>¥K

~-*3

B

'■ rr"

% >■

-*$

B

_

9

MEMOIRS V

OF THE

NATIONAL ACADEMY OF SCIENCES.

Volume X.

FOURTH MEMOIR.

WASHINGTON:
GOVERNMENT PRINTING OFFICE.

1905.

^I5l:1

MEMO I US

OP THE

NATIONAL ACADEMY OF SCIENCES.

Volume X.

FOURTH MEMOIR.

WASHINGTON:

GOVERNMENT PRINTING OFFICE.
1905.

3 i4 90

PHOBONIS AE( JHITECTA :

ITS LIFE HISTORY, ANATOMY, AND BREEDING HABITS.

WILLIAM KEITH BROOKS, LL. D.,

Henry Walters Professor of Zoology in the Johns Hopkins University,

AND

RHEINART PARKER COWLES, Ph. D.,
Adam T. Bruce Fellow in the Johns Hopkins University.

x\

r

INTRODUCTORY NOTE BY W. K. BROOKS.

As my name appears on the title-page of this Memoir as joint author, I take this opportunity to say that my
own share in the work has been that of instructor and director only. The investigations are the exclusive work of
Dr. K. P. Cowles, and while I have followed them in detail, and hold myself responsible for their soundness and
ace iracy, the credit for the research belongs to Doctor Cowles alone.

Dry Tortugas, Florida, July S, 1905.

72

TABLE OF CONTENTS.

Page.

Introductory note by Prof. W. K. Brooks 72

Introduction 75

Methods ^ 76

Breeding habits 76

Laying of eggs - 77

Fertilization 77

Segmentation 77

( rastrulation and further changes in the form of the larva 71)

Formation of the mesoderm 80

Further growth of the young larva 82

General account - 82

Nephridial pit 84

.Medullary plate - 84

Trunk cavity 84

Fully developed Actinotrocha - 85

Species A 86

Species B 86

Internal organization of the fully developed Actinotrocha 87

Sul meural gland 87

Oral and atrial grooves - - s7

Neuropore 88

Subneural sinus 88

Stomach diverticula - - 88

Notochords - 88

Nervous system 89

Muscular system 92

Body cavities, mesenteries, etc 93

Nephridia - 95

Rudiments of the adult blood vessels in the Actinotrocha 96

Blood corpuscles and their origin 97

Rudiment of the "adult collar cavity" 98

Metamorphosis 98

Preoral lobe and tentacles 99

( ianglion 99

Ectodermal wall of collar 99

Perianal ciliated ring and ectodermal wall of the trunk 99

Cavity of the preoral lobe 99

Mesentery between the lobe and collar 99

Larval collar cavity 99

Adult collar cavity 100

Trunk cavity and cavity of the ventral pouch 100

Ventral and lateral mesenteries 100

Stomach diverticula 100

Digestive areas 100

Nephridia - - 100

Vascular system 101

Adult Phoronis architecta 102

i leiieral account 102

Lophophoral organs 103

Vascular system 105

Nerviis system - — 106

Xephridia 107

Reproductive organs 107

Ciliated ridge of the alimentary canal 107

Summary 107

References 1 1 1

73

PHORONIS ARCHITECTA: ITS LIFE HISTORY, ANATOMY, AND

BREEDING HABITS.

INTRODUCTION.

The study of Phoronis a/rchitecta was begun in the summer of 1901 and continued in the
summer of 1902 at Beaufort, N. C. We are indebted to the Hon. (i. M. Mowers. United States
Commissioner of Fisheries, for the privilege of working in the Commission's station at Beaufort,
where all the conveniences necessary for scientific investigation are at hand; to Prof. H. V.
Wilson, director of the station in 1901, and to Dr. Caswell Grave, director during l!ti)2, for
many kindnesses.

While the study of the live material was for the most part done at Beaufort, the rest of the
work was pursued in the zoological laboratory of the Johns Hopkins University.

Since the discovery of Phoronis hippocrepia by Wright in 1856, the affinities of this inter-
esting genus have been more or less under discussion. Different investigators have sought to
ally the Phoronidse with the Bryozoa, the Brachiopoda, the Sipunculida, and other groups.

Roule (20) thinks that the Phoronida? should be placed next to the Bryozoa in a natural classi-
fication. He does not consider that they have any affinity to the /:'«/< ropru usta, but from a study
of the early stages of development he finds that they are related to the true Ghordata (tunicates
and vertebrates). He says, "lembryon de Vertebre est une Trochophore renversee."

Lankester and Mcintosh are inclined to consider Phoronis, Oephalodiscus, and Rlinhda/ileurn
as related forms, while Harmer (7) makes a comparison of Phoronis with Cephalodiscus and
thinks that perhaps there may be some affinity.

Masterman (15, 16) in a series of papers made a comparison of the actinotrocha larva of
Phoronis with Balanoglossus and its larva and also with Cephalodiscus. In this paper, he arrives
at the conclusion that there is a close genetic relationship between the Phoronidse, Balanoglossus,
and ( 'ephalodiscus. Since the appearance of Masterman's papers, Ikeda (9) has investigated the
development of Phoronis ijimai and has made a careful study of several Actinotrocha found in
Japanese waters. Shortly after this, Longchamps (12) published a comparative study of the
early development of several species of Phoronis and also of several species of Actinotrochse,
giving a very careful critical resume of the work done by different investigators.

Menon (17) has lately published a short paper on the Actinotrochse, in which he considers
the Phoronidse to he related to the Chordatu, hut thinks the relationship is to be traced through
a form like Rhabdoplt »/".

This study of the development and anatomy of Phoronis architecta was begun before the
publication of the last four papers mentioned, and when they appeared the abandonment of this
investigation was seriously considered. However, since there seem to he specific differences and
since there are several disputed points in the development, it seems best to publish the results of
this study.

It is hardly necessary to enter into an historical account of the work that has been done on
the development and anatomy of the Phoronidse, since there are several papers which have
reviewed the subject exhaustively.

75

76 MEMOIRS OF THE NATIONAL ACADEMY OF SCIENCES.

METHODS.

Most of- the material — eggs, larvee, and adults — was fixed in a saturated solution of corrosive
sublimate, to which had been added 2 per cent of glacial acetic acid. A fresh solution was made
as soon as the tine white precipitate appeared, which is usually present in old solutions. This
fibrins- aerent Brave very erood results. Material tixed in Perenyi's fluid was found more valuable
in some few respects than the acetic sublimate. When segmentation stages were treated with a
5 per cent solution of formaldehyde, the blastomeres stood out almost as distinctly as in the
living material. The larger species of the two Actinotrocha found in Beaufort Harbor is much
more active than the other, and when it comes in contact with the fixing fluid the preoral lobe is
bent upward into an unusual position. Consequently a few drops of 4 per cent solution of
muriate of cocaine in 50 per cent alcohol was added to the water containing the Actinotrochse.
After this treatment they died in their usual form when put in the fixing fluid.

Flemming's fluid, as well as the acetic sublimate, was found to be a very valuable fixing
agent for the Actmotroch.se. Heidenhain's iron hematoxylin was used in staining sections of the
adult, and a secondary stain of alcoholic eosin or rubin gave very good results. The most satis-
factory stain for sections of young larvse and Actinotrocha was found to be a solution of safranin
in anilin water. Since it was very desirable to make a study of the adults throughout the year,
and as it was not possible to remain in Beaufort for this purpose during the winter and spring
months, specimens were collected and sent to Johns Hopkins University at different times.
Here they were placed in aquaria filled with sea water, which was kept in good condition by
a rich growth of diatoms on top of a layer of sand. Not only did the diatoms keep the water
from becoming polluted, but they also afforded abundant food for the Phoronis, so that healthy
individuals with their lophophoral tentacles fully expanded were continually at hand for a live
study. The authors are much indebted to Dr. Caswell Grave, the originator of the diatom method
in rearing Echinoo\ rm larvse, for the use of his aquaria. Drew's modification of Patton's method
for embedding and orienting eggs was used with fairly good success, although a large percentage
of the embryos were broken during the process. Most of the embryos were cut into sections 3
p. thick, but for some purposes sections 2 // thick were used.

BREEDING HABITS.

Andrews's (1) observations on Phoronis architecta bring him to the conclusion that either the
sexes are separate in that species or that if the individuals are hermaphroditic the male and
female elements mature at different times. Many specimens examined by us during May, June,
July, August, September, and October, both by means of sections and when alive, showed in no
ease ovaries and testes occurring at the same time in an individual, but ovaries and testes
undoubtedly occur together in the same individual in /'. australis. Benham (2) has observed
this, as we have also, in material sent to us by Mr. Ikeda.

During the month of January the peritoneal tissues surrounding the blood ereea is very
abundant, but as a rule at this time no eggs or spermatozoa are found in it. In one individual
out of some 20 or 30 a few ovarian eggs were found, however. All of these specimens
collected in January were without lophophoral organs, and we kept many of them in aquaria until
the 1st of May. At this time lophophoral organs began to make their appearance in some,
while in others thev were absent. In all the specimens, however, either ovaries or testes were
present, as was also the case in specimens collected at Beaufort in the early part of May.
Further reference will be made to the lophophoral organs and their relation to the breeding
season under the section which deals with the structure of the adult.

The breeding season of Phoronis archik eta extends from March or April to November or
December. Ikeda (9) has stated that "the breeding season of Phoronis ijimai ranges through
about half of the year, say from November to June or July." There seems to be a surprising
difference in the time of breeding between these two species. The Actinotrocha at Beaufort are
found throughout the summer and autumn, but they arc especially abundant during August and
September. Ikeda has suggested that Phoronis annually " changes its generation." It does not

MEMOIRS OF THE NATIONAL ACADEMY OF SCIENCES. 77

seem probable that this is the case for Phoronis architecta^ because full-sized adults arc found
throughout the year in Beaufort Harbor, and specimens were kept alive for fifteen months in the
laboratory of Johns Hopkins University.

THE LAYING OF THE EGGS.

During low tide in the summer and autumn it was easy to collect from loo to L50 specimens
of P/ioronis architecta during an hour or two. About one-half of these would usually have
male reproductive organs and the rest female reproductive organs. The Phoronis were placed
in glass crystallizing dishes and after about twenty -four hours many of the individuals began to
lay— usually at night — but the eggs were not retained among the tentacles in a mass, as described
by most investigators, hut were swept gently away from the lophophoral crown by the ciliation
on the tentacles and on the anal region, .so that they settled near by on the bottom of the dish.
Sometimes, however, the newly laid eggs were carried up and down the tentacles in currents
caused by the cilia, and occasionally a few eggs were found grouped near the tips of the tentacles,
being held there loosely by a small quantity of mucus-like material. At no time, however, were
eggs and larva' aggregated in definite masses, as described by Ikeda (9), nor were they brooded
among the tentacles, as Masterman (16) has observed in the case of Phoronis iusMi. That eggs
and embryos were not found by Longchamps among the tentacles of ''Phoronis d II> Igoland" is no
doubt due to the fact that the same habit prevails in the above form that does in Phoronis
architecta.

While the adults were laying, they were examined under the compound microscope. They
showed large numbers of eggs which were floating freely back and forth in the body cavity as
the animal contracted and expanded. Sections of adults in this condition show that all these free
eggs contained the first polar body spindle. At intervals of about one minute an egg is extruded
with considerable force from the nephridial opening, and in no case do the eggs at this moment
have polar bodies. The wall of the nephridial ridge is transparent enough to see the eggs
as they slip through the larger part of the nephridium. While passing through, they are
pressed by the walls of the organ until they are about twice as long as broad (tig. 1). The fact
that Phoronis architecta does not keep its eggs in masses within the tentacular crown, together
with the fact that most of the individuals lay them at about the same time at night, makes
it possible to preserve any one stage in the development of the embryo in sufficient quantity for
a thorough study.

FERTILIZATION.

Ikeda (9) and Longchamps (12) made the observation that the eggs in the body cavity of the
parent showed the spindles of the first polar body. This I found to be the case in Phoronis
architecta (tig. 1). Eggs in the nephridia were found to be in the same stage, and in neither case
was there any sign of an entering spermatozoon or a male pronucleus (tig. 1). There is no doubt
of the fact that in Phonmix ///■r/u'tecfa the spermatozoon does not enter the vgg until the latter
has been expelled from the nephridium. Ikeda observed this fact for Phoronis ijimai.

SEGMENTATION.

The eggs of Phoronis architecta while still in the body cavity are somewhat irregular in
shape, and. as mentioned above, are decidedly so while passing through the nephridium.
However, after they are laid they become almost perfectly spherical and average loo /< jn diam-
eter ("tig. L>). thus measuring the same as the egg of "P/toronis <h Naples." (Longchamps (12).)
The egg is very opaque, being heavily laden with small yolk granules. It is surrounded by a
delicate membrane, which, however, is not very conspicuous, being closely applied to the sur-
face, but after fertilization it separates to some extent (tig. •!).

Observations on the segmentation of the egg of Phoronis are conflicting. This part of the
development of Phoronis seems to have been treated hastily by most observers, probably
because it is difficult to obtain sufficient material for its study. It is agreed that the segmenta-
tion is total. Foettinger (5) and F. Schultz C2L) claim that the segmentation is unequal.

78 MEMOIRS OF THE NATIONAL ACADEMY OF SCIENCES.

Caldwell (3) says that in the four-cell stage two smaller clear and two larger opaque cells are
present. Masterman ( L6) finds that in the four-cell stage the blastomeres taper toward one pole,
and that this results, when the third furrow appears, in the upper four being less in hulk than
the lower four. Masterman's description, 1 find, applies to the eggs of Phoronis amtralis.
Ikeda (9) did not discover any appreciable difference in the size of the blastomeres until the
eight-cell stage. At this time, he says, "it will be seen that the upper four blastomeres are
very slightly smaller than the lower four."

In Phoronis architecta the first cleavage plane is meridional and usually divides the egg into
two practically equal blastomeres (tig. 3), although sometimes the division is decidedly unequal
(tig. 4). The cleavage furrow begins in the region of the polar bodies (tig. 5). After the com-
pletion of the first cleavage and sometimes before, the first polar body divides (fig. 6). In fig. 6
is seen the reconstruction of the nuclei after the first cleavage. Immediately before the second
cleavage the two blastomeres, which were closely applied to one another after the first cleavage,
come to overlap. About fifteen minutes after the first cleavage the second cleavage takes place.
It is meridional and at right angles to the first, dividing the two equal blastomeres into four
equal blastomeres. As Ikeda has observed, the cleavage does not occur simultaneously in both
blastomeres nor does it in later cleavages (tig. 7). The blastomeres of the four-cell stage which
at first overlapped soon become applied to one another so that the two meet in a cross furrow
(fig. 8). Shortly before the third cleavage occurs the cross furrow disappears and the blasto-
meres come to overlap. The third cleavage takes place fifteen minutes after the second cleavage,
and it is equatorial. The blastomeres become drawn out into a more or less ovoid shape and, as
division takes place, the upper four blastomeres become rotated in the direction of the hands of
a watch (tig. !»). The eight blastomeres are approximately the same size, as a rule, and there is a
small segmentation cavity present which from now on persists (tig. 10). The three polar bodies
arc distinguishable at this stage sometimes within the blastoeoele and sometimes on the surface
of the blastomeres. The blastoeoele is open at the animal and vegetal poles. The sixteen-cell
stage arises from the eight-cell stage by a meridional division of each of its blastomeres. but
they do not all divide simultaneously (tig. II). although the difference in time is very slight.
After the sixteen-cell stage the individual blastomeres were not followed. The division takes
place rather irregularly, but the blastomeres are all about of the same size.

The so-called •'blastoeoele pore," observed by Ikeda (9), was found occasionally in young
blastuhe, but it does not seem to be of constant occurrence nor definite in position {dgx. 12, 13).

Two hours after the first cleavage the blastula is composed of seventy or eighty cells, and it
is still inclosed in the ego membrane. Four hours later the membrane disappears, and the
ciliated blastula begins to swim (tig. 11).

The blastomeres were so much alike and so uniform in size that their individual history was
not traced. It would seem probable from Masterman's work on Phoronis bushii (16) that the
cell lineage might be followed in that form, for he finds considerable difference in the size of
the blastomeres in the early stages of cleavage at least.

The apical pole of the ciliated blastula is provided with long cilia (tig. 14). The nuclei are
situated nearer the outer than the inner surface, and the inner ends of the cells are filled with
rather dens,, granules. In the segmentation cavity are found the so-called "corpuscles." which
have been observed by most investigators working on the early stages of Phoronis. Caldwell's
(3a) view that they are not mesoderm cells is undoubtedly correct. Our observations agree with
those of Ikeda (!)), for the "plasmic corpuscles" are much smaller than any of the cells of the
blastula. and none with nuclei were found. In Phoronis architecta they do not appear until the
late blastula stage (tig. 11), at which time the inner ends of the cells are densely granular. It
seems very probable that the corpuscles are pushed out from the densely granular part of the
cell, and that, as Caldwell and Ikeda have held, they are an extra supply of nourishment.

The blastuhe. gastruhe. and young larva' of Phoronis architecta are quite similar in appear-
ance to those of I 'Imr, mis d? Helgoland which Longchamps (12) has figured. The development is
more regular than that of most other species, which is probably due to the fact that the eggs
and embryos are not harbored in the tentacular crown.

MEMOIRS OF THE NATIONAL ACADEMY OF SCIENCES. 79

GASTRULATION AND FURTHER CHANGES IX THE FORM OF THE LARVA.

In the blastula, which has just begun to invaginate, the invagination is eccentric, thus giv-
ing the firal indication of the bilateral symmetry of the larva. This is further emphasized in
the young gastrula of Phoronis architecta by a thickening of the ectoderm cells, which becomes
the ganglion of the Actinotrocha. The cells composing this thickening of the ectoderm are
at the apical pole in the blastula and they bear long cilia, but as gastrulation takes place and
as the embryo elongates the thickening comes to occupy a position nearer the anterior end of
the larva (figs. 15, L5a, 155). The published accounts seem to indicate that the ganglion
makes its appearance much later in most specie- than it does in Phoronis architecta, although
Koule (20) figures the "plaque cephalique" at a rather early stage in Phoronis sahatieri. The
changes which take place in the shape of the blastopore are much like those described by
other investigators. At the beginning of gastrulation the blastopore is wide open and circular in
outline. The lateral lips of the blastopore then gradually draw together in the posterior region,
inclosing that part of the wall of the archenteron between them (fig. 20c), which becomes a
solid mass of cells continuous with the cells of the fused lips of the blastopore (figs. 18<2, 18c).
The cells of the solid mass are of the same character as those of the wall of the archenteron
where the blastopore is open. They are quite granular, except at the periphery, where they
project into the cavity of the blastoccele, and their nuclei are quite indistinct. Both of these
facts are characteristic of the cells, making up the archenteric wall (fig. 18c).

As a result of the closing up of the blastopore posteriorly, the blastopore becomes oval in
shape, and an indication of a ventral furrow, first observed by Caldwell and called a •'primitive
groove.-* appears. In l'}i<>r<>n'ix architecta this groove is only to be seen in one or two sections
back of the blastopore, after which the ventral surface is convex (tigs. 18<2-20c). A "primitive
streak," as described by Caldwell (3a) could not be made out. The gastrula, which is at first
circular in horizontal section, becomes slightly elongated when the blastopore takes on aivoval
shape.

Gradually the blastopore lips close up more anteriorly until the blastopore becomes circular
in outline but much smaller than it was originally. At the same time the anterior end of the
larva begins to bend in a ventral direction and the archenteron becomes elongated posteriorly
(fig. 20).

Xow the larva increase- slightly in length (tig. 21), the blastopore assumes the form of a
transverse slit, the anterior end bends farther ventrally, and the posterior end of the enteron
becomes applied to the ectoderm at the posterior end of the larva (tig. -11).

Our observations on Phoronis architecta agree with the description of Masterman (Hi). Ikeda
(9), and Longchamps ( 12) in regard to the closure of the lips of the blastopore and the resulting
change in the shape of the latter, but in Phoronis <ir<-]iit,ct<i the definitive blastopore does not
seem to be pushed farther anteriorly by the special activity in the posterior region of the blast-
opore, as Ikeda has found to be the case for Phoronis igimai. The definitive blastopore seems
to be represented by the anterior part of the wide circular blastopore of the young gastrula. Its
change in position with reference to the anterior end is due to an elongation of the posterior
portion of the embryo and the ventral flexure of its anterior end.

Our studies on the development of Phoronis architecta lead us to agree with Ikeda and
Longchamps as to the fate of the cell- in the posterior part of the blastopore and as to the ecto-
dermal origin of the " posterior pit." The cells of the posterior part of the blastopore become
invaginated by the closure of the blastopore lips and form part of the ventral wall of the enteron.
while the '" posterior pit." which appears in Phoronis architecta shortly before the definitive
blastopore is formed, is of ectodermal origin and has no apparent relation to the ventral groove.
A- Ikeda (9) has stated, tin' pit is the beginning of the nephridium of the Actinotrocha. In a
recent paper Masterman ( L6a) says that he has found the " posterior diverticulum" both in larva'
of /'. buskii and /'. hippocrepia and that he considers them to be the anlagen of the nephridia.

80 MEMOIRS OF THE NATIONAL ACADEMY OF SCIENCES.

FORMATION OF THE MESODERM.

There is considerable difference of opinion among those who have investigated the embry-
ology of Phorojiis as to the origin of the mesoderm, and there seem to be no two whose descrip-
tions agree, although Ikeda (9) and Longchamps (12), in their recent papers, arrive at the same
conclusions, generally speaking.

The study of the eggs and larva' of Phoronis architecta and those of Phoronis australis
show that the great difference in the origin of the mesoderm, as Roule (20) and Masterman (16)
sec it, may lie due. in a great part, to difference in the larvae themselves.

The eggs and embryos of Phoronis mi si rolls, for which we are indebted to Mr. Ikeda,
arc very similar in appearance to those of Phoronis buskii, judging from Masterman's figures
(l(i). Sections of the eggs and larvae of the former show the development to be of the same
general type as that of J 'homo Is ljlno/1. which Ikeda (9) has described.

The eggs and larva? of Phoronis architecta are considerably different from those mentioned
above. They are more regular in form, the blastoccele is much more spacious and the cells
themselves are more regular in shape and arrangement. They are most similar in appearance to
the early stages of Phoronis sahatieri studied and figured by Roule (20) and those of "Phoronis
oV Helgoland" figured by Longchamps (12).

The formation of the mesoderm begins in Phoronis architecta as soon as the flattened side of
the blastula begins to gastrulate. In a few cases round blastulee are found, within the blastoccele
of which are rather large granular spherical bodies much larger than the plasmic corpuscles
described above. Each of these contains an opaque body which takes saffranin stain with readi-
ness. A comparison of these bodies with the nuclei of the cells of the blastula wall convinces
one at once that they are not nuclei. In tig. 1-ta a section through such a blastula is shown in
which these bodies are seen within the wall of the blastula as wTell as inside the blastoccele. They
are embedded in the wall without reference to the limits of the cell and usually occupy the width
of two cells. The cells inclosing these peculiar bodies do not differ from the cells surrounding
them in that region and each has its own nucleus. These bodies are not the cut ends of amoeboid
processes which Caldwell (3a) and Roule (20) observed, for such processes do not occur in the
blastuhe of Phoronis architecta. We are unable to make any positive statement as to their fate,
but it is very probable that they break up into the smaller plasmic corpuscles. Such bodies as
the former might easily be mistaken for mesoderm cells, and we suspect that the "mesoderm
cells" observed by Foettinger (5), Metschnikoff (18), and E. Schultz (21) in the round blastula1
were of the same character.

The work on Phoronis architecta indicates that the mesoderm which forms the lining of the
preoral lobe and the collar cavities of the Actinotrocha arises from the lips of the blastopore.
As to the origin of the lining of the trunk segment, we are still in some doubt, but we are
inclined toward Longchamps's suggestion that some of the cells of the nephridial pit give rise to
it. Caldwell (3a) also holds that the mesoderm arises from the endoderm, assuming that the
"posterior pit" (nephridial diverticulum), which he considers to be one point of origin of the
mesoderm, is of endodermal origin. Roule (20) derives most of the mesoderm from the endo-
derm, but also considers the " bandelettes mesoltlastiques," which Schultz (21) first pointed out to
be the same as the posterior diverticulum of Caldwell, as giving rise to mesoderm.

As is seen on referring to tig. 15, the flattened part of the wall of the blastula has become
more than one cell thick. In fact active cell division lias taken place. Yet most of these cells
are destined to become the wall of the archenteron, and only a few are to give rise to mesoderm.
Careful examination of many sections fails to show that mesoderm cells ever have their origin
from the dorsal surface of the archenteron. In this respect the development of Phoronis archi-
tecta seems to agree with that of Phoronis Icowal&oskii as described by Caldwell (3a) and Long-
champs (12). and that of Phoronis buskii, which Masterman (1(3) investigated. In Phoronis
architecta, as in the form studied by Longchamps (12), the anterior and lateral borders of the
blastopore are most active in giving rise to mesoderm (tigs. 16 <i, b, c, d, e,f). Most of it is

MEMOIRS OF THE NATIONAL ACADEMY OF SCIENCES. 81

proliferated from the anterior end of the archenteron, and the power "f producing the same
seems to diminish gradually toward the posterior end of the blastopore lips in those gastrula
where the round blastopore is just beginning to close up (figs. l»i </. I>. e, d, e,f).

The mesoderm cells are \ ei'y amoeboid in character and are often seen in living specimens and
sometimes in sections sending out long pseudopod-like prolongations, which become attached to
the walls of the gastrula. l'>\ means of these amoeboid movements thej are able to crawl up the
walls of the blastocoele.

We were unable to make out any structure in Phoronis architecta which could be interpreted
as "archenteric diverticula," such as figured by Caldwell (3a) and [keda (9). Fig. 16c might be
interpreted as showing these diverticula, but the condition there is hardly different from the
arrangement of the mesoderm cells, which are being pushed out into the blastocoele in front of
the blastopore (figs. L6, !•'>/'). Caldwell first observed these structures in the gastrulse of Pho-
ronis kowalevskii, hut Lone-champs (12), who has recently carefully' studied the same species, has
been unable to rind them. Ikeda (9), however, rinds very definite diverticula in the gastrulse of
Phoronis ijimai, but he figures them as being in the region of the blastopore, while, according
to Caldwell's (3a) figures, they are found posterior to the blastopore.

Let us return again to the mesoderm cells which lie anteriorly to the blastopore. These
amoeboid cells undoubtedly multiply while in the blastocoele, and in a gastrula where the blasto-
pore lips have closed up somewhat so as to give an oval outline to the blastopore (tie-. 18/) these
cells have become arranged into a definite sac (figs. 19, 20a), which i- later to form the lining of
the preoral lobe. In no case were we able to find the least indication of an anterior unpaired diver-
ticulum, which Masterman (Hi) says exists in the gastrula of l'h<>r<>nis l>uskii. At this stage
the cavity of the sac is small and is present only in front of the blastopore. The walls, however,
are extended on each side into a lateral cord of mesoderm cells, which lies in the blastocoele at the
side of the blastopore (fig. is/'). Some of the cells of the dorsal wall of the sac send out pseu-
dopodia, which attach themselves to an ectodermal thickening, and this thickening will become
the ganglion of the Actinotrocha (tig. L9).

The above condition continues until the oval blastopore becomes smaller and round in outline
(figs. 20, 20< ). which change is also accompanied by further growth of the enteron in a posterior
direction until it almost touches the end of the larva. The cells of the two lateral cords of
mesoderm have now increased in number, have arranged themselves so a- to inclose a cavity,
continuous with the cavity of the anterior one described above, and have become attached both
to the lateral ectodermal wall and the lateral endodermal wall (tig. 205). Anteriorly this sac,
which is now horseshoe shape (tig. 20e), is still only attached to the ganglionic thickening and
the ventral ectodermal wall (tig. 20). The conditions just described are not due to the shrinkage
of the mesodermal lining away from the wall of the larva, for the transparency of the living
larva makes it possible to see the formation of the mesodermal sac. We have followed this
formation step by step many times in the living gastrula and larva, as well as in sections and
surface mounts.

As tlie anterior end of the larva bends farther ventrally and becomes a definite preoral lobe,
the round blastopore assumes the shape of an oval with its major axis transverse to the long axis
of the larva (tig. -_'l). The posterior part of the larva increases in length and the enteron sends
out a posterior diverticulum, the beginning of the intestinal canal, whose blind end fuses with
the ectoderm of the posterior end of the larva. The walls of the mesodermal sac become applied
to the walls of the preoral lobe more generally, thus forming a definite mesodermal epithelium
lor the cavity of the preoral lobe digs. •_':;. 21, 22). Posteriorly, as Masterman ( 1."-) ha- described
for the fully developed Actinotrocha, the cavity is produced "■into two horn- running back
laterally" (rig. 22a), but as yet there is no complete mesodermal lining in the cavity back of this
(figs. ■_'■_' I. c). The posterior wall of the mesodermal lining of the preoral lobe forms a definite
septum (figs. 21, 21a), but it is not as yet, at least, composed of two layer-, a- Masterman finds
in the older Actinotrocha.

82 MEMOIRS OF THE NATIONAL ACADEMY OF SCIENCES.

While the above changes have been taking place in the preoral end of the larva there has
also been some change in the postoral region. It is seen from figs. 16 c, d, e, f that in the
young gastrula with the large circular blastopore mesoderm cells are being pushed out into
the blastoccele along both sides of the archenteric "wall back almost to the posterior border of
the blastopore. At this stage large spherical cells with rather small, deeply staining nuclei
are sometimes seen floating freely in the blastoccele (lie;. 17"). These cells have their origin
in the wall of the archenteron (tig. 17) and are quite different from the bodies found in the
blastoccele of the blastula. They have, a definite nucleus and they seem to be similar cells to
those found by Ikeda in the larva \yith one pair of tentacles. - They certainly do resemble the
Mood corpuscles found in the older larvae, only they are considerably larger. Ikeda (9) came to
the conclusion that these cells were the " mother cells of blood corpuscles which are found as
corpuscle masses in the collar cavity of the Actinotrocha." Since the publication of his paper
Mr. Ikeda has written that he considers his theory concerning the. fate of these cells to be
incorrect. They are easily distinguishable from all other cells by the fact that they are larger
and that the cytoplasm does not stain. They have a nucleus which is rather small. We shall
return to a consideration of these cells when we describe the blood corpuscles of the Actinotrocha.

As the blastopore lips begin to close up posteriorly (tigs. Is ,/. e) the endoderm cells in that
region lose the power of giving rise to mesoderm cells, but they are still found arising in a more
anterior region.

At a little later stage, in which the blastopore has become circular again after the fusion of
the blastopore lips and the enteron has almost reached the posterior end, a few mesoderm cells
are seen lining the ventral ectoderm in the posterior region (fig. 20a*). These cells, however,
do not have their origin from the wall of the posterior part of the enteron nor from the ventral
ectoderm which Caldwell (3a) would call the "primitive streak." The cells forming the ventral
ectoderm are very regularly arranged into a layer one cell thick and all the nuclei are in a resting
state. The mesoderm cells have either migrated from the cells of the lateral cords which are
prolongations of the sac, forming the lining of the preoral lobe, (tig. '2Qe), or from the region of
the blastopore, where some mesoderm cells are still arising. In general, our interpretation of
the fasts bearing on the origin of the mesoderm in the posterior region of the larva agrees with
that of Longehamps (12) for Phoronis kowalevskii.

When the larva reaches the stage shown in tig. 21 where the blastopore is transverse and
the archenteron fuzes with the posterior ectoderm, the mesoderm cells are found to be more
numerous in the posterior region, and in nearly all cases they are applied to the ventral suface
of the blastoccele. At this time the proliferation of mesoderm cells from the endoderm has
ceased in the anterior region and there is no indication of any mesoderm cells being given off
from most of the posterior region. At the extreme posterior end of the enteron. however, a
transverse section across the larva (fig. 22d) show- a mass of cells which might be taken for
proliferating mesoderm cells. Traced farther back, this mass of cells is found to be part of the
wall of the. "posterior pit," or, as Ikeda (9) has called it, ''the nephridial pit" (figs. 22 e,f, </).
The fate of the cells of the nephridial pit will be discussed in the description of the larva with
two tentacles.

Lar\;e like the one just described do not show the least trace of a mesentery between the
collar and trunk. In fact, one could hardly say that a trunk existed at this time. The oblique
strongly ciliated tract of ectoderm which indicates the line of origin of the larval tentacles has
not appeared.

FURTHER GROWTH OF THE YOUNG LARVA.

The flexure of the preoral lobe continues as the larva grows older (fig. 24). In this way a
vestibule is formed and the original blastopore becomes the part which connects the vestibule
and archenteron (fig. 24). This relation between the vestibule and blastopore has been recog-
nized by Masterman (If.), Roule (20), Ikeda (9), and Longehamps (12). Longehamps speaks of it
as a "stoinodseum," and Masterman does also, but the latter adds "oesophagus" after it. (If our

MEMOIES OF THE NATIONAL ACADEMY OF SCIENCES. 83

idea thai a stomodajum is a pitting in of the ectoderm which finally breaks through into the
enteric cavity is correct, then Masterman's and Longchamps's use of the word is incorrect.)

Masterman speaks of a "slight ridge" running around the edge of the preoral hood and
then "(low awards (ill it is lost on the surface of the tentacles." Such a ridge is not present in
the larva of Pharonis architecta and there i- no connection between the ciliated tract alone- the
line of w Inch the larval tentacles arise and the ciliated edge of the preoral hood.

At this stage (tie'. 24) a definite intestinal canal i- seen which, however, does not open yet to
the exterior. The intestine, as described above, is not of ectodermal origin in the larva of Pho-
ronis architecta. There is no proctodeum. On this point our observations agree with those of
Masterman, Longchamps, and Ikeda.

Roule (20) says: •"1*11 anus et rectum se faconnent, aux depens de I'ectoderm, sur I'extremite'
posterieure du corps" (p. L02), and Caldwell (3a) derives the intestine from the remains of the
"primitive streak."

As vet the anal papilla is not at all definite, but the ciliated hand along which the larval
tentacles are to arise has now appeared. This is indicated in the sagittal section (tie-. 24) by a
thickening of the ectoderm.

The mesoderm cells which in tie-. 21 are seen applied to the ventral ectoderm of the larva
have now increased considerably in number and have become arranged at quite definite intervals
(fie-. 24). If the ventral surface of the larva is examined, while the larva is alive, it will he seen
that these cells have become simple muscle cells made up of two rather delicate fibres which
extend from a large nucleus situated near the mid-ventral line. These fibres run parallel to one
another around the wall of the larva (fig. 25).

The whole body cavity hack of the mesentery between the cavities of the collar and lobe
represents the larval collar cavity of the Actinotrocha, and although its somatic walls are not
lined by a perfectly continuous mesodermal epithelium, yef there are indications that such a
lining i- being formed. The ventral and lateral walls of the stomach, however, are perfectly
free from any epithelial covering. In fact, in all the Actinotrocha- examined no mesodermal
epithelium covering the ventral and lateral walls of the stomach in the collar region could be
found. We have never seen any sign of mesodermal sac-like formation such as occurs in the
preoral lobe.

Roule (20, p. 112) ha- described in a considerably older larva than the one with which we
are dealing certain mesodermal cells to which he has given the name "conjunctivo-musculaircs
elements." These he represents as spindle-shaped cells terminated by lone- fibre-like prolonga-
tions and he has figured them as being quite numerous in the ••plasma, transparent et con-
sistant," of the coelomic cavity. While the young larva of Phoronis architecta bears a close
resemblance to that of Phoronis sdbatieri described by Roule (20), yet at no time during the
life of the larva have we seen these cells suspended in the body cavity in such numbers as he has
shown. Spindle-shaped cells with long prolongations are quite numerous, but they are usually
found applied to the somatic walls of the larva.

Although recent investigators have thrown some doubt on the existence of tin- lobe-collar
septum, yel such a septum unquestionably exists in the larva of Plwronis architecta. Ikeda
(9) has shown that it i-- incomplete in the old actinotrocha and our observations agree with his.
but it is a fact, nevertheless, that the septum is continually present throughout the larval life of
Phoronis architecta and that it makes its appearance at a very early stage in the life history.

Longchamps (12) says: " Si une subdivisions plus on moins complete s'etablissait, entre ces
deux regiones, elle ne seriat en tout cas one secondaire, et la cloison s'edifierait aux depends de
mesenchyme.'" It is plain from what has been said that we can not agree with Longchamps
in his statement that the septum is secondary. It must be admitted, however, that the septum
between the lobe anil the collar i- often considerably thinner than that between the collar and
the trunk. If its origin had nol been followed from the earliest stages by mean- of section-
in three diti'erent planes, whole mounts and live material, we should not have been inclined to
consider it a primary and constant organ of the larva.

84 MEMOIRS OF THE NATIONAL ACADEMY OF SCIENCES.

We had confidently expected to find in the larva of the .stage represented in tig. 21 some
indication of the mesentery between the collar and the trunk. but were disappointed. Although
the ciliated tract along which the tentacles are to arise has made its appearance and extends
obliquely around the body proper of the larva, indicating a line just a little above the line of
separation of the cavities of the trunk and collar in the older actinotrocha, yet examination
of sections does not show the least sign of the mesentery or its fundament.

Ikeda (9) has found mesoderm cells connecting the nephridial canals with the splanchnic
walls, and he thinks they are the first indication of the septum between the collar and trunk
cavities. We have not been able to find the paired masses of ''hypoblast cells'1 which
Masterman (16) says give rise to the trunk cavities, although it is true that mesoderm cells
were found lying on the dorsal wall of the intestine. Masterman does not follow the fate of
those masses, but in the larva with three pairs of tentacles he speaks of "the two mesocceles
pushing dorsallv, their walls forming a pair of conspicuous mesenteries with the walls of the
' metacceles" (p. 395).

As has been said, the masses of cells that Masterman (16) speaks of have not been found,
but we do not deny that the condition which he describes may exist in the larva of Phoronis
hnsl.ii with three pairs of tentacles.

" Nephridial pit" (Ikeda).— A structure of the larva of Phoronis which has given rise to
considerable controversy is the "nephridial pit" ("posterior, anal, ectoblastic pit," "posterior
diverticulum"). It seems safe to assume that such a structure exists in the young larvae of all
species of Phoronis. It has been seen by Caldwell (3a) in /'. howalt vsMi, by Ikeda (9) in /'. ijimai,
by Longehamps (12.) in P. himili mkii, by Masterman (16«) in /'. buskii and /'. hippocrepia, and by
us in /'. architecta. Although Roule (20) has not observed the pit in J', sabatieri, it seems
probable that such a structure will be found there on further investigation. It is difficult to
believe that Roule's understanding of the origin of the nephridia from two cell masses of
somatopleure symmetrically placed at the sides of the larva is correct. As stated above, we
consider the pit to be of ectodermal origin (tig. 21). Further study of the structure leads us
to agree with Ikeda that it divides into two lateral branches, each of which becomes a nephridial
canal of the actinotrocha (tigs. 25, 26, 27, 28). In tig. 25, which is a drawing made from a living
larva, the canals, which in a little younger stage were practically the same diameter throughout
their length, have become tipped at their distal ends with a bunch of cells which, we believe,
are later to form the excretory cells of the nephridium.

No positive statement concerning the origin of these cells can be made, since it is difficult
to obtain many larvae of /'. architecta which are old enough to show these bunches of cells in the
process of formation. The sections examined afford no evidence that they are formed by free
mesoderm cells attaching themselves to the internal ends of the nephridial canals, and we are
rather inclined to consider them as arising from the cells of the internal blind ends of the
nephridial canals (tigs. 25-28). Ikeda's description (9) of the way the nephridial tubes arise
from the ectodermal pit — i. e.. by the " reexamination of the distal unpaired portion of the
nephridia] pit" — seems to be correct. Longchamps's (12) interpretation of the change in form of
the "ectodermal pit" agrees quite closely with Ikeda's description.

"Medullary />/</t< " (Roule). — When the young larva of Phoronis architecta has reached the
two-tentacle stage cross sections show that there is a definite ventral ciliated band extending from
the mouth to the ciliated tentacular band (tigs. 2'.), /-. c, </.,)• This ventral ciliated band has
been observed by Roule (20) in the larva of /'. sabatieri, but it has not been described for any
other species. Koule has homologized it with the medullary plate or medullary groove of the
annelid larva.

" Trunk cavity" — Longehamps has drawn attention to a figure of an Actinotrocha published
in Ilatschek's" Lehrbuch der Zoologie." In this figure are represented two ccelomic sacs sur-
rounding the intestine, llatsehek does not describe the origin of these sacs, but Longehamps
ip. .">.">.'») proposes the question, "Si les canaux ne derivaient pas des expansions laterals du
divertieule ectoblastique, ehacun des canaux restant en rapport avec l'exterieur par mi orifice

MEMOIRS OF THE NATIONAL A.CADEMY OF SCIENCES. 85

resultant du dedoublemenl de I'orifice primitif el median, tandis que le restant du diverticule
ectoblastique deviendrait la cavity post^rieure du corps."

Such an origin for this posterior cavity would seem to be a possible one and would give an
easy explanation for the origin of the collar trunk and ventral mesenteries.

We believe thai the cavity of the (funk is formed in the following manner: As the tentacles
grow out and increase in number the posterior region of the larva about the rectum increases
greatly in length. In doing the latter the mesodermal lining of the collar is drawn away from
the somatic wall in the region hack' of the tentacular hand, and a cavity is left containing the rec-
tum, part of the stomach, and the proximal part of the nephridial diverticula. At the same time
(his is taking place certain cells which seem to arise from the base of the nephridial diverticula
give rise to the lining of the cavity of the trunk. As to the manner of origin of these cells
we are still in doubt. We have not found two coelomic sacs which Hatschek (8) seems to have
figured (it is possible that his figure is meant to represent a single sac cut at two places),
and we have hunted for them in larva' where the diverticula are just beginning to form and
also in larvae with two, four, and six tentacles. In one specimen with two tentacles, however
(fig. 30), an arrangement of mesodermal cells on the dorsal side of the intestine which seems
to he the beginning of a sac is found; this, however, is not paired. Whether or not this sac
and its cavity eive rise to the lining and cavity of the trunk we can not say, for we have found
hut one specimen in which this condition exists.

One thing is certain, the fully developed trunk cavity of the Actlnotroc/ut has a distinct
mesodermal lining, consisting of a somatic and a splanchnic layer. As far as we know all Actino-
trochst have a ventral mesentery, which tends to support the view that the lining of the cavity
of the trunk has its origin in a sac which grows around the rectum and posterior part of the
stomach. Whether or not the fact that there is an indication of a dorsal mesentery in the pos-
terior region of some of the fully developed Actinotrochse, Species B., has any bearing on the
double origin of the cavity of the trunk we can not say. for we have never seen the very young
larva' of this form.

The youngest larva taken from the tow had three pairs of tentacles, with beginnings of the
fourth pair. In this larva the tentacles had grown considerably in length, and the posterior
region had become somewhat elongated (tig. 31). Only one specimen of this age was obtained,
and it was only studied while alive. The mesentery between the collar and lobe was plainly
seen, and there seemed to be a thin mesentery between the region of the collar and the younger
trunk region. The nephridial canals were seen with difficulty, but the rounded bunches of
excretory cells forming the internal ends of the canals were plainly visible.

When the larva of Phoronis wrchitecta has five pairs of tentacles (fig. 32) the trunk region is
elongated considerably and constitutes about one-half the length of the larva. In fact, the larva
at this stage looks much like the fully developed Actinotrocha. The ■'retractors" described by
Lkeda (9) are now present and the body wall in the anal region shows a thickening which is to
become the perianal ciliated band. This larva shows clearly the presence of two mesenteries.

In the larva with six pairs of tentacles (fig. 33) all of the organs of the fully developed
Actinotrocha are present. The ventral pouch begins to invaginate (tig. 33) and sections usually
show that the blood corpuscle masses are forming.

FULLY DEVELOPED ACTINOTROCHA.

There are two species of Actinotrocha found in the waters of Beaufort Harbor, and they are
very similar, if not identical, with the two species that E. B. Wilson (24) observed in Chesapeake
Bay. l'n mi the latter part of May until the latter part of September both species are fairly
abundant in the tow7.

Wilson has designated the two species found in Chesapeake Hay as Species A. and Species
B., anfl because of the general agreement between our observations and his descriptions the
Beaufort Actinotrochse will be designated as Species A. and B., although we are satisfied that
Species A. is the larva of P. architecia.
Vol. 10— No. 4—05 2

86 MEMOIRS OF THE NATIONAL ACADEMY OF SCIENCES.

Species A. — Species A. (fig. 34) is somewhat smaller than Species B. and its average length is
L.03 mm. The trunk is quite stout, the intestine is short, and the posterior end of the stomach
reaches as far as two-thirds of the length of the trunk cavity. When about ready to metamor-
phose, this larva usually has 18 larval tentacles and an equal number of young adult tentacles.
The adult tentacles do not usually appear until the larva has 18 larval tentacles (its full number)
and they arise as thickenings on the under side of the bases of the larval tentacles. In this
respect the larva resembles one of the actinotrochfe which Ikeda (!<) has described. The blood
corpuscles are found in two masses usually applied to the ventro-latcral surface of the stomach,
and they make their appearance in the larva with 12 or 14 tentacles. A pair of muscles which
[keda has been the first to describe, and which he has called "retractor muscles," are always
present; although they have not been made out in younger larva' than those with 10 tentacles.
This species is without the so-called "'stomach diverticula." Pigment cells are found rather
irregularly scattered on the wall of the body cavity. There are definite aggregations of these, at
the bases of the tentacles, and a few pigment cells are seen on the surface of the blood corpuscle
masses. Usually there are quite a number in the wall of the posterior portion of the trunk.

This Actinotrocha is not as active as Species B, and it does not, as a rule, turn up its preoral
hood when irritated. Its metamorphosis usually takes place quickly, fifteen or twenty minutes
being required for its completion. Actinotrocha Species A. is, no doubt, the actinotrocha of
/'. architecta.

Species />'. {Jiff. ■>'■>)■ -This Actinotrocha is larger than Species A., and when about ready
to metamorphose it has an average length of 1.22 mm., and has at least 26 tentacles. (Wilson
(24) figures the Actinotrocha Species B., ready to metamorphose, with 22 tentacles.) The differ-
ence in appearance between this larva and Species A. is rather striking. Beside being somewhat
longer, it is slightly narrower in the collar region and decidedly so in the trunk region, which
gives it a much more graceful appearance than that of Species A. The intestine is quite long,
extending throughout the posterior two-thirds of the trunk cavity.

M. Longchamps has kindly pointed out to me that the "adult tentacles appear bilaterally,
the mid-ventral line being, at first, free of the buds." They do not arise, however, as thicken-
ings on the under side of the bases of the larval tentacles as in Species A. They have their
origin at the base of the larval tentacles, but they are separate from them, and they appear first
in the larva with 24 tentacles.

This Actinotrocha differs in three important respects from Actinotrocha Species A. In the
first place it has its blood corpuscles aggregated into four masses, two of which are usually in
the same position as the pair in the smaller species. The other two, however, are found, as a
rule, more anteriorly in the collar cavity, and are applied to the dorso-lateral walls of the stomach.
The posterior pair lying on the ventro-latcral sides of the stomach make their appearance during
the IS or 20 tentacle stage, but the other pair do not appear until about the 22-tentacle stage.
This larva also has retractors extending from the ganglion to the region of the first and second
pair of tentacles.

A second point of difference is the fact that Actinotrocha Species B. possesses a pair of
diverticula at the anterior end of the stomach. These are present as early as the 22-tentacle
stage. This larva can further be distinguished from the other species by the fact that there
is found in the older larva' a sensory papilla on the mid-dorsal surface of the preoral lobe.

Actinotrocha Species B. is much more active when irritated than the other species. The
least irritation causes it to turn up its hood and to assume attitudes like those figured by
Masterman (15). In fact, judging by the figures and text of Masterman's paper, it seems that
there is considerable similarity between this larva and the one he has described. The two larvae
are very much alike in shape and both have the lateral stomach diverticula, but the form that
.Masterman describes has only two masses of blood corpuscles. The two species are not identical,
nor is Actinotrocha Species B identical with Actinotrocha iranchiata from the North Sea, for, as
Longchamps has pointed out to us, the latter has but two blood corpuscle masses. Longchamps
has informed us that in Actinotrocha Species B. the adult tentacles make their appearance in the

MEMOIRS OF THE NATIONAL ACADEMY OF SCIENCES. 87

same special way that they do in Actinotrocha branchiata (found near Helgoland and described
by .1. Muller (19); and Masterman (15) says in Ins paper that the form he worked on "does not
appear to differ in any essential respect" from Actinotrocha branchiata. There seems, however,
to be considerable difference in size between Actinotroehabranchiataaad Actinotrocha Species B.,
for, according to Longchamps, the best-developed specimen that he obtained of this species
measured 2 mm., while the length of Actinotrocha Species B. averages L22 nun.

Although Actinotrocha Species A. seems to metamorphose without any difficulty when
brought into the laboratory, yet we have never been able to induce Actinotrocha Species B. to
do so. Specimens have been kept for ten days or more (the pouch and blood corpuscles being
well developed) and in some cases they succeeded in evaginating the ventral pouch, but they
were never aide to complete the metamorphosis.

A- far as we know, the adult of this Actinotrocha has never been found, hut probably it lives
under quite different conditions from Phoronis architecta, and it is not improbable that it may be
found as a deep-water form.

INTERNAL ORGANIZATION OF THE FULLY DEVELOPED ACTINOTROCHA.

"Subneural gland" (Masterman). — Masterman (15) has described a depression in the dorsal
wall of the buccal cavity which he terms a "subneural gland" and which he compares with the
eland of the same name in the Tunicata and also possibly with the hypophysis of the Vertebrata.

Roule (20) and Ikeda ('.») are of the opinion that this depression is a product of the fixing
method.

Longchamps (12) does not consider it to be an accidental structure, but he does not agree
with Masterman's view as to its theoretical significance.

Menon (17) says that the "subneural gland" first appeai-s in connection with the collar and
that during development it shifts forward into the preoral lobe, but in another part of his paper
he says the oesophagus is often folded transversely (this also the case in the young Phoronis) into
pouches and the ''subneural gland"1 is a diverticulum of its dorsal wall.

While in examining sections we have frequently found a depression in the region that Mas-
terman (15) indicates, we have never found it in the living larva. Only in very poorly killed
larva? have we found the depression to be as deep as Masterman has shown, and in all cases the
structure of the wall is practically like that of the oesophagus.

In the Actinotrochse Species A. and B. there is no depression in the living larva which might
be homologized to the subneural gland of higher animals, and we are forced to agree with Roule
and Ikeda in their belief that the so-called "subneural gland" which Masterman describes is a
product of Fixation.

•• Oral and atrial grooves" (Masterman). — Masterman (15) has observed a mid-ventral
ciliated area leading into the mouth from the preoral lobe in front and a broad ciliated area
depressed into two oral grooves leading into it from the ventral surface of the collar area. He
has also seen two so-called "atrial grooves" leading into the dorso-lateral corners of the mouth.
Masterman says he does not rind gill-slits in the Actinotrocha, nor does he find structures that
he considers to lie their homologues. "The atrial grooves" of the Actinotrocha, he says, how-
ever, are the analogues of gill-slits (ir.. p. 319). On page 35s (15), however, he says that
"tentatively, I would regard the atrial grooves of the Actinotrocha as the early rudiments of
pharyngeal clefts as found in Cephalodiscus."

His ••oral grooves." he says, correspond to the oral grooves in Cephalodiscus.

Roule (20) doe- not rind the "atrial grooves." but finds two lateral grooves, which he con-
siders to be formed by the insertion of the hood on to the collar wall.

Ikeda CO and Longchamps (12) are of the opinion that these grooves do not normally exist.

We have made a careful study of the live Actinotrocha and of surface mounts, but have not

been able to make out these grooves in either Species A. or Species B. Sections, however, show

that the "oral grooves" are present, and that in most preparations where the preoral hood has

been turned upward by violent contraction (due to the fixing agent) there are two short grooves

88 MEMOIRS OF THE NATIONAL ACADEMY OF SCIENCES.

in the position in which Masterman finds the ••atrial grooves." In those cases in which the hood
remains in its normal position we have seldom found Masterman's so-called •'atrial grooves," and
even in a few cases when' the hood is turned upward they have been absent.

The ventral wall of the hood just as it passes into the wall of the oesophagus not infrequently
shows a pair of bilaterally situated grooves which are similar to those found on the ventral collar
wall.

•• X> uroport " (Masterman). —See section on the nervous system.

"Sub-neural sinus" (Masterman). — Another organ which Masterman (15) has described is a
sinus immediately below the nerve ganglion, caused by the want of contiguity between the
raesoblastic walls of the preoral cavity and the collar cavity. This sinus, he claims, is closed
except for a fissure which leads eventually into the dorsal blood vessel. He compares this sinus
to the heart of Balanoglossus.

Menon (19), as far as we know, is the only other worker on the Actinotrocha who claims that
there is a definite vesicle beneath the ganglion and he has discovered no connection between its
cavity and the dorsal blood vessel.

Roule, Longchamps, and Ikeda do not find this organ, but the latter recognizes the existence
of a space ("posterior recess") free from mesenchymatous fibres, which is the posterior part of
the preoral lobe. This, however, he says, does not connect with the dorsal blood vessel.

From the study of the early development of Phoronis architecta and the origin of the
mesentery between the hood and the collar, we have come to the conclusion that no vesicle is
formed in that species between the two layers of the mesentery (if two layers exist). The
mesentery which forms the posterior wall of the preoral lobe cavity is found attached just back
of the ganglion in the median line and there is not the least sign of a vesicle other than the
cavity of the preoral lobe (tig. -M).

In neither Actinotroclm Species A. nor Actinotrocha Species B. have we found a vesicle below
the ganglion, although in both cases there is a space such as Ikeda (9) has seen, free from mesen-
chymatous fibres. The anterior boundary of this space is rather sharply defined and occasionally
among longitudinal sections a fibre with a nucleus is seen running vertically from the dorsal to
the ventral wall of the hood, giving the appearance of an anterior wall to the space. These
tibres, however, are very much more delicate than the wall of the collar lobe septum, and what is
more, they occur only occasionally and are evidently not sections through a membrane.

In the Actinotrochse,, which we have examined, there does not exist any vesicle beneath the
nerve ganglion nor any structure which could be likened to the heart vesicle of Balanoglossus. For
the supposed relation of the dorsal blood vessel to the "subueural sinus," see Blood system.
"Stomach Diverticula1'' (Ikeda, Longchamps, and Menon).
" JVbtochords" (Masterman and Roule).

Ever since Johannes Midler (19) saw the paired "blinddarme" in Actinotrocha branchiata
nearly all of those who have studied Actinotrochse have observed the same structures. Some
have considered them to be liver diverticula, others have described them as dark masses with
globules and as brown specks. Wilson calls them "glandular lobes of the stomach."

Ikeda (9), Longchamps (12), and Menon (17) speak of them as "stomach diverticula," but
they do not ascribe any function to them. Masterman (15) and Roule (20) look upon them as
rudimentary notochords. Roule, Ikeda, and Longchamps have studied larvae in which the
diverticulum was not paired and lateral, but unpaired and medio-ventral. The latter investigator
has observed larvae of both t\ \«<.

We find that in Species A. the diverticulum is undeveloped even at the time of metamorphosis
while in Species B. the diverticulum is paired, well developed, and ventro-lateral.

Longchamps has very justly objected to Masterman's use of the name " Diplochorda" nndei
which the latter includes the Phoronidse and Cephalodiscus.

The diverticula of Species B. do not show the regularly arranged vacuoles which Masterman
has described for the Actinotrocha from St. Andrews Bay. In fact, we agree with Longchamps's
(12) observations in finding the histological characters absolutely different in Species A. and B.
from the histology of notochords, and there is not the least indication of supporting tissue.

MEMOIRS OF THE NATIONAL ACADEMY OF SCIENCES. 89

However, larvse which we have examined, fixed in Flemming's fluid, have aotshown the vacuoles
to l>c filled with Eat droplets as Longehamps states.

Sections through the diverticula of quite old larvse (fig. 35) stained with iron hematoxylin
show columnar cells, nearly every one <it' which contains a deeply staining body about one-quarter
the size of the nucleus. The bodies are not found in the wall of the stomach proper, and we
believe that they give the yellowish- brown color to the diverticula of the live Actinotrocha.

In some cases we have found old larvse in which the cells of the diverticula were vacuolated,
liut in these cases we have also found that the entire stomach wall was vacuolated. The vacuoles
were never large enough or numerous enough to alter the natural position of the nuclei.

According to Masterman's description, the first vacuoles are formed at the distal ends of the
cells and more vacuoles arise later between these and the inner ends of the cells. As far as we
know, the origin of vacuolated tissue in vertebrates is the reverse of this, the vacuolization
beginning at the center of the cord and traveling outward.

The specimens of Actinotrochx of Phoronis sabatieri which we have examined show the struc-
ture of the stomach diverticulum to be very similar to that of the diverticulum in the Actino-
trocha Species B. The diverticulum, however, is somewhat more vacuolated in the former than
in the latter, but it does not show the peculiar structure which Masterman has described for the
"notochord" of the species from St. Andrews Bay.

It i> hard to see what use the Actinotrocha has for any organ of support in the region where
the diverticula are found and it seems much more probable that they have a glandular function.

Nervous system. — It is generally admitted among investigators who have studied the.
anatomy of the Actinotrocha carefully, that the creature has a subepidermal layer of nervous
tissue throughout the body which is fibrillar in character. This nervous tissue assumes the form
of quite definite tracts in certain parts of the body in Actmotrocha Species B. and fairly well-
developed nerves can be said to exist. The most conspicuous ones are found in the median
dorsal line of the preoral hood as three distinct longitudinal bundles of nerve fibres extending
from the ganglion to the anterior edge of the hood. There are other tracts which, though they
are not as definitely marked out as the above, are undoubtedly nerves.

Masterman (15) in his work on the anatomy of the Actinotrocha from St. Andrews Bay has
described a complicated nervous system, but the investigations of lloule (20), Ikeda ('J), and Long-
champs (12) have thrown considerable doubt on the correctness of his observations. Whether
these differences have been due to differences in the Actinotrochse studied by these workers or
whether they are due to the technique it is impossible to say, but, judging from the difference
in the degree of development between the nervous system in Species A. and Species B., we are
led to believe that the disagreements are due partly to the fact that no two of these investigators
have studied the same species of Actinotrocha.

While the nervous system of Species A. can with careful study be shown to be very similar
to that of Species B., yet it is so feebly developed that without first having studied Actinotrocha
Species B. we should not have been able to see the similarity in the disposition of the different
nervous tracts. The ganglion with its three dorsal longitudinal nerves running along the median
line of the hood is easily seen in the live larva of Species A., but in sections we have found it
impossible to trace the latter. The sensory papilla mentioned in the. description of the Actino-
trocha Species B. is absent in this species.

We are pleased to be able to confirm, to some extent, Masterman's (15) description of the
nervous system of the Actinotrocha, especially since a shadow of doubt has been cast upon his
work by some who have studied the Actinotrocha.

Partly because Species B. seems to be a much more highly developed Actinotrocha than
Species A., and partly because of its similarity to the one that Masterman studied (which i^ of so
much theoretical interest), we shall confine the description and figures to the nervous system of
Species B., although we are convinced that this Actinotrocha is not that of Phoronis architecta,
but of an adult that has not been discovered.

90 MEMOIRS OF THE NATIONAL ACADEMY OF SCIENCES.

We must admit that we have been very unsuccessful in the attempt to study the nervous
system of the Actinotrocha by means of methylene blue and ammonium molybdate. Cold
chloride has given no better results than staining with iron hematoxylin.

If the dorsal surface of the hood of a live Actinotrocha Species B. be examined, one will find
that there are a great many fibres which run in more or less definite tracts (tip;. 36). Many of
these fibres have nuclei along their course and are undoubtedly muscle fibres, while others run to
the edge of the hood and there seem to be continuations of certain cell-like bodies which Ikeda
was the tirst to describe (fig. 37). Although we have seen these bodies on all occasions in surface
\ iews stained with methylene blue, yet in sections we have never been able to make them out, if
they are nerve cells. It must be mentioned, however, that in transverse sections through the
edge of the hood every 3 m section shows at least one nucleus closely applied to the ring of
nervous tissue running round the edge of the hood (fig. 43). These occupy the same position
with reference to the edge of the hood that the cell-like bodies do which are seen in surface views,
but we take them to In' the nuclei of muscle cells, and frequently7 we have traced deeply stained
muscle fibres arising from them (fig. 43). Within the nervous tissue of the preoral ring we have
found no structures which we could consider to be the cell-like bodies mentioned by Ikeda (9).

Ikeda has figured a great many fibres arising from the ganglion, but in the Actinotrocha
that we have examined we have not been able to see the connections; however, we do not wish
to deny that they exist.

The three median nerves arising from the anterior side of the ganglion and running forward
to the edge of the hood, and two longitudinal tracts of nerve fibres arising from the posterior
side of the ganglion, can be easily made out, but the large majority of fibres which compose the
broad tract shown in fig. 36 are not connected with the nerve ganglion. There are some indi-
vidual differences in the arrangement of the above tracts, but in general they are about as shown
in tig. 36.

On each side of the medio-dorsal line in the region of the youngest tentacles a tract of fibres
can be seen running longitudinally. In the region where the edge of the preoral hood is inserted
into the collar a small tract made up of a few fibres branches off' from the dorsal longitudinal
tract and passes into the edge of the preoral lobe. Somewhat farther forward each dorsal longi-
tudinal trunk spreads out sometimes into three rather indefinite tracts, most of whose fibres seem
to reach the edge of the hood. Many of the fibres of the anterior branch appear to end in the
region at the sides of the ganglion, but no connection with the latter could be found.

Immediately posterior to the ganglion a tract of fibres (tig. 36) is seen which runs for a short
distance transversely to the long axis of the Actinotrocha. On both sides the fibres of this tract
soon diverge from one another and in this way distribute themselves over the anterior part of
the hood, ending at the edge of the latter (tie-. 36).

Masterman (1">) has figured (PI. XVIII. tig. 2) certain nerve tracts to the right and left of
the three nerves arising from the anterior end of the ganglion and finds that these "run
forward and outward and then bend backward and take a course to the posterior corner of
the hood."

A lateral view of the hood of Acti/notroclm Species B. shows sometimes fibres gathered
together in trunks, but these never take the direction as shown by Masterman. They diverge
rather regularly and end all along the edge of the hood instead of at the posterior corners of the
same (tig. 37). They are in no way associated with the ganglion and do not have the appearance
of being even when the hood is turned upward out of its usual position.

For several reasons we believe that the complicated tracts of fibres seen in a surface view of
a live Actinotrocha Species B. are not nerve fibres but muscle fibres. First, many of them show
along their course nuclei resembling nuclei of muscle cells. Second, cross sections through the
hood show that there is a rather heavy lining of muscle fibres which run in the same general
direction as do the fibres shown in the surface view. Third, there is no connection between
these fibres and the nerve ganglion.

MEMOIRS OF THE NATIONAL ACADEMY OF SCIENCES. 91

From the posterior side of the ganglion two tracts of nerve fibres pass oul and can be traced

backward sonic little distance, but they are .soon lost to view, as Lkeda (9) lias found to be the
case when studying methylene-blue preparations.

Sections through Actinotrocha Species B. bring out quite plainly certain nervous tracts which
appear as thickenings of the subepidermal nervous tissue and which correspond in a large part
to the principal nerves described by Masterman.

Anterior to the ganglion a section through the hood shows the parallel nerves which run
from the anterior side of the ganglion to the anterior edge of the hood. The boundary of these
nerves, as shown in tig. 44. is a little too definite. The subepidermal nerve tissue, which forms
a thin layer below the ectoderm cells, is not shown in the series of sections to be described.

Following the sections posteriorly we come to the ganglion, which in this specimen has
become invaginated, together with the overlying epidermis, so as to form a pit. A cross section
through this pit is shown in tig. iib. The cavity of the pit is lined by epidermis, while periph-
erally the wall of the pit consists of the ganglion cells and the nerve fibres of the ganglion (tigs.
38— !47>). The nuclei of the ganglia are easily made out, but it is only after staining very deeply
with iron hematoxylin that the cytoplasm can be seen. The invagination in the region of the
ganglion is unusual and is brought about by the violent contraction of the hood when immersed in
the fixing fluid. This undoubted^ is the same condition that Masterman (15) has described and
the same structure that he has homologized to the "neuropore" of the Chordata, or that he has
compared to the tubular dorsal nervous system of the same type as that of Balamoglossus. (Q. J.,
Vol. XL, page 295, 296.) It should be mentioned, however, that Masterman (16&) in his answer
to Koule has admitted the error of his rather hasty conclusion.

Menon (17) has recently described a tubular nerve ganglion for a certain Actinotrocha, but
the structure is probably due to fixation.

Immediately posterior to the ganglion a cross section shows two thickenings of the sub-
epidermal nervous system. These thickenings are what Masterman has described as the
dorsal longitudinal nerves and they can be traced from the ganglion. They are almost exactly
between the dorsal muscle tract and the epidermis of the dorsal wall. A little farther back these
so-called nerves are not quite as distinct, but when the region of the first pair of tentacles is
reached they become more prominent again and diverge, passing down the lateral walls along
the bases of the tentacles (tig. 41). They meet in the ventral region and thus form a ring-like
thickening of the subepidermal nervous system, which is undoubtedly the same that Masterman
has described as the " collar nerve ring" (fig. 42). Ganglion cells are demonstrable in this nerve
ring by staining deeply with iron hematoxylin (tig. 39). As we shall show in the account of
the muscular system there is a ring of muscle fibre which follows the nerve ring.

Masterman says that "fibers pass mid dorsally as a pair of tracts, giving off branches to
the body wall and terminating in a 'nervous ring just anterior to the perianal band." In his
figures of sections, however, the pair of tracts does not show back of the most dorsal pair of
tentacles. In Actinotrocha Species B. there are no definite tracts of nerve fibres running longi-
tudinally from the region where the collar nerve ring passes obliquely downward from the dorsal
surface of the collar. Nerve fibres are undoubtedly present all along the dorsal wall, but these
are not massed together in tracts and are simply the fibres of the ordinary subepidermal nervous
tissue. The nervous ring in front of the perianal band is not present in the Actinotrochse that we
have studied.

Masterman (15) finds that part of the nerve ring around the edge of the hood passes up to
the nerve ganglion when it reaches the insertion of the hood, and that numerous fibres also appear
to pass on to the ventral surface of the collar region. Live Actinotrochse (Species A. and Species
B.), when examined under the microscope, do not show a branch of the nerve ring of the lobe
passing upward to the nerve ganglion. Sections also fail to show this condition, which is very
necessary to Masterman's comparison of the nervous system of Balanoglossus and the Actino-
trocha. Fibres from the nerve ring do, however, pass on to the ventral surface of the collar
region.

92 MEMOIRS OF THE NATIONAL ACADEMY OF SCIENCES.

Numerous fibres, which Masterman .speaks of as passing down on to the ventral collar wall,
are massed in the Actinotrocha Species B. into two definite thickenings which are seen in fig. 44c
and fig. 40. These thickenings of the nervous tissue gradually approach one another as we trace
the sections backward and come to run along the same line as do the two ventral muscle tracts
of the collar, but before the line of insertion of the ventral tentacles is reached these thicken-
ings are lost in the subepidermal nervous tissue.

We have not been able to make out either in sections or in surface mounts any definite nervous
tract running from the collar nerve ring along the ventral region of the trunk, although, as
before said, there is a subepidermal network of nervous tissue throughout the wall. It
will be remembered, how7ever, that above we have described two longitudinal dorso-lateral tracts
of muscle fibres, and there are quite numerous longitudinal muscle fibres in the ventral wall of
the trunk. There is also a fairly well-developed layer of circular muscles, and these, together
with the longitudinal muscles, give the appearance in surface views of the longitudinal tracts
giving off branches.

The nervous system of the Actinotrocha of Phoronis sdbatieri is less highly developed,
judging from the specimens we have examined, than that of either Species A. or Species B. The
ganglion, or, as Roule (20) calls it, the " plaque cephalique," contains ganglion cells like those we
have found in other Acti/notrochse, which shows that there is something more present than a
simple subepidermal nervous system, such as Roule has described, in the Actinotrocha of Phoro-
nis sabatit /•/'.

Muscular system. — There is no doubt but that there is some diversity in the arrangement of
muscle fibres in the different species of Actinotrochse. A study of the two species, A. and B., as
well as the description of different species by other investigators, convinces us of this.

In the study of the muscular system the best results were with material fixed in Hem-
ming's strong solution and stained with Haidenhain's iron hematoxylin. These solutions make
the muscle fibres stand out very distinctly, whereas material fixed and stained with other fluids
shows them so feebly that the muscle tracts might easily be overlooked.

Ikeda (9) has described a pair of bundles of muscle fibres springing from "the hind lateral
corners of the ganglion and running divergently downward until they insert themselves in the
collar walls between the first and second tentacles." These muscles, to which he has given the
name of "retractors," are present in the Actinotrochse Species A. and Species B. (tigs. 34, 35, 4.">.
45a).

The "retractors" that Ikeda figures in the trunk cavity of one of the Japanese Actino-
trochse were not found in either Actinotrocha Species A. or Species B.

Another pair of bundles of muscle fibres is found in Species B. They spring from the
wall of the hood at the sides of the ganglion, traverse the cavity of the hood and become
inserted on its ventral wall directly under the ganglion (fig. 45^).

Certain tracts of muscle fibres are very highly developed in Species B. Transverse sections
(stained with iron hematoxylin) through the wall of the hood in front of the ganglion show
black dots spread over the internal dorsal surface of the hood, and these seem to be embedded
in the mesodermal lining. These dots are the cut ends of muscle fibres, and as the sections are
followed posteriorly, these dots gradually become massed about halfway between the ganglion
and the sensory papilla and represent the sectioned ends of a pair of longitudinal muscle tracts
which are bilaterally placed on the right and left of the median dorsal line (fig. 44). These two
thick tracts of muscle fibres extend posteriorly in the dorso-lateral regions of the Actinotrocha
and do not disappear until the perianal ring is reached. They are very characteristic structures
in Species B. (figs. 44 to 44A), but we have not been able to make them out in Species A.

These muscle bands, no doubt, serve to draw the anal end of the body of the Actinotrocha up
to the oral end during the metamorphosis. They are the most highly developed muscle tracts in
the body of the Actinotrocha and their course is almost identical with the course of the '•dorsal
nerve-." that .Masterman describes.

Examination of cross sections of Species B. in the region of the vestibule shows the cut ends
of numerous muscle fibres which are spread over the ventral surface of the collar. Passing

MEMOIRS OK THE NATIONAL ACADEMY OF SCIENCES. 93

posteriorly, these fibres become massed into definite muscle tracts, aboul halfway back from the
vestibule to the ventral insertion of collar trunk mesentery (figs. Il<. t4rf). These two ventral
longitudinal tracts, which arc bilaterally placed one on each side of the ventral median line,
become separated, in most cases al least, from the ventral body wall in the region of the posterior
pair of blood corpuscle masses and the latter become rather closely associated with them (fig. 4tl).
We could not discover that these fibres were in any way related to the nephridia as has been
described for some specie-.

In the region of the insertion of the ventral tentacles the muscle fibres of the ventral tracts
become again applied to the ventral body wall, hut definite tracts are no longer present. How-
ever, in the trunk region these fibres form a definite tract, which is confined to the ventral body
wall, and it does not disappear until the perianal ring is reached (ties. 44 <j, /*, i).

Another tract of muscle fibres present in species B., which does not seem to be developed in
other Aebinotirochm, judging from existing descriptions, is that found in the region of the bases
of the tentacles. From the dorsal muscle tracts, where the most dorsal and anterior pair of
tentacles arises, muscle fibres are given off, which follow the bases of the tentacles and which
form a well-developed ring of muscle fibres. In other words, there is a ring of muscle fibres
which follows the line of insertion of the mesentery between the collar and trunk cavities
(fig.Me).

A tract of muscle fibres, which also seems to occur only in Species B., is one composed of
only a few fibres, which are found running around the edge of the hood on the internal wall of
the same (tigs. 44 ■', b). Where the edge of the hood passes into the wall of the collar cavity
these fibres arc seen to run on to the internal surface of the lateral wall of the collar and to
mingle finally with the fibres of the dorsal tract. The direction these fibres take wThen they pass
on to the wall of the collar reminds one very much of the fibres which Masterman (15) figured as
nerve fibres.

On the internal ventral surface of the hood in both species of Actinotrochse there is a system
of muscle fibres arranged concentrically. The3r run almost parallel with one another and with
the edge of the hood (tigs. 47 and 44^).

Beside the tracts of muscle fibres which have been described there are, lining the walls of the
collar and trunk, circular muscle fibres lying between the longitudinal muscle fibres and the
ectoderm. These have been generally observed by previous workers as have also the muscular
covering of the ventral pouch and the muscle cells of the dorsal blood vessels.

Body cavities, mesenteries, etc. — Much difference of opinion exists as to the origin and limits
of the body cavities in the Actinotrocha and also as to the value of these cavities in determining
the phylogenetic history of Phoronis.

Koule (20) stands alone in considering the Actinotrocha to have but one body cavity, which
is lined by an epithelium formed from mesenchymatous cells. He absolutely denies the presence
of any mesenteries.

Through the kindness of Mr. Longchamps we have been able to study the Actinotrocha of
Phoronis sabatieri, and have found that the mesentery between the collar and trunk is present,
although it is less highly developed than in other species. We are unable, with the material at
hand, to give any opinion as to the presence of a mesentery between the preoral lobe and collar
cavities.

Caldwell (3) claims that there are but two body cavities, and that these are separated by a
mesentery (collar-trunk mesentery of Masterman).

Longchamps i L2) is inclined toward the view of Caldwell, while Ikeda (!») finds the mesentery
dividing the lobe and collar, which, however, he says, is incomplete. Both of these investigators
recognize the presence of the ventral mesentery.

Menon (17) finds three body cavities (preoral. collar, and trunk), a ventral mesentery, and
indications of a dorsal mesentery in the trunk.

Masterman (16) considers that the Actinotrocha have live body cavities — an unpaired lobe
cavity, a paired collar cavity, and a paired trunk cavity. This idea is based on his study of the

94 MEMOIRS OF THE NATIONAL ACADEMY OF SCIENCES.

early development of the body cavities and not on the adult organization of the Actinotrocka,
for in the collar he finds only :i dorsal mesentery (no other investigator has seen this), and in the
trunk only a ventral mesentery.

While it is possible that the mesoderm arises as diverticula from the enteron, as Caldwell
and Masterman have described, yet Longchamps (12), who lias recently reinvestigated the early
embryology of the form that Caldwell worked on {Phqronis kowalevsMi), denies the origin of the
cavity in front of the collar trunk mesentery from enteric diverticula. Ikeda (!»). who recognizes
the ••anterior diverticula" of Caldwell in the Japanese species, nevertheless holds that the body
cavities do not arise from anterior diverticula, but are simply produced by mesoblast cells apply-
ing themselves to and forming the lining of the ectoblastic and entoblastic wall.

In the section on the mesoderm we have stated that in the embryo of Phoronis architt <-i<i we do
not find that the mesoderm arises from enteric diverticula. There can not be the least doubt,
however, that the preoral lobe at an early stage becomes lined by a sac of mesoderm cells and
that the wall of this sac gives rise to the mesentery. Furthermore, this sac is extended postero-
lateral^ into two horns which are characteristic of the cavity of the preoral lobe, according to
Masterman, Ikeda, and Menon. It must be admitted, however, that this sac does not seem to
retain its character as a sac. but that the cells become separated and apply themselves here and
there to the walls of the preoral lobe. The mesentery remains intact and can not be considered
as a secondary structure as has been suggested by Longchamps. Although we agree with Ikeda's
statement that the mesentery between the lobe and the collar is incomplete laterally in the fully
developed Actinotrocka, yet in the Actinotrocka of Phoronis arckitecta, at least, it must be
considered as a definite mesentery.

The fully formed Actinotrockse (Species A. and Species B.) do not show a complete epithe-
lial lining to the preoral lobe, but the mesoderm cells are arranged as described in the young
larva.

It is stated above in the part on the mesoderm that we do not find that the lining of the collar
cavity is of enteroccelic origin in Phoronis architects. However, in the fully formed Actino-
trocha there is an undoubted mesodermic ephithelium lining the somatic wall. This layer is very
conspicuous immediately before metamorphosis, because it becomes separated from the somatic
wall prior to becoming transformed into the ring vessel of the adult (tigs. 51 </, 51//).

The splanchnic wall of the collar cavity in the Actinotrochse that we have examined is devoid
of a mesodermal lining, and the occurrence of mesoderm cells on the wall is very infrequent.
This condition of affairs in the well-developed Actinotrocka is what one would expect from the
disposition of the mesoderm cells in the very young larvae of Phoronis architecta, where it is
only very seldom that any are found on the stomach wall (tig. 24).

The absence of a mesodermal lining on the splanchnic wall of the collar cavity is made all
the more evident by the examination of cross sections showing the collar-trunk mesentery
(tigs. 51<7, 51//). When the mesentery reaches the stomach wall, instead of dividing into two
layers, one of which would be continued into the mesodermal lining of the stomach wall of
the collar cavity, it turns abruptly upon itself and becomes the lining of the stomach wall
of the trunk. We have never found the least indication in the collar cavity of a dorsal mesentery
such as Masterman (15) has described in the Actinotrocka from St. Andrews Bay. The trunk
cavity is lined throughout by a sac of mesodermal epithelium, and the mesentery is plainly seen
to be continuous with the lining of the somatic wall and with the lining of the wall of the gut.
The ventral mesentery of the trunk is present in Species A. and Species B., and while there
is no dorsal mesentery we have found indications of it in two specimens only (Species B.) at the
posterior end of the trunk. We can not say, however, that it has any ontogenetic significance
( lies. 4.x, 44;, 44//). We have also found the ventral mesentery to be present in the Actinotrocka of
J 'I,, i, ■<m is saiatu ri.

The ventral pouch fills a large part of the trunk cavity in the fully formed Actinotrocka, and
just before metamorphosis it frequently pushes the collar trunk mesentery well forward into the
collar cavity, thus making the study of the relation of the different parts quite difficult. Both

MEMOIRS OF THE NATIONAL ACADEMY OF SCIENCES. 95

the external opening of the ventral pouch and the nephridial openings are found on the ventral
wall of the trunk just posterior to the insertion of the mesentery as in other species.

Nephridia. — Wagener (23) was the first to observe the "nephridial bouquets," bul Caldwell
(3) was the firsl to publish a careful study of the nephridia of the Actinotrocha. Goodrich" (6)
has recently published a paper on the excretory organs of Amphtoxus, and he adds a note on the
nephridium of the Actinotrocha which confirms Caldwell's view. The two latter investigators
agree that the nephridium ends blindly without funnels; that there are tubular processes, each
one containing a lumen and tipped with an excretory cell, and that these processes radiate out
from the blind inner end of the nephridial canal.

Longchamps (12) is inclined to accept Caldwell's view of the subject. Roule (20) and Ikeda
(9) seem to hold the view that the nephridial canal ends blindly without branching', and that the
blind end is tipped with excretory cells, which, however, are not perforate.

Masterman (15) and Menon (17) have another view. They both think that the nephridial
canal terminates internally as two (Menon) or more (.Masterman) funnels, and they recognize the
existence of long processes without lumens attached to the ends of the funnel.

We have not been able to make a study of the nephridia of the living Actinotrocha, but we have
investigated them by means of sections in Species A. and Species B. For this purpose we have
used material fixed in Flemming's fluid and also in corrosive acetic. The sections were stained
with iron hematoxylin. Our work has been done with very high powers (Zeiss obj. fa and No.
L2 Zeiss compensating occulare).

The nephridia of the two Actinotrochse have much the same structure, but in Species A.
we have been unable to find that the internal end of the nephridial canal blanches, while in
Species B. the internal end divides into two short branches.

Figs. 52, 52a, 525 represent three transverse sections through the anterior part of the
nephridium of Species B. Fig. 52 is through the nephridial canal just posterior to its inter-
nal end. Darkly staining dots seen in the lumen represent cross sections of long flagella such as
Goodrich (6) has described in the " solenocytes " of Amphioxus.

Fig. 52a shows a section through the nephridial canal a few sections anterior to that of fig.
.'c_\ and at the same time it shows the lower or most posterior branch with a few excretory cells
and their processes.

In fig. 523, which is a section through the tip of the upper or anterior branch of the nephrid-
ium. the lumen of the nephridial canal is reduced to a very small clear space.

If a section is taken in a longitudinal direction through the nephridial canal and its excretory
processes dig. ■">-</). it is seen that the distinct walls of the nephridial canal disappear when the
"bouquet" of excretory cells is reached, but that the end is blind and that it is merely a thin
walled bulb from whose surface radiate the processes of the excretory cells. The structure of
these processes i- the same in both species that we have examined except that in Species A. they
are much shorter, which might account for the different descriptions we find in the literature.
We are convinced that in both species the excretory process - contain lumens, that these lumens
are continuous with the lumen of the nephridial canal, and that they contain nagella.

Each of the excretory processes is tipped by a body the distal end of which is drawn out
into a sharp-pointed process. The outline of only one cell could be seen in the body. Some-
times it contains but one nucleus, which may be oval in shape or bent almost at right angles
(tig. .V_'./i. but in the majority of eases there are undoubtedly two nuclei (tig. 52c). Fig. 52(
shows a cross section through one of these bodies. If a transverse section of the nephridial
processes is taken (tig. 52/), it is seen that each has a definite wall and that inside there is a
definite dot which we take to be across section through the flagellum. This flagellum has its
origin from the cell body at the end of the process, and the indications are that the flagellum
extends throughout the length of the process into the lumen of the nephridial tube.

"Since writing this paper a description of the nephridia of the Actinotrocha by Goodrich (6a) has come to our

notice. Our account agrees to a large extent with his.

96 MEMOIRS OF THE NATIONAL ACADEMY OF SCIENCES.

We are not prepared to say that the cell bodies at the end of the excretory processes arc
composed of two cells, but it. is a fact that two nuclei exist, and this conclusion is not based
on sections through bent nuclei which might lead one to think that there were two when only
one existed. It will be seen from this description that the anatomy of the excretory cells and
processes of the Actinotrochat which we have studied resembles that of similar structures
described by Goodrich for Amphioxus.

There is nothing new to add concerning the nephridial canal except that a longitudinal
section through it, which has been stained with iron haematoxylin, shows long flagella extending
some distance from the distal end into its lumen (tig. 52;/).

Sections of the Actmotrocha of Phoronis sabatieri show that the nephridia resemble those of
other Actinotrochse studied, although they do not seem to be as well developed as those of
Species A. and 15. The specimens at hand show that the internal opening in the collar cavity is
situated at about the same level as in other Actinotrochse and not at the level of the oesophagus,
as Koule (20) has indicated.

The course which the canal takes is like that which has been described by other investigators,
and we agree with the observation of others in regard to the nephridial canal opening to the
exterior at the sides of the ventral pouch opening.

Masterman (15) describes a pair of ciliated pores opening to the exterior on the dorsal surface
of the preoral lobe. These, he finds, lead into tubes closely similar in cross section to the cross
section of canals of the collar nephridia, and these tubes have an internal opening into the preoral
lobe cavity. He compares the external- pores to the proboscis pores of Cephalodiscus. In the
Zoologisher Anzeiger, 1901, Volume XXIV, page 231, he admits that their occurrence is variable.
No other investigators have found these organs. Ikeda mentions the fact that flask-shaped
glands on the upper fact1 of the preoral lobe occur in one larva studied by him, but he denies
that they are such organs as Masterman describes.

Masterman (15) also finds thin-walled organs lying in the hsemocoele space immediately below
the anal ciliated band. Speaking of these (15) he says: "A pair of organs which I have not fully
made out, but they may be the rudiments of the trunk nephridia." Masterman, however, denies
their existence in a later paper, and no one else has seen the organs, as far as we know.

Rudiments of tht adult Hood vessels in tht Actinotrocha. — Many investigators of Hh& Acti-
notrocha have recognized the beginnings of the adult blood vessels, but. E. B. Wilson (21) is the
first one who clearly states the fact that the cavity containing the blood corpuscle masses give-
rise to the ring vessel of the adult, 'although Metschnikoff seems to have had some such idea.
Caldwell (3) and Ikeda (!») confirm the statement of Wilson with reference to the origin of the
ring vessel of the adult.

While Masterman (15) describes a much more complicated vascular system for the Actino-
trocha from St. Andrews Bay than that of all the Actinotrocha examined, yet we agree with
him in his view that the cavities of the blood vessels may be considered as vestiges of the
segmentation cavity.

Above we have given our opinion that the "subneural sinus" (Masterman) does not exist in
the Actinotrochat that we have examined, and that although there is a space beneath the ganglion
it has no connection with the dorsal blood vessel.

The blood vessels of the adult are represented in the A<tht<itr<ieh;i Species A. and H. by a
dorsal vessel (figs. 34, 35) extending along the median dorsal line of the intestine, from the
mesentery between the collar and trunk almost to the posterior end of the stomach, where there
are small csecal outpushings of the splanchnic mesodermal walls of the end of the stomach. This
dorsal blood vessel, although it is a completely formed vessel, has arisen from a proliferation of
the cells id' the splanchnic mesodermal wall along the dorsal median line of the stomach, and its
lumen is really a part of the blastoccele — i. e., it is a part of the space between the splanchnic
mesodermal lining and the wall of the stomach. Posteriorly, the dorsal blood vessel becomes
indefinite and passes into the ordinary splanchnic mesodermal lining, thus really being open pos-
teriori}' into the space between the wall of the stomach and the mesodermal lining.

MEMOIRS OF THE NATIONAL ACADEMY OF SCIENCES. 97

At the time of metamorphosis in the Actinotroch.se Species A. and Species B., there is no
sign of ;i ventral blood vessel along the stomach, such as Masterman (15) and Rome (20)
describe.

We have been unable to find the "ring sinus" which, according to Masterman, connects the
dorsal vessel with the ventral vessel al the end of the stomach, nor have we seen the " postoral
ring sinus" connecting the dorsal vessel with the ventral vessel.

Masterman's "postoral ring sinus." ••ventral blood vessel," and "ring sinus" (situated at
the junction between the stomach and intestine) will be discussed in the section on the metamor-
phosis.

There is undoubtedly a space between the wall of the perianal ring and its mesodermal lining
(fig. 49) in preserved specimens which seems to be what Masterman calls the " haemal ring,"
luit it docs not become any organ of the adult.

As stated above, we believe, with Wilson, Caldwell, and Ikeda that the cavity of the collar
and its somatic mesodermal lining become the ring vessel of the adult.

We shall continue the discussion of the further development of the dorsal blood vessel into
the efferent and afferent vessels of the adult in the section on the metamorphosis.

.Masterman speaks of haemal sinuses passing down the tentacles, hut says that they are not
very decided. Ikeda has. however, investigated these structures carefully, and we thoroughly
agree with his view that the cavity of the collar, together with its somatic lining, extends into
the tentacles, and that these prolongations become the tentacular vessels of the adult. This
condition is shown very plainly in a dorso lateral section of the Actinotrocha Species A. (fig. 50).

111,,,,,! corpuscles and their origin. — E. B. Wilson (24) has touched upon the origin of the
blood corpuscles, and according to him they develop in solid masses adhering to the stomach
walls near the base of the tentacles. Caldwell (3) finds that the corpuscle masses •'arise from
the niesoblast cells in front of the septum," but he has nothing further to .say about their
position or origin. Ikeda (!») describes the blood corpuscles as arising fronW*gigantic nieso-
blast cells in the body cavity of the larvie with one or two pairs of tentacles." Since the
publication of this paper, Ikeda has rejected this view, although he has published nothing on the
subject. Menon (17) thinks that the blood corpuscles arise from the splanchnopleure covering
the stomach and its diverticulum. According to Cori (4), the blood corpuscles in the adult are
formed from the endothelium of the blood vessels.

In the Actinotrocha Species A. (probably that of Phoronis architecta) the blood corpuscles
usually make their appearance duriug the 14-tentacle stage, as in "Type A" described by Ikeda.
although we have found larva- of this stage in which definite blood corpuscles were not present.

Actinotrocha Species A. with If, tentacles invariably has blood corpuscles, and they are
present in the so-called collar cavity as two masses more or less closely applied to the ventro-
lateral walls of the stomach (lie's. 51 </. //). In some cases, however, they are separated from the
wall by a considerable space.

The transverse section of a larva with 12 tentacles in a plane just posterior to the base of
the tentacles, but anterior to the mesentery, always shows two masses of cells bilaterally placed
and closely applied to the mesoderm lining the ventro-lateral somatic wall (fig. 5:3). Occasionally
cells are found in these masses, situated very close to the mesodermal lining, which are
decidedly spindle-shaped in form and whose nuclei resemble those of the cells of the mesodermal
lining, both in shape, size, and internal structure. These cells are not very rich in cytoplasm.
Most of the cells, however, are almost three times the size of the cells lining the somatic wall,
the cytoplasmic part of the cell having increased in size to a greater extent than the nucleus.
Mosl of the nuclei have large deeply staining nucleoli (tig. 54).

In some specimens parts of these masses of cells are apparently in the act of wandering
across the body cavity to the position the blood-corpuscle masses occupy in the fully formed
Actinotrocha.

Some 15 or 20 larva' with L2 or 14 tentacles have been sectioned, and with one exception we
have found that when the mesodermal masses are present on the ventral body wall there arc no
blood-corpuscle mas>es present in the larva, and that when the blood-corpuscle masses arc

98 MEMOIRS OF THE NATIONAL ACADEMY OF SCIENCES.

present there are no mesodermal masses. In this exception small blood-corpuscle masses were
round applied to the stomach wall, and masses of cells bilaterally placed were found on the
ventral somatic wall, but these colls had already taken on the character of blood corpuscles.

Ikeda (9) has described a "mesoblastic cell mass"" which he evidently considers as giving
rise to the adult body cavity, and its position is very similar to that of the mesoblastic masses
described above. They are both products of the mesoblastic lining of the ventral somatic wall
and are situated between the plane of the bases of the tentacles and the plane of the somatic
insertion of the mesentery between the collar and trunk. Although Ikeda docs not touch upon
the very early origin of the adult body cavity, yet it seems probable that he considers it as arising
from a single mass of cells. The mesoblastic masses described above are paired and bilaterallv
placed, and they are present only in the young larva of 12 or 11 tentacles. Furthermore, in
the larva with L2 or 11 tentacles there is no sign of the beginning of the adult body cavity.
Although these mesodermal masses which, according to our observations, give rise to the blood
corpuscles have a similar position to the fundament of the young adult body cavity, yet we are
convinced that they do not give rise to it.

In Species A. there is no intimate relation between the masses of blood corpuscles and the
uephridia. such as has been described by Masterman (15) for the species from St. Andrews Bay.
and by Longchamps (12) for Actinotrocha branchiata. In the larva of L6 tentacles the blood-cor-
puscle masses are, however, closely applied to the stomach wall in the region of the digestive
area. There is no mesodermal epithelium covering that part of the surface of the stomach which
lies within the collar cavity, and the blood corpuscles seem to be so intimately related to the
digestive areas that we are inclined to believe that they receive nourishment from them.

While the blood corpuscles vary in size and undoubtedly multiply by karyokinetic division,
yet we have never found the ''large and somewhat coarsely granular" and the '•smaller finely
granular" corpuscles that Ikeda (9) speaks of, nor in this species have we found any "gigantic
mesoderm cells'-' in the region of the blood-corpuscle masses. Very large cells in close relation
to the blood corpuscle hkhmn are found in some specimens of Arfiimtiuicha Species B. (tig. 44/').
These cells resemble the cells described in the old gastrula of Species A. as arising from the wall
of the archenteron, only they are not as coarsely granular as the latter. While in Actinotrocha
Species B. the cells are found in most cases closely associated with the blood corpuscles, we have
never seen them in the process of division and do not believe that they give rise to blood
corpuscles. Their occurrence is quite variable, hut as far as has been observed they are not
present in the Actinotrocha which are ready to metamorphose. They are not phagocytes, nor are
they pigment cells, and the only name which we feel justified in giving them is large free
mesoderm cells. Frequently they are also found in the posterior end of the trunk cavity (tig. 44/').

Roule (20) holds that the nephridia end internally at the level of the oesophagus, and he
shows this in a figure. We have made cross sections through this region and have found masses
of cells in much the same place as Roule has shown. These cells seem to be blood corpuscles.
but very few specimens have been examined, and only one of these showed these masses of cells.

Rudiment of tin "adult collar cavity" (Ikeda). — Ikeda has observed a mesodermal cell mass
on the ventral somatic wall just under the second tentacle in rather young specimens of all the
Japanese Actinotrochae. He has traced the development of this mass of cells and finds that a
cavity arises in it which, before metamorphosis, becomes quite spacious ami extends into the
tentacles. We are able toconlirm Ikeda's view that this cavity istherudimentof the " adult collar
cavity," or "supraseptal cavity" of the adult, as it is usually called (tigs. 50, 48, 51A, 455, 4»>).

METAMORPHOSIS.

Several investigators have carefully described the external characteristics of the metamor-
phosis of the Actinotrocha, so it is unnecessary to enter into a detailed description.

Wilson (24) studied the metamorphosis of Actinotrocha?. Species A. and Species B., which are
found in Chesapeake Bay, but he did not cut sections of his material. Ikeda (9), however, has
investigated the internal changes which take place during metamorphosis and has added a vai-

MEMOIRS OF THE NATIONAL ACADEMY OF SCIENCES. 99

liable contribution to the subject. The behavior during metamorphosis of Actinotrocha Species
A. and Species B. from Beaufort Harbor seems to be quite the same as that <>l' the two Actino-
trocha which Wilson has observed, and there is little doUbt but that thej are of the same species.

A.s Wilson has stated, the metamorphosis of Actlnotrocha Species A. (fig. 34) takes place
much more quicklj than that of Species B. (fig. 35). In fact, we have never obtained a completely
metamorphosed specimen of the latter, although many times we have found specimens of this
species with the ventral pouch well evaginated (fig. •">•'>). We have tried to make the conditions
favorable for the completion of the metamorphosis by covei ing the bottom of the aquarium with
a layer of sand rich in diatoms and also by changing the water frequently. Under these
conditions the larva' (Species B.) would invariably sink to the bottom and move around on the
sand apparently in search of a favorable place to finish the metamorphosis. The latter never
occurred, however, although sometimes the Iarvas would attach the end of the ventral pouch to
the bottom of the dish. In this way the creature would often remain for days and although the
preoral lobe and larval tentacles would degenerate the anal end of tin1 larva would never become
turned upward so as to lie in close proximity to the mouth.

A- we have said before, we are inclined to think that the Actinotrocha Species B. belongs to
an adult which lives under different conditions from that of Phoronis architecta, and we should
not be surprised if it were found to be the Actlnotrocha of a deep-water form. Although
( ', rianthus occurs in Beaufort Harbor, we have never found Phoronis australis associated with it.

Actinotrocha Species A., a- a rule, metamorphoses in about twenty minutes (figs. 56, 56a,
."it',/.), and usually just before this takes place it sinks to the bottom of the dish, but occasionally
metamorphosis occurs on the vertical side of the dish near the surface of the water, the young
Phoronis remaining fixed where the metamorphosis takes place.

/',; oral Inh, ,n,, 1 1, ntacles. — Usually the larval or distal part of the tentacles (Species A.) and
the preoral lobe are swallowed during metamorphosis. The proximal parts of the tentacles
become directed upward and constitute the tentacles of the adult. They always number IS in
the very young Phoronis (Species A.) and there is an indication of the horseshoe arrangement
which is found in the adult (fig. 565).

The preoral lobe does not give rise to the epistome of the adult, for as Menon (17) has
correctly observed, this structure is not present in the very young Phoronis. However, the
epistome, which is of ectodermal origin, soon makes its appearance, and when the creature has 30
tentacles it is a very conspicuous organ (fig. 57).

Ganglion. — The ganglion on the dorsal surface of the hood is lost when the preoral lobe is
swallowed, and hence does not give rise to the so-called brain ganglion of the adult.

Ectodermal wall of. collar. —Although the preoral lobe degenerates, the wrall of the collar
does not, but becomes drawn inside the body of the young Phoronis and forms the wall of the
oral end of the gut.

Perianal ciliated ring and ectodermal wall of the trunk. — When the critical point in the
metamorphosis is reached —that is. when the posterior end becomes drawn up to the region of the
mouth — the perianal ciliated ring is usually seen as a protuberance in that region (Wilson's tig.
1l'). but shortly after this it becomes drawn in, and, together with someof the ectoderm, becomes
the lining of that part of the rectum which is near the anal opening. The drawing in takes place
to such an extent that most of the ectodermal wall of the trunk of the Actinotrocha becomes
incorporated in the wall of the rectum, as Caldwell has observed.

This process, together with the drawing in of the ectodermal wall of the collar to form the
wall of the oral end of the gut, seems to cause a change in the position of the nephridial canals.
(See section on nephridia.)

( 'avity "ft!,. /,,;<,r<il lnh. .—Since the preoral lobe is lost d urine- the metamorphosis, its cavity
does not take part in the structure of the adult.

Mesentery between //" lob< and cottar. This mesentery does not persist.

Larval collar cavity. As has been stated by other investigators, and as we have observed, the
larval collar cavity with its mesodermal wall becomes the ring vessel of the adult. This organ
will be discus>ed further in the section on the vascular system.

100 .MEMOIRS OF THE NATIONAL ACADEMY OF SCIENCES.

" Adult collar ca/oity" (Ikeda). — This cavity, which is found in the well-developed Actinotrocha,
undoubtedly becomes the "adult collar cavity" or supraseptal cavity of the adult, as Ikeda says.
In the young Phoronis it is seen as a cavity which occupies all the region anterior to the trans
verse septum and which is prolonged into the tentacles. It is lined by a mesodermal epithelium
and contains the ring vessel with its tentacular vessel.

Trunk cavity and cavity <>f1ln ventral pouch. — These cavities become the infraseptal cavity
(if the adult.

Ventral* mesentery. — The ventral mesentery of the Actinotrocha no doubt becomes the mes-
entery in the adult which Cori calls the " hauptmesenterium " and which Benham names the
"oesophageal" ami "■rectal" mesenteries. Mesenteries are present in the very young Phoronis
(just after the completion of the metamorphosis), which are found in the exact position that one
would expect the ventral mesentery of the Actinotrocha to assume after metamorphosis. Ikeda's
figures indicate that he considers the longitudinal mesenteries of the very young Phoronis to he
tin' same as the ventral mesentery of the Actinotrocha, for he gives them the same name.

We can not oiler any observation on the origin of the lateral mesenteries of the, adult except
that they are not present in the very young Phoronis that we have examined. They undoubtedly
arise later in the life history.

Stomach div( rticida. — This structure is not present in the Actmotrocha Species A., so we are
not able to give any information on the subject. It seems to be the general opinion among those
who have studied the metamorphosis that it does not persist as an organ in the adult.

Digestivi areas. While these organs persist for some little time after metamorphosis, they
are not evident as organs in the adult Phoronis architecta.

Nephridia. -Caldwell (:!), Ikeda (9), Longehamps (12), and Menon.(17) have all observed the
change in position of the larval nephridial canals which is due to the changes taking place during
the critical period of the metamorphosis, and it is a quite well-established fact that the external
ends of the larval nephridial canals come to be situated near the anal opening.

.lust after the critical period a cross section through the anterior end of the young Phoronis
cuts the transverse septum (" diaphragm, " "collar trunk mesentery"), which runs obliquely and
passes through the supraseptal and infraseptal cavities. It shows a transverse section through the
nephridial canals, which are still attached to the mesentery, as in the Actinotrocha. Following
the sections anteriorly, the canals are seen to open into the supraseptal cavity ("'larval collar
cavities," " ring vessel of the adult"), and they are still found in possession of their excretory
cells. Posteriorly the sections show that the nephridial canals leave the septum and pass between
the wall and the mesodermal lining of the infraseptal cavity (tig. 59). At this time their external
openings are situated on the lateral epidermal wall in a transverse plane which is somewhat below
the transverse plane of the anus, and they are by no means as near to the latter as they are in
the adult Phoronis. It is seen from this description that during the critical period there is very
little change in the structure of the larval nephridia or in their position, although the evagination
of the ventral pouch and the drawing in of the ectoderm of the trunk to form the end of the
rectum causes the anus to become rather closely approximated to the external nephridial openings.

Caldwell (•>) says that the whole of the larval nephridial canals remains as the paired nephridia
of the adult, while Ikeda thinks it probable that only the parts of the nephridial canals lying in
the wall of the trunk persist. He assumes that the nephridial funnels of the adult, which both
open into the infraseptal cavity, are secondary outgrowths of the above remnants of the nephridial
canals.

As the metamorphosis continues, sections show that the excretory cells and that part of the
nephridial canals situated in the larval body cavity have become obliterated, together with the
portion of the nephridial canals running in the septum. While we do not wish to deny that
the remnants of the nephridial canals and their external openings, situated originally in the
trunk cavity of the Actinotrocha, become part of the nephridia of the adult, yet in the stage
under consideration they could not be found. So far as we know, Ikeda is the only investigator
who has given us figures illustrating the relation between the larval nephridia and the nephridia
of the adult. While his tig. Qie shows the larval nephridial canals, his tig. 66, which is a cross

MEMOIRS OF THE NATIONAL ACADEMY OF SCIENCES. 10 1

section througb the anterior end of a young Phor&nis and which .shows a section through the
young nephridium of tin1 adult, doc- nol prove thai he is dealing with the same structure.

Vascular system. It will be remembered that the vascular system of the fully developed
Actmotrocka Specie- A. consisted of a dorsal blood vessel (ties. 51 f, g, h) running along the
median line of the stomach from the dorsal insertion of the mesentery, between the collar and
trunk. t«> the posterior end of the stomach, its lumen being a part id' the segmentation <-a\ ity; a
hunch of blood caeca formed at the posterior end of the stomach as e\ aginations of its splanchnic
mesodermal covering and a loose sac of mesodermal tissue arising on the somatic wall of the collar
segment and inclosing the larval collar cavity (ties. 50, 51,/, g, k). (See below for discussion of
the " post-oral ring sinus," ventral vessel and the ki ring sinus" at the junction of the stomach
and intestine.)

There are several important points in the vascular system of the Actinotrocha which must
be taken into account in order to understand its metamorphosis into the vascular system of the
voung Pkoronis. First, that the dorsal blood vessel, which is formed from the splanchnic meso-
dermal lining of the trunk cavity, incloses a part of the space between the lining and the wall of
the alimentary canal — i. e., the segmentation cavity — second, that this vessel dwindles away pos-
teriorly and opens into the space between the lining and the wall of the alimentary canal; third,
that the wall of the stomach in the collar segment is practically free from mesodermal lining
(ties. 51 </. /'). and that the larval collar cavity, with its somatic mesodermal lining, is a blood
sinus: fourth, that the larval collar cavity is a part of the segmentation cavity; and, fifth, that
during metarnorphosi- the act of drawing the stomach and intestine into the cavity of the ventral
pouch causes pressure to be exerted on the larval collar cavity.

When the critical stage is being passed through, the blood-corpuscle masses break up and
they are driven by the pressure on the collar cavity to the points of least resistance. As a rule
some of the blood corpuscles are squeezed up into the dorsal region of the collar cavity where
the dorsal blood vessel ends, and invariably some of the blood corpuscles pass from the larval
collar cavity into the cavity between the wall of the alimentary canal and its mesodermal covering.
In fact, a- soon as the critical stage occurs, the splanchnic mesodermal lining in all regions
becomes separated from the wall of the alimentary canal and thus allows the blood corpuscles t i
move about between these two layers throughout the extent of the alimentary canal.

The dorsal blood vessel (" mediangefasz " (Cori), "afferent vessel" (Benham), and the ring
vessel with its tentacular vessel are completely formed structures at this stage. The dorsal
vessel is still freely open posteriorly into the space or sinus between the stomach wall and its
mesodermal covering and blood corpuscles are carried back and forth from it to the sinus by the
contraction and expansion of the former. Anteriorly the dorsal vessel can plainly be seen
opening into the ring vessel (larval collar cavity.)

The origin of the connection between the dorsal vessel and the ring vessel and the manner
in which the blood corpuscles find their way into the former are questions which have not been
ver\ satisfactorily elucidated. Actinotrocha Species A., does not present any great difficulties
in the way of understanding how- these processes take place. The dorsal blood vessel opens
posteriorly into the sac-like sinus around the loop of the alimentary canal, and it seems probable
from an examination of sections of the critical stage that it is also open anteriorly. Assuming
that such is the condition, it will open into the space between the mesodermal lining and the wall
of the gut. This space, however, is in free communication with the larval collar cavity (adult
ring vessel) which contains the blood corpuscles. Under these condition- the blood corpuscles
can pass into the dorsal blood \essel from either end.

Masterman (15) and Roule (20) both describe a vessel on the ventral stomach wall of the
Actmotrocka. Wehave not found this yes-el in the Actinotrocha, nor do we find it in sections of
the critical stage.

At this time there is but one ring vessel in the supraseptal cavity, but we consider that it
represents both the receiving and distributing vessels of the adult Pko?'onis.

Shortly after the critical point in the metamorphosis, the mesodermal lining on the left side
of the oral limb of the U-shaped alimentary canal begins to -how indications of becoming a blood
Vol. 10— Xo. 4—05 3

102 MEMOIRS OF THE NATIONAL ACADEMY OF SCIENCES.

vessel, and when the metamorphosis is completed ;i definite vessel is seen, which opens posteri-
orly into the spacious blood sinus around the loop of the alimentary canal. Anteriorly before
reaching the transverse septum, it divides into two branches, which run obliquely upward along
the sides of the alimentary canal, almost encircling the same; these finally open into the ring
vessel of the supraseptal cavity (fig. 60). The vessel described becomes the efferent vessel of the
adult (figs. 78, 77. 75, 66) and its branches become part of the recipient vessel.

A> Ikeda has pointed out, the efferent vessel of the adult corresponds to the ventral vessel
which Masterman (15) and Roule (20) have found in the Actinotrocha before metamorphosis.

In all the completely metamorphosed Aetin<itr<><Ii;< that have heen sectioned there is but one
ring vessel, hut the young 1 Vmrmi is, when it is 12 hours old, possesses both the recipient and
distributing vessels; these vessels, we believe, arise from the single ring vessel of the metamor-
phosing Actinotrocha by the fusion of its walls and by the subsequent separation of the two
parts along the line of fusion.

Masterman, in his description of the blood system of the Actinotrocha, speaks of a "ring
sinus" at the anterior end of the intestine which connects the dorsal and ventral vessels. He
also says that there are two lateral branches of the dorsal vessel in the region of the pharynx
which pass downward around the oesophagus ("'post-oral ring sinus") and become continuous
with the ventral vessel.

The former undoubtedly represents the sinus surrounding the loop of the alimentary canal
in the young Phoronis, while the latter, no doubt represents the branches of the efferent vessel
which become part of the recipient vessel of the adult. Masterman says that these branches
open into the dorsal blood vessel, but such is not the case in the completely metamorphosed
Actinotrocha.

From a comparison of Masterman's description of the vascular system of the Actinotrocha
and Ikeda's and our own description of the same before and after metamorphosis, it is seen at once
that this system develops more precociously in the form that Masterman studied. This condition,
together with the facts that the lumen of the blood vessels are parts of spaces between the wall
of the gut and its mesodermal lining and that the mesodermal lining of the alimentary canal tits
loosely while the blood system is developing, gives additional weight to Masterman's statement
that the dorsal vessel opens into the so-called "subneural sinus." However, in the Actinotrochse
that we have examined such a connection does not exist, and, as stated above, a "subneural sinus"
or cavity caused by a lack of contiguity between the mesodermal wall of the preoral lobe and
that of the collar cavity is not present.

THE ADULT PHORONIS ARCHITECTA.

Phoronis architecta was discovered by Andrews (1) in June, 1SS5, at Beaufort, N. O. and
he described it as a new species, giving it the specific name "architecta," on account, no doubt,
of its building a beautiful, straight tube. He finds that the tubes are made up of a clear, firm,
chitin-like membrane covered with small, clear grains of sand, and he thinks that these grains
are selected by the animal. Specimens collected from different localities in Beaufort Harbor
vary considerably in regard to the character of the sand grains and quite often small fragments
of dark shells are found mixed in with the latter. Occasionally two tubes occur cemented
together; but this condition is rare, for they are usually isolated and embedded perpendicularly
in the sand. When the specimens are brought into the laboratory and put into aquaria with
sand and water, they usually ciawl out of their tubes and begin to form new ones. Long-
champs (13) has lately pointed out that the tube is formed by a secretion from the posterior end
of the animal and not from the anterior end, as Cori has said. This is the case for P. architt «■/</.

Above it is stated that the tubes are straight, but where new tubes are formed in the aquaria
they are always twisted to a considerable extent, and they are attached firmly to the bottom
of the jar. In its natural habitat. Phoronis iir<-hit<<-tn does not have a firm substratum to which
to cement its tube, but it is seen from the above observation that when a solid surface presents
it-elf. til.' tube may take on the condition found in some of the other species of Phoronis which
are attached to rocks and shells.

MEMOIES OF THE NATIONAL ACADEMY OE SCIENCES. 103

Phoronis architecta lives at about low-water mark on the sand shoals, which are very
numerous in Beaufort Harbor, and, as a rule, the individuals occur in patches. Three or four
hundred specimens are often found within a radius of 4 or 5 feet, bul one is very apt to find
isolated specimens while digging in the sand anywhere in the harbor.

( >nl\ rarely do the tubes project a hove the surface of the sand as Andrews ( 1) has described,
and in these cases the condition was due to disturbances of the surface of the sand, such as
hollows made by Oallinectes. Usually the upper end of the tube is from 3 to 5 cm. below the
surface of the sand.

The average length of these tubes is 13 cm., and the average width a little over 1 mm. The
adult when removed from its tube is about 1 mm. in diameter in the posterior one-third, and
slightly less in the anterior two-thirds (fig. 61 ). The length of specimens taken out of the tubes
varies with the amount of contraction from 20 to 25 mm., which figures are considerably lower
than the length given by Andrews (about 50 mm.). The specimens which Andrews described
must have been considerably more extended than any we have preserved. When the animal is in
its natural habitat and undisturbed, however, it is capable of great extension, stretching the whole
length of the tube and even considerably farther, so that its lophophoral end may project above
the surface of the sand and reach for some considerable distance along its surface. We have not
been able to preserve specimens in their extended condition, and they usually contract to from
20 to 25 mm. in length.

The anterior two-thirds of the living specimen has a flesh color, while the posterior one-third
i- dark-yellowish red and quite opaque, which is due to the fact that the gonads and blood caeca
arc situated in this region. In preserved specimens, the body wall is annulated (fig. 61), but
such is not the case probably in the fully extended individual.

The crown of tentacles is quite simple compared to the crown of tentacles in P. austraMs,
P. Jmskii, and P. pacifica. A cross section shows that it is crescentric and that the ends are not
spirally coiled (figs. 62. 63, 64).

Andrews (1) lias given us a description of the principal points in the anatomy of l'h<>r<>nix
architecta, which he has undoubtedly made brief because of the resemblance to the anatomy of
Phoronis australis as described by Benham. (2). In general our observations agree with those
of Andrews, but there are a few points which merit discussion.

Lophophoral organs. — These peculiar organs (fig. 62) have been observed in several different
species of Phoronis, and although functions for them have been suggested, the observations do
not seem to have extended over a long enough period in the adult life of the worm to warrant a
definite statement as to their function.

The lophophoral organs (tig. 62) lie one on each side of the median line within the concavity
of the lophophore. They are outgrowths from the base of the inner row of tentacles, and. in
8ome species at least, are quite conspicuous organs, but they do not arise until the Phoronis has
reached its adult size. Organs located in the above region have been described for eight species,
but the size and shape do not seem to be the same in all. Whether these differences are specific
or whether the observations have been made at different periods in the adult life it is hard to
say. Lophophoral organs like those present in Phoronis architt eta arc found in /'. psammophila,
P. pacifica, J', rnulleri, and, no doubt, in some other species also. It seems, however, prob-
able from the description of the anatomy of P. huskii and /'. <hisir<ilix that in these species
the lophophoral organs are much less highly developed than in the smaller species with fewer coils
in the lophophoral crown. We have examined several specimens of J', australis with and without
genital products, but in no case have we seen organs such as are present in P. architecta.

Various functions have been assigned t<> the lophophoral organs. Mcintosh (14), working
with /'. buskii, considers that they are sensory in function, while Masterman (16), who has studied
the same species, says that they are glandular and that they give rise to mucus which serves to
hold the embryos together in masses. In other words, he consider- them to be "subsidiary repro-
ductive organs." Benham (2), who worked on P. australis, and Cori (4), who investigated
/'. psammophila, both give these organs a glandular function, while Andrews (1) thinks that

104 MEMOIRS OF THE NATIONAL ACADEMY OF SCIENCES.

in /'. architecta they are used in building- the tubes. H. B. Torrey (22) has made the interest-
ing observation that in some specimens of P. pacifica the lophophoral organs are like those in
/'. mis! riil 'is, while in others they are like those in /'. iirrliiti <iu. It is probable that in part of
the specimens which Torrey examined the lophophoral organs were not full grown, while in the
rest they had reached their full development.

Our observations indicate that Andrews' supposition concerning' the tube-building function
of the lophophoral organs can not be held. Individuals without the lophophoral organs
build new tubes covered with sand grains just as do those with the organs, and young specimens
of Phoronis architecta, which are in possession of tubes, never have lophophoral organs. These
two facts prove quite conclusively that the lophophoral organs of Phoronis architaiu are not
used for tube building.

Masterman's view that the lophophoral organs of /'. oushii arc glandular, and that they fur-
nish mucus to hold the eggs and embryos in masses can not be applied to P. arcltit<cta, for in
this species the eggs and embryos are not held in masses.

Specimens of Phoronis architecta have been examined during almost every month in the year
in order to discover whether or not there is any relation between the lophophoral organs and the
breeding season. During the months of June, July, August, September, and October examina-
tion of many specimens of Phoronis architecta shows that more than one-half are with lopho-
phoral organs — i. e., with the "carpel-like organs" and the "spherical sense lobes" (Andrews).
Examination of specimens taken during these months shows that some contain ovaries and eggs
while others do not, and that all contain spermatozoa in the body cavity, but that only those
without ovaries contain testes. Occasionally an egg floating freely in the body cavity is found
in specimens with testes. These facts are correlated with the presence or the absence of the
lophophoral organs, for these organs are present in specimens with testes and without ovaries
and absent in specimens with ovaries and without testes.

During the latter part of December and the first part of January specimens of /'. architecta,
some of which were collected in Beaufort Harbor at that time and some of which had been kept
alive in the laboratory of Johns Hopkins University since the summer months were examined.

Many of these specimens (30 or 40) were examined by crushing the posterior end and also
by cutting sections, but with one exception all of these individuals were found to be without
ovaries and testes. In the case of the exception, a few ovarian eggs were present, but the ovaries
were still very young. The blood ewea at this time are surrounded by a great abundance of the
peculiar peritoneal tissue, which later gives rise to the reproductive organs. Lophophoral organs
were absent both in specimens collected at Beaufort during the first part of January and in speci-
mens collected during the summer and kept in the laboratory.

During the months of February, March, and April the specimens in the aquaria at Johns
Hopkins University were examined quite frequently, but until March or April there was no sign
of lophophoral organs. Then they began gradually to develop in some specimens until by the
first of May they were full size. At this time another lot of live material was received from
Beaufort which afforded some very interesting observations. The number of individuals with
and without lophophoral organs were in about the same ratio as during the summer months.
At this time, in specimens with lophophoral organs, the testes are present but the ovaries are
not. while specimens without lophophoral organs possess ovaries and contain ovarian c^'^.
Quite often specimens with lophophoral organs have large bunches of spermatozoa floating freely
in the body cavity, and in sonic cases these occur inside the nephridia. In one individual a large
bunch of spermatozoa was found lodged in the end of the lophophoral organ's pocket-like
cavitx .

Judging from our observations, it seems that the relation of the lophophoral organs to the
breeding season is as follows: Some adults are giving rise to eggs throughout the months of
May, -lime, July, August, September, and October. None of the individuals arising from
these eggs become sexually mature until March or April of the next year. Those which are
the oldest i. e.. those bom in the early months of the year before — develop testes and lopho-
phoral organs in .March or April. Then they lose their lophophoral organs, the testes disap-

MEMOIRS OF THE NATIONAL ACADEMY OF SCIENCES. 105

pear and ovaries begin to develop about the. first of May. While this is going on, individuals
which were born later in the summer of the year before begin t<> develop testes and lophophoral
organs, and thus we have individuals with lophophoral organs and testes occurring at the same
time of the year as individuals without lophophoral organs and with ovaries.

While there is do absolute proof, the upper part at least of the lophophoral organs prob-
ably functions as a kind of seminal receptacle. We are led to this conclusion by these facts:
First, that the organs appear only when the testes are present; second, that large bunches of
spermatozoa have been found in the body cavity, in the nephridia, and in the cavitj* of the lopho-
phoral organs; and, third, that there are ciliated grooves leading from the nephridial pores to the
cavities of the lophophoral organs.

Vascular system. — Nearly all of the early investigators of the anatomy of Phoronis recognize
the existence of an efferent and afferent vessel which are in connection with vessels running up
into the tentacles.

Caldwell's (3) description, although brief, is complete, and differs very little from later ones.
Cori's (-1) account seems to be about the same as Caldwell's; however, he recognizes one ring
vessel instead of two and describes in more detail the relation between the tentacular vessels and
the ring vessel.

In Phoronis austral/is, Benham (2) finds the circulatory system much the same as Caldwell
does in the form that he worked on. Practically the only point of difference is that he describes
the tentacular vessels as dividing into two branches, one opening into the distributing vessel
(inner) and the other into the recipient vessel (outer).

Andrews (1) finds that the vascular system of P. architecta, as far as he has determined, is
like that of /'. australis, while Ikeda (9) says that Benham's description holds good for /'. ijimai
and P. hippocrepia.

A transverse section through the lophophoral crown of /'. arvhitecta (fig. 63) shows that the
cavity of each tentacle contains a blood vessel which is attached to the inner surface of the wall.

At the base of the tentacles a cross section shows that there are two blood vessels running
parallel to one another through most of their course around the cavity of the lophophore (tigs. 65
to 71). These vessels are distinct, although closely applied to one another, thus differing from
what Cori finds in /'. psammophila. The outer vessel and inner vessel (figs. 65, 66) are, respec-
tively, the "recipient" and " distributing" vessels which Benham describes. In tig. 83 is shown
a ci'o-s section through the base of the tentacles. Throughout most of the section the tentacular
vessels open into the outer or recipient vessel, but at one end the tentacular vessels open into
the inner or distributing vessel. This section, together with sections anterior and posterior to
it, show conclusively that the tentacular vessel has two separate openings, one into the distribu-
ting vessel, the other into the recipient vessel, and that the distributing and recipient vessels are
completely separate. A longitudinal section through the anterior end of Phoronis architecta
shows conclusively that the tentacular vessel divides into two branches, one opening into the
recipient vessel and the other into the distributing vessel.

A little moi-e posteriorly the ring-like distributing vessel opens into a median longitudinal
vessel lying between the oesophagus and rectum but close to the wall of the former (tigs. 67, 68).
This vessel, which is the afferent vessel, pierces the transverse septum (tig. 69) and runs
posteriori}* (within the rectal or posterior chamber) between the two arms of the alimentary
canal. At the point where the vessel passes through the septum there is a thick layer of muscle
fibres surrounding flic former which undoubtedly has the power of shutting off the blood supplj
to the tentacles and which may be very necessary to prevent the animal from bleeding to death
when the lophophoral crown is cast away (tigs. 68 to 71).

The two sides of the ring-like recipient vessel do not pass into a single vessel while they are
within the supraseptal cavitj (tigs. ti7 to 71), but after they have pierced the transverse septum
the right side of the ring is seen to pass diagonally across the (esophagus and to meet the left

side of the ring (figs. 75 to 7S). From this point the two becom< vessel, the efferent vessel,

which runs posteriorly within the left body cavity. In the posterior part of the body, where
the alimentary canal makes a loop, the efferent and afferent vessels are continuous and open into

106 MEMOIRS OF THE NATIONAL ACADEMY OF SCIENCES.

a sinus around the stomach. Along most of the course of the efferent vessel blood ceca are
given off. and a large bunch also arises from the sinus.

Nervous system.— We have found that the nervous system of the Actmotrocha is more
highh developed in some species than in others and that it is subepidermal in character. In the
different species of the adult we also find that there is considerable difference in the degree of
development of the nervous system and that it is largely subepidermal.

Caldwell (3) was the first to give a good description of the nervous system, although
Kowalevsky (11) recognized the existence of a lateral nerve and ganglion. Caldwell found the
cing nerve, a hollow nerve cord on the left side, and he speaks of two ciliated pits consisting of
nerve cells, ganglion, and nerve fibres. The description is so brief, that one can not say whether
or not the ganglion that he speaks of represents the ganglion that Kowalevsky (11) and Cori (4)
describe.

Benham (2) finds no ganglion in P. austiraUs, but describes two small areas which, it seems
probable, are the same as Caldwell's " ciliated pits." He is the first to recognize the existence
of a lateral nerve on the right side as well as on the left, and he finds a nerve ring with a nerve
to each tentacle arising from it.

Cori (4) describes a definite ganglion, a lateral nerve on the left side only and tentacular
nerves. He is the onlj- investigator who has published anything on the distribution of the ner-
vous tissue in the lophophoral organ.

Andrews (1). Torrey (22), and Ikeda (9) have given very brief descriptions of the nervous
system, but the two former recognize the existence of a short lateral nerve on the right side as
well as a long one on the left side, while the latter speaks of a so-called brain ganglion and
nerve ring.

The account which Andrews gives of the nervous system of P. wchiteeba is very brief,
since his paper deals only with the description of a new species. He only speaks of the lateral
nerve and makes no mention of a brain ganglion, ring nerve, tentacular nerves, or nerves to the
lophophoral organ.

In general our observations on the lateral nerve of P. a/rchitecta agree with those of Andrews
and Torrey. The lateral nerve of the left side is quite conspicuous and extends from the anterior
end to a point about one-third from the posterior end of the animal. It runs along the lateral
body wall until it is almost in the region of the transverse septum, then it gradually passes
obliquely upward in close proximity to the left nephridial canal, and finally is seen embedded in
the ectoderm at the side of the anal papilla. From this point it passes around the base of the
anal papilla between the anus and the mouth, and then it begins to take the same course close to
the nephridial tube on the right side as it did on the left side, but it soon grows much smaller in
diameter and finally disappears (tigs. 78 to 67). A longitudinal section passing through the mouth
and aims shows the relation which the nerve cord bears to the ganglion and nerve ring (fig. 84).
Cori (4) figures such a section through P. psammophila, but he seems to have overlooked the
nerve cord or axis cylinder in this region. It is closely associated with the cells of the ganglion
and lies just a little, below the latter. In an oral direction from the ganglion is seen a section
through the nerve ring.

If a cross section (fig. 85) is taken through the ganglion so as to cut longitudinally through
the nerve cord and if the section is stained deeply with iron hsematoxylin and eosin, it will show
plainly that there is no cavity in the cord, but that it is made up of a mass of fibres surrounded
by a nucleated sheath. Caldwell (3) considers the structure to he a hollow nerve cord; Benham
(2) says that it has semifluid contents and that he has been unable to make out any punctated
nerve substance; and Cori (4) states that it is an axis cylinder.

We have endeavored to find some connection between the cord and the ganglion, but have not
been very successful. In the region of the ganglion — i. e., between the mouth and the anus —
the sheath of the nerve cord does not seem to differ in thickness or character from the same
structure in other parts. The cells of the ganglion, however, send out processes which in
sections are frequently seen applied to the sheath, but no connection between the fibres of the
nerve ring and those of the cord could be made out.

MEMOIRS OF THE NATIONAL ACADEMY OF SCIENCES. 107

Kowalevsky (11), Cori (4), ami Torrey (22) have all found the nerve ganglion, while Benhain
(•_') denies its existence in /'. avstralis. It undoubtedly exists in /'. architecta, is situated at the
base of the anal papilla between the anus and the mouth, and lies above the nerve cord between
the anal papilla and the nerve ring (fig. 84).

The ganglion consists of nerve fibres and nerve cells and the latter have at least two
processes. While it is a definite structure back of the anal papilla, on the sides it diminishes in
size until its cells become indistinguishable from those of the nerve ring. In fact, all of the
ectoderm forming the sides of the groove between the anal papilla and the base of the lophophore
is rich in nerve fibres and cells.

The nerve line' follows the base of the lophophore on the outer side throughout its extent,
and in the inner part of the horseshoe it is quite rich in nerve cells whose processes can be seen
penetrating into the mass of fibres (figs. 67 to 74). This ring represents the collar nerve ring of
the Actinotrocha.

There is a definite tract of nervous tissue running up the inner side of the tentacle, but we
are not prepared to say that it is a nerve running from the ring, although it is nervous tissue
which is undoubtedly continuous with that of the nerve ring.

Cori (4) has carefully studied the anatomy of the lophophoral organ of /'. psammophila and
we have nothing to add to his description at present. We are also unprepared to say whether or
not the second layer of the lophophoral organ consists of nerve cells. As he has described, they
have long prolongations which extend from the cells of the inner layer to the outer, and these
processes form a rather marked layer just below the epidermis on the outer surface of the organ.
At the base of the lophophoral organ these prolongations seem to be intimately associated with
nerve fibres which can be traced to the nerve ring.

Throughout the body wall of the trunk there is a subepidermal layer of nervous tissue.

Nephridia. —We have nothing newto add concerning the adult nephridia, but our observations
on P. architecta confirm those of Benham (2) for P. austraUs. The nephridial canals lie embedded
in the ectodermal wall in the region of the rectum. Each opens to the exterior through a pore
at the side of the anal papilla. Following the canal from the nephridial pore, we see that it
passes downward — i. e., posteriorly — for a short distance and then bends upon itself running
upward parallel to the descending arm. A short distance above the bend it opens by one funnel
into the lateral body cavity (tig. 72) and by another into the rectal body cavity (tig. 70).

Reproductivt <>r<i<iin<. — Ikeda's recent paper (10) on the reproductive organs of Phoronis
gives a good account of the anatomy and development so we shall not enter into a description of
them. We are able to confirm Andrews's observations that the male organs develop at a different
time from those of the female.

Ciliated ridgt oftTu alimentary canal. — Andrews has described a ridge running along the
inner wall of the oral branch of the alimentary canal (fig. 81). H. B. Torrey has found the
same structure in P. pacifica, and we can confirm Andrews's observation for P. architecta. This
ridge does not seem to have any rudiment in the Actinotrocha, and it is not present just after
metamorphosis.

SUMMARY.

The male and female reproductive organs do not develop at the same time in /'. architecta,
and the indications are that it is a protandrous animal.

Fertilization is external and the eggs are not held in lophophoral masses by the tentacles.

Segmentation is holoblastic and equal, but cleavage does not occur simultaneously in all the
blastomeres. During the division of the four-cell stage into the eight -cell stage, the upper four
blastomeres rotate in the direction of the hands of a watch. The sixteen-cell stage arises from
the eight-cell stage by a meridional division of each blastomere.

The blastopore is eccentric from the beginning of gastrulation and the ganglion of the
Actinotrocha makes its appearance at this time. As development proceeds, the blastopore gradu-
ally closes up from the posterior end toward the anterior end of the larva until finally it becomes
a transverse >lit.

108 MEMOIKS OF THE NATIONAL ACADEMY OF SCIENCES.

The "primitive streak" of Caldwell does not seem to be present in the larva of P. architecta.

The "nephridial pit" is of ectodermal origin.

The mesoderm arises, for the most part, from the lips of the blastopore. Archenteric
diverticula are not present in the larva of /'. architecta, but there is a sac-like formation of
mesoderm cells in the anterior end which forms the lining of the preoral lobe and which gives
rise to a mesentery between the lobe and collar cavities.

The lining of the collar cavity does not arise from a mesodermal sac. It is formed by
isolated mesoderm cells which arrange themselves on the somatic wall leaving the splanchnic wall
practically without any lining.

In the larva of /'. architt eta the mesodermal lining of the trunk cavity is complete, covering
both the somatic and splanchnic walls, and it seems probable that it arises from cells forming the
base of the nephridial diverticula. There is a mesentery between the cavities of the trunk and
collar.

A stomodseum and proctodeum are not present. The blastopore becomes the mouth, the
anus arises quite late in the early life of the embryo, and the rectum is formed as an outgrowth
of the blind end of the archenteron.

The nephridial canals, at least, have their origin in a single median pit which soon branches
into two intercellular tubes. We have not found anv evidence that the excretory cells of the
nephridia are formed from free mesoderm cells attaching themselves to the blind end of the
nephridial canals.

The " neuropore" and "subneural gland," which Masterman has described, do not exist in the
Actinotrocha examined, although imperfectly preserved specimens show unusual structures which
might lie taken for these organs.

Masterman's "subneural sinus" is not present either, although there is a space below the
ganglion which is free from mesodermal strands. The "atrial grooves''' which Masterman says
exist are present in the larvae we have studied, but we can not consider that they have the signifi-
cance that he assigns to them. Occasionally grooves are found which might be comparable to
Masterman's "oral grooves, '" but they are due to imperfect fixation. The stomach diverticula
exist in one species that we have examined, but they do not impress us as being of notochordal
nature, as Koule and Masterman have claimed.

There is a subepidermal layer of nervous tissue throughout the body. Extending anteri-
orly from the ganglion, which is situated on the median dorsal surface of the hood, are three
longitudinal nerves, which finally become continuous with a nervous ring running around the
edge of the hood. From the posterior side of the ganglion two parallel tracts of nerve fibres
issue and pass posteriorly along the dorsal collar wall until they reach the circle of tentacles,
where most of them follow the line of insertion of the collar trunk mesentery, and give rise to a
collar nerve ring. The nerve fibres from the edge of the preoral hood do not pass up to the
ganglion from the point of attachment of the hood on to the collar wall, as Masterman has
described, but they make a sharp turn, running posteriorly and obliquely along the lateral and
ventral wall of the collar, where they form two definite nerve tracts which become lost in the
region of the collar nerve ring. While there may be nerve fibres passing from the ganglion out
in all directions over the surface of the hood, we have not been able to make them out, nor do we
find any definite nervous tract running along the dorsal or ventral wall of the trunk segment.

There is one pair of retractor muscles extending from the region of the ganglion to the
collar walls, in the region of the first and second pairs of tentacles, and besides these, in one
Actinotrocha that we have examined, there is another pair extending from the sides of the ganglion
to the. ventral w all of the hood. In this latter Actinotrocha there is an extensive layer of muscle
fibres in the wall of the hood and also a ring of fibres around the edge of the latter. A pair of
longitudinal muscle tracts extend from the region of the ganglion, along the dorsal wall of the
Actinotrocha, to the perianal ring, and there is a similar pair of tracts extending along the ventral
wall of the collar and trunk. A ring of muscle fibres run parallel with the ring nerve, between
the collar and trunk segments. Beside these muscle tracts there is a layer of circular fibres in

MEMOIRS OF THE NATIONAL ACADEMY OF SCIENCES.

109

the \v:ill of the collar and trunk, lying between the longitudinal fibres and the ectoderm. There
is also a covering of muscle cells on the ventral pouch and on the wall of the dorsal blood vessel.

The nephridia have much the same structure as those of Amphioxus, as described by Good-
rich. In one of the Actinotrocha from Beaufort Harbor the nephridial canal branches, but in the
other it does not. Nephridial funnels do not exist, but the ends of the canals open into tubular
cells, and the lumen of each cell contains a flagellum. The nephridial canals open to the exterior.
at the sides of the orifice of the ventral pouch. The. nephridia, which Masterman describes for
the preoral hood and trunk, are not present in any Actinotrocha that we have examined.

The blood vessels of the Actinotrocha are formed from the splanchnic mesodermal lining and
t hey inclose part of the blastoccele. There is a dorsal blood vessel opening (J) anteriorly into the
space between the stomach wall and the splanchnic lining. At the posterior end of the stomach,
where the dorsal vessel ends, there are csecal vessels formed as evaginations of the mesodermal
lining of the stomach. The dorsal vessel becomes the afferent vessel of the adult, while the
efferent vessel does not arise until just after metamorphosis. The collar cavity, which is a part
of the blastoccele, becomes the ring vessels and tentacular vessels of the adult. There is no
connection between the dorsal blood vessel and Masterman's "subneural sinus."

The blood corpuscles arise from the somatic mesodermal lining of the ventro-lateral collar
wall just in front of the septum. They make their appearance as two masses of cells bilaterally
placed, one on each side of the median ventral line, and as they develop they migrate across the
collar cavity and become applied to the naked walls of the stomach.

The rudiment of the supraseptal or collar cavity of the adult makes its appearance in about
the same region as do the blood corpuscles but a little later in the life history of the Actinotrocha.

During metamorphosis the following organs are lost: The preoral lobe, the ganglion, and
the larval tentacles. The ectodermal wall of the collar cavity, the stomach diverticula, the
digestive areas, and the perianal ciliated ring are not destroyed, but they lose their identity.
The subepidermal nervous layer of the trunk and ventral pouch becomes part of the same
tissue in the adult, but the larger part of this tissue, as well as the lateral nerve, the ganglion,
and the nerves to the lophophoral organs are new formations.

All of the nervous structures of the collar and trunk are lost during metamorphosis, except
the collar trunk nerve ring, which persists as the nerve ring of the adult.

The ventral mesentery becomes the (esophageal and rectal mesenteries of the adult, and the
cavities of the trunk and ventral pouch are transformed into the infraseptal cavity.

At least the greater part of the nephridia is lost during metamorphosis.

The lophophoral organs arise late in the adult life and are present only in individuals which
are with testes and without ovaries. They probably serve as seminal receptacles.

The vascular system of the adult consists of an efferent and afferent vessel, which are
continuous posteriorly by means of a sinus around the loop of the alimentary canal; of caeca!
\ essels as outgrowths from the afferent vessel and the blood sinus; of a distributing and recip-
ient ring vessel, the former opening into the afferent vessel and the latter into the efferent vessel ;
and of tentacular vessels, each of which divides into two short branches, one opening into the
distributing vessel and the other into the recipient vessel.

There is a ciliated ridge extending along part of the inner wall of the alimentary canal.

The nervous system of the adult is to a great extent subepidermal. There is a nerve with
a nucleated sheath extending along the left side of the animal. Anteriorly it bends around the
anal papilla and continues a- a short nerve on the right side. There is a ganglion between the
mouth and anus. -V nerve ring extend-, around the base of the lophophore and i! gives off
nerves to the lophophoral organs. There is nervous tissue in the walls of the tentacles.

The excretory organs are paired and each nephridium consists of a tube bent upon itself.
One end opens to the exterior, while the other is continued into two funnels, one communicating
with the rectal and the other with the lateral body cavity.

The reproductive organs arise from the lining of the csecal blood vessels, and the male organs
develop at a different time from those id' the female.

Johns Hopkins University, March, 1.904-

REFERENCES.

1. Andrews, E. A. On a New American Species of the Remarkable Animal Phoronis. Ann. and Mag. of Nat. Hist.

6th series, vol. 5, L890.

2. Benham, W, B. Tlic Anatomy of Phoronis Australis. Quart. Journ. Micro. Science, new series, vol. 30, 1889.

;'. Caldwell, W. II. Preliminary Note on the Structure, Development, and Affinities of Phoronis. Proc. of the

Hoy. Soc, vol. 34, 1882-1883.
3a. Caldwell, W. H. Blastopore, Mesoderm, and Metameric Segmentation. Quart. Journ. Micro. Science, vol.

25, 1885.

4. Com, C. .1. [Jntersuchung fiber die Anatomie und Histologie der Gattung Phoronis. Zeit. fur Wiss. Zoologie,

vol. 51, 1891.

5. Foettingeh, A. Note sur la- Formation du Mesoderme dans la Larva de Phoronis hippocrepia. Arch, de Biologie,

vol. 3, 1882.

6. Goodrich, E. S. On the Structure of the Excretory Organs of Amphioxus. Quart. Journ. Micro. Science, vol.

45, 1902.
Oa. Goodrich, E. S. On the Body Cavities and Nephridia of the Actinotrocha Larva. Quart. Journ. Micro.
Science, vol. 47, 1903.

7. Harmer, S. P. Appendix to Report on Cephalodiscus dodecalophus, by W. C. M'Intosh. Chal. Report, vol.

20 — Zoology.

8. Hatschek, B. Lehrbuch der Zoologie. 1888-1893.

9. Ikeda, I. Observations on the Development, Structure, and Metamorphosis of Actinotrocha. Journ. of the Col.

of Sci., Imperial Univ. Japan, vol. 13, pt. 4, 1901.

10. Ikeda, I. On the Development of the Sexual Organs and of their Products in Phoronis. Annot. Zoo. Japan,

vol. 4, pt. 4, 1903.

11. Kowalevsky, A. Anatomy and Development of Phoronis. St. Petersburg, 1867. See Arch, fur Naturgesch.,

vol. 2, 1807; Leuckart's Bericht.

12. Loxgchamps, M. de Selys. Recherches sur le Developpement des Phoronis. Arch, de Biologie, vol. 18, 1902.

13. Longchamps, M. de Selys. Beitrage zur Meeresfauna von Helgoland. Oldenburg, 1903.

14. M'Ixtosh, W. C. Report on Phoronis buskii. Chal. Report, vol. 27.

15. Masterman, A. T. On the Diplochorda. Quart. Journ. Micro. Science, vol. 40, 1897.

It;. Masterman, A. T. On the Diplochorda: III. The Early Development and Anatomy of Phoronis buskii
.M'Intosh. Quart. Journ. Micro. Science, vol. 43, 1900.

16a. M vsterman, A. T. Review of Mr. Iwaji Ikeda's Observations on the Development, Structure, and Metamor-
phosis of Actinotrocha. Quart. Journ. Micro. Science, vol. 45, 1902.

166. Masterman, A. T. Professor Roule upon the Phoronidea. Zool. Anz., No. 642, 1901.

17. Menon, K. R. Notes on Actinotrocha. Quart. Journ. Micro. Science, vol. 45, 1902.

18. Meisi iimkoi i , E. Dbei die Metamorphose einiger Seethiere. III. Uber Actinotrocha. Zeit. fur Wiss.

Zoologie, vol. 21, 1871.

19. Miller, J. Bericht iiber einige neue Thierformen der Nordsee. Arch, fiir Anat. und Physiol., 1840.

20. Roule, L. Etude sur le Developpement Embryonnaire des Phoronidiens. Ann. des Sci. Naturelles de

Zoologie, 8th series, vol. 11, 12, 1900-1901.
■_'l. Si m it/. E. Uber Mesodermbildung bei Phoronis. Trav. Soc. Natural. St. Petersburg, vol. 28, 1897.
22. Torrey, II. B. On Phoronis paciflca, sp. nov. Bio. Bui., vol. 2, No. 6, 1901.
I'M. W vgener, K. Uber den Bau der Actinotrocha branchiata. Arch, fiir Anat. und Phys., 1847.

24. Wilson, I'.. B. The Origin and the Significance of the Metamorphosis of Actinotrocha. Quart. Joum. Micro.

Science, vol. 21, 1881.

25. Schultz, K." Aus deiii Gebiete der Regeneration. III. Uber Regenerationserscheinungen bei Phoronis

miilleri Sel. Long. Zeit. fur Wiss. Zoo., Bd. 75, 1903.

26. Chvi.es, R. I'. Origin and Fate of the Body Cavities and the Nephridia of the Actinotrocha. Johns Hopkins

University Circular, April, 1!>04.

27. Cowles, R. 1'. Origin and Fate of the Blood Vessels and Blood Corpuscles of the Actinotrocha. Zoo. Anz.,

Bd. XXVII. No. 19, vom. 3. Juni 1904.

a Keceive< I after completion of paper.

Ill

REFERENCE LETTERS FOR FIGUKES.

r. b
r. c
r. d

v

in

;i anus.

a. c. c adult collar cavity (Ikeda ).

af. v afferent blood vessel.

a], c alimentary canal.

a. p anal papilla.

art ' artefact.

a. t adult tentacle.

a. t. m transverse septum of adult.

I>. c blood corpuscles.

l>. c. m blood corpuscle masses.

b. t basement tissue ( Benham).

cse blood caecum.

c. c collar cavity.

eg ciliated groove i Andrews).

c. w collar wall.

d. a digestive area.

d. m. t dorsal muscle tract.

d. v dorsal vessel (afferent).

d. ve distributing vessel.

ef. v efferent vessel.

eg egg.

e. h edge of hood.

ep epistome.

ex. c excretory cell.

fl flagellum.

g ganglion.

g. c ganglion cell.

g. in. c giant mesoderm cell.

in. c infraseptal cavity.

int intestine.

hit. n lateral nerve.

1. 1. c left lateral cavity.

1. 1. in left lateral mesentery.

1. 1. n lateral longitudinal nerve.

1. in longitudinal muscle.

1. r. in ring muscle tract of lobe.

1. r. n ring nerve tract of lobe.

1. t larval tentacle.

m mesoderm.

ill.1 mesentery between preoral

lobe and collar cavities.

m.'' mesentery between collar

and trunk cavities.

m. c. in mesodermal cell mass giv-
ing rise to blood corpus-
cles.

in. f muscle liber.

m. 1. n median longitudinal nerve.

m. s mesodermal sac of preoral

lobe.

n. c nephridial canal.

neph nephi'idiuni.

EXPLANATION OF FIGURES.

neph. o nephridial opening.

n. f nerve liber.

n. f.1 nephridial funnel into rec-
tal body cavity.

n. f.2 nephridial funnel into lat-
eral body cavity.

n. p nephridial pit.

nu nuclei of ciliated cells of

perianal ring.

oes. in (esophageal mesentery.

p. an. s perianal space.

pig pigment.

p. o. c preoral body cavity.

p. o. 1 preoral lobe.

p. r perianal ring.

p. re posterior recess (Ikeda).

subneural sinus (Master-
nian).

r rectum.

...ring blood vessel.
...rectal cavity.
...rudimentary dorsal mesen-
tery.

ret retractor muscle.

r. he right lateral cavity.

r. m rectal mesentery.

r. n ring nerve.

r. ve recipient vessel.

sen. p sensory papilla.

sp. in sphincter muscle.

st stomach.

sup. c supraseptal cavity.

t tentacle.

t. b. v tentacular blood vessel.

t. c trunk cavity.

th ventral thickening of ecto-
derm.

t. in. t tentacular ring muscle tract.

t. n. t tentacular ring nerve tract.

t, r tentacular ridge.

v vestibule.

ves glandular vesicles in ecto-
derm.

v. in ventral mesentery.

v. in. t ventral muscle tract.

v. n. t ventral nerve tract.

v. p ventral pouch.

v. p. o ventral pouch opening.

v. v ventral blood vessel (effer-
ent i.

v. w. h ventral wall of hood.

The objectives used in this work are the \ and \ Spencer and the ,'. Zeiss Oil Immersion. The OCCUlares used

are the •' <4"and" < 8 " Spencer and the No. 12 Zeiss Compensating.

PLATE I.

115

PLATE I.

Fig. 1. — Transverse section through nephridium of adult, fa Ob. X 4 Or. Camera. X440.

Fig. l\ — Unsegmented egg with one polar body. (From life.) ! Oh. X 8 Oc. Camera. X 360.

Fig. 3. — Two-cell stage, showing equal blastomeres. (From life. ) J Ob. X8 0c. Camera. X360.

Fig. 4. — Two-cell stage, showing unequal blastomeres. (From life.) J Ob. X8 0c. Camera. X 360.

Fig. 5. — Beginning of two-cell stage. (From life. ) 5 Ob. X 8 Oc. Camera. X360.

Fig. 6. — Section through two-cell stage. Chromosomes have lost their identity and granular vesicles have made
their appearance. Granular character of the yolk only shown in part of one blastomere. fa Oil
Immersion. > 8 0c. Camera. ■ 704.

Fig. 7.— Beginning of four-cell stage. (From life.) I Ob. X8 0c. Camera. X 860.

Fig. 8. — Four-cell stage, showing polar furrow. (From life.) 'Ob. X 8 Oc. Camera. X360.

Fig. 8a. — Section through four-cell stage, showing granular vesicles which make their appearance after the dis-
appearance of the chromosomes, j'j Oil Immersion. X 8 Oc. Camera. X 704.

Fig. 9. — Four-cell stage passing into eighth-cell stage. Seen from above and showing the twisting of the blasto-
meres. I Ob. X 8 Oc. Camera. X 360.

Fig. 10. — Eight-cell stage, showing rotation of blastomeres. (From life.) I Ob. X 8 Oc. Camera. X 360.

FlG. 11, — Section of eight-cell stage ready for sixteen-cell stage. Position of mitotic figures indicate meridional divi-
sion, fa Oil Immersion. XSOe. Camera. X 7(14.

Fig. 12. — Young blastula showing "blastocrele pore." (From life.) 5 Ob. X 8 Oc. Camera. X360.

Fig. 13.— Section through young blastula showing blastoccele pore. iV Oil Immersion. X 8 Oc. Camera. X 704.
116

MEMOIRS OF THE NATIONAL ACADEMY OF SCIENCES, VOL X

PLATE I.

Kit.'. :t.

Fig. 8.

Fig. 6.

Fig. 7.

Fig. i).

Fig. 10.

Pig -

Pig. 13.

1

PHORONIS ARCHITECTA

PLATE II.

Vol. 10— No. 4—05 i n:

PLATE II.

Fig. 14.— Optical section of an old blastula. (From life.) J Ob. X 8 Oc. Camera. X 360.

Fig. 11". — Section through an old blastula showing endodermal bodies in blastoccele. fa Oil Immersion. X 8 Or.

Camera. ;• 704.
Fig. 15.— Longitudinal section through older blastula than that of fig. 14k Bhowing proliferation of endoderm cells.

rV Oil Immersion. ■ 8 0c. Camera. X 704.
Fig. 15a. — Transverse section through a gastrula which is just beginning to gastrulate. Taken through ganglionic

thickening, fa Oil Immersion. X 8 Oc. Camera. X 704.
Fig. L56. — Transverse- section through same specimen as that of fig. L5a. Taken through the middle of the gastrula.

[■', ( )il Immersion. X 8 Oc. Camera. X 704.
Fig. 16. — Longitudinal section through a gastrula which has begun to elongate. -,', Oil Immersion. >. s Oc.

Camera. X 704.
Fig. 16a. — Transverse section through anterior end of a gastrula which was a little younger than that of fig. 16.

Taken just back of the ganglion. fa Oil Immersion. X 8 Oc. Camera. ■ 704.
Fig. 166. — Transverse section through same specimen as that of fig. 16a. Taken just in front of anterior end of the

archenteron. Shows mesoderm arising from the latter. jV Oil Immersion. X 8 Oc. Camera. X 704.

Fig. 1 tic. — Continuation of above series. Taken through anterior part of archenteron. fa Oil Immersion. X8

Oc. Camera. X 704.
Fig. 10'/. — Continuation of above series. Taken through middle of archenteron. TV Oil Immersion. X 8 Oc.

Camera. X 704.
Fig. 16e. — Continuation of above series. Taken through posterior part of archenteron. fa Oil Immersion.

X8 0c. Camera. X 704.
Fig. 16/.- Continuation of above series. Taken through the region where the lips of the blastopore are closing up.

i iil Immersion. • 8 0c. Camera. X 704.
Fig. Hi;;. — Horizontal section through a gastrula of the same age as that of liu'. 16. ,'. < >il Immersion. • s Oc.

Camera. X 704.
Fig. 17. — Transverse section through a gastrula showing peculiar granular cells arising from the endoderm. rV Oil

Immersion. X 8 Oc. Camera. X 704.
118

MEMOIRS OF THE NATIONAL ACADEMY OF SCIENCES, VOL. X.

PLATE

FIB. 18 (o)

Fig-- 17.

PHORONIS ARCHTECTA

"FS.Stl.

PLATE III.

119

PLATE III.

Fig. 17'/.— Same as fig. 17, showing one of the peculiar cells set free in the blastoccele. fa Oil Immersion. X 8 Oc.

Camera. X 704.
Fin. 18. — Transverse section through the anterior end of a young larva, which is slightly younger than the specimen

nf which fig. 19 is a longitudinal section. Blastopore is slightly oval in outline. Taken through the

ganglion, /.oil Immersion, x 8 Oc. Camera. ■ 704.
Fig. 18a. — Continuation of above series. Taken just in front of archenteron. fa Oil Immersion. • 80c. Camera.

X 7(14.
Fig. 186. —Continuation of above series. Taken through anterior part of blastopore, fa Oil Immersion. X 8 Oc.

Camera. • 704
Fig. I8e. Continuation of above series. Taken through posterior part of blastopore, fas Oil Immersion. X 8 Oc.

Camera. ■ 704.
Fig. 18d. — Continuation of above series. Taken immediately posterior to the blastopore. Lips have grown together

and there is a straight furrow corresponding to Caldwell's "primitive groove." faOU Immersion.

X 8 Oc. Camera. X 704.
Fig. 18e. — Continuation of above series. Taken posteriorly to section 18d. No sign of Caldwell's " primitive groove

or streak." -fa- Oil Immersion. X 8 Oc. Camera, x 704.
Fu;. 18/. — Horizontal section through same stage as 18 a, b, c, etc. fa i lil Immersion. X 8 Oc. Camera. X 704.
Fig. 19. — Sagittal section through larva which is a little older than that of series 18 a, b, c, etc. Shows mesodermal

preoral sac. TV Oil Immersion. X 8 Oc. Camera. • 7(i4. •
Fig. 20.— Sagittal section through a larva somewhat older than that of fig. 19. Blastopore has become circular again,

but much smaller than originally, fa- Oil Immersion. X 8 Or. Camera. X 704.
lui,. 20a. — Transverse section through larva of same age as that of fig. 20. Taken just posterior to the ganglion.

fa Oil Immersion: • 8 0c. Camera. X 704.
Fig. 206. — Continuation of series. Taken through blastopore. fa oil Immersion. 8 0c. Camera. X 704.

Fig. 20 < ontinuation of series. Taken just posterior to the blastopore. .Slight indication of groove, fa Oil

Immersion. ■ 8 Oc. Camera. X 704.
Fig. 20d.— Continuation of series. Taken near posterior end of larva. No groove, fa Oil Immersion. X 8 Oc.

Camera. ■ 7m.
Fig. 20e.— Horizontal section through a larva of the same age as that of figs. 20 a, 6, etc. Shows mesodermal sac

with posterior prolongations, fa Oil Immersion. X 8 Oc. Camera. X 704.
120

MEMOIRS OF THE NATIONAL ACADEMY OF SCIENCES. VOL X

PLATE

Big. 17 («)

Fig. 18.

Fig. 16

Fig. 18 (6)

Fig. 18 (c)

Big. 18 {d)

Fig. 18 (e)

Fig. It).

Fig. 1* /■

Fig. 20.

Fig. 20 (a)

Fig. !

Fig. 20 (C)

Fig. 20 (<J)

Fig. 20 (e)

PHORONIS ARCHITECTA

PLATE IV.

121

PLATE IV.

Fig. 21.-

Fig. 21a.

Fig. 22.-

Fig. 22a.
Fig. 226.

-Sagittal section through a larva somewhat older than that of fig. 20 a, b, c, etc. Blastopore oval and

transverse. Nephridial pit present at this stage. Archenteron has fused with posterior ectodermal

wall. T'j Oil Immersion. X 8 Oc. Camera. X 704.
—Continuation of above. Taken longitudinally and lateral to that of fig. 21.

Camera. X 704.
-Transverse section through a larva of the same age as the one shown in figs.

terior to ganglion. ,'.. Oil Immersion. X 8 Oe. Camera. X 704.
—Continuation of above series. Taken through blastopore, fc Oil Immersion. X80c. Camera. X 704.

Taken just posterior to blastopore. -,1., Oil Immersion. X 8 Oc. Camera.

T'r Oil Immersion. X 8 Oc.
II and 21d. Taken just pos-

Continuation of above series
X 7fi4.
Fig. 22c. — Continuation of above series. Taken halfway between the blastopore and the posterior end.
mesoderm cells on the ventral ectoderm. T\, Oil Immersion. X 8 Oc. Camera. X 704.
— Continuation of above series. Taken through rectum, fa Oil Immersion. X 8 Oc. Camera.
— Continuation of above series. Next section posterior to that of fig. 22tf. ,', Oil Immersion

Camera. X 704.
— Continuation of above series.
t't Oil Immersion. X 8 Oc.
— Continuation of above series.

X 704.
-Horizontal section tli rough a larva of the same age as that of fig. 21

X704.
— Same as fig. 23, but more ventral. TJj Oil Immersion. X 8 Oc.
-Longitudinal section through a larva somewhat older than that of tig. 21.
just made Its appearance. ,', Oil Immersion. X 8 Oe. Camera. X 704
Fig. 25. — Larva with two pairs of tentacles. (From life. ) J Ob. X4 0c. Camera.
122

Fig.

22d.

Fig.

22«,

Fig.

22/.

Fig.

22,/.

Fig.

23.-

Fig.

i'::-.

Fig.

L'4.-

Nest section posterior to that of fig. 22e.
Camera. X 704.
Section through nephridial pit

Shows the

X 704.
X 8 Oc.

Through wall of nephridial pit.
rl2 nil Immersion. X 8 Oc. Camera.
T'j Oil Immersion. X 8 Oc. Camera.

Camera.

■ 704.
Not quite sagittal.

X225.

Anus has

MEMOIRS OF THE NATIONAL ACADEMY OF SCIENCES. VOL X.

PLATE IV

Fig. 21 (a)

Pig. 21.

Pig. 22 lb)

Pig. 22 [a)

Pig. 22 (e)

Fig. 22 (d) „,.,

;///

Fig. 2H.

Fig. 24.

«.//.

Fig. 22.

Pig. 22 (e)

Vig. 22 (/)

Pig. ■-'

PHORONIS ARCHITECTA

PLATE V.

123

2»

PLATE V.

Pig. 26 —Horizontal section through a larva somew hat older than that of fig. 24, but younger than that of fig. 25.
Shows tli.- separation of the eel Is of the nephridial pit into two wings, ,'j oil Immersion. X 8 Oc.
Camera. ■ 7iM.
Fig. 27. — Horizontal section of posterior end showing slightly older stage in the development of the nephridia than

that of fig. 26. rj Oil Immersion. ■ 8 0c. Camera. X 704.
Fig. 28. — Horizontal section through a larva with beginnings of two tentacles. Shows nephridia. ,'. Oil Immersion.

X8 0c. Camera. X 704.
FlG. 29a. — Transverse section through posterior end of a larva with two pairs of tentacles. ,': Oil Immersion. X 8

Oc. Camera. X 704.
Fig. 296. — Continuation of series 29a. Taken through the region of the rectum. TV Oil Immersion. > 8 0c.

Camera. • 704.
Fig. 29c. — Continuation of series 29«. Taken through the middle of the larva. r'3 oil Immersion. X8 < >c.

Camera. X 704.
Fig. 29e. — Continuation of series 29r<. Taken through the body proper and the lower part of the hood. ,'. Oil

Immersion. X 8 Oc. Camera. -.704.
Fig. 29d. — Continuation of series 29a. Taken through body and hood near the region of the mouth. ,'_. Oil

Immersion. X 8 Oc. Camera. X 704.
Fio. 30. — Almost a sagittal section through a larva with two tentacles. TV OH Immersion. X 8 Oc. Camera. X 704.
Fig. 31. — Larva with three pairs of tentacles. Outline drawing from life. J Ob. X 4 Oc. Camera. X 225.
Fig. 32. — Larva with five pairs of tentacles. Outline drawing from life. 5 < >b. X 8 Oc. Camera. X 202.
Fig. 33. — Larva with six pairs of tentacles. Outline drawing from life. f Ob. X 8 Oc. Camera. X 202.
124

MEMOIRS OF THE NATIONAL ACADEMY OF SCIENCES, VOL X.

PLATE V.

Fig. 20.

n.),.

Fig. 28.

Fig. 28 {d)

p.o.l.

Hi.

Fig. 29 16)

l.i:

Fig. 28 (e

t.r.

Fig. 80.

Fig. 81.

PHORONIS ARCHITECTA

:rs.Djx

PLATE VI.

125

PLATE VI.

Fig. 34. — Actinotroeha Species A. (Drawn from life. ) §01). X 8 Oc. Camera. X 135.
Fig. 35.— Actinotrocha Species B. (Drawn from life.) § Ob. X 8 Oc. Camera. X 135.
126

MEMOIRS OF THE NATIONAL ACADEMY OF SCIENCES, VOL X

PLATE VI

<

I-
O

I

o

<

w

z
o
cr
o

I

CL

PLATE VII.

127

*s£&>

^S-

VJ

PLATE VII.

Fig. 36. — Nervous and muscular tracts of the dorsal surface of the hood. Actinotrocha Species B. (Drawn from

living specimen. )
Fm. :-!7. — Lateral view of anterior part of Aclinotrocha Species 1!., showing muscle tracts. (Drawn from living

specimen. )
Fig. 38. — Longitudiual section through the ganglion of an Actinotrocha. -,'_. < >il Immersion. 12 Zeiss Occulare.

Camera. X665.
Fig. 39. —Section through a ganglion cell in the collar nerve ring. Actinotrocha Species B. T:,- Oil Immersion. X8l

Oc. Camera. ■ 469.
Fio. 40. — Transverse section through the nerve tract of the ventral collar wall. Actinotrocha Species B. -/, Oi

Immersion. 12 Zeiss Occulare. Camera. X665.
Fio. 41. — Transverse- section through the dorsal nerve tract where it passes down along the bases of the tentacles.

Actinotrocha Species B. rV Oil Immersion. 12 Zeiss Occulare. Camera. • 665.
Fig. 42. — Transverse section through the collar nerve ring. Actinotrocha Species B. yV Oil Immersion. 12 Zeiss

< leonlaiv. Camera. ■ 665.
Fig. 43. — Transverse section through the nerve tract around the edge of the hood. Actinotrocha Species B. r\ oil

Immersion. 12 Zeiss Occulare. Camera. ■ 665.
Fig. 44. — Transverse section through the hood of .1 tinotrocha Species B. Taken through the sensory papilla. Hoo.l

aattenedout. ^ Ob. <40c. Camera. <300.
Fig. 44a.— Continuation of scries 44. Taken through the anterior part of the ganglion, i Ob. X 4 Oc. Camera.

■ 300,
Fie;. 44//. — Continuation of series 44. Taken through the ganglion which is invaginated by the action of fixing

agents. I Ob. X4 0c. Camera. X300.
128

MEMOIRS OF THE NATIONAL ACADEMY OF SCIENCES, VOL. X

sell p.

\

v.ir.li.
/

ret.

Ii.i-.ni .

ml /

\ 6

Pig. 36.

HI/.

r.li.l.

...•

•

i

-■/■■

Fig, 48.

* * ;

Fit.'. 40.

•v.

/.»./.

/./

" *^ ,.,.

(/.!»./.

)>.<>.<-.

III. I. II.

nh^0

Fig. 39.

Pig. 41

ll.lll.t.

\

1,1. f.

Fig. 4:i.

Pig. II.

</.;»./.

ll.lll.t.

Pig. 44 f„

Fig. 44 tfr)

PHORONIS ARCHITECTA

PLATE VIII.

L29

PLATE VIII.

Fie. 44r. — Continuation of series 44. Taken through the region where the edge of the hood passes into the collar

wall. ' f, < )b. X 4 Oc. Camera. X 300.
Fig. 44rf. — Continuation of series 44. Taken through the middle of the collar segment. I Ob. X 4 Oc. Camera.

X300.
Fig. 44. . — Continuation of scries 44. Taken through the bases of the ventral tentacles of the collar. J Oh. X 4 Oc.

Camera. X 300.
Fig. 44/. — Continuation of series 44. Taken immediately posterior to that of fig. 44c. } Ob. X4 Oc. Camera.

X 300.
Fig. 44;/. — Continuation of scries 44. Taken in the anterior part of the trunk segment. J Ob. X 8 Oc. Camera.

X 4S0.
Fig. 44/i. — Continuation of series 44. Taken through the rectum, i Ob. X 8 Oc. Camera. X 480.
Fm. 44/'. — Continuation of series 44. Taken through the anterior part of the perianal ring. £ Ob. X 8 Oc.

Camera. / 480.
Fig. 44/'. — Continuation of series 44. Taken through the posterior part of the perianal ring. J Ob. X8 Oc.

Camera. X 480.
130

MEMOIRS OF THE NATIONAL ACADEMY OF SCIENCES, VOL. X.

PLATE VIII

r.n.l

cp.o

/>■»./.

Fiji. 44 (,•]

d.m.i. si.

\\\»0 Flf. 44 («)

il.m.t.

Kit.'. 44 (/,.)

r il in

Fig. 4 I ./

Fig. 1 1 -

PHORONIS ARCHITECTA

PLATE IX.

131

PLATE IX.

Fio. 4r>. — Longitudinal section through the ganglion, showing lobe collar mesentery. Tj Oil Immersion. X4 Oc.

Camera. X 293.
Fig. 45a. — Longitudinal section through ganglion, showing retractor and mesentery, ,'j Oil Immersion. X 4 Oc.

Camera.. X 293.
Fig. 456. — Longitudinal section through the Adinotrocha, showing incomplete part of lobe collar septum. \ Ob.

- X 4 Oc. t 'amera. X 200.
Fig. 46. — Horizontal section through A dinotrocha Species B. ,', <>l>. ■ 4 Oc. Camera. X 200.
Fig. 47. — View showing muscles of inner surface of hood. From living specimen. § Ob. x 8 Oe. Camera.
Fio. 48. — Sagittal section through Actiitatmrhfi Species B. J Ob. X 4 Oc. Camera, x 150.
Fig. 49. — Longitudinal section through the posterior end of Adinotrocha Species B. rV Oil Immersion. • 4 <>,•.

Camera. ■ 293.
Fig. 50. — Longitudinal section through Adinotrocha Species \., showing relations of larval collar cavity and adult

collar cavity. ,' Ob. X 4 Oc. Camera. X 200.
132

MEMOIRS OF THE NATIONAL ACADEMY OF SCIENCES, VOL X

PLATE IX.

ret.

p.o.c.

Flk'. 45 (6)

Klff. 40.

Fl(?. 50.

PH0R0N1S ARCHITECTA

PLATE X.

Vol. 10— Xo. 4—05 5 • 133

PLATE X.

Fig. 51. — Transverse section through Actinotrocha Species A.
225.
Continuation of series 51.
— Continuation of series 51.
— Continuation of series 51.
— ( lontinuation of series 51.
— Continuation of series 51.
— Continuation of series 51.
--( lontinuation of scries 51.
ectodermal wall. Ob

Taken through ganglion. \ Ob. X4 0c. Camera.

Fig. 51a.
Fig. 516.
Fig. 51c.
Fig. 51A
Fig. 51e.
Fig. 51/
Fig. r»i,;.

Fig. -"-I/,.
Fio. 52.-
Fig. 52a.
Fig. 526.
Fig. 52c.
Fig. 52d.
Fig. 52e.
Fig. 52/
Fig. .vj;;.

\ Ob.

x 4 Oc.

< 'anient.

■ 225.

■ .11,

XlOc.

Camera.

■ 225.

' Ob.

X 4 Oe.

Camera.

225.

Ob

10c

Camera.

■ 225.

1 Ob.

■ J Oc.

Camera.

X 225.

i Ob.

■ 4 Oc.

Camera.

X 225.

Showing the lining drawn away from the

Taken through tin liar segment.

X4 Oe. Camera. ■ 225.

Continuation, of series 51. Taken through collar region just anterior to bases of the ventral tentacles.
,' lili 1 Oc. Camera. < 225.

Transvefee section through the nephridial canal. Actinotrocha Species 1*>. ^ Oil Immersion. 12 Zeiss

Comp. Occulare. Camera. X 1400.
-Continuation of series 52. Taken through the lower branch of the nephridial canal. ,'_. < »il Immersion.
12 Zeiss Comp. Occulare. Camera, x 1400.
Continuation of series 52. Taken through the end of the upper branch of the nephridial canal. rV Oil

Immersion. 12 Zeiss Comp. Occulare. Camera. > 1000.
Longitudinal section through one of the cellular processes at the end of the nephridium. Showing two
nuclei. rV Oil Immersion. 12 Zeiss Comp. Occulare. Camera, x 1000.

— Same as fig. 52c. Showing one nucleus. TV Oil Immersion.

xiooo.

Transverse section through the end of a process, /.nil Immersion.

■ 1000.

Transverse section through the proximal half of the processes.

Occulare. Camera. X 1000.

tudinal section through anterior end of a nephridium. Actinotrocha Species B. t\ Oil Immersion

12 Zeiss Comp. Occulare. Camera. X1000.
134

12 Zeiss Comp. Occulare. Camera.
12 Zeis> Comp. Occulare. Camera.
,'. nil Immersion. 12 Zei>s Comp.

MEMOIRS OF THE NATIONAL ACADEMY OF SCIENCES, VOL. X

PLATE X

,2 a)

PHORONIS ARCHITECTA

PLATE XL

135

PLATE XI.

Fig. ">o. — Transverse oblique section through Actinotrocha Species A. Taken through region of origin of blood

• corpuscles. 14 tentacles. rV Oil Immersion. <4 0c. Camera. <293.
Fig. r.4. — Transverse section through ventral cellar wall. Same specimen as that of fig. 53. Showing origin of blood

corpuscles. T'j Oil Immersion. 12 Zeiss ( 'oinp. Occnlare. Camera. ■ 935.

'.rocha Species B with ventral pouch evaginated. | Ob. X8 0c. Camera. X45.
Fig. 56. — Actinotrocha Species A. Immediately before metamorphosis. | Ob. X40c. Camera. X56.
Fig. 56a. — Actinotrocha Species A. Shortly after the beginning of metamorphosis. | Ob. X40c. Camera. X56.
Fig. 566. — Metamorphosed A ctinotrocha Species A. | Ob. X4 0c. Camera. <56.
Fig. 57.— Young specimen of Phoronis archiiecta ('.') with 30 tentacles. H Ob. X 4 Oc. Camera. X28.
Fig. 59. — Metamorphosed Actinotrocha Species A. Transverse section through the region of the transverse septum.

,', Ob. • 8 Oc. Camera. X 240,
Fig. 60. — Completely metamorphosed Actinotrocha Species A. Transverse section in region of branching of efferent

vessel. J Ob. X8 0c. Camera, x 240.
Fig. 61. — Adult Phoronis architects removed from tube. X8.
136

MEMOIRS OF THE NATIONAL ACADEMY OF SCIENCES, VOL X

PLATE XI.

lit jilt.

Fig. 58.

„/:,:

af.v.

Kl«. 57.

l\

J

d.v.

Pig. 80.

1

•

'<.

Pig. 54.

urn ./i.

Fig, 55.

h.f.m.

' ,l

u.t,

Pig. 56.

Fig. 56 (J)

I'.n.l.

r ..

in ph.

\l

""-V

p.r.

i.e.

Pig. 58 (a)

- sll/l.t.

Pig. 61.

Fig. 50.

PHORONIS ARCHITECTA

IPCowzss.fisL

PLATE XII.

137

PLATE XII.

Fig. 62. — Tentacular crown of Phoronis architecta. Showing lophophoral organs, epistome, and mouth. ,Drawn from
a living tentacular crown which had been constricted off from the animal. X 100.

138

MEMOIRS OF THE NATIONAL ACADEMY OF SCIENCES, VOL X

PLATE XII.

ft

w

.%

f

jj

V

I

S~)

Fig. 62

PHORONIS ARCHITECTA

PLATE XIII.

L39

PLATE XIII.

Fig. 63. — Transverse section through Phoronis architecta. Taken through tentacles, 'j Ob. X 8 Oe. Camera.

X130.
Fig. 64. — Continuation of series 63. Taken near base of tentacles. | Ob. <8 0c. Camera. X130.
Fig. 65. — Continuation of series 63. Taken through epistome. 'i Ob. X 8 Oe. Camera. > 130.
Fig. 66. — Continuation of series 63. Taken through anal papilla. § Ob. X 8 Oe. Camera. X 130.
Km. (17. — Continuation of series 63. Taken through nephridial opening. § Ob. X80e. Camera. X130.
Fig. 68. — Continuation of series 63. Taken through the transverse septum and below the nephridial openings.
3 ( )b. X 8 Oc. Camera. X 130.
140

MEMOIRS OF THE NATIONAL ACADEMY OF SCIENCES, VOL X.

PLATE XIII.

&

op

Fig. 63.

d.ve

05.

i/.n .

^. Fijr. 87.

</.rr.

(/.'» .

dan

sup,c.

Fig. 64.

d.ve.

iie/ili.

PHORONIS ARCHITECTA

PLATE XIV.

141

PLATE XIV.

Fig. 69. — Continuation of seriea 63. Taken a little posteriorly to that of fig. 68. § Ob. X8 0c. Camera. X 130.
Fig. 70. — Continuation of aeries 63. Taken through the nephridial funnel that opens into the rectal cavity. 1 Ob.

X S ( >c. ( Jamera. X 130.
Fig.. 71.— Continuation of series 63. Taken a little posteriorly to that of fig. 70. j| Ob. X8 0c. Camera. X 130.
Fig. 72.— Continuation of series 63. Taken through the funnel opening into the lateral cavity, 'i <>b. X 8 Oc.

Camera. X 130.
Fig. 73. — Continuation of series 63. Taken through the loop in the nephridium. ii Ob. X80c. Camera. X130.
Fig. 74. — Continuation of series 63. Taken through the oral side of the nerve ring. § Ob. X8 Oc. Camera.

X130.
142

MEMOIRS OF THE NATIONAL ACADEMY OF SCIENCES, VOL X

PLATE XIV

ill. in. I;

,,/:,:

ll .11 I.

Fig. 71.

V Fig. 70.

§~~ "/

Fig. 72.

latji

lat.il .

Fig. 78.

hit. 1 1

Fig. 74.

lain.

PHORONIS ARCHITECTA

PLATE XA

T

143

PLATE XV.

Fig. 75. — Continuation of series 63, Taken a little posteriorly to that of fig. 7"*-. £ <>!>. X 8 Or. Camera. > L30.

Fig. 7ii. — Continuation of series (i,;. Taken through the region where-the branches of the efferent blood vessel pass

around the oesophagus. \ Ob. ■ 8 0c. Camera. ■ 130.

Yu.. 77. — Continuation of series 63. Taken a little posteriorly to that of li;;. 7t>. £ Ob. X 8 Oc. Camera, x 130.

Fig. 78. — Continuation of series 63. Taken a little posteriorly to that of fig. 77. s Ob. ■ s Or. Camera. ■ 130.

Fig. 79. — Continuation of series 63. Longitudinal muscles begin to appear. £ Ob. < 8 Oc. Camera. X 130.

Fig. 80.— Continuation of series 63. Taken a little posteriorly to that of fig. 79. | Ob. X 8 Oc. Camera. > 130.
144

MEMOIRS OF THE NATIONAL ACADEMY OF SCIENCES, VOL X.

PLATE XV

lnt.li.

Big. 'Jr..

lat.ii

Fig. 76.

Iilt.lt

I, it. II.

Fig. 78.

Fig. 79.

PHORONIS ARCHITECTA

PLATE XVI.

145

PLATE XVI.

Fig. 81. — Continuation of series 63. Taken a little posteriorly to that of fig. 80. s Ob. x 8 Oc. Camera. X 130.
Fig. 82. — Continuation of scries (>:i. Typical transverse section through the middle of the trunk. 1 Ob. X 8 Oc.

Camera. X 130.
Fig. 83. — Longitudinal section through the recipient and distributing vessels. J Ob. X 4 Oc. Camera. X450.
L46

MEMOIRS OF THE NATIONAL ACADEMY Of SCIENCES, VOL X

PLATE XVI

Fie. 81.

1. 1. in.

Fig. 82.

PHORONIS ARCHITECTA

PLATE XVII.

147

PLATE XVII.

Fig. 84. — Longitudinal section through the anal region. Showing the ganglion and its relation to the lateral nerve

cord. I Ob. X 4 Oc. Camera. X 450.
Fig. 8o. — Transverse section through the region of the anal papilla. Showing the relation of the nerve ring to the

lateral nerve. }. -Ob. X 4 Oc. Camera. -'450.

lis

o

MEMOIRS OF THE NATIONAL ACADEMY OF SCIENCES, VOL. X.

PLATE XVII

Fig:. 84

Fig:. 85.

PHORONIS ARCHITECTA

3¥

\