be seen from the curves
that in each of these solutions the loss of water is very slight
at a temperature of i° C., and that the decrease in weight in-
creases steadily as the temperature is raised. But at a tempera-
ture of 50° the physical changes in the protoplasm overbalance
the specific action of any solution and the muscles lose practi-
cally the same amount of water in all solutions. It appears that
in the case of these solutions we are dealing with specific ion
effects, as Loeb1 has already suggested, and that the curve may be
interpreted as follows: The speed of any chemical combination
varies directly with the temperature. At a temperature of i°C.
the reaction between Ca, Ba or Sr and the muscle proteids is
so greatly slowed that the solution has no effect on the muscle,
and the small loss of water is due entirely to the physical changes
in the muscle produced by the low temperature, as in LiCl.
m 1 6 NaCl, and the other solutions which have no specific action
on the muscle substance. As the temperature is increased, the
reaction between the muscle proteids and the ions is accelerated,
and this chemical action of the Ca, Ba or Sr ion overcomes the
effect of the physical changes produced by the temperature, and
the loss of water steadily increases because these ion proteid com-
pounds, like Ca-soaps, hold very little water. It is worthy of
notice, however, that at about 25° C. there is a break in the con-
tinuity of the curves, corresponding with the rapid absorption of
water in the other solutions, which indicates a change in the
physical condition of the protoplasm that neutralizes temporarily
the specific ion effects.

In distilled water the amount of absorption by the muscle is
decreased by lowering the temperature, as is shown by the fol-
lowing result for a one-hour exposure to distilled water : Per-
centage of absorption

at 2° C. 40.7

at 30° C. 53.1

1 Loeb, loc. cit.

54 ARTHUR \V. GREELY.

SUMMARY.

1. In Hydra as well as Protozoa, a lowering of the tempera-
ture brings about certain definite structural changes that result in
the formation of an undifferentiated resting stage.

2. The inhibition of cell division and reversal of vital phe-
nomena by a reduction of the temperature is shown in Hydra by
the fact that at a temperature of 6° C., the growth of a new bud
ceases, and the partially formed bud is gradually absorbed into
the body of the parent animal.

3. A lowering of the temperature and an extraction of water
both bring about the same physical changes in the protoplasm
which serve to accelerate the development and metamorphosis of
the chrysalids of Cecropia.

4. The absorption of water by the gastrocnemius muscle of
the frog in those salt solutions which, when used at dilutions
isotonic with its own substance, either have no chemical effect on
the muscle at the room temperature, or cause an increase in
weight, varies directly with the temperature, until the critical
point is reached at which the muscle proteids begin to coagulate.
In solutions of the same osmotic pressure, which cause the muscle
to lose water at the room temperature, this loss of water varies
directly with the temperature. Above the critical point of tem-
perature the muscles lose practically the same amount of water

in all solutions, regardless of their initial effect on the muscle.
ZOOLOGICAL LABORATORY,

WASHINGTON UNIVERSITY, March 18, 1903.

THE EMBRYONIC DEVELOPMENT OF THE OVARY
AND TESTIS OF THE MAMMALIA. (PRELIM-
INARY ACCOUNT.)

BENNET M. ALLEN.

The following is a preliminary account of work begun in the
spring of 1900, upon the subject of the development of the ovary
and testis of the mammalia. The rabbit and pig have served as
the subjects for this work. The results will here be very briefly
set forth, a more detailed account being reserved for a later paper.
The material studied includes various stages in the development
of the ovary and testis of the rabbit, from the thirteen-day embryo
to and including adult stages. The pig material includes only
embryonic stages, but is more complete for the period covered
than is the rabbit material.

The work was carried on with the aim of solving the following
problems: (i) the origin and development of the seminiferous
tubules and their homologues in the ovary ; (2) the origin, devel-
opment, and homologies of the rete tubules, also their relations to
the Malpighian corpuscles of the mesonephros on the one hand,
and to the seminiferous tubules of the testis and medullary cords
of the ovary on the other ; (3) the origin, development and
homologies of the connective tissue elements and interstitial cells
of the ovary and testis.

Incidental to the solution of these problems, the work has in-
volved to a greater or less extent, a consideration of the follow-
ing allied problems: (i) the development of sex cells; (2) the
morphological phases of sex differentiation ; (3) cell degenera-
tion in the sex gland and rete region, (4) the degeneration of
the mesonephros, and the development of the Wolffian and Miil-
lerian ducts.

The results of work in so large a field can be, only to a limited
extent, new. Certain of the following results are confirmatory of
the work of other authors, to whose results I shall refer in a
later paper. The very contradictory opinions met with in the
literature on the subject call for confirmatory evidence upon these
problems.

55

56 BENNET M. ALLEN.

The earliest rudiment of the sex gland is situated in the genital
ridge, which consists of a zone of thickened peritoneum, running
the entire length of the mesonephros, parallel and close to the
mesentery which unites the latter to the body wall. The rete is
formed from the anterior part of the genital ridge and extends, in
the pig, from about the region of the sixth glomerulus to a point
about opposite the twentieth glomerulus, as shown in a number
of series. However, these limits are variable. In the rabbit the
limits are more constant, but still variable, the rete extending
approximately from the sixth to the twelfth glomerulus. The
anterior end of the sex gland rudiment slightly overlaps the pos-
terior end of the rete region. Posterior to the sex gland is a
section of the genital ridge that does not develop beyond the very
early formation of a region of thickened peritoneum. In each of
these three zones of the genital ridge are found scattered cells
with distinct cell walls, clear cytoplasm, large round nucleus,
centrosphere and centrosome — the so-called primitive sex cells.

In its early stages of development the genital ridge consists of
a thickened layer of peritoneum overlying a loose mesenchyma ;
the cells of the latter are to all appearances identical with those
of the peritoneum, from which they undoubtedly originate. This
resemblance applies both to the character of their nuclei and to
their lack of definite cell boundaries. The peritoneal layer and
underlying mesenchyma are separated from one another by the
basement membrane of the former. This is formed by the inter-
lacing of protoplasmic fibrils given off by the cells of both layers.
In later stages, one finds a progressive crowding of the peritoneal
nuclei. In the rete and sex gland regions this results in the for-
mation of tubular peritoneal invaginations, which are limited from
the surrounding mesenchyma by the persistent membrana propria.
In the sex gland region these tubular cords of cells may be known
as the sex-cords. At this indifferent stage they are closely massed
together side by side, and the very narrow interspaces between
them contain scattered mesenchyme cells, which from now on,
may be considered under the general term of stroma. This is
used to designate the loose connective tissue of both ovary and
testis. Invaginations, in all respects similar to these sex-cords,
arise from the peritoneum of the rete region. They lie further

DEVELOPMENT OF OVARY AND TESTIS. 57

apart than do the sex-cords, and penetrate more deeply into the
mesenchyme. Both rete-cords and sex-cords are at this stage
devoid of a lumen. In cross-section they show a limiting mem-
brana propria, within which is a single layer of cells arranged
with their bases attached to the membrana propria and their
apices meeting at a common point in the center — the rudiment
of the lumen.

The mesentery of the sex gland is formed by the proliferation
of connective tissue cells from the peritoneum, in areas of re-
stricted width immediately ventral and dorsal to the sex gland.
The ventral area is by far the more important of the two sources.
The albuginea is in large part formed from the cells comprising
the proximal parts of the sex-cords. They are formed at the
time when the sex gland has begun to assume definite shape.
At this time the rapidly dividing cells of the attached ends of the
sex-cords have become differentiated from those of the more
distal portions in that they elongate, become smaller, and acquire
the property of staining more deeply. They then break away
from the peritoneum on the one hand, and from the sex-cords on
the other. They may still for some time be found attached to
portions of the membrana propria that ensheathed them. They
become mingled with certain exactly similar intertubular mesen-
chymal elements, to form the albuginea, which is essentially one
with the remaining connective tissue or stroma of the sex gland.
Sexual differentiation is first manifested by the cessation of growth
of the sex-cords of the ovary. We can then distinguish them as
medullary-cords. The peritoneum of the ovary begins to in-
crease in thickness, and eventually forms the cortex of the adult
ovary in a manner to be briefly indicated below. The albuginea
of the ovary forms a broader, looser, and more irregular layer
than does that of the testis. In the testis the peritoneum ceases
to grow, in large measure at least, its cells becoming flattened,
and in later life practically disappearing. The rete cords grow
backwards from their points of origin, and enter the anterior
part of the sex gland. They branch and anastomose through-
out their course, sending branches to the Malpighian corpuscles
on the one hand, and on the other to the seminiferous tubules of
the testis. The branches passing to the Malpighian corpuscles

58 BENNET M. ALLEN.

meet evaginations from the capsules of Bowman, with which they
fuse. Such evaginations are irregular in number, as many as
three having been counted upon the same glomerulus. Some
glomeruli may send out none at all. The tubuli recti connecting
the rete-cords with the seminiferous tubules are likewise irregular
in number, being apparently called forth wherever needed.

Seminiferous tubules, medullary-cords, and rete-cords are
homologous structures. Not only are they of similar origin, but
their component cells show similarities. Two kinds of cells are
found in all three structures : (i) primitive sex cells, which have
already been described ; (2) cells more or less variable but
agreeing with one another in not having clearly marked cell
limits, and also in the absence of centrosphere and centrosome.
The cells of this second class form the germinative cells of the
seminiferous tubules, the follicular cells of the medullary-cords,
and the rete cells proper of the rete-tubules.

Returning to the subject of the ovary, the peritoneum at the
time of separation of the medullary-cords, contains no differen-
tiated sex cells. ' Such may exist, but they are at all events in-
distinguishable from the remaining peritoneal cells. The peri-
toneum or germinal epithelium, as it may now be termed, next
begins to give rise to the cords of Pfliiger, which branch and
anastomose in a similar manner to the medullary-cords. Some
of these cords of Pfliiger may contain a well-defined lumen, in
the case of advanced embryos of the pig (15 cm. length). In
these later stages the inner ends of the cords are broken up to
form follicles. Follicles are likewise formed in medullary-cords.
These however, have nevermore than one layer of follicular cells
in the forms studied.

The rete-cords come in contact with the medullary-cords, and
are then scarcely distinguishable from the latter in the case of
the rabbit. They contain no sex cells in later stages of the
ovary of that animal, although such are present in the rete
tissue when it is first laid down. By this criterion alone can
one, in a very general way, distinguish between medullary-cords
and rete-tissue lying within the rabbit's ovary. In the pig, on
the other hand, the rete-tissue shows some very interesting
characteristics. The portions of the rete-cords lying within the

DEVELOPMENT OF OVARY AND TESTIS. 59

ovary undergo similar differentiation to the medullary-cords
and the cords of Pfliiger forming the cortex. The rete-tissue
within the ovary of the 18 cm. pig embryo is found to contain
young follicles, each with a single layer of follicular cells ; the
enlarged oocyte in the center having passed through the synap-
sis condition, characteristic of one stage in the development of the
young oocytes. All such follicles subsequently degenerate. In
the testis the intra-glandular portions of the rete-tubules are simi-
lar to the seminiferous-tubules, but differ from them in their
much smaller diameter and in the earlier acquisition of a lumen.
They contain the sex cells characteristic of the seminiferous-
tubules. These are at first present in the extra-glandular region
of both ovary and testis, but disappear more or less completely
in later stages. No attempt was made to study out the fate
of the sex cells of the rete-tubules of the pig testis. They are
stjll present in the 25 cm. pig embryo. In the rabbit they are
found in the rete of the testis twenty-four days after birth, but
are not to be found in that of a rabbit killed 140 days after birth.
The rete-tubules are so completely united by anastomosis that
their connected lumen forms a large irregular cavity divided here
and there by irregular partitions formed by the walls of the sev-
eral rete-tubules.

The connective tissue elements of ovary and testis are derived
from the peritoneum. In early stages they are not distinguish-
able from the cells that make up the sex-cords, except that the
latter are marked off from the stroma by their membrana propria.
As before stated, the albuginea is largely formed by actual trans-
formation of the basal part of the sex cords into connective tissue
elements.

The interstitial cells are characterized by a large nucleus, dis-
tinct cell boundaries, a centrosphere and centrosome, and very
granular cytoplasm. They first appear in the stroma of both
testis and ovary of the pig of 2.5 cm. length. They are far more
numerous in the testis than in the ovary. Their appearance is
coincident with that of a large number of fatty globules in the
peritoneum and sex cords. In the testis they persist for a long
time. In the ovary, however, the few cells appearing at this
stage speedily disappear. In both organs they divide by mitosis.

6O BENNET M. ALLEN.

This process soon ceases in the ovary, while in the testis, on the
other hand, division figures are found in the interstitial cells at a
stage as late as the 7.5 cm. embryo. In the testis of the 15 cm.
embryo, they have begun1 to degenerate. This process manifests
itself in a shrinkage of the cytoplasm. Interstitial cells first form
in the testis of the rabbit embryo of a stage between seventeen
and twenty-one days. They are found to be still dividing by
mitosis eight days after birth. They are very rare, however, in
the stage of twenty-four days after birth.

No interstitial cells were found in the ovary of the embryo
rabbit, they being first met with in females killed forty-five days
after birth. Here they are scarce, but unmistakable. Consider-
able light is thrown upon their origin by a study of the eighty-five-
day rabbit. In the ovary of this stage they are very common,
their origin from the cells of the theca interna of atretic follicles
being clearly shown. This, taken in connection with the addi-
tional fact that they make their appearance in the 2.5 cm. pig
embryo coincident with the fatty degeneration of the germinative
cells of the seminiferous tubules and their ovarian homologues,
together with that of many cells of the germinal epithelium,
would lead us to conclude that cell degeneration offers the stimu-
lus or condition that brings about the formation of the interstitial
cells.

Interstitial cells do not develop from unmodified connective
tissue cells, such as those comprising the theca externa and the
general ovarian stroma. Such stroma cells must be transformed
into cells of the theca- interna by the direct or indirect influence of
the growing follicles, before they are again susceptible to the influ-
ences exerted by the process of cell degeneration. Atresia of
small follicles that are not surrounded by a theca-interna does not
bring about the formation of interstitial cells. Many such small
follicles are found to degenerate early and late in the history of
the ovary.

No evidence has been found favoring the theory of the early
segregation of the sex cells, but I am not prepared to say that
my work has in any way tended to disprove such a theory. Sex
cells appear in the very youngest stages studied (pig embryo,
0.6 cm. length and rabbit embryo of 13 days' age). They

DEVELOPMENT OF OVARY AND TESTIS. 6 1

are most common in the region where the sex-gland will even-
tually form, occurring both in the peritoneum and among scat-
tered subperitoneal cells of mesenchymal nature. They are prom-
inent in the sex-cords of a later stage. In the 1.8 cm. pig
embryo, immediately after the separation of the sex-cords from
the peritoneum, the latter is found to contain no sex-cells dis-
tinguishable as such. If the sex-gland be an ovary, they soon
(2.5 cm. pig embryo) make their appearance in the .peritoneum,
and especially in the cords of Pfliiger growing inward from it.
These cords of Pfliiger increase by growth at their bases, i. e.,
their points of connection with the peritoneum. Hence there is
a continual development of peritoneal cells to form the primitive
ova distinguishable as such. The case of the seminiferous tub-
ules is not so clear. Well-developed sex cells are found in them
from the start. On the other hand, all stages of transition are
found to link the germinative cells with the sex cells. These
transitional cells are found in the testis of the pig at as late a
stage as the 13 cm. embryo.

Whether the sex-cells that appear in the very early stages of
embryonic development ever produce functional sex products in
the testis, is a question that cannot easily be solved in this form.
Certain it is, however, that true sex products do form in both
ovary and testis from apparently undifferentiated cells of peri-
toneal origin, and that those which are functional in the ovary
form exclusively from this source. The sex-glands and rete tissue
are the seat of extensive processes of cell degeneration. I shall
not here enter upon a discussion of the different forms which
this process assumes, but shall defer treatment of these considera-
tions to the more complete account.

This piece of research has brought up many interesting facts,
bearing upon questions touching upon the action of trophic
stimuli in embryonic development. Perhaps the most striking
example of this is the formation of follicles in that portion of the
rete tissue lying within the ovary. The extra-ovarian part, or
that remaining in the mesonephros, does not contain follicles,
although it is of precisely the same origin as the intra-ovarian
portion. The influence of the ovary reaches out a short distance
into the mesonephros, as can be seen by the presence there of a

62 BEN NET M. ALLEN.

few sex cells, which are more numerous in the regions nearest to
the ovary. There is a definite interaction between the capsules
of Bowman and the rete-cords lying nearest to them. Each
sends forth a process to meet the other. In the testis the rete
cords send out processes (tubuli recti) to meet the seminiferous
tubules at their inner ends. Each tubule receives its rete branch.
A large number of tubuli recti can arise, from a single rete-tub-
ule at different points in its course. The connection between
cell degeneration and the formation of interstitial cells has already
been discussed. The uriniferous tubules of those glomeruli with
which the rete-cords come in contact persist as the rete efferentia
of the male, while the remaining ones disappear wholly, or in
large part at least. A few rudiments of these rete efferentia
tubules persist in contact with the rete ovarii. Such rudiments
are very rare in the pig embryo of 25 cm. length.

HULL ZOOLOGICAL LABORATORY,

UNIVERSITY OF CHICAGO, April 20, 1903.

Vol. V. July, I9°3- No. 2

BIOLOGICAL BULLETIN.

BUNODERA CORNUTA SP. NOV. : A NEW PARASITE

FROM THE CRAYFISH AND CERTAIN FISHES

OF LAKE CHAUTAUQUA, N. Y.

HENRY LESLIE OSBORN.

\

A trematode is met frequently at Chautauqua, New York,
which though already known seems never to have been critically
studied and described. While generically identical with B. nodu-
losa Zeder, of. Europe, it cannot be referred to the same species.
A form mentioned by Kellicott, '83, and Wright, '84, and Lin-
ton, '92, may be identical with it. I have not yet had access to
the articles of the first two writers, but Linton, '92, regards the
form they mention as identical with the one which he describes
from cysts from the ovary of crayfish from Alma, Michigan,
which, while it is much like B. nodulosa of Europe, he regards
as distinct, on account of the two lateral papillary appendages
projecting from the oral sucker and of triangular shape. Ward,
'94, reports at Ann Arbor, Mich., the form mentioned by Kelli-
cott, Wright and Linton, and considers it probably identical
with B. nodulosa of Europe. I am at present inclined to think
that of these four cases at least that of Linton is identical with
the Chautauqua form, and that the others may be. My knowl-
edge of B. nodulosa is almost entirely drawn from the account
of it in Looss' ('94) admirable monograph of the fish and frog
distomes, as I have not had access to specimens of that form.
A related trematode is described and figured by Linton, '98, from
the intestine of the lake sturgeon, and referred to B. auriculata
of Wedl, '57. A single specimen of the material on which Lin-
ton's account was based has been loaned me by the U. S.
National Museum through the kindness of Dr. C. W. Stiles,
and from such examination of it as I could make without injur-
ing it I was able to see that externally it is essentially like B.

63

O4 HENRY LESLIE OSBORN.

cornnta, excepting as regards the oral papillae. On this point and
in the figures of Linton, there is a divergence from either B.
nodnlosa, for the ventral papillae are transverse and in the form
of a horn, and from the Chautauqua form for the four anterior
papillae characteristic of both nodnlosa and cornnta are wholly
wanting. If the absence of these papillae is a constant character,
as at present it must be assumed to be, we then must accept
three species for this genus. The coarser features of the organi-
zation of the Chautauqua form is described in the following

i.mm

FIG. I. B. cornuta, ventral view, compressed, camera lucida X 3°- d.fj., duc-
tus ejaculatorius ; g.po., genital pore; I.e.. Laurer's canal; oes., resophagus ; ov.,
ovary; ph., pharynx; r.sem., seminal receptacle ; sp., spermary ; ttt., uterus ; T.S.,
ventral sucker; v.sem., seminal vesicle ; y.rec., yolk receptacle.

pages. An account of some points in the minuter structure is re-
served for a later article. It will be necessary and convenient at
least till more is known of B. anricnlata to adopt a name for the
Chautauqua species, and I propose for it the name B. cornuta.

NEW PARASITE FROM THE CRAYFISH.

The adult stage of B. cormtta is found at Chautauqua in the
stomachs of black-bass, rock-bass and cat-fish or bull-heads,
caught near the Assembly grounds, and earlier stages are found
encysted in crayfish, caught near the shore just above the grounds.
These localities are near the head of the lake. I have not ex-
plored the lake in other places and cannot say how generally the
fluke is found in it. The crayfish is clearly the host immedi-
ately prior to the fish, as partly digested crayfish are present
in the stomachs of fishes where the cysts and the young just
escaped flukes are found. The infection of the crayfish is prac-

o.lmtn

FIG. 2. Young worm still enclosed in cyst, X I2O> 'ne shaded area was opaque,
and white by reflected light.

tically universal. The flukes are always found encysted, never
free. They are located in the parts immediately related to the
reproductive system, most constantly in the muscles, especially
those running from the thorax to the abdomen, also in the heart
itself, and in the gonads. Remoter organs are not infected.
This mode of occurrence indicates that the infection may be
through the ducts of the gonads, but I have no observations to
decide this point. The number of cysts per individual varies
considerably, in one case 40 were found, distributed as follows :
25 in the muscles, 6 in the walls of the heart.^9 in the spermary;
in another case : 16 in the muscles, none in the heart or gonad ;
in still another a few were seen in the muscles and none in the
heart or gonads. The cysts vary somewhat in size and structure
with the season. In early July they are 0.9 mm. in diameter
and consist of a soft fleshy grayish enveloping portion about 0.2

66

HENRY LESLIE OSBORN.

mm. thick, enclosing a central mass, dark yellowish-brown and
hard, as if perhaps chitinous, of a diameter of 0.5 mm. By
manipulating the cysts with little knives made from specially
ground needles I found it possible to extract from them a very
immature fluke (see Fig. 2) recognizable as B. cornuta by its oral
sucker. A pair of eyes is present, but the inner organization
showed no traces. I suppose the dark granular mass at the pos-
terior end to be a supply of food for the developing worm.
Some of the cysts differ by having in place of the hard grandular
inner cyst a thin homogeneous covering, enclosing a worm so

o.p

o.i mm

ex.po

FIG. 3. Sagittal section, cam. luc. X 60.
o.p. , oral papilla ; pph,, pre-pharynx.

dr., cirrhus ; ex.bl., excretory bladder ;

much more advanced in development, that the alimentary and
excretory systems ^were formed and the genital organs well
advanced. In early August cysts having a diameter of I mm.
or over were found which contained fully matured worms, con-
taining embryos numerous enough to impart a distinctly brown
tinge to the parent. These facts are of very considerable interest,

NEW PARASITE FROM THE CRAYFISH. 67

for they indicate that the young worm develops actively during
.encystment, and that here self-fertilization must take place. A
fuller study of this point is desirable.

In the fish the parasite has been found only in the stomach.
Both cysts separated from the crayfish, and the free worms are
found. B. nodulosa is reported from the intestine of fishes and
B. auriculata is also an intestinal parasite.

The body form is nearly cylindrical, in contrast with the elon-
gate neck and almost leaf-shaped body of B. nodulosa. This
contrast is well seen by comparing Figs. I or 7 with Fig. 10 of
Looss. The latter is a young stage in which the vitellaria are
not as yet developed, while both of the Chautauqua specimens pos-
sess them and the uterus contains eggs. My specimens differ
considerably in length, owing to the fact that they go on growing

o j mm

FIG. 4. Transverse section passing through the oral sucker in the level of the lat-
eral processes. Cam. luc. X 5°- l.o.f., lateral oral papilla ; b.w., body wall,/<2r.,
parenchyma; o.s. , oral sucke'r.

longer after maturity. The longest one that I have seen meas-
ured 3.0 mm. in length by 0.9 mm. in width (in the preserved
and mounted state). I have seen specimens fully developed
sexually measuring only 0.9 mm. in length by 0.2 mm. in width.
The oral sucker is very large, so that it fills completely the an-
terior end of the body. It is furnished with remarkable muscular
processes, six in number which give the worm a very character-
istic appearance. Four of these processes or papillae are blunt,
and extend forward from the dorsal and anterior end of the
body. In a ventral view of the animal they are seen extend-
ing slightly beyond a thin layer of the body wall which forms the
anterior boundary of the body. The other two papilla; are at the
posterior level of the oral sucker, and ventral, on the opposite

68

HENRY LESLIE OSBORN.

side from the four blunt anterior papillae, and they are extended
transversely to the animal. In form they are tapering and
pointed, and slightly curved backward, in the form of a horn,
extending considerably beyond the contour of the side of the
animal. The oral sucker itself has a diameter of 0.4 mm. The
ventral sucker, while large, is smaller than the oral sucker, its
diameter being 0.3 mm. Its position in Fig. I is strikingly far
forward ; in Fig. 7 it is more nearly in the center of the body.
This difference is due to contraction of the neck in Fig. I, shown
also by the winding course of the oesophagus of that specimen.
The genital pore is located in front of and near to the ventral
sucker, in the middle line. Eyes are present in younger speci-
mens, but older ones do not possess them, though in these it is

- mu dv
par

mu czr
mu Ion

oj mm

FIG. 5. Transverse section in level of the posterior spermary. Cam. luc. X II5-
vt., vitellaria ; mu.d.v., dorsiventral muscle; mti.cir., circular muscle of the body
wall ; mu.loH., longitudinal muscle of the body wall.

possible to find scattered grains of pigment in the region of
the pharynx, indicating their late disappearance. Looss repre-
sents eyes in both the specimens figured by him, so that if they
are not exceptional cases, the eyes persist to a much later time
in the life history of B. nodulosa than of B. cornuta.

The body wall presents the usual cuticle, destitute of spines.
The usual muscle layers are present, .the fibers of the outer circu-
lar layer are very fine indeed ; those of the longitudinal and oblique
layers are exceptionally large. Parenchymatous muscle is some-
what specially collected in each side of the body running dorso-

NEW PARASITE FROM THE CRAYFISH.

69

ventrally, and marking off a lateral area, containing the vitellaria
and the intestine, from the'center (see Fig. 5). There are no
horizontal parenchyma muscles. Cells of the parenchyma directly
underlying the body wall are especially numerous and glandular
in appearance, as often in trematodes.

The oral sucker opens widely downwards and forwards. It is
composed of the usual muscular masses, enclosed within a fine
structureless membrane marking it off from the parenchyma.
The detailed struct me of the papillae is indicated in Fig. 4, which
is a camera lucida drawing from a section passing through the

gpo

•ar

ov-

sp

FIG. 6. General view of the reproductive organs, seen dorsally.

oral sucker in the level of the lateral papillae and through their
length. Sections through the four anterior papillae show the
same things. The wall of the oral sucker, consisting of a layer
of parenchymatous tissue and masses of radial fibers, is pushed
out at the bases of the papillae, the connective tissue portion be-
ing directly continuous and the muscle fibers pushed aside at
that point, and passes up to the summit of the papilla, a new set
of radial fibers being added in the papilla similar to those of the
general wall of the sucker. The papillae are thus not merely

HENRY LESLIE OSBORN.

surface features of the animal, but deep-seated in their origin,
and are entitled to be regarded as of considerable importance
from a taxonomic point of view.

The lateral papillae are unmistakable organs not likely to be
overlooked by an observer, whereas the ventral papillae of B.
nodulosa are inconspicuous and might easily escape notice, a
point discussed by Looss ('94, p. 34). In B. anriailata I looked
very closely for the dorsal papillae without finding them and I am

FIG. 7. Ventral view of a specimen in which the uterus is most fully developed,
the vitellaria are omitted from the right side.

convinced that they are absent from the specimen I saw, in which
respect my observations confirm those of Linton as indicated in
his figures, '98, PI. XLV., Figs. 1-7.

There is a short pre-pharynx, a small pharynx, about 0.05
mm. long, a short oesophagus, its walls very strongly muscular
and surrounded by glandular cells. The forking of the intestines
is thus close to the pharynx, a point different from B. nodulosa.
The intestines are simple and long, reaching to near the hinder
end of the bod}-. They are lined with epithelium cells whose
outer ends are elongate and whose tips extend into the cavity of
the organ. Circular and longitudinal muscle fibers are present
in the wall.

NEW PARASITE FROM THE CRAYFISH. /I

The excretory pore is terminal. Close in front of it is the excre-
tory bladder, which in sections can be seen running dorsally for-
ward at least as far as the level of the front of the anterior testis
(see Figs. 2, 5). I have not been able to recognize more than the
most posterior portion in living animals. According to Looss
there is in B. nodulosa a bladder wholly posterior to the hinder
testis containing concretions, and from which vessels run forward
on either side. I have not seen such concretions in B. cornuta
and the bladder is much more extensive than that.

There are two large testes, 0.3 mm. across, lying one directly
in front of the other. In B. nodulosa the testes are smaller,
farther apart and oblique. The testes are crowded with active
sperm cells, many of them in the last stages of spermatogenesis,
and with numerous fully formed spermatozoa. The seminal vesicle
and spermatic receptacle are also filled with them.

Long and slender vasa deferentia run dorsally to the other
genital organs, and meet at the posterior end of the large cirrhus
sack which is located some distance behind the posterior border of
the ventral sucker. The cirrhus sack is very large indeed, much
larger than in B. nodulosa. It has a definite outer wall, strongly
muscular, enclosing a tubular passage subdivided into two por-
tions, a posterior thin-walled part, the seminal vesicle, and an
anterior ductus ejaculatorius. This latter is surronded by glan-
dular "prostate" cells, is very strongly muscular, having both
' circular and longitudinal fibers. The ductus ejaculatorius is not
coiled. I do not know whether it is eversible or not.

The ovary is generally located on the right side, but not infre-
quently it is found on the left (cf. Figs. I and 7). It is always
near the ventral sucker, a large and conspicuous organ. There
is a short ciliated oviduct, soon joined by first a duct from the
seminal receptacle, then one from the yolk receptacle. Certain
glandular-looking cells which lie around the oviduct may per-
haps represent a shell gland, but a distinct and well-marked
organ is not present. Nearly all of my specimens appear to be
quite young, and though the uterus contains eggs it is not fully
developed. In one, however (Fig. 7), the uterus is longer and
evidently more as in fully matured individuals. In this case the
uterus is distinctly tubular and winds down and back, passing

72 HENRY LESLIE OSBORN.

between the testes in its course, in B. nodulosa the uterus is a
large sack containing old and young eggs indiscriminately, the
uterus is saccular even in young individuals of B. nodnlosa, as
seen in Looss' Fig. 10 of a specimen before egg production has
begun. The terminal part of the uterus differs decidedly from
the rest so as to form an entirely distinct though continuous
organ (cf. Fig. 2.) Its wall is very thick indeed and consists of
a strong muscular coat quite unlike the wall of the deeper parts
of the tube, and within the wall is furnished with a clothing of
very peculiar numerous long slender bluntly ending processes,
which are free at tip in the cavity of the organ. These structures
do not look like cilia, being too blunt. They do not seem certainly
to be protoplasmic, at least the bases do not seem — as far as I have
been able to study them - - to be nucleated cells, as we should ex-
pect. The histological structure of this part will have to be left
for a subsequent study. This organ is further surrounded by
parenchyma cells having much the same appearance as the pros-
tate cells of the cirrhus. The eggs measure 0.07 mm. in length
instead of o. I mm. as in B. iwdiilosa.

Laurer's canal is present, passes dorsally and opens to the ex-
terior on the left side. The seminal receptacle is large and dis-
tinct ; it lies close to and just behind the ovary. It is in all cases
of adults filled with spermatozoa. The vitellaria are large, and
located as above described laterally and so as partly to envelope
the intestines. They extend from near the phaynx to near
the hind end of the body and consist of very numerous small
follicles uniformly distributed. A duct from each crosses the
body in front of the anterior testis and behind the ovary and
seminal receptacle and the two joining from the yolk receptacle
which reaches the oviduct by a short duct close to the ovary.

The points made in the foregoing pages are summarized in the
following table of comparisons :

B. conntta. B. nodnlosa.

Total length, 3 mm., 3 mm. Looss, 4.5 mm.

Olsson.

Body form, cylindrical, leaf-shaped.

Neck, not prominent, prominent and distinct.

NEW PARASITE FROM THE CRAYFISH. 73

Lateral papillze, transverse and hook- longitudinal and blunt,

shaped, not hook-shaped.

Eyes, not persistent, persistent in adult.

Oesophagus, short, long.

Excretory long, short,

bladder,

Testes, close together in me- wide apart, and oblique.

dian line,

Uterus, tubular, saccular.

Ova, 0.07 mm. long, o. i mm. long,

Residing, in stomach, in intestine of host.

BIOLOGICAL LABORATORY, HAMLINE UNIVERSITY,
SAINT PAUL, MINN., March 10, 1903.

LITERATURE CITED.
Kellicott, D. S.

'83 Trematodes of the Crayfish. Proc. Am. Micros. Soc., p. 115.
Linton, E.

'92 Notice of Trematode Parasites in the Crayfish. Am. Nat., XXVI., pp.
69-70.

'98 Notes of Trematode Parasites of Fishes. Proc. U. S. Nat. Mus., XX., pi.
xvi., Figs. 1-7.

Looss, A.

'94 Die Distomen uns. Fischen u. Frosche. Bibl. Zool. Leukart u. Kuhn, XVI.

'99 Weitere Beitrag. Kentn. Trematoden Fauna. yEgyptens. Zool. Jahrb. Syst.,
XII.

Osborn, H. L.

'02 Notes on the Trematodes of Lake Chautauqua, N. Y. Science, XV., p. 573.
Pratt, H. S.

'03 Synopsis N. A. Invert. Trematodes. Am. Nat., XXXVI.

Stiles, C. W., and Hassal, A.

'98 An Inventory of Fasciolidoe. Arch. Parasitol., I., p. Si.

Ward, H. B.

'94 On the Parasites of the Lake Fish. L, On D. opacum. Proc. Am. Micros.
Soc., XV., pp. 173-182.

Wedl, K.

'58 Anat. Beob. u. Trematoden. Sitz. a. k. Akad. Wien., XXVI., p. 242.

Wright, R. R.

'84 Trematode Parasites of the Crayfish. Am. Nat., XVII., pp. 429-430.

ON THE BLOOD VESSELS, THEIR VALVES AND

THE COURSE OF THE BLOOD

IN LUMBRICUS.1

J. B. JOHNSTON.

In a previous paper - an account has been given of the experi-
mental study of the course of the blood flow in Lumbricus
The most important result there set forth was that the circula-
tion is not segmental but strictly systemic. The course of the
flow is as follows : forward in the dorsal vessel for its whole
length ; downward in the hearts ; both forward and backward
from the hearts in the ventral vessel ; outward from the ventral
to the body wall, nephridia and intestinal wall ; toward the lat-
eral neurals from the body wall ; backward in the subneural ;
upward to the dorsal vessel in the parietals from the subneural,
the nephridia, and the body wall, and in the dorso-intestinals
from the intestine. Thus, there is no circuit of blood in each
segment to which a sytemic circuit for part of the blood has
been superadded, as all previous authors have maintained, but
all of the blood flows in a single systerryc circuit. In the head
region the blood is carried forward beyond the hearts by both
dorsal and ventral vessels and is returned to the dorsal behind
the hearts in larger part by the lateral cesophageals, and in smaller
part by the subneural and the parietals of XII. and succeeding
somites. The lateral cesophageal system is considered to repre-
sent the parietals in the somites anterior to XII.

This view of the circulation raised two important questions
which further work has answered : (i) What happens when the
hearts are removed from the circulation by decapitating the
worm ? Do the conditions which obtain in the regenerating

1 Studies from the Zoological Laboratory of West Virginia University, No. 8,
February 28, 1903. A part of the work reported here was done by my former stu-
dent, Miss S. W. Johnson. For the conclusions reached the present writer alone is
responsible.

'2 " The Course of the Blood Flow in Lumbricus," by J. B. Johnston and Sarah
W. Johnson, Amer. Naturalist, April, 1902.

74

BLOOD VESSELS OF LUMBRICUS. /5

worm confirm the above results ? (2) What is there in the
structure of the blood vessels to determine and control the
course of the blood ?

The first question has been answered by a series of regenera-
tion experiments carried out upon large and small specimens of
Lwnbricits. Operations removing from eleven to twenty somites
from the anterior end were performed upon 171 worms. These
were examined alive from time to time and eventually 20 were
hardened for sectioning. The time that the worms were allowed
to live varied from ten days to three and a half months. In a
few worms regeneration progressed well, but the majority died
after a few days or weeks. A detailed report upon these experi-
ments would not be profitable for our present object. Although
there were very great variations in the condition of the blood
vessels, the following may be said to be true in greater or less
degree of all the worms studied alive or sectioned. The vessels
in the anterior one fourth to one half of the worm were greatly
crowded and distended with blood. The anterior portion of the
worm was usually a bright red to the naked eye and under a
lens many small vessels not usually visible were distinctly seen.
Sections showed that all the vessels were more or less crowded
with blood, while the dorsal, subneural, and the vascular plexus
of the intestinal wall showed the greatest distension. The ven-
tral vessel was seldom stretched much beyond its normal size,
while the subneural was often as great in diameter as the ventral.
Occasionally the subneural was much larger than the ventral and
sometimes its cross-section was equal to that of the nerve cord.
In several cases the vascular layer of the intestine was very
greatly crowded and, considering its great capacity in normal
conditions, it is probable that it always held the greatest accu-
mulation of blood. The posterior portion was very poor in blood
in all worms.

These conditions are readily explained in accordance with the
scheme of circulation above summarized. The fulness of the
dorsal, intestinal and subneural vessels is due to the pressure
from the dorsal which is deprived of the normal outlet for the
blood carried by it, and forces the blood downward in the dorso-
intestinals and parietals contrary to its usual course. The small

76 J. B. JOHNSTON.

amount of blood in the ventral is due to the absence of the hearts
and the inability of the dorsal to drive the blood through the
capillary systems to the ventral. The absence of blood in the
posterior end is a further result of the small amount of blood
received by the ventral. If there were a segmental circulation
in Lnnibricns there would probably be no great accumulation of

FIG. I. General scheme of circulation in body region, all the vessels of one seg-
ment being projected upon the plane of a transverse section. The vascular layer of
the intestine is shown by a broad black line. In the typhlosole the vascular plexus
thickens at one place to form the typhlosolar sinus which varies greatly in size in
different worms and in different parts of the same worm. From this sinus three dorso-
typhlosolar vessels in each segment carry blood to the dorsal vessel. These vessels
and the branching of the dorso-intestinal vessels shown in this figure have not before
been correctly described or figured for Lunibritiis.

blood at the anterior end in these experiments, since the seg-
mental circulation would tend to relieve the systemic and the even
distribution of the blood would be maintained in accordance with
the law of least resistance. These regeneration experiments,

BLOOD VESSELS OF LUMBRICUS. 77

therefore, seem to confirm the results of physiological experi-
mentation. No effort was made to trace the development of
hearts in the regenerated heads and the final reorganization of
the circulation, and it is doubtful whether the worms would have
lived long enough for this purpose.

It is probable that the failure of the blood vessels to adjust
themselves to the new conditions is at least one of thec chief
causes of the death of worms under these experiments. The
continued strong pulsations of the dorsal vessel after the
removal of the hearts force the blood out through vessels which
normally empty into it. In some cases the reversal of flow
through the vessels of the body wall and intestines is produced
readily enough to allow the worm to survive the operation, but in
most cases less blood would reach the ventral vessel than is
necessary to supply the posterior end of the worm and an insuf-
ficient amount of blood would pass through the respiratory plexus
beneath the hypodermis. The blood which leaves the dorsal
vessel in the anterior part of the worm either settles in the vas-
cular layer of the intestine, which readily expands to receive it,
or passes directly through the parietals to the subneural, which
is consequently greatly expanded ; and avoids the respiratory
plexus because of the resistance in that quarter. A similar with-
drawal of blood from the respiratory plexus of the posterior end
of the worm also results indirectly from the small amount of
blood in that region, so that the whole worm is seriously deprived
of needed oxygen. In the normal circulation the blood is driven
to the respiratory plexus from the ventral under direct pressure
from the hearts, and there is no other way of less resistance by
which the blood may return from the ventral to the dorsal.
Upon the view of the circulation held by Bourne ' and Harring-
ton,2 namely, that the dorso-intestinals empty into the dorsal ves-
sel and the parietals carry blood away from it, it is evident that
the path of least resistance from the pulsating dorsal vessel is
through the parietals directly to the subneural and that there
would be nothing to drive the blood through the respiratory

1 Bourne, A. C., "On Megascolex cceruleus and a Theory of the Course of the
Blood in Earthworms," Q. J. M. S., Vol. 32, p. 49, 1891.

2 Harrington, N. R. , " The Calciferous Glands of the Earthworm, with an Appen-
dix on the Circulation," Jour. Morph., Vol. 15, Suppl., p. 105, 1899.

J. B. JOHNSTON.

FlG. 2. A diagrammatic horizontal section of the dorsal vessel and those emptying
into it. One somite and part of a second are shown, and the last part of a pulse
wave and the greater part of a following contraction wave are represented. The
arrows show the course of the blood and the position and changes of form of the valves
are shown. The chloragogue cells covering the walls of the vessels are not drawn.

BLOOD VESSELS OF LUMBRICUS, /Q

capillaries of the body wall. This is perhaps an insuperable
objection to that theory of the circulation. This objection does
hold against the view of Perrier1 and Benham,2 according to
which the blood flows to the dorsal vessel in the parietals and
out from it in the dorso-intestinals.

The study of the structure of the vessels shows that the move-
ment of the blood is determined by the structure of the walls
and by definite valves which several of the vessels possess. The
wall of the dorsal vessel consists (Fig. 2) of a lining endothelium
of very thin cells whose nuclei alone are usually visible ; a con-
nective tissue layer containing a few longitudinal (muscle?) fibers,
and a well-developed layer of circular muscle fibers. Outside
these are the chloragogue cells. To the layer of circular muscle
fibers are due the pulsations of the dorsal vessel, and thickenings
of this layer at certain points assist in the action of the valves,
as will be described below. The wall of the ventral vessel has
no circular muscle layer. Its lining endothelium is more con-
spicuous than that of the dorsal vessel and the connective tissue
layer is very thick. This is a strong fibrous layer and gives
great rigidity to the wall of the ventral vessel. Outside of the
connective tissue layer are a few (4 to 6) strands of longitudinal
fibers which take the same stain as the muscle fibers in the
sheath of the neighboring nerve cord. Outside these fibers is a
layer of peritoneum closely similar to that covering the inner
surface of the body wall.

The subneural consists of only the endothelium and connective
tissue layer, outside of which is the sheath of the nerve cord.
This is the structure of the lateral neurals also, and of all the
smaller vessels. The dorso-intestinals and parietals present an
intermediate condition between those with and those without a
circular muscle layer. The dorso-intestinal vessels are devoid
of muscle fibers except at their dorsal ends where there is a thin
extension of the circular layer of the dorsal vessel for a short
distance. The parietals are provided with a thick band of circular
fibres close to their connections with the dorsal and the layer is

1 Perrier, Edw., " Recherches pour servir a I'histoire des Lombriciens terrestres,"
Nonv. Arch, du Mus. d' Hist. Nat., Paris, Tome 8, 1872.

2 Benham, W. B., "The Nephridium of Lnmbricus and its Blood Supply," Q. J.
M. S., Vol. 32, p. 293, 1891.

8O J. B. JOHNSTON.

continued along the vessel for about a third or half its length

O

These muscle fibers in the dorsal portion of the parietals produce
the active pulsations which have been described in an earlier
paper (loc. cit., p. 323).

The walls of the hearts have the same structure as that of the
dorsal vessel, except that they are covered with chloragogue
cells only in their dorsal portion, elsewhere by peritoneum. The
circular muscle fibers are large and the layer somewhat stronger
than that in the dorsal vessel.

The structure of the vessels determines whether they shall pro-
pel the blood by their pulsations or only carry it, and the ac-
count of the structure accords with the well-known facts con-
cerning the pulsations of the vessels. Pulsations in the dorsal,
parietals and hearts are well established ; pulsations in other
vessels, described by Harrington, have not been seen by the author
and to whatever extent they occur they must be produced with-
out muscle fibers.

Valves are present in the dorsal vessel and in all the vessels
connected with it, namely, the dorso-intestinals, dorso-typhloso-
lars, parietals, lateral oesophageals (?) and hearts. The valves
in the dorsal are a pair of large thick flaps attached to the lateral
walls of the vessel at a point a short distance behind each septum
and immediately behind the openings cf the parietals. These
valves are always directed forward and allow the free passage
of blood during the pulse wave. As the contraction wave ap-
proaches, the valves are brought into contact and at the moment
of greatest constriction the two flaps are tightly pressed together
and completely close the lumen of the vessel. The efficiency of
the valves is secured and increased by a considerable thickening
of the circular muscle layer at the valve (Fig. 2). The valves
do not act in the ordinary manner of flap valves, but the two
fleshy flaps are pressed together and form a large mass which
fills the vessel. In the region of the hearts a pair of valves is
found in the dorsal vessel a short distance in front of each pair
of hearts.

The valves in the dorso-intestinal, dorso typhlosolar and pari-
etal vessels are essentially the same in form and position. In
each of these vessels (Fig. 2) a pair of small fleshy flaps are sit-

BLOOD VESSELS OF LUMBRICUS. 8 1

uated at the opening of the vessel into the dorsal. In the dorso-
intestinal and parietal vessels the flaps are attached one to the
anterior and one to the posterior wall of the vessel, and the body
of the flap projects into the lumen of the dorsal vessel. In the
dorso-typhlosolars the flaps are lateral in position, are situated
deeper in the vessels, and do not project so far into the dorsal.
It is evident that pressure from the dorsal toward any of these
vessels would tend to close the valves. The closing of the pari-
etals is further secured by a thickening of the circular muscle
layer as in the dorsal ; and in the dorso-intestinals a thin exten-
sion of the muscle layer of the dorsal serves the same purpose.
Muscle fibers have not been observed in the dorso-typhlosolar
vessels. The valves in these vessels allow the blood to flow
from them into the dorsal only, and this accords with the results
obtained by the earlier experimental investigation. In the case
of the decapitated worms the valves in all these vessels near the
anterior end must have been forced.

The hearts are better supplied with valves than are any of the
other vessels. In each heart are four pairs of valves (Fig. 3);
one situated close to the dorsal vessel, one between the first and
second thirds from the dorsal end, one between the second and
third thirds, and the fourth in the ventral end of the heart at the
opening into the ventral vessel. The three pairs in the body of
the heart are like those in the dorsal vessel but are smaller in
proportion to the diameter of the heart. They are inclined
downward and are large enough to close the heart during its
contraction. The presence of these valves might seem unneces-
sary in view of the fact that the contraction waves pass along the
heart from above downward. However, if from any cause the
contraction becomes modified or irregular or if the whole heart
contracts at once, the functional importance of these valves is
evident. It is a matter of common observation that such irregu-
larities in the contractions of both the hearts and the dorsal ves-
sel do appear in worms dissected alive under an anaesthetic, and
it is probable that such irregular contractions and the influence
of movements of the body make necessary the valves in the
hearts and the dorsal vessel in the normal worm. The valves in
the smaller ventral ends of the hearts fill the lumen and project

82

J. B. JOHNSTON.

into the ventral vessel very much as the valves in the parietal s
project into the dorsal. Thus, with the valves in the dorsal be-
tween each two pairs of hearts and the four valves in each heart,
regurgitation of blood during the strong cardiac contractions is
effectively guarded against.

The study of the fine structure of the valves has presented
great difficulties because methods of fixation which give satisfac-

Ti3.3.

FIG. 3. A diagrammatic cross-section through one of the hearts to show the posi-
tion of the valves. The chloragogue and peritoneal epithelium are not drawn.

tory preparations of all other tissues give very imperfect pictures
of these valves. This itself indicates one fact regarding their
structure, namely, that they are composed of very soft-bodied or
watery cells which may appear vacuolated or shrunken, or even
macerated. In many preparations the valves appear only as
masses of granular or coagulated material containing many ovoid
nuclei. In the most successful sections, however, the valves

BLOOD VESSELS OF LUMBRICUS. 83

show indistinct cell boundaries which produce an appearance of
striations running from base to free edge of the valve. In most
preparations, especially in longitudinal sections of the dorsal ves-
sel, which are often oblique owing to the curves of the vessel, the
substance of the valves appears to be sharply delimited from the
connective tissue layer. This would indicate that the valve is
formed by a thickening of the endothelial layer. It is difficult to
disprove this first supposition because the endothelial cells are so
broad that one can seldom expect to find an endothelial nucleus
on the surface of a valve. However, in some cases in the hearts
flattened nuclei similar to those of the endothelial cells are found
on the surface of the valves. Cross-sections of the dorsal vessel
through the base of the valves show a radial striation running
from the valves through the connective tissue and muscle layers.
From these facts it appears that the valves are composed of elon-
gated cells which run through the connective tissue layer and
securely anchor the valves. Since they are covered internally by
endothelial cells they must be regarded as belonging to the con-
nective tissue layer. Essentially the same structure is presented
by all the valves, although those in different positions differ
greatly in size and form in relation to the function which they
have to perform. The largest valves are those in the dorsal ves-
sel. These are thick flaps attached by broad bases to the lateral
walls of the vessel. When the vessel is distended the valves are
nearly semilunar in form. When the vessel is contracted the
valves become greatly compressed against one another and the
soft substance of the valve is forced both forward and backward
from its point of attachment. When the valve extends far for-
ward it overlaps the opening of the parietal vessel and might ap-
pear to function to close that vessel. Such a condition seems to
have been seen by Benham (loc. cit.). The valves in the dorso-
intestinal and parietal vessels are also paired flaps, but owing to
the small size of the vessels the flaps are small at their bases and
are longer than they are broad. Often these valves' have a
balloon form as they project into the dorsal vessel. The valves
in the dorso-typhlosolar vessels are situated somewhat farther
within the vessels and are more nearly simple semicircular flaps.
The valves at the ventral ends of the hearts are relatively large

84 J. B. JOHNSTON.

and project so far into the ventral vessel that they might be mis-
taken for valves proper to the ventral vessel itself.

The course of the blood flow is determined by the disposition
of the valves as well as by the direction of the pulsations, and
there is evidently entire agreement between the results of the
physiological experiments and anatomical investigation. It is
obvious that in small vessels or in such as receive blood from a
capillary system so that there is no great pressure in the usual
course, there may occur temporary reversals of flow due to
movements of the body or other causes. Such reversals might
most readily take place in the subneural vessel and such phe-
nomena are probably the basis for Harrington's statement that
the blood flows now forward, now backward in the subneural.
However, the general course of the blood flow is strictly deter-
mined, as shown by the consistent experimental and anatomical
results, and no considerable or long-continued reversal or inter-
ruption of the usual current are possible except as the result of
violent interference such as decapitation of the worm.

The valves in the vessels have received very meager notices
heretofore. The mention of valves in the dorsal vessel by Ben-
ham has been noticed above. A recent writer 1 has mentioned
the presence within the dorsal vessel of cells similar to the chlora-
gogue cells. These are also doubtless the valves of the dorsal
vessel.

EXPLANATION OF FIGURES.

Abbreviations : b.w., body wall ; c.t , connective tissue layer of blood vessels ; d.,
dorsal vessel ; d-i., dorse-intestinal vessel ; d-t., dorso-typhlosolar vessel ; end. , endo-
thelial lining of vessels ; i.v.p., vascular plexus of intestine ; /.«., lateral neural ves-
sel; ;«., layer of circular muscle fibers in walls of vessels; ;//>//., nephridium ; /.,
parietal vessel ; s., septum ; s-n., subneural vessel ; t.s., typhlosolar sinus ; ?'. , ven-
tral vessel; va., valve; i>.i., ventro-intestinal vessel.

'Rice, Win. J., "Studies in Earthworm Chloragogue," BIOL. BULL., Vol.
III., Nos. 1-2, 1902.

TWO NEW GENERA OF MALLOPHAGA.

VERNON L. KELLOGG,
STANFORD UNIVERSITY, CAL.

There have come to me recently specimens of Mallophaga,
taken from birds from mid-ocean islands, which demand the
founding of two new genera in this interesting but little-studied
order of parasitic insects. In the order, as at present known,
there are about 1,500 species, comprising twenty -three genera.
The small number of genera is striking in itself, but is made more
amazing when it is remembered that eleven of the genera com-
prise but thirty of the species, leaving thus nearly the whole bulk
of the species included in the twelve remaining genera. The ad-
dition of two new genera is, therefore, rather notable in the de-
velopment of our systematic knowledge of the group. Although
about two hundred new species of Mallophaga have been de-
scribed from North American birds but one new genus (my Gie-
bclia, with only one species, from shearwaters) has had to be
established, all the other North American species being referable
to genera founded on Old World species and specimens. The
following revised key to the known genera of the order (includ-
ing the two new genera described in this paper) is presented for
the use of beginning students of the group, or of general entomol-
ogists :

ANALYTICAL KEY TO SUBORDERS OF MALLOPHAGA.

With filiform, 3- or 5-se.; nented, exposed antennae ; no labial palpi ; mandibles ver-
tical ; cesophageal sclerite and accompanying glands usually present and normal ;
meso- and metathoracic segments fused ; crop a saclike diverticulum ; ingluvial
glands present ; testes four ; egg tubes five IsCHNOCERA.

With clavate or capitate, 4-segmented, concealed antennae ; with 4-segmented labial
palpi ; mandibles horizontal; oesophageal sclerite and accompanying glands absent
or modified ; meso- and metathoracic segments with sutural line usually visible ;
crop simple ; ingluvial glands absent ; testes six ; egg tubes three to five.

AMBLYCERA.

ANALYTICAL KEY TO GENERA OF THE SUBORDER ISCHNOCERA.

A With 3-segmented antennas ; tarsi with one claw ; infesting mammals (family

Trichodectidae) Ti-ichodectes Nitzsch.

AA With 5-segmented antennae ; tarsi with two claws; infesting birds (family Phil-
opteridse).

85

86 VERNON L. KELLOGG.

B Antennas similar in both sexes.

C Meso- and metathoracic segments not fused Nesiotinus Kellogg.

CC Meso- and metathoracic segments fused.

D Front deeply angularly notched Akidoproctus Piaget.

DD Front convex, truncate, or rarely with a curving emargination,

but never angularly notched.
E Species broad and short, with large, movable trabeculae (at the

anterior angle of antennal fossa).

F Forehead with a broad transverse membranous flap pro-
jecting beyond lateral margins of the head in the male,

barely projecting in the female Giebelia Kellogg.

FF Without such membranous flap Docophorus Nitzsch.

EE Species elongate, narrow, with very small or no trabeculos.

Xirtnus Nitzsch.
BB Antennae differing in the two sexes.

C Species wide, with body elongate-oval to suborbicular.

D Temporal margins rounded ; last segment of abdomen roundly
emarginated ; antennae of male without appendage ; third seg-
ment very long Eurymetopus Taschenberg.

DD Temporal margins usually angulated ; last segment of abdomen

convex, rarely angularly emarginated, with two points.
E First segment of antennae of male large, sometimes with an ap-
pendage ; third segment always with an appendage.

Goniodes Nitzsch.

EE First segment of antenna of male enlarged, but always with-
out appendage ; third segment without appendage ; last
segment of abdomen always rounded behind.

Goniocotcs Nitzsch.
CC Species elongated narrow, sides subparallel.

D Third segment of antenna of male without an appendage.

Ornithobius Denny.
DD Third segment of antenna of male with an appendage.

E Front deeply angularly notched.. Bothriometopits Taschenburg.
EE Front not angularly notched.

F Forehead with a broad transverse membranous flap or fold
projecting beyond lateral margins of the head.

Philnteanus Kellogg.
FF Without such membranous flap.

G Antennae and legs long; a semicircular oral fossa.

Liptums Nitzsch.

GG Antennae and legs short ; oral fossa narrow, elongate,
extending as a furrow to the anterior margin of the
head Oncophorits Rudow.

ANALYTICAL KEY TO GENERA OF THE SUBORDER AMBLYCERA.

A Tarsi with one claw ; infesting mammals (family Gyropidae) ... Gyropus Nitzsch.
AA Tarsi with two claws ; infesting birds (excepting Boopial) (family Liotheidae).
B Ocular emarigination distinct, more or less deep.

C Forehead rounded, without lateral swelling ; antennae projecting beyond
border of the head Coipocephalum Nitzsch.

TWO NEW GENERA OF MALLOPHAGA. 87

CC Forehead without strong lateral swellings.

D Antennre projecting beyond border of the head ; temporal angles

projecting rectangularly ; eye large and simple Boopia Piaget.

DD Antenna; concealed in groove on under side of the head ; tem-
poral angles rounded or slightly angular ; eye divided by an
emargination and fleck.
E Mesothorax separated from metathorax by a suture.

Trim ton Nitzsch.
EE Meso- and metathorax fused ; no suture.

Lamobothrium Nitzsch.
BB Ocular emargination absent or very slight.

C Sides of the head straight or slightly concave, with two small projecting

labral lobes Physostonmm Nitzsch.

CC Sides of the head sinuous ; forehead without labral lobes.

D Ocular emargination filled by a strong swelling ; sternal markings
forming a quadrilateral without median blotches.. Nitzschia Denny.
DD Ocular emargination without swelling, hardly apparent or en-
tirely lacking ; median blotches on sternum.

E Very large ; with two 2-pointed appendages on ventral aspect
of hind head ; anterior coxae with very long lobe-like appen-
dages Antisiroiia West wood.

EE Small or medium ; without bipartite appendages of hind
head Alenopon N itzsch.

PHILOCEANUS gen. nov.

In a collection of Mallophaga taken by Mr. Rollo Beck from
birds of the Galapagos Islands (the collecting of birds and para-
sites was done by Mr. Beck in the summer of 1901), are five
specimens, including one male, three females, and one young, from
a single specimen of Pi'occllaria tethys (Wenman Id.) of a Mallo-
phagan species not assignable to any of the known genera. The
shape and habitus of whole body and the secondary structural
differences between the sexes, shown in antennae and abdominal
segments, are those of Lipeurus, while the well-developed and
unusual transversal membranous clypeal flap is that of Giebtlia.
The curious prolongation of the postero-lateral angles of the
mesothorax is a character peculiar to the new genus. As
Gicbelia with its short, broad body and antennae similar in both
sexes stands to Docophorus, so the new genus, which may be
called Philoceanus, with its elongate body, and differing antennae,
stands to Lipcurns.

The characteristics of the new genus may be given as follows :
body Lipeuroid, elongate ; head, thorax and abdomen of about

88

VERNON L. KELLOGG.

equal width (in widest places) ; antennae differing in the sexes,
that of male having an appendage on third segment ; abdomen of
male narrower than in female, parallel-sided, and with segments
6-8 each about twice as long as each of preceding segments ;
head with a broad, thin, transvenal, membranous clypeal flap
projecting far on each side of forehead in an angulated and folded
process ; metathorax with postero-lateral angles conspicuously

FIG. i. Philoceanus becki, male. FIG. 2.) Philoceanus becki, female.
(Length, 1.6 mm. (Length, 1.5 mm.)

produced into tapering, blunt-pointed, backward-projecting proc-
esses.

PHILOCEANUS BECKI sp. nov. (Figs, i and 2.)

Five specimens (one male, three females, one immature) taken
from Procellaria tethys (one specimen) Wenman Id. of the Gala-
pagos group, summer of 1901, by Mr. Rollo Beck.

Description of Male. — Body, length 1.6 mm., width .27 mm. (abdomen),
pale yellowish brown, with darker to blackish-brown marginal and trans-
verse bands which cover so much of the surface as to give the posterior
half of the body a general dark brown coloration.

Head, length .4 mm. width .3 mm., large in comparison with rest of
body, wider than any other part of body, and conspicuously large, /. e. ,
wider and longer than the thorax ; clypeal front broad, flatly convex and
with distinct thin uncolored rounding margin ; clypeal sutures distinct,
broad, and with two conspicuous hairs at the marginal termination ; these
clypeal sutures form a V-shaped figure enclosing the distinct clypeal signa-
ture between the anterior prongs ; the clypeus bears a conspicuous mem-

TWO NEW GENERA OF MALLOPHAGA . 89

branous flap or fold, thin and uncolored, which rises from about in trans-
verse line with the mandibles and projects forward to the point of the cly-
peal sutures, and laterally conspicuously beyond the margins of the head ;
in these lateral extensions the flap is folded back (towards the head margin)
on itself ; eye with rather long hair ; the temples are not much swollen
and each bears two long and a few short hairs ; the antennae (Lipeuroid)
have the first segment as long as all the others combined and the third
segment with an appendage ; the ground color of the head is pale translu-
cent yellowish-brown with the clypeal signature, a broad submarginal an-
gulated band on each side of head, extending from clypeal suture to base
of antennae, darker brown.

Thorax small ; prothorax with rounded postero-lateral angles with two
separated longish hairs in each ; metathorax a little wider and about twice
as long, with postero-lateral angles conspicuously produced as thick, taper-
ing, blunt pointed, finger-like processes, a long hair rising from base of
each process and another not so long and two or three short ones rising
from general postero-lateral angular region ; posterior margin of metathorax
slightly angulated in the middle and slightly concave in the space between
this median angle and the postero-lateral angle ; color pale translucent
yellowish-brown with darker rather broad lateral margins.

Abdomen elongate, rather narrow, subparallel-sided ; segments 1-5
each about one half as long as segments 6-8 ; long, flexible curling hairs
in postero-lateral angles of segments 2-7, and terminal segment with many
short fine hairs ; pale yellowish-brown ground color almost wholly obscured
by strong dark to blackish-brown lateral and transversal bands.

Female. — About same size as male but with abdomen wider (.4 mm.) in*
the middle and thus not parallel-sided ; ground color of whole body less
pale and translucent than in mafe ; head with transversal clypeal flap as in
male ; antenna; without appendage on third segment and with first seg-
ment shorter than second ; thorax with postero-lateral finger-like processes
of meta-segment and with three or four long hairs in postero-lateral region ;
abdomen with second segment longest, others about equal among them-
selves, and segments 4-6 (in middle of abdomen) wider than others, so
that the whole abdomen is elongate elliptical in outline ; last segment with
slight angular median emargination on posterior margin.

NESIOTINUS gen. nov.

A single female Mallophagan specimen of well-defined char-
acter received from Dr. G. Enderlein, of Berlin, proves to be a
form which it is impossible to ascribe to any known genus ol
the order. This specimen was taken from Aptenodytes longi-
rostris, a new penguin species from Kerguelen Id., collected by
the German Deep-sea Expedition in 1899.

VEKNON L. KELLOGG.

This new Mallophagan form unites in striking manner the im-
portant antennal characters of the family Philopteridae with the
general habitus and body characters of the family Liotheidae.
The shape of head, and the distinctly free metathoracic segment
are characteristics heretofore peculiar to the genera Menopon and
Trinoton (of the Liotheidae), but the short, slender, five-segmented
antennae not lying in special antennal cavities identify the species
as a Philopterid, but one not assignable to any known Philopterid
genus. The new form represents a Menopon- and Trinoton-\\\<Q

genus in that family to which
Menopon and Trinoton do not
belong ! The only other Mal-
lophagan species taken from the
penguin genus Aptenodytes is
Goniodcs brci'ipes, a small spe-
cies very unlike this new form,
described by Giebel (from a fe-
male specimen) in the Phil.
Trans. Roy. Soc., Vol. 168, extra
Vol. This specimen also came
from Kerguelen Id.

The distinguishing characters
of this genus are its Menopon-
like form, the small suborbicu-
lar head with slightly-produced
subrectangular temples, the dis-
tinctness of the meso- and meta-thoracic segments in a degree
unequalled elsewhere among the known Mallophaga unless it be
in Trinoton, the very small characteristically Philopterid antennae,
the sharp division of each eye into practically a pair of eyes, the
large size of the hind body in comparison with the head, the
heavy transverse blotches of the abdomen and the five pairs ot
abdominal spiracles instead of the usual six pairs.

NESIOTINUS DEMEKSA sp. nov. (Fig. 3.)

Ffinalt'. — Body, length 5 mm., width 2.1 mm.; head, length .75 mm.,
width 1.15 mm.; thorax, length 1.25 mm., width of prothorax .8 mm.,
width of mesothorax 1.30 mm., width of widest segment, the first, 2.16
mm.; chestnut brown, with large blackish-brown blotches on thorax and
abdomen.

FIG.

3. Xesiotes deinersa, female.
(Length, 5 mm. )

TWO NEW GENERA OF MALLOPHAGA. 9 I

Head small in comparison with rest of body, hardly as wide as meso-
thorax, with flatly rounded front, no orbital sinus, temples slightly swollen,
rounded, but with postero-lateral angle slightly obtusely produced, occipital
margin slightly curving ; eyes divided so as to give the effect of one pair on
each side ; antennae short, slender, tapering ; pustulated hairs on temporal
margins and two small hairs with large pustulation on dorsal surface of each
temple, also four smaller pustulations on postero-median dorsal surface, and
one mesad from each eye pair ; color chestnut-brown with blackish eye
flecks arid dark brown markings along temporal margin and in postero-
mesial angles of each temporal region.

Thorax of three distinct segments regularly widening posteriorly, the
meta-segment being nearly as wide as first (widest) abdominal segment and
resembling an abdominal segment ; prothorax with slight median angulated
point on anterior margin, with parallel straight lateral margins and rounded
antero-lateral and postero-lateral angles, anterior half dark brown, posterior
half light brown ; mesothorax with diverging lateral margins, small pustu-
lated hairs in angles and flatly rounding posterior margin : anterior four
fifths of segment dark brown with series of weak hairs in demi-pustulations
along the hind margin of this dark region ; metathorax with diverging
lateral margins, and with large lateral transverse dark brown blotches leav-
ing a rather narrow light brown median space. Legs with heavy short
femora and long slender tibiae with few short, weakly pustulated spiny hairs
on each segment ; two terminal tibial spines ; femur darker than the trans-
lucent pale brown tibiae.

Abdomen forming with meso- and metathorax an ellipse ; segments i
and 2 widest and others tapering slowly posteriorly ; hairs few and incon-
spicuous ; segments 1-5 with conspicuous spiracles each showing as a
small brown spot in a large clear circular pustulation ; segments 1-4 with
large lateral transverse dark brown blotches leaving a lighter median space
which is narrower on each successive segment posteriorly : segments 5-7
with dark-brown transverse bands extending clear across body ; all trans-
verse blotches and bands blacker and slightly wider at lateral ends, with
slight anteriorly projecting process ; indications of demi-pustulations in
lateral portions of posterior margin of each blotch and band ; posterior
margin of terminal segment flatly rounded, and longest hairs of the body
in lateral angles.

NOTE. — In a paper published while this paper was in press, on the Mallophaga
from Birds of Costa Rica (Univ. Studies, Vol. 3, pp. 123-197, 1903, Univ. Nebraska)
M. A. Carriker, Jr., describes two new genera of Mallophaga, under the names Or-
nicholax and Kelloggia.

EXPERIMENTAL STUDIES ON THE DEVELOPMENT

OF THE ORGANS IN THE EMBRYO OF

THE FOWL (CALLUS DOMESTICUS).

FRANK R. LTLL1E.

I. INTRODUCTION.

The results to be described under the above title relate to the
morphology, functions and power of regeneration of various
embryonic organs, and to the influence that certain embryonic
parts exert on the development of others. They represent the
application of a particular experimental method, viz., the destruc-
tion of definite parts, and study of the subsequent development.
Thus the particular organs studied are those most accessible to
operation, which form a rather heterogeneous assemblage. Nev-
ertheless, taken as a whole, the results form a contribution to
the subject of correlative differentiation of organs.

The Principle of Correlative Differentiation in Embryology (i. e.,
influence of the intraorganic environment in development)1 is
that the rate, degree or mode of differentiation of any embryonic
rudiment is dependent on some part or parts of the same organism
(individual) external to itself; that is, that component parts of an
embryo determine mutually to a greater or lesser extent, their
respective lines and grades of differentiation. Much more is
meant by this than that any embryonic part can develop only in
its normal environment, which offers the prerequisites of its very
existence. The principle of correlative differentiation in fact im-
plies a distinction between a determinative and a non-determinative
environment, and the problem of correlative differentiation is so
far resolved when this is ascertained for all the organs (cf. Roux).

Any part, the entire environment of which is non-determina-
tive, is said to develop by self -differentiation (Roux).

These two principles do not stand in the relation of rival
theories but rather, probably, of cooperative factors in every

1 Environment may be defined as conditions that influence dynamic processes in
protoplasm, and may be divided into extraorganic and intraorganic, the former being
external to the individual and the latter within its bounding surfaces.

92

ORGANS IN THE EMBRYO OF THE FOWL. 93

embryonal differentiation, for any process of self-differentiation
of a structure might be analyzable into correlative differentiation
of its parts.

For the development of the higher animals at least the extra-
organic environment is non-determinative. The development of
the ovum as a whole is therefore a process of self-differentiation.
But it is usually assumed that it is otherwise with the differen-
tiation of its constituent parts ; the extreme view being that each
influences the mode of differentiation of all the remainder. From
this standpoint the complexity of the correlative processes of
differentiation must increase in proportion to the increase in com-
plexity of structure.

Theoretically, at least, the determinative value of correlative
differentiation in any case may be (i) absolute, /. c., the mode of
development of a part being determined entirely from without
itself; (2) partial ; (3) wanting, /. c., absolute self-differentiation.

Our present knowledge is enough to exclude the first theoreti-
cal possibility. No principle in embryology is better established
than that sooner or later the embryo is a mosaic of embryonic
rudiments, each of which is to a certain extent self-determining.
This mosaic of rudiments may become visible very early, as in
those ova exhibiting a definite cell-lineage of organs, or it may
appear later. In some cases, at least, the unsegmented ovum
itself is a simple mosaic (ovum of ctenophores according to
Fischel ; ovum of Unio, Lillie ; ovum of sea-urchins, Boveri ;
ovum of frog, Roux, Schulze and others). Indeed it is quite
probable that all ova are more or less simple mosaics of embry-
onic rudiments.

Unless, therefore, we wish to beg the entire question we must
proceed on the second hypothesis. This is the writer's stand-
point, and the problem is to determine as many definite correla-
tions as possible and to investigate their nature.

There is probably no conception in embryology so vague as that
of correlative differentiation, as the following citations may serve
to show :

Hertwig : " Zelle und Gewebe," II. :

" Die Wechselwirkungen (Correlationen) zwischen den Zellen
eines Organismus und ihren Derivaten bilden sich mit dem

94 FRANK R. LILLIE.

Beginn des Entwicklungsprocesses aus, andern sich von Stufe zu
Stufe und compliciren sich in demselben Maasse, als die Entwick-
lung fortschreitet.

" Im Geeensatz zum Mosaiktheorie von Roux und der keim-

^5

plasma theorie von Weismann stellt die Theorie der Biogenesis
den Grundsatz auf, dass vom ersten Beginn der Entwicklung an
die durch Theilung des Eies sich bildenden Zellen bestandig in
engster Beziehung zu einander stehen, und dass dadurch die
Gestaltung des Entwicklungsprocess sehr wesentich mit be-
stimmt wird. Die Zellen dcterminiren sich ztt Hirer spatcrcn Eige-
nart niclit sclbst, sondern werden nacli Gcsctzcn die sich aus dan
Zusammenwirkung alter Zellen auf den jeweiligen Entwickh/ngs-
stitfen des Gesammtorganisiinis crgcben, dctcnninirt."'

Herbst : " Formative Reize in der Tierischen Ontogenese : "

" Die Aufgabedes zweiten Teiles meiner Abhandlung iiber die
formativen Reize war es also, in der tierischen Ontogenese, ab-
gesehen von der Namhaftmachung jener wenigen Falle von Ge-
kommen von formativen Reizwirkungen, die von irgend einem
Teil des Organismus auf einen oder mehrere andere ausgeiibt
werden, festzustellen und eventuell die Moglichkeit der vollstan-
digen Auflosung der ganzen Ontogenese in einer Reihe von sol-
chen Induktionserscheinungen nachzuweisen.

" So ist es zum Beispiel zum mindesten ungenau, von der
' weitgehenden Wechselbeziehung ' zo sprechen, ' die zwischen
alien Teilen eines Organismus auf alien Stadien seiner Entwick-
lung besteht' (Hertwig : ' Evolution und Epigenesis ') ; denn
das Ektoderm der Echiniden entwickelt sich unabhangig vom
Entoderm, und auch abgeschniirte Hautstiicke, etc., konnen sich
selbstandig differenzieren, wie dies das Vorkommen der Teratome
beweist (Roux). Die Annahme einer ganz allgemeinen Korrela-
tion zwischen alien Teilen des Organismus auf alien Stadien der
Ontogenese ist deshalb ebenso falsch wie jene von der qualitativ
ungleichen Kernteilung der Mosaiktheorie."

Most of the real illustrations (/. e., experimentally determined)
of this principle must be taken from plants and plant-like animal
colonies. One need only glance through Herbst's recent " For-
mative Reize in der Tierischen Ontogenese " to realize that, so
far as egg development is concerned, the application of the prin-

ORGANS IN THE EMBRYO OF THE FOWL. 95

ciple rests very largely on inference, analogy and a few doubtful
pathological conditions.

Discussion of this subject belongs, however, to the conclusion
rather than to the introduction, and the foregoing remarks are
intended only to define the problem.

II. METHODS OF OPERATION.

In making the operations one must work as far as possible
under antiseptic conditions. Instruments, etc., must be steri-
lized ; this is most readily done by passing the needles, knives,
scissors, etc., through a flame immediately before each is used.
In spite of all precaution a great many eggs are infected. In
my experiments only about 20 per cent, of the eggs remained
alive until the time of examination for the results of the experi-
ment, two to five days after the operation. The causes of the
mortality in the remaining 80 per cent, are two: (i) Fatal in-
jury of the operation (about 40 per cent. ?) ; (2) infection with
mould or bacteria (about 40 percent. ?). There is a very notice-
able difference between different lots of eggs ; some bear opera-
tions much more readily than others and are less prone to infec-
tion. These differences in the relative powers of resistance of
different lots of eggs are due to the relative freshness of the eggs
when incubation is begun, and also to the time of year. It is
noticeable that in a lot of eggs in which a relatively large pro-
portion, over 50 per cent., fail to develop in the incubator, the per-
centage of failures in the actual experiments is usually very high.

The method of procedure in my experiments was as follows :

1. The eggs are not turned in the incubator, so that one may
be sure of locating the position of the embryo in the unopened
egg exactly. The upper side of each egg is marked with a pencil.

2. A small opening is made through the shell and membrane
over the embryo.

3. The operation is then made. For cauterization I employ
either a needle heated red hot in the flame, or an electric cauter-
izing needle. The heated needle cools very rapidly, so that the
operation must be hastily performed, and it is difficult precisely to
delimit the injury. The electric cautery, on the other hand, is
apt to give too intense heat. Each method possesses certain
advantages.

96 FRANK R. LILLIE.

4. The opening in the egg is closed as follows : A piece of
the shell with membrane attached is cut from a corresponding
part of another fresh egg, so as to be slightly larger than the
opening in the operated egg. This is placed over the opening so
as to close it completely ; and the albumen adhering to the mem-
brane acts as cement. To ensure perfect closure strips of the
egg-membrane are plastered on so as to overlap all edges of the
foreign shell. The advantages of this method of closure are
that the foreign surfaces are perfectly aseptic if fresh eggs are
used, and that the conditions are as nearly like the normal as
possible. It is, morover, the simplest and easiest method. This
method of closing the opening was first used by Miss Peebles.1

III. EXPERIMENTS ON THE AMNION AND THE PRODUCTION OF
ANAMNIOTE EMBRYOS IN THE CHICK.

A. The Normal Development of the Amnion.

The purpose of this section is to give a brief statement of some
facts concerning the formation of the amnion before taking up
the analysis of the processes by experiment. This is necessary
because the facts are at least partly new, and without knowledge
of them the mechanics of formation of the amnion cannot be
understood. For a recent review of the literature on the whole
subject of the amnion in the Sauropsida, see Schauinsland ('O2a
and '02$) ; the latter paper I regret not to have seen.

In the somatopleure on each side of the axis of an early embryo
of the chick three zones may be distinguished on the basis of the
subsequent differentiation, (A) for the body-wall ; (^) for the
amnion ; (C] for the chorion (serosa) (Fig. i). It is important to
trace the origin of the differentiation between the amnion and
serosa on the one hand, and amnion and body-wall on the other,
for the conditions that determine the development of the amnion
must be antecedent to such differentiation.

i. The Ectamnion. — The differentiation of the amniogenous
from the choriogenous somatopleure is always preceded by the
appearance of a thickening of the ectoderm along the external
margin of the former. This thickening, for which I propose the

1 Rome's Arc/iiv, VII., 1898.

ORGANS IN THE EMBRYO OF THE FOWL.

97

name ectamnion, precedes by a little the formation of amnioge-
nous folds in any region, and indeed it induces the origin of the
entire system of folds. It has been described by many embryol-
ogists at the stages immediately preceding fusion of the limbs of

e.a.

FIG. I. Embryo of chick with 13 mesoblastic somites. University of Chicago
Embryological Collection, No. 555. e.a., ectamnion; a.c., inner margin of amnio-
cardiac vesicles ; A, region of the somatopleure destined to form the body-wall ; B,
amniogenous somatopleure ; (7, choriogenous somatopleure.

the amnion (cf. Schenk, '/i), and it forms the ectodermal sero-
amniotic connection of Hirota ('94). But no one, so far as I
know, has traced it back to its origin and recognized the fact
that it is the earliest formed part of the amnion, which is thus
primarily ectodermal in the chick, as in Chelonia and some other
primitive Sauropsida.

The ectamnion may first be distinguished at about the stage
with nine mesoblastic somities, where it appears as a median thick-
ening of the ectoderm in front of the head near the anterior
boundary of the proamnion. Along the line of this thickening
there is a fusion, between ectoderm and entoderm. The thick-
ening is extended right and left and turns backwards along
opposite sides of the head to about the region of the middle of

98 FRANK R. LILLIE.

the heart, gradually becoming more peripheral in position and
slowly fading out (Fig. i). This line represents the junction of
the amniogenous and choriogenous somatopleure, and thus cor-
responds to the angles of the future amniotic folds.

The head of the embryo lies in a depression bounded in front
by the ectamnion and on the sides by the ammo-cardiac vesicles
of the body cavity, along the inner upper margin of which the
ectamnion runs for a short distance. The floor of the depression
is the proamnion.

In a stage with 14-15 mesoblastic somites the ectoderm of the
proamnion is much more thickened in front of the head, and has

c .

FIG. 2. Transverse section through the anterior angle of the ectamnion, a few
sections in front of the tip of the'head. 14-15 mesoblastic somites. University of
Chicago Embryological Collection, No. 215. b.c. , body-cavity; f., large cavity in
the entoderm ; e.ti., ectamnion.

outer surface in consequence of irregularity in the thick-
ening 1 (Fig. 2), which may be traced back to the level of the
heart, and on one side to its hinder end ; there is also a very
short ectentodermal fusion beneath the tip of the head. In this
series the ectamnion marks the boundary between two distinctly
differentiated parts of the extraembryonic somatopleure, the more
central of which is the amnion.

In another embryo with fourteen mesoblastic somites, the tip of
the head is surrounded by the amnion, and the proamniotic partis
represented only by a short median strip extending eight sections
back to a point where the limbs of the amnion have not yet
closed. The ectamnion is continued only for a short distance
along the angles of the amniotic fold, and then passes peripher-

1 In examining the section one receives a strong impression that the irregularities
may be due to amceboid movements ; but it is not possible to confirm this by actual
observations.

ORGANS IN THE EMBRYO OF THE FOWL. 99

ally. How has the head-fold been formed ? The great expansion
of the body cavity (amniocardiac vesicles) on each side causes
an elevation of the anterior angle of the ectamnion and a pocket
is formed by fusion of its opposite limbs, which have a strong af-
finity for each other ; fusion proceeds along the median dorsal
line so long as the energy of fusion is sufficient to draw the so-
matopleure up. The head of the embryo is rapidly elongating at
this time and slips into the pocket thus formed, being guided in
part by the cranial flexure (Fig. i). It is interesting to note how
far the ectodermal thickening stretches ahead of the mesoderm of
the fold near the point of closure, and that the apical cells are
elongated into pseudopodium-like processes.

The histological differentiation of the amniotic area of the
somatopleure from the chorionic portion precedes the elevation
of the fold.

This brief inquiry, then, suggests that the order of events in
the formation of the head fold of the amnion is :

1. Thickening of the ectoderm on the outer margin of the am-
niogenous somatopleure, beginning in -front of the head of the
embryo and extending back on each side (ectamnion).

2. Great expansion of the body cavity on each side opposite
the head of the embryo and consequent elevation of the anterior
bay of the ectamnion to the level of the dorsal surface of the
embryo.

3. Fusion of the right and left limbs of the ectamnion, begin-
ning at the angle, to form a pocket, the head-fold of the amnion.

4. Pushing of the head of the embryo into the fold.

There may be, however, considerable variation in the time of
formation of the head-fold. I have, for instance, one series with
17-18 mesoblastic somites (ser. 175), where the head-fold is not
yet formed.

Extension of tJic Ectamnion. — The ectamnion differentiates
backward more rapidly than the lateral folds, and always pre-
cedes their origin. In the 48-hour stage (21—22 somites) (Fig.
3) the ectamnion from in front has joined that from behind
formed in connection with the tail-fold. There is a place, corre-
sponding nearly to the final meeting place of anterior and poste-
rior lateral folds, where it becomes very faint. It would appear

TOO

FRANK R. LILLIE.

then that behind the tail there is actually a new starting-point for
the ectamnion as well as the amniotic folds. The primary posi-
tion of the ectamnion is near the boundary of the pellucid area ;
towards the posterior end it bends in very sharply, nearly joining
the body wall proper, and terminating in the posterior rudiment.

FIG. 3. Embryo of chick with 21 mesoblastic somites. University of Chicago
Embryological Collection, No. 99. e.a., ectamnion; s.jf., secondary folds of the
amnion on the right side. The dotted line continuing e.a. represents the continua-
tion of the ectamnion beyond the region of folding The dotted area at the angle of
the folds represents the ectodermal sero-amniotic connection of Hirota.

Origin of the Tail-Fold. - - The tail-fold proper arises from an
ectodermal thickening lying in a depression just beneath the
rudimentary tail-bud. The depression is caused by the enlarge-
ment of the body cavity on each side of the middle line. These
enlargements may be called the amriio-allantoic enlargements, as
they are associated with the formation of the allantois. I would
venture the hypothesis that the existence of a separate tail-fold of
the amnion is associated with the time of development of the al-
lantois, which is represented in the embryo under consideration
(i) by a shallow entodermal evagination and (2) a mass of meso-
blast.

ORGANS IN THE EMBRYO OF THE FOWL. IOI

At the time of formation of the tail-bud a very shallow pocket
forms behind it. This owes its origin to the elevation of lateral
folds of the somatopleure and progressive fusion beginning at the
posterior angle of the ectamnion. The floor of the pocket in-
cludes a thick posterior prolongation of the allantoic mesoblast
which furnishes a firm floor to the pocket and thus determines
the form of the folds.

2. The Amniotic Folds.- -The subsequent development in-
cludes the elevation and fusion of the anterior and posterior lat-
eral folds. The final closure takes place opposite the buds of the
hind limbs. The order of events in these processes is as follows :

1. The growth of the amniogenous somatopleure behind the
head-fold and in front of the tail -fold.

2. The uprising of the amniotic folds, and their growth in a
definite direction around the embryo.

3. The fusion of the right and left folds along the line of the
ectamnion in such a way that the external limbs unite to form
the chorion, and the internal to form the amnion.

Study of the morphology of these processes suggests the fol-
lowing physiological conclusions :

1. The growth of the amniogenous somatopleure may be a
result of the traction exerted in it by the progressive fusion of the
folds already formed in front and behind.

2. The uprising of the lateral folds is determined by the head-
and tail-folds, the progressive fusion of the right and left ectam-
nion dragging the amniogenous somatopleure into place.

It remains to test these conclusions by experiments, but before
proceeding to a description of these, I wish to describe the influ-
ence of the rotation of the embryo on the amniogenous somatopleure.

Practically all of the somatopleure of the pellucid area is amni-
ogenous with the exception, naturally, of that part internal to
the limiting sulci that forms the body-wall. What effect has
the turning of the embryo on its left side on the amnrogenous
somatopleure ? We will suppose that the latter is primitively of
equal width on both sides ; we will furthermore assume that the
somatopleure cannot be drawn in from the vascular area, because
it is here attached to the splanchnopleure. (The fusion of the
somatopleure and splanchnopleure at the margin of the pellucid

IO2

FRANK R. LILLIE.

area is shown by the fact that the splanchnopleure is often drawn
up with the outer limb of the amniotic fold, making a fold of the
splanchnopleure at this place) (Fig. 5). Finally let us assume
that the notochord represents approximately the axis of rotation.
During the process of rotation the embryo sinks and the lateral
limiting sulci become deeper. A direct consequence of the rota-
tion must be therefore a strong tension on the somatopleure be-
longing to the under (left) side, a-b, and practically none on the
upper (right) side, c-d, (see Fig. 4, A, B, C}.

a

Hi*

Or

a

d

C

FIG. 4. A, B and C. Diagrams to represent the effect of rotation of the embryo
on the amniogenous somatopleure. a represents in all figures the position of the
ectamnion on the left (lower) side ; d represents in all figures the position of the
ectamnion on the right (upper) side, b and c represent the junction of amnion and
body-wall on left and right sides respectively. In Fig. A, a-b and c-d are equal.
In Fig. B, rotation of the embryo is assumed to have taken place without formation
of the amnion ; the distance a-b has become greater than c-d. In Fig. C is repre-
sented rotation of the embryo with synchronous formation of the amniotic folds, as is
actually the case ; c-d is inevitably thrown into secondary folds. The vertical lines
at the extreme right and left represent the margins of the pellucid area.

Even though the difference may be partly compensated for by
drawing of the embryo to the left, the tendency would be to
stretch a-b. If there were no such compensation and a and b
were practically fixed points, the length of a-b at the conclusion

ORGANS IN THE EMBRYO OF THE FOWL.

103

of the rotation would much exceed that of c-d (Fig. 4, //) ; and
if during this process there were actual independent growth of
a-b and c-d, the latter would of necessity be thrown into folds,
but not the former. Finally, if the amniotic folds were forming
at the same time (as is actually the case) the right one would
inevitably be thrown into secondary folds by the approximation
of points c and d (Fig. 4, C).

Study of the fusion of the amniotic folds in actual section
shows (i) that the line of fusion of the opposite amniotic limbs
is over the dorsal surface of the embryo only so long as the latter
lies flat on tJie yolk, and does not follow the turning of the embryo
on to (usually) its left side ; the consequence is that after rotation
of the embryo the line of fusion lies over the upper (right) side
of the embryo, often opposite the horizontal level of the intestine

/

FIG. 5. Transverse section of an embryo of about 48 hours (Duval) showing the
position of the ectamnion on the right and left sides. University of Chicago Embry-
ological Collection, No. 689. e.a., ectamnion ; /., left ; s.f , secondary fold of am-
nion on the right side. The great differences in the thickness of the amnion of the
right and left sides should be noted.

(Fig. 6). Thus one fold of the amnion passes all the way from
the under side over the back of the embryo and around on the
other side to the line of fusion, and thus is several times as long
as the opposite limb. (2) Moreover, the amniotic fold of the
right side is invariably thicker than that of the left side, and is
always thrown into secondary folds at the place of turning (Fig.
5 and Fig. 6). These conditions are satisfactorily explained, as
noted above, by the mere turning of the embryo on its side.

One must therefore distinguish in the upper limb of the am-
nion two kinds of folds: (i) The ordinary amniotic fold induced
by the fusion of the right and left rudiments and (2) secondary

104

FRANK R. LILLIE.

folds formed simply by the process of twisting of the embryo.
This distinction is of importance in interpreting the results of
the experiments.

Hirota (94) notices the secondary fold on the upper side and
says : " It seems to owe its origin to the presence of the sero-
amniotic connection. ... It is always on the right side of the
connection, and is pushed on towards the left. There takes place
no folding before the allantois appears, and the longitudinal ex-

e.a.

•J.

FIG. 6. Section of the same embryo as the preceding, 10 sections (150^) in front
of Fig. 5. The section passes through the place of fusion of the right and left folds.
The secondary fold of the amnion is well shown on the right side. Letters as in
Fig. 5-

tent of the fold depends on the extent of the sero-amniotic con-
nection." " Its form and extent are variable." " It is not clear
what significance this fold has." "At both extremities of the
sero-amniotic connection the amnion is also slightly folded longi-
tudinally."

These secondary folds of the amnion are very transitory ex-
cept in two regions : (i) Above the hind end of the heart (apex
of ventricle) and continuing a short distance behind it ; (2) in the
region immediately in front of the allantois, at 60-70 hours, thus
in the neighborhood of the final closure of the amniotic folds.
The former are of very constant occurrence and persist a long
time (Fig. 3). The latter are relatively slight and inconstant.
Hirota is thus mistaken in saying that these folds do not appear
until the formation of the allantois.

The secondary folds in the neighborhood of the heart are
always on the upper (right) side ; they first appear at the time

ORGAN'S IN THE EMBRYO OF THE FOWL. IO$

of rotation of the embryo, and are coincident with the closure
of the amnion (Fig. 3) ; they persist until the body-wall is com-
pleted behind the entire heart. They are not, in my opinion,
exclusively folds of the amnion, but extensions of the body-wall
for enclosure of the region of the heart and liver. The direct
cause of their formation is, however, the rotation of the embryo
with extreme growth of the body-wall contiguous to the amnion,
and fixation of the outer end of this limb of the amnion by the
amniotic suture.

Elsewhere the effect of the twisting of the embryo is rapidly
compensated so that the secondary folds of the right half of the
amnion do not persist long except in the region of the allantois,
where slight inconstant secondary folds may continue longer.

B. Experimental.

I. Experiments on the Head-fold of tlic Ainnion.

Experiment No. 57.

Age of the embryo at the time of operation, 33 hours1
(Duval).

Operation. — The blastoderm was cauterized lateral to the right
optic vesicle with a needle (Fig. 7) so as to make a large open-
ing. At the time of the operation only the most anterior horse-
shoe-shaped segment of the ectamnion was present (cf. Fig. i),
and this was destroyed only on the right side of the embryo. On
the left side, therefore, the amniotic fold was free to form to the
extent that it is independent of the opposite fold. The right optic
vesicle was slightly injured, as the results of the experiments
show. In opening the egg for the operation, the blastoderm was

1 In describing the various experiments, the age of the embryo at the time of the
operation will not be given as the actual number of hours in the incubator, because
the variations in point of actual development after the same period of incubation are
so extreme. It is not possible either to make accurate measurements of the living
embryo or to determine the number of somites present, on account of the loss of time
and danger of exposure of the embryo. A rough sketch of the embryo was always
made at the time of the operation, and this is sufficient to identify it with the various
staaes fio-ured in Duval' s atlas. The age is based on this identification. Thus the

r- •

given age at the time of operation in these experiments represents a certain dehmte
stage of development. On the other hand, the length of time that elapsed from the
experiment to the time of reopening the egg is always given literally.

IO6 FRANK K. L1LLIE.

also inadvertently torn just back of the embryo, and this opening
also appears in Figs. 8 and 9. This, however, was without any
noticeable effect on the subsequent development.

FIG. 7. Experiment 57. Operation diagram. Outline of embryo of chick 01
about 33 hours, after Duval. The ruled area to the right of the head indicates the
area of the blastoderm destroyed by the heated needle.

Examination of the Resulting' Embryo. — The egg was reopened
48 hours after the operation. The heart was beating vigorously ;
the hole made in the blastoderm by the operation had not closed,
and a good deal of yolk had escaped through this and overlay the
blastoderm. The embryo was well developed, corresponding to
the stage of 70-80 hours (Duval), and apparently normal in all
essential respects. (A defect in the right eye was evidently av
direct result of the operation.) The head of the embryo had
slipped through the hole in the blastoderm and was suspended
in the yolk (Figs. 8 and 9).

The embryo was cut into 250 sections of 1 5 // thickness.
Around the edges of the opening made by the operation the
somatopleure turns over and becomes continuous with the
splanchnopleure, ectoderm with entoderm, and mesoderm with

ORGANS IN THE EMBRYO OF THE FOWL.

IO/

mesoderm. In places one cannot determine where the ectoderm
leaves off and the entoderm begins.

Anmiotic Rudiments of tlic Left Side. — A short distance in
front of the margin of the opening there is a sharply defined fold
of the somatopleure capped by an ectodermal thickening that

e.a.

FIG. 8. Experiment 57. Upper surface of blastoderm, op. , aperture in the blas-
toderm made by the operation ; e.t!., amniotic rudiment of the left side ; /./. , tail-fold
of the amnion. The stippled area behind the embryo represents an aperture in the
blastoderm accidentally made in opening the egg for the operation.

represents the head-fold and left lateral fold of the amnion. The
extent of this fold is indicated by the line e.a. on Fig. 8. It
begins as a sharply marked fold at the most anterior angle of
the opening, and passes back, at first along the edge of the open-
ing, later a short distance from it, to the left of the embryo. It
very distinct (Fig. 10, /.<?./".) to the point where it is indicated as
is a broken line ; in this region the fold has disappeared, but the
thickening of the ectoderm (ectamnion), may be traced back to
the tail-fold with which it becomes continuous as indicated in the
drawing (Fig. 8). At no place, until the tail-fold is reached, is
the somatopleure internal to this line thrown into folds. By

io8

FRANK R. LILLIE.

reference to the figure and to the description of the operation it
will be seen that the line of this fold represents the continuation
of the left amniotic rudiment, which was not injured by the
operation.

I conclude, therefore, that when the amniotic rudiment of one
side is left free to develop after destruction of the rudiment of
the other side just prior to the formation of the head fold,

FIG. 9. Experiment 57- Under surface of the blastoderm. There is no amnion.
The right eye is defective, x marks the location of the secondary amniotic fold
shown in Fig. 10. A-B, plane of section shown in Fig. 10. Letters as in Fig. 8.

the ectamnion is propagated in the normal fashion and induces
the formation of a low fold, but that the amniogenous somato-
pleure is unable to raise itself around the body of the embryo.
The growth of the amniogenous somatopleure appears to be less
than normal.

Amniotic Rudiments of the Right Side. --On the right side, on
the other hand, a well-developed fold appears at the place where
the extra-embryonic somatopleure becomes continuous with the

ORGANS IN THE EMBRYO OF THE FOWL.

body-wall (section 138, Fig. 10) and extends to section 167, where
it suddenly ceases, a distance of about 0.5 mm. The location is
indicated by x on Fig. 9, and Fig. 10 shows it in section.

The formation of this fold is not induced by the ectamnion be-
cause the line of the latter (Fig. 10, r.e.a.} may be recognized
some distance lateral to the fold, through it is very slightly
developed. The fold in question is immediately back of the
heart on the right side of the body. It is not, in my opinion, a
true amniotic fold, but belongs to the category of normal sec-
ondary folds of the amniogenous somatopleure produced by the
turning of the embryo, with which it agrees precisely in position
and appearance. This conclusion is reinforced by the following
consideration : in this embryo the roots of the vitelline veins are
prolonged forward to an abnormal extent, and the right vein is
fused to the somatopleure lateral to the fold (Fig. 10). As the

l.a.J.

r.e.a.

FIG. IO. Section through the embryo of experiment 57 along the line A-B of Fig.
9. l.a.f., left amniotic rudiment ; r.e.a., ectamnion of the right side ; s.f. , secondary
fold of amnion on the right side ; v.?>., vitelline veins.

embryo turns, therefore, the somatopleure between the vitelline
vein and the body-wall must be folded to the extent that the
turning approximates the body-wall to the vein, because the
fusion prevents the somatopleure from being pushed peripherally.
As already said, therefore, this is not a true amniotic fold.

The prevention of the formation of the head-fold, by destruc-
tion of the rudiment of one side, operates to prevent the normal
elevation of the amniotic fold on the opposite side ; and thus it is
experimentally demonstrated that the cooperation of right and
left folds is necessary for the normal mode and direction of

I IO FRANK K. 1. 1 LI. IE.

growth of the amniotic rudiments. The height of the fold on
the uninjured side is a measure of the power of independent
elevation of a single amniotic fold.

On the other hand the existence of the ectamnion on the right
side, though in a rudimentary state, and the differences in finer
structure of the somatopleure on the two sides of this line in-
dicate that the distinction between amniogenous and choriog-
enous somatopleure is attained by the normal development of
the somatopleure as a whole, and not simply as a result of their
separation after fusion. However, the relatively rudimentary
condition of the ectamnion on the injured side shows that the
earlier stimulate the growth of the latter formed parts ; otherwise
we should expect to find the ectamnion equally developed on
both sides. The ectamnion of the right side does not exactly
join the tail-fold.

Tail-fold. - - The tail-fold of the amnion may be well seen in
Fig. 8. So far from compensating in any way for the absence of
head and lateral folds, it is of even less than its normal extent, a
fact indicating (possibly) that normally its growth is stimulated
by the traction of the anterior section of the amnion.

Experiment No. 36.

Age of the embryo at the time of operation forty-six hours
(Duval).

Operation. - - The operation consisted in the insertion of a heated
needle just in front of the heart (see Fig. 1 1). Examination of the
sections of the resulting embryo shows that the injury involved the
left optic cup slightly, and that the head-fold of the amnion which
extends back beyond the heart at this stage, stuck to the needle
and was stripped off, carrying with it a certain amount of the ad-
jacent somatopleure. This was not observed at the time of the
operation, but the conclusion is rendered positive by the subse-
quent examination of the embryo.

Examination of the Resulting Embryo.

The egg was reopened and the embryo preserved forty-eight
hours after the operation. The embryo (Fig. 12) appeared like
a normal embryo of about the ninety-sixth hour. The limb-
buds were well started, and the allantois extended out beyond

ORGANS IN THE EMBRYO OF THE FOWL.

I I I

the embryo, but towards the dorsal surface ; the flexures were
normal. The striking thing was the apparent entire absence of
the amnion ; the embryo lay naked on the surface of the blasto-
derm, to which it was attached, in the same manner as a selachian
embryo by a very broad somatic and splanchnic umbilicus.

In the normal embryo of this age the amnion is completely
closed, and the body-wall of the embryo has, therefore, lost all
connection with the chorion.

FIG. II. Experiment 36. Operation diagram. Outline of embryo of chick of
about 46 hours, after Duval. The ruled area shows the site of the operation with the
heated needle. For description of the operation see text.

This embryo was cut into 625 transverse sections. These
confirm the general absence of the amnion, and at the same time
furnish additional data. Back to about the 354th section (forty
sections behind the heart), the somatopleure beneath the embryo
is entirely missing ; evidently it had been torn away by the
operation and had not been replaced. Throughout this region
the extra-embryonic somatopleure begins on each side of the
embryo with a free edge. A short distance behind the heart,
folded portions of the original amnion appear lying in the gap in
the somatopleure, and continuous with the midventral line of the
body- wall. Beginning with about the 3/ist section (see Fig.
13) the body wall is open ventrally, and is continuous with the

I 12

FRANK R. LILLIE.

extra-embryonic somatopleure on one side, while on the other
the original gap in the somatopleure is still open (see Fig. 13).
In this region, the somatopleure for some distance external to

FIG. 12. Experiment 36. Surface view of embryo ; upper surface of blastoderm.
The embryo is anamniote, except for a rudimentary tail-fold, all., allantois ; /. , pel-
lucid area. A-B, plane of section shown in Fig. 13 ; /. , fold of somatopleure.

the part destined for the body-wall is thrown on both sides into
irregular folds that obviously represent the lateral amniotic
folds. They rapidly decrease in size posteriorly, and almost
completely disappear in the region extending from^the 42Oth sec-
tion back, /. c., a short distance back of the fore-limbs. Begin-
ning opposite the hind-limbs the folds again increase in size.
They are very irregular and do not form the normal investment

ORGANS IN THE EMBRYO OF THE FOWL. 113

of the tail. But beneath the latter they form a closed pocket,
the usual tail-fold.

Over the entire region, extending from about the posterior
edge of the fore-limb to the beginning of the hind-limb, there are
no folds in the amniogenous somatopleure. This would indicate
that the normal rapid growth of this region is progressively in-
duced under normal conditions by the extension of the lateral
angles of the head-fold backwards. The folds shown in the
figure are only from about 354—430 and may be explained as
remnants of the original head-fold, the postero-lateral prolonga-
tions of which were probably not entirely removed by the oper-
ation. These folds have not, however, united over the embryo
nor have they induced formation of folds behind them. The

FIG. 13. Experiment 36. Section of embryo along the line A-B, Fig. 12. The
irregular and incomplete amniotic folds are well shown, e.a., ectamnion of the right
side.

reason for this is clear when we consider that the normal process
involves continuous traction on the somatopleure back of the
advancing folds, for the latter are continually fusing along the
dorsal line with those of the opposite side and thus are con-
stantly, so to speak, gathering in the slack, and causing tension.
In the drawing of the entire embryo, the left side is upper-
most, but at the time of the operation the right side was up.
Evidently the embryo was turned over after removal of the blas-
toderm in the process of preparation. This explains why in the
section the lower amniotic fold has the usual appearance of the
upper fold. The ectamnion is visible only on the left side of the
drawing; on the right side no trace of it could be found, except
in the region of the tail-fold.

114 FRANK R. LILLIE.

The results of the destruction of the head-fold of the amnion
in the stage of 46 hours are: (i) Inhibition of the progressive
differentiation of the amniotic zone of the somatopleure ; (2)
failure of the parts of the lateral folds left to unite around the
embryo. The failure of the amniotic folds to unite in the region
where they are best formed and are of more than sufficient length
for enclosure of the embryo shows that the normal union of the
folds is due to the guidance and support of the earlier formed
parts of the amnion.

The tail-fold, however, forms in a fairly normal manner. The
actual abnormalities in this fold are probably secondary, that is,
probably due not so much to direct disturbance of the amnion
itself as to the freedom of movement of the embryo permitted
by the absence of the head-fold, resulting in the withdrawal of
the tail of the embryo from the forming tail -fold.

The body-wall is unenclosed for 1 1 3 sections ; in a normal
embryo of about the same age the body-wall is unenclosed for
about 55 sections. Thus it would appear that the closure has
been delayed.

Experiment 60.

Age of the embryo at the time of operation about 33 hours.

Operation. — The blastoderm was cauterized just lateral to the
right optic vesicle, as in experiment 57, producing a large open-
ing (Fig. 14).

The egg was reopened 72 hours after the operation, and a
large, finely developed vascular area was seen with apparently
no embryo. But more careful examination revealed the naked
hind quarters of an embryo sticking up near the center of the
vascular area, the whole trunk and head of which were plunged
through the blastoderm into the yolk. The head and trunk of
the embryo had slipped through the hole made by the operation
into the yolk-sac, and the edges of the blastoderm around the
original opening had fused in such a way as to close around the
hinder part of the embryo. A large part of the vascular area
was cut out and the embryo was gently floated into a watch
crystal of physiological salt solution. Turning over the blasto-
derm, the embryo was revealed entirely without an amnion (Fig.
15). Not even the tail-fold was found.

ORGANS IN THE EMBRYO OF THE FOWL. I 15

The embryo is represented in Fig. i 5 as it lies on the reversed
blastoderm, the entodermal face of which is up. The allantois
is well developed and lies in a special enlargement of the body
cavity behind the embryo.

FIG. 14. Experiment 60, operation diagram. Outline of embryo of chick of about
33 hours, after Duval. The ruled area indicates the region of the blastoderm destroyed
by the operation.

In this experiment, as in experiment 57, only the right limb of
the ectamnion of the prospective head-fold was destroyed ; and
the consequence of this is in both cases the suppression of the
amnion with the exception of the tail-fold. In this case the vari-
ous membranes have been so confused by the curious position of
the embryo and by various secondary fusions that it is quite
impossible to determine the behavior of the uninjured rudiment
of the amnion of the left side. A single section may serve to
illustrate one of the very peculiar conditions (Fig. 16). Lying
above the embryo is seen the blastoderm composed of the
somatopleure and splanchnopleure. The body-wall of the
embryo has fused with the splanchnopleure in such a way that
the two are directly continuous on both sides, and the body wall
may be traced directly into the wall of the intestine. The result

n6

FRANK R. LII.LIE.

of this fusion must have been an opening on each side into the
yolk-sac ; but this has been roofed over by extension of the

B

FIG. 15. Experiment 60. Under i.e. entodermal, surface of the blastoderm.
The embryo is anamniote, but otherwise quite perfect, all, allantois ; e.b.c., extra-
embryonic body-cavity ; vitelline arteries and veins shown A-£, plane of the sec-
tion shown in Fig. 1 6. The embryo was suspended within the yolk-sac, as described
in the text.

blastoderm surrounding it. Farther back the wall of the intes-
tine becomes continuous with the extra-embryonic splanchno-
pleure.

In the region of the tail rudiments of the tail -fold of the
amnion are found.

Two other completely anamniotic embryos (numbers 1 12 and
124) were produced by experiments similar to those already
described. Both of these had passed through the hole made in
the blastoderm and were suspended within the yolk-sac. One of
these was much farther developed than number 60. They confirm
the general results of the dependence of amnion formation on the
presence of the head-fold. They possess other definite lesions,
the effects of which will be described in another paper.

ORGANS IN THE EMBRYO OF THE FOWL.

117

Condition of t/ic Allantois in Anamniotic Embryos.
The allantois is well formed in four of these embryos ; one
(No. 57) was too young to show it externally. It is obvious
that in the absence of the amnion the growth of the allantois

FIG. 16. Experiment 60. Section along the line A-f>, Fig. 15. bl. , blastoderm
overlying the embryo; sow, somatopleure ; spl., splanchnopleure. On the right side
there is a break in the continuity of somatopleure and splanchnopleure ; this was evi-
dently produced in the preparation, as the continuity is perfect some distance in front,
and also behind.

must be attended with difficulties. When the amnion is normally
formed a large free space is created above and around it, into
which the allantois can freely spread. The absence of this space
causes compression of the allantois, and changes the direction of
its growth, but I do not think that the latter is much impeded.
The mechanical force of the expansion of the allantois causes
separation of the somatopleure and splanchnopleure to proceed
more rapidly in its immediate vicinity than elsewhere (see Fig.
15). In experiment 124 the greater diameter of the allantois
exceeds the greatest length of the embryo. I see no reason why
this process might not provide all necessary space for its expan-
sion. It might be, however, that the resistance offered would
tend to cause accumulation of the products of excretion in the
body of the embryo, and thus gradually poison it.

2. Experiments on the Tail -Fold of the Amnion.
I have also made a number of experiments on destruction ot
the tail-fold of the amnion. The results are in most cases com-

u8

FRANK R. LILI.IE.

plicated by conditions that do not properly belong to the subject
of this paper. There is but one uncomplicated case (exp. 18).
In this experiment the hind-end of the embryo was cauterized
immediately after the appearance of the tail-bud (Fig. 17), thus

FIG. 17. Experiment 18, operation diagram. Outline of embryo of chick of
about 52 hours, after Duval. The ruled area represents the part destroyed by the
heated needle.

destroying the tail-fold of the amnion. When the egg was re-
opened forty-eight hours later, a well-developed embryo of about
five days was found in which the amnion ceased with a free edge
immediately in front of the hind-limbs (Fig. 18).

The conditions of the membranes in this embryo are other-
wise very complicated and difficult to understand. Thus there is
in addition to the amnion a fold of the blastoderm surrounding
both amnion and embryo (Fig. 18). In the posterior half of the
embryo the body-wall is directly continuous with the wall of the
intestine as in 60. As this embryo will come up for description
elsewhere, I shall not dwell further on this topic.

The fact that stands out distinctly is that the tail-fold of the
amnion has not regenerated and that the head-fold has not
compensated for the absence of the tail-fold by continuing its
growth backwards. However, I have a number of embryos in

ORGANS IN THE EMBRYO OF THE FOWL. 119

which a complete amnion has been found without any tail-fold.
These embryos, are, however, defective at the hind end, so that

FIG. 18. Experiment 18. The embryo 48 hours after the operation. The tail-
fold of the amnion has not regenerated. The amnion ends with a free edge in front
of the hind-limbs. A fold of blastoderm is wrapped around the embryo and amnion.
Under surface of blastoderm.

one has not to attribute any work of supererogation to the ante-
rior lateral folds to explain the complete closure. This also will
be discussed elsewhere.

GENERAL DISCUSSION.

The formation of the amnion of the chick seems to be a proc-
ess with extraordinarily slight power of regulation.1 A slight
injury to part of its early rudiment sets the whole process astray.
It is thus an extremely good example of correlative differentia-

1 Barfurth ('02) notes incidentally in one of his experiments " die Amnion war
regenerirt." As I understand him, he means by this simply that an aperture made
in the amnion in the course of an experiment on the eye closed up. I can confirm
this from my own observations. I have found that even considerable tears made in
the amnion after its formation may close completely.

I2O FRANK R. LILLIE.

tion. The correlations in the development of the amnion are of
three kinds :

1 . Mechanical. — Under this head I class the elevation of the
lateral amniotic folds, which takes place only after the establish-
ment of the head-fold, and which is omitted, if for any reason the
head- fold fails to appear or is destroyed.

2. TropJiic Stimulation. — Under this head I class the influence
of the traction exerted by the union of the right and left amniotic
folds on the amniogenous somatopleure, and the influence of the
turning of the embryo on the amniogenous somatopleure of the
left side. The influence of the traction in either case is to in-
crease the extent of the amniogenous somatopleure, in part (pre-
sumably) by stimulating its growth, in part undoubtedly by mere
stretching. If, owing to failure of formation of the head-fold,
such traction is not exerted on the somatopleure it does not ex-
pand nearly to the normal extent.

3. Differential Stimulation. — Under this head I class (doubt-
fully) the propagation of the ectamniotic thickening along the
somatopleure ; though this may be a process of self-differen-
tiation.

Self -differentiation of the Formation of the Amnion. — The for-
mation of the original rudiments of the ectamnion may be a
process of self-differentiation, though the definite relation of the
anterior and posterior rudiments to the head and tail respectively
suggests correlation with their formation. Moreover, a slight
histological differentiation appears between the amniogenous and
choriogenous somatopleure, before, and even in the absence of, the
formation of folds, which is apparently not correlated with any
other of the processes observed.

Beyond this mere classification I do not desire to go at present,
but will reserve a general discussion of principles until the com-
pletion of other parts of the present series.

In conclusion I simply summarize the results :

I. MorpJwlogical.

i. The amnion is primarily an organ of the ectoderm in the
chick. The ectamnion first forms in front of the head and dif-
ferentiates progressively backwards towards the posterior end,

ORGANS IN THE EMBRYO OF THE FOWL. 121

where it is met by the posterior ectamnion differentiating for-
wards. Thus the amniotic zone of the somatopleure is marked
off from the chorionic zone.

2. The head-fold is formed from the ectamnion with the coop-
eration of the amnio-cardiac vesicles and of the proamnion which
is depressed between the former. The immediate prolongation
of the head-fold is produced by the progressive fusion of the
ectamniotic rudiments backwards, and it includes only an ex-
tremely small part of the proamnion.

3. The tail-fold is likewise formed primarily by the ectamnion
with participation of the amnio-allantoic enlargements of the
body-cavity.

4. There are certain constant secondary folds in the upper
(right) limb of the amnion produced by the turning of the em-
bryo. These persist longest in the region of the heart and im-
mediately behind it.

1 1 . Experimental.

1 . Destruction of the anterior ectamniotic rudiment of one side
prior to the formation of the head-fold of the amnion results (<?)
in permanent absence of the amnion back to the hind-limbs
(exp. 57); (b) in inhibition of the growth, and almost complete
suppression of the folds of the amniogenous somatopleure of the
uninjured side; from which we may conclude -

2. That the growth of the amniogenous somatopleure is nor-
mally induced by the traction exerted on it by the progressive
fusion of the folds, and that the uprising of the folds is due to
the lifting power of the same process of fusion.

3. The tail-fold and posterior lateral folds cannot replace the
anterior lateral and head-folds, nor can the latter replace the
former.

4. Not only the initiation, but also the progress of the forma-
tion of the anterior lateral folds is dependent upon the perfection
of the head-fold (exp. 36).

5. The absence of the amnion has, at least for a time, only a
limited effect on the development of the allantois.

6. Inasmuch as the embryo may develop perfectly normally
to the stage of five or six days without the amnion, it is obvious
that the functional significance of the latter must be slight during

122 FRANK R. LILLIE.

this period. It yet remains to be determined how far the embryo
may develop without the amnion (see quotation from Dareste
below).

7. There is a certain relation of interdependence between the
formation of the amnion and the body-wall. In the absence of
normal formation of the lateral folds of the amnion the closure
of the somatopleure to form the body-wall proceeds more slowly
than usual.

Dareste ('79) has observed total absence of the amnion in em-
bryos of the chick. The condition was not, however, produced
experimentally. His observations and conclusions are given in
the following quotations :

"J'ai signale, depuis longtemps, 1'arret de developpement de
1'amnios et les anomalies nombreuses que cet arret partiel deter-
mine chez 1'embryon. C'est la cause la plus frequente des mon-
strouosites simples. II y a des cas, beaucoup moins nombreux,
il est vrai, dans lesquels 1'amnios fait completement defaut.
L'embryon est alors en continuite directe, par son enveloppe cu-
tanee, avec le feuillet sereux du blastoderme, qui ne s'est pas
plisse pour former la poche amniotique. J'ai vu, dans plusieurs
de ces cas, 1'embryon se constiteur d'une maniere parfaitement
normale. La paroi thoraco-abdominale s'etait completement for-
mee, et la continuite de 1'embryon avec le feuillet sereux consti-
tuait une sorte de cordon ombilical. L'allantoide sortant de
1' abdomen par se cordon s'etait engage entre le feuillet sereux
et le feuillet vasculaire."

" Les embryons, ainsi prives d'amnios, peuvent vivre pendant
un temps assez long. J'ai constate 1'absence complete de 1'am-
nios sur en embryon de treize jours, qui etait plein de vie et par-
faitement normal. Rien ne pouvait faire penser qu'il mourrait
prochainement. II est tres-probable cependant qu'il n'aurait pas
atteint 1'epoque de 1'eclosion. L'absence de 1'amnios aurait mis
obstacle au developpement complet de 1'allantoide : ce qui aurait
produit 1'asphyxie de 1'embryon, comme je 1'ai montre depuis
longtemps. Le plus ordinairement 1'absence de 1'amnios amene
la mort precoce de 1'embryon. Souvent aussi elle determine,
dans son organisation, des modifications teratogeniques pro-
fondes."

ORGANS IN THE EMBRYO OF THE FOWL. 123

" Toutes des observations nous font connaitre le role physiolo-
gique de 1'amnios dans la vie embryonnaire. II est bien evident
que 1'amnios protege 1'embryon centre toutes les actions meca-
niques qui tendraient a le comprimer."

HULL ZOOLOGICAL LABORATORY, UNIVERSITY OF CHICAGO,
April, 1903.

POSTSCRIPT.

After the foregoing paper was fully printed, my attention was
called to an article by Weldon in which anamniote embryos of
the fowl were described, and which I had overlooked owing to
the fact that the observations were included in an article entitled
"Prof, de Vries on the Origin of Species" (Biomctrika, Vol.
I., Part III., April, 1902). Partial or complete suppression
of the amnion resulted from experiments to replace the water
lost by evaporation in the incubator without preventing the
process of evaporation itself. "A hole was made in the broad
end of the egg-shell and the subjacent membranes, into which one
end of a siphon, filled with water, was fitted. The other end of
the siphon was placed in a reservoir of water, and the whole ap-
paratus placed in an incubator. In from 20 to 30 per cent, of
the embryos treated in this way the amnion was largely or en-
tirely absent after incubation for three or four days."

Weldon does not discuss the mechanics of formation of the
amnion, but treats the result simply as an example of a definite
relation between the environment and an extremely stable char-
acter. Apparently the immediately effective factor in the experi-
ments was the increased pressure within the shell, which, pre-
sumably, forced the embryonic area into immediate contact with
the shell membrane, and thus prevented the uprising of the

amniotic folds.

F. R. L.

124 FRANK R. LILLIE.

LITERATURE CITED.

Barfurth, Dietrich und 0. Dragendorff.

'02 Versuche iiber Regeneration des Auges und; der Linse beim Hiihnerembryo.
Anat. Anz. Erganzungsheft zum XXI. ,Bd., 1902. Verb, der anat. Ges.
auf der 1 6 Versamml. in Halle a/S, 1902.

Dareste.

'79 Sur 1'absence totale de Tamnios dans les embryons de Poule. Comptes ren-
clus Acad. des Sc., LXXXVIII., 1879, pp. 1329-1332.

Hirota, S.

'94 On the Sero-Amniotic Connection and the Fcetal Membranes in the Chick.

The Journal of the College of Science, Imp. Univ. Japan, Vol. II., part IV.,

1894, pp. 337-370- Plates XV.-XVII.
Schauinsland, H.

'023 Die Entwickelung der Eihaute der Reptilien und der Vogel. In Handbuch

der Vergl. und Exper. Entwickelungslehre der Wirbeltiere herausgegeben

von Oscar Hertwig. Kap. VII., pp. 177-234.

Schauinsland, H.

'oab Beitrage zur Entwickelungsgeschichte der Wirbeltiere II. Beitrage zur
Entwickelungsgeschichte der Eihaute der Sauropsiden. Bibliotheca Zoolo-
gica, 1902.

Schenk, S. L.

'71 Beitrage zur Lehre vom Amnion. Archiv fiir mikr. Anat., VII., 1871, pp.
192-201, Taf. XVIII.

CROSSOBOTHRIUM LACINIATUM AND DEVELOP-
MENTAL STIMULI IN THE CESTODA.

W. C. CURTIS.

In the spiral valve of the "sand shark" (Carcharias littoralis)
taken from the Woods Roll region there is found in a large
majority of the specimens examined the Cestode, Crossobotlirinin
laciniatinn. This genus and species was first described by Linton
(" Rept. U. S. F. Com." for 1886), and in subsequent papers
appearing in the same publication or in the " U. S. F. C. Bul-
letin," he has added further important notes, the whole making
an accurate and satisfactory systematic description.

A striking feature of the species is the remarkable clearness
with which the important features of Cestode structure can be
demonstrated. The water vascular system, main trunks and
flame-cells can be seen in the fresh specimen with the greatest
ease. Almost every detail of the complicated reproductive or-
gans is seen in well-stained whole mounts of the motile proglot-
tids and much of this in specimens freshly prepared. The mode
of using the suckers on the head, the activities of the motile pro-
glottids and their mode of egg-laying and the development of
these eggs in sea-water as far as the six-hooked embryo are all
easily demonstrated. Moreover, there occurs in the cystic duct
of the squeteague (Cynoscion rcgalis], a not uncommon food of
the "sand shark," a tetrabothrian larva which, if not the larva of
Crassobothrium laciniatum, probably belongs to some very closely
related form. This larva, which was first described and figured
by Linton ("Rept. U. S. F. Com.," 1886), is again an extremely
favorable object for study.

If it is possible to obtain conclusive evidence that this tetra-
bothrian larva of the squeteague is indeed the larva of C. lacin-
iatum, we shall have but one gap in the life history of this species,
viz., the transfer to the squeteague of the six-hooked embryo
which develops in the open ocean.

Such favorable material it seemed to me might present, upon
careful examination, facts which would be suggestive along the

I25

I 26 W. C. CURTIS.

line of some ot the general problems involved in Cestode para-
sitism and development in addition to the possible opportunity
for fixing the life history of this particular form. With this iir
mind I have been collecting all the data bearing upon the life his-
tory and during the summer of 1902 I made the first of a series
of experiments in infection which I hope to continue and which
may lead to more precise knowledge concerning the identity of
the larva found in the squeteague.

I wish in this paper to describe the important features in the
structure of the motile proglottids, its egg-laying and other activ-
ities, to give some observations on the larva from the squeteague
and to discuss the view point which my study of the development
in this and other Cestodes has suggested to me.

THE MOTILE PROGLOTTIDS.

When an incision is made in the spiral valve of an infected
" sand shark " the Cestode is frequently found in such abundance
that, as the elongated bodies and the motile proglottids writhe
about in the chyle, one often wonders how there can be enough
nourishment left for the host. I can confirm Linton's record, of
" sand sharks " taken at different times, that in the great majority
of individuals there are literally hundreds of this parasite in the
spiral-valve to the exclusion of all others. When the parasites
are examined in sea-water the alternate protrusion and retraction
of the bothria, as described by Linton, can be observed for hours.
When a scolex is compressed on a slide the flame cells of the
water vascular system can be observed for a considerable time
before they succumb to the abnormal conditions.

The ripe proglottids which can be pulled from the long strobilae
or found loose in the intestine are very active and constantly
changing their shape. A typical outline in a partially extended
condition is represented in Fig. I, and the fully elongated condi-
tion is represented on a smaller scale in Fig. 2. At the anterior
tip I have found in preserved specimens minute projections which
have the appearance of cilia (<r), but which will probably prove
upon examination in the living specimen to be minute spikes sim-
ilar to the larger ones on the penis (/>).

The four ear-like flaps at the posterior end which give the

DEVELOPMENTAL STIMULI IN THE CESTODA.

127

strobila its characteristic appearance are frequently curled back
and outward, giving the posterior end a quite different outline.
In the living specimen I have frequently seen masses of sperm
ejected from the penis, but my records of this do not mention

wr

u

YQ

VDl

VD

uo

ED

SR

YD

YQ

YG

SG O

FIG. I. Reproductive organs of a motile proglottid of C. laciniatum. C, cilia,
like spikes at anterior tip ; ed, egg duct from shell gland to uterus ; gp, genital pore ;
o, ovary ;/, penis ; fs, penis sheath ; sg, shell gland ; sr, seminal receptacle ; t, testes ;
tt, uterus ; no, uterus opening through which the eggs escape ; vd, vas deferens ; I'd' ,
denser inner end of same ; i>a, vagina ; wt, large water vascular tube ; yd, yolk duct ;
yg, yolk glands.

128

W. C. CURTIS.

the condition of the female organs in the proglottids thus ob-
served. There are four main water vascular tubes. The larger
pair lie on the same side of the body as the uterus opening and
in the majority of cases one or both of them can be traced to a
bulb-like enlargement on either side near the anterior tip of the
proglottid. Posteriorly each one seems to end in the angle of
the broadly wedge-shaped concavity formed by the projecting
flaps. They here seem to end blindly against the cuticle which is
perhaps perforated. There is no cross connection between these
two vessels nor any common posterior opening such as is fre-
quently stated to occur in Tcenias. The second pair of main
trunks are vessels of much smaller diameter and lie on the other
side of the flat body immediately under the larger pair. Ante-
riorly they can sometimes be traced a little farther forward than
the bulbs of the larger vessels, but do not seem to end in an en-
largement. It is almost impossible to follow these smaller trunks
for any distance posterior to the penis as the yolk glands are
here closely packed together and obscure everything else.

FIG. 2. Ripe motile proglottid fully extended, showing full uterus and the larger
pair of water vascular trunks.

When the proglottids are examined alive much of their struc-
ture is obscured through the presence in the parenchyma of the
highly refractive and closely packed granules of calcium carbon-
ate. A very easy way of ridding the proglottid of this and pre-
paring it for immediate examination is to use ten per cent, nitric
acid and the pressure of a cover-glass. This dissolves the cal-
cium carbonate and leaves the specimen quite transparent. This
is a valuable method for the rapid examination of the principal
organs, but for the finer details one of course needs more careful
fixation and a good stain. I have found corrosive sublimate
with about five per cent, acetic acid followed by Czokor's alum
cochineal an excellent combination for the demonstration of the
features given below.

DEVELOPMENTAL STIMULI IN THE CESTODA.

129

The cirrus (/, Fig. i) is eversible, working on the same prin-
cipal as a Nemertean proboscis, a type common in Cestodes.
From its base the much-coiled vas deferens (fd) leads away and
is found throughout these coils crowded with sperm. At its
inner end it has a denser wall and is of less diameter for a short
distance (i>d) and then divides into the vasa efferentia which can
be seen radiating to the area in which the testes are located and
in favorable cases followed down to the testicular follicles them-
selves (Fig. i).

The vagina (W) which opens on the genital papilla just above
the penis will be seen in the figure to pass inward and curve
around backward, passing behind the mass of finger-like follicles
which constitutes the ovary (o). It is here enlarged into the

SR

YD

SG

FlG. 3. Ducts of female complex with ovary lobes left out and ducts reflected
slightly to show connections, cd, duct from ootype to inner end of vagina ; cyd, com-
mon yolk duct ; eJ, egg duct ; o, ovary ; of, ootype ; sg, shell gland ; sr, seminal
receptacle of vagina ; v, inner end of vagina ; x, meeting place of ova and yolk ; yd,
right and left yolk ducts.

seminal receptacle which will be found full of sperms. Lying
among the posterior lobes of the ovary is the shell gland (sg) to
which the yolk is delivered from a common yolk duct formed by
the union of a single yolk duct (yd) coming from either side.
Extending anteriorly from the shell gland and beneath the ovary

130

\V. C. CURTIS.

in this figure is the egg duct (ed~) which conveys the eggs to the
uterus. The complex of ducts in this region is shown in a recon-
struction from sections represented in Fig. 3. The lobes of the
ovary which are packed closely around the ducts are here omitted.
This figure may be compared with what is shown in Fig. i,
where some of the same parts appear. The lobes of the ovary
all converge upon a right and left portion (Fig. 3, o) and these
main parts, on uniting, open posteriorly into a spherical cavity
(of) with thick walls, which is probably where the ova and sperm
meet. A duct (cd} passes from this cavity to the inner end of
the seminal receptacle (sr) and thence straight back to the shell
gland (sg). Into the shell gland the common yolk duct (cyd)
opens and from this common meeting place of the yolk and fer-

YQ

FIG. 4. Branching of yolk duct to yolk glands, yd, yolk duct ; yc, yolk cells in
duct ; yg, yolk glands ; wt, large water vascular tube.

tilized ova, after the acquisition of a shell, the fully formed egg
passes up the egg duct (ed] into the uterus.

The right and left yolk ducts branch as they reach the areas
of the yolk glands on either side and some of the branches may
be seen going to individual yolk follicles (Fig. 4). These main
branches are often found closely wrapped around the large water
vascular tubes (wt] of either side. Yolk cells (yc) may often be
found on their way down these ducts and accumulated in large
numbers at their median ends. They are also seen in the short
ducts which run from one yolk gland to another in all parts of

DEVELOPMENTAL STIMULI IN THE CESTODA. 131

the mass. The yolk-producing organ consists of follicles densely
packed with yolk cells and distributed in the proglottid as the
figure indicates (yg, Fig. i).

Fig. i represents a specimen killed under pressure and in
which the uterus had been ruptured and the eggs squeezed from
the oval hole represented by the dark outline in the center of the
proglottid. Very much the same sort of a hole is left when the
proglottid ruptures itself at this point in the normal egg-laying.
The extent of the uterus cavity is indicated by the outline (it) in
the figure. The condition of the intact uterus and the place of
its rupture will be explained in describing the egg-laying of the
ripe proglottids.

ACTIVITIES OF THE MOTILE PROGLOTTIDS.

The proglottids of Crossobothrium laciniatum are an extreme
case of what is usually termed the "motile" condition. So
definite are their movements and activities that one is constantly
thinking of them as though they were individual animals of a
species entirely distinct from the parent scolex. • When observed
in the chyle they are seen writhing about, contracting and elon-
gating rhythmically and bending their bodies into an arch along
the axis of breadth, now one way, now another. If we measure
the maturity of a proglottid by its size and the number of eggs
accumulated in the uterus the conclusion is reached that the pro-
glottids as taken from the spiral-valve are of diverse ages, for one
finds a considerable variation in the number of eggs accumulated
in the uterus and a correlated variation in the size of the pro-
glottids.

When placed in clean sea-water the smaller proglottids do not
lay their eggs, while the large ripe ones will almost immediately
do so. These facts seem to indicate that the proglottids may be
shed off from the strobila some time before they are ripe and
remain in the shark's intestine until they are fully loaded with
eggs and ready for the laying. The enormous number of pro-
glottids usually found in a single spiral-valve is another fact in
favor of this conclusion. On the other hand fully mature pro-
glottids are frequently found on the end of a strobila (Fig. 5),
showing that they may mature while still attached to the parent
stock.

132 W. C. CURTIS.

When carefully examined the ripe proglottids at the posterior
end of a strobila (Fig. 5) or the mature motile proglottids found
free in the chyle show a breast-like protuberance upon that face
on which the uterus opens. The resemblance of this to a breast
is heightened by the existence of a nipple-like prominence at the
summit, as is shown in the side view given in Fig. 5. The general
protuberance is caused by the distension of the uterus, though it
sometimes seems to be enhanced by a concavity on the opposite
face of the segment as the dotted line of the figure indicates.
Motile proglottids when in this ripe condition show, if examined
in the chyle, the ordinary writhings and indefinite locomotor
movements noted above. If, however, a number of these ripe
and full proglottids are transferred from the chyle into clean sea-

FlG. 5. Side view of a ripe proglottid.

water the egg-laying will presently be observed. In making
observations on this process I was accustomed to select carefully
ten or a dozen proglottids which seemed fully ripe and transfer
them all together into a dish of clean sea-water. When this was
done it was found that about eight out of ten thus selected laid
their eggs in three or four minutes. Any of those remaining
might lay after a little longer period or not at all. A similar
reaction of whole chains of proglottids is recorded by Schauinsland
(Jena, Zeitsch. 1886), for Botlirioccplialits latns, Trianophorus iiodu-
losns and Lignla simplissima.

When proglottids of Crossobothrium are taken at random and
thus placed in sea-water only a small proportion, no more than
one fourth or one fifth, will ever lay their eggs. When the small
proglottids which have only a few eggs in the uterus are thus
taken no egg-laying follows in any case.

That the proglottids as found in the spiral-valve at any one time
are not all of same maturity is thus clearly shown and I think we
are justified in the conclusion that these immature proglottids
tend to remain in the spiral-valve until they become fully ripe

DEVELOPMENTAL STIMULI IN THE CESTODA.

133

and then to pass out with the faeces, even though their early de-
tachment from its strobila may have been premature and caused
by the outward passage of the excreta or the contraction of the
intestine.

The manner in which the egg-laying proceeds in any single
proglottid thus placed in sea-water is a very interesting thing to
watch. Extreme writhing movements of a quite definite sort
begin at once. The proglottid bends along its axis of breadth
until it is almost a closed ring, the pointed anterior end some-
times passing into the angle made by the posterior flaps (Fig. 2)
and thus reminding one of an acrobat who could bend backward
until his head should be thrust between his legs from behind,
then the proglottid straightens and the bend is reversed, it

uo

uo

FIG. 6. Proglottid in the act of egg-
laying seen from the side. Lettering same
as for Fig. 7.

FIG. 7. Proglottid at close ot egg-
laying seen from anterior end. a, an-
terior end ; /, penis ; g, groove marking
outline of uterus ; uo, ruptured area
through which the eggs escape.

straightens again and bends into the first position and so on ;
these motions continue until the nipple-like prominence of the
protruding body bursts and the liberated eggs gush forth. When
the break occurs the extreme violence of the writhing ceases and
the proglottid bends backward rather more than forward (Figs. 6
and 7) until all the eggs have been expelled, when a gaping
hole is presented (Figs. 6 and 7) where once was the distended
uterus. Even when there is hardly an egg left within it, the
straining movements of the proglottid continue as though it were
making sure that not a single egg remained. Figs. 6 and 7
represent proglottids in which the egg-laying had been almost

134 w- c- CURTIS.

accomplished and show the characteristic attitude of the proglottid
from a front and side view.

Proglottids which have thus stripped themselves of their eggs
may continue to live in sea- water for a day or two, but I have
not experimented with them to ascertain how long their existence
may be prolonged.

The rupture of the uterus may be very readily produced in a
proglottid having any considerable accumulation of eggs, if a
little pressure is applied with a cover-glass or otherwise. But
the process above outlined is something brought about spon-
taneously by the proglottid itself after it is transferred to sea-
water. Proglottids which have been artificially compressed in
killing, for whole preparations almost always have the uterus rup-
tured and the eggs discharged. They then present the appear-
ance shown in Fig. I, of a large oval hole opening into the
uterus cavity, while the boundaries of the latter can be traced as
a very delicate outline still conforming to the general outline of
the full uterus. A proglottid which is in the last stages of egg-
laying after the spontaneous rupture of its uterus shows the same
sort of opening, but perhaps more widely distended. Such
specimens which were killed without compression are shown in
Figs. 6 and 7.

My observations are then that the proglottids when large and
having the uterus full of eggs (Figs. 2 and 5) will, by a quite
definite series of muscular contractions and writhings, rupture the
nipple-like prominence at the summit of the protruding uterus
(Fig. 5) and allow the eggs to gush forth, the proglottid con-
tinuing its writhing movements in a less pronounced degree even
after all the eggs have been shed (Figs. 6 and 7). The fact
that this egg-laying occurs immediately after the ripe proglottid
is transferred from the chyle to clean sea-water will, I think, con-
vince any one that the same process occurs when a ripe proglottid
of Crossobothrium passes in the normal course of its existence
out of the shark's cloaca into the water of the ocean. We may
conclude I think that such a proglottid, upon coming into contact
with the sea-water outside, goes through muscular contractions
similar to those observed in the laboratory and lays its eggs free
in the open ocean, and that these pelagic eggs are thus widely

DEVELOPMENTAL STIMULI IN THE CESTODA. 135

scattered. The short period between the first contact with the
outer water and the egg-laying indicates that the infection of the
intermediate host is, by means of countless embryos, developed
in the open ocean and not by the eating of the intact proglottids
with their contained eggs.

EGG DEVELOPMENT.

On collecting the eggs laid by proglottids in the laboratory
and placing them in dishes in which the sea-water can be kept
reasonably pure, development ensues as far as the six-hooked em-
bryo stage which I have represented in Fig. 8, drawn from the
living specimen. I did not succeed in obtaining embryos be-
yond this stage, and therefore cannot say whether the embryo

B

FlG. 8. Six-hooked embryo of Crossobothriuni laciniatuw. m, egg membrane ;
r, remains of outer envelop ; e, ectoderm of Schaunisland ; />, six-hooked embryo.

enters the next host in this condition or as a ciliated larva (Schau-
insland, " Bothriocephalidae,'' '86) which subsequently hatches
from the embryo figured.

In the common Tsenias and those Cestodes which have simi-
lar hosts and conditions of life-history, the fertilized eggs on
passing into the uterus develop there into six-hooked embryos
and remain in that stage until they reach the tissues of the in-
termediate host. In the Bothriocephalidae, Schauinsland de-
scribes B. rugosits 'as having such an intra-uterine development
as far as the six-hooked embryo, and B. latus, Trianophorns
nodnlosus and Ligula siniplissiuia as producing eggs which de-
velop only when they reach the water. In the last three species
the eggs accumulate in the uterus in the condition of resting
oosperms surrounded by their yolk and the egg-shell, but do
not develop until they are laid in the water by the parent pro-
glottid after this has left the host's intestines.

136 W. C. CURTIS.

My material was prepared for the general anatomy rather
than for this particular point, and the egg capsules are greatly
shrunken, but as nearly as I can make out from a careful exam-
ination of the uterine eggs they are all in the condition above
noted, viz., a resting fertilized ovum plus the yolk cells and egg
capsule. In Cestodes having this mode of development, there-
fore, the eggs accumulate in the uterus, but do not develop until
they are laid by the proglottid. The resting stage is comparable
to the resting condition known in many forms which lay winter
eggs or eggs which develop only after a considerable time. In
any given proglottid such Cestode eggs are of diverse ages, de-
pending upon how long they have been in the uterus, but are all
alike inhibited from development until the proper conditions are
present. Upon contact with the external water this resting stage
is stimulated, or, we might say, some inhibition is removed, and
development ensues, continuing as far as the six-hooked embryo.
Whether there is some specific thing in the sea-water which can
be fixed upon as the stimulus to development I have not ascer-
tained, but it has seemed to me quite possible that the stimulus
can be located in some specific chemical constituent of the sea-
water.

From what we know of other Cestodes it is unlikely that the
six- hooked embryo of Crossobothrium can develop further in the
water, and that if this embryo does not find the appropriate stim-
ulus for further development by meeting with its next host it must
perish. What this next host is one would perhaps discover as
much by accident as by the most persistent work. Could I
clearly demonstrate that the larva found in the squeteague is the
young of Crossobothrium ladniatmn — and there is a good deal ot
general evidence that this is the case-- 1 think that the nature of
the squeteague's food (young herring, adult herring, menhaden,
etc. --Peck, "Sources of Marine Food," U. S. F. C. Bull., '95)
would lead us to suspect that the six-hooked embryo, instead of
passing directly to the squeteague, might have an intermediate
host like the menhaden, herring, or some fish which feeds upon
the microscopic elements in the sea-water.

DEVELOPMENTAL STIMULI IN THE CESTODA. 137

DEVELOPMENTAL STIMULI IN THE CESTODA.

To one examining closely the present trend of opinion regard-
ing the process which our nomenclature still designates as fer-
tilization it is, I think, quite apparent that the evidence and
conclusions so logically and convincingly set forth by R. Hert-
wig l are gaining a wide acceptance both among special workers
and among those who are viewing the data with a somewhat
broader perspective. The details of Hertwig's paper had been
already accepted by some of the investigators whom he cites, and
so far as they concern the Protozoa they are mentioned by
Calkins in his recent book as well-established facts, but this
cannot detract from the able manner in which Hertvvig has sum-
marized these facts and indicated the important conclusions to be
drawn from them. The work which to most persons offers con-
vincing evidence of the twofold nature of the process we have
been calling normal fertilization is the work in " artificial par-
thenogenesis " initiated by Loeb and Morgan. One could hardly
ask for a more convincing proof that the union of the germ
plasms and the "developmental stimulus," as Hertwig calls it,
are distinct and separable phenomena although they have become
almost indissolubly connected with one another throughout the
Metazoa.

It having been shown in these experiments so far as they have
now gone that a stimulus to development may be operative en-
tirely independent of any union of the germ-plasms, as for example
in the development of eggs of echinoderms or worms upon treat-
ment with salt solutions and other stimuli, we should I think seek
no less for the converse proposition, viz., tlie union of the genn-
plasm and the absence of developmental stimulus (as evidenced by the
absence of development) while the oo sperm continues its life. We
should then seek for the proper stimulus to this resting condition
and by means of this stimulus initiate at will the developmental
changes. To illustrate by a hypothetical case, suppose we could
have a form where it were possible to bring about at will the union
of the ovum and spermatozoon and their nuclei but by subtracting

1 " Mit welchen Recht unterscheidet man geschlechtliche und ungeschlechtliche
Fortpflanzung ?" Sitz. Ber. Gesel. f. Moi'ph. und Physiologic in Mi'uichen, Nov.,
1899. Translated in Science, N. S., Vol. XII., no. 312, Dec. 21, 1900.

138 \V. C. CURTIS.

the " developmental stimulus " to have the resulting oosperm go
into a resting state which would result eventually in its death unless
the " stimulus to development " intervened to start a new cycle.
If this developmental stimulus, or some substitute for it, and its
action were accurately known we might apply it at will at any
time after the union of the germ-plasms was completed, and so
long as the oosperm remained alive it would respond by begin-
ning the new cycle of development.

In the thorough understanding of the two apparently distinct
processes involved in fertilization as hitherto understood we
should I think be greatly aided by the investigation of the prob-
lem just outlined which is the converse of that brought out by the
work on " artificial parthenogenesis."

We have, it seems to me, cases of normal development which
are parallel to this converse proposition in just the same way that
normal parthenogenesis is a parallel to the artificial development
induced by salt solutions and other stimuli. These cases are
seen in the development of those forms in which the fertilized
egg has a long resting stage. Where freezing or desiccation is
necessary for such subsequent development, this condition may
be regarded as one of the developmental stimuli, although if one
tries to picture how a state of affairs necessitating this particular
condition could have arisen in the past he must, I think, feel cer-
tain that though extreme cold or dryness may now be a neces-
sary factor it must originally have been unnecessary if not a thing
fatal to the further existence of the organism.

Any one familiar with biological literature can readily recall
numerous cases of eggs with longer or shorter resting periods
following upon the union of the germ plasms. The particular
case I call attention to is that presented by those Cestoda which
in their development have resting fertilized eggs gradually accu-
mulating in the uterus as in the case in BotJiriocepJmlus /afns, Tri-
anopJionis nodulosus, Ligulasimplissima, etc., and in Crossobothrium
Inclination. On referring to the description which has been given
of the production of the eggs and their extra-uterine development
it will be seen that this illustrates particularly well what I have
termed the converse of artificial parthenogenesis and that the
hypothetical case which I set forth on page 137 might be substi-

DEVELOPMENTAL STIMULI IN THE CESTODA. 139

tuted almost word for word for a description of what occurs in
one of these Cestodes.

In addition to the phenomenon of a resting stage in their early
development the Cestodes just mentioned, and indeed all other
Cestoda, present in their subsequent life history a feature which
I have found interesting when considered in connection with
the primary "developmental stimulus" which starts the oos-
perm upon its course. Such a consideration of the subsequent
facts of Cestode life history may perhaps widen our concep-
tions regarding the nature of one of the two phenomena
which exist side by side in normal fertilization. How this is
will be easily apparent if we recall the life history of Crosso-
bothriuin or any Cestode having a similar extra-uterine develop-
ment.

The female reproductive organs of each proglottid produce
ova which on being fertilized become surrounded by their yolk
supply and encased in a tough shell. Without undergoing any
developmental changes they accumulate in the uterus where they
remain in this condition until the time of egg-laying. They are
thus of very diverse ages if we date the age of each from the
time of the entrance of the spermatozoon, but all are in the same
resting unicellular state. We have here the union of the germ
plasms, but the stimulus to development delayed for a period
which is long or short, depending upon the age of the individual
oosperm. The stimulus to development is normally found in the
contact with the outside sea-water when the eggs are shed, for
the cleavage begins only when they are thus set free. Develop-
ment proceeds so far as the six-hooked embryo stage when death
ensues unless the proper host is found. In the case of Crosso-
bothrium there is perhaps a primary intermediate host between
the six-hooked embryo and the squeteague in which case the
six-hooked embryo which infects this intermediate host receives
a stimulus to development which sends it so far as the resting
stage which is attained in that particular host, and here it stops
and eventually comes to naught unless it is carried into the next
host, the squeteague, where it finds a new stimulus to further
developmental changes and attains in the cystic duct of this fish
its development to the full structure of a tetrabothrian larva.

I4O W. C. CURTIS.

Here again death ensues unless the next stimulus in the series is

o

forthcoming, viz., the contact with the digestive juices of the
"sand shark's " stomach. When this stimulus is furnished the
tetrabothrian surviving the wreck of its teleost host develops into
the final adult condition.

The foregoing is stated in sufficiently general terms to be ap-
plicable, mutatis mutandis, to any Cestode having an extra-uterine
development and whether or not the life history outlined for
Crossobothrium is the correct one does not affect the general
conception of Cestode life history which I am attempting to
portray.

In stating the above I have spoken of the reaction of the embryo
of a given stage to its particular stimulus just as I spoke of the
reaction of the oosperm to the stimulus which initiates the whole
cycle. If we ask ourselves what is the essential nature of the re-
action of the oosperm to the delayed developmental stimulus we
must designate it as primarily a reaction manifest in so many cell
divisions and subsequent differentiations. In what does the re-
sult produced by any one of the stimuli, which, if the embryo
runs its whole course, become applied to each of the successive
stages, differ from the result produced by the stimulus which acts
upon the oosperm ? Cannot the result in each instance be for-
mulated in the same way, viz., that the stimulus causes cell
divisions and subsequent differentiation ? And should we not
speak of all of them as " developmental stimuli " ?

The exact nature of the stimulus at each stage is a thing which
in spite of many technical difficulties would be open to investiga-
tion and something which we may hope eventually to understand,
but however diverse the stimuli might be we should still have,
as above stated, changes of the same nature resulting from each
successive stimulus. The stimulus which is in the first instance
something in connection with the sea-water is probably in the
other cases a change in nourishment incident to the change
of hosts, but in every case the result of the stimulus may be
stated in the same general terms, viz., cell division and subse-
quent differentiation which ends in a condition of stable equilibrium
in which the animal finally perishes unless the next stimulus is
forthcoming.

DEVELOPMENTAL STIMULI IN THE CESTODA. 14!

Viewing then the life history of such a Cestode from this point
of view we have first the union of the germ plasms followed
by a resting stage of varying duration. A stimulus furnished by
the contact with the sea-water when the eggs are laid brings about
the changes resulting in the six-hooked embryo. This embryo
when it receives a certain stimulus (condition of nourishment or
otherwise) from the intermediate host goes as far along the
course of development as the mid-larval stage and stops again.
On reaching the stomach of the final host the last stimulus of the
series is furnished and the adult condition attained.

In those Cestoda which have an intra-uterine development, i. e.,
forms in which a six-hooked embryo develops in the uterus,
we find the primary developmental stimulus intimately associ-
ated with the fusion of the germ plasms as in the fertilization of
most Metazoa, though in Bothriocephalus rugosns (Schauins-
land, '86) the intra-uterine development may not begin for sev-
eral weeks after the eggs have begun to pass into the uterus cavity.
In such cases the comparison between the primary " develop-
mental stimulus" and the developmental stimuli which follow is
of course not so patent, although it is none the less legitimate.
Whether what I have called a "developmental stimulus" in
these several cases shall be found eventually to be some new
condition which the egg or embryo meets or be found to be the
removal of some existing condition which has been inhibiting the
development is of no consequence here since the removal of an
inhibition may be spoken of as a stimulus, and since the impor-
tant thing is not the nature of the stimulus but the similar re-
action of the animal in each case.

In conclusion I may say that no facts not already familiar to
students of Cestode life history have been set forth in the section
of this paper just concluded, nor can it be claimed that the ap-
parently two-fold nature of fertilization has not been in recent
years more than once promulgated. The comparison between
the reaction of the oosperm to the primary developmental stim-
ulus and the reaction of the larval stages each to its special
stimulus has interested me and it has seemed to me worth while
to attempt the formulation of Cestode life history from this point
of view. I also believe that the two-fold nature of fertilization

142 W. C. CURTIS.

has not yet reached such a point of general acceptance or rejec-
tion, but that it will bear further illustration in any case in
which there are facts particularly germane to the question and with
these two points in view I have attempted the foregoing for-
mulation.

UNIVERSITY OF MISSOURI,

COLUMBIA, Mo., May i, 1903.

Vol. V. August, 1903. No. 3

BIOLOGICAL BULLETIN.

ON THE CONDITIONS GOVERNING THE PRODUC-
TION OF ARTIFICIAL PARTHENOGENESIS

IN ARBACIA.

S. J. HUNTER.

In a previous paper * it was shown that sea-water concentrated
by evaporation to a definite volume would produce partheno-
genesis in eggs of the sea-urchin, Arbacia, subjected to its
influence for a given length of time. During the continuation of
these experiments for the purpose, primarily, of observing the
morphological phenomena there has become evident a series 01
conditions necessary to a high ratio of parthenogenetic develop-
ment. These conditions are briefly -- purity of solutions, stage
of development of ovarian eggs when placed in the concentrated
sea-water, length of time the eggs are to remain in this solution,
temperature of solutions. This article is based upon observa-
tions on eighty-three experiments, fourteen of them performed
between July 25 and August 14, 1901, and sixty-nine between
July 4 and August 15, 1902. The work of both seasons
was pursued at the Marine Biological Laboratory, Woods
Holl. In every instance the eggs for each experiment were
taken from one female only. In this brief preliminary account
detailed references to individual experiments are in some cases
given, not to illustrate the condition peculiar to that experiment
alone, but rather to set forth prevailing conditions.

Solutions. — These are rendered ineffective, (i) by the presence
of foreign substances ; (2) spermatozoa ; (3) excessive number
of eggs. The eggs of Arbacia are extremely sensitive to foreign
substances liable to be introduced by the use of glassware not
thoroughly cleansed or glassware previously used as receptacles

1 Hunter, American Journal of Physiology, VI., 1901, p. 177.

H3

144 s- J- HUNTER.

for chemicals. For this reason it was found advisable to use
only new glassware.

Much time was consumed in the work of sterilization to pre-
vent contamination from spermatozoa. This, however, is essen-
tial. In the eighty-three experiments referred to sterilization was
performed in accordance with the plan mentioned in the paper
cited. If, after this treatment, there were eggs that developed
through normal fertilization, such escaped notice. In five other
experiments made to determine the necessity of this sterilization,
the sea-water was not sterilized nor was the sea-urchin carefully
washed in hydrant water. In these five experiments a few nor-
mally developing forms were noted.1

The relative proportion of eggs to concentrated sea-water is an
important factor in determining the percentage of development.
In the paper referred to mention was made of the necessity of
placing comparatively few eggs in the solution. Tabulated re-
sults of some experiments on this point make the relative value
of this condition more apparent. In two bowls containing equal
amounts of condensed sea-water there were placed in the first
bowl a greater number of eggs and in the second bowl very few.
Out of the first bowl 14 per cent, reached the swimming gastrula
stage ; out of the second bowl 87.5 per cent, reached the same
stage. Notes taken on early stages of the culture showed that
in the bowl containing the less number, segmentation began in a
greater number of eggs.

In other experiments an endeavor was made to measure ap-
proximately the number of eggs placed in a given amount of sea-
water. One of these experiments is placed in tabulated form.

No. i, i pipette full of eggs in 50 c.c. concentrated sea-water.

No. 2, 5 pipettes full of eggs in 50 c.c. concentrated sea-water.

No. 3, 10 pipettes full of eggs in looc.c. concentrated sea-water.

After two hours, transferred to sterilized sea-water, frequently
changed at first --examined twenty -seven hours later with re-
sults as follows :

!The test of purity of culture was based on the facts, (i) that cleavage in parthe-
nogenetically developing eggs of Arbacia at no time prior to blastula stage, resemble
normal processes ; (2) the shortest time in which any culture under constant observa-
tion reached the active swimming stage was nine hours and seven minutes, under
like conditions normally fertilized embryos become active in about six hours; (31
the absence of t.he perivitelline membrane.

ARTIFICIAL PARTHENOGENESIS IN ARBACIA. 145

No. I, 7 gastrulae out of 16, 43 per cent.

No. 2, 9 gastrulae out of 90, 10 per cent.

No. 3, 10 gastrulae out of 109, 10 per cent.

Examined again twenty-six hours later with the following
results :

No. i, 8 plutei and 12 gastrulae out of 27, 74 per cent.

No. 2, i pluteus out of 42, .02 per cent.

No. 3, no living forms out of 39.

This difference in the ratio of development is probably due to
the noxious effects of the undeveloping eggs in the solution.
This being the case a frequent change of the concentrated solu-
tion might raise the percentage, for while the sterilized sea-water
was changed repeatedly the concentrated water was not changed
at all during the period. The difference in results was evident
at this stage, that is, the two lots of eggs when removed from
the condensed solution showed a difference in behavior. As just
noted, the eggs from the cultures containing the smaller number
showed a larger percentage of segmentation. Cultures No. 2
and No. 3, having same ratios, gave similar percentages.

State of Development. - - Wilson x has observed that the eggs of
Toxopncitstcs which would not fertilize with spermatozoa gave
some of the best results obtained with the magnesium solution.
Delage ~ notes in Strongylocentrotus that frequently eggs which
will not develop by artificial parthenogenesis are readily fertilized
by spermatozoa. A number of observers have noted the wide
variation in the behavior of the eggs of different females. Some
eggs do not develop at all, others give large percentages of active
larvae. A case in point : An experiment with a large female, the
eggs of which by their number, color and the freedom with which
they came from the ovary — the ovaries in such cases when
placed in sea-water lose form and become a mass of eggs - - seemed
to signify that the eggs were fully matured (ootids). As further
proof, the greater part of these eggs were fertilized with sperma-
tozoa, this resulted in the normal development of all the eggs ob-
served. The remainder of the eggs were subjected to the influence

1 E. B. Wilson, "A Cytological Study of Artificial Parthenogenesis in Sea-urchin
eggs," Archivf. Entwickelun^smeth., XII., 4, 1901, p. 535-

2 Delage, Y., " Etudes Experimentales sur la Maturation Cytoplasmique Chez les
Echinodermes," Archiv d. Zool. Exper. et Generate, 3, Sen, IX., 1901, p. 300.

146 S. J. HUNTER.

of condensed sea-water for two hours and four minutes, resulting
in the appearance of many cytasters (Wilson) and the segmenta-
tion of many eggs, but no active larvae. Segmentation and de-
velopment ceased after four hours in the sterilized sea-water.
The ovaries of another female were teased in sea-water and a
small number of pale eggs were obtained. From the condition of
the ovaries, the color and the number of eggs it was evident that
the eggs were not mature (oocytes). These were placed for the
same length of time in concentrated sea- water and then transferred
to normal sea-water. More than 90 per cent, of these eggs reached
the active larval state. Eggs of females brought directly from
the bed of the ocean gave better results than those kept in the
laboratory aquarium for a time. The higher temperature of the
sea-water in the laboratory probably hastened maturation. It
became evident throughout the later series of experiments that
oocytes gave satisfactory results and ootids gave negative results.
It seems probable, therefore, that concentrated sea-water is effec-
tive in producing development in Arbacia only when its influence
is brought to bear upon the oocyte.

The interesting question naturally arises concerning the exact
stasre at which the solution is effective. If this influence causes

o

retention of the second polar body and its assumption of the role
of the spermatozoon the subject is at once brought into direct
relation with Boveri's l theory of natural parthenogenesis. In
Delasre's2 observations on the influence of carbon dioxide on

o

Astcrias he gives results to show that the moment of suscepti-
bility of the eggs lies between the time when the nuclear mem-
brane of the germinal vesicle begins to dissolve and the beginning
of the resting period of the egg nucleus ; and that the immedi-
ate cause does not concern the polar bodies but rather the
suspension of maturation for a given period. Upon resumption
the polar karyokinesis is not confined to one region of the egg,
but instead becomes general and includes the whole egg. Con-
sideration of this phase is curtailed by Delage's 3 own statement

!Th. Boveri, " Zellstudien," L, 1887, p. 73.

2 V. Delage, " Nouvelles Recherches sur la Parthenogenese Experimentale chez
Asterias glacialis," Archiv de Zool. Exper., 1902, p. 217.

3Y. Delage, "Etudes Experimentales sur la Maturation cytoplasmique chez les
Echinodermes," Archiv de Zool. Exper., 3 Ser., IX., 1901, p. 295.

ARTIFICIAL PARTHENOGENESIS IN ARBACIA. 147

that phenomena manifested in the starfish must not be assumed
to occur in the sea-urchin, and further according to the same
author ' the eggs of Strongyloccntrotus are mature before being
subjected to the solution. From this it would seem that there
is a difference in the behavior of the eggs of Arbacia and Strongy-
locentrotus. In Arbacia it did not appear that in the development
of the egg there was only one opportune moment when the con-
centrated solution was effective, but rather that ovarian eggs
placed in the concentrated solution, were influenced to maturate
and that maturation brought about in this way resulted, when the
eggs were removed to normal sea-water, in segmentation and
subsequent development. Experiments with eggs apparently
mature frequently give small percentages, one to five per cent,
of larval development. This may be accounted for by the pres-
ence of a few oocytes in the ovaries. The difficulty in the case
of Arbacia is that owing to the opacity of the eggs it is not
possible to ascertain their exact state when placed in the con-
centrated solution.

There is some evidence which probably bears upon the ques-
tion to be found in the examination of sections. Without enter-
ing into a detailed description at this time we find in iron-haema-
toxylin sections a heavily staining body in contact with the
nuclear membrane. In some cases astral rays extend out from
this dark body. In others these rays are absent. Later proph-
ases occur, such as the elongation of the nucleus with an aster
at each pole, followed by the mitotic figure in its various phases.
In other words, there appear to be processes closely resembling
normal karyokinesis. This conspicuous dark body shows its
attitude towards the nucleus in cases where the dark body has
failed to divide. In such cases, the nucleus is elongated on the
side of contact and the chromatin is aggregated on the same
side. It seems reasonable, therefore, to say that in these parthe-
nogenetic eggs there is a force whose behavior approximates that
of the spermatozoon.

Briefly, then, in the sea-urchin egg maturation takes place in
the ovary before normal oviposition.2

i Ibid., pp. 296, 301, 324.

2E. B. Wilson, "The Cell," 1900, p. 236.

148 S. J. HUNTER.

The eggs used in these experiments were not deposited natu-
rally, but were from ovaries removed from the female.

The ovaries thus taken were of two kinds : first, dark red in
color, delicate in structure ; when placed in sterilized sea-water
eggs flowed freely from them without cutting or teasing ; second,
light red in color, firm in structure ; comparatively few eggs ob-
tained even after ovaries are cut and teased.

The eggs from ovaries of the first class gave unsatisfactory re-
sults when subjected to influence of concentrated sea-water, satis-
factory results when fertilized with spermatozoa.

The eggs from ovaries of the second category have given per-
centages as high as 80 to 90, of parthenogenetic swimming forms,
when subjected to influence of concentrated sea-water for the
proper period.

Sections through ovaries, typical of this second class, reveal
large numbers of oocytes determined as such by the presence of
the prominent germinal vesicle. Sections through thirty-two
different follicles were examined. Only those showing germinal
vesicle (oocytes) or egg-nucleus (ootids) were counted. In these
thirty-two sections of follicles there were 183 oocytes and 85
ootids.1 Oocytes much smaller than normal eggs were not
counted. The percentage of forms developed parthenogenetic-
ally is thus shown to bear a direct relation to the number of
oocytes in the culture.

It seems reasonable, then, to infer that the concentrated sea-
water acts effectively upon the oocyte only. The exact nature
of this action it is hoped subsequent study will determine.

Duration. — The eggs of Arbacia, as is well known, are not suf-
ficiently transparent to permit close observation upon the activity
of the cell contents. For that reason I have been unable to note
in the egg any definite appearance which would signify the proper
moment for transference from condensed sea- water to sterilized sea-
water. In a few cases I have found wide variations in the time that
the eggs can be transferred and yet develop. The shortest time
was one hour and twenty-two minutes. The limits within which the
eggs from a given culture could be removed and yet develop were

1 These follicles were from the ovaries of one female. Of the utilized eggs from
this female fully 75 per cent, became swimming larva;.

ARTIFICIAL PARTHENOGENESIS IN ARBACIA. 149

relatively narrow. It seems that the critical moment does not
lie, as in the case of Delage's observations on Asterias, in the time
when the eggs are placed in the solution, but rather when they
are removed from the solution.

The stage of development of eggs when placed in solution evi-
dently has some bearing on the time required for development.
The differences in the states of ovarian eggs would seem to ac-
count for the differences in the time required for development, not
only for the eggs of different individuals, but as in the experiment
given below, for the eggs of the same individual.

The culture just referred to, the one in which larval develop-
ment was obtained after an hour and twenty-two minutes in the
concentrated solution, was one of a series of experiments to de-
termine the proper length of time and also the time in which eggs
of a given female will develop. This experiment also presented
the longest period of time within which eggs of the same indi-
vidual could be removed from the concentrated solution for sub-
sequent development. The eggs were placed in the concentrated
sea-water and allowed to remain one hour. Watch-glass cultures
of approximately equal number of eggs, the standard being three
pipette drops of eggs in each watch glass, were removed every
two minutes from the concentrated solution and placed in steril-
ized sea-water. The length of time that eggs remained in the
concentrated solution is given and opposite are the observations
made, beginning seven hours and twenty minutes later.

Minutes in Con-
centrated So-
lution. Notes Taken Seven to Ten Hours Afterward.

62. No segmentation.

64. A number of fragments from a few eggs that had seg-
mented and then broken.

66. The same.

68. The same.

70. Fragments more abundant but nothing in the nature of
a cluster of blastomeres.

72. The same.

74. The same.

76. Not so many fragments, a few eggs segmented into two
and three blastomeres.

I5O S. J. HUNTER.

78. Very few whole eggs, nearly all in fragments of halves
and less sizes.

80. Blastomeres remaining together, few fragments.

82. Three active well-formed blastulae (examination made
ten hours after removal from concentrated sea-water).

82—98. The eight cultures taken out during this time showed
about the same percentage of development as 82.
IOO. No segmentation in this culture nor in any of the sub-
sequent cultures.

This experiment shows a duration of sixteen minutes within
which eggs were removed and larval development ensued. This
was the widest range of the series. In many experiments a dif-
ference of five minutes on either side of the optimum moment
determined the life of the culture. In all cases, as noted by other
observers, eggs removed from the concentrated solution after a
brief period begin to segment but do not continue to develop
until they reach the swimming blastula stage. Eggs permitted
to remain too long plasmolyze when placed in sterilized sea-water.
As a result of this series of experiments the optimum period was
determined at two hours. In each case three cultures were formed
of the eggs, one of five minutes before the period, one at the period,
and the other five minutes after the two hours.

Temperature. — The most favorable temperature obviously is
the normal temperature of sea-water. Sudden changes caused
by the use of water of a different temperature for replenishing
cultures is detrimental. • Greeley1 has shown that blastulse can
be developed parthenogenetically in concentrated sea-water at
a temperature of 2°, 1 1° and at the room temperature of 23°. I
am convinced that uniform results cannot be obtained from cul-
tures kept on the laboratory table. The changes in temperature
which occur between day and night materially affect the behavior
of the eggs. For this reason towards the close of the season
the bowls containing the solutions were surrounded by running
sea-water. This insures constancy of temperature as well as
approximates the normal temperature.

'A. W. Greeley, BIOL. BULLETIN, IV., No. 3, p. 132.

ARTIFICIAL PARTHENOGENESIS IN ARBACIA. 151

SUMMARY.

1. The conditions governing the production of artificial
parthenogenesis in Arbacia by the use of sea-water concentrated
by evaporation to a definite volume, are purity of solutions, stage
in development of ovarian eggs, duration in concentrated solu-
tion, temperature of solutions.

2. The efficacy of solutions is subject to the presence of for-
eign substances, spermatozoa, relative number of eggs in a given
amount of concentrated sea-water, and temperature. Foreign
substances are excluded through extreme care in the preparation
of solutions ; spermatozoa are eliminated by raising normal sea-
water to 70 degrees, by sterilizing all instruments in the flame,
by washing thoroughly the body of the sea-urchin and the hands
of the operator for three brief periods under stream from the hy-
drant. Results are most constant at normal temperature of sea-
water. Development is obtained at room temperatures 22° to
24°. Variations in temperature of solutions materially affect the
development of the culture.

3. The concentrated solution appears to be effective in pro-
ducing development in oocytes only. By reason of the opacity
of the egg it is difficult to ascertain the exact stage or subsequent
behavior in concentrated solution.

4. The average optimum period for eggs in concentrated solu-
tion lies between one hour and fifty-five minutes and two hours
and five minutes.

•

UNIVERSITY OF KANSAS,
April 4, 1903.

HETEROGENY AND VARIATION IN SOME OF THE
COPEPODA OF LONG ISLAND.

ESTHER F. BYRNES.

In the spring of 1898, my attention was attracted to certain of
the Copepoda that occur in large numbers in the fresh-water
ponds in some of the outlying districts of Brooklyn. The
material, which contained many Cyclops, was collected soon after
the ice had disappeared from the surface of these shallow pools
and even at this early season most of the Cyclops were large and
carried eggs in all stages of development.

I isolated individuals with eggs, and subsequently observed
numerous color-changes, which accompanied the rapid growth
and extrusion of eggs into the egg-sacs. A single instance will
suffice to show the rapidity of these changes, and the fertility of
the individuals. On the iQth of April, 1898, a Cyclops, carrying
dark blue eggs, was isolated. On the 2Oth dark bluish ova could
be seen through the transparent body-wall, making the body ap-
pear dark, while the dark eggs in the egg-sacs had developed into
embryos of a reddish tint. On the 22d the copepod carried dark
eggs again, and the body was again almost colorless, with a faint
streak on either side, still marking the position of the ovaries.
On the 23d it remained unchanged. On the 24th the body was
again dark but no eggs were attached. On the 25th the dark
eggs were carried in appended sacs and the body was again col-
orless. On the 26th the dark eggs became detached. On the
2 /th the body again appeared dark. There is no record in my
notes for the next two days, but when I again looked at the
copepod the body was colorless. While it carried no egg-sacs,
the ova must have been discharged since the last record on the
2/th instant.

I attempted to identify the form, which agreed with C. parcns
(Herrick), in most of the points that are regarded as species-
characteristics but it differed from C. parcns in the number of
its antennal segments.

152

HETEROGENY AND VARIATION IN COPEPODA. 153

The chief morphological features by which species of Cyclops
are recognized are the following :

1. The number of joints in the antennae.

2. The number of joints in the rami of the four swimming feet.

3. The armature of the swimming feet.

4. The number of joints in the fifth foot, which is rudimentary.

5. The shape and armature of the segments of the fifth foot.

6. The structure of the abdomen with the caudal stylets and
the armature of the caudal stylets.

7. The shape of the receptaculum seminis.

8. The armature of the maxillipeds.

9. The relation between the length of the antennas and the
cephalothorax.

The characteristics of C. parcus are as follows :

1. Seventeen-jointed antennae.

2. Three-jointed rami in the swimming feet.

3. Armature of the last segment of the swimming feet.

FIRST FOOT. SECOND FOOT.

Outer Rainus. Inner Ramus. Outer Ramus. Inner Ramus.

2 outer spines. I outer seta. 2 outer spines. I outer seta.

2 apical setae. I apical spine. I apical spine. I apical spine.

I apical seta. I apical seta. I apical seta.

2 inner setae. 3 inner setae. 3 inner setse. 3 inner seta;.

THIRD FOOT. FOURTH FOOT.

Outer Ramus. Inner Ramus. Outer Ravins, Inner Rainus.

Like second. 2 outer spines. I outer seta.

I apical spine.

I apical seta. 2 apical spines (equal).

3 inner setae. 2 inner spines.

4. Two-jointed fifth foot.

5. The basal joint short and broad with a single seta on the
outer margin.

A long, cylindrical, distal segment with a blunt, inner spine
and a long, outer seta, but very slightly plumose.

6. The abdomen is composed of segments, the first of which
is as long as the remaining segments combined. The caudal
stylets are long.

7. The receptaculum seminis is broadly oval.

8. There are four hairs on the distal segment of the larger
branch of the maxillipeds. The second segment has a large, im-
movable dactyl with a row of teeth along the edge, and with a

154 ESTHER F. BYRNES.

small hair at its base. Attached to the immovable dactyl is a
small, movable one.

9. The antennae are about the length of the cephalothorax.

The points in which the Long Island Cyclops that I have
studied differs from C. parcus are : In the number of antennal
segments, there being 13 instead of 17, and in the occasional
variation in the armature of the outer ramus of the fourth foot,
there being but one outer spine and one seta, where C. paints
has typically two spines ; as well as in the armature of the ter-
minal joint of the large ramus of the maxilliped, where two
small hairs replace one large one ; also in the armature of the
distal joint of the fifth foot, which carries an outer hair, in place
of the unserrated spine which is present in the form with seven-
teen joints in the antennae.

As the correlated characteristics of species occur with great
regularity in the Cyclops, and as the form under consideration
seemed, both on account of its relatively large size and its fertil-
ity, to be a mature form, I searched for similar individuals but
for a long time failed to find them.

In the summer of 1899, I had the opportunity of collecting
large numbers of Cyclops at Cold Spring Harbor, L. I., where sev-
eral fresh-water ponds afford excellent opportunities for the study
of a variety of species. Though I have worked over some of this
material with great care, I have never met with a single instance
of a thirteen-jointed antenna.

In March of the present year, 1903, I again met with a num-
ber of Cyclops having thirteen-jointed antennae. This material
was collected in one of the large, shallow, fresh-water ponds at
Jamaica, Long Island. The copepods were found in great num-
bers hidden beneath the fallen leaves along the edges of the
pond. Again I noticed marked color-changes incident to the
development and laying of ova. Some were red in the body and
carried blue eggs in their paired sacs, while many were dark in
color and carried about the partly developed reddish embryos.

Associated with these larger forms were smaller Cyclops, often
without eggs, and emerald green to the naked eye, owing to the
numbers of green protozoa that had attached themselves to the
cuticle and almost concealed the host. The larger Cyclops with

HETEROGENY AND VARIATION IN COPEPODA. 155

the pink bodies and the blue eggs, or vice versa, were compara-
tively free from the one-celled forms. I believe this fact is
important as pointing to the strong probability of a recent moult.
Further study revealed the fact that the larger forms had invari-
ably seventeen segments in the antennae and that they agreed in
all essential details with the species known as C. parcus (Her-
rick).

After formulating data gathered from the study of species-
characters in many different individuals from the same locality, I
was able to clearly distinguish three groups, in all of which, all
the leading species-characteristics of C. parcus (Herrick) were
combined with a varying number of segments in the antennae,
which, however, all belonged to the same type (Fig. i).

Group I. comprised individuals with thirteen antennal seg-
ments.

Group II. comprised individuals with fourteen antennal seg-
ments.

Group III. comprised individuals with seventeen antennal
ments.

Nearly all of the Cyclops referred to as covered by protozoa
and hence appearing green, belong to Group II. or are inter-
mediate between Groups I. and II., and are characterized by an-
tennae with fourteen segments either fully formed, or in process
of forming. I have studied no less than ten individuals which
show clearly that the fourteen-jointed antenna is derived from the
thirteen-jointed one, by the division of the tenth segment — the
fourth from the distal end of the antenna — which is divided
almost equally into halves by a transverse partition.

It is always the tenth segment which is dividing at this stage,
and in all cases recorded, when the two antennae are not in the
same stage of division, it is without exception the left that is in
advance of the right, in which division can still be seen in prog-
ress, as in Fig. i, B.

I know of no explanation of the retarded division in the right
antenna, and it may be a mere coincidence that all of my ob-
servations agree on this point. One Cyclops in which the four-
teen segments were perfectly formed in both antennae, proved
particularly interesting, for I believe it furnishes positive proof

FIG. I. Shows the antenna and fifth foot of a cyclops with thirteen-jointed an-
tennae, A, A', Group I. The antenna and fifth foot of a cyclops with fourteen
jointed antennre, B, B/, Group II. The antenna and fifth foot of a cyclops with
seventeen-jointed antennae, C, C/, Group III. D shows the abdomen, the recepta-
culum seminis and the caudal stylets characteristic of all the forms with thirteen,
fourteen or seventeen-jointed antennre. E shows the large ramus of the maxilliped
characteristic of the three forms. B also shows the tenth antennal segment in the act
of dividing, thus giving rise to the fourteen-jointed antenna.

As compared with the length of the cephalothorax all the antennae A, B and C
shown in Fig. I are relatively long, extending to the first segment irrespective of the
number of segments they contain.

HETEROGENY AND VARIATION IN COPEPODA. 157

that this apparently stable individual with the fourteen-jointed
antennae represents but a temporary condition in the develop-
ment of a form with seventeen antennal segments. In the case
referred to, the long eighth joint, that is characterized by three
rather widely separated setae, showed distinct, transverse lines
across the segment at the level of each of the two lateral setae.
Half way up the remaining section, a slight indentation in the
cuticle marked the position of the wall that completes the sepa-
ration of this long segment into four small ones of almost equal
size.

The breaking up of the eighth segment in the manner indi-
cated by these markings gives to the seventeen -jointed antenna
a short eighth segment with a single distal seta ; a short ninth

o o o

segment with a distal seta ; a short tenth segment without any
armature, and a short eleventh segment with one distal seta.
These are precisely the conditions which prevail in the seventeen-
jointed antennas.

In his report on " The Entomostraca of Minnesota," Herrick
describes a Cyclops strikingly like the one from Jamaica, Long
Island, ivitJi fourteen-jointed antenna, three-jointed raini ivith the
armature of the last joints like that given for C. parcus, and u'ith
a two-jointed fifth foot " with the armature like C. strenmis , which
also resembles C. piilc/icHits." The stylets are very long. These
correlated peculiarities of structure are recognized as constitut-
ing a distinct species known as Cyclops insignis (Claus). Her-
rick mentions that "in a previous edition it was suggested that
this is but an atavistic form of C. pnlchelhts - - C. strenmis. " If
C. strenmis is to be regarded as practically the same form as C.
abyssontm, as Schmeil suggests, the Long Island form can hardly
be brought into relation with it, for the armature of the swim-
ming feet, which is remarkably constant in forms of equal size,
differs markedly in the two cases. Schmeil, however, seems to
attach little importance to this fact.

That the Long Island form with the fourteen-jointed antennae
represents a transitional stage in the development of a seventeen-
jointed form, there can be little doubt, though the determining
of the species in the terms of an old and confused classification
is by no means an easy matter. The length of the caudal stylets

158 ESTHER F. BYRNES.

is relatively greater in the Cyclops with the fourteen-jointed an-
tenna than in the adult C. parcus, though in C. parcus the stylets
are characteristically long. Another slight difference is seen in
the presence of a hair in the form with the fourteen-jointed anten-
nse, in place of a small spine on the inner angle of the distal
joint of the fifth foot of the seventeen-jointed form. Compare A' ,
B', C, Fig. i. Moreover, the distal joint of the so-called C. insignis
is strikingly long, longer than the corresponding joint in C. parcus.
The difference between the two forms seems to be almost entirely
one of proportion and size, the insignis-\\V.& form being slightly
smaller than C. parcus, and often with fewer or no eggs.

In favor of the existence of a separate species for those forms
with fourteen-jointed antennae, and against the suggestion made
by Herrick that C. insignis represents a transitional stage in
development, Schmeil urged the occurrence of the Cyclops in
large numbers, and its relatively large size, both of which obser-
vations I can confirm. I can not, however, agree with Schmeil's
interpretation ; although the form is abundant and moderately
large, it is often, though not always, without eggs either in the
body or attached, when older forms associated with it are re-
markably prolific. Moreover, if studied at the right stage, the
form with fourteen segments in the antennae gives frequent signs
of being still in a period of growth characterized by morphologi-
cal changes. The fact that the smaller form is densely covered
by foreign growths indicates that it has not very recently moulted.
In this connection it may not be irrelevant to allude to a few
observations made on isolated copepods.

I separated a number of Cyclops in a small watch crystal.
All were about the same size, some green to the naked eye,
some dark, and others carrying eggs. A few days later my
attention was drawn to a bright red Cyclops with a perfectly
clean cuticle. It had seventeen segments in the antennas, and
from the absence of protozoa on its surface it must have moulted
quite recently. I then looked about in the dish for cast-off
skins and found one still well covered with protozoa and having
fourteen-jointed antennas.1

1 Inasmuch as there were other individuals in the watch crystal, this is by no
means conclusive proof that the seventeen-jointed form had shed the fourteen-jointed
skin, but I could find no other explanation of its presence in the dish and I offer the
fact for whatever it is worth.

HETEROGENY AND VARIATION IN COPEPODA. 159

I then set aside six Cyclops with fourteen-jointed antennae,
giving them clean hydrant water containing but little food and
some fresh-water plants. At the time of -their separation two
had fourteen segments only in the left antenna, while the right
antenna of each contained a dividing segment, the tenth from
the base of the antenna, or the fourth from the distal end. Two
weeks later the division of the segment was still incomplete,
showing that in this case at least, the formation of partition walls
is not very rapid. The bodies looked lighter and clearer than
before, and I examined them again to see if any changes had
taken place, but none had occurred.

In his explanatory notes accompanying Plate XXXIV.1 which
shows the species-characteristics of C. parcus (Herrick), Herrick
shows "caudal stylets of an elongate form," in Fig. 3, with
which my own drawings agree perfectly. It is quite possible
that the elongated distal segment of the fifth foot may be a mere
variation correlated with the elongation of the caudal stylets in
Herrick's ' elongated form ' of C. parats which he suggests " is
to be regarded as a post-imago."

A single characteristic which Herrick describes for C. pul-
chcllus, but of which no mention is made in the characterization
of C. parcns, to my knowledge, is the presence of serrations on
the distal margins (ventral and lateral) of the last abdominal
segment, while the remaining margins of the abdominal segments
are free from such markings. All of the individuals of the three
groups — i. c., of the thirteen-, the fourteen- and the seventeen-
jointed antennae — agree with C. pulclicllus in having these ser-
rations, while Groups I. and II. also agree in having " two rather
long setae" which are not at all or only slightly plumose on the
terminal segment of the fifth foot. But they all differ from C.
pulclicllus in not having the basal joint of the fifth foot longer
than \vide ; the basal joint is unequivocally wider than it is long,
and in this respect agrees with C. parcus.

Although the armature of the appendages is very constant in
the Cyclopidae, it is quite common to meet with similarly placed
spines and setae of different lengths. A notable instance of this

1 " Copepoda, Cladocera and Ostracoda of Minnesota," Zoological Series, II.,
1895, of the Geological and Natural History Survey of Minnesota.

l6o ESTHER F. BYRNES.

occurs in the armature of the third joint of the large ramus of
the maxillipeds of the fourteen -jointed and seventeen-jointed
forms. The armature usually consists of three large hairs and
two very small ones growing close together at the base of one of
the large hairs (Fig i, E\ In the fourteen-jointed forms, these
two small hairs are strikingly shorter than they are in the seven-
teen-jointed form. With this single exception, the maxillipeds
are precisely alike in both groups.

I am aware that Herrick describes the armature of the terminal
segment of the larger branch of the maxilliped of C. parcus, as
consisting of four hairs. I have found an instance in which four
large hairs of almost uniform size occur, but a more frequent
condition in the Long Island Cyclops is seen in those instances
which show three large hairs and two short ones, in place of the
four hairs of Herrick (Fig. I, R}.

Among the many Cyclops I have studied, I have seen but one
with eighteen segments in the antennae. In this case the eigh-
teenth segment is derived from the seventh segment, by transverse
division, at the level of the seta. In both right and left antennae
the division is incomplete, extending but half way across the
segment.

I have studied this Cyclops with great care, and in every
detail of structure, it agrees perfectly with the forms associated
with it in showing the chief species-characteristics of C. parcns.

I have repeatedly made written records of body-segments and
appendages showing the complete armatures, and have made
many outline drawings of those parts that are correlated in the
determination of species, and I believe no room for doubt remains
that the Cyclops with thirteen and with fourteen antennal seg-
ments, as well as the form with eighteen segments, are all to be
referred to the type with seventeen segments in the antennae.
Those having thirteen and fourteen segments, known as C.
insignis, though very abundant forms and though sexually mature,
do not represent a group of sufficient permanency to warrant
us in regarding them as representatives of a distinct species.
They are rather to be considered as transitory stages which,
though capable of producing young, have not as yet attained
their maximum growth, or their highest degree of complexity.

HETEROGENY AND VARIATION IN COPEPODA. l6l

The Cyclops with the eighteen-jointed antennae agrees with
Clans' description of Cyclops clongatus, so far as Herrick has
quoted Claus. Nevertheless, its close agreement in all species -
characteristics with C. parcits, with which it was found, and the
very exceptional occurrence of so many antennal segments, make
it highly probable that we are dealing here with a case of vari-
ation rather than with a species-character.

The Cyclops from Cold Spring Harbor, Long Island, were
collected at the surface of a very shallow pond along a road-side
near the laboratory of the Brooklyn Institute. The pond was
choked with water-plants and a scum of duck-weed floated on
the surface. From the extreme shallowness of the pond, any
life there must have been exposed to rapidly changing conditions.
The material collected in this pond was all taken from one
locality within a radius of a few feet, where the copepods were
in among the duck-weed.

I attempted some statistical studies in variation on these forms,
but the work was soon interrupted by the comparatively small
number of individuals belonging to the same species, or to species
closely enough related to warrant any use of them in obtaining
data. Most of the forms I have been wholly unable to identify,
for while they agree with well known species in certain character-
istics, they differ from them in others which are apparently no
less important.

Certain combinations of characters occur so frequently, that,
in the absence of transitional forms, one is often tempted to be-
lieve that in the bewildering array of forms before him, he is
dealing with new variations, of which it is almost impossible to
say whether they have a species value or not. Whether the
forms met with illustrate paedogenesis, or whether the season
was connected in any way with the morphological aspect of the
copepods, I cannot say, not having been able to collect from
this vicinity at any other season. But I have not seen any tran-
sitional stages in an individual such as would warrant the linking
of it with any well known species.

One Cyclops frequently met with, combines the* following
characteristics: Antenna nine-jointed; rauu of siviinining feet
two-jointed ; rudimentary fifth foot one-jointed.

162

ESTHER F. BYRNES.

ARMATURE OF THE SWIMMING FEET.

First Foot.

Outer Ramus. Inner Ranuts.

3 outer spines. I outer seta.
I apical spine.
I apical seta.

4 inner setae.

I apical spine.
I apical seta.
5 inner setae.

Second Foot.

Outer Ramus. Inner Ramus.

3 outer spines. I outer seta.
I apical spine.
I apical seta.

4 inner setae.

I apical spine.
I apical seta.
5 inner setre.

Third Foot.

Fourth Foot.

Older Ramus.

3 outer spines.
I apical spine.
I apical seta.

4 inner seUe.

Inner Ramus.
I outer seta.
I apical spine.
I apical seta.
4 inner setie.

Outer Ramus.

3 outer spines.
I apical spine.
I apical seta.

4 inner seize.

Inner Ramus.

1 outer seta.

2 apical spines.

3 inner setae.

The antenna and fifth foot of this form are seen in Fig. 2.
The pravalence of the form alone is not sufficient reason for

FlG. 2. Shows the antenna and the fifth foot of a Cyclops with nine antennal seg-
ments. The fifth foot is two jointed and resembles the fifth foot of the Cyclops with
the ten-jointed antennae.

regarding it as a distinct species, and the probability is that we
are here dealing with a transitional stage in the development of
a species with a greater number of antennal segments, as seen in
the case of the fourteen-jointed form, for no species in its mature
condition is recognized as having nine antennal segments, while
the fact that the mini arc two-jointed and the number of seta on
the last joint of the inner minus is exceptionally large, suggests
that the rami may subsequently acquire a third joint. Moreover,
the armature of the feet is strikingly like the armature of another
Cyclops having ten antennal segments.

This second form which occurs frequently in the same locality,

HETEROGENV AND VARIATION IN COPEPODA.

163

combines the following characteristics : Antennce tcn-jointcd ;
raini of swimming feet t-ivo-jointcd ; rudimentary fiftli foot two-
jointed.

ARMATURE OF THE SWIMMING FEET.

3 outer spines.

2 apical spines.

3 inner setae.

First Foot.

Outer Ramus. Inner Rattnts.

I outer seta.
I apical spine.
I apical seta.
5 inner setae.

Third Foot.
Outer Raimis. Inner Ramus.

3 outer spines. I outer seta.

I apical spine. I apical spine.

I apical seta. I apical seta.

4 inner setae. 4 inner sete.

Second Foot.
Outer Ramus. Inner Ramus.

3 outer spines.
I apical spine.
I apical seta.

4 inner setae.

i outer seta.
I apical spine.
I apical seta.
5 inner setae.

Fourth Foot.
Outer Ramus. Inner Ramus.

3 outer spines. I outer seta.

I apical spine. I apical spine.
I apical seta. I apical seta.

4 inner setae. 3 inner setce.

FIG. 3. The antennae, the abdomen and caudal stylets and the two types of fifth
foot correlated with the lo-jointed antenna;. C shows the fifth foot correlated also
with the 9-jointed antennae, while D shows the fifth foot correlated with the II-
jointed antennas. B represents the type of abdomen and stylets correlated with 10
antennal segments irrespective of the form of the fifth foot.

Wherever these forms with the nine- and ten-jointed antennae
occur they show the same striking similarity in the armature of
the swimming feet. The nine-jointed forms are perfectly constant

164

ESTHER F. BYRNES.

throughout tJie group, but the ten-jointed forms vary considerably
within the group, occasionally combining three-jointed rami with
a two-jointed fifth foot, and occasionally two-jointed rami with a
one-jointed fifth foot.

According to Herrick's classification of the Cyclopidae, there
is but one species having ten-jointed antennae, /. c\, C. pJialcratus,
which may combine either ten- or eleven-, usually eleven-jointed,
antennae ivith three-jointed rami in the shimming feet, and with a
one-jointed fifth foot. I have found this combination in a single
case, and the antennae contained each eleven segments. Herrick
gives only the formula for the fourth foot of C. plialcratus, with
which the above form also agrees. The entire armature of the
terminal joints of the four swimming feet in the Cold Spring
Harbor form is shown below.

First Foot.

Outer Ramus. Inner Ramus.

3 outer spines. I outer seta.

I apical spine. I apical spine.

I apical seta. I apical seta.

3 inner setae. 3 inner setae.

Thiid Foot.

Outer Ratnns. Inner Ramus.

3 outer spines. I outer seta.

I apical spine. I apical spine.

I apical seta. I apical spine.

Second Foot.
Outer Ramus. Inner Ramus.

3 outer spines. I outer seta.

I apical spine. I apical spine.

I apical seta. I apical seta.

4 inner setae. 3 inner setae.

Fourth Foot.
Outer Ramus. Inner Ramus

1 outer seta.

2 apical spines.

4 inner setae.

3 inner setae.

2 outer spines.
I apical spine.
I apical seta.
4 inner setae.

2 inner setse.

The length of the antenna in C. phaleratus as compared with
the cephalothorax is short, whereas in the Cold Spring Harbor
form the antennae are relatively long, extending nearly to the
second thoracic segment. Moreover, in a single instance the
long second joint of the antenna showed a light, transverse band
near its proximal margin, suggesting the characteristically short
second segment of the eleven-jointed antenna.

The chief characteristics of the cy clops with the eleven-jointed
antenna are three-jointed rami in the swimming feet combined with
a two-jointed fifth foot (Fig. 4).

Herrick recognizes three species having eleven antennal seg-
ments ; one of these is a European form of marked peculiarity ;
a second is C. diaphranus, whose species-characteristics are
eleven-jointed antennce, two-jointed rami in the swimming feet, and
a one-jointed fifth foot, witJi a long seta and one short spine.

HETEROGENY AND VARIATION IN COPEPODA.

i6S

I have not found a single Cyclops combining these characters.
The eleven-jointed antennae are, with one exception, so far as
my studies show, always correlated with tlire ^-jointed rami in the

FIG. 4. Represents an eleven-jointed antenna, B, correlated with a two-jointed
fifth foot B' and short caudal stylets with very long, plumose setae, A. C and Cf
represent an eleven-jointed antenna and a correlated one-jointed fifth foot with the
same abdomen and stylets as are seen in the form with the eleven-jointed antenna;
and the two-jointed fifth foot.

swimming feet, and the armature of these forms is precisely like
that of C.phaleratus, whether the fifth foot be one-jointed or two-
jointed. The third species having eleven-jointed antennas which

1 66

ESTHER F. BYRNES.

Herrick recognizes, also combines a one-jointed fifth foot with
two-jointed rami. It is known as C. affinis and is like C. pliale-
ratits, " which it closely resembles."

A fourth and a last type to which I shall refer, is seen in a not
infrequently occurring form which combines twelve-jointed antenna:
with three-jointed rami in the swimming feet and a two-jointed
fifth foot (Fig. 5).

Herrick recognizes three species as having these characteris-
tics, namely : C. capillatus and C. crassicaudis, both European

FIG. 5- Shows a twelve-jointed antenna which is relatively very long as compared
with the cephalothorax, notwithstanding the relatively small number of antennal seg-
ments present.

forms, and C. varicans, an American form. The two former are
described as Scandinavian forms only. Of the third species C.
varicans, Herrick says that it is " the American species most
nearly resembling the European form with twelve antennal seg-
ments and a two-jointed fifth foot." " Unhappily," Herrick also
remarks, " this species was taken but once." On Plate XXX.1
Herrick figures the first foot of C. varicans, which he pictures as
having two-jointed rami in the swimming feet. Herrick explains
that the last joint is homologous to two fused segments, and that
the separation might take place " at the next moult." The form
I have studied shows the armature when the rami have reached
1 " Copepoda, Cladocera and Ostracoda of Minnesota."

HETEROGENY AND VARIATION IN COPEPODA. 1 67

the three-jointed condition, and the reduction in the number of
spines and setae in the armature of the fourth foot might seem to
bear out Herrick's suggestion.

C. VARICANS. COLD SPRING HARBOR CYCLOPS.

Fourth Foot. Fourth Foot.

Outer Ramus. Inner Rannis. Outer Ratnus. Inner Ratmts.

3 outer spines. I outer seta. 2 outer spines. I outer seta.

I apical spine. I apical spine. I apical spine. 2 apical spines.

I apical seta. I apical seta. I apical seta.

4 inner setse. 4 inner setee. 4 inner setce. 2 inner setae.

The armature of the Long Island form suggests C. plialeratus,
though in the first foot it is not identical.

I have found three of these forms among a relatively small
number of individuals and they agree very closely with one
another, the only difference being in a slight variation in the
armature of the swimming feet, a spine occasionally appearing in
place of a seta.

Supposing that these individuals represent C. varicans, the
Cold Spring Harbor form is very evidently in a later stage of
development than the individual figured by Herrick. Any
appeal to relative ages as an explanation of differences, requires
the supposition that some of the segments of the feet have an
adult armature while other segments have not. But there is no
reason for supposing that the number of spines and setae in the
fourth foot is incident to the breaking up of the rami into three
segments instead of two, for the armature of the first foot is not
reduced by the presence of the additional joint in the rami.

SUMMARY.

The Long Island Cyclops (C. insignisf), having fourteen-
jointed antennas, three-jointed rami in the swimming feet, with
two-jointed fifth feet and elongate caudal stylets, is a transitional
stage in the development of a seventeen-jointed form C. parcus
(Herrick ?). The eighteen-jointed antenna is derived from the
seventeen-jointed form by division of the seventh segment.

Out of fifteen individuals taken at random, none of whose
antennal segments exceed twelve, five precisely similar individ-
uals constitute a group having nine antennal segments, two-jointed
rami and two-jointed fifth feet.

1 68 ESTHER F. BYRNES.

Four individuals constitute a second group having typically
ten antennal segments, two-jointed rami, and tivo-jointed fifth feet.
Two of these individuals show marked variation, one in having
three-jointed rami in the swimming feet, the other in having a
one -jointed fifth foot.

Four individuals constitute a third group, characterized by
eleven-jointed antenna, three-jointed rami, and tzvo-jointed fifth
feet. One member of this group has a one-jointed fifth foot, and
this is the only individual out of the thirteen that can be given any
place among species, i. e., C. phaleratns, as combining well recog-
nized species-characters.

Three individuals constituting a fourth group combine the fol-
lowing characteristics : twelve-jointed antenncc, three-jointed rami,
and two-jointed fifth feet. These forms suggest C. varicans,
with which they have much in common, but from which they
differ considerably in detail.

Some facts point to the probability that the Cold Spring Har-
bor forms with the ten-jointed antennae are morphologically unde-
veloped. Especially does the variation within the group consist-
ing of but few individuals point to the instability of these forms-

What the true nature of these correlated peculiarities in Cy-
clops may be, can only be determined by following the life his-
tory of each individual. The relatively large size of these forms,
and the frequency with which they occur, as well as the con-
stancy of the correlated characteristics, suggest on first acquain-
tance with the Cyclopidae, that they represent distinct species,
but a fuller acquaintance warns us to look further for an expla-
nation of these most perplexing variations which are doubtless
largely due to the acquiring of sexual maturity while the mor-
phological changes in the body are still incomplete, and to the
varying external conditions to which they are subjected.

BROOKLYN, NEW YORK,
March 30, 1903.

ON THE OCCURRENCE AMONG ECHINODERMS

OF LARVAE WITH CILIA ARRANGED

IN TRANSVERSE RINGS, WITH A

SUGGESTION AS TO THEIR

SIGNIFICANCE.

CASWELL GRAVE.

In this paper a short account is given of some observations
made at the laboratories of the United States Fish Commission
at Woods Hole and Beaufort on the larvae of various echino-
derms. The attempt is also made to show that these observa-
tions, taken together with those made by other students of the
group, have a direct bearing upon one phase of the problem of
the early ancestry of the echinoderms.

It would be quite impossible to give an intelligible discussion
of the bearing these observations are interpreted to have upon this
subject without first recalling the hypotheses which have been
put forward by other students of the group to account for its
origin and present organization.

The hypothesis which now seems to have the most general
acceptance is not the work of any one mind but represents the
work of many. It would be difficult, therefore, in giving a hasty
review of its most important points, to credit each of its authors
with just his contribution, so I shall make only such comments in
passing as will serve to explain the changes and additions which
seem to me to be warranted.

OBSERVATIONS.

Holothurians.

The barrel-shaped pupae of Holothurians have been long
known, having been described by Muller,1 Semon 2 and others.
They arise in each case by the breaking up and rearrangement of

1 J. Miiller, " Abhandlungen iiber die Larven und Metamorphose der Echinoder-
men," Abl. Kgl. Akad. Wiss. Berlin.

2R. Semon, " Die Entwicklung der Synapta digitata, und die Stammesgeschichte
der Echinodermen," Jena Zeitschr., Vol. XXII., 1888.

169

170

CASWELL GRAVE.

the ciliated bands of the fully formed auricularian larvae at the
time when the metamorphosis into the adult form is about to take
place. Semon's figures of the auricularian larva and the pupa

M

-— H

FIG. I. Auricularian larva of Synapta digitata. After Semon. M, mouth ; ff,
hydroccele ; A, anus. The ciliated bands stippled.

of Synapta digitata are reproduced in outline in Figs. I and 2.
During the pupal stage the mouth shifts from a ventral to a ter-
minal position and the tentacles and tube feet first become func-

FIG. 2. Pupa of Synapta digitata. After Semon. A, anus; T, tentacles; I,
2, 3, 4 and 5, ciliated rings.

tional. The ciliated rings of the pupa are five in number and are
arranged transversely to its long axis.

SIGNIFICANCE OF CERTAIN LARV/E OF ECHINODERMS. 17!

Selenka1 has studied and figured the larva of Cucumaria
doliolmn which, although totally unlike an auricularian larva, can
be well compared with a pupa. It is an elongated free swimming
creature with four, sometimes five, transversely arranged ciliated
rings, in addition to which, at the anterior end, there is a ciliated
field. This ciliated field is one of the first of the larval structures
to disappear as development progresses. In Selenka's figure of
this larva, reproduced in outline in Fig. 3, five tentacles and two

F---

-4

--- 5

FIG. 3. Larva of Cucumaria doliolmn. After Selenka. F, tube feet ; T, ten-
tacles ; 2, 3, 4 and 5, ciliated rings.

tube feet are shown to be developed and the rotation of the
mouth and tentacles to the terminal position has begun. The
eggs of C. doliolmn are quite large and well supplied with yolk,

1 E. Selenka, " Zur Entwicklung der Holothurien," Zeit. f. wiss. Zoo!., Vol.
XXVII., 1876.

1/2

CASWELL GRAVE.

thus differing widely from the small transparent eggs of Synapta
digitata. The efficient locomotor and feeding apparatuses with
which the larva' of the latter species is provided, enabling it to
care for itself, are not needed by the larva of Cucumaria doliolum
for whose care provision has already been made. The larva of
Cucumaria can, as it were, give its whole attention to the pro-
duction of a creature with the structure of the adult while the
larva of Synapta must make this secondary to food getting.

Crinoids.

In Antedon rosacea, the only species of crinoid the develop-
ment of which has been studied, the eggs are supplied with con-
siderable yolk and for a time the developing larvse are brooded.

X"

FlG. 4. Larva of Antedon rosacea. After Seeliger. Internal organs shown in
posterior end. I, 2, 3, 4 and 5, ciliated rings.

The free-swimming period is of short duration and the develop-
ment is more or less direct. The larva is elongated and cylin-
drical and is encircled by five transverse ciliated rings. An

SIGNIFICANCE OF CERTAIN LARVAE OF ECHINODERMS. 1/3

apical tuft of longer cilia is also present. Seeliger's ' figure of it
is reproduced in outline in Fig. 4. No pore canal is developed
at this stage but the point on the hydrocoele at which it will ap-
pear later, I have indicated by a small x.

Ophiurids.

For a long time the larvae mentioned above were the only ob-
served cases in which the ciliated bands are arranged in trans-
verse rings, and they were considered to have no special signifi-
cance. Since 1899, however, I have found three other cases

^"""••^rviTr ••«"•- —

FIG. 5. Ventral view of the young larva of Ophiiirn brevispina. Original. 2,
3, 4 and 5, ciliated rings.

which exhibit the same peculiarity and which represent two other
classes of echinoderms.

The larva of Ophinra brcvispina, which I2 described in 1899,
is well supplied with yolk and very early in its development it
sinks to the bottom and clings to grass blades where it under-
goes its late larval stages and final metamorphosis. It is a larva
without arms or processes of any kind and no skeletal rods such

1O. Seeliger, " Studien zur Entwicklungsgeschichte der Crinoiden," Zool. Jahrb.,
Bd. VI., 1892.

2Caswell Grave, " Ophiura brevispina," Mem. Natl. Acad. Sci., 1900.

1/4 CASWELL GRAVE.

as are found in ophiuran plutei are developed, although at one
time I mistook the beginnings of the skeletal plates of the adult
for such. The anterior end of the larva is produced into a long
preoral lobe about which two ciliated rings are developed. The
posterior end is enlarged and contains the various internal struc-
tures of the larva and developing ophiurid. The mouth is ven-
tral and interrupts the third ciliated ring of the larva (numbered
4 in Fig. 5). The fourth ring (5) surrounds the posterior end.
The dorsal pore is situated at the point indicated by the small
x between ciliated rings 3 and 4. As development progresses
the preoral lobe diminishes in size until finally it is entirely ab-

3

4

FIG. 6. Older larva of Ophiura brevispina. Original. The change in position
which takes place in the ciliated ring (5) is shown.

sorbed. During late larval life a change in the arrangement of
one of the ciliated rings also takes place. The ring numbered 5
becomes interrupted on the ventral side and takes on a more
definite relation to the mouth (see Fig. 6).

In 1900 I found a second ophiuran larva at Beaufort which, in
its metamorphosis from the pluteus to the radial form, showed the

SIGNIFICANCE OF CERTAIN LARV.E OF ECHINODERMS. 1/5

same tendency to rearrange the ciliated bands into transverse cil-
iated rings which is found among the holothurians. The outline
of the pluteus is shown in Fig. 7. When the developing ophiurid
has become quite large and the tissues of the pluteus are being
absorbed, the ciliated bands of certain of the arms become applied
to the disc in a quite definite manner, viz., about the madreporic
interradius which had an anterior position in the larva, a com-
plete ring is formed ; an interrupted ring is laid down between
rays 5 and 4 on one side and I and 2 on the other. A third
ring crosses the base of ray 3. Not until I had examined a

H

FlG. 7. Ophiuran pluteus (sp.? ) from the "tow" at Beaufort.
mouth. H, hydroccele. Ciliated bands stippled.

A, anus; JIS,

number of these metamorphosing plutei was I satisfied that this
arrangement of the ciliated areas was not accidental but in all
cases examined (a dozen or more) the arrangement was practi-
cally the same as that shown in Fig. 8.

Echinoids.

During the summers of 1900, 1901 and 1902 I succeeded in
rearing large broods of the larvae of Mellita tcstndinata from the
fertilized egg to the form in which the adult structure is attained.
The larva is a typical highly specialized pluteus as will be seen
from the outline of Fig. 9. The just metamorphosed Mcllitas all
showed three parallel transverse ciliated rings ; the middle one
of which is interrupted by the mouth (see Fig. 10). The func-
tion of these ciliated rings in the young Metlitas is probably to

176 CAS WELL GRAVE.

assist them in feeding until the tube feet have grown sufficiently
to assume the function.

THE HYPOTHETICAL BILATERAL ANCESTOR.
Although numerous papers have been written on the subject
of the phylogeny of the echinoderms there are but few which
retain their vitality at the present time. In these, notwithstand-

-P.A.P.

-V

FIG. 8. Outline of the young ophiuran which metamorphoses from the pluteus
shcfwn in Fig. 7. Original. I., II., III.. IV. and V., arms of the ophiuran;
3, 4 and 5, ciliated rings. P. PL, remnant of the posterior end of the pluteus. P.
A. P., long posterior arms of the pluteus which are never absorbed but are finally
dropped. P., madreporite.

ing the fact that many differences in detail exist, there is a very
great similarity in the views set forth and I may state in this
connection that the facts of this paper and many of my unpub-
lished observations are an additional support to the hypothesis
which has been gradually developed by Bury, McBride and

SIGNIFICANCE OF CERTAIN LARVAE OF ECHINODERMS. \JJ

Bather, and serve to carry it one step further. Each of these
students has reconstructed the hypothetical ancestor both in its
bilateral free swimming stage and the stage during which it be-
came radially symmetrical. The same plan is followed in this
paper.

The papers of Bury,1 McBride 2 and Bather 3 in which the
hypothetical bilateral ancestor of the echinoderms is recon-

FIG. 9. Pluteus of JMellita testudinata. Original. A, anus; Ec/i., developing
sand dollar ; HI, mouth. Ciliated bands stippled.

structed and figured, are so well known and the reasons for every
detail of the anatomy of the creature are therein so well set forth
that it would be a waste of the reader's time to again do more
than give an outline of the supposed structure of the hypothetical
organism, discussing such points only in which a change is made.

1 Henry Bur}', " The Metamorphosis of Echinoderms," Q. J. Mic. Sc. , No. 149,
1895.

2E. W. McBride, "The Development of Asterina Gibbosa," Q. J. Mic. Sc.,No.
151, 1896.

3F. A. Bather, "A Treatise on Zoology." Part III., " The Echinoderma."
Edited by E. Ray Lankester, 1900.

178

CASWELL GRAVE.

Briefly then, the earliest ancestor of the group of echinoderms
of which there is much trustworthy evidence, was a free-swim-
ming organism of microscopic size with an elongated body and a
long preoral lobe. At the tip of the preoral lobe a sense organ and

FIG. lo. Young Mellita testiidinata. Original. 3, 4, and 5, ciliated rings.

nerve center was located. The contours of the body were plain,
no arms or processes of any kind being present. The alimen-
tary tract occupied the posterior part of the body, the mouth and
anus opening ventrally. Three pairs of body cavities, arranged
symmetrically with reference to the alimentary tract, were present.

SIGNIFICANCE OF CERTAIN LARV^ OF ECHINODERMS. 1/9

The anterior pair extended into the preoral lobe where it may
have been united into a single cavity. Posteriorly its cavities
were placed on the right and left of the oesophagus and each
cavity opened to the exterior, on the dorsal surface of the animal,
through a ciliated duct. The posterior end of each anterior cavity
was connected with the corresponding cavity of the middle pair
by a second duct, also ciliated. The middle cavities were situated
on either side of the point of union of the oesophagus and stomach.
The posterior cavities were larger than those of the anterior and
middle pairs and were applied to the stomach, forming a mesentery
on the dorsal mid line.

If Fig. II a, of this paper is compared with Bury's Fig. 45,
McBride's Fig. 157 and Bather's Fig. I it will be seen that the
general plan is the same with differences in detail only.

Bury's idea that the hydroccele (left middle body cavity) encir-
cled the oesophagus (the right cavity having entirely disappeared)
even during the period of the free-swimming existence of the ani-
mal, is, in the light of recent observations, an unnecessary assump-
tion and one for which no explanation has been made. The
changes which take place in the posterior pair of body cavities of
echinoderm larvae, by which the left one becomes horseshoe-shaped
and encircles the stomach, are almost exactly similar to those by
which the left middle body cavity takes on the form of a ring sur-
rounding the oesophagus. If to explain the former it is neces-
sary, as Bury and others believe, to assume a shifting of the posi-
tion of the mouth and oesophagus incident to a life on the bottom,
then a similar explanation for the latter is also called for. I agree
with the more recent writers in the assumption that both the hy-
drocele and left posterior body cavity acquired their circular
shape and position around the alimentary canal, at the same
time, viz., during the period when the entire organization of the
animal was being readjusted to its new conditions of life on the
bottom.

According to McBride's hypothesis, each of the middle body
cavities possessed, during the free swimming stage of the an-
cestor, five tentacles which were used in the capture of food.
There is good evidence for the existence of two hydrocoeles
(middle body cavities), as McBride has shown in his work on

iSo

CASWELL GRAVE.

a,

o

.5 ^
S

o

C ft

2 15

c/3 t/:

" j ~

o O

C T3
oi

*"C CX,

D ^1

•5 .o o

S ^

J3 S «

o .2

<s s

&l

o

5

tn r°
o

p, M

m
B.

vit

M

rt

IW

o

&

ifl

^

oi

CJ

dj

lH
P

e

<L)
<U

^

>>

T3

-^

D

u

• f.

o

P

•^

f^^

J3

•^

u-

' ^

OJ

t-t

O

O

0

"8

_O
"j^

"d

£

D

^

o

p

<a

+-i
in

C

'c

5,

_o

'o

^
>-,

'O

>%

^3

8,

— i

CJ

"~v

r^H

"~*

(J

>

fli

^»

•2«

c

<D

^

."t-

ui

0

T3

r~^l

,—

«

s °

n . fli

C3 C.

5 *
'•S c
a> °

f) '^J
t* JS
s ai

U) i~

O D

2 •§

i rt

a o

ni C

<U

^3

O
«

S -0

c o

M o •= cq

c
£

as

SIGNIFICANCE OF CERTAIN LARV/E OF ECHINODERMS. iSl

Astcrina, and as is further demonstrated in the pluteus of Mcl-
lita, but that they possessed tentacles or assisted in capturing
food at this time is not, I think, supported by evidence. The
structure and function of the left hydroccele as an organ of lo-
comotion and feeding was, in my opinion, acquired during the
period of sedentary life after the animal had so increased in size
that ciliary action alone was not equal to the work of food
gathering. This point will be more fully discussed, however,
further on.

To the bilateral ancestor as above described I would add, in
the place of the general coat of cilia with which it is usually pro-
vided, a locomotor and feeding apparatus consisting of five trans-
versely-placed ciliated rings and an apical tuft of sensory cilia.
The position of each is shown in Fig. 1 1 , a.

Semon [ in his discussion of the larva of Synapta digitata con-
cludes that the transverse ciliated rings of both the holothurian
pupae and the larva of Antcdon are to be considered as secon-
dary structures. Lang 2 also states that it is not at all likely that
the ciliated rings (of echinoderm larvae) have any phylogenetic
significance. From the facts then available this was the only
conclusion warranted and its has been accepted almost without
exception by zoologists. It seems to me however that the ob-
servations recorded in this paper, which have been made since
the publication of the works of the authors just mentioned, make
it worth while to again call attention to the question and to dis-
cuss the bearing which the accumulated facts have upon our
conception of the structure and history of the hypothetical pelagic
ancestor of the echinoderms.

Larvae such as Auricularia, Bipinnaria, Brachiolaria and
Plutei have never, to my knowledge, been seriously consided to
be primitive although the attempt to establish a relationship be-
tween echinoderms and Balanoglossus, on account of the general
similarity of the movements and external characters of the Auri-
cularia and Tornaria larvae, comes very near to an implied belief
in their primitiveness. In each of the above-mentioned larvae we

'Richard Semon, "Die Entwicklung der Synapta digitata und die Stammesge-
schichte der Echinodernen." Jena Zeitschrift, Bd. XXII., 1888.

2 Arnold Lang, "Text Book of Comparative Anatomy," p. 546. (Eng. trans.)

I 82 CASWELL GRAVE.

have to do with highly specialized organisms, there being a long
period in the life of each during which it is thrown upon its
own resources. Its existence during this period depends upon
its ability to procure food and escape from its enemies. The
rigorous selection which must take place under these conditions
can not have failed to have had a profound effect upon the whole
organization of the larvae and especially upon their external
characters.

The whole tendency has been however to look for primitive
characters among free-swimming larvae, throwing aside those
which are brooded or otherwise cared for as much more likely to
be modified and secondary. If my suggestion as to the signifi-
cance of the larvae of echinoderms with transverse rings is correct
then this view is incorrect. On the other hand we would expect
to find the least modified development among larvse which are
freed, to a greater or less extent, from the task of caring for
themselves, provided in such cases the eggs have not been so
crowded with nutritive materials as to become greatly enlarged
or that, during the brooding, no connections with the mother are
established or protective structures developed. Kenogenetic
characters are no doubt found in both types of larvae and the
problem is to ascertain which has remained truer to the ances-
tral form.

The final, sudden and complicated metamorphosis into the
adult form which is so characteristic c& free-swimming larvae is
good evidence that they have been carried far out of the path of
phylogeny. In larvae without a long independent existence the
metamorphosis is gradual and such as might be expected if it is
in any way a true picture of the past history of the race.

No very great similarity is shown in the external forms of the
familiar types of echinoderm larvae and it is difficult to think of
any one of them as having been the type from which the others
originated, but it is possible to think of them all as having arisen
from a type of larva such as is found in Antedon, Cucumaria and
Ophiura.

Owing to the similarity in position of the ciliated rings with
reference to the other organs of the body of the larvae of the
above-named species, and all other cases in which ciliated rings

SIGNIFICANCE OF CERTAIN LARVAE OF ECHINODERMS. 183

have been found, a very definite homology can be shown to
exist between them, as I have endeavored to indicate by the
numbers which have been placed opposite the rings in each of
the drawings. It is conceivable that the long arms of Aitriai-
laria, Bipinnaria, and the various plutci may have arisen and
been developed from elevations of the ectoderm beneath certain
parts of the ciliated rings, the result of which would have been an
increase in their length and thereby an increase in their efficiency
in locomotion and feeding. The relation which the type of
directly developing larvae of Antcdon, Cncmnaria and Ophiura is
suggested to bear to the familiar types, is shown in Fig. 1 1 ; the
larvae with transverse ciliated rings being considered the primi-
tive condition from which the other larvae have been specialized
and carried far out of the path of phylogeny, as a result of their
independent life. To this type of development the specialized
larvae tend to return at the time when their free-swimming life is
given up for one on the bottom, as is indicated in holothurian
pupae, a certain ophiuran larva and young Mcllitas in all of which
the transverse ciliated rings reappear at the time of metamorphosis
in the form in which they were of functional importance to the
common ancestor during the early period of its life on the bot-
tom.

All larvae have not deviated to the same extent from the direct
line, as is shown not only in their less complicated structures but
also in the less radical readjustment of organs which takes place
during their metamorphosis. In ophiurid plutei, for example,
the larval mouth and oesphagus are taken over as such into the
adult form, which, as has been pointed out, must have been the
case in phylogeny. In echinoid plutei, however, the specializa-
tion which has taken place in these organs has been carried so
far that it is impossible to readapt them to the needs of the adult
and new ones must be formed.

THE ATTACHED FORM AND THE ORIGIN OF RADIAL

SYMMETRY.

With almost no exception, students of the embryology and
anatomy of the echinoderms see no other way, at present, to ac-
count for the peculiar asymmetry of certain of the organs and

184 CASWELL GRAVE.

the perfect radial symmetry of other structures, which are char-
acteristic of the adult condition and which arise, in every known
case, by the remodelling of the structures of a larva which is
bilateral in its entire organization, than by assuming that the
group has been derived from a bilateral pelagic organism, similar
to the one described above, which at a very remote period in its
existence exchanged its free swimming life for one on the bottom
during which it became fixed.

Briefly stated, the steps by which the present organization of
an echinoderm are generally accounted for, and for which there
is more or less evidence, are as follows : Pelagic life was given
up for one on the bottom because of less competition and a
greater food supply in the latter place. The preoral lobe be-
came gradually modified into an organ for fixation. The mouth,
at first directed downward as was its position in pelagic life,
gradually moved to the left until it took up a position in which
it was directed upward. This, in a fixed animal, feeding upon
microscopic organisms, is its most favorable position as is shown
by its position in animals which exist at present under these con-
ditions. During the migration of the mouth and oesophagus,
those organs of the left side which would obstruct such a move-
ment (left middle and posterior body cavities) were carried along
and each became drawn out into the shape of a horseshoe and
greatly hypertrophied. In the final position of each, the opening
of the horseshoe was directed anteriorly. The middle and pos-
terior body cavities of the right side also became changed in
position and correspondingly reduced in size. The left middle
body cavity retained its connection with the exterior through the
greatly reduced anterior left body cavity and its duct, but the
duct of the right side disappeared. During this period when
food was plentiful and easily accessible and when no energy was
used in locomotion, a rapid increase in the size of the creature
took place and radial symmetry was developed. There is such
a diversity of opinion, however, as to the details of this process
that I will attempt to give but one ; that which has been sug-
gested by my own observations.

The ciliated rings were useful to the free-swimming animal not
only as organs of locomotion but were used in feeding as well,

SIGNIFICANCE OF CERTAIN LARV.E OF ECHINODERMS. 185

and during the period when it was fixed on the bottom certain of
the rings continued to function as food gatherers. The two
rings which encircled the preoral lobe, being purely locomotor in
function, were lost, but the other three took up a more definite
relation to the mouth and formed six paths for conducting food
to it (see Fig. 1 1, /;). The retention of the ciliated rings among
directly developing larvae and the return to a condition with
ciliated rings among larvae which possess a more complicated
ciliated apparatus during their free-swimming life may, as I have
stated elsewhere, be explained on this ground. The entire number
of rings is not in every ease retained or reproduced because two,
Nos. i and 2, belonged entirely to the locomotor apparatus (pre-
oral lobe) and except in the holothurians and crinoids (for reasons
suggested later) are no longer needed. Only those rings are re-
tained which in the ancestral line had a function in feeding, and
which are needed for the same purpose during the metamorphosis
of the larvae themselves until the developing tube feet are ready
to assume the function.

This ciliated feeding apparatus which had been brought over
by the hypothetical fixed echinoderm from its free-swimming con-
dition and which, in the new surroundings, had, at first, answered
every need in this line, became gradually inadequate to furnish it,
as it increased in size, with enough food. Those portions of the
ciliated sensory epithelium of the mouth situated between the ends
of the ciliated paths were then gradually developed into tentacles
into each of which a diverticulum of the left middle body cavity,
lying below, protruded (see Fig. n,£). In the anterior space
only, did no tentacle develope. This space contained the ex-
ternal opening of the left middle body cavity (madreporite), the
left anterior body cavity (Ampulla) and possibly the reproductive
organ. There was hence no space in which a sixth tentacle
might have developed.

In this way the pentamerous structure of the hydroccele may
be accounted for and I assume, with others, that the hydroccele
formed the basis upon which the entire radial symmetry of echin-
oderms was built. The ciliated tentacles, simple at first, branched
as they grew in length and assumed more and more the function
of food collecting. As the animal increased in size the space

I 86 CASWELL GRAVE.

immediately surrounding it failed to yield a sufficient supply of
food. The tentacles in reaching about over the bottom in search
of more, detached the animal and a crawling habit was developed.
As the tentacles grew in length and complexity a like develop-
ment in the organs which nourish, enervate, support and protect
them would naturally follow. The tentacles being five in number,
we have in them a possible origin for the pentamerous symmetry
which characterizes the nervous and skeletal systems and a con-
siderable part of the ccelomic cavities of all echinoderms. At
the time when fixed life was given up by the ancestor of those
echinoderms which are at present free living, each of its radii
probably contained a five-branched tentacle, since this is the
number which is possessed by many echinoderms at the period
when their metamorphosis is being completed. The period of
fixation was long enough and the changes which took place in the
organization of the animal at this time were so great that all trace
of an anterior or a posterior part, as such, was lost and now, in its
second period of free life, the direction of locomotion depends
wholly upon external conditions.

During the period when the common ancestor of the group
was fixed, differentiations into at least three different types took
place. One line is now represented by holothurians, one by
crinoids and another by asterids, echinoids and ophiuroids.
Among crinoids alone the fixed condition has been retained. In
this group the problem of enlarging its base of supplies was solved
not by becoming free but by the elongation of the organ of attach-
ment, and by the migration of the mouth and tentacles still further
toward the opposite end. In the type which has given rise to
holothurians, the mouth and tentacles migrated in just the oppo-
site direction, viz., into the organ of attachment and were thereby
brought into relation with the bottom. The free-crawling habit
was later acquired. The ancestor of the starfishes and sea-
urchins made no permanent use of its organ of attachment and
no further migration of the mouth took place but it was brought
into direct relation to the bottom by the rotation of the body as
a whole.

ZOOLOGICAL LABORATORY OF THE JOHNS HOPKINS UNIVERSITY,
April, 1903.

Vol. V. September, 1903. No.

BIOLOGICAL BULLETIN.

THE SPERMATOGENESIS OF THE MYRIAPODS.— II.

ON THE CHROMATIN IN THE SPERMATOCYTES

OF SCOLOPENDRA HEROS.

MAULSBY W. BLACKMAN.

In a detailed study of the spermatocyte changes in Scolopendra
heros, now practically ready for publication, the multiplicity of
subjects requiring consideration is such that it is deemed advisable
to prepare a series of shorter papers, in each of which some par-
ticular class of structures may be considered to the practical ex-
clusion of the others. It is hoped that in this manner the con-
fusion which necessarily occurs where the whole subject is treated
at one time may be avoided. In this, the first of the series of
articles, the chromatin structures alone will be treated.

The spermatogonia of Scolopendra are small cells of an elon-
gated, irregular shape lying parallel to the long axis of the follicle,
and containing an oval nucleus (Fig. I). During the resting
stages the chromatin is all aggregated into one rather large,
spherical nucleolus-like body, usually situated at the periphery
of the nucleus and apposed to its membrane. The remainder of
the nuclear space is filled by an irregular network of granular
fibers apparently differing in no way from the cytoplasmic network
without the nucleus. In staining reaction the nucleolar body
mentioned conforms in all respects to a chromatin body as it in-
dubitably is. When stained with Heidenhain's iron-hsematoxylin,
this structure retains the coloring matter after all other morpho-
logical elements of the cell have become almost colorless. In
lightly stained preparations evidences appear which warrant the
assertion that the body in question is not strictly homogeneous
in structure, but probably includes in its composition linin as
well as chromatin. With Flemming's three-color method the
" nucleolus " takes the dense red stain characteristic of closely

187

1 88 MAULSBY W. BLACKMAN.

aggregated chromatin ; and with the Ehrlich-Biondi mixture,
following the action of suitable fixatives, assumes the green color
usual to chromatin treated by this reagent. Numerous other
stains of a greater or less value as micro-chemical tests were
used and with all these the chromatin nature of this body was in-
variably demonstrated.

The character of this nucleolar body which, for reasons late
made apparent, I shall call the karyosphere is still further indi-
cated by its behavior in the prophase of the spermatogonium.
Owing to the advanced development of my material I have beem
unable to study any but the last generations of these cells, but I
believe that the phenomena here observed are common to all gen-
erations of the secondary spermatogonia. In all cases studied, the
active prophase is characterized by the presence within the nuclear
vesicle of 33 small aggregations of chromatin and the complete
absence of the karyosphere (Fig. 2), thus giving a logical basis
to the conclusion that the chromosomes are derived directly from
the substances of the karyosphere. Of these 33 chromosomes
32 are characterized in the earlier prophases by their granular
consistency, while the remaining one is plainly distinguishable
on account of its homogeneous nature and its clear-cut outline.
This modified chromatic element is the accessory chromosome,
first recognized as a specialized chromosome by McClung, '99,
and later found to be probably of universal distribution in the
male cells of arthropods.

It will be noted that the number of chromosomes, 33, given
above as characteristic of the spermatogonium is not a multiple
of two as is generally considered to be necessarily the case of
immature germ cells. The reason for this fact has to do with
the peculiar character of the accessory chromosome, and can
readily be explained when the later behavior of this element is
known.

During the following phases in the mitosis of the last generation
of secondary spermatogonia, nothing of especial interest with
regard to the chromatin occurs until the telophase is reached.
This phase endures for a considerable time as is shown by the
great number of slightly different stages present and by the fact
that more spermatogonia are found in this condition than in any

THE SPERMATOGENESIS OF THE MYRIAPODS.

189

other stage of mitosis. In the early telophase where the two new
cells are almost completely constricted the chromatin is arranged
in a densely packed mass of chromosomes in which the individual
elements are indistinguishable. Later (Fig. 3) these elements

FIG. I. X I>44° dia. Spermatogonium of Scolopendra heros in the condition of
rest. All of the chromatin is aggregated into one mass, the karyosphere.

FIG. 2. X Ij44° dia. Spermatogonium in prophase. The chromatin is all with-
drawn from the karyosphere and is now in the form of 33 small chromosomes all of
which, with the exception of the accessory chromosome, are of a granular consistency.
This element is homogeneous. The centrosomes are to be seen in the cytoplasm
near the nucleus.

FIG. 3. X Ii44° dia. Telophase of the last Spermatogonium. Synapsis. Cyto-
plasmic division nearly complete. All chromosomes with exception of accessory, be-
coming granular. No nuclear membrane. Centrosomes at poles of the cell.

FIG. 4. XI>44°dia. Later telophase. Synapsis. Chromosomes have lengthened
still more. Accessory chromosome still intact. Growth of the cell has begun.

FIG. 5. X I>44° dia. Early spermatocyte. Nuclear membrane beginning to
form. Accessory has taken up a peripheral position. Mass of chromatin has loosened
considerably and is now seen to consist of segments equal in number to one half the
spermatogonial elements, minus the accessory chromosome. Centrosomes have mi-
grated from their polar position.

begin to lose their homogeneous consistency and to lengthen
out into densely granular segments. Owing to the dense mass-
ing of the chromosomes during this and following stages, the ex-

190 MAULSBY W. BLACKMAN.

act nature of the changes taking place cannot be learned. Sev-
eral facts are however very apparent. Of these one of the most
important is this : — At the time when all the other morphological
constituents of the mass of chromatin are undergoing very funda-
mental changes, one of these elements remains unaltered. While
all of the neighboring chromosomes lose their definite outlines
and are changed into elongated threads of a granular structure
one, the accessory chromosome, does not participate in this
metamorphosis but apparently retains all of the properties char-
acteristic of it during metakinesis. While this difference in con-
sistency is the most apparent discrepancy existing between the
accessory chromosome and the ordinary chromatic elements, as
we shall see presently, it is by no means the most important one.
As the telophase advances the chromosomes continue to
lengthen out into long threads. At first, as we have seen, these
filaments form a dense mass which is surrounded by no mem-
brane marking off the nuclear area from the cytosome. As
the chromosomes become more diffuse this mass also becomes
less dense and the individual segments are not so closely ap-
posed to each other. This stage is shown in Fig. 4, where the
chromatin of each of the two daughter cells is in the form of an
irregular, more or less closely knotted, mass of granular fila-
ments. This mass is contained in a large clear vacuole hav-
ing no visible network of linin or cytoplasm and bounded by no
definite membrane. The appearance of the chromatin grouped
in a diffuse mass upon one side of this vesicle suggests very
strongly a comparison between this stage in Scolopcndra and the
" synapsis " in elasmobranchs, as described by Moore, '95,
and later in different objects by numerous other authors. In all
the reported cases with which I am acquainted, however, this
massing of the chromatin upon one side of the nuclear vesicle
occurs at a considerably later stage than the early or mid-telo-
phase. Paulmier '98, and Montgomery '98, both figure it as taking
place after the formation of the chromatic spireme. McClimg, 'oo,
denies the normal existence of any such massing of the chromatin
in the Acridities, referring such appearances to the distorting
effects of the fixing reagents employed. By the majority of in-
vestigators upon male cells this massing of the chromatin is used

THE SPERMATOGENESIS OF THE MYRIAPODS. IQI

as the criterion of the synapsis or pseudo-reduction, but Mont-
gomery, 'oi, apparently abandoning his former views upon the
subject, asserts, probably with very good reason, that in reality
synapsis occurs at a considerably earlier stage. In Pcripatns,
'oo, he is able to study the manner of this union of the chromo-
somes and from observations seems to have good grounds for
the assertion that synapsis is accomplished by an end to end
union, in pairs, of entire chromosomes during the retrogressive
stages of the telophase of the last spermatogonial division.1

In Scolopcndra, owing to the small size of the spermatogonia
and the extreme minuteness of the spermatogonial chromosomes,
as well as their larger number and close aggregation during the
telophase, the manner of union and the details of the process cannot
be studied ; but it can be stated with the greatest certainty that
pseudo-reduction occurs during the telophase of the last sperma-
togonium, and is completed before the reconstruction of the
nuclear membrane. At the time of the formation of this struc-
ture, the nuclear space is occupied by sixteen elongated seg-
ments of chromatin and resembles very closely the nucleus in
insect cells with the exception that the nuclear area is much
larger in proportion to the amount of chromatin and thus the
segmented character of the chromatin is evident (Figs. 5 and .6).
Besides these sixteen diffuse segments of chromatin, the acces-
sory chromosome is also plainly visible within the nucleus. It
still preserves its distinctive characteristics and has changed very
little from its condition during the preceding division. To be
sure, it has increased in size as have all parts of the cell, but this
increase may all be referred to natural growth. This element
takes no part whatever in the process of synapsis. During the
spermatogonial stages it is a simple chromatic structure and in
the following spermatocyte period it still retains its univalent char-
acter when all of the other chromosomes are bivalent.

The completion of cell division and the union of the chromo-
somes occurring during the telophase have occupied considerable
time, as is shown by several facts. Cells in various stages of the

!A late paper by W. S. Sutton upon "The Morphology of the Chromosome
Group in Brachystola magna" contains further and much more convincing proof of
the truth of this process. Mr. Sutton is able to trace plainly the union of the chro-
mosomes and to show that it is undoubtedly an end-to-end union of entire elements.

192

MAULSBY W. BLACKMAN.

telophase are more numerous in the material examined than those
in any other condition of the spermatogonium. A large number
of different stages may be distinguished. The cells in the early
telophase are small, while those in which the nuclear wall is re-
constructed are considerably larger, showing that already the
growth period has begun. (Compare Figs. 3 and 5.)

With the completion of the nuclear membrane after the last
spermatogonial mitosis, the cells no longer belong to the first
division of the spermatogenetic cyde, but now contain the ma-
tured number of chromosomes and are spermatocytes. In insect
material the transformation is not completed until a period ap-
parently considerably later. However, I believe this difference is
merely in appearance, lying in the fact that the nuclear membrane
is reconstructed much earlier in insect cells.

At this stage the cells of Scolopendra enter upon a period re-
markable for the extraordinary changes which take place in their

FIG. 6. X I>44° dia. Slightly later stage. The chromatin segments scattered
throughout entire nuclear space.

FIG. 7. X 1)44° dia. Chromatin partly gathered about the accessory chromosome
to form the karyosphere. Remaining chromatin of the cell present in the form of
very diffuse segments. Spindle remains of last spermatogonial divisions still persist.

structure. At first glance the most striking of these changes
seems to be the enormous increase in the size of the cells (Figs.
6, 7 and 8). This growth I have already described briefly in a
preliminary paper and shall have occasion to describe more in
detail in subsequent communications. In this connection it will
suffice to say that very often the diameter of the larger sperma-

THE SPERMATOGENESIS OF THE MYRIAPODS.

193

tocytes to that of the spermatogonium stands in a ratio of ten to
one.

Striking as this great increase in the size of the cells certainly
is, it is not as remarkable as are the changes which occur in the
cell in general and especially in the nucleus. Shortly after the
formation of the nuclear membrane, the chromatin segments
leave the tangled mass at one side of the nucleus (Fig. 5), and
arrange themselves irregularly throughout the nuclear space
(Fig. 6). At the same time they shorten and thicken and, as the
nucleus is now quite large, the individual elements may readily be
distinguished and their number counted. In all favorable cases

FIG. 8. X I>44° dia. Pseudo-germinal vesicle stage of the spermatocyte of Sco-
lopendra heros. Chromatin all aggregated in karyosphere which here plainly shows ex-
cept at one point a spongy or reticular structure. This dense portion undoubtedly
represents the accessory chromosome. Persisting spindle still visible. Centrosomes
to be seen imbedded in the zone of archoplasm surrounding the nucleus.

in which this count has been taken it has been found that there
are seventeen chromosomes present (sixteen granular segments
and the accessory chromosome), the number later found in the
metaphase. At this time (Fig. 6) the cells resemble insect
spermatocytes more closely than at any other stage. They are
now in a condition apparently comparable in all particulars to
that of the ordinary sperm cell in the "segmented spireme "

194 MAULSBY W. BLACKMAN.

stage. This is true both with regard to the history of the cell
and as regards the morphology of its various structural elements.
But from now on the behavior in Scolopendra differs very mark-
edly from that of corresponding cells in other animals. In other
arthropods at this stage growth is practically completed and the
maturation mitoses immediately ensue. In Scolopendra the sub-
sequent processes are very different. The growth period has
hardlv begun and the maturation divisions do not occur until

* o

considerably later (probably several wreeks or. even months). In
insects the segmented spireme is considered one of the earlier
stages of the active prophase, while in chilopods a condition more
closely approaching a true rest stage than that occurring at any
other time in the history of the spermatocytes, intervenes between
this stage and the first maturation mitosis.

During this intervening stage the history of the spermatocytes
parallels in nearly all respects that of the typical female germ
cell of a like generation, and the changes which take place result
in a structure which if isolated would certainly be mistaken for
an immature egg.

As I have reported in a preliminary paper, this resemblance is
true not only of the cytoplasmic but of the nuclear elements as well.
As the cell continues in its growth the chromatin segments be-
come larger and more diffuse. They no longer retain the stains
with the persistency which has characterized them heretofore.
This is probably due entirely to the fact that the granules are
farther apart and not to a change in the chemical nature of the
chromatin. Gradually they break clown and their subtance is
accumulated about the accessory chromosome, thus seemingly
increasing the bulk of this element greatly (Fig. 7). This proc-
ess continues until finally all of the chromatin of the cell is
aggregated in one large intensely staining body situated periph-
erally in close contact with the nuclear membrane (Fig. 8). The
remainder of the nucleus is occupied by a beautiful regular re-
ticulum, the achromatic character of which is shown by the fact
that it stains even less densely than the cytoplasmic reticulum
immediately without the nucleus.

In a preliminary paper upon Scolopendra spermatocytes I
stated that I believed this nucleolus-like body to be a homo-

THE SPERMATOGENESIS OF THE MYRIAPODS.

195

geneous mass of chromatin. Since then, however, I have studied
this structure under more favorable circumstances, and am able to
demonstrate that this is not true. In my earlier studies sections
six and two thirds micra thick were used and these were studied
under a magnification of one thousand diameters. In arriving
at my later results thin sections two to three micra thick were
used as well as the thicker ones. These were stained in varying
intensities with HeidenhaifVs iron-haematoxylin and were studied
at a magnification of twelve hundred to eighteen hundred diame-
ters. With these improved conditions it is found that this body,
which I shall hereafter call the karyosphere, is by no means a
simple homogeneous sphere of chromatin, but on the contrary is
a rather complex structure consisting of chromatin, linin and

FIG 9. X I544° dia. Karyosphere as seen in various preparations; a, as it ap-
pears in thick densely stained sections ; b, karyosphere in which the chromatin seg-
ments are massed together by the action of the fixing reagents ; c, thin lightly stained
section of karyosphere showing the real normal structure ; d, section through one side
of karyosphere ; e , karyosphere in early prophase shortly before the appearance of the
chromosomes.

Fir,. 10. X I>44° dia. Nucleus of first spermatocyte in prophase, showing the
origin of the chromosomes from the karyosphere. A number of segments have al-
ready become detached and lie free in the nucleus while others are still connected
with the karyosphere. Those detached have already segmented longitudinally.

karyolymph. It is a mass of fine granular filaments of chromatin
so closely gathered about the accessory chromosome as to pre-
sent, under ordinary conditions and amplification, the appearance
of an irregular homogeneous sphere of pure chromatin (Fig. 9, a).

196 MAULSBY W. BLACKMAN.

Upon higher magnification, sections of this karyosphere usually
present a granular or spongy appearance as shown in Fig. 9, c.
In other cases the chromatin is more or less collected into cer-
tain areas forming a coarse cluster in the center from which proc-
esses extend toward the periphery (Fig. 9, $). Here the body
still retains its approximately spherical form, the portion between
the processes not staining with the chromatin stains but showing
the plasma reaction. Quite often, al?o, we find a karyosphere
which presents the appearance shown in Fig. 9, c. This I regard
as the typical form. It consists of very fine and closely aggre-
gated mass of chromatin filaments arranged in the form of a
more or less perfect sphere. Upon one side of this mass when
the section is cut through the right plane is a smaller homo-
geneous body, the accessory chromosome (Fig. 9, d, e). The
remainder of the karyosphere is made up of irregularly arranged
chromatic strands between which minute interstices, undoubtedly
filled with karyolymph, may be discovered by careful focusing.

Thus it will be seen that during the pseudo-germinal vesicle
stage,1 the karyosphere, with the exception of a membrane, pos-
sesses all of the essential elements of a nucleus — chromatin,
linin (upon which the chromatin is arranged) and karyolymph.
It is in fact a "nucleus within a nucleus" similar to that de-
scribed by Carnoy in the closely allied genera of chtlopods,
LitJiobins, Scutigcra and Geopliiliis. This structure which he
calls the " nucleole noyau," behaves similarly in all essential re-
spects during the first spermatocyte to the karyosphere in Scolo-
paidra licros? It is derived from the chromatin of the nucleus
in a similar manner and during the first maturation mitosis be-
haves in a way essentially alike in all respects.

Carnoy by no means stands alone in the assertion that func-
tional chromatin may and does assume the form of nucleolus-
like bodies during resting periods between mitoses, although the
structures found by him in LitJwbins, Scutigcra, etc., are more
highly organized than those reported by others. Among those
who have observed that the "chromatin nucleolus " is derived

1 See former paper.

2 Carnoy failed to find a ' ' nucleole noyau " in S. dalmatica. He considers the intra-
nuclear body in the cells of this animal a true plasmasome in no way related to the
structure found in Lithobius and other chilopods.

THE SPERMATOGENESIS OF THE MYRIAPODS. IQ/

from the chromatin reticulum may be mentioned the following :
Blochmann, '82 (Neritina) ; Van Beneden, '83 (Ascaris) ; Van
Bambeke, '85 (general); Carnoy, '85 (Arthropoda) ; Rabl, '85
(Salamandra) ; O. Schultze, '87 (Rana and Triton] ; Davidhoff,
'89 (Distaplia) ; Hermann, '8<)(Mus); McCallum, 91 (EcJiinoder-
matd] ; Pick, '93 (Axolotl} ; Holl, '93 (Mw) ; Jordan, '93 (Navf) ;
Mertens, '94 (Pica) ; Metzner, '94 (Salamandrd) ; McCallum,
'95 (Necturus, also in plants); Sobotta, '95 (Mits) ; R. Hertwig,
'96 (poisoned eggs of Ecldnodermatd) ; Carnoy and Lebrun, '97,
'98, '99, ' oo (Amphibia) ; Eisen, 'oo (Batrachoseps) ; Wilson, '01
(chemically fertilized eggs of Toxopneustes) ; and Blackman, 'or
(Scolopehdra), In many of these animals the process has been
followed in such detail that no reasonable doubt can exist as to
the accuracy of the results obtained. In other cases the con-
clusions are not so well supported. In several instances all of
the chromatin is not withdrawn from the nuclear reticulum.
This is especially true of the cells of Amphibia (McCallum, Jor-
dan, Pick, Pisen, ct a/.}. In other batrachian cells all of the chro-
matin is at certain stages collected in a number of granular
masses which also contain linin (O. Schultze, Carnoy and Leb-
run, ct al.\ In Mns, Hermann finds that at first there are several
bodies in the spermatid nucleus but these later fuse to form a
single large karyosphere. In this he is confirmed by Sobotta.

Other authors state that all of the chromatin of the cell is with-
drawn from the nuclear network and deposited in one large "chro-
matin nucleolus." Such appearances have been observed and
carefully studied by Blochmann, Carnoy, Davidhoff, Hermann,
Holl, Sobotta, R. Hertwig, Wilson and others. That the results
of such well-known investigators should be descredited or re-
ceived with scepticism seems strange, yet the majority of cytolo-
gists seem not to believe that chromatin may normally be massed
in a nucleolus-like body and later act as the functional chromatin
of the cell.

Now let us inquire whether such scepticism is justifiable? If
it can be shown that in the Protozoa such aggregates of chromatin
are of common occurrence normally, certainly it is allowable to
conclude that at least some metazoan cells should retain this
characteristic. With regard to the intranuclear structures of

198 MAULSBY W. BLACKMAN.

Protozoa, Calkins has this to say: " A distinct plasmosome or
true nucleolus comparable to the analogous structure in Metazoa
apparently exists in no case save possibly in Actinosphczrium,
and even here is limited to a passing phase during mitosis (Hert-
wig, '98). It is probable that the structures which have been
almost invariably but erroneously called nucleoli do not belong
at all to this category of nuclear elements but represent either
the functional chromatin which is aggregated into a central mass
(karyosome) during the quiescent or vegetative period of cell
life, or the intra-nu clear division center." From the work of
Griiber ('83), Rhumbler ('93), Labbe (96), Hertwig (98), Calkins
('98, '01), and others, we must conclude that chromatin bodies
resembling nucleoli more or less closely are of very frequent
occurrence in unicellular animals. From Calkins' ('01) review
of these investigations it is evident that in its primitive condition
the chromatin is present in Protozoa in the form of dense homo-
geneous masses of chromatin (karyosomes) which act as the
nuclei of these undifferentiated cells. In higher types the nuclei
are more complicated. The chromatin may still occur in simple
masses, but these are contained within a nuclear membrane which
also encloses material other than chromatin (karyoplasm and
karyolymph). The spireme condition so characteristic of the
chromatin of metazoan germ cells is not commonly found in
Protozoa and when present, exists for only a short time.

The karyosomes found in some of the higher types of proto-
zoan nuclei (Actinosphcerium, Hertwig) are not homogeneous
bodies of chromatin, but, besides this substance, also contain linin.
This linin often forms a reticulum upon which the chromatin is
deposited in the form of granules, an arrangement very similar to
that found in the nuclei of metozoan cells, and gives rise to a
structure which is similar to the chromatin reticulum of the more
differentiated nucleus. It is, however, still more strikingly like
the spireme structure of the karyosphere in appearance. That
it is different in some respects, however, is shown by comparing
the subsequent behavior of the two structures. The differences
are what would be expected when we take into consideration the
fact that one is contained in a protozoan cell while the other is in
a metazoan cell. The chromatin elements are much more firmly

THE SPERMATOGENES1S OF THE MYRIAPODS. 199

established in the higher animals and hence it is to be expected
that when the karyosphere breaks down, the resulting fragments
should be distinct chromosomes. In protozoa the conditions are
different. The chromosomes are not such definite structures and
hence when the karyosphere of Actinospharium disintegrates it
gives rise to a large number of granules which later collect into
chromosome-like masses. However, the relationship is certainly
sufficiently close to warrant our placing in the same general cate-
gory ; the solid chromatin nuclei of some Sporozoa and Rhizopoda,
the karyosomes of higher protozoan nuclei, and the karyosomes
and karyospheres,1 found in the nuclei of metazoan cells.

" Chromatin nucleoli " being of such universal occurrence in
protozoan cells, it is to be expected that some metazoan cells ex-
hibit the same structure. As I have already shown, such ex-
amples are fairly common in germ cells and seem to be espe-
cially numerous in somatic cells and in the female germinal
elements. So far as I know they occur only in cells which are
undergoing especially long periods of mitotic inactivity. Such is
certainly very evidently true of the germ cells of Scolopcudra
licros where, during the time of their presence, the pseudo-ger-
minal vesicle stage, the cell increases many times in size.

The pseudo-germinal vesicle stage is succeeded by the active
prophase of the first maturation division. This phase is inau-
gurated by modifications in the cytoplasm of the cell and by the
migration of the centrosome to the nuclear membrane. Upon
reaching this structure the centrosome divides and the two parts
begin their divergent courses.

By the time this is well begun the nucleus also commences to
show signs of activity. The linin reticulum becomes more ragged
and the threads are now composed of finer granules. But the most
important phenomena are those to be observed in connection
with the karyosphere. At a casual glance this structure seems
to have undergone no change, but upon careful examination it is
found that its outline is now more irregular and its consistency

1 In the above terminology I have limited the term karyosome to structures other
than chromosomes found within the nucleus which are apparently composed exclusively
of chromatin. The karyosphere is much more highly organized, as it contains chro-
matin (in a granular, reticular or spireme form), karyoplasm, i. e., linin and karyo-
lymph. It is in fact a miniature nucleus.

2OO

MAULSBY W. BLACKMAN.

more spongy (Fig. 9, e]. This continues to become more
marked until in a short time one or several projections may be
seen extending from its surface (Figs. 10, 11). These granular
filaments stain densely and are similar in all respects to the
chromatin segments characteristic of the " spireme " stage. They
continue to lengthen until when they have attained a certain size
they become detached from the karyosphere and lie free in the
nuclear space (Figs. 10, II, 12). These segments continue to
form until they are exactly equal in number to the threads for-
merly seen in the early spermatocytes.

FlG. II. X I>44° dia. Nucleus of about the same stage as seen in a thinner
section. " Spireme" structure of the karyosphere shown. Tetrads in various stages
of formation. One centrosome with rays to be seen upon nuclear membrane, the other
not included in the section.

FlG. 12. X I>44° dia. Later stage, showing the unwinding of the last chromosome
from the karyosphere, thus again disclosing the accessory chromosome.

As this process proceeds, the size of the karyosphere de-
creases proportionately until finally nothing remains except the
body with which the transformation started, the accessory chro-
mosome. From this fact alone we might indeed be justified in
concluding that the chromosomes are derived from the karyo-
sphere, but no such assumption is necessary. Absolute proof of
the truth of this statement is at hand. Actual observations of
all the stages incident to chromosome formation may easily be
made so that it is impossible for the observer to escape the very
evident conclusions to be drawn therefrom. Figs. 10 and II

THE SPERMATOGENESIS OF THE MYRIAPODS. 2OI

are camera' lucida drawings of nuclei which show the origin
of the chromosomes as well as could be done even by the use
of diagrams. In Fig. 10 the karyosphere is very much reduced
in size and of an irregular shape ; from this three filamentous
projections extend, at the distal end of each of which is to be
seen a segment evidently only just detached. This is already
undergoing the process of tetrad formation. Fig. 1 2 represents
a considerably later stage in which the last chromosome is
leaving the karyosphere and the accessory chromosome is again
unmistakably to be seen. From these observations I believe no
other conclusion can be drawn than that stated above.

To sum up briefly : At the time of the pseudo-germinal vesicle
stage, all the chromatin of the cell is aggregated in the karyo-
sphere which consists of a number of fine chromatin segments
closely massed about the accessory chromosome. In the suc-
ceeding prophase, the first change has to do with the loosening
of this mass of filaments. Later several ends become free and
by simply uncoiling, give rise to slender processes extend-
ing out into the nucleus. These become detached and new
threads are protruded until sixteen segments are present, which
together with the accessory chromosome make up seventeen, the
number of chromatin elements characteristic of the spermatocytes
of Scolopcndra.

Several investigators mentioned before have traced in consider-
able detail the origin of chromosomes from nucleolus-like bodies.
Blochmann, '82 (Ncritina) says : " Das die Elemente der Kern-
blatte aus Theilstucken des Nucleolus entstehen, kann bei
unserem Objeke keinen Zweifel unterliegen, da ich alle Ueber-
gangszustande vom unversehrten Nucleolus bis zur angebildeten
Kernplatte beobachtet habe." In the germ cells of Mns a like
condition undoubtedly exists according to the investigations of
Hermann, '89; Roll, '93 ; and Sobotta, '95. Hermann reports
" chromatin nucleoli " as present in the cells at various stages of
spermatogenesis. Holl shows that in the germinal vesicle of the
mouse ovum there is a large nucleolus composed chiefly of
chromatin from the substance of which the chromosomes of the
first maturation mitosis are formed. Sobotta asserts that during
fertilization the chromatin of each pronucleus is in the form of

2O2 MAULSBY W. BLACKMAN.

one or several large nucleoli of pure chromatin from which are
derived the chromosomes of the succeeding division. In the
maturation of the egg of Distaplia, Davidhoff, '89, has observed
similar phenomena.

C. Schleider, '91, believes that the large nuclei found in the
eees of Echmodermata are but reserve masses of chromatin.

o o

That this is true under some conditions at least, is shown by the
recent experiments of R. Hertwig, '96, and Wilson, 'or. Wil-
son finds that, in one series of eggs chemically fertilized with
MgCl solution, the chromosomes functioning in mitosis are ob-
tained by the breaking down of the large densely staining
" nucleolus." " Its contour becomes irregular and its texture
loose. A little later it assumes a spongy appearance and short
irregular processes are extended from its periphery. Enlarging
still more it now gives almost the appearance of a close, broken
spireme from the ends of which chromatin threads here and
there project." These threads later form the chromosomes. As
will be readily seen this process in Toxopncustes is very similar
in many respects to that occuring in Scolopendra.

The chromatin segments as they arise from the karyosphere
in Scolopendra are long, slender, granular filaments usually con-
siderably curved and distorted (Figs. 10, 1 1, 12). They are
arranged irregularly throughout the nuclear area supported by
the limn reticulutn. Very soon after their detachment from the
karyosphere, they are seen to be divided longitudinally along
their entire length. Owing to the length and distortion of these
segments they frequently assume very fantastic shapes. In some
cases the two parts are coiled or twisted about each other like the
strands of a rope (Fig. 11) while the two halves of other chro-
mosomes may be separated by a considerable distance (Fig. 10).
This cleavage of the segment very evidently represents the lon-
gitudinal division of the chromosome, and as the chromosome is
first divided in this manner in the prophase it is, I believe, justifi-
able to conclude with McClung, 'oo, that the first maturation
mitosis accomplishes the equational division of the chromatic ele-
ments. Apparently the next change in the structure of the split
segments is shown in Fig. 13, a. This first becomes apparent as
a weakening of the two parts of the segment at about their mid-

THE SPERMATOGENESIS OF THE MYRIAPODS. 2O3

die. The threads show a tendency to bend at a more or less
acute angle at this point, and this soon results in a transverse
division of each of the parts of the chromosome. Thus each of
the chrornatin segments has been divided into four parts and may
from now on be called a tetrad. Following the terminology
suggested by McClung, 'oo, I shall designate each of the parts
going to make up the tetrad or chromosome of the first sperma-
tocyte, a chromatid. By this system I believe much confusion will
be prevented.

After the cross division has become established the next
change observable is shown in Fig. 1 3 b, c, g. The chromatids
revolve upon each other in such a manner that the ends at the
point of transverse cleavage are drawn out parallel to each other
and an irregular cross-shaped figure is thus formed (Fig. i$,d, c}.

a

*

FIG. 13. X 1,400 dia. Various stages and modifications of tetrads, a, b, c, early
stages in the process of transverse division. </, typical tetrad of mid-prophase. e, J\
g, h, modifications of the tetrad type.

FIG. 14. X I»44° dia. Later stages in the history of the tetrad, a, typical cruci-
form tetrad of later prophase. b, "double V" form of chromosome at the same
stage, c, d, successively later stages of the cross figure, e, /, apparent modifications
of tetrad in later prophase. 2, //, typical chromosomes at beginning of metaphase.
g, tetrad undergoing longitudinal division.

This cross-shaped figure is composed of four arms of about
equal length each of which is split longitudinally. Owing to the
very irregular shape of these arms, the cleavages are masked and
are often very hard to demonstrate. However, in later stages
when the arms are greatly shortened the bipartite structure is
readily seen (Fig. 14, a, b, c\ and is also strongly indicated even
in the earlier stages by the diamond-shaped opening at the center

2O4 MAULSBY W. BLACKMAN.

of the tetrad. When seen en face this opening is always square
or diamond-shaped with the angles directed toward the arm, in-
dicating that it is continuous into each arm.

At the stage represented in Fig. 13, d, the tetrads are often so
distorted that the typical form is lost, but upon studying them
more carefully it is seen that they are always referable to the
same type. Taking d as the type, the more common variations
are shown in b, c,f,g, h. At b the formation of the arms, instead
of occurring in the plane of the threads, has proceeded in a plane
at right angles thereto, resulting in the double V figures first
mentioned by Paulmier. At c, h the long arms of the cross have
been curved around and nearly brought in contact. Such dis-
tortions observed in later stages of tetrads result in a figure simi-
lar in shape to a seal ring, the point of double cleavage repre-
senting the seal and the long arms approximating to form an
apparently closed circle. Fig. 13, e, f, g are but slight or ap-
parent modifications caused by viewing the tetrads diagonally or
in profile.

By later changes the arms of the cross figures are much short-
ened and the divisions between separate chromatids become very
apparent (Fig. 14, <7, b, r). However, this shortening and con-
densation continues and these divisions are entirely obliterated
and the chromosome becomes first a granular mass and later an
apparently homogeneous one. The tetrads even at this stage
vary considerably in shape as shown in Fig. 14, d, c,f. The
typical form is represented by Fig. 14, d, and by numerous
chromosomes in Fig. 15.

During the prophase the tetrads of each nucleus have not de-
veloped synchronously, but at any given time are in various
stages of formation (Fig. 11). This phenomenon is very easily
explained. On account of the dense massing of the chromatin
segment in the karyosphere, but a few elements can separate at
one time and it very naturally follows that those first escaping from
this body exhibit more advanced development than those arising
later. At a short time before the dissolution of the nuclear
membrane, however, the more tardy individuals have overtaken
their fellows and all now appear as homogeneous bodies exhibit-
ing strongly all the chromatin reactions.

THE SPERMATOGENESIS OF THE MYRIAPODS. 2O5

As will be seen from the foregoing description, the tetrads
occurring in Scolopendra are similar to those previously described
by other authors in various arthropods. What may be taken as
the type of these figures was first reported by Paulmier, '99, in
Hemiptera and McClung, 'oo and '02, in Orthoptera. Structures
differing slightly in detail, the apparent divergence evidently
being due more to interpretation than to any essential morpho-
logical variations, have been found in other arthropods by Henk-
ing, '91 (Pyrrhocoris), Vom Rath, '95, '97 (Gryllotalpd), Toyama,

FIG. 15. X i,44O dia. Nucleus of first spermatocyte during late prophase, show-
ing various modifications in the shape of the chromosomes at this time. The acces-
sory chromosome is seen to be split longitudinally. Centrosomes with well-developed
astral rays at opposite poles of the nucleus.

'94 (silkworm), Riickert, '95 (Copepoda), Montgomery,1 '98, 'oo,
'oi (Hemiptera, Pcripatus], Blackman, 'oi (Scolopcndni), P. Bouin,
'02 (Lithobius), Miss Nichols, '02 (Oniscus], and others. In the
tetrads observed by all of these authors, the cleavages universally
represent a longitudinal and a transverse division of the double or
bivalent chromosome of the spermatocyte. Apparent discrep-
ancies are undoubtedly due to mere variations in detail or differ-
ences in interpretation and denote no real important divergence
in the formation of the tetrads.

1 In his earlier paper Montgomery reported two cross divisions as occurring in
Pentatoma (£ut-/n's/us), but in his subsequent publications has denied the accuracy
of this observation and now believes that one longitudinal division invariably occurs.

2O6 MAULSBY W. BLACKMAN.

The results of Wilcox, '95, '96 {Caloptenus}, and de Sinety,
'01 (Orthoptera), however, are indeed radically different. Even
here, however, I believe that the divergence is due either to the
authors' interpretation of observations, or to insufficient or infe-
rior material. Wilcox asserts that the two spermatocyte mitoses
accomplish a double transverse division of the chromosomes.
Such is not the case in the western individuals of the same species
where a longitudinal, followed by a transverse, division invariably
occurs. De Sinety, working upon the cells of several genera of
Orthoptera, asserts that the two divisions are longitudinal. This
also appears to be a mistaken conception, as pointed out by
McClung, '02. Appearances which upon superficial examination
might lead to this view are occasionally met with in Orthopteran
material, but when studied closely a different interpretation must
always follow. In Scolopcndra spermatocytes I believe it would
be impossible to arrive at this conclusion however strong a pre-
conception the observer may have had. The tetrad figures
accompanying this article can by no possibility be logically inter-
preted as representing anything but a longitudinal and transverse
division of the chromosomes. In the interpretation of the first
spermatocyte chromosomes and in the sequence of the succeed-
ing divisions I am gratified to note that P. Bouin, working upon
other genera of Myriapoda, agrees with my conclusions upon
Scolopendra.

The tetrad forms which are of most common occurrence in the
arthropods are modifications of the cross, double V and ring
figures found in Anasa (Paulmier, '98) and Hippisais (McClung,
'oo). It is very probable that all of the other tetrads found in
this group are obtained by a greater or less modification of the
same process. Such is evidently the case in Copepoda (Ruckert,
'94) (Hacker, '95) ; in Gryllotalpa (Vom Rath, '91) and seems
also to be true of other invertebrates, TJialasseina and ZirpJiea
(Griffin, '95), Unio (Lillie, '95), etc.

The typical arthropod tetrad as exhibited in the Insecta,
and in the Myriapoda is obtained in the following manner :
The chromatin segments of the matured number as they arise
from the spireme stage (Insecta) or from the aggregated seg-
ments in the karyosphere (Myriapoda) are long slender threads

THE SPERMATOGENESIS OF THE MYRIAPODS. 2O/

of granular chromatin. Each thread very quickly splits longi-
tudinally, thus giving rise to two long slender segments ex-
tending parallel to each other. Very shortly after this longi-
tudinal split is made, apparent indications of the second cleavage
may be seen. The first indication of this is a bending of the two
halves of the segment at their middle point. This extension may
be in exactly opposite directions when the resulting tetrad is of the
typical cross shape or may occur in, such a manner that the two
angles are drawn out parallel to each other, in which case the
double V figure results. This stage of the two forms of tetrad fig-
ures is shown in Fig. 13, a, b. The bending of the two segments
soon results in a transverse cleavage at the angles as indicated in
Fig- I3. ^» ^. f> £, /*• The short processes thus produced
elongate at the expense of the length of the quadripartite seg-
ment until a cruciform figure is produced, the four arms of which
are of about equal length. Each of these arms is traversed by a
split extending its entire length and thus producing a diamond-
shaped opening in the center of the X figure. Thus it is brought
about that the two adjacent halves of contiguous arms are con-
tinuous and form one of the four chromatids derived by the
double splitting of the chromatin segment (Fig. 14, a). The
structure of the tetrad is best seen in Scolopendra in the later
stages of tetrad formation when the arms have shortened and
when the chromatin granules are more densely grouped together
(Fig. 14, a, b, r). In the late prophase the chromosome becomes
homogeneous and assumes the four-lobed shape represented in
Fig. 14, d, e,f. The diamond-shaped opening at the center and
the splits in the arms are entirely obliterated.

While these fundamental changes have been taking place in
the other elements the accessory chromosome has also under-
gone some alteration. As it emerges from the karyosphere this
element is a homogeneous spherical mass of chromatin. (Fig.
1 2). In the late prophase it is no longer spherical but presents
the appearance of a rod the two ends of which are constricted
(Fig. 1 5). This constriction undoubtedly indicates a longitudi-
nal division.

When its history is considered this divergence in form from
the tetrads surrounding it is very readily explainable and is pre-

2O8 MAULSBY W. BLACKMAN.

cisely what should be expected. Each of the other chromosomes
is derived by the fusion of two of the spermatogonial chromo-
somes during the telophase of the last mitosis of the division
period. On the other hand, the accessory chromosome is de-
scended directly from a single element of the spermatogonium.
This being true, it is but logical to expect it to behave differently.
The primary object of the spermatocyte period is the reduction
of the chromosomes to one half the somatic number. It is
usually, if not invariably, the case, in arthropods at least, that
this period is characterized by two divisions of the chromosomes,
a longitudinal and a cross division. It is generally assumed
that, by one of these divisions — the transverse division — reduc-
tion is accomplished by the pulling apart of the chromosomes
at the point at which they were united in the preceding synap-
sis. Now as the accessory chromosome is not obtained by the
union of two spermatogonial chromosomes, this reducing divi-
sion is not necessary and does not take place. For these reasons
while the ordinary chromosomes are each composed of four
parts, /. e., are tetrads, this modified chromosome is made up of
but two parts, i. c., is a dyad. Furthermore, it is logically to be
expected that the accessory chromosome being dyad in its nature
would take part in only one of the succeeding divisions. This
peculiarity has indeed been observed by many investigators of
insect spermatogenesis and several explanations more or less
supported by observed facts, are offered in explanation thereof.

In Scolopendra, as in other arthropods, the longitudinal
division of the chromosomes occurs in the first spermatocyte
mitosis. Strong indications of the character of this cleavage may
be seen in the metaphase of the first spermatocyte. Fig. 14, i
represents a typical chromosome at the time of the formation of
the first maturation spindle. At g is shown a tetrad of the same
kind undergoing metakinesis. By a comparison of these two
chromosomes it becomes evident that it is a longitudinal division
of the element which occurs. The mantle fibers are attached to
the two ends, and when the force which separates the halves of
the two chromosomes is applied, the two parts glide over each
other and seem to separate with the greatest reluctance. The
strongest proof that we are here dealing with an equation division,

THE SPERMATOGENESIS OF THE MYRIAPODS. 2OQ

however, is to be found in the prophase. As I have already
noted the longitudinal split is the first made manifest at that time,
hence logically would be expected to preceed the transverse
division, which does not appear until later. Further proof of the
sequence of the divisions is found in the second spermatocyte
where, as will be presently seen, a cross division of the chro-
mosomes certainly occurs.

The question as to the sequence of the two spermatocyte di-
vision, while probably of not any vital importance, has been the
subject of considerable controversy. By far the greater number,
however, agree that the equation division comes first, and is suc-
ceeded by the reduction division. Ruckert, '92, Hacker, '92,
McClung, 'oo, '02, Blackman, '01, P. Bouin, '02, in arthropods,
and Bolles Lee, '97, Linville, 'oo, Griffin, '99, Klinckowstrom,
'97, Francotte, '97, and Van der Stricht, '98, in other inverte-
brates, have arrived at this conclusion. While the opposing
view — /. e., that the reduction division precedes — is held by
Vom Rath, '92, '95, Henking, '90, Paulmier, '99, and Mont-
gomery, '98, 'oo, '01, in anthropods and Lillie, 'oi, in molluscs.
In arriving at this latter conclusion the criterion invariably used is
the appearance and behavior of the elements during the two
mitoses. But during the metaphase the chromosomes are always
so compact that the cleavages shown in the prophase are entirely
obliterated, and the manner of division therefore cannot be de-
termined with certainty. An example of the likelihood of mis-
interpretation of the nature of these divisions is shown by Griffin,
'99, TJialasscuia. Here the first division is very evidently longi-
tudinal, and upon superficial observation the second also appears
to be of the same nature. But when the phenomena observed
in the prophase are considered, it is evident that this cannot be
true, as an indubitable transverse cleavage was to be seen at that
time. Upon further study Griffin shows his first impression to
be false, for the second division is in reality a reducing division.

In all of the investigations with which I am acquainted it has
been reported that the longitudinal cleavage is the first to be made
evident in the prophase. Then I believe it is but logical to con-
clude that this division is completed by the first spermatocyte
mitosis, especially when this has been shown to be the case in a

2IO

MAULSBY W. BLACKMAX.

great number of cells. Of course it is possible that the process
varies in different animals, but it is not probable, for if the
sequence of the actual divisions varies, we should naturally ex-
pect the prophase phenomena to vary in a like manner. No such
variation seems to exist.

The chromosomes as they occur in the metaphase are arranged
in no definite equatorial plate but are scattered irregularly through-
out the equatorial region of the spindle (Fig. 16). It is also
noticeable that the chromosomes do not divide synchronously.

FIG. 16. X 9^° dia. Early metaphase of first spermatocyte. Showing the diversity
in shape of the chromosomes, and their irregular arrangement in the equatorial region.

FIG. 17. X 9^0 dia. Telophase of first spermatocyte, showing the unequal division
of the chromatin, the accessory chromosome being present in one cell while it is absent
in the other.

While some still plainly show their tetrad character, others have
completed their separation and have already started toward the
poles.

Owing to the approximately equal size of all the chromosomes
and the diversity of shapes which they present it has not been
found possible to trace the history of the accessory chromosome
during the first metakinesis. However, from an examination of
the telophase succeeding, it becomes evident that this element

THE SPERMATOGENESIS OF THE MYRIAPODS.

211

in Scolopendra undergoes processes analogous to those reported in
insects by a number of investigators. It is found in one of the
cells resulting from the first mitosis and does not occur in the
other (Fig. 17) showing that it takes no part in this division
but goes over to one cell undivided.

With the reconstruction of the daughter nuclei, all of the
chromosomes except the accessory become granular (Fig. 18)
and present the appearance of rather short rods of diffuse chro-
matin, the center of each of which is slightly constricted, thus

FIG. 1 8. X 96° dia. Prophase of a second spermatocyte containing the accessory
chromosome. The ordinary chromosomes are diffuse and of a dumb-bell form, while
the accessory is homogeneous and spherical. Centrosome and persisting archoplasm
visible.

FIG. 19. X96odia. Late prophaSe of second spermatocyte. Chromosomes are less
diffuse. Accessory chromosome seen to be constricted longitudinally, while the others
show indications of a transverse division.

^

producing a dumb-bell-shaped body. In the succeeding stages
these become more dense (Fig. 19) and finally go to the equa-
torial plate as small homogeneous bodies of a distinctly lobate
structure. When arranged in the equatorial region (as in the
first division, there is no true equatorial plate) the lobes of these
bodies are directed toward the poles of the spindle, thus giving
further basis for the conclusion that we have here a cross division
of the chromosome.

212

MAULSBY W. BLACKMAN.

During the metaphase, however, one of the chromatic ele-
ments does not show the dumb-bell-shape characteristic of the
rest, but is very evidently a rod split in the opposite direction,
/. e., longitudinally. This peculiarity is also to be seen in the
preceding prophase where the accessory chromosome is of the
same shape as in the first spermatocyte prophase. As it is seen
during the early metaphase, this element is arranged with the
plane of cleavage at right angles to the spindle (Fig. 20), but
upon the contraction of the mantle fibers which are attached to

FIG. 20. X 9°° dia. Metaphase of second spermatocyte. The difference in shape
and orientation existing between the accessory and the other chromosomes is evident.

FIG. 21. X l>92° dia. High magnification of same stage, showing the differences
exhibited by the accessory chromosome in the relation of the chromatids and in the
attachment of the mantle fibers.

FIG. 22. X 1,920 dia. Slightly later stage ; showing the effect of the contraction
of the mantle fibers on the orientation of the accessory chromosome.

opposite ends of the erement it revolves through an arc of 90°
(Figs. 21, 22) and the component chromatids as they are pulled
apart seem to glide over each other (Fig. 22) in a manner similar
to that already noted as characteristic of the ordinary chromo-
somes during the first mitosis.

It will be seen by consulting the accompanying figures that
the behavior of the other elements is quite different. These are
arranged with their long axis parallel to that of the spindle, the

THE SPERMATOGENESIS OF THE MYRIAPODS. 213

separation of the chromatid occurring along the equatorial plane
at the place of constriction. This very evidently represents a
cross division of the chromosome.

In the division figures of one half of the second spermatocytes,
all the chromosomes are of one type (the dumb-bell form), the
accessory chromosome not being present. Thus it will be readily
seen that the cells arising from two spermatocyte mitoses are di-
vided into two classes of equal numbers — those which possess
the accessory and those which do not. Similar phenomena have
been observed in the cells of a number of insects by Henking,
'90, Paulmier, '99, McClung, 'oo, '02, de Sinety, '01, and others.

Regarding the function of the modified chromosome, two
theories have now been advanced. Paulmier in his paper on
Anasa puts forth the theory that the " small chromosome repre-
sents characteristics which are being eliminated from the race."
He bases this conclusion entirely upon the failure of the ele-
ment to divide in one spermatocyte division. Montgomery in
his later papers adopts the conclusions of Paulmier and believes
with him that it is a chromosome undergoing the process of
elimination.

McClung, '02, however, in a paper in which he considers in
detail all of the reported observations upon the accessory chromo-
some, formulates an hypothesis which ascribes a very different
function to this element. He maintains that the mere fact of the
unequal apportionment to the spermatozoa would not necessarily
indicate that the element is degenerating, and in addition there
are other facts which militate strongly against such a conclusion.
The extreme nicety with which this element is excluded from
contact with the others in most stages, especially in the sper-
matogium, would seem to indicate a very different and much
greater significance. This exclusiveness, taken in connection
with the fact that exactly one half the spermatozoa contain the
accessory chromosome, suggests the theory that it has to do with
the determination of sex, as this is the only respect in which the
progeny are divided into two classes of equal numbers. Although
no positive proof is advanced to support this theory the author
establishes in a very logical manner the probability of the acces-
sory chromosome representing such a function. It seems to

214 MAULSBY W. BLACKMAN.

possess all the characteristics required of such an element. My
observations upon Scolopendra surely lend support to this theory.
Definite proof of the function of this structure can only be
obtained, however, by a study of the process occurring in the
fertilization and cleavage of the egg.

My observations upon the accessory chromosome in Scolopen-
dra have added very little to our knowledge of this element, except
in so far as they help to show its wide distribution and the great
similarity of its behavior in widely separated groups. Indeed in
all important particulars the phenomena accompanying the devel-
opment of this structure are identical in Chilopoda and Orthoptera,
although the minor details of the process vary considerably. In
both groups the element is derived directly from a single sperma-
togonial chromosome, and for this reason takes no active part in
the phenomena of synapsis. During the prophase when the
other chromosomes divide into four chromatids and form tetrads,
this element, as would be expected from its origin, cleaves but
once and that longitudinally. In the two succeeding divisions
it is divided but once and thus is present in but one half of the
spermatids. The differences, although at times puzzling, are in
reality slight and unimportant. Thus, at the time when all of
the chromatin is aggregated in the karyosphere, the accessory
chromosome cannot be distinguished except in the most favor-
able cases ; but from the study of these thin, well-differentiated
'sections we are justified in saying that even in the pseudo-germ-
inal vesicle stage this element retains all its ordinary character-
istics. In the mataphase of the first spermatocyte it cannot be dis-
tinguished from the other chromosomes as it can in Orthopteran
material, because it is of approximately the same size as these.
In the second maturation division, however, it is again very evi-
dent, by reason of the fact that it divides longitudinally while the
other chromosomes divide transversely.

These variations, as has been said, are unimportant modifications
of behavior and do not represent such fundamental differences as
seem to exist between the " small chromosome " (Paulmier) or the
"chromatin nucleolus " (Montgomery) in Hemiptera and the ac-
cessory chromosome in Orthoptera. If the observations of Paul-
mier and Montgomery concerning the origin of this element are

THE SPERMATOGENESIS OF THE MYRIAPODS. 2 1 5

correct, it is indeed doubtful whether the bodies described repre-
sent the same structure as the accessory chromosome. The
chromosome x of Protenor (Montgomery, 'oi) would seem more
closely to approach this modified element in origin and behavior.
I am glad of this opportunity of expressing my gratitude to
Dr. C. E. McClung for valuable advise and criticism throughout
the progress of this work.

LABORATORY OF ZOOLOGY AND HISTOLOGY,
UNIVERSITY OF KANSAS, April n, 1903.

BIBLIOGRAPHY.
Blackman, M. W.

'oi The Spermatogenesis of the Myriapods, I. Notes on the Spermatocytes and

Spermatids of Scolopendra. Kans. Univ. Quart., X., 1901.
Blochmann, F.

'82 Ueberdie Entwicklung der Neritina fluviatilis. Zeit. f. Wiss. Zool., 36., 1882.
Bouin, P.

'oi Sur le fuseau, le residu fusorial et le Corpuscule intermediare das les cellules

seminales de Lithobius. forficatus. C. R., iqoi.
Bouin, P.

'02 Reduction chromatique chez les Myriapods comp. rend, de 1'assoc. des anat.
Calkins, G. N.

'98 The Phylogenetic Significance, of Certain Protozoan Nuclei. Ann. N. Y.

Acad. Sci., II, 1898.

'98 Mitosis in Noctiluca. Journ. Morph., 15, 3, 1898.

'oi The Protozoa. Colum. Univ. Biol. Series. Macmillan, New York, 1901.
Carnoy, J. B.

'85 La Cytodierese chez les arthropodes. La Cellule, I, 1885.
Carnoy, J. B., & Lebrun, H.

'97 La vesicule germinative et les globules polaires chez les Batrachiens. La

Cellule, 12, 1897.

'98 La Vesicle germinative et les globules polaires chez les Batrachiens. La Cel-
lule, 14, 1898.
'oo La vesicule germinative et les globules polaires chez les Batrachiens. La

Cellule, 16, 1900.
Eisen, G.

'oo The Spermatogenesis of Batrachoseps. Journ. Morph., 17, 1900.
Fick, R.

'93 Ueber die Reifung und Befruchtung des Axolotleies. Zeit. f. \Viss. Zool.,

56, 1893.
Griffin, B. B.

'99 Studies on the Maturation, Fertilization and Cleavage of Thalassema and Zir-

phrea. Journ. Morph., 15, 1899.
Hacker, V.

'92 Die Eibildung bei Cyclops und Canthocampus. Spengel's Zool. Jahrb., 5,
1892.

2l6 MAULSBY Y\". BLACKMAN.

Henking. H.

'90 Untersuchungen iiber die ersten Entwicklungsvorgange in den Eieren der
Insecten. 2. Ueber Spermatogenese und deren Beziehung zur Eientwick-
lung bei Pyrrhocoris apterus, M. Zeit. f. Wiss. Zool.. 51, 1890.
Hermann, F.

'89 Die postfotale Histogenese des Hodens der Maus bis zur Pubertat. Arch,
f. mikr. Anat., 50, 1889.

Hertwig, R.

'84 Ueber die Kerntheilung bei Actinosphserium Eichhornii. Zeit. f. Natur. w.

Jena, 17, 1884.
'96 Ueber die Entwicklung des unbefruchteten Seeigeleies. Festsch. fiir Gegen-

baur, 1896.

Holl, M.

'99 Ueber die Reifung der Eizelle bei den Saugethieren. Sitzb. d. Akad. VYis-
sensch. von Wien., 1899.

Jordan, E. 0.

'93 The Habits and Development of the Newt. Journ. Morph., 8, 1893.
Lillie, F. R.

'01 The Organization of the Egg of Unio, Based on a Study of its Maturation,
Fertilization and Cleavage, journ. Morph., 17, 1901.

Linville, Henry R.

'oo Maturation and Fertilization in Pulmonate Gasteropods. Bull. Mus. Comp.
Zool., Harvard Coll., 35, 1900.

Macallum, A. B.

'95 On the Distribution of Assimilated Iron Compounds, other than Haemoglobin
and Hsematins, in Animal and Vegetable Cells. Quart. Journ. Micr. Soc.,
N. S., 38, 1895.

McClung, C. E.

'99 A Peculiar Nuclear Element in the Male Reproductive Cells of Insects

Zool. Bull., 2, 1899.

'oo The Spermatocyte Divisions of the Acrididre. Kans. Univ. Quart., 9, 1900.
'02 The Accessory Chromosome — Sex Determinant? Biol. Bull., 3, 1902.
'02 The Spermatocyte Divisions of the Locustidse. Kans. Univ. Sci. Bull., 1, 1902.

Montgomery, T. H., Jr.

'98 The Spermatogenesis in Pentatoma up to the formation of the Spermatid.
Zool. Jahrb., 12, 1898.

'98 Chromatin Reduction in the Hemiptera, A Correction. Zool. Anz., 22, 1898.

'98 Comparative Cytological Studies, with Especial Regard to the Morphology
of the Nucleolus. Journ. Morph., XV., 1898.

'oo The Spermatogenesis of Peripatus balfouri up to the Formation of the Sper-
matid. Zool. Jahrb., 14, 1900.

'oi A Study of the Germ Cells of Metazoa. Trans. Amer. Phil. Soc., 20, 1901.
Moore, J. E. S.

'95 On the Structural Changes in the Reproductive Cells during the Spermato-
genesis of Elasmobranchs. Quart. Journ. Micr. Soc., N. S., 38, 1895.

Paulmier, F. C.

'98 Chromatin Reduction in the Hemiptera. Anat. Anz., 14, 1898.
'98 The Spermatogenesis of Anasa tristis. Journ. Morph., 15, 1898.

THE SPERMATOGENESIS OF THE MYRIAPODS. 2 I/

Platner, G.

'86 Die Karyokinese bei den Lepidopteran als Grundlage fur eine Theorie der

Zelltheilung. Internal. Monatsschr., Anat. Hist., 4, 1 886.
Rath, 0. Vom

'92 Zur Kenntnis der Spermatogenese von Gryllotalpa vulgaris, Latr. Arch. f.

Mikr. Anat., 40, 1892.
'95 Neue Beitrage zur Frage der Chromatinreduclion in der Samen und Eireife,

IS95-
Rhumbler, L.

'93 Ueber Entstehung und Bedeutung der in den Kernen vieler Protozoen und in

Keimblaschen vom Metazoen vorkommenden Binnenkorper. Zeit. f. Wiss.

Zool., 6l, 1893.

Ruckert, J.

'92 Zur Eireifung bei Copepoden. Merkel and Bonnet's Anat. Hefte, 1892.
Schultze, 0.

'87 Untersuchungen iiberdie Reifung und Befruchtung des Amphibieneies. Zeit.
f. wiss. Zool., 45, 1887.

Sinety, R. de.

'01 Recherches sur la biclogie et 1'anatomie des Phasmes. La Cellule, 19, 1901.

Sobotta, J.

'95 Die Befruchtung und Furchung des Eies der Maus. Arch. f. Mikr. Anat.,

45- lS95-
Stuhlmann, Fr.

'86 Die Keifung des Arthropodeneies. Ber. d. Naturforsch. Gesell. v. Frei-
berg, i. 8, 1886.

Sutton, W. S.

'oo The Spermatogonial Divisions of Brachystola magna. Kans. Univ. Quart.,

9, 1900.
'02 On the Morphology of the Chromosome Group in Brachystola magna. Biol.

Bull., 4, 1902.

Wallace, Louise.

'oo The Accessory Chromosome in the Spider. Anat. Anz., 18, 1900.

Wheeler, W. M.

'97 The Maturation, Fecundation and Early Cleavage in Myzostoma glabrum.
Arch, de Biol., 15, 1897.

Wilcox, E. V.

'95 Spermatogenesis of Caloptenus femur-rubrum and Cicada tibicen. Bull.

Mus. Comp. Zool., Harvard Univ., 27, 1895.
'96 Further Studies on the Spermatogenesis of Caloptenus femur-rubrum. Bull.

Mus. Comp. Zool., Harvard Univ., p. 29, 1896.

Wilson, E. B.

'01 Experimental Studies in Cytology. I. A Cytological Study of Artificial Par-
thenogenesis in Sea Urchin Eggs. Archiv f. Entwm., 12, 1901.

'01 Experimental Studies in Cytology. II. Some Phenomena of Fertilization
and Cell Division in Etherized Eggs, 1901.

III. The Effect on Cleavage of Artificial Obliteration of the First Cleavage
Furrow. Archiv f. Entwm., 14, 1901.

THE EFFECTS OF HEAT ON THE DEVELOPMENT

OF THE TOAD'S EGG.

HELEN DEAN KING.

An extended series of experiments made by Hertwig (1-4)
prove that the maximum temperature at which the eggs of the
frog will develop normally differs for different species. His ex-
periments also show that eggs in the cleavage stages can with-
stand a higher temperature than can unsegmented eggs. These
results have a bearing on the general problem of adaptation ; for
it may be possible to show, after more species have been studied,
that the maximum temperature which the eggs of amphibians
can endure without injury and also the temperature most* favor-
able for their development depend, to a certain extent at least, on
the time of year at which the eggs are deposited.

MATERIAL AND METHOD.

The eggs of the common toad, Biifo lentiginosus, were used
in making all of the experiments recorded in the present paper.
After natural fertilization, the eggs were brought into the labora-
tory where the temperature varied from 18 to 2 1 ° C. Control sets
of eggs from each lot used for the experiments, developing at
the room temperature, all became perfectly normal embryos, and
some of them were kept until metamorphosis.

In making the experiments, small dishes containing about 80
c.c. of spring water were placed in the drying chamber of a large
water-bath, and after the water had become heated, from 50 to 75
eggs were quickly transferred into it and left a given length of
time. The temperature to which the eggs were being subjected
could readily be told from a thermometer that projected into
the chamber through a small opening in the top. Great care
was taken to keep the temperature of the chamber as constant
as possible during the course of the experiments, and in no case
did it vary more than two degrees. After the eggs were re-
moved from the chamber, they were put into fresh water at room

218

EFFECTS OF HEAT ON TOAD'S EGG. 2IQ

temperature and their later development compared with that of
the eggs in the control set.

II. EXPERIMENTS ON UNSEGMENTED EGGS.

Experiment i . — On April 1 6, twenty-five unsegmented eggs
were subjected to a temperature of 28-30° C. for two and one-
half hours. When removed from the chamber, all of the eggs
were in the id-cell stage, while in the control set, developing at
room temperature, the eggs had only reached the 4-8 -cell stage.
The immediate effect of the higher temperature, therefore, was
to increase the rate of development. This result agrees fully
with that obtained by Hertwig in many of his temperature ex-
periments on the frog's egg. The later development of the
eggs in this series appeared to be perfectly normal, and it took
place at about the same rate as in the eggs of the control set.

Experiment 2.-- A number of eggs that had not yet seg-
mented were put into water at a temperature of 30-32° C. on
April 17. Part of the eggs were removed at the expiration of
three quarters of an hour, and when examined they were all
found to be segmenting. In a few cases the first cleavage plane
had nearly cut through the yolk portion of the egg and the
second furrow was appearing. In the control set of eggs, the
first cleavage plane was just coming in at this time, so that, in this
experiment also, the early development became more rapid as an
immediate result of exposing the eggs to a higher temperature.
All of these eggs developed into normal embryos.

Some of the eggs of the above lot remained in the heated
chamber for one hour. The second cleavage plane had ap-
peared in all of the eggs when they were removed to room tem-
perature. Later segmentation was normal, and on the following
day the dorsal lip of the blastopore appeared in all of the eggs
at about the same time that it formed in the eggs of the control
set. On April 1 9, many of the eggs were dead ; some were in
the early gastrula stages, and some showed traces of the medul-
lary folds. Of the seven embryos alive on April 20, three were
abnormal, having a large yolk plug exposed at the posterior end
of the body ; the other four embryos were normal and were kept
for several weeks.

22O HELEN DEAN KING.

The remaining eggs of this lot were kept at the temperature of
30-32° C. for one and one-half hours. At the end of this time
they were in the i6-cell stage, while the eggs of the control
set were only in the 2— 4-cell stage. Later segmentation of
these eggs seemed to be normal, and on April 18 the dorsal
lip of the blastopore appeared in a very few of them. On the
morning of April 19 most of the eggs were dead, and not one of
them, when examined, was found to have gastrulated. In the
eggs still living the blastopore was closing in, but development
was much slower than that of the eggs of the control set in
which, at this time, the blastopore had already closed and the
medullary folds were forming. All of the eggs were dead on the
morning of April 20, and in no case was gastrulation entirely
completed.

In these last two lots of eggs the injurious effects of heat
were not apparent during the segmentation stages and only
manifested themselves when the eggs were ready to gastrulate.
Early development was accelerated ; but later development
lagged behind, or, at most, was equal to that of the eggs in the
control set.

Experiment j. — A number of unsegmented eggs were exposed
to a temperature of 32° C. for two hours on April 22, and when
removed they were in the :6-cell stage. In this lot of eggs
the later cleavage was very abnormal as the upper hemisphere
divided into a number of small cells, while the lower part of
the egg segmented only a few times and, consequently, was com-
posed of a small number of very large cells. Cleavage lines
were very distinct in the upper part of the egg ; but it was almost
impossible to make out the boundaries of the yolk cells. None
of the eggs in this set gastrulated and all of them were dead by
April 24.

Experiment /. — On the morning of April 1 6, a small lot of
eggs was subjected to a temperature of 32—33° C. for one-half
of an hour. The eggs had not segmented when they were put
into cooler water, but in every case the first furrow appeared
in about fifteen minutes. In the control set, the first cleavage
plane came in about half an hour later than it did in the eggs
used for the experiment. All of the eggs of this set developed

EFFECTS OF HEAT ON TOADS EGG. 221

normally, and sections made of later embryos showed them to be
no different from the embryos of the control set.

Experiment 5. — A bunch of about seventy-five unsegmented
eggs was put into water heated to a temperature of 34—35° C. on
April 1 6. Part of the eggs were removed at the end of half an
hour and a few of them at once began to segment. None of the
cleavage planes, with the exception of the first, came in normally,
and in no case did any of them cut through the entire egg.
Part of a section of one of these eggs is shown in Fig. I. All
of the cleavage planes are seen to be parallel and to extend but
a short distance through the upper hemisphere of the egg. De-
velopment did not progress beyond this stage in any case, and
the majority of the eggs never segmented although they appeared
to be living several hours after they were brought into room
temperature.

Some of the eggs of the above lot remained at the temperature
of 34-35° C. for one hour. When put into cooler water and ex-
amined, a slight depression was found in the center of the upper
hemisphere of a few of the eggs as if the first cleavage plane was
about to appear in its normal position. This appearance, how-
ever, proved to be only a wrinkling of the surface as none of the
eggs, when sectioned, showed any true cleavage planes.

The above experiments show that the unsegmented eggs of
the toad can withstand a temperature of 32-33° C. for one-half
of an hour and develop normally, while an exposure to this
temperature for a longer period is very injurious and only a small
per cent, of the eggs produced normal tadpoles. Exposure to a
temperature of 34°, even for a short time, injures the eggs be-
yond the possibility of a recovery. The maximum temperature
that the unsegmented egg can endure without injury is, there-
fore, 33° C. The optimum temperature, a term defined by Hert-
wig (3) as, " Die Temperatur bei welcher sich der Entwicklungs-
process bei alien Eiren mit der grossten Beschleunigung ohne
eine auffallige Storung und Abweichung von der Norm vollzieht,"
for this egg is probably not far from 28° C., judging from the re-
sults obtained in experiments I and 2. In all cases in which the
heat did not kill the eggs, development was accelerated at first,
apparently with no injurious effects on the egg. In later stages,

222

HELEN DEAN KING.

S. C.

FIG. I. Part of a section of an egg that was subjected to a temperature of 34-35°C. for one-half of an hour before cleavage began.

FIG. 2. A section of an egg that was exposed to a temperature of 35-36° C. for three quarters of an hour when it was

in the two-cell stage. S. C., segmentation cavity.

FIG. 3. A section of an egg that was subjected to a temperature oi 33-35° C. for two hours after the first cleavage plane had
appeared.

FIG. 4. A section of an egg that was subjected to a temperature of 31-33° C. for three and one-half hours when it was in the
32-64-cell stage of development. Bl. blastopore.

EFFECTS OF HEAT ON TOADS EGG. 223

however, the eggs of the control sets appeared to be fully as far
advanced in development as were the eggs that had been subjected
to a higher temperature. Increase in the rate of development is,
therefore, but the immediate effect of heat, and after the eggs are
brought into a lower temperature they develop at the same, or a
lower rate, than the eggs of the control set.

III. EXPERIMENTS ON EGGS IN EARLY CLEAVAGE STAGES.

Experiment 6. — On April 17, a lot of about fifty eggs in the
2-cell stage of development was exposed to a temperature of
31-33° C. At the end of one and one-half hours, part of the
eggs were removed. They were then in the 8— i6-cell stage.
The later development of these eggs was perfectly normal in
every respect.

The rest of the eggs of this lot remained in the heated chamber
for two hours. All of these eggs developed normally during the
early cleavage and gastrulation stages ; but later a few embryos
were found with shortened medullary folds and a large yolk plug
at the posterior end of the body. This form of abnormality is
very common among embryos that have been injured by ex-
posure to heat.

Experiment 7. — As the first cleavage plane was appearing, a lot
of about fifty eggs was subjected to a temperature of 35-36° C.
for three quarters of an hour. All of the eggs were segmenting
in a very abnormal manner when they were transferred into water
at the room temperature, and none of them ever gastrulated.
Fig. 2 shows a median section through one of these eggs. With
the exception of the layer of small cells bordering the outer sur-
face of the upper hemisphere, the entire substance of the egg is
seen to be unsegmented and to have a number of different sized
vacuoles scattered through it. A large, irregularly shaped cavity
fills the greater part of the upper hemisphere of the egg. This
cavity is much larger than the segmentation cavity in a normally
segmenting egg, and it appears to be formed of the true seg-
mentation cavity and several large vacuoles which have come to
open into it.

Experiment 8. — On April 22, a lot of eggs in the 2-cell stage
was exposed to a temperature of 35-36° C. for one hour. When

224 HELEN DEAN KING.

removed from the chamber the eggs were in the 8-cell stage, but
development stopped at this point and all of the eggs were dead
inside of twenty -four hours.

Experiment p. — In this experiment, eggs in the 2- and in the 4-
cell stages of development, were subjected to a temperature of
3 3-3 5 ° C. for a period of two hours. At the end of this time the
eggs were segmenting very irregularly in the upper hemisphere
and no cleavage planes were visible in the yolk portion of the egg.
A section through one of these eggs (Fig. 3) shows the entire
upper hemisphere divided into a mass of small cells containing a
considerable amount of pigment which is, for the most part, col-
lected in the middle of the cell around the nucleus. The first
cleavage plane has cut only partially through the yolk portion of
the egg, as its progress was evidently stopped at the beginning
of the experiment. There are no nuclei in the yolk portion of
the egg, and the many vacuoles show the injurious effects of the
heat. The mass of small cells in the upper hemisphere forms
a sort of cap on the unsegmented yolk and make it appear
as if the segmentation of the egg was meroblastic. This same
sort of abnormal cleavage has also been obtained by Hertwig

(I, 2).

According to the experiments in this series, eggs in the early
cleavage stages can endure exposure to a temperature of 31-
33° C. for a longer period than can the unsegmented egg ; yet
they are permanently injured by even a short immersion in water
at a temperature of 35°. The maximum temperature for these
eggs, therefore, is not greater than that for the unsegmented egg.
Hertwig (4) has found that the maximum temperature for the
eggs of Rana fusca in the 8-cell stage of development is 26—28°,
which is 3—4° higher than that for the unsegmented egg.

IV. EXPERIMENTS ON EGGS IN LATE SEGMENTATION AND
EARLY GASTRULA STAGES.

Experiment 10. — On April 1 8, fifty eggs in the 32-64-cell
stage of development were kept at a temperature of 31-33° C.
for two hours. Subsequently all of the eggs developed into
normal embryos and at about the same rate as did the eggs of
the control set.

EFFECTS OF HEAT ON TOAD'S EGG. 225

Experiment n. — Another set of fifty eggs from the same bunch
as the eggs used in experiment 10, was subjected to a tempera-
ture of 3 1-33° C. for three hours. The late segmentation and
early gastrulation stages of all of these eggs seemed to be perfectly
normal. Two days after the experiment was made, 38 of the
eggs were dead, the blastopore not having closed in any case.
Of the remaining eggs four only were normal, the rest had a
large yolk plug at the posterior end of the body.

Experiment 12. - - Twenty-five eggs from the same lot as those
used in the two preceding experiments remained in water at a
temperature of 31-33° C. for three and one-half hours. Fifteen
of the eggs died in the blastula stage. The blastopore appeared
in the other ten eggs, but in many cases it was in an unusual
position at the equator of the egg. When the dorsal lip of the
blastopore was forming in these eggs, the circular blastopore
was already beginning to close in the control set of eggs, there-
fore, in this instance, the heat retarded instead of increased the
rate of development of the eggs. In none of the eggs of this
set did the blastopore ever become circular, and all of the eggs
were dead two days after the experiment was made.

Fig. 4 shows a section of one of these eggs preserved when
the blastopore appeared in surface view as a short, straight line
at the equatorial zone. The dorsal lip of the blastopore rarely,
if ever, comes in as high up as the equator in eggs that are de-
veloping normally ; but it sometimes occupies an unusual posi-
tion in eggs that have been subjected to abnormal conditions.
Morgan (5) has found the blastoporic rim above the equato-
rial zone in eggs of Rana pahtstris that have been subjected to
intense cold. In Fig. 4 the archenteron appears as a shallow
depression with its dorsal wall formed of heavily pigmented cells
as is normally the case. The inner end of the archenteron, in-
stead of turning up towards the black pole as it would do in a
normal egg, here projects downward towards the yolk pole.
The most interesting fact shown by the section is that the normal
position of the large and of the small cells of the egg is com-
pletely reversed. In normally gastrulating eggs, the roof of the
segmentation cavity is formed of two to three layers of small, pig-
mented cells, while the ventral wall is composed entirely of large

226 HELEN DEAN KING.

yolk cells that contain little, if any, pigment. In this egg, however,
the upper wall of the segmentation cavity is made up of a single
layer of heavily pigmented cells which are fully as large as any
other cells in the egg. Below the segmentation cavity, a por-
tion of the yolk is divided into a number of small cells, many of
which contain pigment massed around the nucleus. Some of
these cells are rounded and seem to lie free in the segmentation
cavity, an appearance also noted by Hertwig (4) in eggs oiRana
fitsca that were exposed to a temperature of 29—35° C. after
having reached about the loo-cell-stage of development.

Morgan has also noted the relatively large size of the cells in
the upper hemisphere of gastrulating eggs of Rana palustris that
had been subjected to cold. He suggests that this increase in
the size of the cells " may be due in part to the absorption of
water by the individual cells," and he adds that, " even if this is
the case the cells are fewer in number than in a normal egg
beginning to gastrulate." In the figure shown by Morgan,
the cells of the lower hemisphere are all considerably larger
than those of the upper hemisphere ; the egg, therefore, must
have been much more normal than the one from which Fig. 4
was drawn.

It is evident, in the case of the egg shown in Fig. 4, that the
increased temperature did not injure the yolk region or retard its
development as is usually the case in these experiments ; on the
contrary, it is the segmentation of the upper hemisphere that has
been delayed, while the segmentation of the lower portion of the
egg has continued. No egg in this set of experiments developed
much beyond the stage represented by Fig. 4, and each of the ten
eggs that were sectioned showed abnormalities of the same gen-
eral character.

Experiment ij. — On April 26, about seventy-five eggs in the
late blastula stage were subjected to a temperature of 33—35° C.
A part of the eggs were removed at the end of one and one-half
hours and they all developed into normal embryos.

A second portion of the eggs was exposed to this temperature
for two and one-half hours. All of these eggs developed into
normal embryos, although somewhat more slowly than did those
of the control set.

EFFECTS OF HEAT ON TOAD*S EGG. 22/

A third part of the eggs remained at the temperature of 33-35°
for three and one-half hours. These eggs were all dead when
removed from the influence of the heat.

Experiment /./. — A number of eggs in the blastula stage were
exposed to a temperature of 36-37° C. on April 26. Some of
the eggs were removed from the chamber at the end of one-half
of an hour. The eggs did not appear to be injured in any way
by the experiment and all developed normally.

A second portion of the eggs from the above lot remained at
this temperature of 36-37° C. for three quarters of an hour. All
of the eggs gastrulated normally, but about half of them died
before the blastopore closed. When sectioned these eggs showed
no abnormalities. The rest of the eggs became normal embryos,
although developing very slowly. The medullary folds had
closed in the eggs of the control set when they were only begin-
ning to unite in the eggs that had been subjected to the increased
temperature.

The remaining eggs of this lot were removed to room tem-
perature at the end of one hour. Although the eggs did
not appear to be dead when they were examined, they did not
gastrulate and none of them were alive the day following the
experiment.

Experiment 15. — Twenty eggs in late segmentation stages
were subjected to a temperature of 40-42° C. for one quarter of
an hour. Development was at once stopped by the heat, and all
of the eggs were killed.

Experiment 16. — When the dorsal lip of the blastopore was
just appearing, a lot of about twenty eggs was put into water
at a temperature of 33-35° C. and left there for three hours.
All of the eggs continued to develop somewhat more slowly
than the eggs of the control set and all became normal em-

O C3

bryos.

Experiment 77. --On April 24, a lot of eggs in early gastru-
lation stages was kept at a temperature of 35-37° C. for one
hour. In all of the eggs the lateral and ventral lips of the blasto-
pore formed in the normal manner, but development stopped at
this point and the eggs died. No abnormalities were detected
when sections were made of several of these eggs.

228 HELEN DEAN KING.

Experiment 18. — Eggs in early gastrulation stages were ex-
posed to a temperature of 37-38° C. on April 24. A part of
the eggs were removed at the end of one quarter of an hour.
None of these eggs seemed to be injured in any way by the high
degree of heat to which they had been subjected and all de-
veloped, somewhat slowly, into normal embryos. The rest of
the eggs in this lot remained at the temperature of 37-38° C. for
one hour. They were all dead when removed to room tem-
perature.

The results of the experiments in this series show that eggs in
the 32-64-cell stage cannot withstand a temperature of 31-33°
C. for a much longer period than can eggs that have just begun
to segment. The maximum temperature to which eggs can be
subjected without injury is practically the same for unsegmented
eggs and for those in early cleavage stages, although eggs in the
later stages can remain at this temperature for a somewhat longer
period and still develop normally.

Eggs in late cleavage stages have a much greater power to
withstand high temperature than have eggs in the earlier stages
of development, as they will develop normally after exposure to
a temperature of 36—37° C. for one-half of an hour. The maxi-
mum degree of heat that can be endured without injury is still
higher for eggs in the gastrula stages, as they become normal
embryos after being subjected to a temperature of 37-38° C. for
one quarter of an hour.

The experiments described above are summarized in the fol-
lowing table. The number of the experiment is given in the
first column ; the condition of the eggs when the experiment was
begun in the second column ; the temperature to which the eggs
were subjected in the third column ; followed in the next two
columns by the duration of the experiment and a brief statement
of the results.

The results of these experiments are very similar to those ob-
tained by Hertwig (1—4) in his study of the effects of heat on the
development of the eggs of various species of frogs ; and the
abnormalities produced resemble, in many respects, those which
Hertwig has described and figured. When the unsegmented
eggs of Bufo lentiginosus are subjected to a temperature that

EFFECTS OF HEAT ON TOAD S EGG.

229

TABLE I.

No. of Exp.

Condition.

Temperature.

Time.

Result.

I

unsegmented.

28-30° C.

2^hrs.

Normal development.

2

< <

30-32

# "

Normal development.

2

«

( <

I "

Four eggs developed normally ; the

rest died or became abnormal.

2

ft

it

2/2 "

Most of the eggs died in the bias-

tula stage ; a few gastrulated but

did not develop further.

3

«

32

2 "

All died in the blastula stage.

4

c C

32-33

X "

Normal development.

5

cc

34-35

X "

Irregular cleavage, no gastrulation.

5

( i

i t

I "

Eggs killed.

6

2 cell.

31-33

*l/2 "

Normal development.

6

«

t t

2

Most of the eggs developed nor-

mally.

7

C £

35-36

34: "

Abnormal cleavage, no gastrulation.

8

(I

1 t

I

Development stopped at the eight-

cell stage.

9

2-4 cell.

33-35

2 "

Abnormal cleavage, no gastrulation.

10

32-64 cell.

IT "» ••>
JA~JJ

2 "

Normal development.

ii

t «

t I

o (t
J>

Four normal embryos ; the rest of

the eggs died or became very

abnormal.

12

c<

«

3J^ "

All of the eggs became abnormal,

none of them developed into.

tadpoles.

13

Late seg.

33-35

i^ '

Normal development.

13

t <

tfc

2.y, '

Normal development.

13

1 <

t l

3/2 '

Eggs killed.

H

«

36-37

i/ *
t2

Normal development.

14

«

(t

% '

A few of the eggs developed nor-

mally, most of them died in the

gastrula stage.

H

«

C (

I

Eggs killed.

15

t (

40-42

% •

Eggs killed.

16

Early gas trula.

33-35

-» <

J

Normal development.

17

< t

35-37

I

Development stopped when the

blastopore was closing in.

18

1 1

37-38

I/ <

/T

Normal development.

18

1 (

& t

I '

Eggs killed.

stops their development before gastrulation begins, sections of
the eggs show, in many cases, that the greatest injury has been
produced in the yolk portion of the egg which is frequently
vacuolated and not segmented although the upper part of the
egg has divided into a large number of small cells. Hertwig
has noticed the same phenomenon in some of his experiments,
and in explanation he states as follows : " Dass Froscheier bei
erhohter Temperatur zunachst partiell geschadigt werden und
eventuell absterben, ist offenbar auf die verschiedene Organisa-
tion der animalen und vegetativen Halfte der Dotterkugel zuriick-

230 HELEN DEAN KING.

zufuhren. Die animale Halfte der Dotterkugel ist reicher an
Protoplasma und steht in hoherem Masse unter der Herrschaft
des Zellkerns. Unter der normalen Wechselwirkung von Pro-
toplasma und Kern konnen aber Schaden, welche eine Zelle
erlitten hat, vvie durch verschiedene Experimente festgestellt wor-
den ist, vvieder riickgangig gemacht werden. In dieser Bezie-
hung findet sich die vegetative Halfte der Eikugel unter ungiin-
stigeren Bedingungen. Denn hier ist das Protoplasma nicht nur
sparlicher zwischen den Dotterplattchen vertheilt, sondern ist
auch am ungethielten Ei mehr dem Einfluss des Zellkerns, der
in der animalen Halfte liegt, entriickt ; spater, nach Ablauf der
ersten Furchungsstadien sind die Theilstucke vielmals grosser
als die aus der animalen Eihalfte entstehenden Zellen."

When the injurious effects of the heat are not manifested until
the eggs gastrulate, Hertwig (3) finds, in Rana fusca, that the
abnormalities produced are of two sorts : First, those with a large
yolk plug in the posterior region ; second, those with deformed
heads. In all of my experiments on Bufo, the abnormal tad-
poles, with but very few exceptions, were of the first sort de-
scribed by Hertwig. In some cases the development of the eggs
stopped when the medullary folds were forming and a large yolk
plug was found in the mid-dorsal region ; in three cases only
was the defect in the anterior part of the embryo. My results
are more in accord with Hertwig's experiments on Rana cscit-
lenta than with those on Rana fusca, as in his experiments on
the former species he obtained a much smaller number of spina
bifida embryos than of those with a large yolk plug at the pos-
terior end of the body.

Hertwig (4) finds that the optimum temperature for the devel
opment of Rana fnsca is 20° C. for the unsegmented egg, and
that this optimum rises gradually to 24° C. for eggs in later
stages of development. He adds : " Offenbar hangt diese Er-
scheinung damit zusammen, dass mit der Vermehung der Zellen
die Kernsubstanz im Verhaltniss zum Protoplasma immer mehr
zunimmt und dass so das Protoplasma in hoheren Masse ihrem
Einfluss unterworfen ist." The optimum temperature for the
unsegmented egg of Bnfo lentiginosus is undoubtedly higher
than that for Rana fusca, and it is probably somewhere near 28°

EFFECTS OF HEAT ON TOAD'S EGG. 23 I

C. This optimum in increased 2-3° for eggs in later stages of
development.

In another set of experiments on Ranafusca, Hertwig (4) finds
that the maximum temperature to which the unsegmented eggs
can be subjected without suffering any injury is 23-24° C., while
this maximum is increased to 30° C. for eggs in the late segmen-
tation stages. The maximum temperature for unsegmented eggs
of Rana escitlenta Hertwig finds to be 33° C. This is also the
maximum temperature I have found for unsegmented eggs of
the toad, although eggs in the blastula stage can endure a tem-
perature of 38° C. for a very short time.

Morgan has noted that the blastula stages of Rana pahtstris
can endure extreme cold much better than can eggs in the 2—4-
cell stages, and he also finds that the eggs of Rana tcinporaria
which are laid very early in the spring, can survive the tempera-
of freezing water for several days. This temperature would very
soon kill eggs of Rana pahtstris which are deposited much later
than are the eggs of Rana tcinporaria.

While the eggs of all of these species of Amtra can withstand
a wide range of temperature without injury, there appears to be
an adaptation to temperature corresponding to the different
periods at which the eggs are deposited. Rana fusca and Rana
tcinporaria lay their eggs very early in the spring when the water
is often at the freezing point ; and the eggs of these two species
can stand cold much better than can the eggs of Rana pahtstris and
Rana esculenta which are laid considerably later. Although the
eggs of Bufo lentiginosus are laid but little later than are those of
Rana pahtstris, they are usually deposited in shallow pools of
water exposed to the direct rays of the sun. They must, there-
fore, often be subjected to a comparatively high degree of heat
during the course of their development.

BRYN MAWR COLLEGE,

BRYN MAWR, PA., April 24, 1903.

232 HELEN DEAN KING.

BIBLIOGRAPHY.

1. Hertwig, 0.

'94 Uber den Einfluss ausserer Bedingungen auf die Entwickelung des Froscheies.
Sitzber. der Kgl. Preuss. Akad. der "Wiss. Phys.-math. Abth., Bd. XVII.,
1894.

2. Hertwig, 0.

'96 Uber den Einfluss verschiedener Temperaturen auf die Entwickelung der
Froscheier. Sitzber. der Kgl. Preuss. Akad. der Wiss. Phys.-math. Abth.,
Bd. XIX., 1896.

3. Hertwig, 0.

'97 Ueber den Einfluss der Temperatur auf die Entwicklung von Rana fusca und
Rana esculenta. Archiv f. mikr. Anat. , Bd. LI., 1897.

4. Hertwig, 0.

'99 Ueber das Temperaturmaximum bei der Entwicklung der Eier von Rana
fusca. Cinquantenaire Soc. Biol. Paris, 1899.

5. Morgan, T. H.

'02 The Relation between Normal and Abnormal Development of the Embryo of
the Frog, as Determined by Injury to the Yolk-Portion of the Egg. Archiv
f. Entwickelurgsmech., Bd. XV., 1902.

ON FLOSCULARIA CONKLINI, NOV. SPEC., WITH A

KEY FOR THE IDENTIFICATION OF THE

KNOWN SPECIES OF THE GENUS.

THOS. H. MONTGOMERY, JR.

i. FLOSCULARIA CONKLINI, nov. spec.

Corona with five lobes, the dorsal largest, the ventral next in
size, the lateral very small. The lobes are broad, without knobs
and confluent at their points of insertion upon the corona. Vibra-
tile cilia, of a length not greater than that of the corona and some-
times considerably shorter, line these lobes in a single row, but
are not present between the lobes. Corona usually less than half
the length of the trunk, which is slender and not very sharply
demarcated from the foot. Foot fully two and a half times the
length of the rest of the body, terminating in a peduncle which
is as broad as long. Dorsal and lateral sense organs are present,
but no eyes in the adult. The body cavity is closely filled with
numerous minute floating corpuscles, so that the animal appears
dark by transmitted light. Tube large, gelatinous, usually with
foreign particles adhering to its surface. Length about that of
F. canipamdata Dobie. Two or three ova are frequently found
in the oviduct at once, and from thirty to forty male eggs within
the tube.

This species I found in considerable numbers in a pond on
the grounds of the University of Pennsylvania attached singly to
Myriopliylhun, during the early portion of 1903. It is a pleasure
to me to name it in honor of my friend, Prof. Edwin G. Conklin.
A full description of the anatomy with figures is reserved for an-
other paper upon the morphology of the Flosculariidae. The
new form differs from the closely related F. ainbigua Hudson in
the shortness of the cilia and their vibratile nature (they are not
stiff radiating setas) in the much greater length of the foot and
its very short peduncle, in the rather cylindrical and narrow
corona, in its smaller size, and particularly in its germarium being
rounded whereas mambigua I have found it to be elongate and bent.

233

234 THOS. H. MONTGOMERY, JR.

2. KEY TO THE SPECIES OF FLOSCULARIA.

All the species of F/oscnlaria (Oken, 1815) described up to the
year 1886 are described and figured in the monograph of Hudson
and Gosse. Of those described since that date I have seen the
descriptions of all except uniloba Wierzejski (1892), so that this
species could not be included in the present key. F. brachyura
Barrois and Daday (1894) is considered unrecognizable : their
figure represents a much contracted specimen, and the diagnosis
is simply : " Pede rudimentario, in aculeo curvato exeunti, urcello
nullo." But I differ from Rousselet (1893$) in regarding tenuilo-
bata Anderson as distinct from coronetta Cubitt. F. chimcera
Hudson is included, although it is probable this form will be
subsequently found to belong to another family of the Rotatoria.
Unless otherwise stated all the species entered will be understood
to be sessile and to possess a gelatinous tube.

I. Foot ending in two toes (pelagic ; no tube ; I dorsal eye ; corona with a smaller
ventral and a larger dorsal lobe which overhangs the corona).

L-hiwizra Hudson (1889).

II. Foot without toes or peduncle (pelagic ; foot very slender and whip-like ; a
single large dorsal coronal lobe and two smaller ventral lobes separated by a

very slight constriction) atrochoides Wierzejski (1893).

III. Foot terminating in a peduncle.

A. Corona without lobes.

a, I. Cilia short, in a single row (cilia mainly on dorsal and ventral margins of
the corona ; trunk much larger than corona and little shorter than foot ;
tube large) edentata Collins (1872).

a, 2. Short and vibratile cilia on outer coronal margin, and on 5 prominences of
the inner margin longer, non-vibratile cilia (pelagic ; tubes very slender).

pelagica Rousselet (18933).

B. Coronal margin produced into lobes.

a, I. Short cuticular spines on coronal margin (corona with 5 broad lobes, the

dorsal largest, all bearing long and stiff cilia) spinata Hood (1893).

a, 2. No cuticular spines on margin of the corona.

b, I. Corona with a single (dorsal) lobe (foot much enlarged near its pos-
terior end ; 2 eyes ; a tuft of long cilia upon the dorsal lobe and
shorter cilia upon the remaining margin of the corona ; pelagic).

libera Zacharias (1894).
b, 2. Corona with two lobes, a dorsal and a ventral.

c, I. Lobes short, corona little wider than the trunk (short, non-vibratile

cilia on the lobes only ; 2 cervical eyes) calva Hudson (1885).

c, 2. Lobes large, corona much wider than the trunk (vibratile cilia on the
whole margin of the corona ; 2 eyes near the summit of the dorsal

lobe) mutabilis Hudson (1885).

£>, 3. Corona with 3 lobes.

FLOSCULARIA CONKL1NI. 235

<r, I. Dorsal lobe with two long, flexible, non-ciliated processes (dorsal lobe
much the largest, overarching the corona ; short, vibratile cilia
fringing the whole coronal margin in a double row ; 2 eyes).

hoodii Hudson (1883).

f, 2. Dorsal lobe with 2 short, non-ciliated, horn-like processes on its dorsal
surface (dorsal lobe largest, overarching the corona; entire margin
of corona with an inner row of shorter and an outer row of longer

cilia; no eyes) cucullata Hood (1894).

c, 3. Dorsal lobe without any such processes.

d, I. Lobes small, the dorsal one not overarching the corona.

gosseii Hood (1892/7).
d, 2. Lobes large, the dorsal one overarching the corona.

e, I. Lobes bearing cilia on their tips only (3 rings below the corona).

anmtlata Hood (1888).

e, 2. Entire margin of corona with a double row of cilia (inner row of
short and outer of long cilia ; lobes deeply marginate).

trilobata Collins ( 1872).
b, 4. Corona with 4 lobes (each bearing a tuft of very long cilia).

quadrilobata Hood (1892^).
/', 5- Corona with 5 lobes.

c, I. Lobes very slender, longer than the whole trunk and nearly as long
as the foot (with long cilia on their lateral borders).

millsii Kellicott (1885).
c, 2. Lobes shorter than the trunk.

d, l. A. flexible, slender, non-ciliated process on the dorsal lobe (lobes

knobbed ) cornuta Dobie ( 1 849 ) .

d, 2. No such process on any of the lobes.

e, I. Dorsal lobe trifid at the tip (dorsal lobe much the largest, the
other lobes are slight projections of the coronal margin, and
none with knobs ; cilia rather short, limited to the lobes).

trifidlobata Pittock (1895).
e , 2. Dorsal lobe not trifid.

/, I. Peduncle about one third the length of the extended foot, flexi-
ble (lobes not knobbed, rather pointed, the dorsal the largest
and the lateral the smallest ; cilia long, non-vibratile, along

the whole coronal margin) longicaudata Hudson (1883).

f, 2. Peduncle many times shorter than the foot.

g, I. Lobes very slender, linear, composing almost the whole of
the corona (lobes knobbed, with long cilia on their ends
and short cilia elsewhere). ..fV«ttz70&7/V* Anderson (1890).
g , 2. Lobes not linear, rest of corona distinct.
h, I. Lobes knobbed.

z", I. Cilia along the whole coronal margin.
j, I. Lobes inserted on the coronal margin at some distance
from each other (lobes very mobile and shorter than
the diameter of the corona).

evansonii Anderson and Shephard (1892).
/, 2. Lobes confluent at their bases (fully as long as the
diameter of the corona, not mobile ; cilia long, non-
vibratile ; 2 eyes) coronetta Cubitt (1869).

236 THOS. H. MONTGOMERY, JR.

i, 2. Cilia limited to the knobs of the lobes.
/, I. Cilia longer than the whole animal, extensile and

very mobile mira Hudson (1885).

j, 2. Cilia not longer than the trunk, not clearly mobile.
k, I. Foot three times the length of the body (coronal
lobes very short ; 2 eyes)...fyclops Cubitt (1871 ).
k, 2. Foot barely twice the length of the body (coronal
lobes well developed ; no eyes).

ornata Ehrenberg (1830).
//, 2. Lobes not knobbed.

i, I. Dorsal lobe overarching the corona so that its cilia point
towards the foot (cilia non-vibratile, long).

torquilobata Thorpe (1891).
?',' 2. Dorsal lobe not overarching the corona.
_/, I. Cilia shorter than the corona, vibratile (cilia limited
to the lobes ; lateral lobes very small ; the others
somewhat triangular ; corona usually less than half
as long as the trunk ; peduncle very short).

conklini nov. spec.'
;', 2. Cilia longer than the corona, non-vibratile.

k, I. All five lobes distinct (corona as large as the trunk
with cilia on its whole margin ; peduncle long).
cawpanulata Dobie (1849).
k, 2. Lateral lobes small and indistinct.

/, I. Corona not ornamented with dots, tube distinct
(germarium elongate, extending down the left
side and across the whole diameter of the

venter) ambigiia Hudson (1883).

/, 2. Corona ornamented with dots in symmetrical pat-
terns, apparently no tube (living within an

algal growth) algicola Hudson (1886).

b, 6. Corona with 7 lobes.

c, I. Lobes not knobbed (long cilia around the whole margin of the corona).

diadema Petr (1891).
c, 2. Lobes knobbed (cilia restricted to these knobs ; 2 eyes).

recalls Hudson (1883).

FLOSCULARIA CONKLINI. 237

BIBLIOGRAPHY.
Anderson, H. H.

Notes on Indian Rotifers. Journ. Asiatic Soc. Bengal, Calcutta, 58, p. 345,
1890.

Anderson, H. H., and Shephard, J.

Notes on Victorian Rotifers. Proc. Roy. Soc. Victoria (n. s. ), 4, p. 69,
1892.

Barrels, T., and Daday.

Contribution a 1' etude des Rotiferes de Syrie et description de quelques especes
nouvelles. Rev. Biol. du Nord de la France, 6, No. 10, 1894.

Collins.

New Species of Rotatoria. Science Gossip, p. 9, 1872.

Cubitt, C.

Floscularia coronetta, a new species. Month, micr. Journ., 2, p. 133, 1869.

Floscularia Cyclops, a new species. Ibid., 6, p. 83, 1871.

Dobie, W. M.

Description of two new species of Floscularia with remarks. Ann. Mag.
Nat. Hist. (2), 4, p. 233, 1849.

Ehrenberg, C. G.

Beitrage zur Kenntniss der Organisation der Infusorien und ihrer geogra-
phischen Verbreitung. Abh. Akad. Wiss. Berlin, 1830.

Hood, J.

Floscularia annulata. Science Gossip, 1 888.

Floscularia quadrilobata, n. sp. Internal. Journ. Micr. (3), p. 26, 1892 (a),
Floscularia gosseii, a new Rotifer. Ibid., p. 73, 1892 (b).
Three new Rotifers. Journ. Quekett Micr. Club (2), 5, p. 281, 1893.
On Floscularia cucullata, sp. n. Ibid., p. 335, 1894.

Hudson, C. T.

Five new Floscules (Floscularia), etc. Journ. Roy. Micr. Soc. (2), 3, p. 161,

1883.

On four new Species of the genus Floscularia, etc. Ibid., 5, p. 608, 1885.

Hudson, C. T., and Goose, P. H.

The Rotifera ; or Wheel-Animalcules. London, 1886-1889.

Kellicott, D. S.

New Floscule (Floscularia Millsii). Proc. Amer. Soc. Micr. 8th Annual

Meet., p. 48, 1885.

Oken, L. v.

Lehrbuch der Naturgeschichte, 1815.

Petr, F.

Vernici (Rotatoria) vysociny ceskomoravske. Sitz.-Ber. k. Bohmische Ges.

Wiss. Prag., 2, p. 215, 1891.

Pittock, G. M.

On Floscularia trifidlobata, Sp. Nov. Journ. Quekett Micr. Club, 6, p. 77.

1895-

238 THOS. H. MONTGOMERY, JR.

Rousselet, C. F.

On Floscularia pelagica, n. sp., and notes on several other Rotifers. Journ.

Roy. Micr. Soc. , p. 444, 1893 (a).

List of new rotifers since 1889. Ibid., p. 450, 1893 (b).

Thorpe, V. G.

New and Foreign Rotifera. Ibid., p. 301, 1891.

Wierzejski, A.

Rotatoria (wrolki) Galicyi. Bull. Acad Cracovie, p. 402, 1892.

Floscularia atrochoides, n. sp. Zool. Anz., 16, p. 312, 1893.

Zacharias, 0.

Faunistische Mittheilungen, 2te Forschungsber. Biol. Stat. Plon., 1894.

UNIVERSITY OF PENNSYLVANIA, PHILADELPHIA,
May i, 1903.

Vol. V. October, ipoj. No. 5

BIOLOGICAL BULLETIN.

FORM REGULATION IN CERIANTHUS.

I. THE TYPICAL COURSE OF REGENERATION.
c. M. CHILD.

INTRODUCTION.

During the year 1902-1903 it was my privilege to spend sev-
eral months at the Zoological Station in Naples, as holder of
the Smithsonian table. I take this opportunity to express my
great indebtedness both to the Smithsonian Institution for the
grant and to Professor Dohrn and all other members of the staff
of the Zoological Station. A part of my time at Naples was
devoted to the study of regeneration and other regulative proc-
esses in the Cerianthidae, and an account of these observations
and experiments is begun in the present paper.

So far as I am aware the only work upon regulation in Cerian-
tlnis is that of Loeb.1 A review of this work is unnecessary at
this time since the various points will be discussed in connection
with my own observations as occasion arises.

My observations and experiments upon the Cerianthidae fall
into a number of groups, and, since they are somewhat extended,
the account of the subject will be divided in a corresponding
manner. In the present paper the usual "normal' course of
regeneration resulting in a perfect animal is described. Later
the problem of experimental control of regulation will be taken
up, then variation and abnormalties in regulation and the factors
concerned in their production.

THE NORMAL ANIMAL.

It is necessary to call attention to a number of the features of
the normal anatomy and habits before proceeding to the descrip-
tion of the regenerative phenomena.

1 Loeb, J., " Untersucbungen zur Physiologischen Morphologic der Thiere," I.,
Wurzburg, 1891.

239

240

C. M. CHILD.

Cerianthus solitarius, the species which formed the subject of
most of the experiments, is considerably smaller than C. inon-
branaccus. Owing to the varying degrees of distension and con-
traction accurate measurements of the form are difficult to obtain.
A considerable number of specimens were measured when
apparently fully extended and the body distended with water.
These were all among the larger specimens, for the smaller indi-
viduals were discarded in nearly all cases. These measurements
are of course only approximate and serve merely to indicate the
general proportions of the specimens used for experiment. Under
other conditions of contraction or distension these same indi-
viduals possess very different proportions. In all cases a single
individual was measured repeatedly at intervals and the maxi-
mum measurements taken as representing complete extension.
The foll'owing table presents a few such measurements of different
individuals, the measurements being given in millimeters :

Length of
Body.

Length
of Marginal
Tentacles.

Length of
Labial
Tentacles.

Diameter of
Disc.

Diameter of
Body in CEsoph-
ageal Region.

Diameter
of Body Near
Aboral End

go

30-35

12-15

12

7

7

95

25-30

12-15

12

7

5

60

2O

9-12

10

6

4-5

The specimens used were between these limits of size. A com-
parison of the measurements of the three individuals shows that
the smaller specimen possesses different proportions from the
larger, /. c., its transverse diameters are relatively greater as com-
pared with the length than those of the larger specimens. In
other words, after the individuals reach a certain size further in-
crease is chiefly an increase in length. Without giving the fig-
ures at this time to prove this point, since it will be taken up
later in connection with the discussion of morphallaxis, it may be
said that this difference in proportion between small and large
specimens is of general and probably universal occurrence in
Cerianthus. Smaller specimens are always relatively thicker
than large ones.

In general form the body is nearly cylindrical, expanding or-
ally to form the disc and tapering slightly posteriorly. At the
aboral end is a small pore which under certain conditions permits

FORM REGULATION IN CERIANTHUS. 24!

the exit of water. In the expanded condition the disc possesses
the form of a broad shallow funnel extending from the base of
the marginal tentacles to the margin of the mouth and continued
aborally in the oesophagus. The mouth is slit-like in form with
one siphonoglyph or gonidial groove at one end of the slit.
The disc is marked with radiating lines, slightly depressed, which
correspond to the lines of attachment of the mesenteries beneath
the surface : these continue aborally in the oesophagus. The
oesophagus extends aborally from the disc about -*— 1 the length
of the body when the animal is fully extended.

The marginal tentacles, as their name implies, are borne upon
the margin of the disc, usually in about three rows, the number
varying in grown specimens from about 41 to 71. About the
margins of the mouth are the shorter labial tentacles which are
fewer in number than the marginal tentacles, and form only a
single circle.

The body appears brownish in color, but upon close examina-
tion is found to be marked with light longitudinal stripes or lines
of varying width, some of which extend the whole length of the
body while others are shorter. These are in reality merely
unpigmented areas between the stripes of brown pigment. The
color of the marginal tentacles is in general effect lighter than
that of the body, but they are marked by transverse bands of
dark pigment. The labial tentacles are brownish and usually
unstriped. The disc and oesophagus in large, apparently old
specimens are dark brown without definite striping.

As regards the internal anatomy certain points are of interest
in this connection. It has long been known that the arrange-
ment of the mesenteries in the Cerianthidse differs in some re-
spects from that in the other Actinozoa. In the cesophageal
region all mesenteries extend from the body-wall to the oesopha-
gus and thus divide the enteron of this region into a series of
longitudinal radiating chambers which open into the enteron
aborally. At the oral end each of these intermesenterial cham-
bers opens into the cavity of a single marginal tentacle ; thus
the marginal tentacles are always equal in number to the inter-
mesenterial chambers. The labial tentacles, while corresponding
in position to intermesenterial chambers, are fewer in number.

242 C. M. CHILD.

Abpral to the cesophagus the inner margins of the mesenteries
hane free in the enteron and bear the mesenterial filaments. A

o

single pair of very short mesenteries at that end of the mouth
where the siphonoglyph is situated are known as the directives.
The next mesentery on each side of these extends almost to the
aboral end of the body. From this point to right and left the
mesenteries decrease in length, following a definite, rather com-
plex law which need not be discussed here. On the side oppo-
site the directives, at the opposite angle of the mouth are the
shortest mesenteries, with the exception of the directives ; these
do not extend far aboral to the oesophagus. It is in this region
that all new mesenteries are added, /. e., the region of growth is
opposite the directives. Thus, proceeding from the directives to
the right and left around the body the mesenteries are succes-
sively younger. Each pair of new mesenteries appears between
the members of the last preceding pair formed, thus separating
them. Corresponding to the formation of new intermesenterial
chambers new tentacles appear in this region. It is usually pos-
sible to find at this point in the normal animal one or two pairs
of tentacles much smaller than the others and in process of
growth. Corresponding to the number and arrangement of the
mesenteries there is one unpaired marginal tentacle over the
chamber between the directive mesenteries and known as the
directive tentacle. It is usually somewhat thicker than the other
tentacles since the space between the directives is greater than
that between other mesenteries. The other tentacles are paired
right and left.

In Ccrianthns solitarius the greater number of the mesenteries
about the whole circumference of the body do not extend
aborally far beyond the cesophagus. Only certain mesenteries
extend further, to end at various levels according to their posi-
tion. This is also true of Cerianthus lucmbninaccits.

The muscles of the body-wall consist of a heavy layer of
powerful longitudinal muscles which decreases slightly in thick-
ness toward the aboral end. These are the chief muscles of the
body, circular muscles being absent, and tentacles, disc and
cesophagus possessing only a slight muscular development.

As is well known, the Cerianthidae are found imbedded in the

FORM REGULATION IN CERIANTHUS. 243

sand with the oral end and tentacles protruding. In this position
they secrete about the body a mass of tenacious slime in which
sand-grains and other foreign bodies become imbedded, the whole
forming a tube into which the animal may withdraw. Loeb has
given an interesting account of the geotactic reactions of these
animals and my incidental observations upon this point confirm
his. He has also described a number of experiments concerning
the external conditions which determine the tube-formation.

When specimens are kept in aquaria without sand they creep
about to a considerable extent, often climbing the sides. When
left undisturbed they usually orient themselves as Loeb has noted,
so that the oral end of the body is directed upward, even if this
position necessitates the bending of the body at right angles. In
the jars they secrete a considerable amount of slime and often
form tubes along the sides or bottom, in which they remain for a
longer or shorter time. When handled or otherwise irritated, and
especially when cut, the secretion of the slime is especially,
rapid.

When undisturbed, the body and tentacles are usually more or
less distended with water and the body-wall is always tense. In-
deed, as will be shown later, complete extension of the body and
erection of the tentacles is impossible without internal water-
pressure, /. c.t without water in the enteron. If the body of a
distended animal is opened quickly by a small cut the water
issues with considerable force, and when an individual is made to
contract rapidly by sudden stimulation the water squirts from the
aboral pore with great force. The inability of the animal to ex-
tend to its full length without the aid of water-pressure is due to
the absence of circular muscles in the body-wall. Extension is
passive, not active. The longitudinal muscles are powerful and
under strong stimulus the body may be torn apart if the ends are
fastened.

It was found that the animals could be kept alive for months
without other food than the small forms and organic particles
which the water might contain, and in the present series of ex-
periments no attempt was made to give them food. Of course in
the early stages of regeneration and throughout many of the ex-
periments the pieces were unable to take food ; moreover, the

244 c- M- CHILD.

growth resulting from an abundant food supply constitutes in any
case a complicating factor in the analysis of various phenomena
of form regulation. In such experiments as permit the taking of
food a complete analysis of the phenomena would include studies
of the effect of abundant food supply, but previous experiments
along this line indicate that the results in the lower animals differ
only in degree and not essentially in kind with the presence or
absence of food.

Four species of Ccrianthus were employed for experiment, viz.,
C. solitariits, C. membranaceus and two smaller undetermined
species, one of them almost completely colorless. It was soon
discovered, however, that C. solitarins, a very common form in
the Bay of Naples, was more favorable than the other forms on
account of size, coloration of body and abundance. My atten-
tion was therefore devoted chiefly to this species, though the
other forms, and especially C. membrdnaceus, were used for com-
parative study.

THE COURSE OF REGENERATION.

The cut pieces were isolated in dishes which were placed in
aquaria supplied with flowing water. During the earlier stages
of regeneration the pieces showed little tendency to creep out of
dishes, but later' it was necessary in some cases to cover the
dishes with netting to prevent escape.

In Cerianthus the course of regeneration is complicated in
many cases by various factors, such as the form of the pieces,
the internal water pressure, etc. The simplest cases are those in
which the body is divided by a transverse cut into two pieces, or
a piece is removed by two transverse cuts. In such cases a
nearly cylindrical piece is obtained which regenerates at the cut
end or ends. Such pieces are best fitted for the study of
the "typical1 course of regeneration at the two ends, and
since a knowledge of this is important as a -preliminary to the
study of experimental control of regeneration this paper is de-
voted to a description of the phenomena concerned in such
cases.

A piece cut from the middle region of the body (e. g., between
the lines aa and bb, Fig. i) will serve as an example.

FORM REGULATION IN CERIANTHUS. 245

THE IMMEDIATE EFFECTS OF THE OPERATION.

Individuals which were in good condition and well extended
were chosen and the cuts were made rapidly with sharp scissors.
All parts of the body contract strongly in consequence of the
cut, and of course total collapse of the piece occurs, owing to
the escape of the water from the enteric cavity. Within a few
moments the piece may relax somewhat from the extreme condi-
tion of contraction, but does not attain anything like its original
length. Placed in the jar it lies on the bottom, and the weight
of the tissues causes it to become more or less flattened. The
piece has no power to retain its cylindrical form, though the
mesenteries and mesenterial filaments, especially in pieces cut
from the oral half of the body, partly fill the enteron and so
cause the piece to retain a more or less rounded form. The
body-wall is opaque in these pieces, while in normal specimens
distended with water it is slightly translucent. The opacity is
due simply to its greater thickness in the absence of the tension
caused by internal water-pressure.

Within a few moments after section the cut edges at the two
ends of the piece begin to bend or roll inward, and in an hour or
two this inrolling has proceeded so far that the cut edges are no
longer visible from the ends and the opening is almost completely
closed by the inrolled portions. In Fig. 2 a longitudinal section
through the oral end of such a piece is shown, the ectoderm and
entoderm being indicated by full black lines and the thick mus-
cular layer by fine lines. In this and following figures of the
same kind the mesenteries are not shown ; they of course occupy
practically the whole of the enteric cavity after collapse. A sec-
tion through the aboral end shows conditions similar to those
figured in Fig. 2.

In consequence of the infolding about the whole circumference
of the cut ends the circumference of the body-wall in the infolded
region decreases greatly, although the transverse contraction of
the body- wall during the infolding is not marked. It is, therefore,
thrown into numerous longitudinal folds and ridges at the edge,
and these appear when the piece is viewed from the end as folds
and ridges radiating from what remains of the central opening.

246

C. M. CHILD.

Fig. 31 shows the oral end of a collapsed piece in which in-
folding has occurred. The numerous radiating foldings of the
body wall are evident. Figs. 4 and 5 show the aboral ends
of similar pieces. By this unfolding of the cut edges the open-
ings at the ends of the piece are reduced to slits as is seen from

the figures, and various parts of the circumference of the cut
edge are approximated, though actual contact between parts of
the cut edge cannot occur everywhere, owing to the irregular
wrinkling of the margin as it folds inward. Indeed, since the
margin does not contract transversely to any great extent as
the infolding occurs, actual contact of all parts of the cut edge
is a physical impossibility, as it could occur only by the reduc-
tion of the cut margin to a point at the center of the circle formed

1 In Figs. 3, 4, 5, 6, 7, 8, 9, 10, 12 the longitudinal pigmentation of the body is
not shown. The new tissue, where present, is indicated by stippling.

FORM REGULATION IN CERIANTHUS. 247

by the body-wall. In most cases, however, as the collapsed,
more or less flattened piece lies on the bottom of the jar the in-
folding edges come into contact along the longer margins as in
Figs. 3 and 4, leaving an elongated slit between them. In other
cases the closure may occur as shown in Fig. 5. In general the
form of the end depends wholly upon physical conditions and
especially on the form of the transverse section of the piece after
collapse.

The infolding of the cut margins is undoubtedly the result of
mechanical conditions, though these conditions may themselves
be in part reactive in nature. As Loeb has pointed out, an in-
folding must occur if the inner portions of the body-wall are
under greater longitudinal tension than the outer portions. Such
a condition may possibly be produced in the muscles near the
cut, the inner layers undergoing greater contraction than the
outer, but the elasticity of the fibrillar mesogkea is probably in
part responsible. As will be shown later this infolding produces
in many cases conditions from which a return to the normal form
is impossible. It can scarcely be regarded, therefore, as an
adaptive reaction in the stricter sense. The radiating wrinkles
and folds upon the end are due simply to the fact that the cut
edges do not contract transversely as they are folded in.

As already noted, the result of the infolding is to close the
terminal openings more or less completely. The closure is in
no case perfect since between the irregular wrinkles there are
always numerous interstices which afford communication between
the enteron and the exterior. In most cases, however, these
are soon blocked by the tenacious slime secreted by the ectoderm
and are also frequently more or less completely filled by portions
of the mesenteries or the filaments which happen to extend
through them from within.

THE CLOSURE OF THE ENDS.

The histological changes about the cut margins have not been
fully investigated as yet, but it has been determined that growth
of new tissue upon the edgesvbegins soon after the cut is made.
If after one or two days the infolded end be opened and carefully
spread apart a very thin and delicate whitish membrane of new

248 C. M. CHILD.

tissue will be found extending across parts of the opening.
While growth undoubtedly begins on all parts of the cut surface,
this membrane becomes distinct earlier at those regions where
the cut edges are most closely approximated. Frequently when a
piece is opened in the manner described the membrane will be
found extending across regions corresponding to certain of the
wrinkles about the opening but not yet covering the central area.

This method of formation of the thin membrane closing the
end is well shown in a piece cut from a specimen of C. incni-
branaccus. In this species the body-wall is so thick and stiff
and the diameter of the body so great that in short pieces the
infolding of the ends is often not sufficient to close the opening.
In Fig. 6 a piece of this kind is shown. The new tissue first
became evident along the fold a, and a day or two later a thin
membrane was spread across this fold (Fig. 7, a. The new tis-
sue is stippled). A little later still the fold at b (Fig. 7) also
showed a thin membrane (Fig. 8), which, however, was after-
ward ruptured by contractions of the piece due to the stimulation
incidental to examination. In Fig. 8 it is seen that the new
tissue is gradually spreading over the opening from a. In Fig.
9 the opening is nearly closed. Several days later closure was
complete. The changes in form of the piece as shown in the
figures were the result of stimulation caused by the manipulation
necessary for examination and drawing. In C. soliianns if the
pieces are allowed to remain undisturbed at ordinary summer
temperature the openings at the ends are usually completely
closed by the thin membrane on the third day after operation.
In the piece from C. membranaceus above described closure was
complete after twenty-seven days. In general this species re-
generates much more slowly than C. so/it arias, but here the
closure was exceptionally slow.

The membrane is easily ruptured by the contractions of the
piece when strongly stimulated and great care is always necessary
in the examination of such pieces to prevent rupture. In conse-
quence of contraction the different parts of the margin change
their relative positions or the mass of the mesenteries and fila-
ments exerts pressure from within, thus readily causing rupture.

There is little difference as regards time of closure between the

FORM REGULATION IN CERIANTHUS. 249

two ends, though in general the oral end is slightly in advance of
the aboral end.

DISTENSION OF THE PIECES WITH WATER.

The piece remains completely collapsed during two or three
days in summer and five to six days in winter, and then gradu-
ally becomes distended. At this time the piece is completely
closed at both ends, no mouth or aboral pore being present. It
is probable that the accumulation of water in the enteron is the
result of diffusion through the walls, and especially through the
very thin membranes at the two ends, in consequence of the
accumulation of soluble products of metabolism in the closed
enteron.

In the course of a day or two the piece becomes well filled
with water and attains a degree of distension approaching that of
the normal animal, though not as great. In some cases the ac-
cumulation of water in the enteron occurs so rapidly that the thin
membranes closing the ends are ruptured and collapse occurs
again, though usually the increase in thickness and strength of
the membrane is sufficient to prevent rupture. The piece is
usually well filled with water by the fourth day in summer and
usually by the seventh or eighth in winter.

The immediate result of the renewed distension of the piece
with water is, of course, the resumption of the cylindrical form ;
the body wall becomes translucent and is elastic to the touch
like that of the normal animal.

The most marked effect of the internal water pressure occurs,
however, at the ends of the piece. So long as the piece re-
mains collapsed the thin membrane closing the ends is not visible
since the infolded edges of the body-wall are in close contact.
As the body becomes distended with water, however, the in-
folded portions gradually spread apart and a central area cov-
ered by the new tissue becomes visible. Very small at first, it
gradually increases in size until its diameter is about one third
the diameter of the end. Fig. 10 shows the oral end of a piece
at about this stage. The area within the folded margin of the
old body-wall is covered by the thin membrane of new tissue. In
Fig. 12 the aboral end of a similar piece is shown. There is

250

C. M. CHILD.

little difference between the two ends, except that growth of the
membrane is more rapid at the oral end. In Fig. 1 1 1 a portion
of the oral end is shown more highly magnified. In this case
the abrupt transition from the pigmented body-wall to the almost
colorless new tissues is evident. From this figure it is also seen
that the margin of the old body-wall is somewhat crenated by

fine folds and wrinkles, which, however, are not regular in size
and form, and do not represent the early stages of the new ten-
tacles. The slight folds indicate more or less exactly the re-
gions where the mesenteries are attached and the bulging areas

1 In Figs, ic, 15, 16, 18 the pigmentation of the body- wall is indicated.

FORM REGULATION IN CERIANTHUS. 251

between the intermesenterial chambers, these being now filled
with water and under pressure. Here and there, however, folds
without such significance occur, and moreover some of the
chambers are so situated on the infolded margin that they are
more widely open and thus expand more in consequence of the
pressure than others, hence the irregularity in form and size of
these crenations.

In Figs. 13 and 14 are shown resoectively the oral and aboral
ends of the body-wall at the stage where the infolded portions
begin to separate. The thin membrane closing the end is shown
as a black line. It consists, of course, of ectoderm and ento-
derm, but the muscular layer does not extend into it.

THE FORMATION OF THE MARGINAL TENTACLES AND Disc.

Within the first day or two following the closure of the ends
and the distension of the piece with water the changes leading
to the formation of the characteristic organs of the oral end be-
gin. In pieces cut from the middle region of the body the full
number of mesenteries is not present, since some end anterior to
this region. Regeneration of mesenteries occurs, though the
number of mesenteries in a regenerated oral end from the middle
region of the body is somewhat less than the number originally
present at the oral end of the individual from which the piece is
taken. This point will be considered at another time. It is suffi-
cient for the present purpose to say that the whole oral end of
the piece becomes divided into intermesenterial chambers, in the
manner characteristic of the species, by the regeneration of new
mesenteries, at first very short, between the longer mesenteries
which are present in the piece. Attention was called above to
the crenation of the infolded margin in correspondence with the
position of the mesenteries (Fig. I i).

The first marked change following the closure of the end is
the appearance of a slight ridge on the infolded margin of the old
body-wall as shown in Fig. 15. The ridge is wholly confined to
the tissue of the original body-wall, the thin membrane which
closes the end playing no part in its formation. The crenations
become more distinct and extend in many cases from the margin
of the old body-wall over the ridge, as the regeneration of the

252 C. M. CHILD.

mesenteries beneath advances. In Fig. I 5 the ridge is shown as
slightly lighter in color than the rest of the body. The pigmen-
tation is beginning to disappear. Most of the stripes can still be
followed over the ridge to the margin of the old tissue, but upon
the ridge they are fainter than before. Fig. 16 shows a portion
of the end at a slightly later stage, more highly magnified. Here
the lighter color of the ridge is more distinct. While the body
in general retains its brown color the ridge becomes light yel-
lowish and its pigment disappears completely in the course of a
day or two.

This change in pigmentation indicates that some alteration in
the tissues is occurring, and the nature of the alteration becomes
evident when a longitudinal section through the end (Fig. 17) is
examined. This figure shows that the thickness of the body-
wall and especially of the muscular layer is decreasing consider-
ably in the region corresponding to the ridge. This decrease is
shared to a certain extent by the ectoderm and entoderm as the
figure indicates. The new regenerating mesenteries are minute
folds in the infolded region, ending free aborally (;;/, Fig. 17).

This ridge in which loss of pigmentation and reduction in thick-
ness of the body-wall are taking place may be designated as the
marginal tentacular ridge, since it is from this that the marginal
tentacles arise ; indeed the reduction in thickness of the body-
wall and the division of the ridge into areas corresponding to
the intermesenterial chambers are the preliminaries of tentacle
formation.

The marginal tentacles do not arise from the cut edge of the
body-wall itself but a short distance away from it, viz., at the
highest point of the ridge (/, Fig. 16), /. e., entirely within that
portion which was originally part of the body-wall and not in
the new tissue which closes the end.

Fig. 1 8 shows the oral end of a piece about a day later than
the stage shown in Figs. 15 and 16. Here the new marginal
tentacles are distinct and are evidently increasing in length. The
pigment has disappeared completely from the tentacular ridge
which is now whitish in color and distinctly translucent. Some
of the tentacle buds are slightly broader than others owing to
the fact that in the infolded condition of the margin some inter-

FORM REGULATION IN CERIANTHUS. 253

mesenterial chambers were compressed and others stretched
according to their position on the folds. There is, however, no
marked difference in the length of the new tentacles on the dif-
ferent sides of the body, those in the region of the directives
being no more advanced than those in the growing region oppo-
site. From this figure it is very evident that the marginal ten-
tacles arise from the highest, i. t\, the most oral point of the
tentacular ridge. Moreover they arise in a single circle or row,
although in the normal animal they occur in about three concen-
tric circles.

A longitudinal section of the body-wall at this stage is shown
in Fig. 19. A comparison with Fig. 18 shows marked changes.
The most conspicuous of these is the continued reduction in
thickness of the body-wall upon the ridge. The muscular layer
has almost or quite disappeared in this region and also between it
and the new tissue occupying the central region of the end, and
is reduced considerably in thickness for some distance aboral to
the ridge. At this stage then the whole oral end is closed by a thin,
unpigmented, translucent membrane consisting of ectoderm and
entoderm, but without a distinct muscular layer. The central part
of this membrane resulted from the growth of new tissue at the cut
edge, while the more distal portions forming the tentacular ridge
have arisen by the transformation of a part of the old body- wall
into tissue capable of a large amount of new growth, and of dif-
ferentiation into new structures. In other words, the body-wall
in this region has changed from its differentiated condition to
what is commonly called the embryonic condition. The histo-
logical features of this change are of great interest, but will be
described at another time.

The marginal tentacles now grow rapidly, and in another day
(six days after the operation in summer) the oral end presents the
appearance shown in Fig. 20. Several changes of importance
have occurred since the stage of Fig. 18 : the disc is greatly ex-
panded, the marginal tentacles are much longer, the distinction
between the tissue of the old body-wall and the thin membrane
closing the end has disappeared completely, and finally the
mouth is beginning to appear as an opening between the center
and the periphery of the disc in the directive radius. The disc is

254

C. M. CHILD.

marked with radiating lines, each of which terminates distally
between the bases of two tentacles ; those lines are in reality
grooves and mark the lines of attachment of the mesenteries to
the aboral surface of the disc. It will be seen that a small area

24

in the center of the disc, indicated iirthe figure by stippling, is free
from these lines ; this represents that portion of the thin mem-
brane beneath which regeneration of the mesenteries has not yet
occurred. In the directive radius is situated a small opening,
the new mouth, which gradually elongates in the directive plane.

FORM REGULATION IN CERIANTHUS. 255

The directive tentacle, which corresponds to the chamber between
the two directive mesenteries, is slightly thicker than the other
tentacles in consequence of the fact that the directive mesenteries
are somewhat farther apart than other mesenteries. As regards
the arrangement of the marginal tentacles it will be seen that they
are no longer in a single row, but some appear as if they were being
crowded out of the row, owing to lack of space. In all proba-
bility that is what is occurring. As the tentacles increase in size
there is not sufficient space for them in a single row upon the
margin and some are pushed out, probably in most cases
peripherally.

Fig. 21 is a schematic figure of one half the body after longi-
tudinal section in the directive plane, the directive tentacle being
on the left of the figure. The stage of regeneration is about the
same as that of Fig. 20. As compared with the earlier stages
(e. g., Fig. 19) several points of difference are to be noted : the
marginal tentacles are longer, the difference in thickness between
the reduced body wall of the tentacular region and the thin new
tissue across the disc has completely disappeared ; the reduction
and disappearance of the muscular layer extends further aborally
than before ; the regeneration of the mesenteries has advanced ;
and finally there is a minute mouth, which, as was evident from
Fig. 20, is not centrally placed, but lies near the base of the
directive tentacle.

One or two points of importance as regards the regeneration
of the mesenteries may be noted. In the normal animal a slight
furrow, which appears as a faint longitudinal line on the surface
of the entoderm, extends aborally from the aboral end of each
mesentery. In a piece cut from the middle region many of the
mesenteries lie wholly oral to the cut and so are not present in
the piece, but most of the furrows, extending aborally, are visible '
in the piece. The mesenteries regenerate along these furrows.
Whether regeneration of a particular mesentery aboral to the
end of the furrow representing that mesentery is possible has
not yet been determined. The point to which it is desired to
call attention here is that the mesenterial regions are determined
for some distance posterior to each of the mesenteries them-
selves.

256 C. M. CHILD.

Those mesenteries which extend into the piece undergo re-
gressive changes, losing their thickened margins and filaments at
the oral end, and become united with the new oesophagus.

THE APPEARANCE OF THE LABIAL TENTACLES AND THE LATER
STAGES OF ORAL REGENERATION.

The marginal tentacles continue to increase rapidly in length
and the oesophagus extends further across the disc from the di-
rective side and also becomes deeper.

Fig. 22 is drawn from a stage three days later than Fig.
2O (nine days after the operation). Comparison of this figure
with Fig. 20 shows at once the increased diameter of the disc,
the greater length of the tentacles, and the marked change in the
size and shape of the mouth opening. The tentacles in Fig.
22 are still of about equal size and length, except the directive
tentacle, which is somewhat thicker and longer than the others.
They still retain, to a large extent, the arrangement in a single
row, though here and there a few have been forced out of line.

Upon the disc and forming a circle about the mouth appear
the earliest stages of the labial tentacles. They are at this time
mere buds, less than one half millimeter in length. All appear
nearly simultaneously and develop with equal rapidity. As
noted above, they are fewer in number than the marginal ten-
tacles, some of the intermesenterial chambers possessing none.

A view of half the oral end at the stage of Fig. 22 after longi-
tudinal section in the directive plane is shown in the schematic
Fig. 23. In this case the plane of section passed through one
of two small tentacles in the growing region opposite the direc-
tive tentacle ; the section of this tentacle (on the right of the
figure) is thus considerably smaller than that of the directive
tentacle opposite. Comparison of Figs. 21 and 23 shows the
changes which have occurred during the three days elapsing be-
tween the two stages. The cesophageal invagination is much
deeper, the opening to the enteron is larger and the area of
growing tissue, including the reduced body-wall, is much greater.
From this time on the course of regeneration consists in the
gradual increase in size and the pigmentation of the regenerated
parts in the manner characteristic of the species.

FORM REGULATION IN CERIANTHUS. 257

The problem of " morphallaxis," /. c., the changes in the pro-
portions of regenerating pieces leading to the more or less com-
plete reestablishment of the "normal" form will be considered
elsewhere.

Fig. 24 shows a regenerated disc and tentacles at a later stage ;
in form and general arrangement of parts it is not distinguishable
from the normal animal. The marginal tentacles have not yet
fully attained their final arrangement ; at present they are in two
fairly well marked rows or circles. During the still later stages,
however, as further increase in size occurs, the bases of some are
forced still farther peripherally and so the characteristic arrange-
ment of tentacles is finally acquired. The pigmentation of the
marginal tentacles with dark transverse bands, which appears at
this stage or earlier, is not shown in the figure.

THE DIFFERENTIATION OF THE ABORAL END.

The infolding of the body-wall and the closure of the aboral
end of a piece by a thin membrane have already been described.
It remains to describe the formation of the characteristic aboral
end. The course of regeneration here is much simpler than at
the oral end.

The first marked change from the condition shown in Fig. 14
consists in the protrusion in conical form of the thin membrane
closing the end (Fig. 25). About the margin of this new tissue
the slightly wrinkled margins of the old body-wall are still
clearly marked. In Fig. 26 a longitudinal section of the aboral
end at this stage is shown. The absence of the aboral pore is
to be noted. This new outgrowth at the aboral end does not
become well-marked at once after closure, but only after the
piece is well filled with water and the regeneration is advanced
at the oral end, /. c., it is much slower than oral regeneration.

In Fig. 27 the aboral outgrowth is seen at a somewhat more
advanced stage. The wrinkles and folds upon the margin of the
old tissue are gradually disappearing as this stretches and under-
goes remoulding. A few days later the wrinkles have disap-
peared and there is no sharp distinction between the old body-
wall and the new tissue at the time of union. Fig. 28 shows
the end at this stage; and it is evident that the margins of the

258

C. M. CHILD.

old body-wall are becoming involved in the regenerative changes
in the same manner as at the oral end, for the pigment stripes
are gradually fading out in the region which was before infolded.
Fig. 29 shows a still later stage in which the gradual fading of
the pigment-stripes is clearly seen. The significance of this loss
pigment is made clear by Fig. 30, a longitudinal section of the
aboral end at this stage. Here it is seen that a reduction of the
muscular layer is occurring, /. c., the old body-wall is becoming

27

28

29

involved in the regulative changes for a short distance oral to the
cut end : In other words the new aboral end is formed not
merely from the new tissue which closes the end soon after op-
eration, but, as in the regeneration of the oral end, in part from
tissue derived from the margins of the body-wall near to the cut
surface, by reduction of the muscular layer and growth of the
ectoderm and entoderm. Thus the distinction between " old
tissue" and "new tissue," at first well-marked, gradually disap-
pears in this region.

FORM REGULATION IN CERIANTHUS. 259

Fig. 3 1 shows a still later stage in which the new tissue is
becoming pigmented. The appearance of the pigment corres-
ponds in time with the differentiation of the muscular layer, and
I am inclined to believe that in CcriantJins as in various other
forms, the pigmentation of the body is closely connected with
the presence and arrangement of the muscular layer.

The course of regeneration described in the present paper is
characteristic of pieces cut from the middle half of the body. In
following papers the regeneration of pieces from various regions
will be compared, and experiments determining some of the fac-
tors concerned in regeneration will be described.

SUMMARY.

1. In cylindrical pieces of Ccriautliits obtained by two trans-
verse cuts collapse occurs at once and the cut ends begin to roll
inward soon after section, finally coming into contact and closing
the opening more or less completely. Since little or no transverse
contraction of the infolded margins occurs they are thrown into
numerous radiating folds and wrinkles.

2. Within two to three days after section a thin membrane
formed by the growth of new tissue from the cut surfaces closes
the two ends completely. The piece now becomes gradually
distended with water, probably owing to the accumulation of
metabolic products in the enteron and consequent diffusion of
water into this closed cavity. As distension proceeds the in-
folded margins of the body-wall at the two ends are forced apart
by internal pressure and the area occupied by the thin membrane
increases.

3. The first step in the regeneration of tentacles is the forma-
tion of a slight ridge, the marginal tentacular ridge, on the oral
end. This ridge is formed wholly within the tissue of the old
body-wall, its formation being accompanied by reduction and
disappearance of the muscular layer, disappearance of the pig-
ment and great reduction in thickness. The marginal ten-
tacles first appear as slight upgrowths from the highest — most
oral — point of the ridge, one tentacle corresponding to each in-
termesenterial chamber. The position of the mesenteries is indi-
cated externally on the tentacular ridge by slight furrows which
separate the regenerating tentacles from each other.

260 C. M. CHILD.

4. The regenerating marginal tentacles appear at first in a
single circle and all usually regenerate with nearly equal rapidity,
except in some cases the youngest pair in the growing region.
The directive tentacle is usually slightly thicker than the others
since the directive mesenteries are somewhat farther apart than
the other mesenteries. Rapid increase in length occurs in the
marginal tentacles, and the arrangement in about three circles or
rows is gradually attained in consequence of the fact that there
is not sufficient space on the margin of the disc for all of the ten-
tacles in a single row ; some are forced peripherally by the mu-
tual pressure exerted.

5. As the tentacles grow the disc expands and the distinction
between the thin membrane of new tissue which first closed the
end and the old body-wall with which it was connected disap-
pears completely in consequence of the complete disappearance
of the muscular layer, the reduction in thickness, and the loss of
pigment in the body-wall of the oral end.

6. The mouth appears after the marginal tentacles are well
established near the base of the directive, tentacle, gradually
extending along the directive plane across the center of the disc
until it is symmetrical. The part of the mouth first regenerated
is the region of the siphon oglyph.

7. The labial tentacles do not appear until the marginal ten-
tacles have attained a length of several millimeters. Each ten-
tacle appears as a distinct bud over an intermesenterial chamber,
but some intermesenterial chambers are without labial tentacles.

8. After the aboral end is closed by the new tissue this slowly
acquires a conical form, protruding from within the wrinkled
margin of the old body-wall. The wrinkles on the latter gradu-
ally disappear and the pigmentation slowly fades out for a short
distance oral to the cut end, this change being connected with re-
duction and disappearance of the muscular layer as this region of
the body-wall becomes involved in processes of growth and redif-
ferentiation in the same manner as the oral end. The aboral end
grows out into an elongated conical form at the end of which
the aboral pore appears. As the new muscles differentiate in
this region pigment stripes begin to appear.

HULL ZOOLOGICAL LABORATOR-*, UNIVERSITY OF CHICAGO,
July, 1903.

THE EYES OF THE BLIND VERTEBRATES OF

NORTH AMERICA. VI.1 THE EYES OF

TYPHLOPS LUMBRICALIS (LINN^US),

A BLIND SNAKE FROM CUBA.3

EFFA FUNK MUHSE.

lyphlops lumbricalis? a blind snake, is generally distributed in
the West Indies and Guiana. The specimens examined were
obtained by Dr. C. H. Eigenmann in the neighborhood 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 changed into alcohol. Only one young specimen was ob-
tained, and it was preserved in Zenker's fluid. For decalcification,
the heads of some were placed for at least three days in ten per
cent, nitric acid and others in Perenyi's fluid from one to two weeks.
One series was stained by the iron haematoxylin process, the others
with haemalum and eosin. It was very difficult to obtain satis-
factory 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 covered 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.

I. NORMAL EYES OF SNAKES.

Snakes differ from other animals in having the edges of the
two eyelids entirely grown together. A disk-shaped, conjunctival

1 Contributions from the Zoological Laboratory of Indiana University under the
direction of C. H. Eigenmann.

2The blind vertebrates of Cuba are rated with those of North America.
3Boulenger, G. A., "Catalogue of the Snakes in the British Museum," 1893.

261

262 EFFA FUNK MUHSE.

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 four straight
ones arise in the neighborhood of the foramen opticus, while the
two oblique ones arise from the surface of the prefrontal which
is turned toward the eye socket.

Closely connected with the eye is a gland, Harder's, whose
function is doubtful. Leading from this gland is a single duct,
which either empties into the duct from Jacobson's gland or di-
rectly 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. Ciliary muscles
are not found, but next to the iris is a great bundle of equatorial
muscle fibers running obliquely, which seem to be a continua-
tion of the iris musculature. The ciliary processes are weakly
developed.

The retina consists of the usual layers. The nerve fiber layer
is very thin (.003-. 004 mm.).

The ganglion cell layer consists of a single, rarely two layers
of small cells, each with a very large nucleus (.OI2-.OI3 mm.).
The inner reticular layer contains, at apparently regular inter-
vals, elongated, oval nuclei (.042— .045 mm.).

The inner nuclear layer consists of two kinds of cells (.052-
.054 mm.).

The outer reticular layer is very thin (.004— .005 mm.).
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—6 microns in length and an inner
much larger section, almost completely filled with a larger, pear-
shaped, strongly refractive body, the ellipsoid, 14-16 microns
in length and 8-9 microns across its widest part, which is turned
toward the limiting membrane. The twin cone consists 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.

A BLIND SNAKE FROM CUBA. 263

•

Passing between the limiting membranes are the radial sup-
porting Mullerian fibers.

II. THE EYE OF TypJilops vennicnlaris.

The work thus far on blind snakes has been done by Kohl on
Typhlops vermicularis, a species found in Greece and the south-
western part of. Asia, and on Typlilops braininns, 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 Wir-
belthieraugen."

He found that in depth the eye of Typlilops vermicularis is
equal to about one sixth that of Tropidonotns.

The brille is thicker in Typhlops than in Tropidonotns and, com-
pared with the axial diameter of the respective eyes, is seven
times thicker. In Typlilops the brille is equal in thickness to
about one half that of the ordinary skin of the head. In Tropi-
donotns it is equal to one fourth.

The cornea of TypJilops measures .0052 mm., and compared
with the relative sizes of the eyes is equal to about one half that
of Tropidonotns, which measures .064 mm.

The conjunctiva is thickened at the edge of the disc-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 one half those of Tropidonotns.

Harder's gland in Typlilops is many times larger than the eye-
ball.

The six muscles are present.

The lens is elliptical, while that of Tropidonotns is almost glob-
ular. The ratio of the lens volume of Typlilops to the eye vol-
ume is i : 14.04, while in Tropidonotns it is I : 3.6. The lens epi-
thelium of the former is relatively six times greater than that of
Tropidonotns.

The retina at the back of the eye of Typ/ilops, and the retina
of Tropidonotns bear the actual ratio of 8:13, while compared
with the eye axis in each case the Typhlops-r&tim. is four times
greater. The fovea centralis and area are absent.

'Kohl, Dr. C. "Rudimentare Wirbelthieraugen," Erster Thiel, Heft. 13,
Bibliotheca Zoologica. Verlag von Theodor Fischer, 1892, Cassel.

264 EFFA FUNK MUHSE.

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 ganglion cells are arranged in a single layer ex-
cept 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.

III. THE EYE OF TypJilops lumbricalis.

The eye shows through the large ocular scale, which entirely
covers it. It appears as a black spot surrounded by an unpig-
mented circle. The preocular, also a large scale, overlaps the
ocular and reaches just to the edge of the eye (Figs. I and 2).

General Account of the Eye.

Compared with one of the garter snakes and in proportion to
the size of the head, the eye of Typhhps lumbricalis is located
further from the surface and occupies far less space, while Har-
der's gland, associated with the eye in both, is relatively much
larger in TypJilops. In a specimen of TypJilops lumbricalis 2 \
cm. in length, the eye measured .306 mm. in width, and .387
mm. in depth. The greatest width of the gland of the same was
.711 mm. and the length was 1.067 mm. The gland completely
surrounds the eye up to the edges of the conjunctival sac (Figs.
3 and 4). In proportion to the size of the eyes, the gland of a
garter snake is much smaller than that of Typhlops lumbricalis,
but compared with RJdnenra floridaua J the gland of TypJilops
Iwnbricalis 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 being
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 (Fig. 3).

1 Eigenmann, C. A., "The Eyes of RJnneura fton'dana," Proceedings of the
Washington Academy of Sciences, Vol. IV., pp. 533-548, Sept. 30, 1902.

A BLIND SNAKE FROM CUBA.

265

ro.

•

po.s. o.s.

n.s.

266

EFFA FUNK MUHSE.

A BLIND SNAKE FROM CUBA. 267

A conjunct! val 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 .003 mm. At the edge of the sac, it is
differentiated into glands, the fornix conjunctiva, and measures
.016 mm (Figs. 3 and 4).

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.

Minute Anatomy of Eye.

Clioroid and Sclera. — The dense pigmentation makes it impos-
sible to distinguish 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 to-
gether representing the choroid. The choroidal pigmentary
layer seems to consist of long fibers circularly arranged. The
sclera can be followed by starting 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 cil-
iary 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 (Figs. 3 and 4).

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 little could be
determined about its structure. A well-developed capsule sur-
rounds it (Fig. 7).

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

268 EFFA FUNK MUHSE.

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 Typlilops 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 (Fig. 7). At this point the
arrangement could not be definitely determined in the sections.
At the back the retina, exclusive of the pigment layer, measures
.0725 mm.

Ends of fibers were seen projecting inward from the ganglion
cell layer, but no definite fiber layer could be distinguished
(10 in Fig. 5).

The ganglion cell layer (9 in the figures) consists of a single
row of large nucleated cells, somewhat irregularly arranged
(.008 mm.).

The inner reticular layer (8) consists of a mass of fibers in-
terwoven in a close network. This layer measures, at the back
of the eye, .015 mm.

The inner nuclear layer (6) consists of at least three layers of
cells, loosely arranged. The course of some of the fibers can be
followed among these cells. This layer measures .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 .005 mm.

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, con-
sisting probably of the limiting membrane and ends of the Miil-
lerian 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 .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. 5).

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

A BLIND SNAKE FROM CUBA.

269

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 .290 mm. and .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 elements of each layer are much more numerous
than in the adult, and they are packed much more closely to-
gether (Fig. 6). 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. 6).

COMPARATIVE MEASUREMENT' OF RETINAL LAYERS IN MM.

u C

££

E3

fife

9. >>

•ajj

tsa
°CJ

Inner
Reticular
Layer.

L- .

.. a t-i
o3 v a;
cTJ >>
c s «

-H^J

Outer
Reticular
Layer.

Sensory
Epithelium.

•3-3

"=• D,
^ V

^o

Tropidonotus natrix.
Typhlops veriniiula ris.
Typhlops, lumbricalis (adult).
Typhlops lumbricalis
(young 10 cm. ).

.003

.0018
.005

.012

.OOSl
.008
.OIO

.042

•0155
.015
.024

.052
.0221
.Ol6
.032

.004
.OO22
.005
.008

.0196
.0324
.030
.040

•1331

.0821

.0725

. 1206

RELATIVE PROPORTION* OF EYE PARTS.

Tropidonotus
natrijc.

Typhlops
vernricttla ris.

Typhlops
lumbricalis (adult).

Eye depth.
Brille :
Cornia :
Lens depth :
Coats :
Retina at back :

Eye axis
Eye axis
Eye axis
Eye axis
Eye axis

2.5541 mm.
:: I 177.4
:: I =39.9
:: I : 1.56
:: i :2i.63
:: I :I9.I9

.4399 mm.
I =10.77
I : 84. 6

I : 3-03
1:38.58

I : 5-36

.4032 mm.
I : 12.5
l:85
I : 2.5
I =25.4

I: 5-5

2/O EFFA FUNK MUHSE.

EXPLANATION UF FIGURES.

Figs. I and 2 are from entire specimens. All figures except I and 2 are from
sections. Figs. 7 and 8 are diagrams.

EXPLANATION OF NOTATIONS USED.

b. Brille. lv Second labial scale.

ch. Choroid. ly Third " " .

ci.p. Ciliary processes. I.e. Lens capsule.

cj. Conjunctiva. m.m. Middle member of cone.

fj.s. Conjunctival sac. n.s. Nasal scale.

cor. Cornea. o.c. Ocular scale.

cov. Coverings of eye. /./. Pigment layer.

d. Dermis. po.s. Preocular scale.

e.m. End member of cone. r. Retina.

F.cj. Fornix conjunctiva. ro. Rostral.

H g. Harder' s gland. r.m. Roof of mouth.

i. Iris. s. Sclera.

i.e. Interpolated cells. s.e.l. Sensory epithelium layer.

/. Lens. rit.cav. Vitreous cavity.

/j. First labial scale.

1. Pigment layer. 6. Inner nuclear layer.

2. Cones. 8. Inner reticular layer.

3. Outer nuclear layer. 9. Ganglion cell layer.

4. Outer reticular layer. 10. Fiber layer.

FIG. I. Dorsal view of head of a specimen 21 cm. long.

FIG. 2. Lateral view of head of same specimen.

FIG. 3. Horizontal section of a specimen 20 cm. long, ^-objective, 2-inch eye
piece, camera lucida.

FIG. 4. Transverse section of a specimen 21 cm. long, ^-objective, 2-inch eye
piece, camera lucida. (Scales not shown.)

FIG. 5. Section of retina of an adult specimen 21 cm. long, ^-objective, l-inch
eye piece, camera lucida.

FIG. 6. Section of retina of young specimen, 10 cm. long, ^-objective, l-inch
eye piece, camera lucida.

FIG. 7. Diagrams of eye of adult.

FIG. 8. Diagram of eye of young.

(The region x-r\\\ the sections could not be made out and is consequently left
blank in the diagram. )

SOME EXPERIMENTS IN FEEDING LIZARDS WITH
PROTECTIVELY COLORED INSECTS.1

ANNIE H. PRITCHETT.

During the past year, from October to May inclusive, I have
been experimenting with insects that possess protective, mimetic
and warning colors or that have some disagreeable character-
istics which in a measure are supposed to prevent their being
devoured by insect-eating animals. For this purpose several
species of lizards found in the vicinity of Austin, Texas, have
been kept in separate, convenient cages and fed with the various
insects. Some interesting observations on the habits of the liz-
ards were made incidentally and these are also noted in the fol-
lowing paper.

The species of lizards used for the experiments are the follow-
ing : GcrrJionotns inf emails Baird, C/irotaphytus collaris Say,
Scelopom s floridamt s Baird, Holbrookia tcxana Troschel, Cncuii-
doplwnts scxlincatns Linn., Phrynosoma cornutitui Harl., and an
undetermined species of Euincces.

EXPERIMENTS WITH SC'ELOPORUS FLORIDANUS.
LEPIDOPTERA.

Anosia plexippus Linn. This species is conspicuously colored
in light brown with black and white markings. It is also said
to have a disagreeable taste and is the supposed model of the
mimic Basilarchia disippns. Specimens were introduced Oc-
tober 3 i , November 6, April 2 (two), April 4, April 6. Each
time the butterfly was caught by the wing, or by the wings if
folded, held for a few moments and then eaten slowly. It was
not torn to pieces but held by part of the wings and swallowed
gradually, the lizard often pausing a moment to rest.

Papilio (Laertias) philcnor Linn. Formerly this was included
in the genus Papilio but has been separated because of character-
istic differences, important among which is the supposition that
it is an especially protected form because its larva feeds on Aris-

1 Contribution from the Zoological Laboratory of the University of Texas, No. $2.

271

2/2 ANNIE H. PRITCHETT.

tolocJiia, a poisonous plant of disagreeable taste. On October
30, March 27, March 30 (two), March 31, April I (three), April
4, April 16, April 23, May 4 (four), May 6 (two), butterflies
were introduced into the cage and quickly eaten by the lizards
with evident relish. On May 6 one of the specimens was badly
mutilated and the lizards were not induced to take it for more
than an hour.
• Picris occidentalis Reakirt. October 29, April 20 (three).

Picris protodicc Boisd-Lec. April 23, May I (four). These
forms, white with black markings, were readily eaten.

Colias curythcmc Boisd. November 8 (two), March 3 I (two),
April 9, April 20 (seven), April 23 (five), May I (two). All
quickly eaten.

Colias ariadnc Edwards. April 16 (two).

Colias scnddcrc Reakirt. April 20, May I (two). These
species are of striking yellow or orange marked with black, a
typical warning combination, yet all were eaten eagerly.

PyrrJiamca andria Scudder. This form is admirably protected
by having the under side of the wings an exact imitation of a
dead leaf. The wings are held folded closely together when the
butterfly is at rest, and it remains motionless in this position for
a great length of time. It is one of the most perfect instances
of protective resemblance that I have obtained. Specimens were
introduced November 14, April 22 (two, $ and 9) ar)d April 27.
On April 22 the butterflies were not noticed at first. Several
times they were offered to the lizards ; the male was taken in
about five minutes and the female ten minutes later. On April
27 the butterfly was seized by the wings several times, then
dropped again. It remained motionless unless I moved it and
the lizard would then seize it again. Finally it was abandoned,
but it had disappeared the next day and probably had been eaten
at last.

Pyraineis atalanta Linn. November 29. This is a conspicuous
form, of black, brown, red and white. The lizards ate it eagerly.

Pyraineis lutiitcra Fabr., a similar form but having large eye-
spots underneath the wings. It was eaten May I.

Grapta inter rogationis Fabr. April i. This species also has
the under side of the wings in imitation of a dead leaf, and is

FEEDING LIZARDS WITH COLORED INSECTS. 273

very difficult to detect when at rest. It is in the habit of remain-
ing motionless for a long while. The specimen introduced was
at once eaten.

Papilio cresphontes Cramer. One specimen was introduced
April 23 and four lizards at once seized the outspread wings.
They showed no preference for the body but ate the wings first,
as is usually the case. On May 7 the wings of the specimen
introduced were almost entirely eaten when the lizard happened
to drop it. It remained quiet, and the lizard would only take it
again after I had made the butterfly move several times.1

DeilepJiila lincata. May 5. Two of these Sphingid moths
were introduced and seized at once. They fluttered continuously
and thus frustrated the attempts of several other lizards that
were trying to participate. One moth was held by the head,
the other by the wing for quite a while, till they ceased fluttering,
and were then eaten.

Species unknown. May 4. This small moth is of black and
orange, the typical warning coloration. It was eaten at once
without any symptoms of dislike being shown.

H EMITTER A.

Lygceid. May 5. Just after the above-mentioned moth was
eaten four of these bugs were introduced. They are of the typi-
cal black-and-red or orange warning colors and have a very dis-
agreeable odor. The same lizard that ate the moth at once
seized a bug, chewed it a moment and spit it out, then licked his
mouth for some time as if to remove the bad taste. Another
lizard examined a second bug but made no attempt to take it.
One bug was eaten later by the third lizard and the other two
were gone next morning. May 13 a bug was introduced, seized
at once and then rejected as before. It is evidently quite un-
palatable.

BracJiymena my ops. Three were introduced November 8, but
were never noticed by the lizards. The bug is gray in color,

1 A glass jar containing live butterflies was placed on a chair about two and one
half or three feet from the cage of Sceloporus. A large male lizard immediately
climbed up the side of the cage, eyed the butterflies eagerly and seemed quite excited.
This happened a few days later with several of the lizards. When the insects were
introduced they were seized and eaten at once, several lizards quarreling over a de-
sirable specimen and sharing it among themselves.

274 ANNIE H. PRITCHETT.

quite similar to the bark of trees that it frequents, and possesses
a very unpleasant odor.

Fnlgorid. Introduced November 5, November 6. This lan-
tern fly is almost impossible to detect when at rest upon the
trunks of the cedars and arbor-vitse which it frequents. The
upper wings and exposed portions of the head and thorax are
somber gray, the almost transparent wings showing a tinge of
pink when spread. The under wings are either entirely black or
have a small white spot near the center. The posterior dorsal
portion of the abdomen is bright red or deep orange, the re-
maining portions of the body being black. The insect shows
perfect protective coloration at rest and a rather typical warning
combination in flight. The insects were eaten at once by the liz-
ards when seen in motion.

COLEOPTERA.

Chauliognathus scutellaris Lee. Although this beetle is colored
black and yellow it appears to be palatable. May I five were
introduced. The first was taken by the lizard that sampled the
Lygceid, tasted a little, and rejected. However three others were
eaten by a second lizard and the last beetle by a third. May 4
twenty beetles were introduced and all were eaten without any evi-
dences of unpalatability. On May 5 four were introduced just after
the four Lyg&ids. The first was carefully examined before being
eaten ; the second was tasted and refused by another lizard ; the
others were not noticed, as was also the case when seven were intro-
duced the following day. The lizards were probably too well fed,
for since then, May 1 1 and 1 3, they have eaten all that were offered.

Epicauta sp. November 3. This black blister-beetle was
tasted and rejected immediately. Unfortunately no more speci-
mens were found.

ZopJicrus lialdcinani Salle. This very hard Tenebrionid beetle,
conspicuously colored in black and white, was introduced Nov-
ember 9 and removed alive December 13 during which time no
attempt to take it was seen. Specimens experimented with Nov-
ember 12 and May 5 gave the same results.

Lncanns dauia Thumb. This black, horny beetle was intro-
duced November 17 and died January 7; during this time the
lizards never tried to take it.

FEEDING LIZARDS WITH COLORED INSECTS. 2/5

Harpalus caliginosus Fab. This beetle is large, black and
rather hard, nevertheless one was eaten December 2, one De-
cember 12 and another partly eaten January 8. Four remained
dead at this date. Their odor is offensive.

Brachynus sp. When seized this beetle ejects a strong, volatile
acid with a sharp, audible report. This always surprised the
lizards ; nevertheless, of the four beetles placed in the cage three
were eaten, but the last refused. Two more were introduced Feb-
ruary 26 and one March 5, which afterward disappeared and pre-
sumably were eaten.

Brachynus sp. April 3. This beetle, larger than the pre-
ceding species, was eaten at once.

Calosoma angulatus Chev. and

Pasiinachns depresses Fab. were introduced March 17. The
lizards attempted to catch them, but failed, and soon gave up the
chase.

Chlanius orbits Horn. The odor of this beetle is quite offen-
sive. March 9 one was eaten at once. On March 10 two lizards
tried to catch a specimen but failed repeatedly. They appeared
to notice the odor and gave up the chase. On March 23, how-
ever, the lizard that ate the former now ate another, and still a
fourth was eaten April 3, but with evident disgust.

Cantliaris fiihnpennis Lee. This large blister beetle has the
typical warning colors of black and yellowish-brown and is
further protected by a disagreeable secretion that exudes from
the joints of the legs when the insect is seized and which is
capable of producing blisters. Four of these beetles were intro-
duced May 19 and each was seized at once, then quickly shaken
off. The lizards eyed the beetles intently, but made no attempts
to take them. These specimens were removed and introduced
again the following day. Only one beetle was taken this time
and it was quickly rejected. On May 21 several beetles were
again introduced. One was caught and quickly rejected and no
further notice was taken of them unless they crawled upon the
lizards, in which case they were shaken off violently.

DIPTERA.

Musca doincstica and Stoino.vys calcitrans. A small lizard of
this species (Sccloporns floridanus} soon became so tame that it

2/6 ANNIE H. PRITCHETT.

would lie on my hand and eat the flies which I caught and of-
fered in my fingers. Sometimes he would catch the flies himself
if I held him close to the window where they were crawling. He
also ate a number of small spiders that were just emerging from
the egg case placed in a glass jar. The lizard was kept in a cage
with adults of the same species and was possibly eaten by them,
as no trace of him could be found, and these lizards had, on two
other occasions, been suspected of devouring small lizards.

HYMENOPTERA.

Pogonomyrmex ba.rba.t2is var. nwnefaciens. These ants were
eaten October 29, November 3, November 22 and May 24.
The sting is quite severe.

Pachycondyla liarpax, a stinging Ponerine ant, was eaten Octo-
ber 28.

Polistcs annularis, a formidable wasp, was not noticed Novem-
ber 5.

ORTHOPTERA.

Giylhis abbreviatus. Several of these crickets were eaten
March 7 and March I r. It is therefore probable that those in-
troduced November 9, January 18, and January 19, were also
eaten, since crickets seem to be a favorite food with all the species
of lizards.

NEUROPTERA.

Panorpa nuptialis Gerst. This species has the wings of typical
black and yellow warning colors. A female was introduced
November 9 and a male November 15. Both disappeared in
some way, but were not seen to be eaten.

ARACHNIDA.

Epeira fasciata Hentz. This protectively colored specimen was
eaten October 25 and a second November 6.

SCORPIONS.

Centrums caroliniensis Beauv. On March 23 the specimen
which was introduced stung one of the lizards. He appeared to
be in much pain and was so frightened at the scorpion that the
experiment seemed likely to terminate there, but suddenly he
seized the offending sting in his mouth and spitefully devoured the

FEEDING LIZARDS WITH COLORED INSECTS. 2JJ

whole specimen. The color of this scorpion would seem to
afford it efficient protection. This, together with its flat form,
frequently prevents its being noticed by a casual observer when
the stone under which it rests is overturned.

MYRIOPODA.

Jit/ns (Spirobolus) multistriatns Walsh. The specimen intro-
duced November 15 was not molested, but when two were intro-
duced February 12 a lizard bit off part of the head of one Jitlus.
Both specimens died after a few days, neither being eaten. This
myriopod has a hard integument and is defended by means of an
acrid secretion that is thrown out from the repugnatorial glands
along each side of the body. It has the habit of coiling up and
remaining quiescent whenever it is touched. This action makes
the lizards suspicious of it.1

EXPERIMENTS WITH GERRHONOTUS INFERNALIS BAIRD.

The favorite foods of these lizards are crickets, grasshoppers,
spiders and scorpions. A few Hemiptera were eaten also.

LEPIDOPTERA.

Anosia plcxippns Linn. April i, April 2, April 4 (three).
None of these specimens were eaten.

Papilio (Lturtias) pJiilcnor Linn. March 26, March 30, April
6. All were examined and rejected.

Pyramcis cardui Linn. November 17. Offered and refused.

Pyrrhancca andria Scudder. November 9. Refused.

Colias eurytlicuie Boisd. March 30, April I (three), April 6.
On the latter date the butterfly was taken by the wings but soon
dropped, and all others were refused entirely.

1 Sceloporus floridamis is badly infected with an interesting mite which attaches
itself under the scales of the lizard until sexually mature and then crawls up on
the wooden part of the cage to oviposit. The eggs are placed in a peculiarly con-
structed palisade and hatch as a six-legged larva that appears identical with the ordi-
nary " red bug." The adult has a pubescent black integument ; the head, anus and
four pairs of legs are bright red. The legs are arranged in groups, two pairs being
situated on the anterior portion of the body and two in the posterior region. Mr.
Nathan Banks believes that this form may represent a new genus since it is the only
lizard parasite that has been taken in this country, and appears to be closely related
to the Italian genus Geckobia.

2/8 ANNIE H. PRITCHETT.

ORTHOPTERA.

Acridiuin americamim Scudd. November 15, November 24,
January 28, March I I, March 30. This large grasshopper is of
a very somber, dusty color and extremely quick in flight. When-
ever introduced into the cage it was at once eaten eagerly. The
lizard seized the insect by the thorax, held it thus for some time,
regrasped it more anteriorly several times until the head was taken
into the mouth. The insect was then swallowed slowly, the
lizard chewing a while, pausing to rest, then gulping down another
portion. On one occasion when the grasshopper became some-
what crooked, although it was nearly completely swallowed it
was disgorged, 'straightened, and then devoured again.

Species unknown. On November 29 a large grasshopper was
eaten in the usual way. The body, legs and head were dark
green ; the wings brown. The whole body was ornamented with
white or yellow spots and lines.

Gryllus abbreviates. December 12, January 10 (five), March
7 (several), March 10 (two), April 6 (several). All the speci-
mens were eaten eagerly.

NEUROPTERA.

Panorpa nuptialis Gerst. November 9. Although this warn-
ingly-colored insect remained in the cage six days, no attempt
was made to seize it.

COLEOPTERA.

Lucanus dama Thunb. November 8, was not eaten.

Zoplierns haldemani Salle. November 9, was refused.

Harpalns caligiriosus Fab. December 11 and December 18.
Five specimens were introduced, and all died.

Bracliymis sp. February I 2. Two of these beetles were in-
troduced and were not noticed by the lizards, though offered re-
peatedly. They run very swiftly, hiding at every opportunity,
and the lizards are probably too slow in their movements to catch
so quick a prey.

Pat rob us longicornisSay. The beetle was introduced February
13, and remained until March 5, but no attempt was made to
take it.

FEEDING LIZARDS WITH COLORED INSECTS.

Diabrotica punctata Oliv. February 13. These green-and-
black beetles were probably too small for the lizards to perceive.

Chlcenius orbus Horn. March 7. One of the lizards ran up
to examine the beetle but when near turned aside, evidently dis-
couraged because of the disagreeable odor, and did not try again
to take it.

Pasimachns- dcpressns Fab. March 17. The beetle was ex-
amined and refused.

Calosoina angulatns Chev. April 6. The beetle seemed never
to have been noticed.

Chauliognathns scntcllaris Lee. May 4. The lizards seemed
to pay no attention to the beetle although fifteen1 specimens were
introduced.

Cantliaris fulvipennis Lee. Two specimens of this black-and-
yellovv blister beetle were introduced May 19. One was seized at
once by one of the lizards, chewed a moment, then dropped quickly.
The lizard began writhing and rubbing his mouth in the sand,
appearing much distressed. The second beetle was not noticed
by any of the lizards and was removed. On May 20 they eyed
the beetle that was introduced, but made no attempt to take it.
May 21, the specimen seemed not to be noticed. Others intro-
duced May 26 gave the save negative result as the preceding
experiment.

H EMITTER A.

Brachymena inyops. December i, January 24. This pro-
tectively colored, malodorous form was not noticed by the
lizards.

Lygceid. May 5. Two specimens of this warningly colored
bug were introduced, examined and refused.

Fulgorid sp. November 5, November 6. Several specimens
were eaten with evident relish. The bug was never refused if
alive, but never eaten if dead.

HYMENOPTERA.

Polistcs annnlaris. Linn. November 4, refused.

Caniponotns sansabcanus Buckley. November 29 and Cain-
ponotus fcstinatns Buckley. April 13. These ants were possibly
too small to be noticed.

28O ANNIE H. PRITCHETT.

ARACHNIDA.

Lathrodectes mactans. November 17, November 29, Decem-
ber 6 (two), December 18 (four), January 20 (two), February 2
(two), March 9, March 11, March 17, March 25, March 30,
April 6, April 13 (four), April 20 (three), May 19 (two). These
spiders are of a jet black color conspicuously marked with
crimson or sometimes white, thus exhibiting striking warning
coloration. They are even said to be poisonous, yet they were
always quickly seized and eaten by these lizards.

Attus mystaccus. December 10, December 12 (two). The
somber gray color of these spiders affords them good protection
under the stones where they live. They were eaten eagerly.

Lycosa sp. March 9, March 23 (two), March 25. This spider
resembles very closely in color the underside of the stones where
it is often found. It was eaten at once when introduced.

SCORPIONS.

Centrums caroliniensis Beauv. January 20, March 17, March
23 (two), April 13 (three), April 20 (six), April 27 (five), May 4,
May 1 8 (six), May 19 (two), May 25 (two). All these speci-
mens were eaten with evident relish and no attention was paid to
the sting. The hard integument of the lizard prevents the pene-
tration of the sting.

^5

MYRIOPODA.

Jnlns ( Spirobohis] multistriatus Walsh. The specimen was in-
troduced November 18 and died January 7. It was not noticed
by the lizards, as was also the case with two specimens intro-
duced February 1 2.

EXPERIMENTS WITH CROTAPHYTUS COLLARIS Say.

Two of these lizards were captured November 9 and were not
seen to eat a single insect until February I 2. Various kinds of
insects were placed in the cage, and though the lizards were
quite tame and lively they would not eat. On January 233 dish
of water was placed in the cage and they learned to drink from
the dish and also from the pipette used for refilling it. The water
furnished their only nourishment for three months. A third
lizard was captured April 4 and though very fierce at first,

FEEDING LIZARDS WITH COLORED INSECTS. 28 1

became quite tame in about a week, allowing me to rub its head
and body with my hand. These lizards occupied the cage with
Gerrhonotus infernal is until December I when they were placed
in a separate one. The experiments were as follows :

LEPIDOPTERA.

Meganostoma enrydice Boisd. December i 5, was not eaten.

Papilio ( L&rtias] pliiletwr Linn. March 26, March 30. These
were not noticed and were afterward removed. The specimen
introduced April 2 was found dead and apparently unharmed the
following day. On April 7 the specimen introduced the previous
day was gone, and on April 8 the lizard last caught was seen
eating a butterfly. On April 21 a specimen was introduced and
only a part of the wings remained next day. However, the two
specimens introduced on May 4 remained in the cage two days
and were not eaten.

Colias eurytheme Boisd. March 30. The specimen was not
noticed by the lizards and was removed next day. The two
that were introduced April i were gone the day following and of
the two introduced April 2 one was entirely eaten and only the
torn wings of the second remained. On May 8 one of the lizards
seized the specimen just introduced by the edges of the folded
wings and ate it slowly, often pausing to rest, but never releas-
ing it.

Anosia plexippns Linn. Introduced April 2. Next day the
head and thorax were chewed up and one fore wing was missing.
Others that were introduced afterward disappeared but were not
seen when eaten. But on May 18 the butterfly was seized at
once by one of the lizards and a second lizard bit off part of a
wing. Between them they ate the specimen, but did not take
the two introduced May 25.

Pieris occidcntalis Reakirt. April 16, was eaten.

Grapta interrogationis Fabr. The specimen introduced April
27 was eaten, and those placed in the cage May 7 and May 9
were gone the following mornings. Probably they were also
eaten.

Papilio cresphontes Cramer. April 23. This butterfly did not
seem to be noticed by the lizards.

282 ANNIE H. PRITCHETT.

Anosia berenice var. strigosa Bates. This butterfly has the
same warning coloration scheme as Anosia plexippus. It had
disappeared next day and was probably eaten.

COLEOPTERA.

The following specimens were introduced, but none of them
were eaten and were rarely ever noticed by the lizards, though
offered repeatedly :

Har pains caliginosus Fab. December 2.

Brachynns sp. February 13 (three).

Clilcenins orbits Horn. March 7, March 23, April 3.

Micryxis distinctns Hald. March 7. This beetle was evidently
too small for the lizards to perceive. They pay no attention to
small insects, possibly because their eyes are not capable of per-
ceiving them.

Chauliognathus scutellaris Lee. May 4 (eighteen), May 5
(six). All refused.

A notable exception to this custom of refusing beetles was seen
when three black-and-yellow blister beetles, Cantharis fulvipennis
Lee., were introduced May 19. A lizard seized one of the beetles
and ate it, then seized a second. One of the other lizards tried
to take it from the former, but was unsuccessful, and the second
beetle was eaten. The third was apparently not noticed by any
of the lizards and was soon removed. Specimens were intro-
duced May 20, May 21 and May 26. but did not seem to be
noticed.

Occasionally larvae of beetles were introduced and eaten, but
with the above exception these lizards do not appear to feed on
imaginal Coleoptera. Cantharis probably does not appear in the
natural habitat of the lizard, the latter being a mountain species,
while the beetle is found in the fields on the Mexican poppy
(Argemone mexicand).

ORTHOPTERA.

Gryllns abbreviates. February 12 three specimens were intro-
duced, one of which was dead, and was at once seized and eaten
by a lizard. This was the first food it had taken since its cap-
ture, November 9, and it is the only instance known of a lizard
eating a dead insect. The two remaining crickets disappeared

FEEDING LIZARDS WITH COLORED INSECTS. 283

4

later and were evidently eaten. On March 23 one of the lizards

tried repeatedly to catch one of the five crickets introduced, but
failed, and finally gave up the chase, even refusing the insect
when it was held before him in the forceps. The lizards were
seen to catch and eat crickets on the following days : April 1 3
(two); April 20, April 27 (two), and on several occasions spec-
imens that were introduced in the evening had disappeared by
the following morning. Indeed, crickets seem to form the prin-
cipal food of these lizards.

i

NEMOPTERA.

Panorpa miptialis Gerst. December 1 2. This warningly-
colored insect was apparently not noticed and died soon after-
ward.

DIPTERA.

Hcnnetia illnccns Linn. December 13. This form resembles
a wasp somewhat closely. It was not noticed by the lizards.

HEMIPTERA.

Lygceid sp. May 5. The lizards could not be induced to
take the specimens.

HYMENOPTERA.

No experiments with Hymenoptera were made with these
lizards.

ARACHNIDA.

Attns uiystaceus. December I. This spider was not noticed
though offered repeatedly.

Latlirodectcs mac tans. Specimens were introduced January 20
(three), March 23 (two) but none were eaten.

Other small spiders (names unknown) were introduced at dif-
ferent times but were never eaten.

MYRIOPODA.

Scutigera forceps. December 18. Specimen refused.

SCORPIONS.

Centrums carolinicnsis Beauv. November 15. The scorpion
stung one of the lizards and it seemed to suffer so intensely and
was so frightened whenever the former came near it that the ex-
periment was never repeated.

284 ANNIE H. PRITCHETT.

Three other species of lizards were placed in the same cage
with Crotaphytits collaris, from which the following results were
obtained :

1. Cncinidophonis sexlineatns Linn. One specimen was caught
December I and died January 7 during which time it was never
seen to take any food. This was also the case with two small liz-
ards of this species that were in the cage with Sceloporns. They dis-
appeared mysteriously and are supposed to have been devoured.
The lizard is quite common, but difficult to catch, and it is re-
gretted that more were not obtained for the experiments.

2. Holbrookia texana Trosch. Two of these lizards were
placed in the cage early in April and have never been seen to take
any food.

Emucccs sp. This small lizard was captured March i 2. On
March 30 it tore up and ate the body of a butterfly, Pi en's occidcn-
talis Reakirt. April 6 it caught, tore to pieces and ate a cricket
larger in circumference than itself. April 8 it ate a large house
fly and on April 10 a number of small mantids, Staginoinantis
Carolina, recently hatched. The lizard was very alert, spying
the mantids at a distance of several inches, though the latter
were quite small and exactly the color of the sand on the floor of
the cage. On April 23 and May 8 other young mantids of the
same size were eaten.

Phrynosoma cormttuni Harl. The "horned toads" were kept
in cages with other lizards and also separately and were never
seen to eat anything but ants. They are especially fond of the
large agricultural ant, Pogonomyrmex barbatus Smith var. uwlc-
faciens Buckley.

GENERAL SUMMARY.

1 . Only one instance is known of a lizard eating a dead insect.

2. Insects that move slowly do not attract the attention of the
lizards so much as do the more active forms, hence those that
remain quiescent are rarely even attacked.

3. Insects below a certain size are apparently not perceived by
the large species of lizards. Examples of such insects are Dia-
brotica punctdta Oliv., Micryxis distinctns Hald., and various ants
(Camponotus).

FEEDING LIZARDS WITH COLORED INSECTS. 285

4. Large beetles having hard elytra are seldom eaten.

5. A butterfly with mutilated wings was not taken for an hour
and a half although another perfect specimen introduced at the
same time was eaten at once.

6. If an insect (e. g., a beetle) falls upon its back the lizards
rarely ever seize it until it has gotten upon its feet again.

7. The myriopod Julns was not eaten by any lizard.

8. Although the combinations of black and yellow, black and
orange, or black and red are supposed to serve the purpose of
warning coloration, all insects possessing these colors were, at
one time or another, eaten, with the possible exception of Pan-
orpa nuptialis Gerst and a malodorous Lygcvid bug.

9. Sceloporus floridanus is perhaps the most satisfactory lizard
for these experiments since it eats insects of all groups.

10. Sceleporus seizes any part of the insect, but as a rule only
the wings of the butterflies and large moths.

i i. All the lizards except Enincces seize the insect with the
mouth and swallow it a little at a time, never biting off pieces,
but keeping the insect entire. Eumcc cs swallows its prey thus
if small, but when the insect is large he shakes and pulls it to
pieces with his mouth and eats the separate pieces.

1 2. Sceloporus is very active and is not easily tamed.

13. Gerrhonotus is exceedingly slow in capturing its prey. It
creeps up stealthily, pauses when quite near, examines the insect
by protruding the tongue, rises as high as possible on the toes
of the fore limbs and then seizes the insect by the back with a
sudden spring. If the insect does not move it is frequently left
unmolested. This lizard soon becomes quite tame but does not
enjoy being handled. It was seen to drink water from the dish
by lapping with the tongue, but usually preferred taking it from
the pipette, allowing me to place a drop at a time on its out-
stretched tongue.

14. Eumeces sometimes drinks by lapping with the tongue,
sometimes by sucking up the water. Sceloporus, Crotaphytus
and Phrynosoma drink by sucking the water into the mouth. At
first Sceloporus and Crotaphytus would drink only from the
pipette, but were gradually induced to follow that to the dish
and drink from the latter.

286 ANNIE H. PR1TCHETT.

15. Phiynosoma cor/nttiun, though apparently quite tame,
seems at first rather shy about eating in confinement. Ants,
especially the agricultural ants (Pogonomyrmex}t are its only
known food.

1 6. Crotaphytus is not accurate in seizing its prey. It often
fails repeatedly and gives up the attempt.

17. The larger lizards were several times suspected of having
eaten smaller specimens that had been placed in the same cage.

1 8. Crotaphytus soon becomes quite tame and enjoys being
petted. The smaller ones crawled upon my hand in the cage
and refused to be put down.

19. The largest Crotaphytus shed its skin during the night of
May 6. Next morning the sand in the cage was very much
dug out and heaped up, but no traces of the skin could be found.

20. A Gerrhonotus shed during the night of April 29. The
old skin was turned wrong side out and probably came off nearly
whole, though several parts were broken when it was found next
morning. A second lizard shed May 22 and I watched it pull
the old skin off wrong side out by creeping round and round
the cage close to the sides. The skin was loosened first from
the upper and lower jaws along the sides of the mouth, and be-
gan to peel off backward by the lizard's rubbing its head against
the sand on the bottom of the cage.

BIBLIOGRAPHY.

Beddard, Frank E.

'91 Warning Colors. Nature, Vol. 45, No. 1152.
Blanford, Walter H.

'97 On Mimicry. Nature, Vol. 56, No. 1444.
Coste, F. H. Perry.

'92 On Insect Colors. Nature, Vol. 45, Nos. 1170-71.
Distant, W. L.

'91 Warning Colors. Nature, Vol. 45, No. 1156.
Distant, W. L.

'91 Assumed Instance of Compound Protective Resemblance in an African But-
terfly. Nature, Vol. 43, No. 1113.
Haase, Erich.

'97 Mimicry in Butterflies and Moths. Nature, Vol. 57, No. 1462.
Hampson, Geo.

'98 Protective and Pseudo- Mimicry. Nature, Vol. 57, No. 1477.
Heckel, M. E.

'91 Mimicry in Spiders. Nature, Vol. 44, No. 1141.

FEEDING LIZARDS WITH COLORED INSECTS. 287

Jordan, Karl.

'97 ( hi Mimicry. Nature, Vol. 56, Nos. 1442-53.
Marshall, G. A. K., & Poulton, E. B.

'02 Bionomics of South African Insects.
Mayer, Alfred G.

'97 On the Colors and Color Patterns of Moths and Butterflies. Nature, Vol. 55,

No. 1435.
Mayer, Alfred G.

'02 Effects of Natural Selection and Race-Tendency upon the Color-Patterns of

Lepidoptera. Museum of Brooklyn Arts and Sciences, Vol. I, No. 2.
Peckham, G., & E. G.

'89 Protective Resemblances in Spiders. Occ. Papers of Nat. Hist. Soc. of Wis.
Poulton, E. B.

'go The Colors of Animals.
Poulton, E. B.

'98 Protective Mimicry and Common Warning Color. Nature, Vol. 57, No.

1478.
Poulton, E. B.

'97 Mimicry as Evidence of the Truth of Natural Selection. Nature, Vol. 56

No. 1458.
Poulton, E. B.

'97 Theories of Mimicry as Illustrated by African Butterflies. Nature, Vol. 56,

No. 1458.
Poulton, E. B.

'90 Mimicry. Nature, Vol. 42, No. 1090.
Poulton, E. B.

'87 Protective Value of Color and Markings in Insects. Nature, Vol. 36, No.

938.
Poulton, E. B.

'87 Experiments upon Color Relations between Phytophagous Larvae and their

Surroundings. Nature, Vol. 36, No. 438.
Poulton, E. B.

'87 The Secretion of Pure Aqueous Formic Acid by Lepidopterous Larvae for

Purposes of Defense. Nature, Vol. 36, No. 438.
Sibley, Walter K., & Poulton, E. B.

'90 Protective Colors. Nature, Vol. 42, No. 1092.
Syme, David, & Wallace, A. R.

'91 Topical Selection and Mimicry. Nature, Vol. 45, No. 1150.
Trimen, Roland.

'98 Mimicry in Insects. Nature, Vol. 57, No. 1479.
Wallace, A. R.

'90 Colors of Animals. Nature, Vol. 42, No. 1081.

THE UNIVERSITY OF TEXAS, ZOOLOGICAL LABORATORY,
AUSTIN, TEXAS, June 4, 1903.

SEX RECOGNITION AMONG AMPHIPODS.1

S. J. HOLMES.

How do males of the amphipod Crustacea distinguish the
females ? It is well known that the males of the Gammaridea have
the curious habit of carrying the females under their body for a
considerable time. This act of transportation has probably no
further significance in relation to the fertilization of the eggs than
to secure the proximity of the two sexes when the proper time
for fertilization arrives. According to the observations of Delia
Valle on Gainniarus pungcns the eggs are not fertilized until after
they are laid, oviposition occurring a short time after moulting.
When the moulting of the female has been effected, the male
bends his body beneath that of his mate and deposits spermatozoa
upon the ventral surface of her thorax. The deposit of sperm is
followed within half an hour by the laying of the eggs. After
the act of copulation the male regains his original position and
swims about with the female as before. The same relation of
oviposition to moulting was found by Miss Langenbeck in Micro-
deutopus, the male leaving the female during her moulting proc-
ess but soon resuming his previous position when the moult was
completed.

The instinct of the male amphipod to seize and retain hold of
the female is one of remarkable strength. The male retains his
hold, despite all efforts to dislodge him, with remarkable persist-
ence, and will still cling to the female after the posterior half of
his body has been cut away. My own observations on the sexual
behavior of amphipods relate mainly to three species, AuipJiitJwe
longiinana Smith, Hyalclla dcntata Smith and Gammarus fasciatus
Say. The sexual behavior of these three species is remarkably
similar, athough they belong to as many distinct families. The
female while being carried about keeps remarkably impassive.
Her thoracic legs are drawn up, the abdomen held strongly
flexed, the whole body assuming as compact a form as possible.
She takes little or no part in swimming ; the movement of the

1 From the Zoological Laboratory of the University of Michigan, Ann Arbor, Mich.

288

SEX RECOGNITION AMONG AMPHIPODS. 289

pleopods when the body is strongly bent upon itself serves only
to keep a current of water passing by the gills. She is carried
about like a helpless burden, allowing her vigorous spouse to
assume the entire labor of transportation and the responsibility
of keeping her as well as himself out of danger. The efforts of
the male to seize the female and get her into the proper position
to be carried have the effect of inducing her to throw herself into
the characteristic bodily attitude and remain quiet. The attitude
assumed by the female is similar to that observed in the ordinary
thigmotactic reaction of amphipods and may, perhaps, be but the
same form of response, somewhat modified and specialized in re-
lation to the function of reproduction. When the males are torn
away from the females they soon seize their partners again and
roll them about into the proper position and then proceed on
their way in apparent contentment. The female as soon as seized
by the male curls up and allows herself to be rolled and tumbled
about without a show of resistance or protest. The males, as a
rule, are considerably larger than the females and usually get
their partners into the desired position quite readily ; but when a
small male attempts to carry a large female he experiences much
difficulty. I have observed a male Hyalella endeavoring to carry
a female somewhat larger than himself. After seizing the female
he would turn her around until she finally came into the proper
position for transportation, but owing to the larger size of his
partner the male could not reach around her body so as to carry
her away. No sooner was the female properly adjusted than the
male would lose hold of her round body and the same efforts
had to be repeated. During all this performance the female re-
mained dutifully passive. After watching the further struggles
of the male for over half an hour I became convinced, although
he was not, that he had undertaken an impossible task, and dis-
continued my observations.

In order to ascertain if sight plays any part in sex recognition
in Hyalella, I tore some males away from their partners,
blackened their eyes with asphalt varnish, and placed them in a
dish with several females. It was not long before each of the
blinded males was provided with a mate. Sight, therefore, is
not the determining factor in sex recognition in this species.

2QO S. J. HOLMES.

That the females are distinguished through the sense of smell
seemed more probable, since it has been shown that among
many insects sex recognition is brought about in this way. The
sense of smell in crustaceans is often highly developed and in
some groups probably affords the means by which the females
are distinguished. The sense of smell in the Crustacea is mainly,
although not quite exclusively,1 located in the first antennae. To
determine if the male distinguishes the other sex by this sense,
recourse was had to the experiment which naturally suggested
itself, of removing the first antennae of several males and ob-
serving whether they experienced any difficulty in finding mates.
It was found that after they had recovered from the slight shock
of the operation, the males seized the females as eagerly as
before and carried them about in the usual manner. Even after
both pairs of antennse were removed the females were seized and
carried in the same way. It is very improbable, therefore, that
the sense of smell plays an important part in enabling male
Hyalellas to distinguish the other sex. The experiment was then
tried of placing several females in a small enclosure of wire gauze,
while several males which had recently been torn from females
were placed in the same dish, but outside of the enclosure. The
males paid not the slightest attention to the females within the
gauze ; but soon after the gauze was raised and the females
allowed to scatter through the dish most of the males had ac-
quired a partner.

If one attentively observes Hyalellas as they are swimming
about, it will be seen that the males do not pursue the females, great
as their eagerness may be to seize and carry one of the opposite
sex. Only when the two sexes collide in their apparently random
movements does the male become aware of the presence of the
female. When a male and a female collide, the female curls up
and lies quiet while the male makes efforts to seize her. Should
two females collide, they may curl up for a moment, but as they
are not seized they soon pass on. When two males meet there
is often a lively struggle. Each apparently attempts to seize
and carry the other, but as neither will consent to remain passive
they soon separate. The different reactions of the two sexes to
1 Bethe, Archiv. mic. Anat., Bd. 2, 1897 ; Holmes, BIOL. BULL., Vol. II., 1901.

SEX RECOGNITION AMONG AMPHIPODS. 2QI

contact with other individuals is the factor which effects the union
of the males with the females. Each reacts to the reactions of
the other. The male has a strong instinci to seize and carry
other individuals of the same species. The female has the instinct
to lie quiet when another individual comes into contact with her,
especially if she is seized. The instinctive reactions of the two
sexes are complementary and cooperate to bring about and main-
tain the peculiar sexual association characteristic of the Gamma-
ridea. If the association of the sexes is brought about by their
peculiar modes of reaction to certain contact stimuli, it would
seem probable that the only reason why males do not carry other
males as well as females is that they are prevented from so doing
by the active resistance of their intended mates. I was accord-
ingly led to try the experiment of mutilating some male speci-
mens so that they could no longer make effective resistance to
seizure. The large second gnathopods (the principal means of
defense) of several males were cut off and the mutilated indi-
viduals were placed in a dish with several males which were re-
cently torn from females. The mutilated males were soon seized
and carried about as if they were members of the other sex. In
one case a mutilated male was carried about for over five hours.
The mutilated males were more active than females are under
the same conditions, and did not assume the same bodily atti-
tude, but nevertheless their captors carried them without any
manifest awareness of the deception to which they were subjected.

Male Hyalellas, however, will not carry dead specimens of
either sex, at least for more than a short time. I have observed
males of both Hyalclla and Gammarus struggling for a time
with a dead specimen, but their efforts to carry it were soon dis-
continued. The failure to carry dead individuals may be due to
odor or some sort of chemical stimulation from the object seized,
or to the lack of an occasional movement causing a struggle on
the part of the male to retain his hold. Stimuli of the latter
kind may be necessary to cause the instinctive reaction of the
male to continue.

There can be little doubt that the origin of the instinct of the
male amphipod to seize and carry the female is to be sought in
a modification of the act of copulation. The lower Crustacea af-

2Q2 S. J. HOLMES.

ford many cases in which the association of the two sexes is
prolonged for a considerable period. The males of Artcmia
clasp the females with their peculiarly modified antennae and the
two sexes swim about together for several days (Leydig).
Among the free-swimming copepods the male may continue
clasping the female for some hours after, as well as before, de-
positing the spermatophore (Jurine, von Siebold). And among
the Cumacea Dohrn has observed the males swimming about
upon the backs of the females, much as in Amphipoda. The
tendency for the association of the sexes greatly to exceed the
act of copulation is apparently quite widespread among the
Crustacea ; and although, so far as is known, the mating instinct
of the Gammaridea is much the same throughout the group so
that we cannot trace the successive steps in its development, the
sexual behavior of some of the lower Crustacea presents many
features which may serve to throw some light upon its origin.

REGENERATION OF THE LEG OB AMPHIUMA

MEANS.

T. H. MORGAN.

My object in studying the regeneration of the limbs of Ain-
phiuma means was to discover whether the limbs, which appear
to be of so little use to the animal as organs of locomotion, have
the power to- regenerate as have the limbs of other urodele
amphibia.

The first amphiuma that I obtained (in 1900) was a large in-
dividual, and after several months had begun to regenerate, but
died as the result of an accident before regeneration had gone
very far.1 The next individual that I was able to procure was
also large, but escaped before regeneration had gone any farther
than in the last case. Two smaller individuals have been kept
for more than a year (from March 21, 1901, to May 3, 1902).
The following account applies to them. Each had a fore-leg
and hind-leg of opposite sides cut off through the upper portion
of the leg. In the course of several weeks a knob of new tissue
appeared which continued to elongate for several months, when
further growth seemed to have ceased. To make certain of this,
the animals were kept for six months longer, but no further
change occurred. The new part was shorter than the part re-
moved, and appeared to be a single rod, tapering at the end,
without any external signs of toes.

The normal fore- and hind-foot of the amphiumas that I used
had each three toes. Cope 2 gives a figure of the skeleton of
amphiuma showing a cartilaginous carpus of four or five pieces,
and three ossified metacarpals with ossified phalanges. In the
hind-foot there are three cartilaginous tarsalia, three ossified
metatarsals and three phalanges.

After the legs had regenerated they were cut off, imbedded in
paraffine, and cut into sections. These showed in three of the
four cases that the two bones of the middle part of the limb have

1 This is the case referred to in Towle's paper. BIOLOGICAL BULLETIN, II., 1901.
"Cope, " The Batrachia of North America," Bull. U. S. Nat. Mus. No. 34.

293

294

T. H. MORGAN.

developed. The condition of the carpus and tarsus appears to
be different in each of the four cases, Figs. 1-4. The rough re-
constructions shown in these figures were made from sections.
The figures are not very accurate, but serve to show the number
of bones and their relation to each other. The relative sizes of
the bones is less exact. It will be seen from the figures that the
regeneration has lead neither to the formation of a uniserial row
of skeletal elements, nor is it clear in all cases whether more
than a single toe is represented. It seems probable that the

terminal middle phalanx represents a toe, but whether any of the
other cartilages represent other suppressed toes can not be stated.

In these four cases the legs had been cut off through the
humerus, or the femur. It occurred to me that if the limb
were cut off through the fore-arm or the fore-leg the result
might possibly be different, since two bones are present at the
cut surface. Therefore on May 3, 1902, when the two regener-
ated legs were removed for study, the remaining two legs were
cut off through the fore-leg and fore-arm.

The two amphiuma were kept alive for nearly another year ;
until March 30, 1903. They were occasionally fed on earth-
worms. The limbs that had been cut off through the fore-arm
and fore-leg regenerated, but again produced only a single
pointed, or in one case a somewhat flattened, new part. Serial
sections show that, besides completing the ends of the two bones
at the exposed surface, there have been produced a number of
more distal cartilages. The arrangement of these pieces is irregu-
lar, and different in each case, as also occurred when the leg was

REGENERATION OF THE LEG OF AMPHIUMA. 295

cut off through the upper portion. In other words, no better
regeneration took place here than in the former instances.

It is also of interest to notice that the other two legs that had
been cut off (close to the body) for examination had not regen-
erated. The skin grew over the cut surface, and in several cases
the muscles of the body wall seemed to have grown over the
short piece of the humerus or femur that had been left. At
most, a short protrusion indicated the position of the limb.

How shall we interpret this result. Those who hold that the
power to regenerate a part is commensurate with the value of the
part to the animal, if it is a part liable to injury, will welcome this
experiment as in harmony with their interpretation. On the other
hand, as I have tried to show elsewhere, the evidence is so strong
against this point of view that I think we shall not go wrong if
in this case we deny that the result has any such meaning.

In fact, in other adult amphibia, in the frogs for instance, in
which the limbs are of some importance to the animal they can-
not be regenerated, although in the tadpole stage in which the
limbs are of no importance, and, in the case of the fore -limb at
least, not liable to injury, the power of regeneration is present.
Moreover even in the urodeles the power of regeneration is un-
equally developed in forms that use their legs for purposes of
locomotion. It is said that Triton marnioratus shows only a
slight power to regenerate its legs. In other cases, as I have
observed in Nectiirus, the time required to regenerate a leg is so
long that were the presence of the leg essential to the existence
of the individual it would succumb before the regeneration could
take place.

These considerations make it clear, in my opinion, that the
lack of complete power to regenerate in amphiuma can not be
interpreted as having any connection with the unimportance of
the legs to the animal. It should not be overlooked that it is
not that the leg does not regenerate at all ; in fact it regenerates
quite well, but that the new part is different from the old. It is
at least conceivable that some simple physical or physiological
factor may interfere with the formation of the complete toes,
such, for instance, as the thickness of the skin in relation to the
size of the limb.

296 T. H. MORGAN.

If it could be shown that the leg of amphiuma is a degenerate
structure it might appear that there is some connection between
the degeneracy of the part and its lack of power to regenerate,
but it is far from being established that any such general relation
really exists. In fact, in the male hermit crab I found that the
very small and apparently rudimentary abdominal appendages
have the power to regenerate. It would be interesting, never-
theless, to examine this point further in cases where the degener-
ation and uselessness of an organ are more certainly established,
as in the case, for example, of the appendix of man, which does

not appear to have the power to regenerate after removal.
WOODS ROLL, MASS., June 22, 1903.

Vol. V. November, ipoj. No. 6

BIOLOGICAL BULLETIN.

ABSORPTION OF THE HYDRANTH IN HYDROID

POLYPS.

H. F. THACHER.

In 1900 there appeared a paper1 by Professor Loeb on the
"Transformation and Regeneration of Organs," the first part of
which contained a discussion of the process of absorption in
campanularia hydroids. His results were obtained from a study
of the effects produced on the polyps by placing them in shallow
dishes of sea water, so that they were in contact with the glass ;
under these conditions he found that they were gradually trans-
formed and at length absorbed completely into the stem. To
summarize briefly Loeb's account of this process, he states that
there is noticeable first a contraction of the animal into the cup,
followed by the fusion of the tentacles and later by the with-
drawal of the whole polyp — now a shapeless mass of proto-
plasm - - into the stem. This complete transformation he ascribes
to contact, since it "is certain that contact with sea- water favors
the formation of polyps with their more solid elements, while the
contact with solid bodies favors the formation of the more fluid
material of the stem or stolon." It seemed probable that a his-
tological examination of these changes, in which the hydroid is
represented as transforming and creeping back into the stein, might
prove of interest, since they involved a complete transformation
of well-differentiated structures. Therefore, at Professor Mor-
gan's suggestion, I worked on this subject at Woods Holl during
the summer of 1902. I was able to obtain a table first through
the kindness of the director, and later was appointed to the
Bryn Mawr table.

On examining the literature it will be found that there are
frequent references to the absorption or disappearance of polyps.
Loeb finds for Margelis and Antcnnnlaria that the polyps

1 The American Journal of Physiology-, IV., 1900.

297

298 H. F. THACHER.

"disappear" when their condition of growth is disturbed — /. e.,
the former being brought into contact with a solid, the latter
being suspended horizontally so that its relation to gravity is
changed. . Eudendrium, according to some workers, sheds its
hydranths when brought into the laboratory, but I have also
often found absorption occurring under the same conditions,
and Eudendrium tcnuc, a smaller and more delicate form than
Eudendrium racemosum, responds in this way even more con-
stantly. Pt'/inaria1 has recently been examined by Cerfontaine
who finds that the day after the hydroids have been collected
" ca materialse trourait dans un mauvais etat, les polypes qui
persistaient etaient morts, les parties mollas s'etaient retirees
dans la perisarque et les extremitees du coenosarque reduit
s'etaient cicatricees. Si 1'ou conserve les branches, en mainte-
nant une circulation d'eau de mer, ou les voit souvent reprendre
de la vigueur. . . . Ou peut de cette fac_on determiner experi-
mentalement une repetition de la regeneration spontanee. A la
suite des troubles brusques produits dans les conditions d'etre de
ces organismes, par la recolte, le transport, le changement d'eau,
le changement de temperature, de lumiere, etc., ou determine
rapidement la destruction des polypes ; mais bientot, il semple se
produire une acclimation rapide, et aussitot une nouvelle regener-
ation commence." Tubularia never absorbs its polyps but sheds
them soon after being collected, and after a day or so if undis-
turbed, new polyps grow out from the old stalk, a new growth
of stalk also taking place behind the head.

It seemed possible that the absorption of the heads of Cani-
panularia might be analogous to that in these other forms, in
which case it should occur even when, not in contact with solids.
To test this, I left the hydroids still growing on bits of wood, and
placed them in the dishes, so that they were completely sur-
rounded by water. Nevertheless the polyps began to absorb and
by the end of twelve hours had almost entirely disappeared, while
a few new ones were beginning to form from the old stalks. I
also noticed on examining dishes of unused hydroids that had
been standing over night, a large percentage of absorbing polyps.

1 " Recherches experimentale sur la Regeneration et 1'Heteromorphose chez as-
troides calycularis et Pennaria Carolinii," Archives de Biologie, XIX., 1902.

ABSORPTION OF THE HYDRANTH IN HYDROID POLYPS. 299

These results show that contact cannot in any case be con-
sidered the only factor to which the absorption of campanularian
polyps is due, and that the process closely resembles that in
other polyps in which under similar conditions we find either
absorption or direct shedding of the hydranths with subsequent
regeneration.

The material for study was obtained fresh each day, so that
the animals should be in thoroughly good condition. Pieces of
Campanularia were then cut and laid in watch crystals in contact
with the glass in the way described by Loeb. The stages in the
absorption of Eudendrium and Pennaria, which I used for com-
parison, being more difficult to obtain, were taken whether in
contact or not, according to where they presented themselves.
All the material was killed in cold corrosive acetic, and stained
with Delafield's haematoxylin and congo red.

Within a few minutes after the removal of a piece or stalk, the
cut end closes over, and the digestive current begins to flow slowly
from one end of the hydroid to the other. It passes forward, and
then is driven backward mainly by the contraction of the circular
muscles of the polyps in the region just below the tentacles, but
not involving a contraction of the whole animal ; a slight pause
occurs between each change in direction. The irregularity in
the contraction of the polyps sometimes complicates the course
of the current. At first the polyps remain expanded, and the
only change noticeable is in the digestive fluid which becomes
more and more laden with spherical granules of all sizes. The
current is sometimes driven with such force that the contents
break their way through a newly formed stolon or through the
mouth of the polyp. The animal has up to this time been fully
expanded except for the rhythmic contractions which decrease
only the diameter of the body, but now it gradually contracts
into its cup, and the body becomes shorter and broader, the latter
change being largely due to the thickening of the ectoderm as
can be seen even in the living animals. The tentacles undergo
excessive contraction, becoming a crown of mere stubs, and then
disappear altogether ; their cells passing into the cavity of the
polyp. At the same time, the hypostome absorbs.

These changes take some time and normally occupy at least

300

H. F. THACHER.

two thirds of the time required for the complete disappearance
of the polyp ; sometimes the digestive current may, at this stage,
distend the degenerating polyps and delay absorption for several
hours. The usual time required is from six to twelve hours, but
under the same conditions it may last from one to two days.
The size of the structure left in the cup becomes slowly less and

D

less, and at last the tiny ball of matter is drawn into the stem.
I examined the living material carefully for signs of the breaking
of the protoplasmic threads that stretch from the ccenosarc to
the perisarc just below the cup, but I was unable in most cases
to find any trace of it, until the last stage. At that time the
strands break and the ccenosarc is drawn out in a fine thread.
The protoplasm has been under a strain for the greater part of

ABSORPTION OF THE HYDRANTH IN HYDROID POLYPS. 30 1

the time, due to the growth of the stolon, but the protoplasm of
the polyp cannot apparently be draivn through into the stem
until it has reached a certain stage in its absorption.

The finer structure of normal Campanularia is as follows :
The ectoderm cells which are flat on the body become cubical
on the hypostome ; there are no nettle cells except an occasional
wandering one, until we come to the upper half of the tentacles.
Below the cup lie masses of nettle-forming cells, somewhat
irregular in their position, but never found in an quantity anterior
to the first annulation. The endoderm is well differentiated on
the hypostome into deeply-staining goblet cells and long spindle-
shaped cells ; in the walls of the body cavity there are large,
clear endoderm cells and smaller granular gland cells. The ten-
tacles contain a single row of endoderm cells. These are sepa-
rated from those of the body cavity by a lamella at the base of
the tentacle. Signs of change first arise in the endoderm of the
body and the digestive current becomes filled with degenerating
endoderm and gland cells, pinched-off portions of cytoplasm and
loose nuclei. This process continues for some time without the
appearance of any other change, except that as the endoderm
becomes less, the lamella slowly contracts, becoming corre-
spondingly thicker, and the ectoderm, having less surface to
cover, changes from a thin layer to a much thicker one. The
tentacles have also contracted to an abnormal extent, and at last
by the breaking of the lamella across their base the endoderm
cells round up and pass out into the body cavity. At this stage
the tentacles are crowded together, and, the ectoderm being
thrown into folds by the excessive contraction, frequently give,
in surface view, the effect of being fused, as stated by Loeb. But
by careful study the independence of the tentacles can be traced
in spite of the closeness with which they are pressed together.

Soon after the endoderm has begun to pass out from the ten-
tacles the lamella breaks near the tip and masses of nettle and
ectoderm cells are poured into the cavity. The hypostome also
degenerates, the ectoderm cells passing out rapidly into the diges-
tive current and the lamella contracting after them. Soon the
lamella of the hypostome breaks and disappears and the mass of
ectoderm is also turned in. The polyp is now simply a shell of

3<D2 H. F. THACHER.

ectoderm and endoderm which are separated by the elastic
lamella, which usually meets more or less completely at the
oral end after the material of the tentacles and hypostome has
been absorbed. At this time the lamella breaks in places and
more cells from the ectoderm pass through. There is also a
small amount of degeneration on the outside, and by these
means the amount of ectoderm rapidly diminishes. Gradually
the structure becomes smaller and smaller and finally the last
fragment is drawn out of the cup. If there are many cells loose
in the body cavity of the polyp at this time, they frequently break
through the thin wall and pass out into the water.

The best guide by which to determine the amount of proto-
plasm drawn into the stem, was found to be the masses of nettle-
forming cells before alluded to. The cells really drawn represent
a very small fraction of the original number. The greater ma-
jority have been thrown into the digestive current, from which
many are absorbed by the endoderm cells throughout the entire
colony.

To compare the process in Campamtlaria with that in other
hydroids, I examined both Eudendrium and Pennaria in which
"absorption" also occurs and found the process again one of
degeneration. From the time when the first degenerating masses
are seen in the digestive current to the final drawing through of
the small degenerated mass, the method is almost identical with
that in Campanularia.

Recently there has appeared a paper by Gast and Godlewski,
Jr., on the degeneration of the polyps of Pennaria ' who have ob-
tained results similar to my own.2 It is interesting to note that
their material was taken from polyps which had regenerated'
their heads in the laboratory, and then after two or three days
had begun to absorb again — a different condition from that under
which mine were obtained, yet the process is the same. Since
these investigators have fully covered the ground for Pennaria*
I shall not describe the changes in that form and indeed merely
speak of two or three points in the degeneration of Eudendrium

1 " Ueber den Regulationsersheinungen bei Pennaria carolinii," Archiv Jiir
Entwickehtngsmechanick der Organisnnts, XVI., 1903,

2 See preliminary note, BIOL. BULL., IV., 2, 1903.

3 Probably another species.

ABSORPTION OF THE HYDRANTH IN HYDROID POLYPS. 303

that differ from that in Campanularia. The degeneration of the
endoderm is much more rapid, the cells breaking down more
completely and filling the digestive cavity with fine protoplasmic
granules. Since there is no lamella across the bases of the ten-
tacles, the endoderm can also pass out from them more readily.
The loss of ectoderm is here also accomplished by the passing in
of cells through breaks in the lamella, the edges of which are
apt to draw together again. The complete disappearance of the
lamella does not occur until a very late stage. At the end the
whole of the remaining structure is not always drawn through
into the stalk, but an ill-defined mass of protoplasm is often left
at the end.

The constant position of the ectodermal gland cells near the
beginning of the stalk throughout the degenerative changes show
that there is no drawing of cells into the stem until the final
stages.

The histological evidence thus supports my observations on
the living animals, that in Campanularia we have to do with no
transformation of the protoplasm due to contact, but with a de-
generation of the polyp. Similar changes take place in other
hydroids, and occur apparently when they are subjected to
abnormal or harmful conditions.

I wish to express my thanks to Professor Morgan for his sug-
gestions and kind supervision of my work.

FORM REGULATION IN CERIANTHUS.

II. THE EFFECT OF POSITION, SIZE AND OTHER FACTORS UPON

REGENERATION.

C. M. CHILD.

In the preceding paper (BiOL. BULL., Vol. V., No. 5, 1903),
the course of regeneration in cylindrical pieces from the middle
region of the body was described, since such pieces afford a typ-
ical result and serve as a basis for comparative study. It is de-
sired in the present paper to call attention to certain conditions
which influence the result, either as regards time or quantity.

The principal features in the regeneration of Cerianthns may
be reviewed as follows : the collapse of the piece after section
and the infolding of the ends ; the closure of the ends by new
tissue and the gradual distension of the piece and the increase in
the area of the new tissue at the ends in consequence of the
accumulation of water in the enteron, probably by diffusion
through the body-\vall ; the reduction and disappearance of the
muscular layer and pigment at both ends ; the regeneration of
mesenteries ; the outgrowth from the tentacular ridge of a mar-
ginal tentacle over each intermesenterial chamber ; the formation
of the mouth in the directive radius ; the appearance of the
labial tentacles in a circle upon the disc ; the outgrowth of new
tissue at the aboral end of the piece.

Since each of these processes is gradual it is impossible to de-
termine with exactness the time of its beginning ; moreover, the
various processes overlap and are connected in such a manner
that it is difficult to separate distinct stages except arbitrarily.
For these reasons the comparison of different pieces with a view to
determining the conditions which effect regeneration can best be
accomplished by the examination of these pieces at stated times,
rather than by noting the time at which a given piece arrives at
a particular stage. The former method not only allows direct
comparison of the pieces, and thus often renders the detection of
slight differences less difficult, but it obviates the necessity for

304

FORM REGULATION IN CERIANTHUS. 305

almost continuous observation and the accompanying manipula-
tion necessary to examination, which is a source of irritation to
the regenerating pieces and may often effect the result by caus-
ing rupture of new tissue or other injuries.

In general then the method pursued in the experiments was
that of examining at intervals pieces to be compared and noting
the condition of each. Owing to the number of points to be
observed and the necessity for indicating slight differences any
arrangement of the results in tables is unsatisfactory : they are
given, therefore, in much the same manner as first recorded. The
series of experiments described are selected from a large number
but the results were remarkably uniform in all cases. In
the description of the stages only the most salient features of
the regeneration are mentioned in most instances. In all cases,
however, unless definite statement is made to the contrary, regen-
eration proceeded in the typical manner.

I. DESCRIPTION OF EXPERIMENTS.

SERIES 22. '

September 24, 1902. The oral end, including the cesophageal
region, was removed from twenty-three large specimens of C.
solitarius and the remaining portion of the body was divided by
a transverse cut as nearly as possible into two equal pieces (Fig.
i), oral halves being designated A, aboral B. All of the pieces
A were placed in one aquarium, all of B in another.

September 27 : Three days after section :

A. Most of the pieces are still collapsed, but in a few the ends
are closed and a slight distension with water is evident.

B. All still collapsed.

September 28 : Four days after section :

A. All the pieces are more or less distended with water : three
pieces show the tentacular ridge and the first traces of marginal
tentacles.

B. Three pieces are closed and somewhat distended, the re-
mainder still collapsed.

September 29 : Five days after section :

1 The series retain the numbers given them in my notes.

306

C. M. CHILD.

A. Distinct marginal tentacles are present on eight pieces ;
the remainder all distended and with tentacular ridge.

B. All closed and more or less distended ; in a few distension
is just beginning ; none with distinct tentacles.

September 30 : Six days after section :

A. All with distinct marginal tentacles from 0.5— i.omm. long.

B. The pieces which were the first to close and become dis-

tended show traces of marginal tentacle buds; all
pieces distended with water.

October I : Seven days after section:

A. All with marginal tentacles 1.0-2.0 mm. long.

B. Traces of marginal tentacles on all pieces
except those which were the last to close.

This series was not kept under observation for
the later stages. As regards the earlier stages,
however, it shows clearly that the aboral pieces
regenerate somewhat less rapidly than the oral
pieces, although the latter are cut at both ends,
the former at only one end. The difference be-
tween the two sets of pieces is universal, not even
the most advanced pieces in the set B showing as
rapid regeneration as the least advanced of A. In
general the differences between pieces of the same
set are slight.

SERIES 45.

November 7, 1902. Tentacles and disc were removed from
four specimens and the remaining portion of the body was cut
into four pieces, A, B, C, D, as nearly equal as possible (Fig. 2).
All of the pieces A were placed in one aquarium, all of B in
another, etc. The pieces A contained a part of the oesophagus.

November 9 : Two days after section :

A. Ends closed and piece distended ; as in other similar cases
the cut oral margin of the oesophagus has united with the oral
margin of the body-wall so that the pieces possess a well-
developed mouth-opening.

B, C, D. All collapsed.

November 10 : Three days after section :
A. Marginal tentacular ridge appearing.

B

\J

FIG. i

FORM REGULATION IN CERIANTHUS.

30/

B. One piece beginning to fill with water ; others collapsed.

C. One piece beginning to fill ; others collapsed.

D. All collapsed.

November i 2 : Five days after section :

A. Marginal tentacular ridge distinct, with fading pigment.

B. All distended ; new tissue at ends visible ; ten-
tacular ridge forming.

C. One piece distended ; new tissue at ends visible ;
two pieces partly filled but not sufficiently to show
the new tissue at the ends ; one piece still col-
lapsed.

D. All still collapsed.

November 1 5 : Eight days after section :

A. Marginal tentacles just appearing in all.

B. One piece with marginal tentacles just appear-
ing ; three pieces distended ; new tissue at ends visible ;
tentacular ridge distinct, unpigmented.

C. Two pieces distended ; new tissue at ends visi-
ble; tentacular ridge forming ; one piece still col-
lapsed ; one piece collapsed and completely enclosed
in slime which was removed.1

D. All still collapsed.

November 20 : Thirteen days after section :

A. Marginal tentacles in all I mm. in length.

B. In one piece marginal tentacles i mm., in others about
0.5 mm.

C. All distended ; tentacular ridge unpigmented, marginal
tentacles just appearing.

D. One piece beginning to fill, others collapsed.
November 25 : Eighteen days after section :

A. Marginal tentacles 2-2.5 mm- showing faint traces of pig-
mented bands in two pieces, in other two unpigmented ; labial
tentacles 0.5 mm. At aboral end new outgrowth 2 mm.

B. Marginal tentacles 1—2 mm., some differences in length ap-
pearing in individual pieces, unpigmented ; a few labial tentacles

1 The complete enclosure of pieces in slime, as in a cyst, often occurs when they
remain collapsed for more than four or five days. The slime being secreted all over
the ectoderm unites at the infolded ends and forms a complete cyst from which the
piece is unable to emerge after a few days.

D

FIG. 2.

'

308 C. M. CHILD.

just appearing in each piece. At aboral end new outgrowth
about 2 mm.

C. In two pieces marginal tentacles 0.5-1.0 mm., in other two
slightly less advanced ; in all unpigmented : none with labial
tentacles. At aboral end no distinct outgrowth of new tissue.

D. One piece partly filled, others collapsed.
December 2 : Twenty-five days after section :

A. Marginal tentacles about 5 mm., with distinct transverse
pigment bands ; labial tentacles 1-1.5 mm- Aboral end as before.

B. Marginal tentacles 3.5-4 mm., pigment bands visible but
lighter than in A; labial tentacles 0.5-1 mm. Aboral end as
before.

C. Marginal tentacles mostly 3 mm., a few in two pieces 4
mm., all unpigmented ; labial tentacles just appearing. At
aboral end no distinct outgrowth of new tissue.

D. One piece partly filled, but enclosed in slime which was
removed ; others still collapsed.

December 12 : Thirty-five days after section :

A. Marginal tentacles 6-7 mm., pigment bands dark: labial
tentacles about 2 mm. At aboral end outgrowth of new tissue
2-3 mm.

B. Marginal tentacles 5-6 mm.; pigment bands lighter than in
A; labial tentacles 1-1.5 mm- At aboral end outgrowth of
new tissue 2—3 mm.

C. Marginal tentacles 4-5 mm.; pigment bands visible, but
slightly -lighter than in B ; labial tentacles about I mm. At
aboral end outgrowth of new tissue I mm.

D. One piece closed and partly filled as before, but no traces of
tentacular ridge. Three pieces still collapsed and enclosed in
slime which was removed.

The series as a whole was concluded at this time, since the
only further changes in A, B and C consist of a slight increase
in length of the tentacles and the pigmentation. The pieces D,
however, which had not as yet shown any traces of regenerating
tentacles were kept under observation until January 21, 1903.
Up to this time only the one piece which had become partly
filled showed any signs of regeneration, the others remaining com-
pletely collapsed and surrounded by slime, which was removed

FORM REGULATION IN CERIANTHUS. 309

from time to time in order to permit distension to occur if there
were any tendency. The changes during this time in the one
piece which was closed and partly filled are of considerable in-
terest. At one side of the closed oral end of the piece a few
minute outgrowths 0.2-0.5 mm. in length made their appearance.
They resembled marginal tentacles and were situated where
these organs should appear, but there were only a few of them
close together on one side and no others appeared. Fig. 3
shows the piece as it appeared January 21.
The new tissue closing the end is indicated
by the stippling. At one side are six
small outgrowths resembling tentacles,

but no traces of any others can be found

FIG. 3.

at any point of the circumference. At the

conclusion of the experiment the piece was opened and it was
found that a few of the longest mesenteries extended into the
piece in the radius in which the outgrowths appeared. This
region is then without doubt the region of the directive mesen-
teries, and the mesenteries present are simply the longest mesen-
teries of the body which lie to the right and left of the short
directives and extend nearly to the aboral end. The small out-
growths correspond in position with the spaces between these
mesenteries and there can be little doubt that they represent
marginal tentacles. No other mesenteries are present in the
piece, none having regenerated. It becomes evident from the
history of this piece that the presence of mesenteries is necessary
for the regeneration of marginal tentacles. In pieces from
regions nearer the oral end mesenteries are regenerated, but in
this piece no trace of regenerated mesenteries could be found,
and tentacles have begun to regenerate only in the spaces be-
tween such of the old mesenteries as extended into the piece.

The series as a whole affords several results of importance.
As in the preceding series, the decreasing rapidity of regeneration
with increasing distance from the oral end of the body is clearly
shown. The pieces A regenerate more rapidly than B, B more
rapidly than C, and finally D, the aboral pieces, are capable of
only a slight degree of regeneration or of none at all, the differ-
ence between the one piece which regenerated a few tentacles and

3IO C. M. CHILD.

the others probably being due to the fact that the cut separating
C from D in the one case was slightly more oral than in the
other three.

The differences in the rapidity and the amount of regeneration
are best shown at the oral ends of regenerating pieces, for it is
difficult to determine with exactness the amount of actual new
tissue at the aboral ends of the pieces, since the line of demarca-
tion between the unpigmented tip and the normally pigmented
regions oral to it is not at all sharp, extending in many cases
over two to three millimeters. As regards the aboral ends the
pieces A and B showed little difference, but regeneration at the
aboral end of C was in all cases distinctly less than in A and B.

In general the series seems to indicate that not only is regen-
eration less rapid with increasing distance from the oral end, but
that there is a corresponding difference in the amount of regen-
eration. In the pieces A, B and C the differences are compara-
tively slight, though without doubt present as can be seen by
comparing the data for these pieces thirty -five days after section.
When the pieces D are taken into consideration, however, the
difference between these and all other pieces is marked, for in no
case did these aboral ends show anything approaching complete
regeneration. There is then, according to these results, a rapid
decrease in regenerative power near the aboral end of the body,
and apparently complete absence of this power in an aboral region
representing approximately one fifth of the body-length. As
will be shown below, much smaller pieces than this from other
regions of the body are capable of complete regeneration ; more-
over, the size of the area within which regeneration does not
occur differs according to conditions.

SERIES 54 AND 55.

December i 5, 1902. The tentacles, disc and cesophageal region
were removed from twenty large specimens ; ten of the remain-
ing pieces were then divided by a transverse cut into two pieces,
A and B (Fig. 4), the cut being made near the aboral end so
that the pieces A comprised the greater part of the body aboral
to the oesophagus, while the pieces B represented the extreme
aboral end, about one sixth of the body-length. These two sets
of pieces constituted Series 54.

FORM REGULATION IN CERIANTHUS.

Each of the remaining ten pieces was also divided by a trans-
verse cut into two pieces A and B, but in this case the cut was
made near the oral end of the piece (Fig. 5), so that A repre-

\J

FIG. 4.

V
FIG. 5.

sents one sixth of the body-length from the region just aboral
to the oesophagus, while B is the whole remaining portion of the
body. These pieces constitute Series 55.

In this manner four sets of ten pieces each were obtained.
The oral ends of pieces 54A and 55A represent approximately
the same level in the body of the individuals from which they
were taken, while their aboral ends lie at very different levels ;
moreover, the pieces 54A are about four times as long as 5 5 A.
The oral ends of the pieces 546 and 55B are at very different
levels, while their aboral ends are the aboral ends of the parent
bodies ; 556 is about four times as long as 546.

In 54A and 54B we have pieces differing widely in size and
with oral ends at very different levels ; the same is true of 55 A
and 556, but the relations of size are reversed.

The comparative study of the regeneration of those four sets
of pieces should afford data regarding the rapidity and amount
of regeneration at different levels of the body and in pieces of
different sizes.

December 19 : Four days after section :

312 C. M. CHILD.

54A. All with closed ends and filling with water, but ends not
yet expanded so as to show new tissue.

546. All collapsed.

5 5 A. All with closed ends and filling with water, but ends not
yet expanded. Condition same as in 54A.

55B. Ends closed, but pieces contain less water than 54A
and 5 5 A.

December 22 : Seven days after section :

54A. All distended, ends expanding, new tissue visible ; ten-
tacular ridge just appearing.

546. All collapsed.

5 5 A. Similar to 54A.

536. All filling with water, but none distended sufficiently to
expand the ends and show new tissue.

December 24 : Nine days after section :

54A. Tentacular ridge distinct, with fading pigment.

546. All still collapsed.

5 5 A. Similar to 54A.

55B. All distended, ends expanded and tentacular ridge just
appearing.

December 26 : Eleven days after section :

54 A. Marginal tentacles just appearing in some specimens as
minute outgrowths from tentacular ridge, which is now unpig-
mented.

546. All still collapsed.

5 5 A. Similar to 54A.

55B. Tentacular ridge with fading pigment, but less distinct
than in 54A ; no marginal tentacles.

December 28 : Thirteen days after section :

54A. Marginal tentacles 0.25-0.5 mm.

546. All still collapsed.

5 5 A. Similar to 54A.

556. Tentacular ridge distinct, unpigmented in most cases;
in a few the earliest traces of marginal tentacles visible.

January 3, 1903 : Nineteen days after section :

54A. Marginal tentacles 2—3 mm.

546. All still collapsed, much contracted, rounded in form.

5 5 A. Similar to 54A.

FORM REGULATION IN CERIANTHUS. 313

556. Marginal tentacles 1 — 1.5 mm.; in one specimen with
rather unequal tentacles a few 3 mm.

January 1 1 : Twenty-seven days after section :

54/\. Marginal tentacles 5—6 mm.; transverse bands of pig-
ment distinct; labial tentacles 1 — 1.5 mm. At aboral ends no
well-marked outgrowth of new tissue ; ends slightly lighter in
color at region of closure.

546. All still collapsed, much contracted, rounded in form.

5 5 A. Oral ends similar to 54A. At aboral ends a distinct
outgrowth of new tissue 2—3 mm.

55 B. Marginal tentacles 3—5 mm.; pigmentation of tentacles
slightly less deep than in 54 A ; labial tentacles just visible — I
mm.

January 2 1 : thirty-seven days after section :

54 A. Marginal tentacles 7-8 mm.; labial tentacles 1-1.5
mm. No distinct outgrowth of new tissue at aboral ends.

546. Collapsed, rounded and still further reduced in size.

5 5 A. Marginal tentacles 6-8 mm.; labial tentacles 1-1.5
mm. At aboral ends distinct outgrowth of new tissue 3-5 mm.
At this time the average length of the marginal tentacles in these
pieces is somewhat less than in 54 A. In the latter cases there
are fully as many specimens with tentacles 8 mm. in length as
with tentacles 7 mm. In 55 A, however, only a few pieces, and
these the largest, possess marginal tentacles 8 mm. in length ; in
nearly all the marginal tentacles are 6—7 mm. Moreover, the
average length of the labial tentacles in 55 A is slightly less than
in 54 A. These pieces are evidently falling behind the longer
pieces.

556. Marginal tentacles 5—6 mm., somewhat less deeply
pigmented than in 54 A ; labial tentacles i mm.

At this time the regenerated structures had acquired their
maximum size ; afterward reduction in size, which always occurs
in the pieces kept without food, began. For present purposes it
is not necessary to follow the history of these pieces further.

Comparison of the data afforded brings to light a number of
interesting results. Comparing the rapidity of regeneration in the
different pieces, it is seen that the oral ends of pieces 54 A and
5 5 A, which represent approximately corresponding regions of

3 14 C. M. CHILD.

the parent body, regenerate with equal rapidity except at the end
of the experiment, although pieces 54 A are about four times as
long as pieces 55 A. The oral ends of pieces 556, which repre-
sent a region of the parent body further aboral than those of

54 A and 55 A, regenerate less rapidly than these, although the
pieces are about equal in size to 54 A and four times as long as

55 A. And finally, the pieces 546, whose oral ends represent a
region near the aboral end of the parent, do not regenerate at all.

As regards the aboral ends of the pieces only 54A and 55 A
need be considered, since no regeneration occurs at the aboral
end of a piece when this represents the aboral end of the parent-
body, as is the case in 546 and 556. In 54A and 55A the dif-
ference in the rapidity and amount of regeneration at the aboral
ends is marked; in 5 5 A, where the aboral ends of the pieces
represent a region oral to the middle region of the parent-body
the aboral regeneration was much greater than in 54A, where
the aboral ends represent a region near the aboral end of the
parent-body, even though the pieces 54A were four times as
long as 5 5 A.

From all of these facts it is evident that the rapidity and amount
of regeneration decrease as the cut surface, either oral or aboral,
approaches the aboral end of the parent-body, and that the size
of the piece has no marked influence, at least within the limits
of size of the present experiment. That the size of the piece
does, however, affect the final result in some degree is shown by
the condition of pieces 54A and 55A at the end of the experi-
ment 37 days after section ; while no differences between the two
sets were noted earlier it was found at this time that the smaller
pieces 55A were falling slightly behind the larger 54A. Here
then a slight influence of size is noticeable, though only in the
later stages of the experiment. As will be shown later this re-
sult is confirmed by other cases. In pieces above a certain min-
imal size regeneration is not influenced by the size, except in the
later stages.

SERIES 35.

October 20, 1902. In this case after removal of disc and
tentacles a single specimen was cut into four pieces, A, B, C, D
as shown in Fig. 6. The piece B was much smaller than the

B

FORM REGULATION IN CERIANTHUS. 315

others and masses of the mesenterial filaments protruded from
each end, thus delaying the closure and normal regeneration ; it
is therefore omitted from the present consideration. The pieces
A and C are nearly equal in length and are about
two thirds the length of D.

October 22 : two days after section : All pieces
still collapsed.

October 23 : three days after section:
A. Margin of oesophagus united with body-wall,
aboral end closed and enteron partly filled with
water.

C. and D. Both still collapsed.
October 24 : Four days after section :
A. Sufficiently distended with water to spread
the inrolled margins and allow the oesophagus at
the oral end and the new tissue closing the aboral
end to become visible. I /

C. and D. Both still collapsed.

FIG. 6.

October 25 : Five days after section :

A. Distended with water : tentacular ridge visible and pigment
disappearing from it.

C. Ends closed by new tissue ; distended.

D. Enteron partially filled with water ; distension not yet suf-
ficient to separate the infolded oral margins and permit new tissue
to become visible.

October 27 : Seven days after section :

A. Marginal tentacles just appearing on tentacular ridge.

C. Tentacular ridge distinct ; its pigment disappearing.

D. Distended with water ; new tissue closing oral end ex-
posed by separation of cut margins in consequence of distension :
tentacular ridge visible, with fading pigment.

October 29 : Nine days after section :
A. Marginal tentacles I mm.

C. Marginal tentacles just appearing on tentacular ridge.

D. Tentacular ridge distinct, without pigment ; no tentacles
visible.

October 3 1 : Eleven days after section :

A. Marginal tentacles 2 mm. At aboral end new tissue grow-
ing out in a small point 1.5 mm.

316 C. M. CHILD.

November 6 : Seventeen days after section :
A. Marginal tentacles 5 mm.; labial tentacles 0.5—1 mm. At
aboral end outgrowth of new tissue 2 mm.

C. Marginal tentacles 3-4 mm.; labial tentacles 0.5. At aboral
end of outgrowth of new tissue I mm.

D. Marginal tentacles 1-2 mm.; labial tentacles not yet visible.
November 1 2 : Twenty-three days after section :

A. Marginal tentacles 10 mm.; distinctly marked with the
characteristic transverse bands ; labial tentacles 2—3 mm. At
aboral end margins of old body-wall are becoming involved in
the growth and losing pigment ; the unpigmented area, includ-
ing outgrowths, is about 3 mm. in length from aboral end.

C. Marginal tentacles 6—7 mm.; transverse pigment bands
visible but not dark as in A ; labial tentacles 2 mm. At aboral
end new outgrowth 1.5 mm.

D. Marginal tentacles 5—6 mm. still unpigmented ; labial ten-
tacles i mm.

November 20 : Thirty-one days after section :
A. Marginal tentacles 12-13 mm.; pigment bands dark and
distinct ; labial tentacles about 3 mm. Aboral end as before.

C. Marginal tentacles 8—9 mm.; pigment bands distinct but
less dark than in A; labial tentacles 2—2.5 mm. At aboral end
unpigmented area 2—3 mm. in length.

D. Marginal tentacles 6—8 mm.; pigment bands visible but less
dark than in C ; labial tentacles 2 mm.

December 2 : Forty -three days after section :
A. Marginal tentacles 12—13 mm.; labial tentacles 3— 4 mm.
At aboral end the new outgrowth is becoming pigmented.

C. Marginal tentacles 10 mm.; pigmentation of tentacles
scarcely distinguishable from that of A ; labial tentacles 3 mm.
At aboral end the unpigmented area about 3 mm.

D. Marginal tentacles 12—13 mm.; pigmentation slightly less
dark than that of A ; labial tentacles 3-4 mm.

At this time regeneration is essentially complete in the pieces ;
no further increase in the length of tentacles, or of the new growth
at the aboral end occurs. The marginal tentacles of C and D
are still slightly lighter in color than those of A, and the pigment
has not yet extended over the aboral outgrowth in C as far as in
A, but these slight differences are later eliminated.

FORM REGULATION IN CERIANTHUS. 3 I 7

Examination of the data shows that at all stages except the
final A is more advanced in regeneration than C, and C more ad-
vanced than D.

It will be noted also that the regenerated parts of piece A did
not increase in size after 3 1 days, with the exception of the labial
tentacles which showed a slight increase between 31 and 43 days.
In the piece C a slight increase in the length of all tentacles oc-
curred between 3 1 and 43 days. In the piece D, however,
there was a marked growth during this time. In other words
the piece A completed its regeneration first, then the piece C,
and last of all the piece D.

Throughout this series then there is a distinct relation between
the rapidity of regeneration and the position of the pieces in the
parent-body, the rapidity of regeneration decreasing with increas-
ing distance from the oral end.

One other point requires consideration : the regenerated tenta-
cles of the piece D finally attain the same length as those of piece
A. This would appear at first glance to contradict the results
obtained from other series of experiments where not only the
rapidity but the amount of regeneration diminishes toward the
aboral end. Comparing A and C, two pieces about equal in size,
we find that the amount of oral regeneration in A is greater than
in C, as might be expected from comparison with other series,
since C represents a region farther from the oral end of the
parent-body than A. The piece D, still nearer the aboral end
of the parent-body, but much longer than A and C, while regen-
erating more slowly than either of these finally equals A in the
amount of regeneration. Apparently in this case the influence
of size has counterbalanced the influence of position. If piece
D was of the same size as A and C the amount of oral regenera-
tion would undoubtedly be less than in those pieces, but since it
is much larger, i. e., contains much more available material, re-
generation continues for a somewhat longer time (note the in-
crease in size of tentacles in D between 3 1 and 43 days) and the
regenerated organs finally, though after a longer time, reach a
condition similar to that in A. In this case, as in Series 54 and
55, the influence of size is slight and appears only in the latest
stages of regeneration.

C. M. CHILD.

SERIES 56.

December 15, 1902. Disc, tentacles and oesophageal region
were removed from ten large specimens by a transverse cut ab-
oral to end of oesophagus. The aboral piece was then cut into
two pieces, A and B, of equal length (Fig. 7) which
were kept for comparison.

December 19 : four days after section :

A. Nine pieces with ends closed ; a few dis-
tended, the others partly filled with water ; one
piece still collapsed.

B. All still collapsed.

December 22 ; seven days after section :

A. All distended ; ends well expanded, showing
new tissues in a few pieces the tentacular ridge is
just appearing.

B. All still collapsed.

December 26 : Eleven days after section :
A. Tentacular ridge distinct, with faded pig-
ment ; in a few pieces the first traces of marginal
tentacles distinct.

B. Filling with water but not distended ; new tissue at oral
end not visible.

December 28 : Three days after section :

A. Marginal tentacles 0.5 mm.

B. Four pieces fairly well filled with water ; tentacular ridge
just visible ; six pieces collapsed or only partly filled ; tentacular
ridge not visible.

January 3, 1903 : Seventeen days after section :

A. Marginal tentacles 2-3 mm.

B. One piece distended ; marginal tentacles 2 mm. Nine
pieces partly or completely collapsed ; no tentacular ridge or
tentacles visible.

Circumstances necessitated the conclusion of the series at this
time, so that it was impossible to determine whether the nine
pieces of B would ever have regenerated. The series affords,
however, some interesting results. As in all other series regen-
eration is much less rapid in the aboral pieces ; in only one case
did the aboral pieces regenerate tentacles before the conclusion

FORM REGULATION IN CERIANTHUS. 319

of the experiment. Examination of the data shows that the
pieces B were all filling with water eleven days after section ; that
two days later all but four were collapsed, and that, finally,
seventeen days after section, only one piece was filled with water.
These changes are undoubtedly due to the fact that the growth
of new tissue at the ends of these pieces failed to keep pace with
the pressure of water in the enteron, and so rupture occurred as
soon as the pieces reached a certain point. In only one case
did the new tissue remain intact, viz., the case in which tentacles
appeared.

As is evident from a comparison with other series, viz., series
22, 35, and 45, regeneration was found to occur in other cases in
pieces representing about the aboral third of the body, though
pieces representing the aboral fifth (series 45) or less did not
regenerate. Why then did regeneration fail to occur in the
aboral pieces of the present series ? The difference is un-
doubtedly to be accounted for by the low temperature of the
water. This series was begun in December and continued into
January. The temperature of the water was very much lower at
this time than during the autumn, and several other series begun
on the same date showed similar results. In other words, the area
at the aboral end which is incapable of regeneration increases
as the temperature becomes lower, and in the present series in-
cludes more than the aboral third of the body. This point will
be discussed in a following section where the influence of tem-
perature is considered.

( To be Continued.}

ARTIFICIAL MIXED NESTS OF ANTS.

ADELE M. FIELDE.

Mixed nests of ants are rarely found in nature, and the ants
associated in such nests are always of the same subfamily if not
of the same genus.1

There are two ways of causing ants of different genera, or even
of different subfamilies, to live peacefully together. One way is
that of destroying the sense of smell in the ants by depriving
them of a portion of the antennae. Forel discovered, in the
seventies, that the funicles were the organs of smell. I have
had representatives of three subfamilies of ants, all without funi-
cles, living amicably together through several consecutive weeks,
although the members of the group varied in size, from the huge
Camponotus pennsylvanicus to the small Stcnamma fulvuin ; in
form, from the shark-like Stigmatouima pallipcs to the chubby
Lasius umbratus ; in color, from the jet-black Crcinastogaster
lineolata to the amber-yellow Lasius latipes ; and in character, from
the truculent Myrmica rnbra to the patient Formica snbscricca.

In 1901, \\s\v\<gSteuamviafulvum for the experiments, I located2
the appreciation of the nest-aura in the distal segment of the
funicle, the eleventh ; that of the colony, in the tenth segment ;
that of the individual track, in the ninth segment ; that of the
inert young, in the eighth and seventh segments. I have lately
located the appreciation of the odor of enemies in the sixth and
fifth segments.

I cut off the five distal segments of the antennae from seven
queens3 of Steuamma fulvnm, seven queens of Crcmastogastcr
lineolata, five queens of Mynnica rnbra, five queens of Lasius

1 E. Wasmann, " Die zusammengesetzten Nester und gemischten Kolonien der
Ameisen," 1891. William Morton Wheeler, " The Compound and Mixed Nests of
American Ants," American Naturalist, 1901.

2 A. M. Fielde, "Further Study of an Ant," Proceedings of the Academy of
Natural Sciences of Philadelphia, November, 1901.

3 Among the Myrraicine ants, queens only were used for these experiments, because
of the abnormal irritability of myrmicine workers lacking parls of the antennae.

320

ARTIFICIAL MIXED NESTS OF ANTS. 321

umbratus, seven workers of Lasins latipes, five workers of Cam-
ponotus pennsyhanicus, four workers of Formica sanguine a, four
workers of Formica subsericea, and three workers of Stigmatoinma
pallipes, and when these ants had recovered from shock-effect,
with healed wounds, I placed them all in an artificial nest, roomy
for their number, having thirty-two square inches of floor-space.
Duels were constant, and in two hours there were but twenty-
three survivors from the forty-seven ants. Several of the sur-
vivors were disabled.

I then formed a new group, with other ants, having the four
proximal segments of the funicle intact. This group included rep-
resentatives of the Camponotines, Camponotus Pennsylvania is and
Formica sanguinca ; of the Myrmicines, Stenamma fnlvnm and
Cremastogaster lincolata, and of the Ponerines, Stigmatoinma pal-
lipes. These lived peacefully together many days, in one of my
small Petri cells, and ants of different subfamilies often huddled
together. In this cell I saw a queen of the Stenammas lapping
regurgitated food from the mouth of a Camponotus worker.

In another mixed group, made up of ants retaining from three
to six segments of a funicle, I removed and examined every
ant that attacked one of another species, and found that all such
ants retained more than four segments of the funicle.

We may, then, secure peaceful mixed nests by depriving the
inmates of certain segments of the antennae.

I have lately created many mixed nests by another method,
that of educating the ants in ant-odors unlike their own.1 If one
or more individuals, of each species that is to be represented in
the future mixed nest, be sequestered within twelve hours after
hatching, and each ant so sequestered touch all the others with
its antennae during the three ensuing days, these ants will live
amicably together thereafter, although they be of different colo-
nies, varieties, species, genera or subfamilies. For sequestering
the ants, I used artificial nests, made in watch-glasses so small
that the natural movement of the newly-hatched ants would
brine each of them into contact with all the others. In no case

o

did the callows quarrel, and those of most diverse lineage some-

1 The experiments were made at the Marine Biological Laboratory at Woods Hole,
Mass., in July, August and September, 1903.

322 ADELE M. FIELDE.

times snuggled one another. The ant's sense of smell appears
to be perfectly acquired, and its standards of correct ant-
odor to be established during the first three days after hatching.
Any two species or any number of species that I captured for
use in these experiments, became accustomed to each other's
odor, and therefore friendly, if the early association was close
and continuous. This association is more perfect when no inert
young distracts the attention of the callows from one another,
and when the arrangement of the nest offers no place of seclusion
for any of its inmates. Air, humidity and nourishment were pro-
vided as in large nests of the Fielde pattern. When the ants
had been thus segregated for five days or more, the inmates of
several like nests were transferred to a more spacious habitation,
and newly hatched ants from the same colonies could be safely
added thereafter ; but no ant of other lineage nor of greater age
was amicably received in any of the mixed nests.

Each of the groups mentioned in the following list existed
under my care for a month or more after the cessation of addi-
tions of newly hatched ants to their mixed nest.

MYRMICINE ANTS.

Group i. — Six queens of Creniastogaster lineolata with eighty
workers of Stenamma fiihnui. The workers snuggled the
queens as closely as if of the same species as themselves. In
each of two watch-glass nests, the sole queen died on the third
day after hatching. Newly-hatched queens of the same Cremas-
togaster stock were accepted by the bereaved workers,1 in the

1 After these ants in group I had been established for two days in a Fielde nest, a
raid was made upon them by adult workers, of the queens' stock, that had escaped
from the hatchery-nest, and hidden in a crevice in the laboratory. Very early one
morning, I discovered that these adult Cremastoeaster workers had entered the nest

D ' O

in considerable numbers through a rift in the towelling. Some of them were cluster-
ing around the young queens, while others were busily employed in dragging the Ste-
namma callows out of the nest. My arrival thwarted an apparent design of the
Cremastogasters to eject the Stenamma! and dwell in an unmixed nest with queens
of their own.

This first group is noteworthy, because Stenamma fulvum and Creniastogaster
lineolata will each feed their larvae upon the eggs, larvse and pupae of the other.
In one of my artificial nests the Stenammas lately took care, with their own young, of
a great number of Creniastogaster larvae and pupae, during two months ; but every
Creniastogaster that hatched was instantly killed and cast upon the rubbish-heap.

ARTIFICIAL MIXED NESTS OF ANTS. 323

•

one case three days, and in the other case five days, after the
death of their first queen. I know no adult ants that will ac-
cept a queen from another colony of their own species, much
less a queen of a genus not their own.

Group 2. — Myrmica r libra, Stcnamma fulvum and Cremasto-
gaster lineolata, workers of each species.

ONE SPECIES OF CAMPONOTINE ANTS WITH ONE SPECIES
OF MYRMICINE ANTS.

Group j.-- Lasius latipes with Stenamma fulvum ; workers of
each species.

Group 4.-- Lasius umbratus with Stenamma fulvum ; workers
of each species.

Group J. - - Lasius umbratus with Cremastogaster lineolata ;
workers of each species.

Group 6. - - Formica sanguiuca with Cremastogaster lineolata ;
workers of each species.

Group 7. — Formica subsericea with Cremastogaster lineolata ;
workers of each species.

ONE SPECIES OF CAMPONOTINE ANTS WITH TWO SPECIES OF

MYRMICINE ANTS.

*

Group 8. — Formica sanguinca with Stcnamma fulvum and
Cremastogaster lineolata ; workers of each species.

Two SPECIES OF CAMPONOTINE ANTS WITH TWO SPECIES

OF MYRMICINE ANTS.

Group 9. — Lasius latipes and Formica lasiodes l with Stenamma
fulvum and Cremastogaster lineolata ; workers of each species.

Group 10. — Camponotus pennsylvanicus and Formica sanguined
with Stenamma fulvum and Cremastogaster lineolata.

ONE SPECIES OF CAMPONOTINE ANTS WITH THREE SPECIES

OF MYRMICINE ANTS.

Group ii. — Lasius latipes with Stenamma fulvum, Myrmica
rubra and Cremastogaster lineolata ; workers of all four species
with one queen of Cremastogaster lineolata.

1 Kindly identified for me by Dr. W. M. Wheeler.

324 ADELE M. FIELDE.

•

THREE SPECIES OF CAMPONOTINE ANTS WITH ONE SPECIES

OF MYRMICINE ANTS.

Group 12. — Camponotus pennsylvanicus, Formica sanguine a
and Formica subsericea with Stenamma fulvum ; workers of each
species.

ONE SPECIES OF PONERINE ANTS WITH ONE SPECIES OF

MYRMICINE ANTS.

Group ij. — Stigmatomma pallipcs with Stenamma fulvum ;
queens of the former with workers of the latter.

ONE SPECIES OF PONERINE ANTS WITH ONE SPECIES OF

CAMPONOTINE ANTS.

Group ij. — Stigmatomma pallipes with Formica subsericea ;
workers of each.

ONE SPECIES OF PONERINE ANTS, ONE SPECIES OF CAMPONOTINE
ANTS AND ONE SPECIES OF MYRMICINE ANTS.

Group 75. — Stigmatomma pallipes, queens and workers, with
workers of Formica subsericea and of Stenamma fulvum.

In my artificial mixed nests, there is a close affiliation of ants
of different species. Those of different subfamilies sometimes
lick one another. Introduced young is carried about and taken
care of without regard to its origin. Ants of one genus accept
regurgitated food from those of another genus.

Ants appear to associate readily with all harmless familiars.
In the wild nests of Stenamma fuhnim I often see gray sowbugs
roaming about, and they do not molest the ants, nor are they
molested by the ants. On my putting a sowbug into an artificial
nest of these ants, they seemed to treat it sportively, two or three
young ants sometimes mounting upon its back and riding there,
like children making excursions on an elephant. In my artificial
mixed nests, small ants often ride on large ones, or stand on
their backs and lick their heads.

Natural mixed nests probably originate among ants that seek
in their abodes the same degree of moisture and of warmth. The
habitat of each species being determined by the food-supply, the
humidity and the temperature, any two species finding the same

ARTIFICIAL MIXED NESTS OF ANTS 325

habitat a congenial one, might form a mixed nest through an
accidentally close association of their newly-hatched members.

Were the occupants of my artificial nests free to seek the
habitation most agreeable to each species, they would doubtless
soon separate. Perhaps they would never quarrel with each
other on meeting ; but they would certainly fight with all ants
whose age and lineage were not the same as their own, or else
the same as that of their quandam associates in the artificial
mixed nest.

MARINE BIOLOGICAL LABORATORY,

WOODS HOLE, MASS., September, 1903.

A CAUSE OF FEUD BETWEEN ANTS OF THE

SAME SPECIES LIVING IN DIFFERENT

COMMUNITIES.

ADELE M. FIELDE.

If the blood of several ants of the same species be shed upon a
morsel of sponge, the characteristic odor of the species is dis-
cernible upon the sponge, even by human nostrils. The odor
may be pungent, acid, acrid, or musty, or may be like that of an
animal or vegetable oil. Of the thirty-five hundred known
species of ants, probably each has its distinctive odor.

Every ant recognizes its acquaintances through their odor
and its own sense of smell. It is violently hostile to all ants
bearing an unfamiliar scent, and is caressingly friendly with ants
whose odor it has always known.

That ants of unlike species should be inimical one to another
is less strange than the fact that those of the same species and
variety, inhabiting the same localities, but living in different com-
munities, should be as intensely antipathetic as are those of dif-
ferent species. With a view to ascertaining the cause of the an-
imosity between such communities, I made in 1902, many ex-
periments l with Stenauunafiilviun, with results showing that the
odor of the ants changes with their age, and that ants will not
live amicably with those much older than any that inhabited the
nest in which they were hatched.

If an ant be hatched in isolation, and the isolation be main-
tained until the ant has attained its adult strength and color, the
odor of its own body is this ant's sole criterion of proper ant-
odor, and it will affiliate with no ants other than those of the
same lineage and of nearly the same age as itself. It will affiliate
instantly with the queen-mother from whose egg it came and
whose odor it inherits, and will identify and caress that mother
though she be presented among five queens never before en-

1 A. M. Fielde, " Notes on an Ant," Proceedings of the Academy of Sciences of
Philadelphia, December, 1902.

326

CAUSE OF FEUD BETWEEN ANTS OF THE SAME SPECIES. 327

countered. It will also affiliate with any of her progeny of the
same age as itself, or with the progeny of her own sister of the

same age.

A difference of forty days in the ages of two ants produces a
difference of odor appreciable by the ants. If many pupae be
taken from one colony, and the workers hatched therefrom on
the same day be segregated ; and then, later on, more pupae be
taken from the same colony and the workers hatched therefrom
on the same day be likewise segregated and established in a nest
with inert young, the younger group of ants will not permit the
members of the older group to approach the young in their nest,
provided always that there be forty days or more of difference in
the age of the two groups. The degree of animosity exhibited
is in direct ratio to the difference in the age.

An ant hatched in the first brood of a solitary queen associates
during its earliest days only with its queen and with its sister-
ants, all hatched in one summer. These workers know only
ants that are less than a year old, and will never become ac-
quainted in a friendly converse with ants older than themselves.
As seasons pass, and more ants are annually hatched from the
eggs of this queen or the queens among her offspring, the latest
comers know the odors of those of their own year, and of each
year gone by, up to that of the oldest in the common nest. One
might say that the sense of smell in the ant is more highly cul-
tivated if she live in an old community.

I have been personally acquainted for four years with the ants
in a community, the C colony, whose domain is a hundred yards
in its diameter. On August 22, 1901, I took queens, males and
workers from the wild nest of this colony, and segregated a
similar group in each of two Fielde nests, where I kept them
two years. The queens were winged when captured, and were
doubtless less than a month old. The workers were fully colored,
and may have been a year or more older than the queens. No
young was permitted to hatch in either nest, and there was no
communication between the two nests nor with outside ants.
On August 25, 1903, I united the two groups, then numbering
four queens and twenty-five workers in one nest, and two queens
and nineteen workers in the other nest. They all affiliated in-

328 ADELE M. FIELDE.

stantly with no sign of cognizance of their long separation. They
had added years simultaneously and there was no difference of
odor to occasion distrust among them.

I then introduced into the nest of the two united groups several
very young ants taken that day from the wild nest. These cal-
lows were kindly received because the old ants all recognized an
ant-odor with which they had formerly been acquainted, and this
recognition was instant notwithstanding the fact that they had
met no callows during two years. It is probable that an ant
remembers during its lifetime any odor with which it has once
been acquainted.

I then brought queens and workers from the same wild nest,
housed them with their inert young in one of my artificial nests
and left them to establish their nest-odor. A few days later I
introduced into their nest marked queens and workers from the
groups segregated two years previously. The marked queens
were instantly accepted by the queens and workers in the latest
nest. The marked workers were amicably received by all the
queens, and by most of the workers in the latest nest, while a
few nabbed them or dragged them away from the pupae-pile.
They were not killed but were denied by these few, the crown-
ing mark of ant-esteem, permission to share in the care of the
young. It thus appeared that ants as old as were these seques-
tered workers were not common in the summer of 1903 in the
wild nest of the C colony, while queens two years old were known
to all the ants taken from the wild nest.

Difference of food, drink and environment during two years
had not caused a difference of ant-odor between the segregated
ants and their ancient comrades.

T/ie progeny of queens of unlike age but of the same community
are unlike in odor.

Four queens of the C colony, captured by me before their
swarming and while they were still winged, on August 22, 1901,
were segregated with kings of their own colony in one of my
nests which I here refer to as Section A. Two queens of the
same colony hatched on August 5, 1902, from pupae taken from
the wild nest two days earlier. They mated with kings of their
own colony on August 22, 1902, and were later on segregated

CAUSE OF FEUD BETWEEN ANTS OF THE SAME SPECIES. 329

with workers hatched in my artificial nests between August 8
and 28, 1902, from C colony pupae. This nest I here refer to
as Section B.

The ants in the two sections were fed with the same kinds of
food on the same days and had in all respects similar envi-
ronment.

On July 12, 1903, an ant-worker hatched from a pupa that
had been previously removed from Section B, and isolated in a
Petri cell. This worker was kept in isolation until she was six
days old. I then introduced into her cell a worker, the off-
spring of a queen in section A, and she attacked this worker
with great violence, although the worker was of an age precisely
her own and had likewise been isolated from the pupa-stage.
The only difference between the two lay in the age of their re-
spective mothers, one queen mother being two years old and the
other one year old. Neither of these callows had, previous to
their meeting, ever smelled any other ant, and had they had the
same odor they would have affiliated, as do similarly reared ants
that are the progeny of the same queen or of sister queens.

On August 24, 1903, when the ant from Section A, used in
the foregoing experiment, was forty-three days old and was
occupied in the care of introduced larvae, I put into her Petri-
cell, where she had always lived alone, a callow five days old,
reared in isolation from a pupa taken from Section B. The resi-
dent ant at once attacked and dragged the callow. In this case
the offspring of the older queen attacked the offspring of the
younger queen, though that offspring was much younger than
herself.

Other experiments coincided in their results with the two here
recorded.

A cause of feud between ants of the same species living in
different communities is a difference of odor arising out of differ-
ence of age in the queens whose progeny constitutes the commu-
nities, and difference of age in the ants composing the com-
munity.

MARINE BIOLOGICAL LABORATORY, WOODS HOLE, MASS.,
September, 1903.

DIMORPHISM IN BLISSUS LEUCOPTERUS.

J. F. CAREER.

Two forms of the chinch bug are recognized by entomologists
— the one having wings fully developed, the other having wings
more or less abortive. Between the two extremes of fully winged
and almost wingless all gradations exist. Where the short-
winged form occurs it is usually intermixed with long-winged
individuals. Such a mixture appears at certain times in abun-
dance in the timothy meadows of northeastern Ohio. It was
from Trumbull, Portage, Mahoning and Stark counties of this
state that Professor F. M. Webster furnished the principal portion
of the material for the present study.

The study was undertaken with the direction of Professor C.
B. Davenport to determine by quantitative methods the biological
significance of the dimorphism.

METHOD.

The insects examined represented several random collections
from different points. For study they were taken from the vari-
ous bottles with no attempt at selection so those studied are
presumed to present fairly the conditions in the whole group.

Where practicable, the wings were carefully removed from the
body and mounted in a series on glass slides. By means of a
dissecting microscope of low power and a camera lucida the
image of the wing was projected upon a magnified scale and the
length thus read to tenths of a millimeter. With museum
material it was necessary to measure the wings in situ and this
was accomplished by the use of a metal scale divided to fifths of
a millimeter placed against the wing under a lens.

THE FREQUENCY POLYGONS.

The size of a class was fixed at one fifth of a millimeter and
this gave a range of ten classes. The polygon is bimodal, one
mode being at 1.5 mm. and the other at 2.7 mm. The extremes
of the range include from i mm. to 2.99 mm.

DIMORPHISM IN BLISSUS LEUCOPTERUS. 33!

For convenience in calculation the polygon was considered as
two, the first having six and the other five classes, the small
connecting class being divided between the two polygons. Both
polygons are skew, running down very rapidly on their outer
slopes and shading off gradually toward each other to be con-
nected by a very small class. The skewness of the polygon with
the mode at 1.5 mm. is -f .0235 and that of the one with the
mode at 2.7 mm. is — .018.

An examination of short-winged specimens from California and
from Long Island kindly loaned from the National Museum by Dr.
L. O. Howard and others from New York State loaned by Dr.
C. E. Felt gave polygons with the same mode as that obtained
from short-winged material from Ohio. Similar results were ob-
tained by a comparative study of long-winged insects sent from
Urbana, 111., by Professor S. A. Forbes. This indicates that the
tendency of a given form is toward the same mode from whatever
region taken or whether the two forms are mixed or separate.

DISCUSSION OF RESULTS.

The significance of these results is by no means easy to deter-
mine. Looking at the polygons only it seems reasonable to
suppose that the present dimorphic species has been derived from
a parent stock with a mode lying somewhere between the two
present ones. In that case it may be assumed that differences of
environment have permanently impressed themselves, dividing the
parent stock into two evolutionary lines one of which at present
has wings longer and the other wings shorter than the parent stock.

The evidences of geographic distribution appear to negative
this view. The genus is almost cosmopolitan, having been re-
ported from every continent save Asia and from many islands of
the sea. So far as known, it is most abundant and certainly
most destructive in the United States. Nevertheless there are
good reasons for regarding the chinch bug not as a native but as
an immigrant. In his very reasonable hypothesis about the origin
and distribution of the chinch bug in North America, Professor
Webster (1898) l assumes that our stock of chinch bugs has come

i Webster, F. M., "The Chinch Bug," U. S. Dept. of Agriculture, Bulletin No.
15, New Series.

332

J. F. CAREER.

from South America by way of the Isthmus of Panama, Central
America and Mexico. The north-flowing stream was divided
first by the Cordilleran system, one branch of the division follow-
ing the Pacific Coast northward, the other, by far the more im-

I2O

no

IOO

portant one, spreading over the Gulf States, was split again by
the Appalachian Mountain System. One of these latter branches
overflowed the Mississippi Valley ; the other, following the coast
of the Atlantic, finally rounded the north end of the mountain
system and finding a congenial highway across New York State

DIMORPHISM IN BLISSUS LEUCOPTERUS.

333

joined the Mississippi Valley branch in northern Ohio and around
the Great Lakes.

The short-winged form, so far as known in America, is con-
fined to the ocean coasts and the immediate vicinity of the Great
Lakes. The vast interior region from Central America to Mani-
toba abounds with only the long-winged form.

If Webster's theory is the correct one, we can scarcely escape
the conclusion that the short-winged form originated in the re-
gions where it is at present found. No short-winged specimens
have ever been reported from the Gulf States outside of Florida,
from Mexico or Central America, nor west of the Alleghanies,

notwithstanding, the insect is common in those regions and the
short-winged form has been carefully looked for in some of them.
The long-winged insects, then, appear to have been the ancestral
form in America as far as history and hypothesis can give a clue.
There seems to be an inherent tendency in the species to produce
the short-winged form when the proper ecological conditions are
provided. How the species acquired this tendency is a very
difficult thing to understand and it is not the purpose of this
paper to attempt an explanation of a phenomenon that appears
to be older than the division of Heteroptera into the present
recognized families.

According to Saunders l dimorphism is exceedingly common
among British Heteroptera and this caused much confusion

1 Saunders, Edward, F.L.S., "The Hemiptera — Heteroptera of the British Is
lands," 1892.

334 J- F- GARBER.

because long- and short-winged forms were placed in separate
species, certain other correlated characters, c. g., a weaker de-
veloped pronotum in the short-winged form being constant.

In the Family Lygeidae all grades of winged shortening occur
and in some species a fully-winged individual is very rare. In-
deed every important family shows wings shortening to some
extent.

Though the short-winged form occurs in America almost
exclusively near large bodies of water such proximity is not
necessarily a factor in producing and preserving this peculiar
character.

A closely related species, B. dories, is comparatively abundant
in southern Europe and far northward into the interior of Hun-
gary. A long-winged specimen of this species is a rarity and
was not supposed to exist until 1880 when a very small colony
was discovered by Professor Sajo. From his paper, presented in
full in Professor Webster's bulletin previously cited, we get the
facts concerning this species.

The colonies of B. doric? live on the bases of bushy grass near
or even under the surface of the ground, and here the stages of
development are passed through. The species is very widely
distributed on sand drifts and in hilly regions, but long-winged
specimens were found in but a single tiny spot. The bunches of
grass on which the insect lives are isolated in partially bare
ground. During the period of development, great drought
prevails. The long-winged specimens possess a stronger and
broader thorax than the short-winged ones, and it never attacks
cultivated crops.

According to numerous observers cited by Professor Webster,
the habits of B. lencopterus along our coasts are almost identical
with those described for B. dories. Professor C. W. Woodworth
writes me that the chinch bug is found in California chiefly in the
salt marshes.

SUMMARY.

Where short-winged chinch bugs occur in Europe and America
their habitat almost without exception compels them to live
about the roots of tufts of grass on a soil otherwise almost bare.
In California they are found in salt marshes. In Europe it may

DIMORPHISM IN BL1SSUS LEUCOPTERUS. 335

be added that the developmental stages occur at a season of
great drought. Taken all together we have a picture par ex-
cellence of a xerophilous insect which is only another way of
designating a species capable of withstanding hard or unfavorable
conditions of living. Among the hard conditions which are
responsible for dwarfed wings as well as more or less dwarfed
bodies of chinch bugs, I should place first drought and poor food
supply. Latitude and climate do not influence them, but edaphic
conditions that may extend over large areas are the potent factors.

The only recorded observation that seems to oppose this view
is that of Mr. E. P. Van Duzee.1 He states that in portions of
Ontario and New York where the short-winged form usually pre-
dominates, in dry, hot summers they mostly acquire fully de-
veloped wings. It seems possible, however, that a dry hot
summer added to an ordinarily unfavorable habitat may have
destroyed the short-winged form to an extent, only those in the
most favored places being allowed to develop.

That the short-winged form should extend at times beyond
the borders of the particular habitat which served to develop the
dimorphic tendency (as occurs for example in northern Ohio)
may be regarded only as the persistence for a time of a charac-
ter acquired by the race even when the insect is in different sur-
roundings. The mixed forms, however, always cling to old
food habits as far as possible, taking by preference to grass
meadows instead of attacking grain fields as do the long-winged
insects of the interior.

1 Van Duzee, E. P., Canadian Entomologist, Vol. XVII., pp. 209-210, 1886.

UNIVERSITY OF CHICAGO,
June, 1903.

ON TWO CASES OF MUSCULAR ABNORMALITY

IN THE CAT.1

RAYMOND PEARL.

The muscular anomalies here described were found by the
writer in specimens of the domestic cat used for dissection in
class work in the University of Michigan. As both of the cases
presented certain interesting features it seemed advisable to pub-
lish an account of them at this time.

I. A CASE OF ABNORMAL INSERTION OF THE M.
LATISSIMUS DORSI.

In the cat the tendon of insertion of the M. latissimus dorsi
normally is in two parts. One of these parts- is joined by the
muscle and tendon fibers of the M. teres major, and the conjoined
tendon of these two muscles is inserted on the medial side of the
shaft of the humerus. The other portion of the latissimus ten-
don, which may not be always present according to Reighard and
Jennings,2 joins with the pectoralis minor, reaching the bone
along the line of insertion of the pectoralis minor. This line is
along almost exactly the middle of the ventral face of the hum-
erus. As a consequence of the existence of their different lines
of insertion the two portions of the latissimus tendon form an
arch, which makes up a part of the bicipital arch.

In a well-formed, adult male cat dissected by the writer the
very peculiar arrangement at the insertion end of the M. latissimus
dorsi shown in Fig. I was found on both sides of the body.
From the cranial border of the latissimus a slip (Fig. I, .r), about
4 cms. long and 6 mm. wide passed craniad above that portion
of the latissimus which joins the pectoralis minor (Fig. I, y).
This slip was inserted by fleshy fibers on the surface of the M.
pectoantibrachialis on the medial surface of the leg, just beneath

1 Contributions from the Zoological Laboratory of the University of Michigan,
No. 65.

2 Reighard, L, and H. S. Jennings, "Anatomy of the Cat." New York, 1901
p. 121.

336

MUSCULAR ABNORMALITY IN THE CAT.

337

the skin. This band of muscle formed a very distinct, rather
thick slip.1 The relations of all the other muscles of the leg
were normal. The two tendons of insertion normal to the latis-
simus dorsi were present and in their usual relations. The ab-
normal slip was simply added on, as it were, to the muscles
normally present.

clv.br.

pct.mn.

XphK.

FIG. I. Ventral view of left side of the thoracic region in cat, showing abnormal
insertion of the M. latissimus dorsi. dv. br., M. clavobrachialis ; pabr, , M. pectoan-
tibrachialis; epit., M. epitrochlearis ; //. ds., M. latissimus dorsi; xph/i., M. xiphi-
humeralis ; pet. ;«/., M. pectoralis major; pet. ma., M. pectoralis minor; x, ab-
normal slip of M. latissimus dorsi ; y, portion of the latissimus dorsi which joins the
pectoralis minor.

The conditions found in this case of the latissimus dorsi insert-
ing in three portions, one of which does not reach the humerus
at all, is apparently unique. So far as I have been able to dis-

1 In another cat dissected by a student in the laboratory precisely the same arrange-
ment was found, except that the mucle slip was much thinner than in the case here
described. Only a few fibers reached the pectoantibrachialis.

338 RAYMOND PEARL.

cover no record of such a condition has been made in teratologi-
cal literature, nor is such a condition found normally in any
form. In most mammals1 the latissimus inserts by one tendon ;
in some forms (e. g., the cat) usually by two; and finally as a
variation, which apparently occurs with some frequency, it inserts
by two tendons in forms where it normally has only one. This
last is the condition in man.2

The condition found in this abnormality to a certain degree
resembles morphologically what is normally found in many
mammals in the M. epitrochlearis. This muscle, in the majority
of cases, takes origin from the surface of the latissimus dorsi
near its insertion, and is inserted into the superficial fascia of the
forearm and the olecranon. This muscle is usually regarded as
a differentiation product of the latissimus dorsi. It is possible
that the present abnormality may indicate that originally the M.
epitrochlearis had in the carnivora a greater extent at its inser-
tion, extending on to the superficial fascia of the upper as well
as the forearm. Further than this I am not able to make any
suggestion regarding the significance of this abnormality. On
account of the fact that apparently such a case had not been de-
scribed, it seemed desirable to make a record of it.

II. A CASE OF CONNECTION BETWEEN THE M. CLEIDOMASTOI-

DEUS AND THE M. LEVATOR SCAPULAE VENTRALIS.

The M. cleidomastoideus normally forms a distinct muscle in
the cat, taking its origin from the apex and caudal margin of the
mastoid process of the temporal bone. It passes caudad, flatten-
ing during its course, and is inserted on the lateral four fifths of
the clavicle and laterad of the clavicle on the clavicular raphe.
This clavicular raphe is formed between the Mm. cleidomastoideus
and clavotrapezius (= M. cleido-occipitalis + cleido-cervicalis
Streissler) 3 craniad, and the M. clavobrachialis (= Pars claviculi

JCf. Leche, W. , Mammalia, in Bronn's " Klassen u. Ordnungen des Thier-
Reichs," Bd. 6, V. Abth., 1874-1900, pp. 722-725.

2Cf. Le Double, A. F., " Traite des Variations du Systeme Musculaire de
1'Homme," Paris, 1897, T. I., pp. 194-202.

Testut, L.,"Les Anomalies Musculaireschezl'Homme," Paris, 1884, pp. 106-118.

3Streissler, E., " Zur vergleichenden Anatomic des M. cucullaris und M. sterno-
cleidomastoideus," Arch. f. Anat. (u. Physio/.) Jahrg., 1900, pp. 335-365, Taf.
XXI. u. XXII.

MUSCULAR ABNORMALITY IN THE CAT. 339

of M. deltoideus of earlier writers) caudoventrad. At its inser-
tion the cleidomastoid lies entirely beneath the clavotrapezius.
Lying close besides the cleidomastoid (dorsad and in part
mediad) is the M. levator scapulae ventralis ( = M. omo-trans-
versarius Streissler, loc. cit., = Pars ventralis of the M. otno-
cleidotransversarius Leche, loc. cit., = " omo-trachelien " Le
Double, loc. cit.\ This muscle in the cat takes origin by two
heads, one coming from the basis cranii opposite the middle of
the bulla tympani, and the other from the ventral surface of the
transverse process of the atlas.

In a well-developed adult female cat dissected by the writer,
the following abnormal relation of the cleidomastoid and the
levator scapulas ventralis was found on the left side of the body.
At almost precisely the middle point of the levator scapulae ven-
tralis a thick muscle band, approximately 4 mm. wide, passed
from the ventral border of this muscle cranioventrad to the dorsal
border of the M. cleidomastoideus, with which muscle it joined.
The connecting band was throughout its length of approximately
the same thickness as the Mm. cleidomastoideus and levator
scapulae ventralis at the places where it joined them.

In considering the significance of this abnormality the possi-
bility of its representing a case of reversion may be dismissed at
once, because in their comparative anatomy the cleidomastoid
and levator scapulae ventralis are known to be quite distinct
muscles. The M. cleidomastoideus is a differentiation from the
general sternocleidomastoid group of muscles, which in turn is
to be considered as having separated from the trapezius group.1
It belongs to the rather thin, superficial sheet of muscle which
covers the dorsal, lateral and part of the ventral surface of the
neck, and the dorsal surface of the cranial thoracic region in all
the Mammalia. This sheet of muscle breaks up into varying
numbers of separate muscles in different groups. All of these
muscles, however, as has been very clearly brought out by
Streissler (loc. cit.), fall into either a dorsal or a ventral group.
The dorsal group may be characterized as the dorso-scapularis-
trapezius group, and the ventral as the sternocleidomastoid group.
All the muscles of this superficial layer are innervated primarily

1 Cf. Leche, loc. cit., pp. 701-706.

34O RAYMOND PEARL.

by the N. accessorius, with, in some cases, fibers from the cer-
vical plexus going to the muscles of the ventral group. The
levator scapulae ventralis or omo-cleido-transversarius, pars ven-
tralis (Leche) belongs to an entirely different set of muscles than
those just considered. According to Leche1 it is highly probable
that this muscle is a differentiation product of the muscle group
from which the M. levator scapulas comes. It is innervated by
fibers from the ventral branches of the spinal nerves.

Evidently then, since the cleidomastoid and the levator scapulae
ventralis have such different sources the abnormality under dis-
cussion cannot be considered as a reversion.

The abnormality does, however, seem to be suggestive as pos-
sibly giving us light on the meaning of the conditions found in
man with reference to the muscles of the ventral neck region.
In what manner will be apparent if the relations in man are con-
sidered briefly. The M. omotransversarius (i. e., levator scapulae
ventralis) is normally found in some form or other in practically
all mammals up to man. In man it is only occasionally present
as a separate muscle in abnormal cases. It has been a problem
how to account for the absence of this muscle under normal
conditions in man, and no satisfactory explanation for it has ever
appeared so far as is known to the writer. On the other hand
the human sternocleidomastoid is, of course, a complex muscle,
made up by the fusion of elements normally forming distinct and
separate muscles in the lower forms. Streissler2 has shown that
this muscle contains at least the following elements : In the super-
ficial portion a sternomastoideus superficialis, a sterno-occipitalis
and a cleido-occipitalis element ; and in the deep layer a sterno-
mastoideus profundus and a cleidomastoideus element.

The fact that occasionally the omotransversarius appears in
man as a distinct muscle may be taken as strong presumptive
evidence that in all cases the muscle is present in man as an
element in the ventral neck musculature. Why it is not found
under normal circumstances is because it is indistinguishably
fused with some other muscle. In the abnormal cases where it
does appear as a separate muscle we most probably have simply

* Loc. «/., pp. 731-735-

* Loc. cit.

MUSCULAR ABNORMALITY IN THE CAT. 341

a failure to fuse or only partial fusion, where normally complete
fusion occurs.

The abnormality here under consideration has suggested to me

the view that normally in man the omotransversarius element is

fused completely with the cleidomastoid portion of the M. sternoclei-

domastoideits. This view would make the sternocleidomastoid a

complex of six elements, as shown in the following scheme :

f Sternomastoideus superficialis ^
Sterno-occipitalis L Superficial.

Cleido-occipitalis
M. sternocleidomastoideus (Man) -j

Sternomastoideus profundus -»
Cleidomastoideus > Deep.1

^ Omotransversarius

The evidence for this view comes from two sources. In the
first place, the occurrence in anomalous cases in man of a sepa-
rate M. omotransversarius makes it extremely probable that this
element is generally present in man, but in normal cadavers is
completely fused with some other muscle. In the second place,
the anomalous case in the cat just described shows that in a form
lower than man it is possible for a partial fusion of the cleido-
mastoid and omotransversarius muscles to occur as a variation.
This makes it seem probable that the muscle complex with which
this omotransversarius element in man normally fuses is the ster-
nocleidomastoid.

SUMMARY.

1. A case of insertion of a portion of the M. latissimus dorsi on
the M. pectoantibrachialis is described.

2. A case of partial union of the Mm. cleidomastoideus and
levator scapulae ventralis (or omotransversarius) is described.

3. The view is advanced that the human sternocleidomastoid
muscle contains an omotransversarius element. This element is
normally completely fused with the deep portion of the sterno-
cleidomastoid, but, in abnormal cases, it may fail to fuse com-
pletely and consequently then appears as a separate muscle.

MBL WHOI LIBRARY

lilH 17JG