5?«n^w^A^^xio,

l^'^^^-^..e^'^'

'^^.^X:^'

1 01. as

Txhv

lit

it

'?«??e5J«s-->2S«'^

SM«W««^^

ISftys^'^^^^^^l*^

>*v;'''*\>«,.--S

"m^im

■cZS::^rM

sK^^rmmW^

.^AAf^f^

M^

'CCw^i

^m-.^rA,

^!^*:/^A^Ar' ./^^^

«fW*SK

i^f^imm<f\;^Mr'^^^nf

^^ii^:

^ftft*«W^¥W

^^^.A#»<^

w.ii.A^,a^

:;.^r^rJMteeee^^

^^^n^a^^/^W'^'^'^'^

r^^^r^AM

:..j^f&m

^--y^^^^N^

.N^mm^m^^'

^AftC'«oo«^«.^

!w?<^^lA/^^^r^n^^^^'

^^^^^^^^^

fP'M.jjm

mm

TThA&,.

Ai".^^^-UC-^L iv/^

THE
DF,VELOP^',EHT OF

0 N I 0 N E M A H U R R A C H I I

--0O —

by
Henry Farnham Perkins.
DISSERTATION

submitted to the

Board of University Studies of the

JOHNS HOPKINS UNIVERSITY

in conformity v/itii the

requirements for

the decree of

DOCTOR OF PHILOSOPHY

Baltimore
1902.

loun

Contents,

page,

Introduction 1

Note on the Ontogeny of the "Trachornedusae" 4

Embryology 9

A, Dehiscence 10

B, Periodicity 11

C, Egg-envelope 15

Methods 16

D, Segmentation 18

The Planula 21

The hydra Larva 22

A. Tentacles • ■ -- 24

B. Form of Coelenteron 25

Habits 27

■Degeneration Phenomena 29

Budding in the Larvae - Metagensis 30

Transformation of the Polyp 59

Youngest L!edusae; Arrangement of Tentacles

and Sense-Organs 42

Histogenesis of Marginal Organs.

A, In the Lax'va 53

E. In the Adult 54

C. The Sense-Orp;ans 57

Mematocysts 58

Sexual Organs 62

Sutninary 64

Vita 71

Description of Figures • 66

T)E\^ELOPMENT OF TtONIONEMA HURBACHII.

Gonionernus : A.A^assiz, 1862. Contrib. Nat, Hist. U.S.,
IV. , . p. 350.

From y- CO V i 0 angled, and h> rj a^ ol thread-

"kneed tentacles".
Gonynema: Haeckel, 1S79. System der I'edusen.
Gonionemus : Murbach, lvS95, Journal Morph. , XI., 2
Gonionemus Murbachii : Mayer, 1901, Brooklyn Inst. Sci, ,

Bui. , Vol. I. , L.
Gonionema : A. A^assiz, mss.
Gonionema Murbachii . Perkins, J. H. Univ. Circ, May, 1902.

INTRODUCTION,

This genus was established by Dr. Alexander Agassiz
to include a medusa which he discovered in 1802 in the Gulf
of Georgia, V/ashington Territory, Its most striking char-
acter, anatomically, is t\\e peculiar form of the tentacles,
v/hich are bent at an angle near the tip, and at the angle
bear a sucking organ by means of v/hich the medusa makes it-
self fast to any favorable object. This peciilarity in tlie
form of the tentacle suggested to Agassiz the name which he

pi-oposed. Tho form of the word which I use is that which
Professor Agassiz offers in correction of the original one,
which was m error as to its ending.

For a long time the CiUlf of Georgia was the only lo-
cality from which this genus was described. In 1894, how-
ever, another habitat was discovered far distant, at Y/oods
Hole, '^Massachusetts, and since then members of the gemis
have been found at the widely separated localities of the
Fiji Islands and Alaska, and a closely allied genus, Olin-
dias , has been described from the coast of Brazil and the
Bahamas, This last instance was that in which Mayer de-
scribed a species (aphrodite) of Gonionema ; but as a matter
of fact this medusa possesses rather the characters of the
Oliniadae, two distinct kinds of tentacles and papilliform
gonads being present. These far distant localities will in
all probability be connected in the fiiture, during the
course of further research, by a chain of other localities.

The history of the Woods Hole Gonionema is interest-
ing. In spite of the fact that the "eel-pcnd", at the cen-
tre of the village of Woods Hole, connected with the outer
harbor by a narrow inlet, is from its location easy of ac-
cess to collectors, and although many stvidents of tlie group
had investigated the waters around Woods Hole summer after

summer for a mmnber of years , Gonionema was never found in
the Atlantic Ocean until 1394. This season a number of
specimens were taken from the "eel-pond", the creature hav-
ing made an astonishingly sudden appearance upon the scene.
It seeiris incredible that it could have been living in this
small body of water for any time previously, or at any rate
that a number of individuals had been there. But the medu-
sa at once secured a good "foot-hold" and since the first
summer has been very plentiful; its numbers remain undimin-
ished by the wholesale raids of collectors, in spite of the
keen anxiety of some of those especially interested in it.
Luring the summer of 1894 when the jelly-fish was first
found at Woods Hole Professor W.K.Brooks secured a number
of the specimens and made drawings both of live medusae and
of sections of preserved material. Some of these drawings,
figures 29, 30 and 33 are now published, with Dr.
Brooks' generous permission, for the first time.

The first printed account of the Woods Hole Gonion-
ema , since recognized as a distinct species, was published
in 1895 by Dr. L. Murbach. The species has frequently been
mistaken for G^. vertens. It was not until 1901 that the
specific name Murbachii was bestowed upon it by Dr. A. G.
Meyjr (1901). The work which I have done on the life-his-

tory of the species was orip;inally undei'takon and has since
boen prosecuted with Dr. Murbach's kind encouragement, and
I have received from him many favors in the way of material
and helpful suggestions. The work has been carried on dur-
ing the past two years at the U.S. Fish Commission laborato-
ry, whei-e I have had the great privilege of working during
the summer, and under the direction of Professor W. K,
Brooks at the Biological Laboratory of the Johns Hopkins
University, I wish to acknowledge my obligations to T)r,
Bumpus , Dr. II, K, Smith and Dr. V/hitman for courtesies
which they have extended to me in my work.

Note on the Ontogeny of the Trachomedusae.

Gonionema . according to Ilaeckel's classification, falls in-
to his third order, the "Trachomedusae". Haeckel character-
ized this order as follows: "development, hypogenesia
(not metagenesis), but usually with metamorphosis". Subse-
quent research into the life-history of members of this
group has shown that each clause of this statement is open
to emendation. In the first place, the "usually" is super-
fluous, and therefore ei-roneous. The exceptions which
Haeckel supposed to exist and which caused him to say "usu-
ally with metamorphosis" have been shown to be no excep-

tions, but cases of somewhat easily misunderstood metamor-
phosis. Such, for exami'le, v/as the case of Liriope, which
has been studied by Metschnikoff (1874) and Brooks (1886).
T?ie larva is a true hydra, although its free swimming mode
of life and its siiperficial aspect caused it to be mistaken,
formerly, for a gonosome.

My investigations of a jelly-fish which Haeckel in-
cludes in his order "Trachomedusae" show that the first part
of Haeckel's statement also requii-es revision, and that
metagenesis does occur in the "Trachomedusae". Although
there may bo some differ*ence uf opinion as to the exact
definition of the terms "metagenesis" and "hypogenesis" ,the
following notion of the process of alternation of genera-
tions !nay be safely acr-epted as tliat generally held. The
prodTzction by a lai-va of offspring unlike itself, and its
own ultimate death without undergoing metamorphosis, are
frequent accompaniments of the intermediate as of the pri-
mary process of multiplication; but they are by no means
essential to the process of metagenesis or alternation of
generations. Creatures which multiply sexually at one
point of their life-history, and at another point nonsex-
ually by budding or fission, are said to have a metagenetic
development. In Gonionema (Figs. 3 - 23) a large number
of adult individuals are produced from a single egg through

an intermediate process of rmiltiplication; buds which are
developed upon the body of the hydra larva drop off and,
beginning as planulae, just follow exactly the same co\irse
of development as the parent; both parent and offspring la-
ter change into fully developed medusae. "Gonionema"has ,
then, a metagenetic form of development. The mere presence
of a hydrula stage in the life-history is of course not
enough, accor-ding to generally accepted notions of the mean
ing of terms, to constitute alternation of generations
(Murbach, lcS95 , p. 496).

These emendations of Haeckel's description of the
order add evidence to that already put forward (Brooks,
1895, p. 300) and by others that the hard and fast lines
drawn by Haeckel and the Hertwigs between the "Trachylinae"
and the "Leptolinae" , on the r^round of anatomical differ-
ences or of developmental features, are not borne out by
tho facts. The Ifertwigs (1S7S) hold that "the m-arginal
sense-organs (Gehflrorgane) alone furnish characteristics
which enable us in every case to distinguish the 'Trachome-
dusae' (Trachomedusae and Narcomedusae of Kaeckel) from the
i^ampanularian medusae (Vesiculatae) without knowledge of
their development". Dr. Brooks has, however, described a
species of Laodice which unites in its anatomical features

tVie characters of both the Loptolinae and Ti-achylinas , hav-
ing T,he ocelli of the former order and the chitinous p:orran-
gium containinp; the developing medusa buds, (A.":assiz, 1865,
p. 12j) wiiile Professor Bi-ooks has demonstrated (lo9r;) that
it also possesses the true endodermal sense-clubs of the
Trachylinae.

It may bo that the present record of obsei'vations on
Gonionema will be of intei-est as contributing some new
points to the present meagre knowledge of the manifold
forms and types which are exhibited in the development pro-
cesses of this great gi-oup,

I. GONOSOME.

Gonionema is a very attractive feature of the
Woods Hole fauna. Its exquisite glassy umbrella, marked
with a cross of yellow or brown by the four radial canals
and the gonads, a brilliant row of closely set spots of
gleaming phosphorescent green outlining its edge, the
fringe of delicate streaming tentacles strung with bead-
like clusters of thread cells, are all more or less famil-
iar to many American biologists (Fig, 1), On cloudy days
or toward nightfall it is very active, swimi^iing upv/ard to
the top of the water and then floating back to the bottom.

In swimming it propels itself upward with I'ytlimic pulsa-
tions of the bell-mar{;;in, t?ie tentacles shor-tened and the
bell very convex. ("Pig. 2) Upon reaching the surface the
creature keels over almost instan+ly, and floats slowly
downwards with bell relaxed and inverted, and the tentacles
extended far out horizontally, forming a wide expanse of
stinging threads, which carry certain destruction to animals
even larger than the jelly-fish itself, if they are unwary
enough to be caught within their reach. Gonionema contin-
ues this peculiar fishing up and down in the water with
little respite. Occasionally it fas+ens itself to the eel
grass or other object near the bottom, or stops midv/ay in
its course with tentacles extended as in my figure 1, well-
nigh invisible, but a deadly foe to small fish or crusta-
ceans, which cross its path.

II. GONADS.

In the mature Gonionema the sexual organs are
"frill-like lobes, passing from one side to the other of
the chimiferous canal". (Agassiz, 186b). Their form and
position is shown in figures 3 and 4. The free edge of the
ribbon of tissue is thickened and rounded, and is bent back-
wards and forwards across the radial canal. The color of

the i^onads has been supposed to afford means of descrimi-
nating between the sexes, the males differing from the fe-
males in the brigliter yellow of the gonadial tissue. But
this distinction aces not hold, and it is necessary to ex-
amine the individual medusae with a lens in order to sepa-
rate the sexes. The ovarian eq,g5, enclosed in the ectoderm
of the gonads of the female, give them a granular appear-
ance as contrasted with the more homogeneous and translu-
cent tissue of the male. When a large number of the jelly-
fish are separated into t-/o vessels, one containing the
males and the other the females, +he general color tone is,
it is true, brighter and more lively than that of the fe-
males. But the specimens in each dish range all the way
from light straw color up through orange, ochre, sienna, to
dark brown.

III. EMBRYOLOGY.

It is my purpose to give in outline the main
points in the early part of the life-history of Gonionema.
I have not discovered that this species exhibits any nota-
ble peculiarities. I shall lay greater emphasis upon cer-
tain features of the later developmental stages, which have
more significance in so far as they are less familiar.

10

A. DEHISCENCE -

The e.-^gs are embedded in the ectodenmal tissue
of the gonad as in a gelatinous matrix. The round thicken-
ed edge of the ribbon contains the riper eggs, but the thin-
ner portion is v/ell packed with maturing ova. Dehiscence
takes place by the breaking down of the superficial ecto-
dermal investment and the liberation of the eggs and sper-
matozoa imbedded in its substance. The contractions of the
umbrella in swimming piit a strain upon the subumbral walls
and help to rupture the epithelium of the gonads, V/hat the
cytological change is which precedes the extrusion of the
sexual elements or the nature of the causes which effect
this change are matters of uncertainty. We know only that
these conditions can be artificially induced by means to be
mentioned presently.

The process of dehiscence occupies only a very few
moments. Host of the eggs which are ready for fertiliza-
tion are extruded all at once, coming out of the bell-cav-
ity in a cloud at each contraction of the marginal ring.
Two or three minutes after the first eggs are extruded only
a few belated ones are loosened from the gonads and expell-
ed from the subumbral cavity two or three at a time. Fig-
ure 3, is drawn from a sketch made of a medusa in the act

11

of spayming. The specimen was held inverted in a v/atch-
glass under the microscope, Althoixgh not Tree to swim, it
went throush the motions, contracting the bell rythmically.
In this way the softened ectodermal tissue of the gonad was
ruptured, and the e^gs expelled. Little round pits are
left by the ep;gs , like bullet-molds. The earliest date at
which fertile medusae have been found was the first of Ju-
ly; the latest, the last week of September, The period of
maximum sexual activity is from the middle of July to the
middle of August,

B. Periodicity - As stated by Murbach (1895) the
eggs are normally extruded at about eight o'clock P.M. This
is the case during the earlier part of the summer, but lat-
er in the season, when the days are shorter, half-past sev-
en, or even shortly after six is the normal time for spawn-
ing. Further than this, extrusion of the eggs may be arti-
ficially induced in Gonionema, In this respect this form
differs markedly from some other marine animals which ex-
hibit equally definite times of spawning. Dr. Ilurbach
found that during the daytime after being shut up for an
hour in a dark place the ripe medusae would deposit eggs
and sperm. My experiments show that this is true in the

12

afternoon to a much greater extent than early in the day;
before two o'clock in the afternoon the hovir in the dark
would sometimes bring about the deposition of a very small
number of eggs, and if the period of darkening was length-
ened to an hour and a half, a slightly larger number of
eggs were foiind in the water. But after two o'clock an
hour's shutting away from the light brought about an appa-
rently normal spawning. I fovmd that the stimulus of the
withdrawal of light is surprisingly definite in its ef-
fects; the condition of the tissues arrives at the point
requisite for the release of the eggs almost on the minute.
This constancy is not appreciably affected by ra<Jderate
changes in temperature. A large munber of experiments and
observations have been made to educe the exact time of
stimulation (if we may so speak of an influence which seems
to be purely negative) and the results are summarized in
the following table. Record was kept of experiments car-
i-ied on during the whole of the fertile season, partly in
one summer, partly in the next. The stimulation-time va-
I'ies somewliat with the season; the table gives the results
obtained during the last week in July, when the eggs were
being discharged in the greatest numbers.

13

Before 2 P.M. small no, of eggs laid after 90 minutes darken

ing.

2-3 P.M. almost the nox'mal n-omber laid 90 minutes darkening

3 - 4 " " " " 65 " "

4 _ 5 " " •' " 60 " "

5 - 6 " " " " 60 " "

6 - 7 " " " " 50 " "

At 8 P.M. eggs laid normally, without darkening ar-
tificially.

As the hour approached the normal time for the de-
position of the eggs the precision with which they wore dis-
charged became more and more marked. Between four and seven
P.M. the time of darkening necessary to produce spawning va-
ries not more than four minutes on either side of the hour.

Some experiments were tried with a view to inhibit-
ing normal deposition of eggs, or at least of hindering it,
by keeping the eggs under strong artificial illiuninat ion.
The results were not conclusive, as the electric lights in
the laboratory were not in use until after dusk, when part
of the stimulus had already been received. The experiments
showed a certain amount of retarding of the process of spawn-
ing as a result of strong illumination. It woixld be inter-

14

esting to determine whether the use of stronger light begun
as soon as daylight commences to wane would result in com-
plete inhibition of the process.

It is evident from the above statements that Gonio-
nema is exceedingly sensitive to external conditions. Not
all coelenterates are affected in the same degree, some are
apparently not affected at all, by changes in illumination.
Some medusae lay their eggs always in the early morning,
while others of nearly related genera choose the evening or
night (Brooks - Life History of the Hydromedusae , p. 399).
Experiments carried on by Wilson and Donaldson at Beaufort
under Professor Brooks' direction (loc.cit.) showed that in
the case of Renilla and some sea-anemones, at any rate,
changes in light and temperature did not affect the precis-
ion with v^hich the regular physiological activities took
place. It is well known that a great many marine animals
show more or less definiteness in the habit of spawning.
Metschnikoff gives a table on page 25 of his Embryological
Studies showing the time of spawning of a large number of
medusae. In other groups the same tendency is manifest.
This phenomenon is probably the result of the working of
natural selection, the habit of laying the eggs at a cer-
tain definite time having proved of value to the different

15

species. Th9 fact that in some forms this precision of pe-
riodicity is not dependent upon external influences, while
in others there is manifest a marked degree of sensitive-
ness to external influences, seems to me to indicate that
the tendency has been arrived at by quite different proces-
ses, and may be due to quite different requirements in the
various creatureSt

But to return to the dehiscence in Gonionema , not
all the eggs, by any moans, which the ovaries contain are
liberated at one time. Individuals have been seen to de-
posit eggs every night for a week, and while specimens kept
in captivity are not very reliable in drawing inferences as
to natural processes, this time of sexual activity would be
more liable to be shortened than to be prolonged by the un-
natural conditions. After the first three or four days on
which spawning took place, a small number of ova were left
in the gonads, and on the three succeeding evenings these
were extruded a few at a time. Late in the summer the spe-
cimens taken are usually devoid of sexual products, and the
gonads small and shrivelled.

C. Egg-envelope- In freshly laid eggs the polar bodies
are only rarely to be found. They are normally given off

16

and lost in the gonads previous to dehiscence. Before fer-
tilization the eggs float in a cloud throiigh the v/ater,
each one surrounded by a very soft thick gelatinous envel-
ope. If the eggs are not fertilized, the surro'anding mass
of semi-fluid jelly slowly shrinks up, and the increased
specific gravity causes the eggs to sink to the bottom.
Blister-like vacuoles appear in the substance of the egg,
puffing out the envelope, and in the course of several days
the protoplasm becomes disintegrated and the egg goes to
pieces.

When fertilization takes place, the shrinking of the
egg- envelope iti more immediate, and greater in degree, so
that the egg sinks at once and sticks to the bottom by
means of the viscid substance of the surrounding envelope,

METHODS -

The adhesive property above referred to is of great
assistance in making preparations of the segmenting eggs,
as they may be allowed to settle on glass slides, which are
afterwards run up through all the reagents without danger
of washing off. For sectioning, the best method of secur-
ing the eggs was found to be by stirring about in the water
with a camels-hair brush, and preventing them from gluing

17

themselves down to the bottom of the dish. They would then
stick together in masses, and beinf; protected from too much
pressure by the gelatinous envelope they were found to seg-
ment normally. The bunches of eggs were large enough to
see with the iinaided eye, and could be easily transferred
to the killing fluid, and afterwards stained and cut. The
best reagents that were used for killing were corrosive^d«tic
three per cent, glacial acetic in saturate solution of bi-
chloride of mercury; and the full strength (40 per cent.)
solution of formalin. Oori-osive-acetic was satisfactory
for most purposes, both segmenting eggs and adult medusae
being fixed in this solution. They were immersed for from
one to ten minutes, according to the bulk of the tissue.
Pure (40 X) formalin was used very successfully for the
younger stages, giving good cytological fixation of segment"
ing eggs and of larvae. Fifteen to forty seconds is suffi-
cient to fix the tissues thoroughly. In woi^king with Gon-
ionema , I have experienced none of the difficulty that
seems to be met with in other coelenterates in getting uni-
form results with formalin material. I have used this re-
agent not only for fixation but for permanent preservation,
with the best results. For narcotizing the larvae and ad-
ult medusae, I find menthol crystals the most convenient

18

and rapid chemical to use.

It may be well to mention the method of keeping Go-
nionema alive in the laboratory, Rimning water is not de-
sirable, and it is of no benefit to either medusae or lar-
vae to change the water frequently, as I learned after much
laborious effort to keep the specimens alive in this man-
ner. Balanced aquaria furnish the best environment for
these creatures. I succeeded in keeping a large number of
larvae in healthy growing condition for six months in aqua-
rium jars in the laboratory. The quantity of v/ater was
kept constant by adding fresh water to make up for evapora-
tion. Food was furnished in the form of various protozoans
and other microscopic organisms. Oxidation was secured by
means of large quantities of diatot.is which wore reared for
the purpose. Cultures were made from the scrapings of eel-
grass, etc., and the diatoms which accumulated from them,
collecting in cliomps on the bottom of the dish, were scrap-
ed into the water with the larvae. At the end of Janizary,
the polyps which came from eggs laid the preceding August,
died vfithoiit undergoing metamorphosis. Their death was
probably due to a lack of food supply sufficient for the re-
quirements of their growing bodies.

D» Segmentation - The egg is spherical, averaging .07 mm.

19

in diameter. It consists of yellowish rather cloudy proto-
plasm, sufficiently transparent to permit one to observe
the more conspicuous changes which take place in the sub-
stance of the living egg. Segmentation is total and equal,
of the type which is designated by Metschnikoff as "durch-
schneidende Fu re hung". The cleavage-furrow appears at one
side of the egg first and cuts through its substance until
it reaches the opposite side, dividing it into two hemis-
pheres (Fig, 5), The point at which the furrow starts is
that nearest the nucleus, which lies eccentrically in the
granular substance of the egg. The first indication of the
furrow is a shallow groove which deepens rapidly and at the
same time lengthens so as to embrace the egg meridionally.
The furrow is finally completed, superficially, a short
time before it has entirely separated the egg into two dis-
tinct halves. The last point to be cut off corresponds in
position almost exactly with the nucleus, but on the oppo-
site side of the egg. This first cleavage is completed one
hour after fertilization. The two daughter-nuclei now lie
near the plane of fission, and at the same distance from
the surface as the original nucleus was. The second furrow
normally starts at the same point on the surface as the
first, and again divides the egg meridionally, in a plane

20

at right angles to the first. Sometimes tiie second furrow
starts iyrregularly , at a point part way around the egg from
the origin of the first furrow. One of the hemispheres is
thus divided before the other, as in figure 6. "Fifty min-
utes elapse between completion of the first and second fi^r-
rows. Suc':'essive segmentations come in at intervals of 45
to 60 minutes.

With the eight-celled stage rotation of the blasto-
meres occurs. The four upper ceils turn through an angle
of Ard^wpon the lower ones, so that they come to lie in the
valleys between the four lower ones instead of being exact-
ly superimposed upon them. Segmentation continues until a
hollow blastula (fig. 7) is produced, a layer of thick
cells surrounding a small cleavage cavity. The cells are
of unifonn thickness, and their outer ends give rise to
cilia which drive the egg round and round by their motion,
within the membrane, sometimes in one direction, sometimes
in the other. During this stage the formation of the endo-
derm takes place. The inner ends of the blastoneres are
delaminated, the process going on at an equal rate on all
sides, until a unifo):™ layer of endoderm cells lies within
the ectodermal layer ("Fig. 8). By increase in size of these
endodermal cells the cavity of the egg is entirely obliter-

21

ated. During the subsequent life-history of the larva no
cavity exists in the body until several other marked chanj^-
es have taken place.

IV. TIIE PLANULA - By the rupture of the egg-merabrane the
nearly spherical ciliated larva makes its escape and starts
upon the stage in which it is a swimming planula (Fig. 9).
Its shape soon changes, becoming longer, and more narrow
a+ one pole than the other; this narrower pole is to be the
future oral extremity of the larva. The cilia serve to
propel the planula in a slow rotating progression through
the \'/ater, usually not far from the bottom. The larger end
is directed forward in swimining. The time at which the
planula appears is in the morning, about twelve hours after
fertilization of the egg took place. The length of the
larva is now between .1 and .15 mm. ("^ig. 9). This condi-
tion persists for a varying time, usually several days.
Towards the end of this time the first indications of a coe-

lenteric cavity appear in the arrangement of the cells at
the posterior end of +he swimming larva ("Fig. 10). Their
inner margins come to lie in a straight line, following the
long axis of the larva (Fig. 10,G). This is better under-
stood when we notice that in changing its shape from the
spherical morula to the elongated planula the larva also

22

underwent a slight rearrangement of its cells. The endo-
derm was when first formed in the shape of a spherical .
iriass, a-'id its cells were all conical, rauiatinf^ from the
centre to the surface. Eiit as elongation took place in the
fonnation of the planula, the cells were stretched out into
a cylinder, and +heir inner ends overlapped irregularly, as
is sho'^m in the anterior end (A) of figure 10. v/hen the
coelenteron begins to be developed, the inner ends of these
upper endoderinal cells change their position somewhat, and,
as stated, meet along a continuous line. At the same time
a change is to be noticed in the ectodermal cells at the
exterior of the future oral pole (Fig. 10, 0). The cell-
walls at +his point become less distinct, and finally a
disintegration of the boundaries leaves the tissue an un-
differentiated layer of protoplasm. Before separation of
the tissue to form the definitive coelenteric cavity, the
larva stops swira^ning, loses its cilia, and settles do'.vn up-
on the bottom. The larger end, which was directed forward
in swinrriing, is dovmv/ard. Between the free-swimming stage
and the sessile hydra stage there freciuently, though not
always, intervenes a condition which reminds one of a mi-
nute planarian in its shape and movements. The planula
settles down upon the bottom, and slowly glides along by a

23

rytliinic wave-like prof^ression. This condition seems to
take the place of the last part of the ordinary and prob-
ably more normal free-swimming stage, and is perhaps due to
the unfavorable conditions of the laboratory. This condi-
tion is not, hovfever, like the pathological plasmodial
forms mentioned below. Its changes in shape are slight, and
its maTiner of luovement more of a glide than a protoplasmic
flowing. None of the definiteness of structure is lost,
and these larvae change into hydras as soon as those which
transform directly from the free-swimming planulae. It is
then not a phenomenon of degeneration, nor, on the other
hand, an essential phase in the life of the creature, but
rather an intermediate and probably accidental condition.

V. I-r!i''DRA»- As soon as the planula stage has given place
to the settled hydra stage, the coelenteron becomes com-
plete. The mouth appears at the tip of the free end, where
the tissue has previously showed indications of disintegra-
tion, at the end of the axial line formed by the endodermal
cells. At first the mouth is visible only v;hen the speci-
mens are killed and cleared or sectioned. Soon, however,
it becomes large enough to see in the live animal, by fo-
cussing down from above v,'ith a high-power lens; it appears
as a minute pit in the ectoderm. The coelenteron is more

24

distinct at the upper end than below, where it disappears
into the loosely constituted cell-mass in the interior.
The definite cavity of the coelenteron is sometime later in
making its appearance. When finally established it is
lined with a thick layer of columnar endodermal epithelium.
At its bottom it flares out in following the contour of the
body-wall, as it appears in figure 14, which shows a late
stage, but the same condition of the coelenteron as exists
in the newly transi'ormed larva. The figure also shows a
thickened core of endoderm which projects upward into the
coelenteron as a gastric peduncle (G.P.). This conical
mass of cells develops during the later part of the hydra-
stage,

A. TENTACLES. -

In the later transformations of the developing Go-
nionema no definitely determinate periods separate the
times of active change. The development-time is variable,
depending upon external conditions of food, temperature,
etc. In an average larva, however, the first tentacles
make their appearance during the third week after the fer-
tilization of the egg, or a week after the commencement of
the fixed polyp-stage. Two tentacles appear opposite to

25

one another at a level about one-quarter ol" the distance
from the upper pole of the hydra (Fi^, 11). They are knob-
like v/hen they appear, but grow rapidly to a considerable
length, the few rjndodermal cells which form the core of
each increasing in number. Figure 11 shows a vertical sec-
tion of a two-tentacled polyp of the fifth week. The man-
ner of origin of the tentacles will be described in the
section on "Origin of Tentacles," under "The Medusa".

The second pair of tentacles (Fig, 13) appear soon .
after the first, and by their rapid growth soon become as
large as the first pair from which they are then no longer
distinguishable. Irregularities are common in the appear-
ance of the tentacles of the hydra, as in the adult. It
frequently happens that only one of the second pair ever
makes its appearance (Fig. 24); or one may be slow in
arising, and always remain smaller than the other. On the
other hand, an abnormally large number of tentacles is fre-
quently developed; individuals with five or six tentacles
are not rare (Fig. 14).

B. Fonn of Coelentoron. -

The appearance of the tentacles is accompanied by al-
terations in the form of rho coel;nteric cavity. The rapid

26

growth of the cells at thepoint where the tentacles arise,
and the out-pushing of the tissue in the process seems to
affect the tissue over a considerable area, so that diver-
ticula of the coelenteron and of the mouth extend in the
direction of each of the tentacles, A stellate form re-
sults, the mouth being in the shape of ^. cross. This cor-
responds exactly to the condition in the medusa, especially
in young specimens (Fig, 2G ) in which the twisting which
obsc^ires the true relation of parts has not yet taken place*
In a three-or f ive-tentacled hydra the number and arrange-
ment of the oral lobes corresponds with the number and ar-
rangement of the tentacles. Figure 12 represents a polyp
with five radial parts, in which one lobe of the mouth is
bifurcated. This condition is very similar to that fre-
quently met with in the adult medusae. Hargitt figures a
specimen in his "Variations among Hydromedusae" (1901, PI.
Ill, Fig, 11) which is much like this. The whole aspect of
the hypostome of the Gonionema polyp is very similar to
that of the manubrium of the young medusa (Fig, 26), The
ectodexTn at the edges of the mouth becomes thickened and
armed with nematocysts, which have by this time appeared,
in the manner to be described later, over much of the body
of the hj/dra. Below the mouth, the hypostome becomes nar-

27

row and tubular, and distinct from the x-est of the body, a
decided angle separating them at the level of the tenta-
cles,

VI. Habits . - One of the most striking habits of the
adiilt jelly-fish is its prehensile propensity. The adhe-
sive organ at the "knee" of the tentacle is composed of
long slender glandular cells, packed into a thick cushion
vfiiich is enclosed v/ithin a strongly muscular i-im or collar
(Fig. 24). This organ is located on the aboral side of the
tentacle(Fig. 25). When at rest the jelly-fish lies on the
bottom v/ith inverted bell, and tentacles widely extended
horizontally and attached to the bottom by means of the
combined vacuum-cup and cement-gland near the tip. How
this habit came to be acquired by the adult medusa it is
hard to see. But if, as I shall give my reasons for be-
lieving, the medusa arises by direct metamorphosis from the
hydra, the habits of the hydra would naturally be more or
less perrmnent in the adult. It may be that this partic-
ular habit is more likely to be first acquired by the larva
than by the adult. The tentacles of the hydra reach a rel-
ative length greater than in the case of any other known
hydroid polyp, I believe. They frequently stretch out in

28

the water for a distance three or four times the height of
the polyp. Figure 13 shows a hydra with the tentacles
fully extended, and their tips touching the ";round. This
is the habitual attitude. The drooping of the tentacles is
evidently caused by their extraordinary length, and is al-
most as unusual an occ rrence among the hydromedusae. At
the points where the tips of the tentacles come in contact
with the bottom they spread out somewhat, forming a sole-
like surface v;hich is closely applied to whatever object
the polyp is settled upon (Fig. 13 P.T). This smearing out
of the tentacle-tips is like what occurs in live specimens
of Hydra held between slide and cover-glass for examina-
tion. It is not known that the cells at the tip of the
tentacle of Gonionema undergo tr^insformation to form gland-
colls and an adhesive organ, but such a change seems at
least possible. At any rate, the fact is that the habit ot
the polyp is much like that of the medusa; both remain when
at rest with the mouth expanded, the manubrium stretching
upv/ards, the tentacles widely extended and adhering to the
bottom. When an animal swims against one of the tentacles,
the reactions are much the same in polyp and adult. The
feeding habits of Gonionema have been described at length
by Yerkes (Sensory Reactions in Gonionemus" - 1902). His

29

account would apply almost as well to the process in the hy-
dra. The tentacle which comes in contact with the prey is
contracted with a suddenness and vigor which belies the ap-
parent inertia of the moment before. The animal is seen to
be firmly spitted on the microscopic lances of the nemato-
cysts, and it is evident that the first thing that happened
when the animal touched the tentacle v/as the dischar^^e of
all the thread-cells in that region. The tentacle in con-
tracting carries the food, protozoan or minute woiTn or
whatever, towards the mouth. The long manubrium then moves
slightly, as if in search of the morsel. Finally, the ten-
tacle places the food directly upon the mouth, which pro-
ceeds to tux-n itself over the object and work it downwards
until it vanishes into the gastric pouch of the polyp.

VII. DEGENERATION PHENOMENA.

Owing to some peculiar condition in the water of an
aquarium jar, the larvae of one lot, when three months old,
began to exhibit most singular forms and activity. All ap-
pearance of the hydra was lost, ectoderm and endodei'm be-
came indistinguishable, and cell outlines dissolved. The
larvae in this condition had very much the appearance of
amoebae. They slumped down on the bottom of the aquarium,

30

a shapeless mass, and by protoplasmic flowing changed their
shape through an endless variety of forms, and moved slowly
from place to place. Thin pseudopodia were sent out along
which the substance of the organism flowed, and by the
bi-eaking of the connecting band divided into two. The
fragments became smaller and smaller until no longer recog-
nizable. These abnormal larvae remained alive for six weeks
after v/hich no trace of them remained.

VIII. BUDDING IN THE LARVAE - tIETAGENSIS. -

Contrary to Haeckel's statement that in the group of
jelly-fish which he calls the"Trachomedusae" metagenesis
doesnot occur, in Gonionema , which falls into that group,
metagensis does take place. Ey a form of non-sexual multi-
plication diffei'ent from any which has previously been de-
scribed for any member of the Hy dromedusae , an intermediate
process of reproduction is introduced into the life-history
of Gonionema. whereby a large munber of adults --are produced
from a single egg.

Asexual multiplication in the larvae of Scyphomedix-
sae has been known since 1841, when Sars saw and described
the formation of buds in a scyphistoma of uncertain identi-
ty, but regarded as either an Aurelia or a Cyanea. Since

31

that time several analCTous cases have been made kno^/m.
The scyphistoma larvae of Cassiopea , for example, were
found producing buds in larr;e numbers by Bigelow (1900), who
gives a detailed account of the method of budding in his
monograph on this Rhizostome, It may be further stated in
general, that the non-sexual process of production of buds
by the larvae is an important method of multiplication
among the Discomedusae, The buds usually develop, after de-
tachment from the parent polyp, into a second genei^ation of
scyphistomas , identical in form and fate with the original
ones. Buds may arise on the body of the scyphistoma, or
upon stolons from its base, and either singly or several
at a time. In Cotylorhiza the buds develop so rapidly and
remain attached so long that large clusters accumulate
about the base of the scyphistoma. According to some au-
thors, Goette, for example, the distal end of the bud in
Aurelia and Cyanea is destined to become the oral end of
the detached larva, developong mouth and tentacles. Fried-
ma nn , on the other hand, says ("Postembryonal dev, of Aure-
lia aur i t a " 1902) that in Aurelia he has found the opposite
condition, the mouth being invariably developed at the at-
tached end of the bud. This is the common relation in oth-
er forms.

32

In Cunina, which falls into Haeckf^l's order, the
"Narcomedusae" , the ciliated, tentacled larva multiplies by
buds, i^roduced from an aboral stolan. These buds are not
detached until mouth, digestive cavity, and tentacles are
well developed. Several are produced simultaneously, and
are attached to the parent by the oral extremity. The de-
scription of this remarkable process is given by Dr. W, K.
Brooks, in "The Life History of the Hydromedusae" , (1S86),

It is my puii)ose in this section to give an account
of a process of budding in a very diffei'ent medusae from
Cununa, and one in v/hich the asexual multiplication takes
place quite differently. In Gonionema the buds are produ-
ced in a manner which reminds one very strongly of the sim-
ilar process in Cassiopea.

In the course of my general study of the embryology
and development of Gonionema I came upon the budding larvae
(Figs. 14 -23). From a lot of eggs obtained at Wood's Mole
in Aiigust, 1901, a large number of polyps developed and
were kept alive in a "balanced" aquarium for several months.
This lot was left at V/ood's Hole in as nearly natural con-
ditions as possible until the last of November. The water
was kept fresh by frequently r-enewed supplies of diatoms

33

and ulva, and occasionally changed by carefully adding a
quantity taken from the natural habitat of the medusa in
the eel pond. A low temperature was maintained. When
these larvae were received from V/ood's Hole, November 28,
they were apparently thriving well. They were settled upon
Minot watch glasses which were easily removed from the
aquarium- jar without disturbing their contents. These
dishes were numbered and the positions of the polyps care-
fully noted and mapped. Successive examinations shov/ed
that the number of polyps was on the increase, and, on De-
cember 3, it was seen that one or two ol' the largest spec-
imens had rounded knob -like bodies upon their surfaces,
which were recognized as buds. These specimens were exam-
ined as frequently as it was thought safe to remove them
from the jar, and camera drawings were made of the growing
buds. Observations were made of the different stages in
the development of fourteen buds; their phases agreed in
all the main particulars.

The first indication of the appearance of a bud upon
any individual polyp v/as seen as a rounded eminence upon
the hydrocaulus (Fig. 15, b). It was usually located at a
level about half way between the base of the polyp and the

S4

rins of tentacles, as in figure, and interradially , i.e.,
at the end of a radius which bisects the angle between two
tentacles (Fig. 19), Never more than a single bud appeared
at one time on any polyp.

All three body-layers of the parent are involved in
the formation of a bud (Fig. 16). The cells of both ecto-
derm and endoderm multiply rapidly in the region of the
wall of the polyp where the bud is about to be formed. The
endoderm pushes out as a rounded protuberance covered by
the ectoderm in a layer of constant thickness. A thin sup-
porting lamella of mesogloea lies between the two. As the
bud increases in size it bulges out at its base, around the
stalk which connects it with the polyp, and it also devel-
ops rapidly at the tip of the free end. In this way it be-
comes pear-shaped (Fig. 17). As the drawings indicate, the
ectoderm is of the same thickness in the bud as in the pa-
rent polyp (Fig. 16, ect.); indeed, so nicely regulated is
the rate of growth of the two tissue-layers that the thick-
ness of the ectoderm dees not change appreciably during the
entire grov/th of the bad, previous to its detachment. The
cells of this latter layer are iri-egular, loosely constitu-
ted, and coarsely granular, and their walls are hardly dis-

35

cernible. No cavity exists in the bud until considerably
later. The endoderm of the bud now becomes separated
From that of the parent by the constriction of the ectoderm
and the emitting off of the core of endoderm which filled
it. Its appearance is as represented by figure 18, an isth-
mus of clear elastic ectodermal tissue (ect.) uniting the
bud to the parent. By rapid centrifugal growth the bud be-
comes sausage-shaped, and as long as the diameter of the
polyp. Soon after the bud reaches the stage shown in fig-
ure 19 it becomes detached from the polyp. In only one in-
stance was I 3o fortunate as to see this process taking
place. In this individual the bud was drawn out into a
long finger-like body, its distal end drooping almost to
the ground. Soon the ectodermal isthmus began to stretch
out and dwindle in diameter until it was mei'ely a thin stem
of transparent protoplasm (Fig. 20). The bud seemed to be
reaching out and trying to free itself from the limitations
of the connection with the parent. This stretching of the
isthmus was bi'ought about by constriction of the tubular
ectoderm, as if by circular muscle-fibres. When this
stretching had gone on until the isthmus was a quarter as
long as the entire bud, (Fig. 20), it be/':an to grow still

36

thinner at about its middle, and finally, just half an hour
after it first bf3gan to stretch out, it broke in two and
the bud fell away from the parent (Fig. 21). T}ie two ends
of the connectins-stalk shrank back into the tissue of the
bud and the X)arent, appearing for a time as drops or points
of protoplasm (Fig. 21). AJ'tor separation from the polyp,
this particular bud settled down at once upon the previous-
ly free or distal end, and began an independent existence
(Fig. 22). Otlier observations, however, indicate that the
usual course of development is slightly at variance with
this instance, and that it includes a period of from three
to four days, intervening between the detachment of the bud
and its settling down as a hydra, during which it has the
form of a free-swimming planula. A bud which v/as growing
upon the body-wall one day would be gone the next, and for
some time could not be found. Then suddenly it would ap-
pear in some previously vacant spot, at a distance from the
Ijolyp, perhaps in an entirely different watch-glass on the
bottom of the aquarium, with its tentacles just ber^inning
to appear. In one case the bud was drawn and measured , when
it seemed to have reached its full size and to be ready to
drop off; this was done one evening, and the next morning

37

the parent polyp had no bud on it. Threo days afterward a
small creature v/as found upon a spot which had been cer-
tainly unoccupied by any larva up to that time, careful di-
agrams having been made as stated above. This was a larva
like that in figure 23. It v^as measured, and although
changed in shape, as nearly as one could estimate its bulk
it cori-esponded exactly with the bud which had disappeared.
Similar experiences were repeated so many times that it
seemed unavoidable to consider this a normal phase in the
process. This being so, it is an interesting case of re-
version in the non-sexually developed larva to a condition
earlier in point of ontogenetic order than that in which the
parent v/as when it gave rise to the bud. In any case, the
future history of the bud is certain. After settling down
upon the bottom, it repeats the changes which occur in the
sexually produced polyp. The newly arisen larva (Fig, 22)
loses its planula shape, becoming shorter and thicker, es-
pecially at the base, on account of the plastic character
of the tissues (Fig. 23), It has now secured a firm hold
upon the bottom, being so closely applied to it that it can
be dislodged with difficulty. The cells at the base in-
crease in thickness until they form a columnar epithelium

38

(Fig. 20, ect ). After the i'irst day a slir^ht pit indicates
the point at which the coelenteron is to open externally
(Fig. 23, .J ); this process, as observed in a number of cas-
es, is exactly the same as in the sexually produced polyp.
The tentacles also make their appearance in tl^.e same manner
as described for the hydra which came from the egg.

The length of time I'equired for the complete devel-
opment of a bud, from its first appearance on the hydrocau-
lus of the parent as a round knob, until the comp^letion of
the formation of the coelenteron and appearance of the ten-
tacles, is from ten to fourteen days; (a) the first period,
including as far as the detachjnent of the bud, five days;
(b) free swimming planula, two to four days; (c) from at-
attac?unent to appearance of tentacles, three to five days.
These periods refer, of course, to specimens which develop-
ed in captivity.

Figure 24 shows a specimen from an entirely differ-
ent lot of polyps from those which exhibited the budding
phenomena shown in figures 14 to 23. This polyp was killed
when 23 days old. It may not be a normal individual, but
shows a tendency to divide transversely and seems worth
calling attention to. The coelenteron has completely di-

39

vided into two, and the endoderrnal wall of the pouch has
crown in as a solid partition between the two new pouches.
'L'he aboral portion of the body is seen to be considerably
longer than is usually the case.

IX. TRANSFOraiATION OF THE POLYP. -

Up to the present time all efforts to secure
specimens of the larval Gonionema in their natural habitat
have been well-nigh fruitless. Although the eggs are laid
in enormous numbers during four to six weeks of the summer,
and even when kept in the laboratory a large proportion de-
velop , it has yet been impossible to find the polyps in
the eel-pond where the medusae are so plentiful. Many
speculations have been hazarded as to the condition in
which the lai'vae pass the cold months of winter, and no
small energy and time have been expended in attempting to
get at the secret. And yet I am much more ready to believ(
that the difficulty has been with our methods of search
than that any extraordinary transformation in form or
change in habitat should render the success of such search
impossible. This seems the more likely from the fact that
during the season when the medusae are laying their eggs

40

most plentifully, and within a fev/ days from the time when
an e[rg is laid it has developed into a fixed polyp with
tentacles, and such polyps must be present in great mimbers
in the mud and on the stones at the bottom of the eel pond,
not even then have more than a very few specimens been
found, although search has been made by others than myself.
It is quite out of the question to suppose that the larvae
which develop into the great numbers of medusae which ap-
pear year after year in the eel pond have been swept out to
sea to undergo their transformations in deep water, because
in such case the adults v/ould appear in much wider range of
locality, in some of the bays and inlets of the coast where
the conditions seem to be almost the same as in the eel-
pond. The fact is that only a few stragglers are ever
found in the vicinity - as many as would readily be swept
out of the shallow water by the tide. Not only these con-
siderations, but all the other indications seem to point to
a direct transformation of the polyp to the adult gonosome
v^ithout leaving the eel- pond. The habit of the polyp of
resting with tentacles extended and adhering to the bottom,
the feeding reactions, the form of coelenteron, manubrium
and oral opening, and the manner of origin of the tenta-

41

cles, all resemble the corresponding conditions in the
adult so closely that it is easy to regard this as the most
likely theory. May it not be that the same type of meta-
morphosis as that which takes place in Liriope (Brooks,
1885) is passed through in this genus as well? In Liriope
the coelenteric cavity is transformed into the system of
chimiferous tubules by the growth of fusion areas uniting
the upper and lower walls of the cavity except where they
are to be left separate along the lines of the canals. Fig.
25 ±1} a C'imera drawing of a twelve-tentacled gonosome of
Gonionema . which has very much the appearance of the newly
metamorphosed Liriope (Brooks, 1885, PI, 41; Kaeckel, "Die
Russelquallen" , PI. 12), The transformations v/hich are ne-
cessary to bring about the adult form from the larval are
the change in the coelenteron from a pouch to a system of
tubes, the centralizing of the diffuse nervous system to
form the two nerve-rings, the appearance of new tentacles
provided with adhesive disks and of tentacles modified to
the form of sense organs, from the expanded tentacular
ring, and the growth of the velum. The relative size of
the fully developed polyp and the yoimgest mediisa offers no
contradiction to the conception of such a process of direct

42

metamorphosis; if the polyp grows as rapidly in the natural
environment of the eel pond as in the laboratory, even al-
lowing a long period of absolute quiesence during the cold
weather, the discrepancy in size is easily accounted for.

X, YOUNGEST MEDUSAE, - Arrangement of Tentacles and
and Sense-Organs.

During the last of June, 1900, a number of very
small specimens of Gonionema were taken in a tow-net at the
surface of tlie eel pond. Several of these had sixteen ten-
tacles, some had twelve, one had only eight. A careful
study of these very young and evidently recently metamor-
phosed gonosomes has brought out some exceedingly interest-
ing points,

Hargitt (1901), in his paper on "Variations among
Hydromedusae" , discusses the arrangement of tentacles in
Gonionema; he approaches the qitestion as a student of vari-
ations, and unfortunately lacks the young material from
which I have found it possible to educe very definite rules
in the arrangement of marginal organs, and their order of
appearance. As a natiiral result Hargitt forms the conclu-
sion that so much irregularity occurs as to render it im-

43

possible to discover any definite order of appearance or
ultimate arranfrement in these organs. It is true that the
abnormal specimens which he studied most closely do show
very little regularity - as would be expected. But in nor-
mal individuals quite a remarkable degree of precision is
manifest in the position and order of appearance of tenta-
cles and sense-organs, with reference to each other and al-
so to previously arisen organs of the same kind. This is
particularly tn^e in the younger stages.

If we examine the eight-tentacled medusae the follow-
ing points are noticeable: First, the tentacles are evi-
dently of two cycles, in order of appearance. The four at
the ends of the radial canals, or the perradials, are equal
in size, and larger than the four interradials , which are
also equal in size. These tentacles are very similar in
structure and appearance to the larval tentacles, and there
seems little reason why the four larger perradials may not
be the permanent larval tentacles.

Second, the sense-organs are foux' in number and
placed in definite positions, relative to the tentacles.
If we look at the bell-margin from the oral side, the newly
arisen interradial tentacle in each interradial <iuadrant

44

has apparently crowded in between the sense-orn;an and the
perradial tentacle which comes before it, as the hands of a
watch go. Text-figure 3 shows this stage, and is made from
a camera di-awing of the eight-tentacled medusa. And the
relation which is here exhibited in the youngest stage of
the free-swimming gonosome is the same throughout the
growth of the creature: "/her ever a_ rudimentary or n ewl y
arisen tentacle lies on the bell-margin, it will always ,
normally . be found to lie ,1 u s t in frorit qf_ a_ newly arisen
sense-organ and just after a_ large tentacle. i,e, , one of a
much earlier cycle,

r'uch has been written to show that coelenterates ,
and especially the Hydrozoa, sliow bilateral symmetry, eith-
er in the normal condition or when they depart from the
normal form and may be supposed to revert to a more prim-
itive type (Mayei', 1901, e.g, ). Gonionema shows a very
different plan fx-om that of bilateral symnnetry. It is rath-
er a certain sort of radial symmetry v/hich has nothing bi-
lateral about it - one in which the radial parts cox-i-espond
exactly to each other and are superimposable , but none of
which is the reflected image of any other. I shall call
this relation one of cyclic symmetry, With reference to

45

the order of appearance of the marginal orf^ans I shall
speak of cyclic sequence. (These terms wore sugj^ested by
Professor f'or-ley of Johns hop-kins). New tentacles make

their njipearance four at a time, or, so to speak, in quar-

o
tettos; they are 90 apart, so that they occupy identical

positions in the four quadrants of the marginal ring. But

while the tentacles, and sense-organs as well, appear to

arise in fours, the condition in the larvae and in frequent

instances ai^iong the adults indicates that a paired origin

is more primitive and fundamental. It is the universal

rule in the early larvae that two tentacles appear opposite

to one another (Fig, 11) and later the second pair of the

quartette (Fig. 12). And it often happens that in the

adult medusae, two members of a quartette, in opposite

quadrants, are retarded in making their appearance. In

Aurolia, Glaus established the theory that while the larval

tentacles seem to come in fours after the earliest stage,

the first four tentacles appear first two, then two more,

as is the case in Gonionema ; (v. Glaus, 1S92), Goette

(18S7) has examined a great number of specimens of the

younger stages of Aurolia, and has come to the same general

conclusion as Glaus, with regard to the primitive paired

46

condition and the significance of this in the philogeny.
HaeckeJ (ISSl), however, regards the appearance of two ten-
tacles in advance of the second two as an accidental and
insignificant occurrence, and takes four as the primary
niimber. '.Vhile this tendency to a paired origin of the ten-
tacles disappears after the earliest stages in Aurelia .
(jonionema exhibits this tendency in frequent instances dur-
ing the whole life of the animal. Its occurrence in the
appearance of the sense-organs is of the same significance,
because as will be pointed out below, these organs are mod-
ified tentacles. Figure 31 shows this condition in the
sense-organs, quadrants A and C having 5, while in quad-
rants B and D only 4 are yet developed. It is tirie that
other variations than these do occur in the appearance of
the tentacles and sense-organs of the adult, and in the
tentacles of the larva. Polyps with three, five or six
tentacles are not uncommon (Figs. 12. 24). It is notice-
able that departures from the normal number of tentacles
cori-espond very closely in polyps and adults. This would
be expected to be the case from the evidence that the lar-
val tentacles are permanent, and that they determine the
position of the four radial canals in the noi-mal medusae,

47

or of the three, five or six in the specimens v/hich are
aberrant. This inference seems a likely one, from the fact
that in the adult medusa the tentacles which from their
larf^er size are presumably of the first cycle are always lo-
cated at the ends of the radial canals. Wliile a few excep-
tions are found, they fre^iuently occur in cases which have
apparently become abnormal by accident subsequent to meta-
morphosis. The inference is, then, that five-parted adult
medusae were f ive-ter tacled larval polyps. And this is
borne out by comparison of the relative numbers of each
kind of variation among medusae and among polyps, hargitt
(1901) has tabulated the number of medusae that have come
under his notice having three or five oi- six radial canals;
and he finds that about five per cent, of all medusae are
irregular in this regard, i.e., that they vary from the
normal four-parted condition. While I have not had a great
number of specimens of the polyps from which to compute av-
erages, my counts show quite a striking similarity to those
which are given by Dr. Hargitt for the adults.

Among all the varieties of geometrical figures which
appear in the ai-rangement of parts among the various orders
of coelenterates , there is none, so J'ar as I can find,vihich

48

is at all comparable with that which appears in the arrange-
ment of tentacles and sense-organs in Gonionema, The only
suggestion of siich a plan or arrangement as this is given
in a paper on the later development in Aiirelia , by Friede-
mann (1902). In the course of the paper the author de-
scribes the origin of the eight tentacles which follow the
original eight. Four of these appear at once, the other
four later. In the ajjpearance of the fii-st four, two pos-
sibilities arise, according to Friedmann. Either the four
arise in bilaterally symmetrical positions in the foiir
quadrants, the two halves of the tentacle ring being re-
flected images one of the other, and the new tentacles ap-
pearing one on either side of two opposite perradial tenta-
cles (text-figure 1); or else they appear in identical po-
sitions in the four qiiadrants, one appearing next in front
of each perradial tentacle, as the hands of a watch move
(text-figure 2). Friedmann' s figures do not make it clear
that he actually found specimens in exactly this state. It
appears moi'o probable from his descriptions that he inter-
preted older stages by this theory. But it may easily be
true that in other groups than that to which Gonionema be-
longs the tentacles originate according to a plan of cyclic

49

symmetry, or th-^t such a condition may ajjpTar in occasional
instances, as is perhaps the case in Aurelia, In Gonionema
the rule holds with reriarkable constancy.

From a study ol' the successive stages of growing me-
dusae the following table is compiled, to show the I'elation
in time of appearance of tentacles and sense-organs. The
numbers in brackets, in the coliunn of sense-organs, indi-
cate half-quai'tettes , the corresponding pair in each case
having been delayed in appearing. Since the sense-organs
are only half as numerous as tentacles, they appear with
half the rapidity, and are therefore more frequently found
in this incomplete condition. That is, if a jelly-fish

were killed when the tenta-

Tentacles Sense-organs

rval I 2

olyp)| 4

Adult
(Medusa)

'2)

f S (6)

12 8

1^3 (10)

^0 12

24 (14)

28 16

36 20

40 (22)

44 24

48

cles and sense-organs were
in the precise stage indi-
cated by the line A.B, the
fifth quartette of sense-
organs would be found only
half formed - five sense-
organs appearing in two op-
posite ciuadrants, and only
four in the other two. This

50

is just the condition which exists in the specimen sho^vn in
figure 26, The mimbers indicating the s ense-oi'cins are put
at the intex-vals between t'-iose indicating tentacles, to
show that while the eight tentacles between the eight-tenta-
cled and the 16 tentacled stages, for instance, are appear-
ing, the four sense-organs which make up the second quad-
rant are appearing.

Another rule is followed by the marginal organs in
their order of apjjearance; P'ach new tentacle arises not on-
ly just in front of a sense-organ, but in a definite loca-
tion with reference to the tentacles already present. And
the same thing is ti-ue for the sense-organs. It is there-
fore possible, from a study of successive stages, to pre-
dict where each new quartette of tentacles or sense-organs
will arise. The diagrams shown in text-figures 3 to 8
are from camera drawings of mounted whole specimens of me-
dusae. If we examine text-figure 4 v/o see that T (tenta-
cle) III follows T I; and text-figure b shows TIV following
T II. Thus T III and T IV form a series, arising in corres-
pending positions in the (juadrant relative to the tentacles
already present. (By "series" is not meant "cycle", with
the idea of simultaneous appearance; tlie notion is rather

51

one of relative position, simply.) The next series of four
tentacles in each quadrant, T V to T VIII, brinf^s us
to the 32 tentacled staf^e seen in text-figure 6, In this
it will bo seen that T V follows T I, T VI follows T II,
T VII follows T III, T VIII follows T IV, four numb'-rs in-
terveninp- in each case. The next series comprises T IX to
T XVI, which follow the same plan in order of appearance,
T IX following T I, etc., eight numbers intervening in each
case. While this may seem r'lore like a fanciful exercise
of the imagination than an actual condition in nature, the
truth is that the larger the number of specimens in which
one tests the arrangement of the marginal organs by this
riile, the more will one be convinced of the remarkable con-
stant adherence to it. Given a specimen with, say, 28 ten-
tacles, such as that represented in text-figure 7; this is
a drawing of a specimen of Olindias , from the Bahamas, a
genus which follows t?ie same rule in the order of appear-
ance of the tentacles as that which is being described for
Gonionema: the sense-organs are not so numerous in Olindias.
In this specimen the most recently arisen tentacle in each
quadrant is evidently the one numbered VII, lying just af-
ter each perradial tentacle. Then, if the rule which we

52

?iave educed applies in this case, we should expect to find
the eighth tentacle in each quadrant arisinf^ in a corres-
ponding position with relation to the interradial tentacle.
And such we find to be the case. Text-figure 8 sho"/s a
slightly older specimen of the same species, in which we
plainly soe the eighth tentacle in each quadrant lying in
its appointed place (VIII),

It would be singular indeed if there appeared no ex-
ceptions at all to this regular cyclic sin-rmetry. Many va-
riations do occur, but not more than v/o should expect from
the marked tendency to variability in all parts of the me-
dusa. These variations no more obscure the normal definito-
ness of plan than the occurrences of six- or seven-rayed
star-fish obscure the normal pentameroi;s form in echinodenns,
Text-figure 5 shows an irre-^ular condition, the fourth ten-
tacle in each quadrant having arisen aberrantly, following
instead of preceding the first sense-organ (1). In figure
31 one of the latest arisen <iuartette has not piit in its
appearance (see arrow in quadrant A). In the older speci-
mens, the number of irregularities increases. It seems to
me that the bell-margin increases in extent subsequent to
and as a consequence of t?ie increase in the number of ten-

53

tacles, rather than that tlie tentacles arise, haphazard,
wherever there is space enough on the margin to accomodate
them (Ilargitt, 1901), Certain it is that the most crowded
part oi' the bell-m.argin at any particiilar time is that from
which the new tentacles arise.

XI. HISTOGENESIS OF MARGINAL ORGANS. -
A. In the Larva,

The similarity in the appearance of the ru-
diments of a tentacle in polyp and in gonosome make it de-
sirable to deocrihe both in the same section; for that rea-
son the account of the origin of the tentacles in the polyp
was reserved for this place. At first the larval tentacle
is merely a small roi^nded knob in external appearance, and
internally it is made up of a core of two or three endoder-
mal cells. When tPie tentacles make their appearance the
body v,'ail of the polyp is made up of a doixble layer of
cells, the ectoderm and the endoderm, separated by a thin
supporting lamella of mesogloea. These three layers are
pushed out somev/hat in the growth of the tentacle, the re-
gion of greatest activity being in the endodermal layer,
where the core of the tentacle is formed by a rapid out-

54

gx'owth of the cells of the body wall, accornpanied by mul-
tiplication of these s^nvi cells. After some 'weeks the cav-
ity of the coelenteron becories dravm out in a diverticul^^m
in the direction of the axis of the tentacle, so that the
upper part of the gastric cavity becomes stellate in cross-
section, Fi?Ture 14 shows this condition in a five months
old polyp. This cavity does not reach out into the tenta-
cle itself in any of the specimens which I studied, but may
do so before metamorphosis takes place. During the whole
of larval life, the tentacle is made up of a core of endo-
dermal cells in a single row (Fig. 27), as is the case in
hydra. Figure 11 shows a polyp with the first pair only de-
veloped, and the cell-layers are seen as described. The
endoderm cells are filled with a loose protoplasmic net-
work (Fig. 27, end), and the nucleus is evident. The con-
dition which is seen in an adult tentacle v/ith several
cells of endoderm surrounding the central cavity (Fig, 34)
is easily derivable from the larval condition, by longitu-
dinal fission of the endodermal cells repeated until a
cross-section of a tentacle at any level cuts several cells
(Fig. 34).
B, In the Adu.lt. - The regularity with which the tenta-

55

cles and sense-organs iiake their appearance in the adult,
as previously described, makes it possible to locate with
coniparitive certainty the beginnings of one of these organs
on the bell-margin. Figure 18 is from a section of a medu-
sa, cut horizontally at the point of origin of one of the
tentacles. The figure shows tlie aspect at the level of the
tentacle, somewhat above the velum. Both cell-layers are
seen to be concerned in the formation of the new tentacle.
The endoderm (End. ) is pushed out from the region of the
circular canal, and has the shape of a solid plug of tissue
composed of a few cells arranged radially about a central
axis ( T.R,); the nuclei are at the inner ends of the cells.
Outside of this endodermal core is the ectoderm (Ect.)
which is, in the region of the bell-margin, of the charac-
ter of gelatinous tissue, stiffened by the presence of a
large number of concretions (Cor). These concretions and
the nuclei of the cells are more numerous at the point
where the tentacle is to appear than elsewhere. In the me-
dusa, as in the polyp, the greatest activity in the forma-
tion of a new tentacle is manifested by the endoderm. Ac-
cording to Allman (Monograph, 1871), in some hydroids (Cam-
panularia Johnstoni . for example) the first indication of

56

tentacle formation is the thickening of the ectodorm at the
point where the tentacle is to appear. This is contrary to
the condition which we have in Gonionema* But to continue
our description; along with the growth of the endodermal
prpcess v/hich is to be the core of the tentacle the ecto-
derm also increases rapidly, and constitutes an investment
which contains within it numerous nematocysts and concre-
tions Avhich were scattered throughout the ectoderm at the
margin of the umbrella. After the tentacle has grown out
for a little distance beyond the bell-margin the cells on
the upper or aboral surface become modified to form an ad-
hesive organ. (Fig. 27, 2S) . The cells over a disk-shaped
area become elongated until they have t>ie form of a thick
pad. The tissue immediately around the pad grows out in a
flange so that the organ becomes a vacuum-cup strongly mus-
cular around the edge (Fig. 28) .

After the tentacle has grown out to a length of six
to eight mm. and has increased in diameter considerably,
the cavity of the circular canal is drawn into it. The en-
dodermal cells, arranged radially about tlie axis, thicken
until they are forced away from the centre, and a tubular
cavity is left(Fig. 2). As this process takes place first

57

at the proximal end of the tentacle, within the tissue of
the bell-margin (Pig. ;^4) the cavity of the circular canal is
carried out along the axis of the tentacle towards the tip.
In this way the tentacle, which was originally imperforate
as in the lax-val condition, has becone hollow,

r. The Sense-Organs. - The origin of the sense-organs is
very similar to that of the tentacles (Figs. 27 ^: 30), In fac"
it seems clear from a study of these processes in Gonionema
that the sense-organs must necessarily be regarded as mod-
ified tentacles. In the case of these sensory clubs (S.*^.)
the endodermal tissue of the circular canal (End. ) grows
down in a plug into the ectodermal tissue of the bell-mar-
gin ( Ect . ) This latter becomes closely applied to the
outside of the plug, as a thin investigating epithelium,
and it also spi'eads out in a thin lamella over the inner
surface of the capsule which appears in the ectoderm in
front of the developing club. Figures 34 and 55 are draw-
ings by Professor Brooks from sections cut transversely
across the bell-margin, showing the early stage in the for-
mation of a sense-organ. I have not been able to demon-
strate the presence of sensory hairs in tlio cavity of the
capsule. The cells at the tip of the club soon begin to

58

secrete the solid concretion which later attains a consid-
erable size. The concretion is invested v/ith the thin mem-
branous ectodermal covering. In Gonionema the concretions
correspond with the composition which has been given for
similar structures in other medusae - a lime carbonate de-
posit in an organic matrix. Thus it is seen that both ten-
tacle and sense-organ consist of an endodermal core which
appears first and grows out from the lining of the circular
canal. In each case this core becomes invested with a tu-
nic of ectoderm which remains associated with it.

XIII. NEHATOCYSTS. -

In the hydra stage the earliest appearance
of nematocysts was as interstitial cells arising from eith-
er tissue-layer. Their growth in Gonionema is much as it
is in Cordylophora lacustris . as described by Morgenstern
(1901), They are carried out on the tentacles by migration
along with the ectodermal layer in which they arr> set. The
extreme attenuation of the tentacle as it is fully extended
(Fig. 13) gives an admirable chance to study the construc-
tion of the cell-elements, especially the nematocysts. The
tentacle appears as a delicate rod of translucent substance

59

partitioned off at intervals by the transverse v/alls of the
endoderm cells, and studded alonj; its len,";th with numerous
p;listening bead-like bodies, the nematocysts. Above each
of tiiese thread-cells a palpocil projects like a thoi'n
(Fig. 31). The capsule has a slip;htly unusual form, long
and bean-shaped (Fig, 32), Examination with a high power
objective, focussed down into the water upon the extended
tentacle shows with considerable distinctness a ganglion
cell of glistening highly refractive appearance, lying
close to each nematocyst (Fig, 32, g.c). In every case
this ganglion cell is situated distal to the thread capsule
toward tJie free end of the tentacle (Fig, 32 ) . A thin
strand of nervous tissue runs in each direction from the
ganglion cell, towards the nematocyst pi'oxiraally , towards
the free tip of the tentacle distally. It is visible for
only a short distance, ?iowever, soon vanishing into the ec-
todermal tissue, and none of its branches or terminations
are to be followed. It evidently innei'vates the nettling
capsule, near the base of which it can be seen.

In the gonosome the nematocysts are carried out onto
the ectoderm of the growing tentacle in situ, as in the
larva. Further growth in the extent of ectoderm is brought

60

about in two '.vays : by multiplication of the cells already
incorpoi-ated in the epithelium cf the tentacle, and by im-
migration of cells from the thick ectodermal pad at the
base. The cartilaginous tissi^e coia])osing this pad is pecu-
liar in character. The cell walls are aLnost or quite
obliterated, and the gelatinous substance contains the con-
cretions already mentioned, v/hich give tho tissue something
of the nature of a cartilaginous encasement for the nerve
ring lying immediately beneath it. These concretions are
not very different in appearance from the concretions of
the sense-organs - the "otoliths" as they are called, with,
however, quite inadequate evidence that they function as
auditory organs. They are solid masses of hard substance,
partly calcareous and partly organic. They are not made up
of concentric lamellae such as give the concretion in the
sensory capsule its onion-like layers. In this whole group
of mediisae the older tentacles are left stranded as it were
by the growth of the margin of the umbrella beyond their
point of origin. As they are in this way carried up onto
the exumbral surface the cartilaginous tissue grows so as
to fill the space between the base of the tentacle and the
bell-margin, fox-ming a round cushion oi" hard tissue. In

61

sections cut through this tentacle pad (Figs, 33 and 34) it
is seen that the concretions which lie towards the boll-
margin are more dense and homogeneous; that further inwards
they are somewhat less solid in appearance, spaces appear-
ing within their outer walls; and that at the side nearest
the circular canal there are great mmibers of nettling
cells in various stages of formation. All gradations are
present betvfeen the solid concretion and the nettling cell
(Fig. 33), It is therefore evident that the two distinct
kinds of specialized cell products, the one for the protec-
tion of the delicate nerve ring and the other for capturing
and paralizing prey, ai"e produced from the same cells and
in the same locality ; ^hat they are, in fact, homologous.
Fig\^re 33 was drawn by Professor Brooks to show this fact
in Gonionema. Whether the cartilaginous colls actually
change into nettling cells after they have become fully de-
veloped is not clear. At the inner margin of the cartilag-
inous pad the nematocysts lie closely packed together (Fig,
33). From this breeding place the nematocysts work their
way out onto the tentacle along which they migrate until
they reach a spot where t?iey are needed. In young tenta-
cles, which are still elongating, the nematocysts are car-

62

ried out with the ectoderm as it becomes applied to the
tentacle-base. But al'ter a certain time the tentacle in-
creases only very slowly in length and additional nettling
cells are needed to keep up with the inci'ease in diameter.
This migration of nematocysts has been seen and described
by Hurbach in his paper on the nettling organs in Hydroids
(Arch. f. Naturg. 1894 ). After the capsules have become
established, the ectodermal covering becomes modified to
form the cnidocil (Fig. 32, en.). The nerve connection is
developed at an early stage.

XIV. SEXUAL ORGANS. -

In minute specimens of the adult gonosome the
gonads t.re frequently found in their first stage of devel-
opment. They appear as outgrowths of the ectodermal cover-
ing of the radial canals, at first in the for-m of rounded
lumps projecting downwards from the radial canal into the
subumbrella at a point two-fifths of the distance from the
top of the bell to the margin. The rudimentary lump of go-
nadial tissue elongates in both directions from the point
at which it started. (Text-figure B shows the condition in
Olindias , where it is identical at first with that in Go-

nionema ) , The Ronad thus becomes an elongated ridge of
tissue which finally reaches to the extremities of the
chimiferous tubes, and increases in depth until it hangs
down into the subumbrella as a ribbon. Early in its devel-
opment the ribbon is somewhat sinuous, and as the medusa
attains greater diameter the convolutions become more ne-
merous and farther extended on either side of the radial
canal, until ultimately the folds are packed tightly to-
gether in a solid band of tissue, which at the time of ma-
turity is extended with sexual elements. The process of
formation of the sexual organs is identical in the two sex-
es; it is impossible to tell whether a given individual is
male or female until the sexual products begin to mature.

64

1. Observations on the development of Gonionema indi-
cate that Haeckel's sharp distinctions between jelly-
fishes which he groups in his orders "Trachomedusae"
and "Leptomedusae" are not justified.

2. Dehiscence occurs in Gonionema with precise perio-
dicity, and is definitely affected by changes in
light.

3. Segmentation is total and equal; endoderm is formed
by delamination of the blastomeres; a solid morula
results.

4. A planula stage occurs, and later a hydra stage, in
which the polyp develops first two tentacles, then
a second pair.

5. Youngest medusae and oldest polyps show marked ho-
mologies; direct metamorphosis is suggested.

6. Peculiar pathological phenomena appear, the larva
living for weeks in the form of a Plasmodium, with
amoebiform activities.

65

7, Alternation of generations occurs, A non-sexual
form of multiplication appears durinr; larval life;
buds are produced which are detached as planulae and
go through the same changes as the parent.

8, The order and arrangement of tentacles in the go-
nosome follows a definite plan in cyclic sequence,
producing a figure which is cyclically (not bilater-
ally,: symmetrical. Tentacles and sense-organs ap-
pear at determinate points on the bell-margin,

9, Histogenesis of tentacles and sense-organs sho'.vs
their homology,

10, The origin of nematocysts from the ectodermal pad

at the base of the tentacle, and tlieir homology with
cartilaginous concretions, are establis?ied,

11. Cronads arise as enlargements by proliferation of the
ectodermal subumbral epithelium of the radial canal.

66

DESCRIPTION OF FIGURES.
Plate I,

Fig. 1. Adult Gonionema in resting attitude; floating after
a period of active sv/iroming. 2/l.

" 2. f'edusa in act of swimrning; bell contracted, tenta-
cles drawn up at the end of a forward impulse. Pho-
tograph from life. l/l.

3. One radial canal from ripe male, shovfing gonad,
c. circular canal; r, radial canal. 8/1.

" 4. Gonad of female, during dehiscence. 20/l.
" 5, Egg during first segmentation; cleavage fui-row half
completed. n, nuclei. 1370/l,

" 6. Egg during second segmentation, left hemisphere

completely divided, right hemisphere in process of
dividing.

" 7. Hollow blastula 7 hours after fertilization. Opti-
cal section of live egg.

67

Fig. a. Two-l'Hyered blastula, endodenn having arisen by de-
lamination.

Plate II.

Fig. 9, Young planula larva. P, posterior end, a, anterior
end. 675/1.
" 10. Planula larva; posterior end enlarged; endodermal
cells at posterior end arranged along the axis of
the larva, marking line of future coelenteron.

" 11. Two-tentacled polyp, in section; four weeks old.
Kc , thickened basal ectoderm.

" 12. Polyp, four months old, with five tentacles and
five oral lobes, lying in the same planes.

" 13. Polyp, five months old; in typical resting atti-
tude, tentacles expanded 2 mm. P.T., prehensile
tip of tentacles, adhering to bottom.

" 14. Five-tentacled polyp, showing form of coelenteron
and formation of bud.

68

Plate III.

'ig. 15. Five months old polyp, with bud just foi'tning.
" 16. ThG same, bud 8 hours old. "Ret, ectoderm.
" 17. " " " one day old, pear-shaped
" 18 " " " three days old. Endoderm isolated
from parent by constriction of ectoderm-
19. The same bud, four days old, ready to be detached;
showing interradial position on parent

" 20, Another individual, bud in process of detachment,
showing elongated ectadermal isthmus,

" 21, The same, lb -ainutes later. Bud settling on former

distal end,
" 22. Detached larva, after three days of free- swimming

planula life, just attached. Showing nematocysts,

" 23, The same, four days later; basal ectoderm much
thickened, coelenteron developed,

" 24, Larva 23 days old, exhibiting transverse fission of
coelenteron and elongated hydrocaulus,

" 25. Young medusa with 12 tentacles and 4 sense-organs;

69

showing spherical shape and constricted bell-margin

Plate IV.

Fig, 26. 32-tentacled medusa with 14 sense-organs. 7th and
Sth tentacles have appeared in each quadrant ex-
cept quadrant A, where 8th is lacking. 4 sense-
organs have appeared in quadrants E and D , 5 in
ijuadrants A and C,

Fig, 27. Horizontal section of bell-margin at level of ru-
diment of tentacle, T.R. Con. calcarious concre-
tions; N.R. nerve ring; T tentacle; C.C. circular
canal.

Fig, 28, Tentacle tip of medusa, showing rings of nemato-
cysts, angle of tentacle, and adhesive organ on
aboral side.

Fig, 29, Cross-section of adhesive organ, G.C. gland cells
composing cement gland; M.F. muscular flange;
500/1 drawn by W. K, Brooks.

Plate V,
Fig, 30, Radial transverse section of bell, at point of or-

70

rigin of sonse-organ, S.C., showing endodex-mal
origin; Caps, sensory capsule, surrounded, by ecto-
derm; V veliim.

Fig, 31. Tentacle tip of larva from above. End, single
chain of endodermal cells; On, cnidocil; G.C,
ganglion cell. 500/l.

Fig. '62, Nematocyst in detail, shoviing cnidocil, Cn. , gan-
glion cell, Ci.C, nerve fibres, N.F. running to-
wards distal end. 2000/1,

Fig. 33, Ectodermal pad at base of tentacle. Three areas

of cell secretion, C.P. cartilaginous pad, T, ten-
tacle, M. mesogloea; Radial vortical section.

Pig, 34, Transverse section, at bell margin, of base of
tentacle, showing tentacle pad, C,P,

71

VITA.

The writer, Plenry Farnham Perkins, v;as born ^'lay 10,
1S77, in Burlington, Vermont. He is the son of George Hen-
ry Perkins, Ph.D., Yale, 1867, Howard professor of Natural
History at the University of Vermont; and of Mary Harnham
Perkins, of Galesburg, Illinois, A.B., Knox College, 1869.

The degree of A.B. was received from the University
of Vermont in 1898, at the completion of the regular four
years' course in classics, literature, mathematics and the
sciences. During the four years since October, 1898, the
writer has been enrolled as a graduate student in the Johns
}Iopkins University, in the department of Biology, Phys-
iology and Physics were taken as subordinate subjects. The
summers of 1899, 1900, and 1901 were spent in research at
the sea-coast of tliis country and the Bahama Islands. Two
papers have been published giving in abstract some results
obtained while investigating the development of the Hydro-
medusa CTonioneina I'urbachii, the subject of this disserta-
tion. A third paper appeared in the University Circulars
on a jelly-fish found at Nassau, N.P.Id., not previously
known on this coast.

ga*

-v-^: ^-^ "* ^''^^^js.iv^'

*^yy\m

^jjs^V'V ,^^"^ ^^^u.;oy'

iU'^v-'Wv.'V

lywv,. „ iLi'v, „

/^ww^^^Si^?^

^,vJ»'W^"'i^i

Ww«i»':x«^

^^mmrn^

Kf. »

'\,:^^^-

/VVW-jwO^''-

^U^Wi^^W^^^^-"^

^y/V'W^

:^!^^i-^!^

^^@S1S«^

MJM^^

:y;^V/^

.N*^>"^

/vwywW**^'!!^^!^^

mm^^

-^^i..^^^m

^-.^v.^^^^^.

UjMMfi&^'^.Zi^.'^tm-aa^

"^Si«S^^;^£i^^

mm

m^mms^m^

ivw*

*^aii««

?i?^W^/WU^WV^*'^K

/^wwv^^w-'v^"rw&^.

v^^^^w^^w^^^igi

w#^

:^

^"'WiSV*^

'vgw'guMUj^^y

iJ^''-' :uw^'

,-wi

^^^i4;:^^ftft^'^^^'