103




HARVARD UNIVERSITY.

LIBRARY
MUSEUM OP COMPARATIVE ZOOLOGY.
GIFT OF



Proceedings of the Boston Society of Natural History.
Vol. 31, No. t,
p. U1-X)'.), p|. 4-10.
THE METAMOBPHOSES OP THE HERMIT CRAB.
By Millktt T. Thompson.
IiOSTON:
PRINTED for toe SOCIBTT.
September, 1903.



k 4


No. 4.—THE METAMORPHOSES OF THE HERMIT
CRAB.1
BY MILLETT T. THOMPSON.
Introduction.
Tim group of the Pagurids or hermit crabs has always attracted
the attention of carcinologists, not only because of its extent both in
species and in individuals, but also because of the asymmetry which
involves nearly every genus and the habit of protecting the soft
abdomen within hollow objects, typically the shells of Gastropod
Mollusca. Whether the asymmetry — which is of dextral type
except in the genus Paguropsis — owes its origin to this use of
shells for residences, since nearly all marine Gastropod shells are
dextral in coil, cannot be determined with certainty until the phy-
togeny and relationships of the various genera are better understood.
But, nevertheless, it is unquestionable that the modifications found
are in a very great degree correlated with residence in dextrally
spiral shells. This alone makes the ontogeny of the group an
.extremely interesting and important subject for study.
Knowledge of Pagurid ontogeny was in its beginnings practically
contemporaneous with the discovery of the metamorphosis among
the higher Crustacea. During the discussion which followed
Thompson's assertion ('28) that the supposed genus "Zoea" was
a larva, Vigors ('30) in a review of Rathke's study of the develop-
ment of Aatucus fluviatilis, which has no metamorphosis, appeared
skeptical with regard to Thompson's conclusions. This drew a
reply from the latter author (Thompson, '30-'31), communicating a
list of fifteen Decapods in which he had observed the young to be
unlike the parent and "Pagurus" was included in this list.
The further statement by Thompson five years later, in 1835, that
both "Zoe" and "Megalope" were larvae and his reiteration that,
among "Macrura," Attacua marinus, Pagurus, and other forms
underwent a metamorphosis, greatly stimulated research on erusta-
1 From the Anatomical laboratory, Brown university, and the laboratory of
the U. S. fish commission, Woods Hole, Mass.


148 PROCEEDINGS: BOSTON SOCIETY NATURAL HISTORY.
cean development, and in the year 1840, two papers on Pagurid
development appeared. One of these (Rathke, '40) deserves to stand
as the first real contribution to our knowledge of the metamorphosis
in this group. For Thompson did not describe his larva and
Philippi's paper of the same year merely figured in a rough way
the first zoea. Rathke's paper, however, briefly described three
zoeal and a Macruran-like older stage of the European "Pagurus
bernhardus." Two years later (Rathke, '42) it was republished in
more complete form with excellent figures of the zoeae and of the
maxillipeds, tail fan, and pleopods of the "older larva."
Then twenty years passed without any important addition along
this line of carcinological research until the publication of Muller's
"Fur Darwin," in 1864. This article described a "Pagurus" zoea,
called attention to the absence of any gradation in the successive
zoeal stages towards the Macruran-like stage, and, although it is
uncertain whether Muller actually saw this later stage, it is described
and compared to a small shrimp which had received from Milne-
Edwards the name Glaucothoe. "Glaucothoe peronii may be such
a young, still symmetrical Pagurus." The recapitulatory nature of
the stage is also asserted. "The abdomen is truly in the adult a
clumsy [ungeschlachter] sac, filled with liver and sexual organs, but
it is yet fairly powerful [kriiftig] in the glaucothoe-stage and it was
also still stronger when this stage was the permanent form of the
animal."
A few years after Muller's article appeared, Spence Bate (?68)
published an account of two zoeae and a glaucothoe which he col-
lected at the surface and correctly assigned to "Pagurus." Of the
latter stage he says: "In this they probably continue until they find
a suitable molluscous shell. ... I imagine that they may cast their
exuvium and grow during the whole time that they are deficient of
such a shell because I have taken specimens, occupants of shells,
that were still smaller than the ones described and yet further
advanced toward maturity. It would be curious to see if, were they
deprived entirely of a shell as a habitat they would continue to grow
and retain the normal [i. e., symmetrical] form of the pleon gener-
ally."
This query as to the effect of depriving the young of a shell was
in part answered by Agassiz ('75) nearly ten years later. He
showed that the larva might attain asymmetry and a soft abdomen


THOMPSON: METAMORPHOSES OF HERMIT CRAB. 149
before a shell was entered and considered that the desire tor a shell
arose from the anatomical changes. The extreme brevity of his
record, however, has been a cause of much confusion. It was not
clearly shown whether this metamorphosis without a shell was the
typical developmental sequence or only a frequent happening.
Moreover, the amount of the asymmetry attained at this time was
not certainly defined. Hence the record has been interpreted to
mean that the change from the glaucothoe to the adult form was
gradual and covered several ecdyses (Bouvier, '91). Hut I am
convinced that Agassiz intended to indicate the change as occurring
with a single ecdysis and the figures published by Faxon ('82) sup-
port this contention. These show four zoeae, and in less detail a
fifth stage which is identified with the genera Glaucothoe and
Prophylax. Immediately following the fifth stage comes a stage
entitled "age when it takes up its abode in a molluscan shell," and
this figure depicts a crab with almost adult asymmetry, but bearing
minute rudiments of pleopods on the right side of the abdomen.
The main outlines of Pagurid ontogeny had now become clear.
The young passed through several zoea stages and then moulted to
a glaucothoe phase analogous to the megalops of the true crabs, and
as with them the mysis phase was suppressed (Claus, '76) into the
last of the zoea stages, the metazoea (Claus, '85). The anatomy of
the glaucothoe, however, remained almost unknown and the details
of the metamorphosis to the adult form obscure. For as already
noted, considerable uncertainty existed as to whether the larva
passed to the adolescent phase directly or gradually. Since
Agassiz's note, no work has given us much additional data regard-
ing this most important point in the whole development. For the
recent articles by Sana ('89) and Czerniavsky ('84) do not deal with
the postzoeal stage except to depict the external form, although
they are valuable records for several genera and species of hermit
crabs.
The present research, then, was undertaken in the hope that
through a study of the anatomy from the time of hatching until the
adult form was attained, and through examination of the rdie of the
shell in the development, some knowledge of this almost untouched
field might be obtained. The study was carried on during the three
years prior to 1902 at Brown university and at the laboratory of
the United States fish commission at Woods Hole. I wish to


150 PROCEEDINGS: BOSTON SOCIETY NATURAL HISTORY.
express ray especial gratitude to Dr. H. C. Bumpus, now director of
the American museum of natural history, for the assistance and
guidance furnished me throughout the work at both institutions. I
am also indebted to Dr. A. D. Mead, of Brown university, and to
Dr. H. M. Smith, of the United States fish commission, for ample
provision extended to me during my study.
The Adult Crab.
Throughout the region about Woods Hole, the genus Eupagurus
is practically supreme and is represented there in the shallower
waters by four species, viz.: longicarpus, annulipes, acadianns,
which is apparently only a variety of the European bern/tardus,
and pollicaris. Of these, E. longicarpus is the only one generally
distributed along the shore. It is extremely abundant and extends
from tide-water to a depth of about twenty-five fathoms and is
associated over the lower limits of its range with acadianns which
is a deeper-water species, and in the shallow waters with pollicaris
and annulipes, but of these only pollicaris occurs along shore and
then only in a few localities.
Because of its distribution and abundance, E. longicarpus was
selected for the present research. But after much of the work had
been completed, it was discovered that the larvae of annulipes could
not be distinguished from those of the selected species, as their
slightly smaller size furnished no adequate criterion for their separa-
tion. Whenever the annulipes larvae were unusually abundant,
as was shown by the occurrence of their adolescent stages in the
experiments,-1 could note no difference in the ontogeny from the
periods when undoubted longicarpus predominated. In sections
also, the smaller specimens of any stage, presumptive annulipes,
were wholly like their larger companions. This is perfectly in
accord with what we know of the adults themselves. Eupagurus
longicarpus and E. annulipes differ only in specific details. E.
longicarpus is. the larger, has slender chelipeds and its coloration is
diffuse. E. annulipes, on the other hand, has stout chelipeds and
prominent belts of brown pigment on the anterior thoracic limbs.
My research is therefore not invalidated, but rather enriched by the
confusion. It becomes a life history of two, instead of one species.


THOMPSON: METAMOKPHOSES OF HERMIT CRAB. 1")1
The majority of the larvae handled belonged to E. longicarpua and
it is fitting therefore that this species should be described rather than
E. unnidipes. But the following account is applicable to the latter
and to the other species of the genus also.
Like many other Decapods, Eupagurus longicarpua is crepuscular
and during the day a majority of individuals remain buried in the
sand or congregated in the shade. They are omnivorous and must
glean very closely, as they pick up bits of gravel or detritus from
the bottom and brush them over between the maxillipeds. They
also toss sand — usually with the smaller cheliped only — to the
mouth parts, brush it between them and let the grains fall again in
a continuous stream. Probably it is in this manner that they
obtain the diatoms and foraminifera, which- are found in the stomach
and intestine. But although the food is thus very largely diatoma-
ceous, no sort of vegetable or animal matter is refused.
The asymmetry is dextral. The chelipeds and the two following
pairs of limbs are larger on the right than on the left side. The
abdomen is spirally twisted to the right and only the sixth segment,
telson, and first segment are heavily calcified. The remainder and
major portion is soft and bloated. The boundaries of the segments
are lost, except dorsally where the posterior borders of the terga
can be traced, and show that the fourth and fifth segments are elon-
gate, while the second and third are shortened. The sternum of
the first segment or peduncle is indistinguishably fused with the
sternum of the last thoracic segment. The peduncle bears no
appendages and the four succeeding segments have no pleopod on
the right side. On the left side, the female has a large biramous
pleopod on the second, third, and fourth segments and a smaller
pleopod with a minute internal ramus on the fifth (pl. 6, fig. 24);
while the male has a pleopod similar in shape to the most posterior
of the female's series on the third, fourth, and fifth segments, but
no pleopod on the second segment (pl. 6, fig. 25). The uropods
are alike in the two sexes and the left is the larger (pl. 7, fig. 31„).
The greater development of the female's pleopods is probably
correlated with their use during the breeding season when the eggs
are borne on the hairs of their borders. It must not be hastily
assumed, however, that the pleopods of the male have no function.
While watching adolescent crabs that were inhabiting straight glass
shells I have noticed that their pleopods were at times waved in the


152 PROCEEDINGS: BOSTON SOCIETY NATURAL HISTORY.
water. This suggests a possible function for these appendages in
both sexes to reinforce the current in the shell which is primarily
induced by the branchial outflow and movements of the body. Bate
('50) records a similar movement of the appendages in female hermit
crabs which were bearing eggs. The rdle played by the sensory hairs
on these appendages must be nearly equivalent in both sexes.
A large columella prominence is present. Although it has been
suggested that this organ aids the crab in maintaining a firm hold
on its shell, I feel that this cannot be a complete explanation for the
existence of the structure. If it serves this function at all, it must
be mainly passively, in conforming the body more perfectly to the
columella angle of the shell chamber. Moreover, it is not a very
muscular organ even in the genera in which it is well developed and
it is only imperfectly developed in a large number of hermit crabs.
Our species of Eupagurus maintain their hold on the shell chiefly by
the grip of the calcified telson and uropods on the columella,
while the tuberculated areas on the posterior thoracic limbs and
uropods may lend assistance. The chitin over the venter of
the abdomen also is roughened with fine transverse lines. When
an attempt is made to dislodge a crab by traction on its limbs, an
additional resistance is often obtained by the elevation of the rostral
region of the carapace against the roof of the chamber.
The breeding period of E. longicarpus is very long. Females
with eggs attached to the pleopods can be obtained from May until
mid-September. The zoeae begin to appear in the auflrieb in the
latter part of June and are very abundant during July and August.
The glaucothoe can be obtained as late as October. Our knowl-
edge of the breeding period of the remaining species of Eupagurus
is imperfect. E. annuHpes breeds over practically the same period
as lonf/icarpus, ceasing a little earlier in the fall. Zoeae that agree
closely with Sara' ('89) description of the zoeae of E. bernhardus
occur seatteringly throughout the summer. Quite likely these are
the larvae of acadianus. The glaucothoe, however, has eight instead
of the ten telson bristles shown in Sars' plate. The young of
pollicaris have not been identified and most of them are probably
liberated in June, though females with eggs are taken during July.


THOMPSON: METAMOIU'HOSES OF HEHMIT CRAB. 153
General Accocxt of the Metamorphosis.
The eggs and young of Eupagurus are very sensitive. Zoeae
were hatched in confinement with great difficulty and as they invari-
ably died in a few hours it was necessary to collect all material
directly from the auftrieb. Even zoeae collected in this way could
not endure more than one moult and although the later phases were
much more resistant, they were nevertheless delicate as compared
with the young of many other Decapods.
The most vital factors in rearing crab larvae seem to be cleanli-
ness and an even, moderate, water temperature. The following
method for rearing the young, although not original, is recorded
here because it was by far the most satisfactory of those tried. In
the end it was exclusively used. The young were kept in covered
dishes of clean sea-water which was in sufficient volume to render
unnecessary the use of algae for aeration. The water was changed
daily, or in hot weather oftener, and to ensure a constant tempera-
ture the dishes were partly immersed in running water or suspended
in large aquaria. Diatoms, collected with a fine net or scraped from
submerged objects, were the most satisfactory food, but animal food
was also given.
Woods Hole offers especial advantages for the study of crusta-
cean development. Strong currents prevail in the neighboring
waters, and one of these rushes past the wharves of the Fish Com-
mission station during part of each tide and a large "tow net" may
simply be suspended in this current and emptied at convenient inter-
vals. Also, at those times when there is a paucity of animal life in
the water, the numerous "slicks" caused by conflicting currents
and back-sets, may prove an excellent resource, as the plankton is
concentrated in these areas.
Zoea phaae.— The zoeae of lonijicarpu s and annulipes (pl. 4. fig.
1-4) have the characteristic Pagurid form: without carinae or
spines; with long, straight rostrum; hind angles of carapace pro-
duced; swollen compound eyes, and third maxilliped rudimentary at
time of hatching. The transparent body is pigmented with con-
tractile scarlet and yellow chromatophores, and the eyes are black
with yellow pigment diffused over the surface. No trace of a
median eye can be found. The stomach and intestine are usually
clean, as but little foixl is taken during the phase. The livers con-


154: PROCEEDINGS: BOSTON SOCIETY NATURAL HISTORY.
tain strongly refractive, yellowish globules. These larvae are pho-
totactie, though to a less degree than Brachyuran zoeae, and in a
glass vessel range themselves head downwards against the lighted
side. They swim tail foremost, and rather slowly and steadily.
The surfaces of the body, however, are not definitely oriented,
though either the dorsal or ventral are uppermost more frequently
than the lateral. This lack of orientation seems to be correlated
with the complete absence of otocysts in the three earlier stages and
with the undeveloped condition of these organs in the fourth stage
or metazoea. While swimming, the eyes, antennae, and uropods
are invariably held in the positions shown in figure 4 (plate 4).
In captivity, the moults between the successive stages usually
take place at night or in the early morning, and the larvae sink to
the bottom of the aquarium and remain quiescent for a considerable
period previons to the actual- ecdysis. The manner of moulting
under natural conditions could not be determined.
The length of the zoea phase and of its several stages is not
known. The occurrence of the successive stages in the auftrieb
indicates that the phase possibly extends through only two or three
weeks. But little reliance, however, can be placed upon data of
this class, and it is very probable that the period varies with the
external conditions, as is the case with other crustacean larvae.
In our Eupagurids, the zoea phase comprises four stages, which
may be separated as follows: —
First Zoea.
Third maxilliped rudimentary (pl. 4, fig. ]); sixth abdominal
segment not distinct from telson; exopods of maxillipeds
with four feathered setae; thoracic limbs as a simple mass of
undifferentiated tissue. Length, 1.9-2.7 mm.1
Second Zoea.
Exopod of third maxilliped functional as swimming-foot;
endopod barely indicated (pl. G, fig. 142); exopods of maxil-
lipeds with six setae; anlagen of uropods as a baud of tissue
along each side of the telson (pl. 7, fig. 312); thoracic limbs
distinct, rudimentary, fourth pair excessively long, fifth pair
short and concealed beneath others (these proportions are
found also in the two following stages). Length, 2.7-3 mm.1
1 Including as they do the larvae of two species, all the measurements have
an unduly extended range.


THOMPSON: METAMORPHOSES OF HERMIT CRAB. 155
Third Zoea.
Uropods present, and sixth abdominal segment distinct;
exopods with seven setae; rudiments of the gills on second,
third, and fourth limbs. Length, 3.5-4 mm.1
Fourth Zoea, Metazoea.
Limb rudiments very large, dactylus of cheliped distinct;
rudimentary pleopods present: gill rudiments for chelipeds
present; right cheliped obviously larger than left—this
condition of the chelipeds may possibly date from the third
zoea; maxillipedal exopods with eight setae; uropods sym-
metrical ; mandibles without palpus rudiment. Length, circa
4 mm.1
Faxon ('82) described four zoea stages identical with these four,
and it is likely that he studied larvae belonging either to Eupagurus
longicarpus or to annulipes. Other investigators have not recorded
so many stages. Xone of them describes the second zoea, unless one
of the figures, no. 52, in Claus's "Zur Kenntniss der Kreislaufs-
organe" ('84), corresponds to it. But it should not be assumed
that this stage is typically absent from the zoea phase of the various
Fagurids, for it closely resembles the first stage and might readily
be overlooked. My third zoea corresponds to Kathke's ('40, '42)
"young of one and three-fourths lines" and to Clauses ("61)
"spiiteres Stadium." The other writers make no mention of a
similar stage. The metazoea has been repeatedly noted, l'athke
('40, '42) termed it, "young over two lines long"; Bate ('68), " what
we take to be the second stage"; Clans, "Mysisstadinm" ('76),
"Metazoea" ('85); and Sars ('89), "last larval stage," "last stage
before moult to adolescent stage."
Postzoeal phase.— The postzoeal phase includes only one stage,
the fifth or glaucothoe (pl. 4, fig. 5). The larva is now 2.8-3.3
mm. long and is Macruran in form with symmetrical abdomen and
with pleopods on the second to the fifth segment. The uropods
are asymmetrical as in the adult, and the thoracic appendages are on
the whole adult in type and proportions. A few tubercles on the
posterior thoracic limbs and on the uropods (pl. 6, figs. 19, '20; pl. 7,
fig. 31t) represent the future tuberculated areas. The sternum of
1 Including as they do the larvae of two species, all the measurements have
an unduly extended range.


156 PROCEEDINGS: BOSTON SOCIETY NATURAL HISTORY.
the first abdominal segment, which is not very wide at any time, can
no longer be detected in sections. The otoeysts are functional and
orientation is definite. Internally the livers, lateral caeca, green
glands, and sexual glands are thoracic. The unpaired intestinal
caecum is lacking and the muscles and blood vessels are similar to
those of Macrurous Decapoda.
In color, the glaucothoe resemble zoeae except that the branches
of the yellow chroinatophores form a fine network over the limbs
and carapace and produce a grayish effect. As before, the stomach
and intestine are usually empty and transparent and the livers con-
tain refractive globules.
Glaucothoe are found at the surface, either swimming or clinging
to floating seaweed and seem to be more abundant at night. When
swimming, the dorsal surface is uppermost, the abdomen is extended
while the limbs are either extended or hang stiffly down. Though
phototaxis is still present, at times individuals cease swimming and
crawl about on the bottom of the aquarium. These examine the
objects in their path, and if they find a shell or other hollow object,
may enter it and abandon the free-swimming life. Quite frequently,
however, after a brief exploration the larva will recommence swim-
ming. More rarely, a glaucothoe that has already entered a house
will abandon it. As the sixth stage approaches, the desire for a
covering for the body becomes stronger with the alterations in struc-
ture until it is almost impossible to keep the larvae "naked." They
use all available objects or ensconse themselves in crevices. A house
may be taken at any time during the phase, but a short period of
free-swimming life is typical.
Not infrequently a glaucothoe will remove bits of rubbish from a
shell, but I was not able to confirm Agassiz's observations ('75)
where his young tore out and ate dead snails and then used the
shells. Glaucothoe, according to my observations, take but little
food. If, however, his reference is to adolescent crabs, they are
quite voracious and might readily eat a dead snail, though I have
observed nothing of the sort. It is, however, scarcely necessary to
assume or suggest a causal sequence between this act and the use
of the shell for a dwelling.
The glaucothoe stage as a rule lasts only four or five days, but
during this time a profound modification of structure takes place.
The livers, sexual glands, and green glands pass into the abdomen,


THOMPSON: METAMORPHOSES OF HERMIT CRAB. 157
the circulatory system is modified, and the muscles and pleopods
degenerate, so that before the moult to the sixth stage closes the
period, the anatomy has become adult in plan. This metamorphosis
is not dependent on the presence of a body covering, but completes
itself perfectly in larvae which are prevented from obtaining a shell.
The whole animal also becomes less transparent, the chelipeds
become white, and brown bands appear on the posterior pereiopods,
a coloration recalling that found in the adult Eupagurus annidipes.
The moult to the sixth stage is preceded by a brief period of
quiescence, and in the few cases observed, the actual eedysis was
rapid. Either the integument of the thorax is sloughed first and
that of the abdomen later, or the entire integument is sloughed in
one piece. The latter method seems to be the rule for the zoeal
ecdyses, while the former is more frequently the rule in the adult,
and the abdominal exuvium is usually badly torn in the process.
The glaucothoe stage is the "noch altera .Tungern" of Rathke
('40, '42); the "Glaucothoestufe" of Mailer ('64); the "third stage"
of Bate ('68); "stage when it seeks a shell" of Agassiz ('75), and
the "first postlarval," "first adolescent" of Sars ('89). It is not
mentioned by other workers with the exception of Czerniavsky ('84).
Adolescent phase.— "Adolescent phase" is more a convenient
term under which to discuss the development during the earlier
postlarval life, before the adult anatomy is fully attained, than a
definitely limited period. There is also little to be gained by an
attempt to separate the numerous stages which may be included
within it. The crab has the adult structural plan before the close
of the glaucothoe period, but all the organs must still undergo devel-
opment to realize fully the adult structure. The length of this
process varies widely in different parts.
Sixth-stage larvae are of the same size as glaucothoe, and the
specific adult form may be attained before much growth occurs.
After about forty days have passed, the young reach a length of
from five to eight millimeters. The sixth stage lasts from six to
twelve days, but the later moults are irregular and crabs of the
same age may be very unlike in size and development. The
manner of life in all the adolescent stages is that of the adult, and
food is taken abundantly almost immediately after the moult from
the glaucothoe.
The sixth stage retains the annidipes color, handed to it by the


158 PROCEEDINGS: BOSTON SOCIETY NATURAL HISTORY.
glaucothoe (pl. 4, fig. G). With the seventh stage, however, the
longicarpua young become separable from annidipes larvae, as in
them the brown bands are lost. The full adult colors and the com-
plete specific form are not attained until about the twentieth day
from the glaucothoe. The sixth stage also has a nine-jointed
antennal flagellum and certain retrogressive alterations in the
mouth parts, which are legacies from the glaucothoe (pl. 5, figs. 10,
12, 13). The elongation of the flagellum is very gradual, the
nine-jointed condition often persisting till the eighth stage, while
crabs forty days from the glaucothoe have only 17 or 27 joints.
The metamorphosis of the pleopods during this period is of special
interest. At no time are there any traces of appendages on the
peduncle, which is interesting when we remember that some Pagu-
rids, as for example, Sympagurus and Paguristes, have pleopods on
this segment in the adult. Typically, the sixth stage has no append-
ages on the right side of the abdomen, except the uropod, but on the
left the pleopods are well developed on segment three to five and are
of the type found in the glaucothoe, i. e., they resemble those of the
adult male. On the second segment the pleopod is reduced to a
mere rudiment (pl. 4, fig. 6). About nineteen percent of reared
sixth-stage larvae, however, retain rudiments of one or more of the
right hand pleopods. The typical reduction may be expressed in a
formula, by use of R for an appendage which is retained intact, Ku
for a rudimentary appendage, and O to denote the loss of an
appendage. The non-typical reductions are dealt with in another
part of this paper.
Glaucothoe.
Left : Right
. 1 O:
0
tt
2 R:
R
..
3 R:
i:
h
1 R:
i:
tt
5 R:
R
ii
6+R:
R
Sixth Stage.
Left : Right
O : I)
Ru : O
R: 0
]{ : 0
R : ()
+R : R —
Adult Male.
Left : Right
() : 0
O: 0
R: 0
R: O
R : 0
+R: R—
Adult Female.
Left : Right
O : O
R : O
R: <)
R : 0
R : 0
+ R: R—
At the moult to the seventh stage or, more rarely, at the next
following moult, the retained rudiments are lost, except the one on
the second segment in those crabs which will become females. But
although sex can be recognized thus early, not less than a year and


THOMPSON: METAMORPHOSES OF HERMIT CRAB. 159
probably a still longer time must elapse before sexual maturity is
attained. In the young females the rudiment on the second seg-
ment begins to develop into a perfect pleopod at about thirty days
from the glaucothoe. The remaining pleopods, and in many
instances, this one also, do not begin to alter to the female type
until ten days later, and the development of this type requires
several moults for its full completion.
The sixth stage was figured by Faxon ('82), and described by
Agassiz ('75) as "stage when they need a shell." The later ado-
lescent stages have not been recorded by anyone.
Special Account op the Metamokpiiosis.
The account of the development that has already been given
has briefly described the plan of the anatomy in the larval stages
and shown how this gives place to the adult type of structure in
the glaucothoe phase. In the present chapter it is proposed to
discuss in more detail the larval anatomy and the modifications by
which the adult type is produced.
Techni'/ue.— In this work, microscopical examination of living
larvae or of specimens cleared in cedar oil gave only dubious results
and therefore it was necessary throughout to employ serial sections.
No killing fluid was uniformly satisfactory because of the difficulties
of penetration and because the same tissues in different stages do
not react in the same way to a reagent. The best and most reliable
solutions used were a saturated aqueous solution of picric acid, five
percent picro-acetic, and the stronger solution of Flemming. The
brown stain produced in the contents of the liver cells was a serious
objection to the use of the otherwise excellent reagents of the vom
Rath series, and the lime salts in the integument in all stages barred
picrosulphuric. Perenyi's fluid and corrosive sublimate'in its various
forms, the latter admitted by incision, though they frequently gave
excellent fixations, were uncertain in action.
To prevent injury to the delicate tissues after fixation, all the mate-
rial was imbedded in paraffin from xylol as soon as possible. The
integument in this connection offered no obstacle to penetration or
dehydration. In sectioning, however, it was troublesome, forcing
me to cut with a thickness of ten micra. Selective staining was


160 PROCEEDINGS: BOSTON SOCIETY NATURAL HISTORY.
found essential to good work, and iron-alum haematoxylin, counter-
stained with orange G or Bordeaux red proved especially valuable.
It was necessary to cut sections in both sagitto-longitudinal and
transverse planes, for, although transverse sections are the more
generally valuable, the others are essential for purposes of compari-
son and absolutely indispensable in studying the muscles of the
abdomen. It was found advantageous, though not absolutely neces-
sary, to have in addition to these last, a few corono-longitudinal sec-
tions of each stage.
External anatomy.— The development of the form of the body
has already been in the main adequately treated in the preceding
chapter, while the details of the development of the appendages will
be better understood from the figures on plates 5, 6, and 7, than
from any description. It remains, therefore, in this section, to speak
of the telson and gills only.
The telson (pl. 7, fig. 31) in the first zoea bears on each side of
the median marginal notch five feathered spines (1-5), a minute
bristle (ti), and the short, smooth spur of the angle (7). The for-
mula is: 1, 2, 3, 4, 5, 6, 7. In the second stage, a pair of new
spines (1') are added within the older series. The tips of the uropods
as they develop, are sheathed within the angle spur (7) and this is
consequently lost when these appendages become free. The angle
spur of the third zoea is a new structure (x). The third and fourth
stages have the same telson formula: 1', 1, 2, 3, 4, 5, G, x, and in
both, spine 4 is short and smooth. The eight setae on the border
of the telson of the glaucothoe represent spines 1', 1, 2, and 3 of the
zoeal series. The median cleft appears with the sixth stage, and the
adult form is attained with the seventh stage. The moderate length
of spine 4 in the earlier, and its short spur-like form in the later
zoea stages serve to differentiate the zoeae of longicarpus and
annulipes from the zoea that I have assigned to acadianus and
from the larva of the European bemhardu s (Kathke, '42; Sars, '89),
in both of which these spines are smooth and elongate throughout
the zoea phase.
As already noted, the gills become functional with the glaucothoe
stage. At this time they are present in the same number and
arrangement as in the adult crab, viz.: —


THOMPSON: METAMORPHOSES OF HERMIT CRAB.
161
Mxp, Mxp8 I II III IV V
0 0 0 0 0 1 0 1 Pleurobranch.
0222220 10 Arthrobranchs.
000 0 0000 Podobranchs.
The larger posterior gills are provided with two short rows of ova
lamellae, but the smaller anterior ones show at most only two or
three plates each. The gills on the maxilliped and cheliped seg-
ments are simple and the maxillipedal pair are so minute as to be
invisible in surface views. They cannot be detected, either, in some
excellent sections. But since they are very plainly shown in all
sections of mature glaucothoe and most of the sections in which they
cannot be seen are of younger specimens, it may be presumed that
they arise during the period. In this case it scarcely seems possible
to refer their occasional absence to error in interpretation, although
such outpnshings of the body wall usually appear at and not between
the ecdyses. Whenever present they are very distinctly shown in
sections cut in any plane.
The sixth-stage larva has cheliped gills that are divided into
two or three lamellae, but its maxillipedal gills are still simple.
These latter reach a trilamellate condition at about the fortieth
day from the glaucothoe phase. One crab eighty days from the
glaucothoe phase showed an anterior maxillipedal gill with four,
a posterior with twelve pairs of lamellae.
Internal anatomy.— A brief description of the stomach of the
adult crab is a necessary preliminary to the description of the con-
struction of this organ in the larvae, but it may be limited in scope
to an account of the topography. No details of the ossicles are
necessary, since these cannot be worked out in the serial sections of
the developmental stages.
The stomach of the adult (pl. 9, fig. 57) has a cardiac portion of
more than twice the length of the pyloric. The cardio-pyloric valve
is crowned with blunt setae. Each lateral tooth consists of two
rounded tubercles, a comb of transverse rugae, and a hairy terminal
process. The oesophageal opening is guarded by a pair of upper
and lower oesophageal plates. The pylorus is broadened laterally
into a pair of shallow upper, and prominent lower pyloric pouches
(pl. 9, fig. 57, 52, upp, Ipp). The latter have their inner surfaces
densely clothed with long setae, and above, the wall of the pylorus


162 PROCEEDINGS: BOSTON SOCIETY NATURAL HISTORY.
projects into the lumen in a prominent crest, the lateral-valve ridge
(Ivr), which terminates posteriorly as the lateral pyloro-intestinal
valve. These valves are united with the dorso-lateral and the min-
ute, median, dorsal valve for a space (pl. 9, fig. 58, do, dlv, lo) so
that these valves enter the intestine as a continuous curtain.
Throughout the zoea phase, the stomach (pl. 9, fig. 55) has a
deep and narrow cardiac portion, which is not longer than the pylo-
rus, and chitin is poorly developed except on the smooth cardio-
pyloric valve. There is no dorsal tooth; the lateral teeth are simple
and project upward instead of horizontally (pl. 9, fig. 48, It); the
pyloro-intestinal valves are not united and are three in number, viz. '•
the paired laterals and dorsal. In the first zoea the pylorus is with-
out pouches, but with the second stage an area on either side of the
median pyloric valve becomes setose and in the fourth stage these
areas are depressed to form the lateral pouches. The eardiopyloric
and median pyloric valves are at first confluent, but become distinct
with the second stage. A single oesophageal plate appears with the
fourth stage.
The stomach of the glaucothoe (pl. 9, fig. 56) may be regarded as
transitional in type, its more elongate form, horizontal lateral teeth
(pl. 9, fig. 51), and well developed lateral-valve ridges recalling the
stomach of the adult. There are indications of a dorsal tooth; the
pyloro-intestinal valves are united; and a small dorso lateral valve
(dlr) is added on each side. In other respects the larval characters
persist. No metamorphosis occurs during the period except a
change in the mutual relations of the openings from the livers and
lateral caeca.
The sixth stage has a stomach of adult type, but with the parts
less specialized. The upper pyloric pouches are still wanting, and
they are probably of late adolescent development.
In the intestine of the Decapoda, it is generally accepted that the
limits of the ehitinous lining are coincident with the limits of the
post- and mid-guts. But until this is supported by a larger body of
evidence, it is perhaps better to use the purely descriptive terms
"ehitinous" and "achitinous" for the post- and mid-guts respec-
tively. In the adult Eupagurus the achitinous gut is relatively
shorter than usual, as the ehitinous gut only extends into the ante-
rior part of the abdomen. An unpaired caecum — first described by
Swammerdam in 1787 — starts from the right side of the achitinous


THOMPSON: METAMORPHOSES OF HERMIT CRAB. 163
gut in the rear of the thorax, turns back into the abdomen and lies
there in a coil superficial to the livers (pl. 8, fig. 46, tc). The chiti-
nous gut near the point of union with the anterior gut has the usual
series of prominent folds, the methoria, and at its posterior end a
rectum is differentiated. This occupies the sixth segment and the
telson.
In the zoea and glaucothoe phases the achitinous gut is longer
than in the adult and extends back to a point within the fifth seg-
ment of the abdomen, where its short, columnar cells give place to
the larger, more vacuolated cells of the chitinous gut (pl. 9, fig. 54,
ch hit). Methoria are present at this point with the glaucothoe and
during the latter part of this period the unpaired caecum arises as an
outpushing of the dorsal wall of the achitinous gut just eephalad of
these folds. As soon as this diverticulum appears, or occasionally
a little earlier, the chitinous gut begins to encroach on the territory
of the achitinous gut. Unfortunately, however, sections throw no
light on the mechanism of the change, but a series of specimens
merely shows the caecum, methoria, and chitinous lining lying
farther and farther forward in the abdomen, until with the earlier
adolescent stages the methoria reach their definitive position in the
region of the second segment. No mitotic figures can be found and
altbough histolysis occurs at this time throughout the length of the
gut, it is not especially prominent.
The elongation of the achitinous gut which brings the proximal
end of the caecum from the abdominal into jts definitive thoracic
position, must take place late in adolescent life. The diverticulum
was still wholly abdominal in a reared specimen two months past
the glaucothoe. But some small crabs that were collected at
Wareham, Mass., in August, 1900, showed the caecum in its definitive
relations. The age of these crabs was not known, but they were
very small and their development was greatly advanced over that
of larger, reared crabs known to be sixty days past the glaucothoe
phase. So they were probably about a year old; perhaps two
years old.
An elongate achitinous gut which is gradually replaced by the
chitinous gut, is found in the young of other Decapods besides
Eupagurus. For these, however, I have only fragmentary records.
An examination of an immature Virbius, 4 mm. long, and of a
Crangon with a length of 0 mm., shows no methoria, but the


H54 PROCEEDINGS: BOSTON SOCIETY NATURAL HISTORY.
achitinou8 gut extends back as far as the fifth abdominal segment.
A similar condition is found in Homarus even in the first adolescent
— fourth — stage, at which time the animal is adult in form and has
a caecum developed from the anterior gut. Among Thalassinids,
the older zoeae of Gebia affinis and CaUianassa etimpsonihwre the
union of the two regions of the gut in the fifth abdominal segment,
and in all the preadolescent stages of Naushonia (M. T. Thompson,
:03) this relation is maintained with the addition of metboria. The
metazoea of ITippa talpoidea has the methoria in the anterior part of
segment five of the abdomen, but in the first adolescent these have
moved to the second segment. Among the Brachyura, the late
zoea of Pinnotheres displays the anterior limit of chitin at the fifth
abdominal segment and the megalops has methoria at this point.
There is no caecum at this stage. On the other hand, a very young
zoea referable either to Cancer or to Carcinus showed methoria and
a well developed caecum in the second abdominal segment. The
megalops of Callinectes hastatus likewise has these parts in this
anterior segment and in this crab they have moved to the first
segment in the first adolescent stage. Cancer irromlus has the
caecum and methoria in the rear of the thorax in both metazoea
and megalops.
The livers or enteric glands, which open from the lateral pyloric
pouches are very voluminous in the adult crab. Each consists of an
axial tube from which arise slender lateral diverticula. The latter
are long and numerous along the abdominal portion of the axis, but
short and scanty along the thoracic. Both axis and tubules have a
wall of one layer of cells with abundant cytoplasm (pl. 9, fig. 53)
vacuolated and laden with secretions. At intervals, single cells or
groups of cells, either granular, or more usually vacuolated, project
into the lumen and partly occlude it. Proper fixation of these
tissues is difficult and they do not stain readily.
Immediately caudad from the openings of the livers into the
pylorus, a pair of lateral caeca arise. These lie one on either side
of the stomach in an irregular coil. The cells of their walls are
distinguishable from those of the livers by the absence of secretions
and vacuoles, and by the ease with which they may be fixed
and stained.
During the zoea phase, the livers are cephalic in position (pl. 8,
fig. 34—37). Each communicates with the pylorus by an extensive


THOMPSON: METAMORPHOSES OF HERMIT CRAB. 1(55
opening and presents four diverticula: anterior, lateral, dorsal, and
posterior lobes. The cells of these glands are vacuolated and
distended by deposits of a highly refractive yellow substance that
stains black with osmic acid. As in the adult, cells or groups of
cells may partly occlude the lumen, and during the fourth zoea and
glaucothoe stages this phenomenon reaches its climax (pi. 9,-fig. 08).
The lateral caeca at this time are rounded glands which resemble
the livers in histology and in reactions to reagents. Posteriorly they
become approximated and enter the main canal dorso-laterally at
the origin of the intestine, ten or even twenty micra cephalad from
the openings of the livers (pl. 9, figs. 48, 49, 50).
The livers and lateral caeca of the glaucothoe have at first the
same relations as in the zoeae except that they extend within the
newly developed thorax (pl. 8, fig. 38). The lateral caeca are
insignificant in size throughout the phase, and undergo no meta-
morphosis beyond a separation of their proximal ends, so that they
enter the intestine at the sides of the stomach, as in the adult crab.
The livers, on the other hand, pass through a complicated meta-
morphosis.
A description of a selected series of glaucothoe of different ages
will better indicate the order of these modifications as well as the
relations that they bear to the changes in position or structure in the
other organs of the body, than a detailed account. The altera-
tions are of course, subject to slight individual variations.
1. A young glaucothoe just moulted from the zoea (pi. 8, fig.
38). Only the posterior lobes of the livers have increased propor-
tionally to the increased size of the stomach.
2. An older specimen, never in a shell (pl. 8, fig. 39). The
anterior and dorsal lobes of the livers are further reduced; the
lateral lobes have almost disappeared, though the one on the right
still retains a minute lumen; the posterior lobes extend to the last
segment of the thorax and the apex of the right gland lies beneath
the intestine toward the left side of the body. The openings of the
lateral caeca into the intestine are now caudad instead of cephalad
from the openings of the livers. The green glands have begun to
grow back to form the median nephrosac and a new artery is
developing in the abdomen.
3. A still older specimen. The lateral liver lobes have disap-
peared; the dorsal become mere prominences, and the posterior
barely enter the abdomen.


!()() PROCEEDINGS: BOSTON SOCIETY NATURAL HISTORY.
4. Next specimen. The posterior lobes are now abdominal, lying
to the left of the intestine, which is displaced to the right and
dorsally. The canals from the green glands reach the region of the
pericardium but are not yet united. The sexual cells are abdominal.
A rudimentary intestinal caecum is present.
5. With livers completely shifted (pl. 8, fig. 40). The dorsal
lobes have disappeared, and the anterior are greatly reduced. The
chitinous gut is rapidly elongating, advancing the caecum towards
the anterior part of the abdomen. The green glands have formed
the nephrosac. The muscles are beginning to degenerate.
G. Fourth day in the shell. The anterior lobes of the liver are
gone. The cells of the lateral caeca have taken on the adult his-
tology and reactions; the nephrosac is abdominal; the muscles and
pleopods and all the other organs are of adult type, in readiness for
the moult to the adolescent phase.
Sixth-stage crabs retain the simple cylindrical livers (pl. 8, fig. 42)
for only a few days; then diverticula begin to appear along the
borders, first of the right, then of the left gland (pl. 8, fig. 43).
The development of these diverticula seems to follow a fairly definite
plan (pl. 8, fig. 44). Owing to the way in which those from the
right gland pass under the intestine, the adult condition is ultimately
produced and the earlier displacement is obscured to casual inspec-
tion. The livers seem to lie each on its own side of the intestine.
A shift of the latter back again toward the mid-line of the body,
which becomes possible from the eighth stage on because of the
gradual proportional'increase of the diameter of the abdomen, also
aids in confusing the earlier relations. But by careful dissection,
the displacement can be traced in the adult. The main axis of the
right gland will be found to lie beneath or slightly to the left of the
gut for a considerable part of its course in the abdomen. As young
crabs show this better than older ones, two views of the abdominal
contents of the crabs collected at Wareham are appended (pl. 8,
figs. 45, 46).
The green glands of Enpnt/urus longicarpus have the same gen-
eral arrangement as the glands of J'J. bernhardus described by
Marchal ('92), except that the nephrosac is short and broad and the
canals which unite this structure to the cephalic portion of the
gland are without accessory diverticula. The cells of the canals
and nephrosac have a characteristic histological appearance (pl. 9,


THOMPSON: METAMORPHOSES OK HERMIT CRAB. 167
figs. 52, 53). Tlie cytoplasm is scanty and stains faintly; the
nuclei are small, spherical, and prominent.
The green glands cannot be found in sections of the first or
second /.oea stages. Some third zoeae show them, but they are
without a lumen and almost wholly confined within the base of
the antennae. These glands are, however, constantly present in
fourth zoeae and each has the form of a bent tube, 0.2 mm. long,
either simple or with two short, ventrally projecting diverticula at
the proximal end.
The glands of the glaucothoe are relatively longer than those of
the zoeae and extend out of the antenna vertically into the cephalo-
thoracic cavity (pl. 9, fig. bl,gg). Their shape may he compared
to a letter "L "; the orifice being situated at the angle, and the
proximal diverticula forming the shorter arm. About the time
when the livers swing to the left as they pass toward the abdomen,
the tip of the vertical limb of each green gland begins to grow back
as a canal which lies closely appressed against the lateral wall of the
cephalothorax until it reaches the region of the pericardium. Here
the canals swing toward the midline of the body, meet one another
beneath the pericardial septum, and fuse to form a nephrosac, 0.1-
0.2 mm. long (pl. 8, fig. 40), which for a considerable period may
retain an imperfect median partition as a remnant of its double
origin. Toward the end of the glaucothoe phase the nephrosac
passes to its definitive position in the abdomen (pl. 8, fig. 42) and
during the adolescent period attains to the larger proportions rela-
tive to the surrounding structures which it has in the adult crab.
The cells of the basal portion of these glands in all stages, and of
the whole gland in the zoeae and glaucothoe, have granular, homo-
geneous cytoplasm, small, reticulate nuclei, and indistinct cell
boundaries. But the canals and nephrosac from their first appear-
ance show the adult histology. Though the bulk of the gland is
not sufficient for the production of canals and nephrosac without
a multiplication of cells, no mitotic figures could be found, which
recalls the condition of the liver cells during the changes in those
organs, where the nuclei remained reticulate. But there the original
bulk is ample to form the livers of the sixth stage through a remod-
eling. In both series of specimens the tissues seemed perfectly well
preserved.
A shell gland is present throughout the zoea phase. It is lost


168 PROCEEDINGS: BOSTON SOCIETY NATURAL HISTORY.
with the moult to the glaucothoe. It opens on the ventral face of
the second maxilla and in form is a bent tube whose distal end
extends toward the cephalothoracic cavity. These glands must be
very important for the zoea before the development of the green
glands with the fourth stage, and, despite the fact that the green
glands have a lumen and are therefore presumably functional, in
the latter period the shell glands are relatively longer than in the
three earlier stages.
Our knowledge of the development of the sexual system is very
meager. A sixth-stage larva or a mature glaucothoe shows in the
abdomen near the tip of the nephrosac two fusiform clusters of
five or six cells (pl. 8, fig. 42, g\ pl. 10, fig. G3). These are readily
identifiable as the sexual glands. Younger zoeae and glaucothoe
have similar cells lying beneath the pericardial septum in the
thorax (pl. 10, fig. 62). This position recalls the grouping of the
sexual cells in the zoeae of Mysis, (Nusbaum, '87), Palaemonetes
(Allen, '93) and Gebia (Butschinsky, '94) the only Decapod larvae
for which they have been described. I was not able to find sexual
cells in any earlier stage than the fourth zoea, perhaps because of
the thickness of my sections. These cells pass to the abdomen at
the time when the livers shift, but unfortunately no sections showed
them in transition. There must, however, be an increase in their
number at this time. The pericardial group contains less than half
a dozen cells, while the abdominal clusters have five or more cells
apiece.
The time for the appearance of sexual ducts and orifices is
unknown. When adolescent larvae reach an age of about forty
days from the glaucothoe, they show what are apparently sexual
orifices. Hut if these specimens are sectioned, no openings can be
found nor anything that can be interpreted as even the an/age of a
duct. These "pseudo-orifices" must be merely shallow depressions
in the integument over the regions where the true openings will
ultimately be developed. The crabs collected at Wareham, in
1900, which were certainly not less than one year old, had the
sexual ducts well developed (pl. 9, fig. 53, sd), and the sexual
glands were large and complexly coiled, but not quite mature.
This would mean that the production of sexual products would not
have occurred in them before the following year, e., the hermit
crab is probably not mature before the second or third year of its
life (see page 168).


THOMPSON: METAMORPHOSES OF HERMIT CRAB.
Kill
Sex is recognizable at the seventh stage, from six to twelve days
after the close of the glaucothoe phase, when the males lose the
rudimentary pleopod on the second segment (pl. 4, fig. 6, ru). But
the secondary sexual characters in the female pleopods (pl. 6, fig. 24)
do not begin to appear until thirty or forty days of adolescent life
have passed and then their development is quite gradual.
This early differentiation of sex in Eupagurus loiigicarpus and
.annul ipes has an interesting bearing on the subject of parasitic
castration as it exists in the allied Eupagurus bernhardux. There,
Giard ('86) has found that the male crabs if infested with a Bopyrid,
Athelgespaguri, have pleopods of female number and form; while
the females when parasitized with the Cirriped, Peltogaster, bear
the typical female number of appendages, but in type these approxi-
mate those of the normal male. Xo data is at hand with respect to
the adolescent development of E. bernhardus, but if it resembles
that of our species at all closely, the modifications shown by the
parasitized males require either the attachment of the parasite very
early in adolescent life, or a sufficiently potent effect from its presence
to cause a reappearance of the pleopod on the second segment. The
alterations of the parasitized females on the other hand, would be
explainable as arrest of development. The parasite presumably
might attach itself at any time within the first fifty days of adoles-
cent life and yet be able to check the complete development of the
female type of appendage. Xo ecto-parasite has been found on our
Eupagurus annidipes. The only ecto-parasite on E. longicarpus, a
Bopyrid, Stegophryxus hyptius, produces no alteration in the second-
ary sexual characters of the host.
The circulatory systems of our various species of Eupagurus are
.similar in all, and that of E. bem/utrdus (Bouvier, '91) will serve as
the type. The hepatic arteries are wholly thoracic, so that only a
small part of the liver receives blood from this source. The ven-
tral thoracic artery terminates posteriorly with the branches to the
fifth pair of limbs and the abdomen is supplied by the superior
abdominal alone. As this artery enters the second segment of the
abdomen, it divides into two trunks: b and b'. The former is a
small vessel which courses superficially to the left, supplying livers,
sexual glands, and appendages. The latter is a larger trunk which
plunges downward to the right of the intestine, runs caudad along
the dorsal surface of the flexor muscles, and then in the fourth seg-


170 PROCEEDINGS: BOSTON SOCIETY NATURAL HISTORY.
ment divides into a supramuscular branch that continues the course
above the muscles, and an intramuscular branch that pierces them
and runs caudad beneath them in the position of a ventral abdominal
artery.
The blood vessels in the larvae can only be studied satisfactorily
from serial sections and hence there are annoying gaps in my record.
For the arteries are scarcely traceable unless distended with blood,
and only a few specimens of any stage will chance to show the
desired condition. As a rule, however, all the dorsal vessels were
equally well preserved and much could be learned from a single
specimen, but ventrally the shrinkage of the integument against the
ganglia obliterated the arteries to a greater or less degree. These
and other causes also, prevented a study of the venous sinuses.
The only artery passing cephalad from the heart in the earlier
zoea stages is the anterior aorta, and it extends to the base of the
rostrum, lying close beneath the dorsal wall of the cephalothorax.
In the fourth stage its anterior end becomes deflected over the sur-
face of the supra-oesophageal ganglion. No structure which sug-
gested a cephalothoracic sac similar to that described for the zoea of
Palaemonetes (Allen, '93) was found at any stage. A sternal artery
is present at all stages and passes down in its adult relations to the
thoracic ganglia, between the fiber masses for the third and fourth
pairs of limbs (pl. 7, fig. 29, at a). The antennary arteries are first
found in the fourth zoea and as they are not invariably present, they
probably arise during the period. They diverge strongly and give
off a branch to the stomach. Their ultimate distribution, however,
could not be determined for the different stages. The hepatic
arteries first appear during the glaucothoe stage after the livers
have shifted to the abdomen. There is no trace of them in younger
specimens even when the preservation of the heart and adjacent
parts is perfect.
The deferred development of the hepatic arteries can scarcely be
regarded as correlated with the reduced function of these vessels
in the adult crab, although at first this might seem probable. For,
although among Decapods it is usual (Claus, '84) for these arteries to
be present throughout the larval period, this is not an invariable rule.
The Thalassinid, Naushonia (M. T. Thompson, :03), has only anten-
nary arteries and aorta forward from the heart during the zoea and
mysis phases. Claus's figures in his monograph on the circulation in


THOMPSON: METAMORPHOSES OF HERMIT CRAB. 171
Decapods show that in the Thalassinid, Calliaxis, and the shrimp,
Crangon', the hepatic arteries, although appearing in the zoea phase,
are yet later in development than the antennary arteries. Thus the
latter in their development seem to precede the hepatics when both
pairs are not present throughout larval life. And in Eupagurus the
antennary arteries appear^unusually late, not until the fourth zoea.
Unquestionably, the ventral thoracic artery does not enter the
abdomen at any stage, but as already noted, this vessel was difficult
to trace and the distribution of its branches could not be completely
determined. It is certain, however, that in the first zoea, vessels pass
off to the first maxilliped; that in the second zoea this artery sup-
plies the third maxilliped and, in some cases at least, one or more of
the rudimentary limbs; and in the fourth stage the artery extends
to the mouth and gives off branches to the maxillipeds and to the
five pairs of rudimentary limbs. It is only necessary to suppose that
the artery gives off in addition to the branches already enumerated,
vessels to the second maxilliped iu the first, and to the two pairs of
anterior maxillipeds in the second zoea, to make the arrangement for
each stage agree exactly with the distribution of the vessels in the
"Pagurus" zoeae studied by Glaus ('84). The ascending arte-
rioles which pass up through the nerve chain between the ganglia
for the first and second limbs and in the next posterior interspace
(Bouvier, '89, '91) are discernible as early as the second zoea. Two
other ascending arterioles are found even in mature glaucothoe:
one between the maxillipedal and the cheliped ganglia and one
between the maxillipedal and the maxilla ganglia. They may yet be
detected in the adult. Some sections of zoeae suggest the possibility
of the existence during the earlier stages, of additional ascending
arterioles between the individual maxillipedal and maxilla ganglia,
but they are not conclusive.
The superior abdominal artery is present throughout the develop-
mental period. With the earlier zoea it is a simple vessel extending
almost the length of the abdomen in the first, and to the telson in
the later stages. In the fourth zoea and glaucothoe, however, this
artery gives off five pairs of segmental branches, a pair for each seg-
ment from the second to the sixth.
At the time when the livers commence their shift, the adult plan
is evolved from this simpler arrangement (pl. 10, tig. G4). A new
artery arises from the right segmental artery in the second segment,


172 PROCEEDINGS: BOSTON SOCIETY NATURAL HISTORY.
plunges downward toward the flexor muscles and during the last
part of the glaucothoe period, pierces these. Its further develop-
ment could not be followed, as it was impossible to interpret the
blood vessels through sections of the adolescent stages, owing to the
twist of the abdomen and consequent confusion of landmarks. But
it is obvious that this new vessel is to be identified with b' of the adult
crab's arteries and that b of the adult system is the superior abdom-
inal. This latter artery swings to the left with the completion of
the shift of the livers and as the displacement is usually accompanied
by the suppression of the right segmental artery of the second seg-
ment distal to the origin of b', the latter quickly assumes the adult
relations and appears to arise directly from the superior abdominal,
b. Occasionally, however, this segmental vessel will persist in mature
glaucothoe and even into the sixth stage. The fate of the other
segmental vessels is uncertain. The four posterior pairs can be
identified in very mature glaucothoe and occasionally in sixth stage
larvae. Their presence in the latter stage suggests that they may
persist in the adult crab, bnt whether this is actually the case or not
could not be determined. The anterior pleopods of the adult cer-
tainly receive blood from branches of the superior abdominal
(Bouvier, '91). But on the other hand at this time this vessel does
not extend to the rear of the abdomen and the uropods are supplied
by the inframuscular branch of b'.
We have no data with regard to the abdominal arteries of any
Pagurid outside of the genus Eupagurus and whether the peculiar
abdominal blood system of this genus is generally distributed among
the members of the group is not known. In case it is not generally
present, it might furnish a valuable criterion of relationship. Among
other Decapods the artery most nearly analogous to b' occurs in
Gebia deUura (Bouvier, '90). This vessel, however, arises directly
from the superior abdominal in the fifth segment and such a poste-
rior position appears to me effectually to militate against regarding
it.as homologous with the Eupagurid artery, //. The development
of the latter suggests rather that it originated either as a new
structure, or, more likely, as an enlargement of some minor branch
of the segmental artery of the second segment, when the gradual
suppression of the ventral abdominal artery necessitated a more per-
fect connection between the dorsal arterial trunk and the ventral
region of the body. Once introduced, the branch has usurped many


THOMPSON: METAMORPHOSES OK HERMIT CRAB. 173
of the functions which ordinarily belong to a superior abdominal
artery.
The abdominal musculature of the young Eupagurus was studied
chiefly from serial sections. Although only longitudinal sections
proved of value, this does not at all effect the accuracy of the results,
since in examining the muscles of adult Cambarus or Homarus by
means of sections, only those cut in the longitudinal plane are
interpretable. Such sections, however, throw a great deal of light
on the arrangement of the muscles in these Crustacea. And I feel
the more confidence in my conclusions with regard to musculature
in these forms from the fact that they are based on both dissections
and sections, although they differ somewhat from those usually
accepted. In any event, the abdominal muscles of the young
Eupagurus are essentially like the muscles of Cambarus, Homarus,
and other Maerura; and sections of each are mutually comparable
(pl. 10, figs. 59,60, 61).
The muscles of Eupagurus reach their highest efficiency with the
glaucothoe stage. At this time the extensors are well developed
with a generally longitudinal course (pl. 10, fig. 59, ext); the
pleopodal muscles converge from a region of attachment above the
hinge or metacleis where the segments interlock and are independent
of the flexors (pl. 10, fig. 59a, pim). The flexors comprise several
muscles, the descending, transverse, longitudinalis, and loop-envelop-
ing. The arrangement of these various muscles can best be under-
stood from the study of a single segment, selecting segment two as
typical.
The abdomen of the glaucothoe is highly convex and hence the
attachment above the hinge, the metacleis insertion, is more dor-
sal than in the flatter segment of Homarus. The descending and
transverse muscles arise from this point. The former is a broad
band whose ventral end is inserted at the articulation with segment
one; the latter is cylindrical and runs first ventrally and then trans-
versely to come into intimate union with its mate from the other
side of the body. In Cambarus and Homarus the transversalis is
flat and in the latter genus it arises as part of the loop-enveloping
muscles.
The lateralis or lateral longitudinal passes from the anterior to
the posterior borders of the segment at the sides; and ventrally
above the nerve cord the ventralis or ventral longitudinal occupies


174 PROCEEDINGS: BOSTON SOCIETY NATURAL HISTORY.
a similar position. Possibly the latter is a series of muscles rather
than a single bundle, but the sections are not clear on this point.
The loop-enveloping muscles, which form the bulk of the flexors,
arise in common with the ventralis muscles from the anterior
boundary of the segment, but ascend as a broad sheet of fibers
(pl. 10, fig. 59b). They soon become transverse in course and grad-
ually separate into two parts, the loop or circularis and the envel-
oping or oblique. The fibers composing the loop muscles, probably
augmented by fibers from the metacleis insertion, turn longitudinally
so that the anterior end of the muscle lies dorsad of the transversalis,
meeting the posterior end of the loop muscle of segment one, while
its posterior end passes into segment three and there meets the
loop muscle of that segment above the transversalis. _ The middle
portion of the muscle is depressed and the successive arcs are very
characteristic in longitudinal sections. Toward the mid-plane of the
body a slip (pl. 10, fig. 59e, .'.) passes down from the anterior end of
the muscle to become attached at the union of segments three and
four, but whether on the sternum of segment three or into the
articular membrane cannot be determined. Still nearer the mid-
plane, the whole muscle, the anterior fibers changing first, becomes
a longitudinal descending muscle.
The fibers which go to form the enveloping muscle retain a more
generally transverse course as they are separated from the fibers of
the loop muscle; and therefore this muscle lies across the belly of
the loop muscle (pl. 10, fig. 59d). Toward the mid-plane of the body
it becomes in its turn descending, and goes down closely associated
with the descending part of the loop muscles (pl. 10, tig. 59f, x3).
The attachments of the descending portions of these muscles are not
clearly shown in the sections. At times they are apparently at the
articulation between the segments three and four; again the muscles
seem to pass on to the next posterior articulation, a discrepancy
which is possibly due to an attachment in part at both points. The
loop muscle is small in Cambarus and is largely or perhaps wholly
derived from the metacleis insertion. In Homarus it is a large mus-
cle and derived from the metacleis insertion and attached at its
descending end to the sternum of the third segment. In Homarus
the slip x is attached to the sternal plate of segment three. The
enveloping muscle in both of these Macrura and in allied forms is
very large.


THOMPSON: METAMORPHOSES OF HERMIT CRAB. 17")
The foregoing description for the second segment will apply
equally well to segments one and three except that in the former
the anterior attachments are on the eephalothoracic walls. The
fourth segment has a weak loop muscle which barely reaches the
end of the loop muscle from segment three. Posteriorly its descend-
ing portion seems to be attached at the articulation between seg-
ments five and six. The fifth segment has only pleopodal, trans-
versalis, and ventralis muscles. The sixth segment has pleopodal,
transversalis, and ventralis, and the last mentioned muscle lies
beneath the nerve cord.
The muscles of the zoeae are on the whole similar to those of the
glaucothoe, but less well developed and with weak attachments.
The pleopodal muscles are wanting, except those for the uropods
which come in with the fourth zoea. The three earlier zoeae have
also the fourth segment like the fifth from the absence of loop-
enveloping muscles.
Mature glaucothoe, adolescent crabs, and adults have a musculature
of a totally different type. The extensors are extremely weak; a
thin iayer of fibers — the integumentary muscles — lines the integu-
ment beneath the nerve cord; the flexors are bulky and those of the
right side are considerably larger than those on the left. Hut the
flexors on both sides are merely a series of strongly diagonal bands
with the more dorsal fibers runninsr almost transverselv. Minor
peculiarities present themselves in the different segments, but there
is nothing which suggests either transversalis or loop-enveloping
muscles, although descending, lateralis, and ventralis muscles are
doubtfully identifiable. These muscles have been compared to the
chevron-like muscles of Gebia and Callianassa and described as a
crowded series of such "chevrons" (Bordage, '93).
If these muscles are to be identified witli their forerunners in the
larvae, a study of their metamorphosis becomes imperative. But
unfortunately this could not be attained in as complete a form as
might be desired. The changes are crowded into a comparatively
short period near the end of the glaucothoe phase, the degenerative
processes make it difficult to secure good preservation and there
seems to be a slight reconstruction or remodeling near the very end
of the alterations. The following data are at hand, however, from
several individuals, and from them a general notion of the homol-
ogies of the adult muscles can be obtained.


176 PROCEEDINGS: BOSTON SOCIETY NATURAL HISTORY.
1. The transversalis muscles begin to lose their fibers at the
time when the livers shift to the abdomen.
2. After the shift is completed and the nephrosac has passed to
the abdomen, the loop muscle's fibers have become straight, and the
descending portions of this and of the enveloping muscles are weak.
The descending, lateralis, and ventralis muscles remain distinct, but
the transversalis, the pleopodal muscles, and the intrinsic muscles
of the pleopods themselves have disappeared.
3. Then the integumentary muscles appear; the descending and
the ventralis still show fibers, but only a few fibers can be found
in the loop-enveloping muscles. The columella prominence is
developing.
4. Still later, traces of the enveloping and descending muscles
can still be identified, but only with difficult}'. The ventralis and
integumentary muscles are somewhat more distinct.
The flexor muscles of the adult hermit crab then, evidently lack
the transversalis elements, and retain only remnants either of
descending or of lateralis muscles. The relative proportions of the
remaining flexors, the ventralis and the loop-enveloping muscles,
are not readily or surely determinable. Probably the ventralis
plays the larger rdle.
The thin layer of integumentary muscles seems to be derived from
scattered fibers that lie in the same position during the glaucothoe
stage. These may also give rise to the apical fibers of the muscles
in the columella prominence but the basal fibers of this organ are
certainly derived from the ventralis of the third segment and it is
possible that the others are also.
The theory that the flexor muscles of the hermit crab are a
crowded series of "chevron" muscles is scarcely tenable in face of
this evidence from the study of their structure and metamorphosis.
The support that it has received from the chevron-like muscles of
Gebia and Callianassa also fails when a closer examination is made.
In Gebia certainly, the "chevrons" resolve themselves into loop-
enveloping systems with a weak loop and insignificant transversalis,
so that the preponderance of the enveloping element gives a notably
oblique course to the muscles.
The larval hermit crab has, therefore, abdominal muscles more
like those of generalized Macrurous Crustacea, than those of such
Thalassinids as Gebia, and it would be interesting in this connection


THOMPSON: METAMORPHOSES OF HERMIT CRAB. 177
to know the type of musculature among the less specialized Thalas-
sinoids. This might throw some light on the extent to which the
perfection of the Maeruran type in the Eupagurus larvae is palin-
genetic. For it is not possible to regard it as simply correlated with
the active life of the zoea or glaucothoe. The very active megalops
of the Brachyuran, Callinectes, although provided with enormous
pleopodal and well developed descending muscles, has of the
remaining possible flexors only the ventralis bundles; the metazoea
of Pinnotheres and both the young zoea and the metazoea of Cancer
show similar relations, with the addition of a few fibers which may
be doubtfully identified as loop-enveloping elements. Moreover,
Gebia, like other Macrura, has in the larval stages a very perfect
loop-enveloping system of flexors, and these are carried on into the
first adolescent almost intact, although the animal is at this period
quite inactive.
The muscles of the stomach of the zoea and glaucothoe are very
simple. At first only dorsal and ventral supporting muscles are
present (pl. 9, tig. 47). But with the second stage two additional
bundles arise with attachments near the point where the oeso-
phageal plates will later appear and the third zoea adds two more
which extend from the anterior face of the stomach forward to the
cephalothoracic wall. This simple arrangement is retained by the
metazoea and by the glaucothoe until the very close of the latter
period. Then it rapidly gives place to the complex musculature of
the adult. The new muscles are developed from myoblasts which
lie around and above the stomach from the third zoea onward. Of
the zoeal muscles, the pair that extend from the region of the oeso-
phageal plate are alone retained.
It seems probable that the adductors of the mandibles in the
adult are also new structures with the glaucothoe, and not deriva-
tives of the fan of muscles that moves the jaws in the zoeae (pl. 9,
fig. 49). But the moult from the fourth zoea produces so great a
change in the plane of the mouth parts relative to the body axis that
a satisfactory comparison between sections through this region in
zoea and glaucothoe is impossible. The remaining muscles of the
body call for no especial mention. The developing limbs remain
filled with undifferentiated tissue until the fourth zoea and then
fibers begin to make their appearance.
The thoracic portion of the nervous system of the adult hermit


178 PROCEEDINGS: BOSTON SOCIETY NATURAL HISTORY.
crab is symmetrical. The supra-oesophageal ganglion resembles in
its finer structure the same ganglion in Carcinus (Bethe, '95, '97)
or Astacus (Krieger, '80). The ventral ganglia are somewhat con-
centrated (pl. 7, tig. 29). Those that supply the mouth parts form
a quadrate mass which is closely associated with the more posterior
thoracic ganglia. The three interior of these are large with strong
ventral projections of the fiber masses; the two posterior are
insignificant.
The abdominal portion of the nervous system is but little modi-
fied by the asymmetry. The ganglion of the first segment is not
apparent and is presumably fused with the thoracic mass (Bouvier,
'89). The five remaining ganglia are distinct (pl. 7, fig. 30) and
those for segments two, three, and four are displaced to the left.
The displacement of the ganglion for the fourth segment is very
slight and scarcely establishable in the adult crab, but is well shown
in sections of adolescent crabs. Each abdominal ganglion gives off
a pair of ganglionic nerves (gri) beneath the integumentary muscles,
either to a pleopod or to the point where a pleopod morphologically
ought to be situated. The areas supplied by these nerves also
receive a more delicate commissural (coinm n) above the muscles
from the commissure cephalad from the ganglion. A similar
arrangement of sub- and supramuscular nerves is found in (iebia
uffiiiis. There is no obvious difference in the size of the nerves
supplying the right and left sides of the body, and they all show
fibers and bipolar cells. This lack of asymmetry appears less
striking when we recollect that these nerves are probably very
largely sensory and this function would be nearly equal on both
sides of the body through the hairs on the pleopods on the right
and the tufts of hair on the left at the points where pleopods
ought to stand. The function of the muscles also is very simple
and so the larger size of the right flexors cannot greatly disturb
the equality of the two sides of the body with respect to other
structures.
The nervous system develops almost without metamorphosis.
The earlier larval stages have a supra-oesophageal ganglion of
essentially the same structure as that of the adult, but differing
in the smaller relative bulk of the fiber masses and fiber tracts as
compared with the ganglion cells. In the zoea phase these cells
completely surround the fiber masses and tracts, and although less


THOMPSON: METAMORPHOSES OF HERMIT CRAB.
17!)
numerous in the glaucothoe, the cell groups still are indistinguishable
and confluent. But by the time the adolescent stages are reached
the groups though still united, become identifiable. The median
ventral group, the cellulae mediales, is the first to become com-
pletely distinct and separate, in crabs forty days past the glaucothoe.
The adult separation of all groups was found in the Wareham crabs.
The globuli of the brain decrease in size in the earlier stages more
rapidly than do the fiber masses. The diameter of each bears the
following ratio to the width of the latter, measured in coronal or
transverse sections at different periods : —
Diameter of Width of
Stage globulus. intervening fiber mass.
Adult
Sixth stage
Fifth stage
Second stage
First stage
1.42 +
2
2
2.6
3
The infra-oesophageal ganglion has a history similar to that of
"the supra-oesophageal. The zoeae have small fiber masses which
are deeply imbedded in ganglion cells and these latter are especially
prominent on the ventral side of the ganglion. In surface view,
therefore, the ganglia would exhibit as great a degree of concen-
tration as in the adult, but in reality the individual fiber masses are
much more distinct, especially those for the maxillipedal and max-
illa segments. The fiber masses for the limb segments are very
small in the zoea phase. With the glaucothoe, the cheliped gang-
lion alone shows a ventralward projection, this detail appearing for
the other limbs in the adolescent phase. Nerves pass to the rudi-
mentary limbs in the fourth zoea.
The three earlier zoea stages have possible traces of the first
abdominal ganglion in two minute fiber masses that lie outside
of the thoraco-abdominal commissures at their origin. Nothing of
this can be detected later. The length of the segments of the
abdomen is more nearly equal during the zoea and glaucothoe
phases than during later life, and hence the five abdominal ganglia
are nearly equidistant from one another at these periods. Both
ganglionic and commissural nerves are present from the first zoea
,(pl. 7, fig. 29). At the close of the glaucothoe period, after the


180 PROCEEDINGS: BOSTON SOCIETY NATURAL HISTORY.
muscles are already much degenerated, the commissure between the
ganglia of the second and third segments shortens by one third, the
next posterior commissure shortens to a less extent, and commis-
sures 4 and 5 lengthen. These changes bring the ganglia into their
definitive distances from each other, accompanying the changes in
the lengths of the respective abdominal segments preparatory to the
moult to the sixth stage. The displacement of the anterior ganglia
to the left is one of the last changes before the glaucothoe stage
closes, and occurs as the columella prominence arises.
The Shell in the Ojitocexy.
Rathke noted in 1840 that the developing hermit crab became
slightly asymmetrical before the close of the zoea phase and all
observers have recorded asymmetry for the chelipeds and usually
also, dissimilarity for the uropods in the glaucothoe phase. Agassiz
('75) went a step farther and found that a very considerable advance
toward the adult asymmetry might be attained before the larva ever
entered a shell. It seemed clear from his account that this could
not be regarded as the invariable sequence; but how far it might be
looked upon as typical remained uncertain. He says of the change:
"When the moult has taken place which brings them to the stage
when they need a shell [my sixth stage] we find an important change
in the two pair of feet now changed to shorter feet capable of propel-
ling the crab in and out of the shell; we find that all the abdominal
appendages except those of the last joint are lost; but the great dis-
tinction .... is the curling of the abdomen; its rings .... are quite
indistinct and the test covering it is reduced to a mere film. It is
therefore natural that the young crab should seek shelter for this
exposed portion of his body." His figures, published in 1882
(Faxon, '82), corrected and furnished a partial interpretation of this
text. But they seemed to indicate that these changes might be only
a beginning of a gradual metamorphosis to the complete adult form.
It has therefore remained for the present research to show that the
adult structure, so far as its plan is concerned, is completely attained
during the glaucothoe stage, and that typically the metamorphosis
is inaugurated before a shell is taken. Nevertheless the young
crab always enters the shell before the changes are far advanced,


THOMPSON: METAMORPHOSES OF HERMIT CRAB. 181
usually within the first forty-eight hours of its life as a glaucothoe.
\V hat then is the significance of the shell in the ontogeny? It is
clear that it cannot be regarded as an essential factor, but may it
not perhaps influence the development to a lesser extent?
If such an influence were present it obviously might act in two
directions: to affect the rate of the metamorphosis, or to modify
the order of the metamorphosis and possibly, but less probably, the
resulting structures themselves. In addition to this, despite the fact
that the dextral asymmetry is fixed in the ontogeny, appearing before
a shell is taken and completing itself even when the latter is absent,
it is conceivable that the entrance into a dextral shell which nor-
mally occurs at a definite point in the development may have some
controlling influence, so that exposure to other than dextral shells
might not be without effect on the course or integrity of the meta-
morphosis.
Effect on the rate of development.— In connection with our first
possible effect of the shell, that on the rate of the development, the
influence, if any, will involve the glaucothoe more than a later
stage as this is the period of most extensive change and the time
when the shell is normally assumed. Therefore a number of experi-
ments were performed on glaucothoe by which they were exposed
to different environments relative to the time when the shell was
taken and also to the shape of the shell, whether dextral, sinistral,
or uncoiled.
For these experiments the age of the larvae used, had to be known
within as narrow limits as possible. So fourth-stage zoeae were col-
lected and placed in a large jar, where during the night many would
moult to the glaucothoe. New zoeae were used each day, and in
the morning the glaucothoe were removed and reared in one or
more experiments as the case might be. For I made it a rule never
to mix glaucothoe of different ages, but kept each batch separate
from all preceding ones. The experiments were examined daily at
the same morning hour. At first, aii additional mid-period record
was taken, but it was found that the larvae moulted chiefly in the
"small hours" of the morning and this habit rendered a mid-period
record less necessary and as the pressure of other work became
severe it was abandoned. The fairly definite time of moulting also
served to reduce the unavoidable time error which might otherwise
have had a range of twelve hours. The large number of specimens


182 PROCEEDINGS: BOSTON SOCIETY NATURAL HISTORY.
used further counteracted this error and also tended to equalize it in
the different series. So on the whole, the number of days that the
larvae remained in the glaucothoe phase was determinable with a
fair degree of accuracy, and although some error must remain, it
does not affect the comparability of the series.
Series A.
Normal.
The glaucothoe were provided with dextrally spiral shells and given
opportunity to take these whenever they pleased. The shells of the
marine snail, Soyootypus, were used, being abundantly obtained from
the egg cases. Eleven experiments were made, using 183 larvae,
99 of which attained the moult to the sixth stage.
6 glaucothoe 6 percent remained in stage 4 days.
8 •
3
ti
tt
i.
u
44
57
57
t.
tt
u
I.
.-)
2 i
2
ii
tt
tt
54
19
19
u
tt
tt
u
o
14
14
it
tt
64
4'
4
ii
tt
7
2'
2
"
.h
2'
2
it
8
The percentages for the duration of the stage may be arranged in
a curve in which the ordinates represent days, the abcissae the per-
centages (pl. 7, fig. 3'2, continuous line). This gives a very steep
curve with the mode at the fifth day.
Series B.
Delayed Normal.
In these experiments, a batch of larvae was separated into two
parts'. One half was provided at once with dextral shells to
serve as a control experiment; the other did not receive shells
immediately. The figures for the control records have already been


THOMPSON: METAMORPHOSES OF HERMIT CRAB. 183
given as part of the figures for series A. They are separately listed
here for comparison with the figures for the "delay" record.
B 1.
The larvae were kept three days without shells and were then
provided with them, and they entered these within a few hours.
One experiment, using 10 glaucothoe in the control section, all of
which lived to the sixth stage; and 10 in the delay section, 7 of
which attained the moult.
Control. Belay.
1 glau. 10 % r'm'n'd 4 days. 1 glau. 14 % r'm'n'd 5 days.
8 " 80 % " 5" 4" 57 % " 5^"
1 " 10 % " 5J" 2" 28 % " 6
B 2
The larvae were kept four days without shells and were then pro-
vided with them. Three experiments, using 55 crabs in the control,
25 of which attained the moult to the sixth stage; and about the
same number in the delay, 22 of which reached that moult.
Control. Delay.
3 glau. 12 % r'm'n'd 4 days. 6 glau. 27 % r'm'n'd 5 days.
2" 8 %" 4J" 2" 9 %" b\"
11)" 76%" 5" 12" 55%" 6"
1" 4 %" o\" 2" 9 %" 7"
The curves that may be constructed from the percentages of these
experiments, resemble in shape the curves for the experiments of
series A (pl. 7, fig. 32), but have a more restricted range. The
curve for the three days' delay (broken line) has the mode between
the fifth and sixth day; while that for the four days' delay
(dotted line) has the mode at the sixth day. The curves for the
control experiments exactly resemble the curve plotted for series A.


184 PROCEEDINGS: BOSTON SOCIETY NATURAL HISTORY.
Series C.
Naked.
In these experiments, glaucothoe were reared without any covering
for the body. Five experiments, using 123 larvae; 50 of which
attained the moult to the sixth stage. No mid-period examination.
7 glaucothoe 14 percent remained in stage 4 days.
6 " 12" """5"
21 " 42" """6"
15 " 30" """7"
\ *t 2 u u tt "8 u
The curve that may be constructed (pl. 7, fig. 33, continuous line)
has the same range as the curve for the experiments of series A,
but its mode is on the sixth day and the seventh day percentage is
high, giving it a very different shape from the curves for either A
or B. It may be objected that the absence of a mid-period exami-
nation is responsible in large measure for this difference in shape.
But the mid-period moults, if not separately recorded, will merely
increase the record for the following morning and if the curve for
A be plotted with the mid-period moults thus added to the day
records (dotted line), it does not at all resemble the curve for the
present series. So although the sixth and seventh day percent-
ages may readily be somewhat higher than they would have been
with a mid-period count, the difference would be insufficient to
modify the curve essentially.
Series D.
Indifferent.
The glaucothoe were provided with straight tubes of such
diameter that they could not twist the bod}-. Glass tubes proved
the best of the various objects tried. Three experiments, using (54
larvae, 39 of which reached the moult to the sixth stage. No mid-
period examination.
35 glaucothoe 89.7 percent remained in stage 5 days.
4 « jo « " « » (i »


THOMPSON: METAMORPHOSES OF HERMIT CRAB. 185
The curve for these experiments would have a very limited range
and show the mode on the fifth day.
Series E.
Sinistral.
The larvae were provided with small sinistral shells and given
opportunity to take them at will. The shells of the fresh-water
snail, Physa heterostropha, were used. Eight experiments, using
146 glaucothoe, 58 of which attained the moult to the sixth stage.
No mid-period examination.
3 glaucothoe 5 percent remained in stage 4 days.
35 it 60" " "" 5"
12 " 20.7" " "" 6"
8 " 13.8" " "" 7"
The curve which may be plotted for this series is similar to the
curves for A, B, or D except that the percentage for the sixth day is
slightly higher (pl. 7, fig. 33, broken line). But this may be
attributed to the fact that a large number of larvae, 36 percent,
delayed their entrance into the shell and so retarded their develop-
ment. The absence of a mid-period examination probably also
contributes to increase the percentages for the sixth day.
These six series of experiments clearly show that the minimum
duration of the glaucothoe stage and phase under any condition is
four days, and the maximum duration is eight days. In all classes of
experiments,except those of the "indifferent" and "delayed" series,
some larvae remained in the stage the minimum period. The
maximum period was reached not only in the "naked" but also in
the "normal" series. The experiments also show that the shell
exercises an influence on the rate of development proportional to
the time which elapses between the moult from the last zoea
stage and the time when the glaucothoe enters the shell. The
experiments with shells of other than dextral coil, however, so
closely resembled the latter " normal" experiments that no attempt
was made to repeat them under "delay" conditions. They indicate
that the influence of the shell is independent of its form as far as any


186 PROCEEDINGS: BOSTON SOCIETY NATURAL HISTORY.
direct effect is involved. But when shells depart strongly from the
dextral type, an indirect influence is exerted, dependent on the fact
that many larvae are reluctant to enter such shells.
Effect on health.—Another influence of the shell on the growing
larva was brought out by the experiments. Attention has already
been called to the high mortality among Eupagurus larvae. Now
this mortality was especially noticeable in those experiments where
the glaucothoe were either wholly prevented from entering a shell
(C) or only permitted to enter one after a considerable delay
(B 2). Here it reached about sixty percent. In the "normal"
experiments and in the less abnormal experiments of the "delay"
series (B 1), only forty-six and thirty percent respectively of the
glaucothoe perished before the moult to the sixth stage could be
reached. The entrance into a shell early in the glaucothoe phase
has then an important bearing with regard to the health of the crab.
At first sight also this seems to be to some extent correlated
with the form of the shell used. The glaucothoe that lived in
straight glass tubes (D) lost thirty-nine percent of their number, a
death rate less than that for the normally reared crabs. The crabs
in sinistral shells, on the other hand, had a death rate as high as that
prevailing among crabs which never obtained a shell. But of course,
as before, this may be the result of the reluctance of glaucothoe
to enter sinistral shells with the consequent prejudice to their health,
rather than a direct effect of the peculiar twist of the shell. With
my adolescent crabs the death rate among those reared with sinis-
tral shells was not noticeably higher than among those reared in
other kinds of houses. Although this is based on a smaller number
of crabs, it would lend support to the idea that the higher death
rate among "sinistrally reared" glaucothoe was indirect. A partic-
ularly disastrous experiment with adolescents showed a mortality
among larvae living in dextral, indifferent, and sinistral shells, of
fifty-seven, fifty-two, and sixty percent, respectively.
The absence of a habitation is invariably injurious. The mortality
for "naked" adolescents in the above experiment was eighty-one
percent. Even with wild adults death usually follows after a short
continuance of the "naked" condition. Injury to the soft and
defenseless abdomen is generally the immediate cause of death, but
there are occasional indications which may point to some general
physiological disturbance caused by the lack of a body covering.


THOMPSON: METAMORPHOSES OF HERMIT CRAB. 187
Effect of the shell on structure.— Since the shell determines to
some extent the duration of the glaucothoe phase and also affects
the health of the crab, a further influence might be perhaps expected.
But in vain. The order in which the different organs develop from
the larval to the adult type is the same whether the glaucothoe
obtains its shell early or late in the period. It is independent of the
presence or absence of the shell, and is not affected by its shape.
Moreover, no difference in the structure or arrangement of the
parts of the body of sixth-stage crabs reared under the various
conditions can be detected, except in the reduction of the pleopods.
A crab of this age, typically, has no pleopods on the right side of
the body, but this is not invariably the case even among "normally
reared " larvae, and some crabs retain rudiments of one or more of
these appendages. The metamorphosis of the pleopods has been
expressed by the following formulae (see page 158): —
Glaucothoe.
Sixth Stage.
Adult g.
Adult 0.
Segment 1
O : 0
O : O
O : 0
O : O
"2
R : R
Ru : 0
0 : O
R : 0
"3
R : R
R : 0
R : 0
R : 0
4
R : R
R : 0
Ii : 0
R : 0
"6
R : R
R : 0
R : O
R : 0
"6
+R : R—
+R : It-
+R : R—
+R : R—
However, since segments one and six do not vary, the formulae may
be written in briefer form: glaucothoe = K, R, R, R : R, R, R, R;
sixth stage = Ru, R, R, R : O, O, O, O; adult i? = O, R, R, R:
O, (), O, O; and adult ? = R, R, R, R : (), (), (), ().
The nature of the variations from the normal sixth-stage formula
under all conditions was studied in about five hundred crabs
which had been reared from the glaucothoe as follows: 107 in dex-
tral shells, 62 in straight shells, 108 in sinistral shells, and 221 with-
out shells. Of these 493 crabs, 159 or thirty-two percent varied
from the usual pleopod formula. The different variations were
numerous, but they fall naturally into three groups: (a) those
where one or more of the right appendages were retained as
large (Ru +), medium (Ru), or small (Ru —) rudiments, in order
from before backward; (b) those where there was loss in number or
size of the pleopods on the left side; and (c) those where there was


188 PROCEEDINGS: BOSTON SOCIETY NATURAL HISTORY.
irregularity in the retention. The time that each crab was under
the microscope had to be very limited and the dimensions given do
not represent measurements, but are based on comparisons with the
unaltered pleopods of the left side.
Class A.
Rudiments without joints or rami.. Twenty-five combinations,
but only a few common to several specimens, viz.: —
Ru, R, R, R : Ru, O, O, O, 9 examples. All series.
Ru, R, R, R : Ru, Ru, O, O, 8" All series excepting
indifferent.
Ru, R, R, R : Ru, Ru, Ru, Ru—, 12" Naked series only.
Ru, R, R, R : Ru, Ru, Ru, Ru, 14" All series.
Ru+, R, R, R : Ru-f, Ru+, Ru-f, Ru-f, 13" Naked and indiffer-
ent only.
Class B.
Rudiments without joints or rami. The variations of this class
are probably only the result of defective ecdysis or injury during the
glaucothoe stage. Three examples.
Ru-K Ru-f-, Ru-f-, Ru-f-: Ru, Ru, Ru, Ru, 1 example. Normal series.
O, R, R, R :O, O, 0, Ru, 1 " Sinistral series.
Ru+, Ru+, Ru+, Ru-f- :Ru+, Ru-f-, Ru, Ru, 1 " Indifferent series.
Class C.
Rudiments without joints or rami except once, where the append-
age was small, but perfectly formed and of male type (R*). Eleven
examples in all.,
Ru, 11, R, R : Ru, O, Ru, Ru, 1 example. Normal series.
Ru, R, R, R : Ru, Ru, R, O, 1" ""
Ru, R, R, R : O, Ru-f, (), O. 1" ""
Ru, R, R, R : Ru, O, Ru—, O, 1" Sinistral series.
Ru, R, R, R : O, O, Ru, O, 1" ""
Ru. R, R, R : O, Ru—, Ru—, O, 3" Naked and sinistral series.
Ru, R, R, R : Ru, Ru+, Ru, O, 1" Naked series.
Ru, R, R. R : 0, O, R», O, 1" Sinistral series.
Ru, R, R, R: O, Ru-f-, Ru, R, 1" Naked series.


THOMPSON: METAMORPHOSES OP HERMIT CRAB. 189
The distribution of the variations.— The crabs that showed the
variations, the variants, were distributed among the different con-
ditions as follows: —
1.
Normal 107 crabs 21 variants 19.6 percent.
Indifferent 62" 10" 16.0"
Sinistral 103" 23 " '22.2"
Naked 221" 105" 47.0"
The percentages of the actual combinations under each environ-
ment to the whole number of variations recorded for all environ-
ments were: —
2.
Normal 28 percent of all the combinations.
Indifferent 26 " """"
Sinistral 36 " """"
Naked 221 " """"
The actual combinations of each kind presented, as compared with
all the recorded combinations for each class of variations were: —
3.
A
B
C
Normal
24%
33%
30%
Indifferent
32"
33"
—
Sinistral
36"
33"
50"
Naked
94"
—
25"
The distribution of the variant crabs among the classes of varia-
tions was as follows: —
4.
A
1!
C
Normal
21 variants
86%
4.7%
14.0 %
Indifferent
10"
80"
4.7"
—
Sinistral
23"
74"
4.7"
17.0"
Naked
105"
!)(!"
—
3 8"


190 PROCEEDINGS: BOSTON SOCIETY NATURAL HISTORY.
The first two tables closely parallel the results of the experiments
on the rate of development. Here again the emphasis is placed on
the actual presence or absence of the body covering rather than on
its form. The crabs reared in sinistral shells had only a slightly
larger percentage of varying individuals or variants, and only a little
higher degree of variability than normally reared larvae. But
those which had never entered a shell showed more than twice the
normal percentage both of variants and of variability. The crabs
reared in straight shells were less variable both in respect to the
number of variants and to the combinations of characters than the
normal, just as in the earlier experiments nearly ninety percent of
the "indifferent" glaucothoe remained five days in the stage as
against fifty-seven percent of the normally reared glaucothoe. I
cannot suggest any explanation for this phenomenon.
Tables 3 and 4 are somewhat disappointing. The greater varia-
bility of the "sinistral " and "naked " crabs over normally reared, is
not, except with the variations of class A, either united with a
tendency toward a larger percentage of combinations compared to
all the combinations for am' class, or with leanings toward the
peculiar variations of class C, although, a priori, both contingencies
might have been expected. The numerous variant crabs of the
"naked " series were almost all members of class A. Three of the
most striking variations found, viz.: Ru, R, R, R : Ru, R*, Ru, O,
Ru, K, R, R : Ru, Ru, R, O and Ru, R, R, R : O, Ru + , Ru, R,
were presented by crabs from the normal, sinistral and naked series.
Although the presence of a sinistral shell or the lack of any
covering for the body affects the larva only with respect to one
series of organs, the pleopods, it might seem that a continuance of
such a condition might be more potent; at least to conserve the
variations already present. The rudimentary pleopod on the second
segment becomes in the female the largest of the abdominal
appendages. May it not be possible for a similar reconstructive
power occasionally to reside in those other rudiments which are
retained by so many sixth-stage larvae, despite the fact that typic-
ally with the males this appendage lacks reconstructive power and
is lost very early in adolescent life. On experiment, however, it was
found that outside of the pleopods no organ was affected at all by a
continuance of abnormal conditions and only in a single case were
the latter altered.


THOMPSON: METAMORPHOSES OF HERMIT CRAB. 11)1
In the summer of 1900, a number of crabs from "sinistral"
experiments were kept with sinistral shells. No especial attention
was paid them. One of these grew much more rapidly than its
mates and when it was about eighty days past the glaucothoe phase
it was examined. It then had a carapace 3.7 mm. in length, an
antennal flagellum of 35 joints, an ophthalmic scale which was adult
in form, and sexual "pseudo-orifices " on the third pair of limbs, i. e.,
in the female position. The pleopods, however, were of male number
and type with the addition on the right side of the fourth segment
of a pleopod exactly like the corresponding appendage on the left.
The formula was: O, R, R, R,: O, O, R, O.
I am not inclined to attach much importance to this peculiar
specimen and least of all to regard its peculiarities as connected with
its life in a sinistral shell. With the exception of the pleopod
it was internally and externally, a dextrally asymmetrical crab.
Although it had lived for eighty days from the glaucothoe, it had
not lost the male type of pleopod, when normally the female type
should be almost or completely established. Moreover, the pleo-
pods were of male number, while the "pseudo-orifices " were in the
female position. It seems to me that this crab was merely a sport,
resulting very probably from the unnatural conditions of captivity,
and just as likely to have appeared in one series of experiments as
in another. However, it certainly shows that the retained rudiments
of the sixth stage may occasionally be conserved or reconstructed.
An attempt, during the following year, to repeat this experiment
with larvae that were known to be variants, was unfortunately
rendered inconclusive by an unusually high mortality- One hundred
and eighteen variant sixth-stage crabs were collected, chiefly from
"naked " and "sinistral" experiments, and reared to the adult form
under the following conditions: 30 in dcxtral shells, 7 in straight
tubes, 30 in sinistral shells, and 58 without shells. Out of
these only 37 survived to be examined—a dead crab is almost
immediately so mangled by the survivors as to be worthless for
study — and in no case did any crab preserve the variations beyond
the seventh stage and only 16 percent retained them into that stage.
That is, whenever the moult from the sixth stage failed to give the
normal adult pleopod formula, the succeeding ecdysis produced it.
Further investigations ought to be made in this direction. They
require but little care outside of a supply of food, as metazoea may


192 PROCEEDINGS: BOSTON SOCIETY NATURAL HISTORY.
daily be placed in a jar, the glaucothoe removed as they appear and
reared in balanced aquaria, new larvae being continually added and
the young crabs studied as fast as they attain the adult form.
Reliance should be mainly placed on sinistral shells. It is a waste
of material to keep adolescent crabs naked.
The results from the foregoing study of larvae under various
conditions with respect to the shells, effectually dispose of any
probability that the adult crab can be modified by altering the form
of the shell in which it lives. If in the presumably most plastic
period of the animal's life, when there is in addition a slight varia-
bility to work upon, we are unable to alter the structure except in a
most trivial detail and then but rarely —supposing that the retention
of the pleopod by the crab in 1900 was due to the form of the shell
.—we cannot expect for an instant that even a long continuance in
an abnormal shell will modify the adult.
The occasional discovery of a hermit crab (Milne-Edwards &
Bouvier, '91; Marchal, '91) in a left-handed shell has no significance
in this connection. The supply of shells runs, along shore at least,
very close to the demand. Residence in a sinistral shell simply
means that the crab has been dispossessed and is protecting his body
with the best hollow object available. An endless list of such habi-
tations might be compiled. I have collected adult crabs in the shells
of Crepidula and Vermetus; and my glaucothoe used every hollow
object they came across, from bits of algae that were twisted or rolled
and broken float-bladders of Fucus or Sargassum, to fragments of
exuviated crustacean integument and fragments of the fine dirt-tubes
of small Annelids. There is no case where any departure from the
normal form has been found in a crab which was using an unusual
residence.
Choice of shells. — A consideration of the dextral asymmetry of
our hermit crabs with the rdle played by the shell in the ontog-
eny, naturally leads to the question whether these crabs evince
any preference for one type of shell over another. The only
investigator who has attempted to study this experimentally
(Bouvier, '92) has come to the conclusion that they do not show a
preference for dextral over sinistral shells. He experimented by
supplying the crabs with a mixture of shells of different sorts and
then keeping a record of their movements. Approached from this
side, the problem is difficult of solution. When a crab seizes on a


THOMPSON: METAMORPHOSES OV HERMIT CRAB. 193
new shell, it turns it over and over, thrusts the chelipeds within
the chamber and then probably enters it. As far as I can judge,
the preliminary exploration merely tells the crab that the "shell " is
hollow, empty, and clean, and I cannot agree with those who would
see in it a measuring or comparing of the shell. Apparently the
crab does not perceive either the type or the size of the shell until it
has inserted its abdomen into the chamber and tested the shell by
moving about in it, etc., deciding by actual trial whether the new
will prove better than the old house. It almost invariably keeps a
firm hold on its former habitation, so as to be able to return to it if
the new house proves ineligible. This method of selection by trial
leads to numerous changes that are inconclusive and only serve to
confuse an observer.
On the other hand, there seems to me to be evidence which points
to a preference for shells of dextral type. If a single right-handed
shell is dropped into an aquarium where the crabs are in straight
tubes or sinistral shells, it will in the long run be found and used —
unless its size be utterly inadequate — even though the bottom is
encumbered with empty shells of other sorts. Similarly there is
a tendency to abandon sinistral shells for straight tubes. This does
not indicate any deliberate choice. The crabs are constantly
changing shells and if one obtains the dextral shell he does not as
readily exchange it as he would a shell of other type. Hence that
shell will be found in use when the aquarium is examined.
In general, for both larvae and adults, the need seems to be, first
for a covering and then for one of suitable size and shape. The
animal is twisted dextrally. Therefore, other things being equal,
the desire for a well fitting house will tend to bring the crab into
a right-handed shell. Hut there are evidently a number of unknown
factors which enter into a choice, especially a choice between shells
of the same type. And there are many instances where crabs leave
shells that, as far as appearances go, are vastly more suitable than
the ones they substitute for them.
The larvae show the tendency to take dextral shells more strongly
than do the adult crabs. The experiments show that young
provided with shells of unusual form are more restless than those
provided with dextral shells. The number that were found swim-
ming after the first twenty-four hours, which means that they were
either deferring their entrance into a shell or had come out again


194 PROCEEDINGS: BOSTON SOCIETY NATURAL HISTORY.
after entrance, aggregated 36 percent of all individuals in the sin-
istral series, '20 percent for the "indifferent" and 16 percent in the
normal series. The observed cases where glaucothoe abandoned
their shells when disturbed, were 7 percent in the sinistral experi-
ments, but only 1 percent in the normal. In the "indifferent"
experiments no specimen was observed to act in this manner. More-
over, a single dextral shell dropped into an'aquarium will certainly
be found and used.
When several glaucothoe are placed in a dish on the bottom of
which there are scattered various shells, wre find that as each larva
stops swimming, settles down and crawls on the bottom, it examines
the objects in its path and is apt to enter the first that has a cavity
in it. Then, if not satisfied, it either recommences swimming or,
more usually, makes a further search dragging about its newly
acquired house. It is well known (Agassiz, '75) that in the inves-
tigation and use of a shell the glaucothoe handle the shell exactly
in the same manner as do adults. My observations would make
them subject also to the same limitations.
Like other investigators, I have not been able to repeat the obser-
vation (Brooks, '99) that hermit crabs "even make vacant [a shell]
that seems eligible by pulling out its occupant piece by piece and
eating him," although a dead snail is quickly devoured by the crabs
and the emptied shell not infrequently used. If the snail, while the
crabs are pulling and picking at it, gets pulled out bodily, the vacant
shell may indeed be taken by one of the diners, but more usually the
«rabs devote their attention to the meat and leave the shell unno-
ticed. I have never seen anything which would imply an inten-
tional or necessary connection between the eating of the snail and
the use of the vacant shell. When, however, a crab is searching for
a shell he acts in a very different manner, attempting to pull out the
dead meat bodily, just as other rubbish is removed from a shell
that is being examined. The snail may be eaten after removal, and
I have seen a shell-hunting crab eat fragments which wTere torn off
while he was trying to extract the snail. But this is unusual; for,
as a rule, if the snail does not come out readily the crab either aban-
dons the attempt, or if desperately in need of a residence crowds in
between the meat and the shell.


THOMPSON: METAMORPHOSES OF HERMIT CRAB. 195
The Asymmetry.
Although it is evident that the structural modifications possessed
by the majority of the hermit crabs are on the whole in closest
correlation with the use of dextrally spiral shells as residences, there
are two genera in the group which suggest that the asymmetry in
its beginnings was perhaps not connected with this mode of life.
Mixtopagurus certainly might be regarded as pointing to an origin
for the asymmetry prior to, or outside of the use of a coiled shell, as
it combines a slight dextral twist with residence in holes in bits of
wood. The sinistral asymmetry of Henderson's genus Paguropsis
also seems difficult of explanation on any theory which would derive
the asymmetry primarily from the use of shells. For the right-
handed spiral has always been the predominant type of coil among
the marine Gastropod Mollusca. The point of view with regard to
these geflera remains unchanged whether the Pagurids be considered
as a natural or a convergent group.
This question of the origin of the asymmetry seems to me to be
insoluble at the present day. Detail of the adult anatomy is as yet
very scanty and for those genera about which most is known it does
not bear closely on the problem. Probably a large number of the
anatomical peculiarities of such a typically asymmetrical genus as
Eupagurus, for example, may be looked upon as inherited; similar
structures being found among the Thalassinidea, the modern repre-
sentatives of the stock from which the hermit crabs were in all
likelihood derived (Ortmann, :01). For example, the weak integu-
ment of the abdomen finds its counterpart among the members of the
latter group. These live in burrows or crevices and probably their
ancestors had similar habits. So the Pagurids might have attained
a much weakened integument before they began to exchange
stationary for movable "burrows." There is a curious resemblance
between the degenerative details of the abdominal muscles of the
Eupagurids and those of Gebia. Unfortunately, knowledge of the
abdominal musculature of more generalized Thalassinids than Gebia
is wanting. And even if the details shall be found to agree in the
two groups, such a reduction of parts may be merely the usual out-
come where the abdomen is little used by a burrowing species and
not necessarily peculiar to the Pagurids, nor inherited from their
Thalassinoid ancestors.


l!)t) PROCEEDINGS: BOSTON SOCIETY NATURAL HISTORY.
Other anatomical details also fail to throw light on the problem.
The alterations by which the posterior end of the body has been
transformed into a hook-like organ point to life in a movable
residence, but might have been developed in either a straight or a
spiral house. The tuberculated areas on the uropods in the more
symmetrical hermit crabs and on both uropods and posterior thoracic
feet in the asymmetrical forms, do not even indicate life in a mov-
able residence. They are found in Pylocheles, which lives in the
cavities of sponges. The reduction of the ventral abdominal artery
also may have come in either early or late in the development of the
hermit crab group. This vessel is very weak in some Thalassinids,
e. g., Gebia deltura (Bouvier, '90), and is still present in the
anterior abdomen of so asymmetrical a Pagurid as Paguristes
(Bouvier, '90a). The record might be thus continued with respect
to other characters and with the same result. Our knowledge of
the adult anatomy is insufficient to throw light upon the origin of
the asymmetry, although, as already noted, the anatomy has been
profoundly modified for the life in spiral shells.
The metamorphosis is also inconclusive on this point except in one
particular; and this merely points to an early use of a dextrally
spiral shell and does not bear upon the beginnings of the asymmetry.
I refer to the displacement of the right liver to the left of the intes-
tine and mid-line of the abdomen as the livers pass back from the
thoracic position. That this must be interpreted as palingenetic
seems to me to be fairly certain. Not only is it natural to extend
this significance from the liver shift itself to the displacement, but
the change also is an essentially transitory phenomenon and it is
very hard to imagine any cause on other grounds for its introduction
into the metamorphosis. If the displacement be palingenetic, it will
go far to support the idea of an early use of shells. The shift of the
livers in the phylogeny of the group probably occurred very early, for
these organs have had time to attain a complicated structure since
reaching the abdominal region. Further, they lie far back in the
thorax in modern Thalassinids. Now, if these organs were already
abdominal at the time when the shell was first assumed, it is difficult
to understand how the compression of the right side of the body
against the columella of the shell could have caused a displacement
rather than an atrophy of the right liver. But if they were still
thoracic or barely abdominal when life in shells first began, displace-


THOMPSON: METAMORPHOSES OF HERMIT CRAB. 197
ment during the development of the abdominal position would be
the most likely outcome. These considerations make it seem prob-
able that the shell was taken early in the history of the group, and
so indirectly support that theory which derives the asymmetry pri-
marily from use of shells.
Another important consideration adds further support to this
view. At the time when the ancestor or ancestors of the hermit
crabs began to seek other residences than burrows or crevices, the
chances would have been favorable for the immediate use of dextrally
spiral shells. The Pagurid group cannot have originated before the
late Cretaceous or early Eocene (Ortmann, rOl) while the Gastro-
poda are geologically a very old group. So we may suppose that
a supply of mainly dextral shells would have been at hand when the
hermit crabs began to use "movable burrows" and the asymmetry
might have been affected by the shell at the very outset. This of
course leaves the sinistral asymmetry of Paguropsis unexplained.
Thus the whole question falls back on general considerations, and
on such a basis there is no reason to abandon our "shell " theory
despite the lack of positive evidence. The relationships of such
forms as Mixtopagurus and Paguropsis to the other genera of the
group are wholly unknown. In fact, the relationships existing
between the various genera as a whole, are very obscure. The
group is in all likelihood convergent and not natural, although there
is reason to believe that most of the genera are genetically related.
If convergent, the parts of the group may be of different ages, the
characteristic mode of life being assumed at different times by vari-
ous forms. All this confusion must be cleared away before the
asymmetry can be thoroughly understood. And this will require
a vastly more extensive anatomical knowledge than we at present
possess. The extent of the group and the poor preservation of the
internal anatomy, particularly in the abdominal region, have hitherto
turned investigation toward description of species and genera.
While this is necessary, it usually does not lay enough emphasis upon
the anatomy as a whole, but is content with a record of the more
obvious and superficial detail of carapace and limbs. Now research
must go farther afield. We need information with respect to the
finer details of the structure in each species. I believe the clue to the
origin of the asymmetry and to the phylogeny of this group of Crus-
tacea lies in a study of the internal rather than the external anatomy.


198 PROCEEDINGS: BOSTON SOCIETY NATURAE HISTORY.
The Glaucothoes.
In the year 1830, Milne-Edwards described a small, supposedly
adult shrimp under the name " Glaucothoe peronii." No one seems
to have questioned the validity of this form as a true species, until
the year 1864, when Miiller pointed out its similarity to the post-
zoeal stage of "Pagurus" and suggested that it might be a larva.
But since that time the problems arising out of Muller's suggestion
have attracted the attention of several carcinologists. On the one
hand, Bate ('66, '68) asserted that Glaucothoe was "only a larval
Pagurus" and Faxon ('82) simply identified his postzoeal Pagurids
with Milne-Edwards's genus and with the older genus Prophylax
(Latreille, '30); while on the other hand, Glaus ('76) asserted that
the structure of the mouth parts in Glaucothoe was adult rather
than larval; Miers ('81) described a second species, G. rostrata;
and Henderson ('88) described a third species, G. carinuta.
More recently the whole problem has been re-attacked by Bouvier
through a study of the Glaucothoes themselves and, in the year
1891, he published the results of his work. He concluded that
these forms were undoubtedly larval, since they lacked sexual
orifices and opthalmic scales; that they were Pagurid rather than
Thalassinid and belonged to the asymmetrical Pagurids; that they
were not a natural group, for on structural grounds G. peronii was
assigned to Sympagurus, while G. carinata lay nearer "Pagurus or
a related form "; and that they presented "exactly all the essential
characters of the larvae described by certain embryologists under
the name of glaucothoes."
Now my research, as has been the case with the work of other
students of Pagurid development, does not bear directly on the
main problem presented by these peculiar crustacean forms. But
it does involve some subordinate phases of the problem and affects
some of the suggestions which have been put forward during the
discussion. Two main objections have been brought against the
"larval theory": the great size — 10 to 20 mm.— of the Glaucothoe
as compared with all known glaucothoe-stage larvae, and the feeling
that if these forms were really developmental stages they would be
more abundant. Henderson ('88) has especially laid stress on this
latter point. It seems to me that the cogency of both of these
objections has been overestimated.


THOMPSON: METAMORPHOSES OF HERMIT CRAB. 199
Although historical parallels are unsafe, it is interesting to note
the essential likeness of the argument from relative size to one of
the objections brought by Westwood ('35) against J. V. Thompson's
assertion that the higher Decapoda developed through a metamorpho-
sis. He said that he had collected "specimens of zoea that were ten
lines long between the spines, .... far too large to suppose that
they would subsequently put off their zoe form and appear as crabs,
bearing at the same time in mind .... the minute size of the latter
animals in the very young state, although possessing their ordinary
form."
Similarly, the supposed rarity of the Glaucothoe is more apparent
than real. When we remember that these forms were almost all
collected at some depth below the surface, the number already taken
— about twenty — does not seem especially small. And further,
even among littoral Crustacea the larvae are often less abundant
than would be inferred from the prevalence of the adults. The
young of the lobster, Homarus, for example, swim freely for ten to
twenty days after they hatch, but they are not commonly taken in
tow nets. The glaucothoe larvae of the Pagurids themselves have
seldom been recorded. 'The Plankton expedition of 1889 does not
record a Pagurid zoea, and the glaucothoe phase is relatively much
rarer than the zoeae. At Woods Hole, where hermit crabs are very
abundant and the zoeae are obtained in unusually large numbers,
fifteen or twenty glaucothoe would be an excellent haul for one day,
and the average would be much lower than this without considering
the days when the animal plankton is extremely scanty, although
occasionally I have collected, on particularly favorable night tides,
when there was a full moon, nearly or over a hundred glaucothoe.
My research bears directly upon Bate's suggestion ('68) that glau-
cothoe-phase larvae probably continue in the form until they obtain
a shell and perhaps moult and grow. Recently this theory has
been elaborated by Bouvier ('91a) as directly applied to the Glauco-
thoes themselves, in explanation of their size and rarity. "They
are, so to speak, less fortunate larvae than the rest, which continue
to grow until the time that they find a suitable habitation." Now,
I find in the genus Eupagurus a glaucothoe phase that comprises
only one stage, during which there is almost no increase in size.
The duration of this period is short and it is but little affected by
delay in obtaining a habitation. These results are found to be fully


200 PROCEEDINGS: HOSTON SOCIETY NATURAL HISTORY.
in accord with Agassiz's earlier observations ('75) on the same
genus, when his rather obscure note is critically examined. And
they ought to make us cautious in assuming as at all characteristic
of Pagurid ontogeny either the existence of more than one stage in
the glaucothoe phase or growth during that period.
However, the multiform and possibly polyphyletic nature of the
Pagurid group must not be overlooked. So that no general conclu-
sion can safely be based on the ontogeny of a single genus. Espe-
cially will this be true of the Glaucothoes. Despite Bouvier's
conclusions, there seems reason to question whether they can be
the young of forms closely related to those genera for which glauco-
thoe-stage larvae have been described, viz. . Eupagurus (Rathke, '42;
Bate, '68; Faxon, '82; Sars, '89), Spiropaguros (Sars, '89), and
Diogenes (Czerniavsky, '84). The mouth parts of Glaucothoe as
figured by Milne-Edwards appear more mature in character than
the corresponding appendages in the glaucothoe phases of Eupay-
urus anuUipes, lonr/icarpus, or our variety of bernhardus. For the
equivalent stage of the European E. bernhardus we have figures of
the maxillipeds alone (Rathke, '42), but they agree in all respects
with the same appendages in the New England species. The mouth
parts of the other described glaucothoe larvae have not been studied.
The differences in the mandibular palp, in the first maxilla, and
anterior maxilliped are especially noticeable in this connection, for
in Eupagurus these parts have retrogressed from their condition in
the zoeae and are without setae or at most only provided with rudi-
mentary setae, while in Glaucothoe they are at least setose.
The absence of ophthalmic scales also is in contrast to the con-
ditions existing in the larvae studied by Sars ('89) and Czerniavsky
('84) and probably those examined by Rathke. With respect to
his larvae Rathke ('40, '42) only makes the indefinite statement that
the eyes had the same form as those of the adult, but he seems to
have worked upon the same species as did Sars. Of Bate's larva
('68) nothing is known with respect to these structures, and the
young figured by Faxon ('82) do not have the ophthalmic scales.
This last is almost unquestionably attributable to oversight. The
figure was drawn from a rough sketch in which there was no intent
to show more than the general form. Moreover, if the figure
is correct with respect to the scales, it must represent the glauco-
thoe phase of some other genus than Eupagurus. Four species of


THOMPSON: METAMORPHOSES OF HERMIT CRAB. 201
this genus have these scales in the glaucothoe stage. As the figure
was drawn from a specimen collected near Woods Hole, a region
characterized by the practically exclusive predominance of the genus
Eupagurus this would be a most unlikely contingency.
At the present time three '' species" of Glaucothoe are recognized,
but Bouvier ("91) has noted that the specimens of G. peronii
collected by the "Talisman" are not wholly like Milne-Edwards'
type specimen. If as I believe, the glaucothoe phase in the crabs
from which these larvae are derived, is short and without any appre-
ciable growth, these discrepancies can no longer be explained as
differences in age and development. They indicate rather that the
j-oung of more than one species of hermit crab have been grouped
under the single name.
Summary.
The adult Eupagurus has a dextral asymmetry which not only
affects every organ in the abdomen but extends into the thoracic
region, so that scarcely any system of organs in the body escapes
some modification or displacement. However, with the exception
of the flexor muscles and arteries of the abdomen, the homologies
with corresponding parts in other Decapoda are clear. But the
diagonal muscle bands and the peculiar division of the superior
-abdominal artery into the two trunks, b and b', are interpretahle only
from a study of the larva. The muscles are then shown to be a
greatly degenerated loop-enveloping system from which the trans-
versalis muscle has been lost. The arteries resolve themselves into
-supra-abdominal and a new vessel, primarily derived from the second
segmental artery of the right side. This latter artery is probably
peculiar to the Pagurids and without homologue among the higher
Macrura.
The development is concentrated. There are four stages in the
zoea phase, the last of which is a metazoea. The postzoeal or glau-
cothoe phase consists of one stage which is Macruran in general
form and from the first presents a mingling of adult and larval char-
acters. The external anatomy, especially of the cephalothoracic
region, recalls the adult. The appendages — except the pleopods
—the gill formula, the asymmetry of the chelipeds and uropods,
the structure of the stomach, the concentration of the nerve ganglia


202 PROCEEDINGS: BOSTON SOCIETY NATURAL HISTORY.
and the absence of a distinguishable nerve center for the first abdo-
minal segment are more adult than larval. On the other hand,
the anatomy and positions of the livers, and lateral caeca, the mus-
culature of the stomach, the lack of an intestinal caecum, and. the
great elongation of the achitinous portion of the gut, the position of
the sexual cells beneath the pericardial septum, and the preponder-
ance of ganglion cells over fiber masses in the nervous system, are
essentially as in the zoeae. Also many organs may be interpreted
as presenting palingenetic as well as developmental characters. The
cephalic position of the green glands, the thoracic position of the
livers, the structure of the abdominal muscles and arteries, the com-
plete set of pleopods, and the segmented abdomen may possibly
serve as examples of this.
The metamorphosis by which the structures attain the adult
type of arrangement, commences before a shell is taken, although
the larva usually enters a shell within the first forty-eight hours
after the moult from the zoea phase. Through it, the livers, sexual
cells and green glands become wholly or in part abdominal, the
arterial system in this region is reorganized, and its muscles and
pleopods degenerate. The anatomy becomes completely adult in
type before the glaucothoe period ends.
The stimulus of a shell is not necessary for the completion of this
metamorphosis, any more than for its inauguration. The glaucothoe
that has never entered a shell attains the adult structural type
exactly like the glaucothoe which may have taken a shell immediately
after the moult from the zoea phase. The shell is, nevertheless, very
important. The length of the glaucothoe period bears a direct rela-
tion to the time that elapses between the moult from the zoea phase
and the entrance into a shell. Under normal conditions the period
ranges from four to eight days, lasting five days for a majority of
individuals; but deferring or preventing the taking of a covering
for the body prolongs it, so that it may reach six or seven days for
a majority of individuals. The range, however, is not altered.
This effect is derived from the presence or absence of the shell, and
is not dependent, except indirectly, on any peculiarities of its form.
The shell also deeply affects the health of both glaucothoe and
adult. Among glaucothoe which do not obtain a covering for the
body or only obtain one after a delay of some days, the death rate
is higher than among more normally reared individuals. The form


THOMPSON: METAMORPHOSES OF HERMIT CRAB. 203
of the shell here plays only a very subordinate part. Adolescent
crabs which are kept without shells show an increase of mortality
similar to that among glaucothoe.
The anatomical modifications that appear during the glaucothoe
stage are, with but one exception, uninfluenced by either the
presence, absence, or form of the shell. The exception is found in
the retention of rudimentary pleopods on the right side of the body
in the sixth stage, though typically at this period appendages should
be absent from this side. The percentages of crabs that vary from
the typical arrangement and the number of the variations displayed,
are larger among larvae which have been reared without shells.
The form of the shell is important to a small degree, and sixth-stage
larvae reared from the glaucothoe in sinistral shells are slightly more
variable than individuals reared with dextral shells.
In a single instance a crab reared in a sinistral shell reconstructed
or retained one perfect pleopod on the right side of the abdomen.
But this crab was otherwise so abnormal that there is reason to
doubt whether the variation was connected with the special environ-
ment. Outside of this case, there is no evidence that the anatomy
can be modified by a longer or shorter residence in shells of peculiar
shape. Typically the retained pleopod rudiments are lost in the
moult that closes the sixth stage.
There seems to be no reason to infer that hermit crabs learn the
shape or size of a shell, except by trial and use. But on the other
hand there is evidence for the assertion that they show a preference
for dextral shells over those of other forms, based on the fact that if
a dextral shell is dropped into an aquarium it will, in the long run,
be found and used, even when the bottom is encumbered with shells
of other kinds.
It is not possible at present to determine whether the asymmetry
of the Pagurids is primarily attributable to life in spiral shells.
The relationships of the genera are too imperfectly understood.
Nevertheless, the asymmetry in its structural details is very closely
adapted to the conditions imposed by this mode of life, which raises
a strong presumption in favor of the view that the asymmetry was,
from the first, a result of life in dextrally spiral shells. If the dis-
placement in ontogeny, of the right liver to the left of the intestine
points to a similar displacement in phylogeny the latter might
well be the outcome of a compression of the right side of the body


204 PROCEEDINGS: BOSTON SOCIETY NATURAL HISTORY.
at the time when the livers were passing from the thoracic to the
abdominal position such as a shell would cause. There is reason
also to believe that this shift took place early in the history of the
Pagurid group. Finally, the chances that the ancestral hermit crabs
would use dextral shells rather than other objects at the time when
they began to seek movable residences, are rendered very great by
the age of the Gastropod Mollusca as a group, while the Pagurids
are of comparatively recent origin.


THOMPSON: METAMORPHOSES OF HERMIT CRAB.
205
LITERATURE.
Agassiz, A.
'75. Instinct ? in hermit crabs. Amer. journ. sci., series 3, vol. 10. p. 290-
291.
Allen, E. J.
'93. Nephridia and body-cavity of some Decapod Crustacea. Quart. journ..
microscop. sci., new series, vol. 34, p. 403-42(1, pi. 36-38.
Bate, C. S.
'50. Notes on Crustacea. Ann. and mag. nat. hist., series 2, vol. 6, p.
109-111, pi. 7.
'66. Carcinological gleanings.— No. 2. Ann. and mag. nat. hist., series 3,
vol. 17, p. 24-31, pi. 2.
(Preliminary note for next article.)
'68. Carcinological gleanings.—No. 4. Ann. and mag. nat. hist., series 4,
vol. 2, p. 112-121, pl. 9-11.
(Two zoeae and the glaucothoe of "Pagurus.")
'76. On the development of the crustacean embryo, and the variations of
form exhibited in the larvae of 38 genera of Podephthalmia. Proc royal
soc. London, vol. 24, p. 375-379.
(A list of species whose zoeae were seen by W. H. Powers. He asserts
that the young of Birgus latro " certainly spend their larval life in the
sea.")
Bethe, A.
'95. Studicn ttber das centralnervensystem von Carcinus maenas nebst
angaben ilber ein neues verfahren der methylenblaufixation. Archiv f.
mikroscop. anat., vol. 44, p. 579-622, pi. 34-36.
'97. Das nervensystem von Carcinus maenas. Ein anatomisch-physiolog-
ischer versuch. Archiv f. mikroscop. anat., vol. 50, p. 460-546, pi. 25-30.
Bordage, E.
'93. Note surTetude comparee du systeme musculaire des Thalassinides et
des Paguriens. Compte rendu sommaire des séances de la soc. philo-
mathique, Paris, no. 10, p. 3-5, text fig. 1-2.
Bouvier, E. L.
'89. Le systeme nerveux des crustaces Decapodes. Ann. des sci. nat., zool.,
series 7, vol. 7, p. 73-106, pi. 7.
'90. Sur l'organization de la Gebia deltura. Bull. soc. philomathique,
Paris, series 8, vol. 2, p. 46.
'90a. Variations progressives de l'appareil circulatoire arteriel chez les
crustaces Anomoures. Bull. soc. philomathique, Paris, series 8, vol. 2, p.
179-182.


20(5 PROCEEDINGS: BOSTON* SOCIETY NATURAL HISTORY.
'91. Recherches anatouiiques but le systeme artériel des crustaces D4ca-
podes. Ann. des sci. nat., zool., series 7, vol. 11, p. 197-282, pi. 8-11.
'91a. Les Glaucothoes sont-elles des larves de Pagures? Ann. des sci.
nat., zool., series 7, vol. 12, p. 65-82.
(Anatomy of Glaucothoe and discussion.)
'92. Observations sur les raoeurs des Pagures, faites au laboratoire mari-
time du Saint-Vaast-la-Hogue pendant le mois d'Aout 1891. Bull. soc.
philomathique, Paris, series 8, vol. 4, p. 5-9.
Brooks, W. K.
'99. The foundations of zoology, viii + 339 pp.; New York.
Butschinsky, P.
'94. Zur entwicklungsgeschichte von Gebia littoralis. Zool. anzeiger, vol.
17, p. 253-256.
Claus, C.
'61. Zur kenntniss der Malacostracenlarven. WUrzburger naturwissensch.
zeitschr., vol. 2, p. 23-46, pi. 2-3.
(First and later zoea of "Pagurus." Regarded by the author as possibly
a Dromia larva.)
'76. Untersuchungen zur erforschung der genealogischen grundlage cles
crustaceen-systems. viii-|-123 pp., 25 text figs., 14 pls.; Wien: 4to.
(The previously described larvae were "Pagurus" and now the mysis
stage is described. Discussion of the metamorphosis).
'84. Zur kenntniss der kreislaufsorgane der Schizopoden und Decapoden.
Arbeiten a. d. zool. inst. d. univ. Wien, vol. 5, p. 271-318, pl. 1-9.
(Figures ventral arteries of two zoea stages, pi. 9, fig. 53-55.)
'85. Neue beitrflge zur morphologic der crustacean. Arbeiten a. d. zool.
inst. (I. univ. Wien, vol. 6, p. 1-108, 1 text fig., pi. 1-7.
(Contains a figure entitled "Metazoea of Pagurus" but regarded by
many as representing a Galathea larva.)
Czerniavsky, V.
"84. Crustacea Decapoda Pontica littoralia; materialia ad zoographiam
l'onticam comparatam. 2. Schr. nat. gesellsch. Charkoff, vol. 13, auppl.,
268 pp., 7 pls. (Russian.)
(Two zoeae and postzoeal larva of "Diogenes mrians," teste Bouvier,
'91a.)
Faxon, W.
'82. Selections from embryological monographs. 1.— Crustacea. Mem.
mus. couip. zool., vol. 9, no. 1, 11 pis. and explanation.
(Four zoeae, the glaucothoe, and the sixth stage of "Pagurus." probably
Eupagurun longkarpus; pi. 12, tig. 18-30, pi. 13, fig. 5-6.)
Giard, A.
'86. Sur la castration parasitaire chez VEupagurus bernhardus (Linne) et
chez la Gebia stellata (Mont.). Comptes rend, de l'acad. des sci., vol. 104,
p. 1113-1115.


THOMPSON: METAMORPHOSES OF HERMIT CRAB. 207
Goodsir, H. I). S.
'42. On a new genus and six new species of Crustacea, with observations
on the development of the egg and on the metamorphoses of Caligus, Car-
cinus, and Pagurus. Edinburgh new philosoph. journ., vol. 33, p. 174-192.
(Note on the first zoea of "Pagurus." )
Henderson, J. R.
'88. Report on the Anomura collected by H. M. S. Challenger during the
years 1873-76. Report on the scientific résulte of the voyage of II. M. S.
Challenger during the years 1873-76, zoology, vol. 27, xii + 221 pp., 21 pis.
(Description of Glaucothoe carinata, p. 83-84, pi. 9, fig. 1-la.)
Hesse, E.
'76. Description des crustacés rares on nouveaux des côtes de France.
Ann. des sci. nat., zool, series 6, vol. 3, art. 5, 42 pp., pl. 5-6.
(Curious description of the first zoea of 'Tagurus m'manthropos.")
Krieger, K. R.
'80. L'eber das centralnervensystem des flusskrebses. Zeitschr. f. wis-
sensch. zool., vol. 33, p. 527-594, pl. 31-33.
Latreille, P. A.
"30. Cuvier's Le règne animal, distribué d'après son organisation, pour
servir de base à l'histoire naturelle des animaux et d'introduction à
l'anatomie comparée. 2d éd., 5 vols. ; Paris.
(Description of Prophylax, vol. 4, p. 78.)
Marchai, P.
'91. Sur un Pagure habitant une coquille sénestre {Xeptunea contraria
Chenu). Bull. soc. zool. de France, vol. 16, p. 267-269.
'92 Recherches anatomiques et physiologiques sur l'appareil excréteur des
crustacés Décapodes. Archiv. de zool. expérim. et gén., series 2, vol. 10,
p. 57-275, text fig. 1-20, pl. 1-9.
Mayer, P.
'77. Zur entwicklungsgeschichte der Dekapoden. Jenaische zeitschr. f.
naturwissensch., vol. 11, p. 188-269. pl. 13-15.
(Embryology of " Pagurus striatus."]
Miers, E. J.
'81. Account of the zoological collections made during the survey of H. M.
S. "Alert" in the Straite of Magellan and on the coast of Patagonia.
Crustacea. Proc. zool. soc. London, 1881, p. 61-79, pi. 7.
(Description of Glaucothoe rostrata.)
Milne-Edwards, A., & Bouvier, E. L.
'91. Sur les modifications que subissent les Pagures suivant l'enroulement
de la coquille qu'ils habitent. Bull. soc. philomathique, Paris, series 8,
vol. 3, p. 151-153.
Milne-Edwards, II.
'30. Description des genres Glaucothoe, Sicyouie, Stïgeste et Acete, de


208 PROCEEDINGS: BOSTON SOCIETY NATURAL HISTORY.
l'ordre des crustaces DCcapodes. Ann. des sci. nat., zool., series 1, vol. 19,
p. 333-352, pi. 8-11.
(Description of Glaucothoe peronii.)
Mttller, F.
'64. Fiir Darwin. 8vo; illus. Leipzig.
(Zoea of "Pagurus," p. 35-3", fig. 26.)
Nusbaum, J.
'87. L'euibryologie de Mysis chameleo (Thompson). Archiv. de zool.
experim. et gen., series 2, vol. 5, p. 123-202, 13 text figs., pl. 5-12.
Ortmann, A. E.
: 01. Bronn's Klassen und ordnungen des thier-reichs. Arthropoda. Vol.
5, part 2, no. 60-62.
(See p. 1317, pi. 128.)
Philippi, A.
'40. Zoologische beuierkungen. 2. Das genus Zoe ist der erste zustand
von Pagurus. Archiv f. naturgesch., vol. 6, pt. 1, p. 184-186, pi. 3, fig.
7-8.
(First zoea of " Pagurus hungarus.")
Ratlike, H.
'40. Zur entwickelungsgeschichte der Dekapoden. Archiv f. naturgesch.,
vol. 6, pt. 1, p. 241-249.
(Description of three zoeae and the glaucothoe of "ragurus bern-
/tardus.")
'42. BeitrSge zur vergleichenden anatoinie und physiologie. 2. Zur ent-
wickelungsgeschichte der Dekapoden. Neuste schr. d. naturforseli.
gesellsch. in Danzig, vol. 3, pt. 4, p. 23-55, pi. 2-4.
(An expansion of the foregoing account.)
Sars, G. 0.
'89. Bidrag til kundskaben om Decapodernes forvandlinger. 2. Lithodes,
Eupagurus, Spiropagurus, Galathodes, Galathea, Munida, Porcellana
(Nephrops). Archiv f. math, og naturvid., Christiania, vol. 13, p. 133-201,
pi. 1-7.
(Zoea, metazoea, and glaucothoe of "Eupagurus hernhardus," "E.
pubescens," and "Spiropagurus chirocanthus." Note on the larvae of "S.
forbesii.")
Smith, S. I.
'73. The metamorphoses of the lobster, and other Crustacea. In VerriU's
Report upon the invertebrate animals of Vineyard Sound and adjacent
waters, with an account of the physical characters of the region. Report
U. S. comm. fish and fisheries, 1871-1872, p. 522-537, text fig. 4.
(Record of the young of Eupagurus in Vineyard Sound, p. 530.)
Swammerdam, .T.
1737-38. Bybel der natuur. 2 vols. ; Leyden.


THOMPSON: METAMORPHOSES OF HERMIT CRAB. 20't
Thompson, J. V.
'28. Zoological researches and illustrations. No. 1, 36 pp., 4 pis. Cork.
'30-'31. On the metamorphoses of Decapodous Crustacea. Zool. journ.,
vol. 5, no. 19, p. 383-384.
(Asserts a metamorphosis for "Pagurus " among other genera.)
'35. On the double metamorphosis in the Decapodous Crustacea, exempli-
fied in Cancer maenas, Linn. Philosoph. trans, royal soc. London, 1835,
pt. 2, p. 359-362, pi. 6.
(With respect to "Pagurus," merely a repetition of his former assertion.)
Thompson, M. T.
:03. A rare Thalassinid and its larva. Proc. Boston soc. nat.hist., vol. 31,
no. 1, p. 1-21, pi. 1-3.
Vigors, N. A.
'30. Untersuchungen ueber die bildung und entwickelung des flusskrebses
von Heinrich Rathke. Zool. journ., vol. 5, no. 18, p. 241-255.
West-wood, J. O.
'35. On the supposed existence of metamorphoses in the Crustacea.
Philosoph. trans, royal soc. London, 1835, pt. 2, p. 311-328, pi. 4.
Willey, A.
:00. Zoological results based on material from New Britain, New Guinea,
Loyalty Islands and elsewhere, collected during the years 1895, 1896, and
1897. Pt. 5. 4to : Cambridge, Eng.
(The young of Birgus latro hatch as zoeae.)
Printed, September, 1903.


Thompson.—Metamorphoses of Hermit Crab.
PLATE 4.
All the figures are reduced about two thirds in reproduction.
Fig. 1. First zoea, ventral. X 52. mxp„ third maxilliped.
Fig. 2. Second zoea, lateral. X 40. mnd, mandible; ul. upper lip.
Fig. 3. Third zoea, lateral. X 34. ych, yellow, rch, red chromatophores.
Fig. 4. Fourth zoea, dorsal, swimming position. X 25. ht, heart; ch int,
chitinous gut.
Fig. 5. Glaucothoe or fifth stage, dorsal. X 25.
Fig. 6. Sixth stage, dorsal. X 25. at, stomach; ic, caecum; int, intestine;
ru, rudimentary pleopod on segment two; eij, right liver.


Thompson. — Mftamokhiiuses ok Hkhmi r ("kah.
I'l.AI K +.


Pm,i\ liosros Sim- Nat. Misi Vin.. ,\.








Thomi'Son.— Metamorphoses of Hermit Crab.
PLATE 5.
Development of the appendages, showing the structure in the first (1), sec-
ond (2), third (3), and fourth (4) zoeae, the glaucothoe (5), and the adult (a).
For purposes of comparison, the appendages in each series are drawn to a
common size.
Abbreviations Common to all Figures.
Fig. 7. First antenna, dorsal.
Fig. 8. Second antenna, dorsal.
Fig. 9. Mandible, ventral.
Fig. 10. First maxilla, ventral.
Fig. 11. Second maxilla, ventral.
Fig. 12. First maxilliped, ventral.
Fig. 13. Second maxilliped, ventral.
en— endopod.
ex— exopod.
ie — lacinia externa.
(t —lacinia interna.
sc. — scaphognath i te.
prot — protopod.


Thompson.— Metamorphoses of Hermit Crab.
Plate 5.

M, T. Tliumpson dei.
I'hoc. Boston Sue. Nat. Hist. Vol. 81.








Thompson. - Metamorphoses of Hermit Crab.
PLATE 6.
Development of the appendages (continued,!.
Abbreviations as before; subscript numbers mark the stages and Roman
numerals indicate the thoracic limbs and the abdominal segments.
Fig. 14. Third maxilliped, ventral.
Fig. 15. Thoracic limbs, first zoea, lateral.
Fig. 16. Thoracic limbs, second zoea, lateral.
Fig. 17. Thoracic limbs, third zoea, lateral.
Fig. 18. Thoracic limbs, fourth zoea, lateral.
Fig. 19. Fourth limb, glaucothoe, dorsal.
Fig. 20. Fifth limb, same, dorsal.
Fig. 20a. Tip of same, adult.
Fig. 21. Abdomen of fourth zoea, showing pleopods, ventral.
Fig. 22. Pair of pleopods of fourth zoea more'enlarged, lateral.
Fig. 23. Pleopods of glaucothoe, drawn in their respective proportions.
Fig. 24. Pleopods of adult female, drawn in their respective proportions.
Fig. 25. Same of adult male.


Thompson.—Metamorphoses op Hermit Char.
Plate (i.








THOMrsos. — Metamorphoses of Hermit Crab
PLATE 7.
As before Roman numerals indicate thoracic limbs and abdominal segments
and subscripts mark the stages.
Fig. 26. Curves showing amount of variability and range of variations in
sixth-stage larvae. The ordinates indicate the conditions of rearing:
the abcissae the percentages. Continuous line = amount of varia
bility; broken line = range of variations. Conditions are: n»\
normal: na, naked; Bin, sinistral; ind, indifferent. See page 18i).
Fig. 27. Part of abdomen, first zoea; living material, showing ganglia and
nerves.
Fig. 28. Development of ophthalmic scale, a. first antenna; e, eye.
Fig. 29. Ventral surface of infra-oesophageal ganglion, adult, ar, ascending
arterioles; comm., thoracic commissures; oes, oesophageal comiuis
sures; st a, sternal artery.
Fig. 30. Eupagurus polticuris; abdominal ganglia to show arrangement of
these parts in the genus Eupagurus. gn, ganglionic nerve from fifth
ganglion; comm n, commissural nerve from fourth commissure;
cp, columella prominence; no, nerve to fifth thoracic limb.
Fig. 81. Development of telson and uropods; 31,, 31„, and 314 show the
ventral, the others the dorsal surface. Drawn to common size. ur.
anlagen of uropods.
Fig. :!2. Curves showing duration of the glaucothoe stage, normal and delayed.
Ordinates represent days, abcissae percentages. Continuous line =
A, normal; broken line = HI, delay of three days; dotted line = I3-.
delay of four days.
Fig. 33. Same, naked and sinistral. Continuous line = C, crabs kept with-
out shells; broken line= E, sinistral shells; dotted line = A, plotted
with the day and half day percentages added together.











Thompson'.— Metamorphoses of Hermit Crab
PLATE 8.
Development of the livers.
All the figures are schematic, but each represents the condition in a single
individual, depicted either from a "plotting " on profile paper, or from a wax
reconstruction. Allowance must be made for shrinkage and shiftings in the
loosely attached organs, especially in figure 34-37, and 44. Figure 34-37 are
drawn in relative proportions; figure 38-42 with a common length of thorax.
The slender liver caeca are crowded together in the living animal, but sepa-
rated here for clearness.
ABBREVIATION'S CoMMOS TO ALL FlGL'RES.
an — anus. lat I — lateral lobe.
ant I — anterior lobe. k — lateral caecum.
tlor I — dorsal lobe. oes — oesophagus.
eg — liver. post I — posterior lobe.
g— gonad. sd— sexual duct.
gg — nephrosac. .it — stomach,
ic — intestinal caecum. sta—sternal artery.
int — intestine.
Fig. 34. Thorax of first zoea, reconstructed; right, lateral. X 120.
Fig. 35. Thorax of second zoea; right, lateral. X 200.
Fig. 36. Thorax of third zoea, reconstructed; right, lateral. X 200.
Fig. 37. Thorax of fourth zoea; right lateral. X 200.
Fig. ;!8. Thorax of glaucothoe, example 1, page 165, reconstructed. X 120.
Kig. 39. Thoiax of glaucothoe, example 2, page 165, reconstructed. X 160 (?).
Fig. 40. Thorax of glaucothoe, example 5, page 16C, reconstructed. X 271 (?).
Kig. 41. Thorax of glaucothoe, example 6, page 166, reconstructed. X 129 (?).
Fig. 42. Sixth stage, reconstructed; right lateral. X 280 (?).
Kig. 43. Abdomen of mature sixth stage, showing earliest growth of liver
caeca; reconstructed.
Kig. 44. Abdomen of adolescent crab, twenty to thirty days past the glau-
cothoe; reconstructed.
Kig. 45. Abdomen of Wareham adult, see page 166. Dorsal; from photograph
of reconstruction.
Fig. 46. Same; lateral, with intestine partly cut away.











Thompson.— iMetamnrphoees of Hermit Crab.
PLATE 9.
Ahbkeviations Common to all Figures.
a— anterior aorta. int— achitinous gut.
aa — antennary arteries. lat t — lateral tooth.
ant I — anterior lobe of liver. Ic — lateral caecum.
car — carapace. Ipp — lower pyloric pouch.
rh int — chitinous gut. Ic—lateral pyloro intestinal valve
rpv — cardio-pyloric valve. Ivr— lateral-valve ridge.
dlv — dorso-lateral pyloro-intestinal mpv — median pyloric valve.
valve. mua — muscles passing to limbs.
dor I — dorsal lobe of liver. oes — oesophagus.
tit — dorsal tooth. oes comm — oesophageal commissures.
dv — dorsal pyloro-intestinal valve. op — oesophageal plate.
gg — green gland. post I — posterior lobe of liver.
ht—heart. upp — upper pyloric pouch.
Fig. 47-50. Transverse sections, first zoea.
Fig. 47. At beginning of mandibles. X 85. gang, posterior cells of supra-
oesophageal ganglion; at, stomach.
Fig. 48. Thirty micra farther bark, showing approximation of lateral caeca.
X 130. Ipp, site for lower pyloric pouch; //, lateral tooth.
Fig. 49. Ten micra beyond last. X 85.
Fig. 50. Thirty micra farther back, showing entrance of livers into pylorus,
tips of pyloro-intestinal valves and beginnings of floor of intestine,
x. X 85.
Fig. 51. Transverse section, glaucothoe.showing vertical limb of green gland.
X 85. gl'ili, globulus: «/, upper lip.
Fig. 52. Transverse section, sixth stage, through the pylorus, X 115. gang,
ventral thoracic ganglia; gg, canal from green gland to nephrosac;
upp, site of future upper pyloric pouch.
Fig. 53. Wareham adult, transverse section of abdomen, showing three liver
tubules, the ventral wall of the nephrosac {gg), sexual duct {sd), chit-
inous gut, caecum {ic), and upper part of flexor muscles {flet). X 180.
Fig. 54. Intestine of glaucothoe, longitudinal section through methoria; show-
ing chitinous gut. achitinous gut, and mesodermal sheath of gut, ma.
X 525.
Fig. 55-57. Development of stomach, somewhat diagrammatic,
from plottiiigs on profile paper.
Fig. 55. Stomach of fourth zoea, (cardiac part too short). X 184. Ic,
entrance of lateral caecum into pylorus; eg, entrance of liver into
pylorus.
Fig. 56. Stomach of glaucothoe, (cardiac part too short). X 138.
Fig. 57. Stomach of adult, X 43. lop, lower oesophageal plate; uop, upper
oesophageal plate.
Fig. 58. Transverse section, glaucothoe, to show union of the pyloro-intes-
tinal valves in the intestine and the partial occlusion of the lumen
of the liver. X 180.


Thompson. Metamorphoses or Hkhmit Ckak.
Plate 9.








Thompson.— Metamorphoses of Hermit Crab.
PLATE 10.
Koman numerals indicate thoracic limbs and abdominal segments.
Abbreviations Common to all Figures.
car— sinus of carapace. I seg a — left segmental artery.
dese m — descending muscle. nc — nerve cord.
en m — enveloping muscles. pI m— pleopodal muscles.
ext — extensor muscles. post I — posterior lobe of liver.
flex — flexor muscles. r set/ it — right segmental artery.
g — gonad. st a — sternal artery.
gang — ganglia. vent —ventralis muscles.
int — achitinous gut. x— descending portion of loop muscles.
lat m — lateralis muscles. x* — descending portion of enveloping
ten m — loop-enveloping muscles. muscles.
I m — loop muscles.
Fig. OOa-f. Muscles of abdomen, glaucothoe, slantingly sagittal; the 1st, 2d,
3d, 4th, 7th, and 10th, of a series of sections each ten micra thick.
The transversalis muscles are blackened.
Fig. 60. Diagram of the flexor muscles of glaucothoe; lateralis and ventralis
omitted except where they come in contact with the other muscles.
Viewed from within.
Fig. 61. Flexors of Cambarus; series of four sections for comparison with
sections in figure 59 ; descending, ventralis,and lateralis omitted for
clearness.
Fig. 62. Sexual cells, glaucothoe, reversed by compound microscope, sa, pos-
terior aorta, x 370.
Fig. 63. Sexual cells, sixth stage. See plate 8, figure 42. rns, nuclei of
mesodermal sheath. X 370.
Fig. 64a-d. Development of artery b', as shown in one individual.
F.xample 2, page 16"). . .sa, superior abdominal or artery b.
Fig. 64a. Dorsal aspect, plotted from sections. X 220.
Fig. 64b-d. The three successive sections indicated in fig. 64a. X 60.
Fig. 64b. Showing the outer end of the left segmental artery cut longitudi-
nally, and the right artery in cross section.
Fig. 64c. Showing the tips of the segmental arteries passing into the pleo-
podal muscles on each side.
Fig. 64d. Showing the tips of the segmental arteries in the muscles and
artery 6' running beneath the extensors before it dips toward the
flexors.



Desc m Car


Phoi\ Boston
So. . Nat Hist. Vol. at.


I















Price list of recent memoirs. 4to.
Vol. V, No. !). The skeletal system of Nccturus maculatus Rafinesque. By
H. H. Wilder, Ph.D. 53 pp., 6 plates. $1.00.
No. 8. Observations on living Brachiopoda. By Edward S. Morse. 73 pp.,
23 pis. $2.00.
No. 7. Description of the human spines showing numerical variation, in the
Warren Museum of the Harvard Medical School. By Thomas Dwight
76 pp. SI.50.
No. 6. The anatomy and development of Cassiopea xamachana. By Robert
Payne Bigelow. 46 pp., 8 plates. $1.40.
No. 5. The development, structure, and affinities of the genus Equisetum. By
Edward C. Jeffrey. 36 pp., 5 plates. $1.00.
No. 4. Localized stages in development in plants and animals. By Robert T.
Jackson. 65 pp., 10 plates. $2.00.
No. 3. Synapta vivipara: A contribution to the morphology of Echinoderms.
By Hubert Lyman Clark. 35 pp., 5 plates. $2.00.
No. 2. Notes on the dissection and brain of the Chimpanzee "Gumbo." By
Thomas Dwight. 21 pp., 4 plates. $1.00.
No. 1. On the reserve cellulose of the seeds of Liliaceae and of some related
orders. By Grace E. Cooley. 29 pp., 6 plates. $1.00.
Vol. IV, No. 14. A bibliography of Vertebrate embryology. By Charles
Sedgwick Minot. 128 pp. $2.50.
No. 13. Fusion of hands. By Thomas Dwight. 14 pp., 2 plates. 75 cts.
No. 12. The insects of the Triassic beds at Fairplay, Colorado. By Samuel H.
Scudder. 16 pp., 2 plates.
No. 11. Illustrations of the Carboniferous Arachnida of North America, of the
orders Anthracomarti and Pedipalpi. By Samuel H. Scudder. 14 pp., 2
plates.
No. 10. New Carboniferous Myriapoda from Illinois. By Samuel H. Scudder.
26 pp., 6 plates.
No. 9. New types of cockroaches from the Carboniferous deposits of the United
States. By Samuel H. Scudder. 16 pp., 2 plates. Nos. 9-12, $3.25.
No. 8. Phylogeny of the Pelecypoda; the Aviculidae and their allies. By
Robert Tracy Jackson. 124 pp., 8 plates. $3.00.
No. 7. The flora of the Kurile Islands. By K. Miyabe. 74 pp., 1 plate. $1.75.
No. 6. The Entomophthoreae of the United States. By Roland Thaxter. 70
pp., 8 plates. $3.50.
No. 5. The Taconic of Georgia and the report on the geology of Vermont. By
Jules Marcou. 28 pp., 1 plate. $1.00.
No. 4. A Study of NorUi American Geraniaceae. By William Trelease. 34
pp., 4 plates. $1.25.
No. 3. The introduction and spread of Pieris rapae in North America, 1860-
1885. By Samuel H. Scudder. 18 pp., 1 plate. 50 cts.
No. 2. The development of the ostrich fern, Onoclea struthiopteris. By Doug-
las H. Campbell. 36 pp., 4 plates. $1.50.
No. 1. The significance of bone structure. By Thomas Dwight. 17 pp., 3
plates. $1.25.



Boston Society of Natural History.
RECENT PUBLICATIONS.
Proceedings. 8vo. (For price list of Memoirs, see third page of cover.)
Vol. 31, No. 2. Proceedings of the Annual Meeting, May 6, 1003. 37 pp.
10 cte.
No. 1. A rare Thalassinid and its larva. By M.T. Thompson. 21 pp., 3 pis.
25 cts.
Vol. 30, No. 7. The life history, the normal fission and the reproductive
organs of Planaria maculata. By W. C. Curtis. 45 pp., 11 plates. 50 ets.
No. 6. Monograph of the Acrasieae. By K. VV. Olive. 68 pp., 4 plates. 25 cts.
No. 5. Proceedings of the Annual Meeting, May 7, 1902. 15pp. 6ets.
No. 4. Memorial of Professor Alpheus Hyatt. 20 pp. 10 cts.
No. 3. The origin of eskers. By W. O. Crosby. 36 pp., 15 cts.
No. 2. The Medford Dike area. By A. W. G. Wilson. 21 pp., 4 plates
25 cts.
No. t. Systematic results of the study of North American land mammals to the
close of the year 1900. By G. S. Miller, Jr., and J. A. G. Rehn. 352 pp.
95 ets.
Vol. 29, No. 18. The Polychaeta of the Puget Sound Region. By H. P. John-
son. 56 pp., 19 plates. 55 cts.
No. 17. Proceedings of the Annual Meeting, May 1, 1901. 33 pp. 10 cts.
No. 16. Bermudan Echinoderms. A report on observations and collections made
in 1899. By U. L. Clark. 7 pp. 5 ets.
No. 15. Echinoderms from Puget Sound: Observations made on the Echino-
derms collected by the parties from Columbia University, in Puget Sound
in 1896 and 1897. By H. L. Clark. 15 pp., 4 plates. 30 cts.
No. 14. Glacial erosion in France, Switzerland and Norway. By William Morris
Davis. 50 pp., 3 plates. 50 ets.
No. 13. The embryonic history of imaginal discs in Melophagus oviuus Ljj
together with an account of the earlier stages in the development of the
insect. By H. S. Pratt. 32 pp., 7 plates. 75 cts.
No. 12. Proceedings of the Annual Meeting, May 2, 1900. 18 pp. 10 cts.
No. 11. A revision of the systematic names employed by writers on the morphol-
ogy of the Acmaeidae. By M. A. Willcox. 6 pp. 10 ets.
No. 10. On a hitherto unrecognized form of blood circulation without capillaries
in the organs of vertebrata. By Charles Sedgwick Minot. 31 pp. 36 cts.
No. 9. The occurrence of fossils in the Roxbury conglomerate. By Henry T.
Burr and Robert E. Burke. 6 pp., 1 plate. 20 cts.
No. 8. The blood vessels of the heart in Carcharias, Raja, and Amia. By
G. H. Parker and F. K. Davis. 16 pp., 3 plates. 25 cts.
No. 7. List of marine mollusca of Coldspring Harbor, Long Island, with
descriptions of one new genus and two new species of Nudibranchs. By
Francis Noyes Balch. 30 pp., 1 plate. 35 cts.
No. 6. The development of Penilia schmackeri Richard. By Mervin T. Sudler.
23 pp., 3 plates. 30 cts.
No. 5. Contributions from the Gray herbarium of Harvard university. New
series, no. 17. 1. Revision of the genus Gymnolomia. 2. Supplementary
notes upon Calea, Tridax, and Mikauia. By B. L. Robinson and J. M.
Greenman. 22 pp. 25 cts.
No 4. Studies in Diptera Cyclorhapha. 1. The Pipunculidae of the United
States. By Garry de N. Hough. 10 pp. 10 cts.
No. 3. Notes on the reptiles and amphibians of Intervale, N. H. By Glover
M. Allen. 13 pp. 15 cts.
No. 2. Variation and sexual selection in man. By Edwin Tenney Brewster.
17 pp. 25 cts.
No. 1. Proceedings of the Annual Meeting, May 8, 1899. 43 pp. 15 cts.
Vol. 28, No. 16. Moniloporidae, a new family of Palaeozoic corals. By Ama-
deus W. Grabau. 16 pp., 4 plates. 25 cts.
No. 15. Studies in the gold-bearing slates of Nova Scotia. By J. Edmund
Woodman. 33 pp., 3 plates. 50 cts.
No. 14. North American wood frogs. By R. H. Howe, Junior. 6 pp. 10 cts.
No. 13. Some Hydroids from Puget Sound. By Gary N. Calkins. 35 pp., 6
plates. 50 cts.
No. 12. The Odonate genus Macrothemis and its allies. By Philip P. Calvert.
32 pp., 2 plates. 50 cts.
No. 11. Proceedings of the Annual Meeting, May 4, 1898. 26 pp. 15 ets.
No. 10. On the veins of the Wolffian bodies in the pig. By Charles Sedgwick
Minot. 10 pp., 1 plate. 25 cts.
No. 9. Notes on a Carboniferous boulder train in eastern Massachusetts. By
Myron L. Fuller. 14 pp. 15 ets.


3 2044 107 170 48