Peta ta ne
Sian

io
iy.
“

<<

Death

aie be

Sy on a
aan

ava '
Wile ka

Son

_ into)
Vea

‘

9 ’
ee oui

%
ASD)
pea in

Sarmimeneerces
SEAR NR

vy
978 pity Bia ey tee
RASS IN ae Gee
Aine vaca
HAAR rhe eA
Wav ies

9,
Rane

ey ae
ig ey ae th

aati

ris

eee

.
et, hee

bias

ase

4
‘An tiem he
Ty Atte om no

mata te
a Spas,

Aette
pee
aot d :!
pee

ary
Pat!

er ars 4).
Sac ab

Robey
Roberts:
a

2: eats
poeta peat

CK Ie ag
if GbE WSS dee
ve

vr
a

be

Ret ast

43

‘I t acai oe ne ‘
“o ty sae pA fomeetys sa ers Bae
| : fot bare ees ie ae i ba ma
: : riteirscaue cet c Ruantpieteh op ox eee
! Yr Ce AIRES A ey ite ate ae sores
eaten

FelsGe"

ae tid ape

ss
Pe sSarimns ye
otis

+
bid

=
: rt Pras
i 1 vie

28.

+

weet,
Dara s 1a

SMITHSONIAN
LIBRARIES

Cetoek)
Rem eet
ie gers tty
Cree

if i
vee ites
ees aie

ests
ne

SY = < trig 48 ee = < oy, ‘ys Sh MBO) ‘Tr ‘ yy t io ar =a La, ee? tate Fe ud

wn — w = 22) — w = 120) ’
ION’ NOILALILSNI NVINOSHLINS S3IUVUYGIT LIBRARIES SMITHSONIAN INSTITUTION NOILMLILSNI_ NVINOSHLINS S31 uy

n z 2) < . ” ra 72) z= 7)

= < = HW); st ; = ae ~\ = = s

3 ie > Ys Ne: 2 \ai O: :

i) a2) wn GJ KC MQ _@ n BS NY w iD fy na wn
AS 9 < 9° fii = WAY xc BA IRY 2 WY tj, XZ fe)
Pe, : ia TaN ENS ae 2,
| =? = = = ~S = = AS: = = =

ra a ee 2) y ” #s 4 7) =e
G17 LIBRARIES. SMITHSONIAN INSTITUTION NOILNLILSNI_NVINOSHLINS S31YVUEIT_LIBRARIES SMITHSONIAN _INSTIT

5 “ 5 2 Nan Z gre tee

” wn ra n : pals Yn = w

= w Set a = a = foal =

5 < - < 2 SY = 5 : 5

: : : é 5 AN = : = :

Z = Zz a3 2 a a = Zz
JON NOILMLILSNI NYINOSHLINS ~SAIYVUAIT LIBRARIES SMITHSONIAN INSTITUTION NOILMLILSNI NVINOSHLINS S31¥)

= z is z Lose = = z

S oO ° o 2 ace) 2 = fo)

= oO — E 4 = o LY

Ie > iz oe i = E = =

ra ee) = aes & = = ay =

” =, Ww m 2) m 7p) = wo

ze ie . 2 n cs no Zz 5 . 2
git LIBRARI ES SMITHSONIAN INSTITUTION SE CS ee 1uvud NILLIBRAR I ES SMITHSONIAN INSTI]
pee SE “ ee KX EF 2 ¢
yy 9 MW 5 $e} ARR RMAM AACA A WN oO WSs

aR 2 : IE Re 8 2
EM 2 E : bo we E

>

a . 2 = 5 ees =
TION NOILALILSNI NVINOSHLINS ! ,NOILALILSNI NVINOSHLINS S3IY

wn = (ép) = “ ep)

ax a uw U & RNS Ww

a = a ] rc = WAS a

< S < ef rg = RW az

= = E ; SQ E

© 3 o | tae EX LIBRIS SNS

Tas » =) if) = aN =5
aime EB RAR IES SMITHSONIAN _| nM LIBRARIES. SMITHSONIAN SINS

= Zz “V7° Nl 5 c

= o = _ William Healey Dall | ; é hy, >

zy 5 Ane Whi, WAN AN ) S ogy, ®

> = i > on. > = Wi Gi>

2 ae a Division of Mollusks P ee Miia =

o z ys Sectional Library Q Z m
TION NOILALILSNI NVINOSHLIWS § 9 NOILNLILSNI NVINOSHLINS S31

wn Fa on > z wn

=e ds! = = : &, = = 4
\ 2 Fy = 1 Wiy & 2 9%
A Jel m2) % p JEU a iw lp
ER ae g Sa fle 3 g
ote = By ; = >

2 ”) z 4 7) Zz

{/€@11 LIBRARIES SMITHSONIAN LIBRARIES SMITHSONIAN INSTI

= w = ve a uw —s w —

a ws - us & & uw & us a

: = = : ES 5 E =

< SMG

c s e = S ENS nS S e S

= mM.” me faa) _ SS SS oO _ mM. —

= = eS Sie es Beir 3 = =

and

TION NOILALILSNI NWINOSHLINS S3IYVYEII LIBRARIES SMITHSONIAN INSTITUTION NOILNLILSNI NVINOSHLINS S314

z me z = 6 = S : 5

= W SS — o oon aw — wo —

5 aN Wee = 5 Op 5 2 E

Kb > SMa = > = > = > =

= Pe) SS, = >a) = * a = > 8] =

7) ra : : i = 2) ni 7) = i

Z m Oe ys z o z a _ 2
Ya T_LIBRARIES | SMITHSONIAN TUTION | Se ar NBS EINE LBS R SIMA cI INSTI
at z 2 az :
<x aN = xt z x = << eae = x N
Te WN N\ _ — > = iS SN = = §

y z a \ 4 iS = ‘S

ip ro WS z =e.” y 5 i 4 l by ro) SS 5 5 Wy
, 2X & i PM ° bid 2 Re 3 2
Dp ENS 2 9 Yjy = 2% = SX 2 =a

es = > = ee | = Re s

7) z z ” z 2) o: Zz ”

JTION NOILALILSNI NWINOSHLINS Saly¥vuaIT LIBRARIES SMITHSONIAN INSTITUTION NOILNLILSNI NVINOSHLINS S318

117 LIBRARIES SMITHSONI

wo — == w =a ” — w
‘ Zz

= & RNG & a +0 ea B ES

ox =u WAS = o a « =I “

< 2 a < z < 2 <
YS 6 = = = = S S s

—! z * Zz 25 Zz - Zz Saeed
SPUR ENE RAR TES SMITHSONIAN INSTITUTION | NOILALILSNI NVINOSHLIWS Sil uvuall LIBRARIES SMITHSONIANZINST

>

Sa a sono ote) Beat Oe, ey ae Ot PastON We ke sans OS ee ae |

‘ "4 ‘“ A Cel =," ee aa he a Ww Sf “ sey \% ef
ee oO See = eR ae = UR on SB WY o “ty
Wast'’ m TAN WASH’ m iNOS. m JNOSHY Wass AN > Iasis®
; o ase 2 wo = o = oA = _

RARIES SMITHSONIAN INSTITUTION NOILNLILSNI NVINOSHLINS SSIYVYEIT LIBRARIES SMITHSONIAN INSTITUTIOS

m
wn
wn ; z i ”
z q z < z em ie = <
= — . — — z . NO
z s Nai 92: ; s Nae 5 \w
de AQ: : ro) I SN OO WIG
NEL SNE eg ON ON
aE a Mey 5 ey date
ALILSNI_NVINOSHLINS, S31YVUGIT LIBRARIES SMITHSONIAN INSTITUTION NOILNLILSNI_NVINOSHLINS "S31NV¥ a
= ¥ = wn =. ~~
a w o w 5 w 5 wl
= a = = =t oa = [4
au os =f < oS < 4 VSS <
= cc iS fre Cc oc = SAS ac
= [eo] = mM. pal am = ss N (oa)
fe) AX = fo} = fo) = fo) SS =
za ty 4 = ra pay 4 all
ZARIES SMITHSONIAN INSTITUTION NOILMLILSNI NVINOSHLINS S31NVuaIT LIBRARIES SMITHSONIAN INSTITUTIO
5 = 6 = ge SG = 6 :
— 1s ° ea oe ) ° a
S 2 = = S EF a = 2
Es > = = = > Ey =
=u 4 = Z = Bs Be 2
Se rareituse 2 m Re 2 ne OE
MLILSNI_ NVINOSHLINS SA1YVUAIT LIBRARIES SMITHSONIAN NOILNLILSNI_NVINOSHLINS, Sa1uyu a
& z = ae g = =
ra =) se =] =| Rs z =|
re) Sel On Se rr L% re) a1)
2 B 2 g é x 2 2
“ 2 = 2 2 Wy = 2
= > = > > : = >
(79) Fo! ” os ma nw Fad

RARIES SMITHSONIAN INSTITUTION NOILNLILSNI LIBRARIES SMITHSONIAN _INSTITUTIO!

LIBRARIES

NOILNLILSNI
NOILNLILSNI

=~
RE
NOILNLILSNI

cil *
NLILSNI NVINOSHLINS S3IYvYdIT LIBRARIES SMITHSONIAN_INSTITUTION NOILNLILSNI NVINOSHLINS S3IYVUSI

Set uvud 17 LIBRARIES SMITHSONIAN

SSIYVYEa!II_LIBRARIES
INSTITUTION NOILNLILSNI

INSTITUTION
S31uvudl)
INSTITUTION
ey
INSTITUTION

RARITIES SMITHSONIAN INSTITUTION NOILNLILSNI NVINOSHLINS S3SIYVYEIT_LIBRARIES SMITHSONIAN

x

INSTITUTIO!

NLILSNI_ NVINOSHLIWS

NVINOSHLINS S3I1YVYEITILIBRARIES
Ip

NVINOSHLIWS
NVINOSHLIANS
NVINOSHLIWS
Sy
SMITHSONIAN
SMITHSONIAN
fg

LIBRARIES SMITHSONIAN INSTITUTION NOILNLILSNI_NWINOSHLINS S31YWvUd!

ARIES SMITHSONIAN

NOILNLILSNI

NOILNLILSNI
LIBRARIES

RARIES SMITHSONIAN INSTITUTION NOILNLILSNI NVINOSHLIWS SMITHSONIAN _INSTITUTIO}

a

SMITHSONIAN INSTITUTION

Saiuyvuai7_LIBRARIES

S3alY¥vu¥di1l LIBRARIES

INSTITUTION NOILNLILSN!

INSTITUTION

INSTITUTION NOILNLILSNI

INSTITUTION

LNLILSNI NVINOSHLINS SJIYVYUGIT LIBRARIES NOILNLILSNI NWINOSHIINS S31YVua!

\.
SMITHSONIAN INSTITUTION NOILNLILSNI NVINOSHLIWS

gs

SN
\

+

\N

SS
SS

NVINOSHLINS S3IYVYd!IT_ LIBRARIES SMITHSONIAN

NVINOSHLINS S3I14YVNY9I7 LIBR
NVINOSHLINS Saluvugt

Sh.

NS R

SYN
SMITHSONIAN

SMITHSONIAN
NVINOSHLIWS

RARIES

S SMITHSONIAN

saluvygi LIBRARIES SMITHSONIAN INSTITUTIO!

Pép
gy

NVINOSHLIWS

NOILNLILSNI

NLILSNI

717
BSTITUD.

Salyvudii

STir,
Sl oN

LIBRARIES
ALL

INSTITUTION _ NOILNLILSNI

SMITHSONIAN
¥ Aan ws.

ey.

NVINOSHLINS S3IIlYVy al”

<aSTITty

J1/1 LIBRARIES SMITHSONIAN

Gi1_LIBRARIE

ION NOILNLILSNI
J17 LIBRARIES
JON NOILALILSNI
ION NOILNLILSN
Bili See ES

ION

ae Awe

r ey bi of y a1 ve i ae
“y a ie 7 Tay ay . i i ! j C A
Roy ri f :
r aa he
; eh

, ; : f Naty , ; - é
eae ek » on it ee a
mn a) rr 7 : had

Rg Se.
ts ore

H A if el es j
Ap CAE
Uae Ba

| eo

Whe it | ne |
0 i ee

i
j

i
Wy

j
ht Ve
Pak

BOSTON SOCIETY OF NATURAL HISTORY;

~

VOLUME IV, NUMBER VIIL.

PHYLOGENY OF THE PELECYPODA,

eUW

Zi

THE AVICULIDH AND THEIR ALLIES.

PLATES XXIII—XXxX.

By ROBERT TRACY JACKSON, S. D.

sa?

BOSTON:
PUBLISHED BY THE SOCIETY.
Juty, 1890.

WILLIAM H. DALL
(teen LIBRARY
DIVISION OF, MOLLUSKS

VIL. PHyLoGceny or THE PELECYPODA.

Tue AVICULIDH AND THEIR ALLIES.

Division of Mollusks

Sectional L
By Ropert Tracy Jackson, S.D.

CONTENTS.
I. Introduction. p. 277 IX. Pecten: anatomy and habits. p. 333
II. Sources of material and riettode of X. Pecten: the shell : ; ‘ p. 342
work : ; p. 284 XI. Genera allied to Pecten . : p. 350
III. Classification of beeen of biomth: and XII. Anomia: anatomy and habits . p. 354
decline . p. 288 XIII. Anomia: theshell . . . ~~ p. 359
IV. Ostrea: development of the soft par urts ; XIV. Studies of a few other genera . pp. 362
embryology. . p- 294 XV. On the genetic relations of the Pro-
V. Ostrea : development of the soft pare dissoconch . : 4 p. 375
the spat . : p. 303 XVI. Genealogical connection of the
VI. Ostrea: development of the shell p. 311 Aviculidz and their allies 4 p: 379
VII. Gryphza, Exogyra and adult Ostrea, XVII. List of papers quoted or referred
with a consideration of the mechani- to in the text : : : p. 394
cal origin of the ostrean form of shell p. 317 XVIII. Description of plates. : : p. 395
VIII. Perna, Avicula and near allies. p. 327

I. InvTRopDuUCcTION.

My paleontological studies have been carried on at the Museum of Comparative Zo-
ology as a pupil and assistant, first of Professor Nathaniel S. Shaler and later of Pro-
fessor Alpheus Hyatt. To these gentlemen first are due my acknowledgments in this
paper, which was offered as a thesis for the degree of Doctor of Science at Harvard Uni-
versity May 1, 1889.’ Acknowledgments are next due to Mr. Alexander Agassiz for
the opportunity to study the collections in the Museum.

The investigations on which this paper is based were made while associated with Pro-
fessor Hyatt in the Paleontological Department of the Museum.? I have applied some
of the principles which he has deduced as a result of his scientific work, but I wish to
state that he has never urged his views upon me, or pointed out their application in the
Pelecypods. At Professor Hyatt’s special request, I have used entire freedom in reject-
ing or criticising his views whenever they seemed not in accordance with the facts pre-
peated

‘On account of new material obtained, some addi- have been made to sections vim and xiv, and are indi-
tions have been made to this paper since it was read be- cated by the date of the collecting of the material.
fore the Society on April 3, and since it was offered as *Three years have been spent in that department as-
a thesis on May 1, 1889. The additions for the most part sisting Professor Hyatt arranging the great collections of

Fossil Invertebrates.

_ MEMOIRS BOSTON SOC. NAT. HIST., VOL. IV. 38 (277)

278 ROBERT TRACY JACXSON ON THE

For the freest opportunity to study collections under their charge and for frequent as-
sistance while studying those collections my obligations are due to Dr. C. A. White and
Dr. William H. Dall of Washington; Professor Angelo Heilprin of Philadelphia; Pro-
fessor R. P. Whitfield of New York; Dr. J. Walter Fewkes and Mr. Samuel Henshaw.
For material borrowed for study, I am indebted to Dr. White, Dr. Dall, the Museum of
Comparative Zodlogy, the State Museum of New York and the Boston Society of Nat-
ural History. I have to thank Colonel Macdonald, United States Commissioner of
Fisheries, for the electrotypes of figures of Ostrea edulis and Professor A. E. Verrill
for electrotypes of Vermetus, Pecten and Anomia.

Tn order to unravel with any reasonable hope of success the relationships of the mul-
titude of fossil forms of animal life in which only the hard parts are commonly preserved,
and to demonstrate the connection between the living and fossil representatives of allied
groups, it is most important to study with minute care the hard parts in living genera.
It is necessary to know as far as possible the structure, development and functional ex-
pression of the hard parts, both by themselves and in their relation to the soft parts. To
attain this end, studies of any given animal should be made on all stages of growth from
the embryonic upward, and in so far as the young is considered as a living expression
of the adults of ancestral groups, this period of the life-history must be studied with special
care.

On the value of anatomy and embryology to the paleontologist studying the affinities
of groups I would quote the following passage from Professor Huxley’s Anatomy of In-
vertebrated Animals, p. 587. “ Already indications are not wanting that the vast multi-
tude of fossil Arthropods, Mollusks, Echinoderms and Zodphytes, now known, will yield
satisfactory evidence of the filiation of successive forms, when the investigations of palze-
ontologists are not merely actuated by the desire to discover time-marks and to multiply
species, but are guided by that perception of the importance of morphological facts which
can only be conferred by a large and thorough acquaintance with anatomy and embry-
ology. But, under this aspect, the paleontology of the Invertebrata has yet to be cre-
ated.”

The results of my investigations give abundant proofs that a study of the soft parts
and especially studies of life habits and environing influences will throw light on the
meaning of the form of hard parts. They all lead toward the conclusion that external
hard parts are built conforming to the needs of the animal and are modified by cireum-
stances of environment. Many facts discussed in this paper are directly opposed to the
assumption that hard parts are of prescribed form, to which the animal adapts itself, as I
have found most active response in the form of hard parts to changed or changing con-
ditions of the anatomy or habits of the animal. The changes in form thus introduced
might well be called cases of “mechanical genesis,” adopting Professor Ryder’s apt phrase
used of the development of toothed forms in Vertebrates.’ When the meaning of feat-
ures of shell form in living animals is ascertained, it is recognized that they make im-
portant factors in the consideration of fossils and the serial connection of groups.

To unite living and fossil forms of a group of Pelecypoda in a common study is the

1 Facts bearing on this point are especially considered andin the discussion of Perna (section vitt), Pecten (sec-
in the origin of the ostrean form of shell (Section vil), tions rx and x), and Anomia (section xt1).

PHYLOGENY OF THE PELECYPODA. . 279

aim of this paper. It is hoped by this means to make connections and differences exist-
ing between some genera better understood, and thus to further the knowledge of the
relationships of those genera. So much time has been given to the study of living forms
that fossils do not figure as largely as they should in a paper of this nature. It is
hoped, however, that in future this want may be supplied by investigations in that direc-
tion.

The main line of my research is the study of shell structure and form. Studies of
anatomy and habits are subservient and correlative to those of the shell. Facts observed
that have no bearing on shell structure are however frequently given as they may yet be
of interest or of serial value. In studying the shell I have sought especially for young
shells hoping in them to find characters of interest which would be valuable in attempt-
ing to trace the genetic connection of groups. It is to this study of the young, that
Professor Hyatt has especially devoted himself in his researches among the Cephalo-
pods, and I owe a debt of gratitude to him for the suggestion to try the same method of
research in Pelecypods, as well as for the encouragement and sympathy he has given
me in my work.

The young shell of Pelecypods has been very little studied either by itself or in its re-
lation to later stages, and I know of no studies embracing observations on the develop-
ment of the shell in related genera, or any attempt to trace the connections between the
changes of shell form and anatomy during the development of the individual. In Zit-
tel’s Paleontology (Barrois’ translation) we find in the consideration of the Pelecypoda
the following statement: “Quant aux modifications des coquilles 4 leurs différentes ages,
elles Wont été que bien peu étudiées jusqw ici, et ces observations n’ont pas encore mises
a profit dans la systématique.” The researches of this paper may therefore be consid-
ered as opening up a new field of investigations, as I consider not only the changes in
form and structure observed at different periods of development in several genera, but
also trace in many cases the serial connection between those changes in form and the
phylogenetic history of the group.

Pelecypods are commonly divided into three groups accordingly as the adults possess a
single adductor muscle, Monomyarians; two equally developed adductor muscles, Dimya-
rians; or two adductor muscles, one large and functionally the most active, the other
smaller, Heteromyarians. This classification, I believe, is misleading, as by it are grouped
together genera having no close relation and similarly genera closely akin are sepa-
rated. I show that the size and number of the adductors are closely connected with the
relative position of the axes of the body to the hinge axis (section v), and refer to the
fact that the single adductor condition is adopted by several genera not related and
which it would be entirely disregarding the sense of systematic classification to group to-
gether. On this basis of adductor muscle classification even different species of one ge-
nus would in one case have to be placed in the Hetcromyarians and in another in the Mon-
omyarians (see Mytilus, section v, note). The words monomyarian, etc., are occasionally
used in the text, but as general descriptive terms, not with the classificatory value ordi-
narily applied to them.

In the text frequent reference is made to the law of concentration and acceleration of

280 ROBERT TRACY JACKSON ON THE

development. This law was enunciated separately and almost simultaneously by Pro-
fessors Hyatt (29)' and Cope. The law is, that characteristics acquired in the adult, if
inherited at all, are inherited at earlier and earlier periods in successive generations, until
they exist only in embryonic stages, or are completely lost through disuse and concentra-
tion and acceleration of development. This law finds strong support in the facts pre-
sented in the development of the oyster and other Pelecypods as well, where many pe-
culiarities occur which are well explained by it. A corollary of the law of concentration
and acceleration of development is that, in the young stages are found the equivalents of
which are to be sought in the adults of related groups.

T have met with difficulty in procuring young specimens of Pelecypods to study, for the
most part depending on what I collected in the field or found adhering to masses of shells
in the museums. Conchologists commonly pay little attention to the young in their col-
lections as they are not considered characteristic. They certainly are not characteristic,
in that they have not a full expression of features found in the adult, but they are char-
acteristic of their period in the development of the individual and from this point of
view they are well worth collecting and studying carefully.

Tn section ut Professor Hyatt’s classification of stages of growth and decline is con-
sidered in its application to mollusca. I have there qualified some of his definitions
and introduced a new term, phylembryo, which is applied to the stage in development
where features exist which first render the embryo referable to the class or phylum to
which it properly belongs. In considering stages of growth it is a valuable assistance
to have an established nomenclature by which any given period may be briefly and defi-
nitely expressed, and that given in this paper seems to fill the requirements of such a
nomenclature.

The phylembryonic stage of developing Pelecypoda presents some striking features
which have not been previously considered as far as known. The shell of this period is
characterized by having a straight hinge line as shown in many genera(fig. 25, p. 300, Pl.
xxii, figs. 7-11 and PI. xxrv, figs. 13-15), although later succeeded by a curved one.
An important point is the fact that at this age but one adductor muscle exists, and that
one is destined to become the anterior adductor of adults which retain it(dimyarians).
The posterior adductor is developed later. The mouth and anus in early embryonic
stages of Pelecypods are closely related on the ventral border of the embryo (fig. 23,
p- 299) and the anterior adductor muscle develops early on the dorsal side of the ali-
mentary canal (fig. 25, p. 300; Pl. xx, figs. 9-11) which position it retains through-
out life, in dimyarians. The posterior adductor is not developed until the anus has re-
volved dorsally from its early ventral position, and then the adductor is formed on the
ventral side of the alimentary canal, which position it retains throughout life in all
Pelecypods. We see, therefore, in this an explanation of the relation of the adductor
muscles to the alimentary canal, one adductor overlying, the other underlying it, Pl.
xxv, figs. 9-12 (for full discussion see section v).

1A list of publications quoted or otherwise referred to is as in this case corresponds with a number on the refer-
given in section xvir. Where the reference in the text is ence list.
not perfectly evident, 2 number in brackets is used which

PHYLOGENY OF THE PELECYPODA. 281

One of the most important discoveries described in this paper is the existence in
young Pelecypods of an embryonic shell, or prodissoconch comparable to the embryonic
shell or protoconch of cephalous molluscs. The shell of Pelecypods develops from a
preconchylian gland, as does the shell of cephalous molluscs; therefore, the two valves
of a Pelecypod are together homologous with the single valve of univalves as was shown
by Brooks (7). The adult shell of cephalous molluscs has been named the conch, or
true shell, and the first formed embryonic shell on account of its differences the pro-
toconch; as the adult double shell of Pelecypods is the homologue of the adult single
shell of univalves I have named it the dissoconch (double shell) ; similarly the early em-
bryonic double shell on account of its differences I named the prodissoconch (early
double shell). A prodissoconch is described in this paper in twenty-four genera of Pe-
lecypods,! figures of most of which are given in the plates. It has been found in many
species of some genera. These genera are widely separated so that it may be considered
as proven that a prodissoconch is a characteristic feature of developing Pelecypod shells
(discussion and descriptions, sections vi to xv inclusive).

In the discussion of the prodissoconch it will be seen that stress is laid on the direc-
tion of the umbos, as to whether they are inclined anteriorly or posteriorly relatively to
the anatomy of the animal at the close of the prodissoconch period. I find that they are
inclined posteriorly in Avicula, Perna, Ostrea, Pecten, Spondylus, Plicatula and Anomia.
On the contrary they are inclined anteriorly in the other genera described except the
Unionide. The value of this feature may be overrated, but I think it is of importance
in considering the connection of genera and the derivation of the prodissoconch (sec-
tion Xv).

In considering what value is to be attached to inclination of the umbos, attention
should be called to two singular, ancient paleeozoic genera, Antipleura and Dualina,
described by M. Barrande. In single species of these genera series of specimens may
be picked out in which the umbos are inclined in one direction, and other series may
be picked out in which the umbos are inclined in the opposite direction; the two series
being apparently identical in every other character. This is a puzzling case, seemingly
opposed to all natural laws. M. Barrande figures large numbers of specimens illustrat-
ing the two series, and I have studied them in the Schary collection of fossils from
Bohemia, now in the Museum of Comparative Zoélogy). Ordinarily, however, the in-
clination of the umbos is a constant character and therefore I think it is not a mistake to
consider it a feature of serial value as [ have done in the text.

The prodissoconch of Ostrea, Perna, Pecten, Anomia and allies, is composed of
homogeneous sub-nacreous lime showing no prismatic clelular layer, fig. 27, section v1,
the first stages of the dissoconch growth on the other hand in these gencra possesses a
prismatic cellular structure making a marked contrast in histological structure with the

1Previous to the publication of my preliminary pap2r, e2eding dissoconch shall in Pelecypods. In this paper a
an early embryonic shell and its relations to the succeed- prodissoconch is described in the following genera: Ostrea,
ing shell had been described at some length by Ryder in several species; Gryphra; Exogyra; Perna; Avicula; Pec-
the oyster, briefly by Morse in Anomia, and briefly by ten, several species ; Spondylus; Plicatula; Lima; Anomia,
Lacaze-Duthiers in Mytilus. With these exceptions I be- two species; Mytilus; Modiola, Crenella; Argina; Unio;
lieve my descriptions are the first concerning the early Margaritana; Anodon; Sphirerium; Petricola; Scrobicu-

embryonic, prodissoconch shall and its relation to the suc- laria; Venus; Tottenia; Mya and Saxicaya.

*)

282 ROBERT TRACY JACKSON ON THE

condition in the prodissoconch. In Mya, Modiola, Venus and most other genera studied
(section xtv) no prismatic layer was observed in the first stages of dissoconch growth
and this difference, with others considered, is deemed sufficient to separate widely these
two divisions of genera.

A discovery of much importance is the existence of two adductor muscles in the pro-
dissoconch stage of Ostrea, Pl. xxiv, figs. 1-2. From the form of the shell and from
analogy, a similar condition is predicated in Pecten, Avicula, Perna, Anomia, ete. All
these genera possess but one adductor muscle in the adult and it has not been demon-
strated before (excepting in Ostrea in my preliminary paper) that two exist in the young.
Professor Huxley in 1883 showed that the first formed muscle of the oyster (fig. 25,
p- 300, this paper) could not be that of the adult and that a new, second muscle to be re-
tained throughout life must develop. On this basis and knowledge of early types of
Pelecypods, he argued that Ostrea must have developed from a dimyarian ancestry and
my discovery of the early two-muscled stage, therefore proves the truth of his assumption
(discussion, sections rv and xv, also briefly in other sections on Perna, Pecten, etc.).

There is commonly little consideration given to the true relations of the axes of the
body to the shell in describing Pelecypods. In dimyarians the hinge is dorsal and in
monomyarians the hinge area is also considered as dorsal almost universally. Professors
Hyatt (28), Brooks (9), Ryder (60) and Lacaze-Duthiers (39), have pointed out the true
condition of affairs in Ostrea and Anomia, where on account of the changed position of
the axes of the body the hinge line is anterior, not dorsal. I have endeavored to main-
tain this more careful consideration of the true relation of parts in my paper and this
must be borne in mind in reading the text. For instance, in Pecten the hinge line being
anterior, the free ends of the valyes are posterior and the ears are truly ventral and
dorsal, not anterior and posterior as usually described. This changed position of the
axes is intimately connected with the relative size of the adductor muscles of the valves;
a change in position coinciding serially with the decrease and final loss of the anterior
adductor, together with the increase and final exclusive retention of the posterior ad-
ductor, as described in section v.

The Ostread is a group having a peculiarly modified highly irregular shell which has
been a perplexing problem to naturalists in seeking the affinities of the group. As the
result of my studies I have come to the conclusion that the ostrean form of the shell is
due to the mechanical conditions of direct cemented fixation. The proof of the conclu-
sion is that all Pelecypods which are directly attached by the cementation of one valve
assume a similar form of growth. The form claimed as the result of cementation of one
valve is a concave, highly modified, attached valve and a less modified, commonly much
flatter free valve, often differing so widely from the attached valve, as not to be recog-
nizable as belonging to the same species unless found in place. Cementation also appar-
ently induces a camerated structure of the shell. (See Pl. xxv, fig. 8 and Pl. xxv1,
figs. 1-15, as proof of the effects of cemented fixation.) Zhis statement contains im-
portant conclusions, which are discussed in the text (sections vm and xT) but they may
be briefly stated here. The normal Pelecypod is bilaterally equivalvular, but when at-
tached this feature is lost, asymmetry replacing the typical condition. The cemented
valve is highly modified, the free valve less markedly so; therefore it is the free valve

PHYLOGENY OF THE PELECYPODA. 283

which retains more of the ancestral features, and to it and the free young, before mod-
ifying cementation takes place, we must turn in studying the relationships of attached
Pelecypods. It is an animal with one side modified, losing ancestral features; the other
side, less modified, therefore retaining ancestral features. The actual change from an
unmodified free form to a modified cemented form may be studied in Ostrea (Pls. xxtv,
xxv), Hinnites and Mulleria (Pl. xxv1, figs. 3-7) and Spondylus (Pl. xxvut, fig. 4).
The irregularity of attached shells being due to the condition of fixation we may study
their relationships with an entirely new light on the subject, as the irregularity is a char-
acter superimposed on any regular form of Peleeypod by special given conditions. The
aberrant forms of the Chamidez and their greatly specialized allies the Rudistz: (PI. xxv1,
figs. 8-12 and 20) are considered as extremes of modification in the ostrean line of va-
riation. What the mechanical causes of this line of variations probably are, and what
they lead to as their highest expression is considered in section VI.

Finding that irregularity commences with the introduction of the fixed period we have
a powerful argument in explaining the absence of features in the young dissoconch
stages of Ostrea which might, if free, have retained the form of related ancestral groups,
as is the case in the young of Perna, Pecten, Spondylus, etc. Ostrea is cemented at the
close of prodissoconch period so that stages of shell growth later than that are all mod-
ified. The free valve should retain more ancestral features than the attached valve and
this is the case with the oyster. ‘The prodissoconch, the free valve and anatomy of the
adult, all unite in characters which refer the Ostreade to a type, which is Perna, or some
close ally of that genus (sections V-VI1).

T have described the development of the gills of the oyster with some care, as it is but
slightly treated of by other investigators of this mollusc. In order to understand the
serial development of the several lamellz and to show the relations of the divect and re-
flected borders of the filaments, diagrams are given (PI. -xxtv, figs. 6-12) of the de-
velopment of the gills of Ostrea, together with diagrams of adult gills in some related
genera. .

In Perna (PI. xxvu, figs. 1-3) I find a prodissoconch very closely like that of Ostrea;
but the initial stages of dissoconch growth show indications of a byssal notch not seen
at allin the prodissoconch. This character of a byssal notch originating in the initial
stages of dissoconch growth, but not seen in the prodissoconch, is described also in Avic-
ula, Pecten and Spondylus (Pls. xxviI-xxviil, sections VIII-Xr).

In Pecten, interesting facts are presented in regard to the life habits of crawling, bys-
sal attachment and mode of swimming, new points are discussed and the shell is described
at length in the young of five species. In all these young I found a prodissoconch with
important characters. The first stages of dissoconch growth in four species were found
to possess well defined prismatic structure in the right valve whereas this is not known
in the adults of these species, and in the extensive researches of Professor Carpenter on
shell structure, prismatic tissue was found in only one species of Pecten. The Pectens
are allied to a group of Pelecypoda which has prismatic structure, in the shell, and it is
a highly interesting fact in accord with other conclusions, that in the extreme young
this feature exists, though early lost and not seen at all in adults. In the young disso-
conch stages of Pecten (PI. xxvuir) in many species I find characters in the form of the

284. ROBERT TRACY JACKSON ON THE

shell which are referable in the closest manner to ancestral groups in early geologic times,
from which the Pectens were doubtless evolved. The life-history of Pecten shows most
strikingly how the phylogenetic history of a group may be traced by the series of stages
presented in the early development of the individual (sections Ix—x).

Studies were made of young Anomias (PI. xxtx) growing on glass slides and yielded
facts of high interest, especially in connection with Professor Lacaze-Duthier’s observa-
tions on the adult. In the young the three muscles, adductor, posterior pedal and byssal,
are widely separated, though in the adult they almost come in contact, rendering their
nature ambiguous. The byssus of Anomia was seen to originate from a eleft in the foot
and in the young consists of separate threads. The calcareous plug is formed by the
aggregation into parallel lines of separate dise-like centres of calcification, each centre
being comparable to a separate byssal-thread attachment found in ordinary Peleeypoda
which have a byssus. These and other points of habits, anatomy and shell structure
are discussed in sections xi and XuIt.

The section x1v, entitled Studies of a few other Genera, considers the young of Mytilus,
Argina, Venus, Mya, etc., and is intended principally to show the extension of the pro-
dissoconch to other genera besides those with which this paper is more especially con-
cerned. Brief accounts are given of the young of several genera and a comparatively
extended deseription of the development of the siphon and byssal attachment in Mya.
The considerations of this section are for the most part of a preliminary nature and it
is hoped to give a full description of these and other genera in a future publication.

After studying the prodissoconch in Ostrea and its allies, Pecten, Perna, ete., I have
summed up m section xv the evidence of anatomy, shell strueture and paleontological
occurrence, all of which points towards Nucula or a nuculoid form as the probable an-
cestral typeof which the prodissoconch is the representative in the development of these
several genera. Passing to the next section (xvr),I have attempted to give in tabulated
form the serial relations existing between the genera studied. These two sections may
be considered as the culmination, to which the studies of shell growth at different ages
of the individual and in related genera have gradually led.

As amatter of technique I would call attention to the benefit received from the use of
drain-pipe traps containing glass, as described in section un. By its means attached
forms were studied on a transparent medium and free forms of small size were entrapped
and easily found.

For the guidance of readers attention is called to the italicized paragraphs on pages 295
and 300. Every care has been used to give full references to all observations borrowed
from authors and equal pains have been taken to give full credit to investigators where
their observations seemed similar to mime, or to have bearings on the various facts and
conclusions discussed. The facts described and inferences from them not specially ered-
ited in this paper are my own.

II. Sources or MATERIAL AND Mrtuops or Work.

In the summers of 1887 and 1888, several weeks were spent at a fine collecting local-
ity on Buzzards Bay, which is situated on the southern shore of Cape Cod, from which
I walked, or rowed, to my laboratory not far off.

PHYLOGENY OF THE PELECYPODA. 285

The water of the bay is remarkably clear and pure. It is warmed by the Gulf Stream,
and also by rising tides which flow over great areas of sun-heated flats. The water is
freshened by Monument River, and by brooks and percolating water entering the bay
from the low, gravelly shores.

The site chosen for operations is alow-lying neck of land in the township of East Ware-
ham. At this locality a sand bar is formed by the tide as it rises and flows on the one side
to Onset and on the other to Buttermilk Bay. The force of the current is very great
and it is a noted spot amongst the fishermen for a “set” or “fall” of oyster spat. On the
north side of the neck of land a small sinuosity in the shore line forms an almost en-
closed pond which is bare at low tide. The greater part of the young Pelecypoda stud-
ied, were obtained on the bar and in this pond. For the use of a temporary laboratory,
the bar and the pond, I am indebted to Dr. E. B. Larchar. Iam also indebted to him
for occasional assistance in prosecuting my work.

In studying forms of animal life that become fixed to foreign bodies during their early
stages of development, it is an advantage if they can be induced to attach themselves to
transparent media, such as glass or mica, so that they may be studied by transmitted
light without disturbance under the microscope.

Several attempts have been made by Ryder (62), Horst and Winslow, but with only very
partial success, to secure oysters on glass. As I was anxious, if possible, to get oysters
growing on glass, in 1887 I tried several methods to accomplish the desired end, which
were described in my paper on “Catching Fixed Forms of Animal Life on Transparent
Media for Study.” The following method was tried and found successful. On the 15th
of July, 1887, at the suggestion of Dr. Larchar, earthenware drain-pipes were partially
sunk on the sand bar in an upright position, and loosely filled with broken glass. Six-
inch and four-inch pipes were used, and at different elevations above the sand. I did
not secure any spat in the four-inch pipes. A six-inch pipe sunk nearly to the level of
the bar was successful, but in an unlucky day was filled with sand, and the young oys-
ters buried. Six-inch }ipes, reaching about ten inches above the bar, were perfectly suc-
cessful, and on the glass in them I-got large numbers of young oysters. During the
summer of 1888 the attempt to get oysters on glass was repeated. On the 18th of July,
four six-inch drain pipes were sunk perpendicularly on the bar at the lowest point avail-
able at low tide. They stood about fifteen inches above the sand. The last of July,
four eight-inch pipes were sunk in a similar position, all were filled loosely with broken
glass, and met with entire success. The eight-inch pipes were preferred to other sizes
as large pieces of glass could be put in them and the spaces between contiguous plates
could be more easily maintained. Long strips of glass, two to four inches wide, the
trimmings of glaziers, were used. They were arranged perpendicularly in the pipes, and
were readily taken out, examined and cleaned. They were cut as required by one of
the rotary, hard steel wheels sold by hardware dealers for that purpose. This size of
pipe and glass gave such satisfaction that I should use it in future if occasion demanded
to the exclusion of other methods that were tried. Suitable conditions for the success
of the undertaking were therefore as follows: (1) a strong tide-way, which would bring
plenty of free-swimming fry, and afterwards abundant food for their maintenance; (2)

MEMOIRS BOSTON SOC. NAT. HIST., VOL IV. 39

I86 ROBERT TRACY JACKSON ON THE

an area of quiet water within the pipe, in which the fry could settle and attach them-
selves to the smooth glass; (3) the absolute shutting-out of violent currents which would
detach them from their precarious abiding places.

I first found spat on the glass on the 25th of July, 1887, and on the 28th took out a
pane 6 by 41 inches square on which I have counted eighty-two young oysters, as it is
still in my possession intact. This was only exceptional in the size of the glass, as other
pieces bore nearly or quite as many spat proportionally to their area. Spat were secured
in the pipes until the 29th of October, 1887, when few were left, large numbers having
dropped off by the natural dissolving action of the sea-water upon the organic cement
by which they are attached. In 1888 the first oysters were found on glass on August
1st, and they continued to set into the month of September. The lateness was probably
due to the comparative coldness of the summer.

Besides oysters, the young of many invertebrates and alge attached themselves to
the glass, affording admirable opportunities for studying them alive and undisturbed.
The following attached invertebrates were noted: Molgula manhattensis, Verrill; Balanus
balanoides, Stimpson; Serpulas, and other tubicolous worms; Crepidula fornicata, Lam. ;
C. plana, Say; Ostrea virginiana, Lister., in profusion; Anomia glabra, Verrill, nearly
as abundant as the oysters; Pecten trradians, Lam., attached by a byssus, both very
young and quite large individuals, varying from 4 to 13 mm. in height; Modvola plica-
tula, Lam., attached by a byssus; Argina peaxata, Gray, attached by a byssus; Mem-
branipora pilosa, Farre, often in large sheets; Hscharella variabilis, Verrill.*

The drain-pipes formed a sort of trap, into which many free-moving, crawling, or swim-
ming molluses found their way. When in the pipe they would have some difficulty in
getting out, and much valuable material was thus obtained. The following Gasteropods
were found entrapped: Scalaria lineata, Say; Urosalpinx cinerea, Stimpson. This gas-
teropod destroys many oysters both young and adults by boring into the shell and
killing the animal. I have not seen holes bored by Urosalpinx in Anomias, though they
are quite as exposed as are oysters to its attacks. Littorina litorea, Menke, was found
entrapped, both adults and young, in great abundance. In moving about on the more or
less dirt-covered glass, this animal makes a trail by the lingual ribbon with which it as-
siduously scrapes the glass for its food. The tracks consist of zigzag series of short,
parallel lines of cleaned glass, arranged with the longer axes in the direction of the trail.
The short lines are almost confluent with the lines of the next succeeding row; being
at first nearly confluent on one side and then on the other, producing a zigzag pattern as
the animal moves along. Jlyanassa obsoleta, Stimpson, was found in great abundance
in the pipes. This species makes a track similar to that of Littorina, but much smaller,
as might be supposed from the size of the animal. The differences are that the individ-
ual, parallel, short lines are less distinctly separable one from another and there is a
much larger angle between the successive zigzag series of lines in the track of Ilya-
nassa, than is the case with the track of Littorina. The last two gasteropods did good

1 Our sea-anemone, Metridium marginatum, attaches it- probably successful methed to adopt in studying the de-
self freely to glass slides when in confinement. The plan velopment of corals where very fine results might be ob-
of placing glass in enclosed traps would be a desirable and tained by its employment.

PHYLOGENY OF THE PELECYPODA. 287

service in eating vegetable growth on the glass, that would have been injurious to the
young oysters and other fixed forms, which I was studying.

The following free Pelecypods were also caught in the pipes as entrapped forms;
many specimens of young Mya arenaria, L., varying from 0.5 mm. to 6 mm. in length,
and very young Venus mercenaria, L., 0.5 mm. to 2mm. in length. Argina pexvata
appears again under this list of entrapped forms, as I found a very young specimen
0.75 mm. long which was free. Pecten irradians also belongs to both the lists of free
and attached forms, a number of unattached individuals varying from 0.75 mm. to 1.5
mm. in height being secured. It is a stage of development preceding the attached bys-
sated stage as described in the text. Occasionally adult Pectens and Levicardium
mortont, Perkins, were entrapped.

A bottle, bearing many oyster spats and Anomias, was washed up on the shore dur-
ing a storm, showing that they may be caught on glass under other conditions than those
afforded by the drain-pipes. Probably the oysters set in the bottle when it was in com-
paratively deep water, where moderate quiet prevailed.

The nature of my researches was such that young oysters with shells as clean and
perfectly preserved as possible were desired. Year-old oysters perfectly cleaned were
found very beautifully preserved in the inside whorls of a dead Sycotopus sheli, where
they were completely protected from all eroding action. T’o imitate these conditions, in
the salt-pond above referred to, a large number of inverted flower-pots, two-and-a-half
and three inches in diameter, were suspended from galvanized wire stretched between
stakes driven firmly into the sand. The pots were raised about six inches from the sand.
This met with entire success, the pots in many instances being almost literally covered
with spat. On the outside of the pots very little, and on the inside no sediment was de-
posited, as the pots hung like suspended bell-jars, so that the oysters were clean and well
preserved. Further, on account of the porosity of the earthenware, the oysters had less
hold than on natural cultch of stones and shells, and were easily removed for study.
The owner of the pond in the winter of 1887 built a dam across the entrance so as to hold
about two feet of water in the pond. Under these changed conditions, inverted pots
were again suspended on the 17th of July, 1888. Oysters set on pots quite as suc-
cessfully as in 1887, when the pots were exposed at low tide, and they grew much
more rapidly. .

In four large inverted pots microscope slide boxes filled with slides were suspended.
The boxes were strengthened by brass nails, as they were simply glued and would have
fallen apart in the water. One hundred slides were thus exposed in an horizontal posi-
tion, but they were not as successful as glass exposed in the drain-pipes. They became
coated with sediment on their upper faces and covered with minute vegetable and animal
growths on their lower faces. The last is ascribed to the fact that the slides were so
close to one another that the scavenger gasteropods, Littorina and Ilyanassa, could not
get at them and destroy the objectionable growths as they did in the loosely-filled drain
pipes.

In 1887, a number of wooden plates, such as are used at picnics, were exposed in the
salt-pond. Holes were punched in the plates which were then strung on galvanized iron

288 ROBERT TRACY JACKSON ON THE
wire and suspended between stakes. 'The plates were placed in pairs with the edges
apposed so that the centres were widely separated and chambered pockets were thus
formed. They were fairly successful. In 1888 when the pond was continually full the
plates worked more successfully, the oysters growing with extreme rapidity. The
‘apid growth was probably due to the fact that the oysters were suspended and there-
fore kept very clean and also to the increased warmth of the water in the shallow pond
as compared with water in the open bay. In the chambers formed by the apposed plates,
similar forms to those found in the drain pipes were attached or entrapped. ‘The value
of obtaining fixed forms like oysters on thin wood is that they are easily removed for
study or may be decalcified and sectioned without removal. Should wooden plates be
desired, it is best to secure those which are cut from the block in a dish-form, not those
which are cut flat and then pressed into a dish-form; because the latter flatten out when
exposed in water, whereas the former retain their curved shape.

Cocaine was found a most valuable reagent in paralyzing Pelecypods which it was
wished to kill distended or to dissect alive. A four per cent solution was used and was
added gradually to a small volume of sea-water containing the animal. By this means
only was I successful in killing young Pectens with the byssal attachment intact. Ad-
ditional material was collected at various localities near Boston and some was loaned me
by museums as noted in the text.

III. CLASSIFICATION OF STAGES OF GROWTH AND DECLINE.

In successive stages of growth of the individual, peculiarities are often found which
are characteristic of definite periods. These periodical characters have been studied,
especially in their later phases, by Professor Hyatt in fossil Cephalopods, and as a con-
venience in considering the several periods, he has (33-4) devised a nomenclature of the
stages of growth and decline,! which he believes is applicable throughout the animal
kingdom.”

We will now consider this classification in its application to the mollusca. The no-
menclature devised for early embryonic stages consists of names very broadly recalling
the grand divisions of the animal kingdom. They are protembryo, mesembryo and met-
embryo. The following names, neoembryo and typembryo, are applied to later stages
of embryonic development.

Professor Hyatt considers that the neoembryo molluse is the trochosphere stage, and
that it has not yet acquired characters peculiar to a mollusc. The trochosphere stage
_ (PL. xxi, figs. 3-4), before the formation of the shell-gland, certainly has not yet ac-
quired characters which would define it as a mollusc, and we would limit the neoembryo
stage to that period of development.’ A comparison of this stage in molluses with Pro-

1 Besides the references cited, this paper was published

3 The ideal trochosphere of Professor Lankester (44) as
in the American Naturalist October, 1888. An abstract

that author says: ‘‘ exactly agrees with the lar-

was published in Science, Vol. x1, No. 260.

2 Professor Hyatt says that his nomenclature is similar
to that adopted by Professor Haeckel; but is also supple-
mentary, and based on morphological, rather than on phys-
iological grounds. In his paper he discusses the relations
of the two systems of nomenclature.

val form of many Chsetopod worms and other Coelomata ;
most remarkable is its agreement with the adult form of
the wheel animaleules or Rotifera . . . .” Again he
says: ‘So far the young molluse has not reached a defi-
nitely mollusean stage of development, being only in a con-
dition common to it and ‘other Caelomata.’” Some of

PHYLOGENY OF THE PELECYPODA. 289

fessor Hatschek’s figures (36-53) of developing Eupomatus (Serpula) shows a close
similarity in structure, proving the identity of the two types at this early stage.

The name typembryo Professor Hyatt applies to the period in developing animals when
the characteristics of the greater divisions begin to appear; when the embryo has char-
acteristic features of the sub-kingdom to which it belongs. He considers the typembryo
of molluses to be the veliger stage, at which period the embryo is sufficiently far ad-
vanced to be recognizable asa mollusc. He speaks of the protoconch (completed embry-
onic shell) as a typembryo.

I think Professor Hyatt has not given the necessary limitations to his typembryo
period in the mollusca. ‘The shell-gland which produces the first rudiments of the prim-
itive shell is considered characteristic of all molluses (vide p. 297). It is the first defi-
nitely mollusean feature to appear, and as it is singularly alike in all molluscs where
known, the embryo at the period when the shell-gland exists is referable to the mollusca;
but has no characters yet which refer it to the class or phylum to which it properly be-
longs. The veliger stage, however, is referable, not only to the sub-kingdom mollusca
as claimed by Professor Hyatt, but also to that class of mollusca to which the embryo
really belongs. The veliger Pelecypod, fig. 25, p. 300, Pl. xxi, figs. 9-11, has two valves
and an anterior adductor muscle; the veliger Gasteropod, Pl. xxu, fig. 19, has a spirally
twisted shell and characteristic foot, tentacles and eyes; the veliger Scaphopod, Pl. xxmt,
figs. 13-14, has a saddle-shaped shell known nowhere else in developing molluscs; the
veliger Pteropod has an anatomy peculiar to itself. I would therefore suggest an alter-
ation in Professor Hyatt’s definition of the typembryo. In the mollusca the typembryo
is the period characterized by the existence of a shell-gland and the plate-like beginnings
of a shell; init normally the ciliated trochosphere dise is becoming transformed into the
velum and the foot is developing. Hxamples of typembryo molluses are given in the
summary of stages later.

_L have shown that the veliger stage of molluscs is referable to the class or phylum to
which the embryo really belongs. As it is a period most distinct and important, I sug-
gest for it the name phylembryo.” The phylembryonic mollusc is an early period in de-
velopment, yet having characters in the shell and anatomy which make it referable to the
class to which it belongs. Commonly a velum exists at this stage; but the stage is not
meant to include the whole veliger period.*?, Examples of phylembryos are given in the
summary of stages.

The completed protoconch and prodissoconch of molluscs is a period later than that
at which the phylembryonic features are fixed and yet is very distinct from the period

Lankester’s and Balfour’s figures of the trochosphere stage
have already acquired a shell gland and in regard to this
character cannot be considered as purely trochospheres ac-
cording to Lankester’s definition, for, with the shell gland
they acquire a feature which makes them referable to the
molluseca. They are trochospheres in which characters
typically occurring later are very early inherited so as to
cause an overlapping of stages.

lgvhov, a race, tribe stock, phylum, and “Eyipy07,
an embryo.

* The existence of a velum cannot be used very closely
as a diagnostic character in considering stages of growth.
It is frequently nearly or quite absent, as in Unio, Anodon,
Pl. xxumi, figs. 7-8, Cyclas, and Spherium, Pl. xxx, fig.
4, or it may exist long after the period of its normal
disappearance. In studies of developing Sycotopus canal-
iculatus, Pl. xxi, fig. 20, I haye found that a large,
highly developed velum exists for a considerable time af-
ter the formation of the true shell, although it is lost be-
fore the young leave the egg cases.

290 ROBERT TRACY JACKSON ON THE

of growth which immediately succeeds it (see figures referred to in the summary). In
view of the fact that it is the close of the embryonic shell period and is expressible as a—
completed protoconch or prodissoconch, I do not think it is necessary to name the pe-
riod of growth. The period in Pelecypods and Gasteropods, where both anatomy and
shell are known in many genera, is referable in origin to the adults of known allied
forms.

In the early period of growth of the true shell, which succeeds the first formed embry-
onic shell, stages have been found which are referable to adult ancestral forms. Pro-
fessor Hyatt studied these stages in Cephalopods and named them the silphologic (grub)
stages. Ina footnote to this paper’ he has given up the use of the term silphologic and
supplied in place of it the word nepionic?

My inyestigations of young Pelecypods and Gasteropods, and also studies of pub-
lished figures of Scaphopods, lead to the conclusion that nepionic stages in molluscs have
characters referable to ancestry in most cases quite close to those naturally grouped as
near allies of the adult. In the nepionic period (see figures referred in the Summary),
it is quite common to find a shell with characters almost inseparable from those of the
adult, or at least so near that the genus can be correctly determined. In other cases,
on the contrary, wide differences from the adult condition may be observed at this early
stage. :

Periods later than nepionic are named by Professor Hyatt successively as nealogic
(youthful), ephebolie (adult), and geratologic (old age) stages.

We will now give a brief summary of the stages of growth and decline as applied to
the mollusca, with figures or references to known forms which may be used as types il-
lustrative of the various stages. The illustration of the stages is a new feature, not given
in Professor Hyatt’s paper. Definitions in this summary are taken from that author’s
paper with the exception of such alterations as are suggested in the foregoing pages.

A classification of stages of growth and decline, in its application to the various groups
of the animal kingdom, must be separately worked out for each group. It is believed,
however, that the following classification contains the essential features necessary for
such applications. New terms may be needed in some groups, but those here used will
probably be applicable to the principal periods of growth in all animals.

Protembryo. The ovum and stages of segmentation of the egg preceding the forma-
tion of the blastula cavity. Examples: figs. 10-15, p. 295, of developing Ostrea, also
Professor Brooks’ figures 1-25 of a developing oyster and Dr. Patten’s figures 2-14 of
developing Patella.

1The use of the word ‘‘silphologic” derived from
cihen, ® grub, when applied to the younger stages of all
animals is likely to lead to confusion of ideas. I have as-
certained, by recent studies of insects, the truth of Brau-
er’s opinion, that the grubs, maggots and caterpillars
common among insects are not normal progressive stages
of development, such as are most common in other types
of animals; but are, strictly speaking, cases of retrogres-
sive or more or less degraded specializations. I desire,
therefore, to give up the use as a general term of the word
Silphologic and propose in its place Nepionic, or young an-

imal, from Nyztos, an infant. This word also is more in
accord with other terms of the nomenclature proposed in
my paper on ‘‘Values in Classification of the Stages of
Growth and Decline,” Proc. Bost. Soc. Nat. Hist. Vol.
XXII, 1888. Signed, ALpneus Hyarr.

* The word nepionic is particularly suitable for early
yet post embryonic stages. Besides its meaning of infant,
applied to the young of man, I am told by Mr. Ernest
Jackson that it is used twice in the Iliad, of young ani-
mals, and by Theophrastus of young plants.

PHYLOGENY OF THE PELECYPODA. 291

Mesembryo. The hollow blastula stage comparable to the condition found in the
adults of Volyox and Eudorina types of the Mesozoa.’| Examples: Pl. xxi, figs. 1, 1a,
of developing Patella.

Metembryo. The gastrula, comparable to the lower Porifera in which three cell-layers
exist, as in the lowest Hydrozoa. Examples: Pl. xxumt, fig. 2, of developing Paludina,
also Hatschek’s figure 5, of developing Teredo.

Neoembryo. The trochosphere. Stages not yet possessing the essential diagnostic char-
acters of the mollusca; comparable to the embryos of Chxetopod worms and other Coelo-
mata. Examples: Pl. xxut, figs. 3-4, of developing Patella, also Professor Butschli’s
figure 6, of developing Paludina.

Typembryo. The period at which an essential molluscan feature, the shell-gland, and
plate-like beginning of the shell are discoverable, and yet the embryo is not referable to
the class to which it properly belongs. Examples: fig. 22, p. 297, of developing Ostrea;
PI]. xxi, fig. 5,of developing Pisidum; PI. xxi, fig. 6, of developing Pleurobranchidium,
also Lankester’s (44), fig. 72, of developing Limnzus. A comparison of the molluscan
typembryo and neoembryo stages with Hatschek’s figs. 36-53, of a developing Eupo-
matus (Serpula) shows graphically the value of these stages in classification.

Phylembryo. arly veliger stages in which the characters of the class subdivisions are
indicated, and the structure of the shell and other features render the embryo referable
to the group of molluscato which it belongs. Examples: p. 299, figs. 23-24, of developing
Ostrea; Pl. xxi, fig. 19, of developing Sycotopus and Pl. xxi, fig. 15, of developing
Dentalium, also Lankester’s (43), Pl. 8, fig. 22, of developing Apysia, and our figs.
7-11, Pl. xxim, of developing Anodon, Modiola, Cardium and Montacuta.

Completed protoconch and prodissoconch. 'The completed first-formed shell of molluses,
the ovisac, fry shell, embryonic shell and larval shell of authors. It finds its representative
in the globular, cup-shaped, or spirally coiled protoconch of Cephalopods. Examples: PI.
xxii, figs. 15,16, of developing Goniates. Also see Professor Hyatt’s (27) figures of
developing Ammonites and Zittel’s figures of young Ammonites, Belemnites and Spirula.
The existence of a cicatrix at the base of the conch in Nautiloids is claimed by Profes-
sor Hyatt (27) as proof that a protoconch existed in the young and a parallel fact in
favor of this conclusion is the finding of a cicatrix at the tip of the conch in Acmza as
described below.

A withered, irregular bulb, found at the apex of the conch in some Orthoceratites,
figs. 1-8, Professor Hyatt considers as the remains of a protoconch, probably originally
of conchyoli structure. M. Barrande (3) figures cicatrices and withered bulb-like
protoconchs in many Nautiloids. See his figures of Cephalopoda, Pls. 487-490, inclu-
sive. Several of his figures are taken from unpublished drawings of Professor Hyatt’s.
~ In some Pteropods a bulb-like expansion of the tip of the shell seems to be the equiy-
alent of the protoconch im other mollusks. Examples: Pl. xx, fig. 17, also several
figures in Bronn’s Thier-reichs. In well-preserved Tentaculites a globular or elongate
protoconch is sometimes found. Examples: Tentaculites acuarius, P|. xxi, fig. 18; also
see figures in Hall’s Palzeontology of New York and other authors.

For the consideration of Volvox and Eudorina as mesozoa see Hyatt, Reference No. 30.

292 ROBERT TRACY JACKSON ON THE

The protoconch in Gasteropods is represented by a spirally twisted embryonic shell.
Examples: Sycotopus canaliculatus, Gill, Pl. xxi, fig. 19. Older specimens than this
from egg cases show a well-marked protoconch at the tip of the conch and I have also
observed a well-defined protoconch in young specimens of Hulgur earica, Con., Pur-
pura lapillus, Lam., Vermetus radicula, St., Buecinum undatum, L., and Crepidula for-

Fias. 6-8. Orthoceras politum, Klipst. Apex of conch
and protoconch viewed from the sides and below.

a, protoconch; e, apex of conch; b, shoulder of apex.

All from the Triassic of St. Cassian (after Hyatt).

and protoconch, viewed from the side, below and in front.
Figs. 4-5. Apex and protoconch of another specimen.

nicata, Pl. xxi, figs. 21 and 23. Holis has been figured with an embryonic shell and
it is important that this genus which possesses no shell in the adult has a typical, spiral
protoconch when young. A very interesting condition is that which I recently found in
the young of Aemea testudinalis var. alveus, Verrill, Pl. xxi, fig. 24, which I collected
at Vineyard Haven, Martha’s Vineyard. No protoconch exists at the apex of the conch;

The embryonic shell or protoconch of Gasteropods is
commonly spoken of by embryologists as a ‘‘ Nautiloid”
shell; but this is not strictly correct. A Nautiloid shell
is spirally coiled in one plane as in the Nautilus from
which the name is borrowed; but the protoconch of Gas-
teropods, in all the numerous cases which I have seen,
either as specimens or figures, is coiled in a spiral plane
and should not be confounded with the Nautiloid form
from which it is totally distinct.

°The spirally twisted protoconch of Crepidula proves
that the almost conical form of the adult is not due to

a primitive, simple condition in which the spiral has not
yet become developed; but rather toa modified condition in
which the ancient spiral form, shadowed in the protoconch,
is almost entirely lost in the adult. Similar observations
have been made on Fissurella in an exhaustive paper by
M. Boutan. The protoconch is yery sharply defined and
is beautifully clear in the young of Crepidula fornicata,
Lam., and C. plana, Say, both of which species are com-
mon on our shores and may be collected in abundance in
July and August. A short paper on this shell is now in
preparation for publication.

PHYLOGENY OF THE PELECYPODA. 293

but, instead, a well-defined cieatrix marks the position where the protoconch has become
separated off. The cicatrix consists of granular lime and fills as a plug the space which
otherwise would be a foraminal opening in the shell. In form the cicatrix simulates the
outline of the base of the separated protoconch as will be evident by comparing the scar
of fig. 24 with the protoconch of Crepidula in Pl. xxi, fig. 23. The cicatrix of Acmza
is made up of two areas, one circular and surrounded by a ring of similar nature, the
other roughly triangular ovate. I find that a similar outline is formed by laying a small,
slightly muddy Gasteropod ona clean glass slide. The closely related genus Patella
from Patten’s studies is known to have a typical spiral protoconch. Therefore, there can
be no question as to the meaning of the cicatrix in Acmea. Professor Iyatt assumed,
as noted above, that the existence of a cicatrix in Nautilus predicated the presence of a
protoconch inthe young and it gives me great pleasure to offer this corroborative evidence
of the truth of his assumption from another class of molluscs.

In Seaphopods the embryonic shell has been named a periconch by Professor Hyatt
(33) on account of its peculiar saddle-shape form. See Dentalium, Pl. xxut, figs. 13-14.
In Pelecypods the bivalvular embryonic shell or prodissoconch is figured in many genera
in this paper. Examples: see almost every plate.

Nepionie period. The period of the first formation of the true shell which succeeds
the embryonic shell and is normally retained throughout the rest of life. The period is
commonly characterized by marked stages. Examples: figures of developing Ostrea,
Pls. xxiv and xxv, figures of developing Perna, Pecten, Anomia, Arginaand other young
Pelecypods figured on Pls. xxvii to xxx, inclusive. Also Professor Hyatt’s (27) Pl. rv,
figs. 7-9, of a young Nautilus koninckii. ‘These last figures have a smooth, nepionic
period, succeeded by a fluted, neologic stage. The nepionic stage is shown in Crepidula
fornicata and in Acmea testudinalis var. alveus, P|. xxiu, figs. 21, 23, 24. It may be
observed in Sycotopus, Fulgur and Buecinum when removed from their egg cases, in
which condition the shell is very finely preserved.

Vealogic period. The period succeeding the nepionic and preceding that period which
may properly be considered as the adult. It is frequently characterized by marked
stages, being a period in which the growing animal often differs widely from the adult,
and as it is of considerable size, the differences and stage are easily recognized. Ex-
amples: Hinnites cortesi, Pl. xxv1, figs. 3-4, has a pecteniform nealogic stage sharply
marked off from the ostreaform adult condition. Slava jfibrosa from the paleozoic Pl.
xxi, fig. 12, is one of the most striking cases of definition of the nealogic period that I
have ever seen. The nealogic stage as shown in the figure is concentrically plicated and
is highly areuate, whereas the succeeding ephebolic period is radially plicated, spread
out in a fan-like fashion and is much less arcuate. Many species and individuals of this
remarkable genus are figured by M. Barrande. The period is strongly marked in the well-
known Clavagella and Aspergillum and the last mentioned genus has a definitely marked
nepionic stage as well. The young shell of these two genera has up to the close of the
nealogic stage two equal valves. In the adult the bivalvular condition is lost; the two
valves having united so as to form a single tubular shell, which by itself would not be
recognizable as a Peleceypod. In the Gasteropoda, Vermetus, fig. 9, and Magillus, the
nealogic period is close coiled, whereas the adult is loosely coiled and vermiform haying

MEMOIRS BOSTON SOC. NAT. HIST., VOL. IV. 40

294. ROBERT TRACY JACKSON ON THE

lost nearly all the characteristics in the shell of the Gasteropod form. Marked nepionic¢
and Beslonic stages of growth are a striking feature of many Gasteropods as seen in
Yetus, Cypreea, Strombus and Pteroceras. These and other stages
are also strongly marked in many Cephalopods, where they are de-
scribed by Professor Hyatt (384 and other papers).

Ephebolic period. That period best characterized by saying that in
it the adult characters find fullest expression; it is often separable by
marked stages from the earlier nealogic period, and also from the later
or senile period. As examples we would refer to the adult (middle
age) condition of the figures referred to as illustrating the nealogic
period and also the next or geratologic period. In Slava fibrosa, P|.
xx, fig. 12, the radially plicated area is referable to the ephebolic
period. In Hinnites cortesi, Pl. xxvi, figs. 3-4, the ostreaform stage
is ephebolic and in Vermetus radicula, fig. 9, the vermiform stage is
the ephebolic period.

ae +e Renae rad Geratologic period. 'The period of decline of the individual, often
cula, St. After Ver- ee o C 5 eae
rill.) marked by distinct stages. The geratologic period is subdivided by

Professor Hyatt into the clinologic and nostologic periods, signifying the early and
later periods of decline of the individual. As examples of geratologic stages, I would
refer to Venus Tridacnoides from the Miocene Tertiary of this country and to Cardita
planicosta, Say, from the Eocene Tertiary of Claiborne, Ala. In both forms old age is
strongly indicated by a loss of costations and other characters found in the ephebolic
period, the new period assuming a strikingly different appearance. In the old age of a
Brachiopod, Atrypa rugosa, may be observed a most marked change; the regular even
plications of the adult become irregular and spread out in a fan-shaped fashion. ‘This
is shown in many of M. Barrande’s figures and is more or less completely developed in
some specimens of any large collection of American or European specimens.

TV. Ostrrea: DEVELOPMENT OF THE SOFT PARTS.
EMBRYOLOGY. ,

The early embryonic development of the oyster may be followed in the writings of
Dr. R. Horst on the European oyster, O. edulis, L., and Professor W. K. Brooks on the
development of our oyster, O. virginiana, Lister. The later stages of development have
been described by Professor Huxley and Dr. R. Horst in the European oyster, and Pro-
fessor J. A. Ryder and the author in our species.

The European species, Ostrea edulis, L., is hermaphroditic according to European in-
vestigators and Professor Ryder, and the embryos undergo considerable development
in the gill chambers of the parent. Our species, O. virginiana, Lister, on the contrary,
is not hermaphroditic, according to Brooks (9), and Ryder (61), and differs from O. edulis
in that the eggs are discharged from the parent and undergo their whole development
in the sea. It produces many more eggs than the European species, and the embryos
are much smaller than those of O. edulis.

PHYLOGENY OF THE PELECYPODA. 295

LI have not followed the early embryonie development of the oyster, and therefore, for
the stages considered make reference to the publications of Professor Brooks (9), Dr.
Forst (23, 24) and Professor Huxley (26).'

As will be seen in studying the oyster both the soft parts and the hard, we have to deal
with a member of a highly peculiar and aberrant group, varying much from the charac-
ters commonly met with in molluses. Many of the characters, and especially their period
of occurrence in the life of the individual, are well explained by the law of concentra-
tion and acceleration of development, which is considered in the introduction. By this
law a highly modified or “ concentrated” type is one which acquires characteristics very
early in its development, earlier than the normal, so to speak, and such is found to be
the case in a marked degree in the oyster.

The egg after fertilization undergoes a period of segmentation until it becomes a solid
mass of cells, see Brooks’ figs. 1-29 and our figs. 10-15. These stages, following our
classification, are considered as protembryo stages. (Nomenclature of stages, section IIL.)

Fic. 16,

DEVELOPMENT OF OSTREA EDULIS (AFTER HORST).

Fic. 10. Ege of oyster, with germinal vesicle and spot Fia. 15. More advanced stage seen from above, where
Ss! y s p

visible within the vitellus.

Fic. 11. Beginning of development; the germinal ves-
icle has become invisible and polar globules appear.

Fig. 12. First stage of segmentation; the egg is divided
into two spheres of unequal size (animal and vegetative).

Fig. 13. More advanced, the egg being divided into four
spherules.

Fic. 14. Stage when the egg shows one large vegeta-
tive sphere and several animal spheres.

the vegetative sphere is divided into two spherules.

Fic. 16. Embryo side view at the beginning of invagi-
nation (gastrular).

Fic. 17. Embryo more developed; optical section, with
invagination of the entoderm and beginning of the pre-
conchylian gland: ec, ectoderm; en, entoderm; 0, blasto-
pore; sk, preconchylian gland.

Fig. 18. Embryo a little older, side view: v, pediform
appendix; other letters as in fig. 17.

The next stage in the development of the embryo, taking as a normal that which has
been commonly observed by embryologists, should be the formation of a morula stage,
the cells arranging themselves in a peripheral layer around a hollow sphere. According
to the investigations of Professor Brooks and Dr. Horst, there is no true morula stage

1 In the consideration of these early stages such points an abstract or even a discussion excepting in a limited

are selected from the writings referred to as serve to il-
lustrate this paper. I do not pretend in any sense to give

sense of the early development.

296 ROBERT TRACY JACKSON ON THE

in the oyster, the gastrula invagination commencing before the formation of a hollow
sphere. We have, therefore, no stage comparable to our mesembryo stage, Pl. xxit,
figs. 1, la, its absence being explicable as a case of extreme concentration of develop-
ment in the oyster.

In the later segmentation stages, Brooks’ figs. 22-28, also our fig. 14, we find that
those cells which are of ectodermic and entodermic nature are clearly defined. The en-
toderm consists of a single macromeric cell surrounded in an increasing degree by mi-
cromeric cells up to the period shown by Brooks’ fig. 26. The single macromere then
divides into two, Brooks’ fig. 28, then into several cells, his fig. 30. Comeident with
the splitting up of the entodermic cells there is, according to both Brooks’ figures and
our fig. 16, a flattening of the embryo from above downwards, and a pushing in of the
entodermic cells, which commencing as a slight depression, Brooks’ fig. 31, our fig. 16,
soon forms a deep invagination, the gastrula cavity; see Brooks’ figs. 82-35 and our
fig. 17. :

From this point on there are some important differences in the observations of Pro-
fessor Brooks and Dr. Horst. As Horst’s observations were made on embryos of
O. edulis, which ave much larger than those of O. virginiana, and as they are more in
accordance with observations made on the development of other Pelecypoda and mol-
lusca I have for the most part followed them where differences exist.

In our classification of the stages of growth, the metembryo mollusc represents the
gastrula stage, Pl. xm, fig. 2, being a period in which two well-formed and an incipient
third layer exist. Professor Brooks’ figures 32-35 represent different aspects of the gas-
trula oyster according to his observations. His fig. 33, an optical section, shows an em-
bryo with a primitive gastrula cavity and body cavity. It also has a velum slightly devel-
oped. These figures answer well the requirements of the metembryo stage and may be
said to represent that period in the development of Ostrea virginiana.

Dr. Horst, considering the stages (our figs. 16-17) which are referable to the gastrula
period, says, “ there can however be no question of a genuine invagination because there
is no cavity of segmentation; we have here, so to speak, an intermediate form between
an embolic gastrula and an epibolic gastrula.'"” He notes the remarkable fact that coinci-
dent with the gastrula invagination at the vegetative pole, there is also a very distinct
depression at the other pole, a little below the dorsal side, our fig. 17sk. This second in-
vagination by further development (our figs. 18-19) he shows is the beginning of the
preconchylian invagination.

The gastrula invagination, according to his figures, begins before the preconchylian in-
vagination; but the latter exists before the completion of the gastrula invagination.
Horst’s figures do not show any trace of a velum at this period, which is somewhat pecu-
liar, as he represents the oyster as so much more advanced in its mollusean characters than
does Brooks, who figures the velum at the gastrula period. The early absence of the velum
in O. edulis may be connected with the brooding habit of that species, for in Pelecypoda,

1 According to Hyatt (Larval Theory of the Origin of a process much shorter than is usual in embryos of the
Tissue, p. 156), in the formation of epibolic gastrule embolic type. He accounts for the epibolic gastrula by
the growth of the ectoblast cells is so rapid that the endo- heredity and extreme concentration of development.

blast cells become enclosed, and the gastrula is formed by

PHYLOGENY OF ‘THE PELECYPODA. 297

which brood the young, as Unio and Pisidium, the velum is highly reduced or wanting.
From Horst, we may gather the conclusion that the metembryo (gastrula) stage in the
European oyster, fig. 17, is not a pure gastrula, but has already acquired by concentra-
tion of development, some of the characters which would normally appear later.

Professor Brooks says that the primitive blastopore closes over, becomes the dorsal as-
pect of the embryo and gives rise to the shell; the permanent mouth and anus forming
on the opposite side of the embryo and independent of the blastopore. This differs so
essentially from the observations of Dr. Horst in the European oyster and from those of
other investigators except Rabl, in the development of various groups of mollusca, that
we prefer to follow Dr. Horst’s observations. Professor Brooks himself says (9), p. 76,
that the whole evidence furnished by comparative anatomy and embryology tends to
show thatthe mouth, anus and shell-gland can be homologized perfectly in all the classes
of true molluscs and that they are not only homologous with each other, but must be
perfectly homologous also with similar structures in the ancestral form of which the
molluses are modifications. It is evident that these several organs would not be “ per-
fectly homologous ” if in one case the shell-gland arose from the closed blastopore and in
another from the opposite pole of the body, and the same would be true of the derivation
of the mouth and anus.

Dr. Horst in considering the development of the preconchylian gland! says that when
the embryo is viewed from the side the glandular depression is seen as inour fig. 18, sk, and
in fig. 17 in optical section. During the further development, our figs. 19 and 21, a small

Fig. 19. Fic. 20.

DEVELOPMENT OF OSTREA EDULIS (AFTER HORST).

Fig. 19. Embryo same stage as fig. 18, optical section ; Fic. 21. The same embryo, optical section.
me, mesoderm; d, primary gastrula invagination ; si, shell- Fig. 22. Embryo a day older, with a wreath of vibra-
gland. tile cilia, a stomachic cavity and the beginning of a shell, s.

Fig. 20. Embryoa day older, front view, with the prim-
itive mouth opening.

invagination is formed bordered by walls of high cylindrical cells with a narrow opening.
Considering fig. 19, Horst says that the gastrula cavity has deepened, and behind the
mouth are cells, probably mesodermic. ‘The blastopore continues to be very distinct, and

as far as he could ascertain, it did not disappear but remained and was transformed in-
to the mouth.

1 The preconchylian gland was discovered by Lankester Nassa (see Balfour); by Patten, in Patella; by Brooks
(44) in 1871, and has been figured by him (44) in Loligo, in a pulmonate Gasteropod; by Hatschek, in Teredo; by
Aplysia, Pleurobranchidium, Limneeus and Pisidium. It Horst in Ostrea and in other genera by various inves-
has also been observed and figured by Fol in a Heteropod tigators. This array of genera confirms the opinion of
(see Balfour); by Rabl in Planorbis (see Claus); by Lankester that it is typical of all developing molluscs.

Butschli, in Paludina (see Claus); by Bobretsky, in

298 ROBERT TRACY JACKSON ON THE

Dr. Horst says that great changes take place during further growth of the embryo.
The preconchylian gland gradually loses its primitive character of a glandular invagi-
nation and forms a thickening of the ectoderm composed of long, conical cells, fig. 22, sk.
A thin, cuticular membrane secreted by these cells is the first indication of the shell, and
at this position in the adult the hinge is found.’ Horst maintains, therefore, that the
shell of the oyster originates as a single plate in accordance with the observations of
Brooks in Anodon and Hatschek in Teredo2 Meanwhile in fig. 22 a body cavity has
formed by the separation of the ectoderm and entoderm, a velum has developed and the
gastrula invagination has formed a spacious cavity in the base of which is a diverticu-
lum, now closed, but which later communicates with the ectoderm and forms the poste-
rior part of the alimentary canal. The neoembryonic mollusc in our classification 1s rep-
resented by the trochosphere stage before the development of the shell-gland, Pl. xxm,
figs. 3-4. In the developing European oyster the shell-gland, according to Horst, orig-
inates early, closely following the gastrula invagination and before the formation of a
normal trochosphere stage. I think, therefore, that we must consider the neoembryo
stage as skipped in the oyster by the overlapping of stages due to extreme concentration
and acceleration of development.’ Fig. 22, on the other hand, is the nearest approach to
the ideal typembryo stage that we have in the developing oyster (definition of typem-
bryo, see p. 291). The primitive gastrula cavity has not yet become connected with the
ectoderm by the secondary opening, the anus, as is ordinarily the case, but the existence
of the shell-gland, with the beginnings of the velum, entitles it to be considered as a
typembryo. The existence of these comparatively high characters in the young oyster,
at a period when the primitive oral invagination still retains almost its early simplicity
and no second alimentary opening has been formed, must be considered as strong proof
of the highly concentrated development of this type.

Taking up Horst’s account of the developing oyster again, he says that on the follow-
ing day (our fig. 23) the shell, which has grown very much, covers a large portion of the
body. It already contains carbonate of lime, as shown by treatment with acids. Of
fig. 24, Horst says, the velum now forms a very distinct part of the body, which is al-
most entirely covered by the shell. The surface of the velum surrounded by a wreath
of vibratile cilia already shows in its centre a thickening which is the beginning of the

cephalic shield. An cesophagus in the shape of a funnel leads to a large, pear-shaped

“which occupies so precisely the required position, the
“homologue of that production.”

2 The primitive shell in Cyclas and Unio has also been
observed by investigators to begin as a simple, unpaired

14 horny plug, the product of the cells of the shell-
gland, has been described in Gasteropoda and Pteropoda
(Lankester (44), Balfour). The typical universality of
the shell-gland cannot be doubted (note, p. 297) and the lig-

amentous union of the valves of Pelecypoda occupies the cup.
precise position of the shell-gland. It seems reasonable 3Lankester, in his article Mollusea (44), refers to

Horst’s fig. 14 (our fig. 23) as the trochosphere stage. But
as the shell has not only considerably developed, but con-

to follow Lankester’s suggestion, based on Pisidium (43)
and homologize the horny ligament of Pelecypods with

the horny plug of embryonic Gasteropods and Pteropods.
To quote Lankester, ‘‘In Pisidium as will be seen from
“Pl. ry, figs. 38a and 39, the two calcareous valves do not
“make their earliest appearance in close contact one with
“the other. The central portion of the shell-gland is not
“*eoncerned with them; and since it is precisely that point
“« which in the Gasteropods cited gives rise to a chitinous
““plug, may we not see in the ligament of the bivalve,

sists of two valves, we think this figure is referable to
the class of Pelecypoda, a period later than the tro-
chosphere stage. Lankester’s own definition of the tro-
chosphere is that at this period the embryo has not yet
acquired molluscan features and is comparable to the
larval forms of many Chetopod worms and other Coelo-
mata, which is not the case in the figure cited.

PHYLOGENY OF THE PELECYPODA. 299

stomach which communicates with the outside through the intermediation of an in-
testinal canal.

Our fig. 23 is distinctly the essential equivalent of Brooks’ figs. 88-42. Our fig. 24
again is sees ae ae close to Brooks’ fig. 44. At this last stage there are two well-devel-
oped valves, a stomach, mouth and anus, a ciliated foot and a large velum. In fig. 24a

Fic. 23. Fic. 24.

DEVELOPMENT OF OSTREA EDULIS (AFTER HORST).

Fic. 23.—Embryo at more advanced stage than fig. 22, Fig. 24.—Embryo still more developed; v, velum; @,
side view, with shell more developed. cesophagus; st, stomach; a, anus; and kp, cephalic plate.

body is represented which, although Horst does not mention it, appears to be the be-
ginning of the adductor muscle of the valves. It is in the same position as that oecu-
pied by the adductor in Horst’s fig. 16, Brooks’ fig. 44 and our fig. 25, p. 300.

T have described as the phylembryo an early Read of development (see section 1m)
where, however, the class characters are recognizable, and I consider figs. 24—25 as phyl-
embryonic stages of the oyster. They have two equal valves and an anterior adductor
muscle, characters which render the embryo referable to the phylum or class Pelecypoda,
to which it properly belongs, although not referable to any known adult form.

The mouth and anus when Seve knee arose in close proximity on the ventral bor der
of the embryo, Brooks’ fig. 38, our fig. 23, which is a generalized characteristic of a mol-
lusean embryo. It is seen in fig. 24 that the first adductor muscle develops anteriorly
and close to the dorsal margin of the velum. The retractor muscles of the velum and
liver then develop. The anus, from its early ventral position, revolves dorsally in the
plane of the valves (see Brooks’ figs. 38-44). The intestine by interstitial growth
makes a single loop-like curve on the left side; this with some other slight changes
brings us to the stage, fig. 25, described by Huxley (26). Ryder (63), remarking on
this stage, notes that the intestine is already flexed on itself in much the same manner
as in the adult, though it does not extend as far anteriorly, and its position is modified
later by flexions of the stomach and esophagus.

The shell of the stage, fig. 25, is symmetrical, equivalvular, and of about the period of
development of Pl. xxtv, figs. 13-16, of the developing shell of this same species. The
mouth has yet no palps, a sac-like process of the stomach on either side represents
the liver, two pairs of muscles serve to retract the velum, and the single adductor muscle
closes the valves. Huxley observes that this early anterior adductor cannot be the
equivalent of the single adductor of the adult as it is on the ventral side of the alimen-
tary canal whereas that of the adult is on the dorsal side. He says that a second posterior

300. ROBERT TRACY JACKSON ON THE

adductor must be developed and this early anterior adductor lost in later development.
This was all that was known concerning the development of the adductor muscles of
the oyster when I published my preliminary paper in which it was shown that two adduc-

tors exist at the completed prodissoconch stage.
A stage of O. virginiana of similar age to our fig. 25 of QO. edulis has been figured

FIG, 25.

EMBRYO OF OSTREA EDULIS
L. (AFTER HUXLEY).

The phylembryo stage,
viewed from the left side. a,
ad, anterior adductor muscle,
which alone exists at this
stage; m, mouth; @, cesopha-
gus; @,anus; 7, intestine; st,
stomach; 71, left lobe of the
liver; v, velum; 7s, 77, Supe-
rior and inferior muscles,
which retract the velum into
the shell, sh; the hinge, h, is
straight. Compare with Pl.
xxin, figs. 7-11 of phylem-
bryonic stages of developing
Pelecypods, where as here the
hinge line is straight and the
anterior adductor alone ex-

by Ryder (62). The valves were nearly disc-shaped, but flattened
at the hinge area. A ciliated velum existed, and as Ryder noted,
“the embryos appeared to be attached to the glass by a marginal
extension of the mantle over the edge of the lower valve, forming
a sort of lip,” but they were not firmly fixed as “the embryos
swung to and fro.” He did not succeed with later attempts in
getting similar attachments, and from my own observations I
have every reason to believe that the fixation was of a transitory
nature, not to be confounded with the fixation which young oys-
ters later acquire when they become fixed for life.

T have thus far considered the early development of the oyster,
using facts observed by other investigators. This is done in order
to make clear the later portion of the embryological development
and also to bring out some facts elucidating the development of
the group of Pelecypoda. All later observations are my own, ex-
cepting where specially credited. .

Between the stage fig. 25, and our next stage, Pl. xxrvy, figs.
1-2, there is a blank in the knowledge of the development of the
oyster. It has not been described in the European species, and
all attempts to obtain it in our species have failed. In artificial
confinement the oyster dies at this stage.

The next step in the anatomy which I have to consider is that
of the oyster when it has completed the prodissoconch growth

ists.

and just preceding the beginning of the dissoconch, or “spat
growth.” This period is the last of the embryonic stages, using as bases of consideration
the structure of the soft parts and the hard as well as a comparison with a similar period
in other developing Pelecypoda. In my preliminary paper (36) I showed the important
fact that the oyster at this stage has two adductor muscles, situated relatively to the ali-
mentary canal as they are in adult dimyarian Pelecypods. The specimen then figured
was from alcoholic material collected at Buzzards Bay, and the parts had contracted so
that some features were not successfully made out; the muscles however were rendered
clear by treating the specimen with dilute acetic acid, and thus dissolving the calcareous
shell. Living material of this age was obtained later growing on glass slides, so that it
could be studied more satisfactorily.

On the third of August, 1888, clean glass was put into a drain-pipe trap on the sand-
bar. The following day it was examined and a young oyster, Pl. xxrv, figs. 1-2, was
found attached. No spat growth had taken place, but the embryo was firmly attached
by the ventral margin of the left valve. On driving a current of water forcibly over
the glass from a pipette it did not rock or move from its position, thus demonstrating

PHYLOGENY OF THE PELECYPODA. 30]

the firmness of the fixation. As the oyster had been attached less than twenty-four
hours, the anatomy and shell here represented must have been developed while it was a
free, unattached embryo. The fact that the embryonic shell is always fixed by the ex-
treme ventral margins only, is further proof that fixation takes place at the close of this
period.

The embryo was yellowish-green and lively, constantly extending and retracting
the mantle and gills and moving the palps. The important feature of the two adduc-
tor muscles was clearly made out as shown in Pl. xxtv, fig. 1, viewed from the right side,
and in fig. 2, viewed from the left side through the transparent glass to which the em-
bryo was attached. The intestine wraps around the dorsal aspect of the posterior ad-
ductor, as in all later stages. A large, lobed, ciliated velum still exists and is situated
on the ventral aspect of the anterior adductor muscle, lying close up to that body. The
ciliated and lobed palps are situated close up behind the velum. Club-shaped, tubular,
ciliated gills, Pl. xxrv, fig. 1, also fig. 6, with the tips not recurved upon themselves, exist.
The mantle border is broad and thick, bears no tentacles and is capable of considerable
extension beyond the margins of the valves. Within the umbo of the lower left valve
was a rhythmically, rapidly pulsating body, as of a large blood-vessel. The liver forms
a large, light green mass in the upper part of the body and extending into the umbos.

About three and a half hours after the specimen was first seen a slight spat growth
had taken place along the ventral margins of the valves. It was kept alive several days,
during which growth proceeded rapidly and was studied at intervals; Pl. xxrv, fig. 3,
being three days older than figs. 1 and 2. We will now discuss this stage, which is of
great importance, both in the development of the oyster and related Pelecypoda as well.

The anterior adductor muscle of the early development, fig. 25, p. 300, is still existent in
this stage, Pl. xxrv, figs. 1-2, a, ad, but there is another muscle, the posterior adductor,
as well. We have then the interesting feature of a monomyarian Pelecypod in its early
stages possessing two muscles, situated relatively to the alimentary canal as they are in
typical adult dimyarians. he anal part of the intestine revolves dorsally during the
early development of the oyster and the posterior muscle when formed must have origi-
nated on the ventral aspect of that organ. If the posterior muscle had developed on the
dorsal side of the intestine it would always have retained that relative position as the
permanent anterior adductor of dimyarians does in relation to the mouth and oesophagus.
One of the striking characteristics of the ostrean prodissoconch is, that the umbos point
posteriorly as shown in PI. xxty, figs. 1-2. This is contrary to the condition observed
in developing dimyarians, as in them the umbos point anteriorly, but an exception is found
in Nucula, the embryo of which, according to Lovén, has umbos directed posteriorly. This
character of the umbos is discussed in its serial connections in section xv.

In Ostrea, fig. 25, indisputably the anterior adductor is first developed; also in Car-
dium, Modiolaria and Montacuta, Pl. xxmr, figs. 9-11, in Pisidium according to Lankester
(43) and in Anodon,! Pl. xxur, figs. 7-8, according to Schierholz’, the anterior ad-
ductor is first developed. I have found only two possible exceptions to the rule of the

1Tn the embryo of Anodon, Pl. xxm, figs. 7-8, the first first formed adductor is probably in part, at least, the be-
formed adductor muscle is in the larger side of the shell. . ginning of the permanent anterior adductor as in the other
According to Schmidt’s figure of a later stage, the larger Pelecypoda above considered.
side is the anterior end of the embryo. Therefore the 2Schierholz’ paper, see Schmidt’s reference list.

MEMOIRS BOSTON SOC. NAT. HIST., VOL. IV. 41

302 ROBERT TRACY JACKSON ON THE

anterior adductor being developed first. These are Mytilus and Unio. As the nearly re-
lated Modiola and Anodon follow the rule these two genera may be considered as excep-
tions, or perhaps illustrate errors of observation as discussed in my preliminary paper,
(35), p. 535.

In Ostrea the mouth and anus develop ventrally, the anal extremity of the intestine re-
volves dorsally and the anterior adductor is developed. Later, the posterior adductor de-
velops on the ventral side of the intestine. When one compares this with the figures and
descriptions of the development of the above-mentioned genera (excepting Mytilus and
Unio), together with the anatomy of the adults, it seems that there is a close uniformity
of plan. The genera mentioned are not numerous, but they are taken from widely sep-
arate groups of Peleeypoda. There is then strong evidence in favor of the development
of the anterior adductor first in Pelecypods as the typical mode, and the development of
the posterior adductor later, after the intestine has revolved into place, from its early ven-
tral position. This theory easily explains the constant relative positions of the mouth
and anus to the two adductors in the adults of dimyarians, or to the single adductor,
where only one exists, as in monomyarians.

Tt is seen that there are two distinct stages of development of the muscles of the em-
bryonic oyster. First, a single-muscled stage, the early monomyarian, fig. 25, p. 300,
pointing towards a problematical ancestor, which in the adult condition had only one ad-
ductor muscle, and that one in the anterior portion of the body.’ The second stage is the
two-muscled stage, dimyarian, Pl. xxrv, figs. 1-2, pointing to an ancestral adult form
which had two muscles like the typical dimyarians. The figure of this stage shows the
early formed velum to be still existent, but it is not supposed that the early dimyarian,
adult ancestor possessed a velum. It exists here presumably because of the lapping
over of stages, due to concentrated development, which is most marked in the early em-
bryonic development of the oyster. ‘The shell which covers these stages is continuous
in outline.

As stated, a velum still existed at this dimyarian stage, Pl. xxrv, figs. 1-2. The long
vibrating cilia were in active motion and some motion of the velum as a whole was
noted; but it was not seen to extend beyond the margins of the shell. Attempts have
been made by Horst and Ryder (62) to discover the method by which the young oyster
attached itself to the object of permanent fixation and both investigators surmised that
it might have a brief byssal attachment. Ryder, in considering the fixation of the em-
bryos described by him,’ thought that it was effected by means of the border of the man-
tle reflected over the edge of the undermost valve as shown in his figure. The nearest
approach to a foot known in the developing oyster is that shown in fig. 24, p. 299, and I
discovered no traces of a foot in my youngest specimens. As the byssus is an organ de-
veloped in the ventral portion of the foot, the high reduction or almost complete absence
of that organ is in itself strongest evidence against the suggestion that the attachment

'Tt is possible that the first formed muscle does not ° The fry described is comparable inage tomy Pl. xx1v,
point to an ancestral single-muscled Pelecypod, in which figs. 18-16, and is not to be confounded in its attachment
only the anterior adductor was developed. It may point with the permanent attachment made at the close of the
to. some muscle in the primitive ancestral mollusc from prodissoconch stage, Pl. xxiv, fig. 17.

which the Pelecypeda were derived, and which muscle be-
came the anterior adductor of the Pelecypoda.

PHYLOGENY OF THE PELECYPODA. 303

of the young oyster is effected by means of a byssus of however short a duration. In my
studies of the young of Avicula, Perna, Pecten and Anomia, Pls. xxviI—xXxIXx, it is shown
that a notch for the projection of the byssus, or foot, is marked in the dissoconch shell im-
mediately after the close of the prodissoconch stage. On most careful study no such indi-
cations of a byssal attachment were seen on the lower valve of attached oysters, and as
they were growing on transparent media, fullest opportunity was enjoyed to make such
observations. In view of the evidence it is therefore safe to conclude that the oyster does
not have a byssus at any period of its development. The preliminary fixation is probably
effected by means of the reflected mantle border, as-described by Ryder, and is then im-
mediately succeeded by a cementing conchyolin attachment of the extreme edge of the
lower left prodissoconch valve.

The fact that a velum, or swimming organ, exists up to the period of permanent fixa-
tion, accounts for the great reduction of the foot, because that organ is unnecessary while
the animal is provided with another locomotive organ, and is useless for progression af-
ter the animal is permanently attached. _'The reduction of the foot is clearly attributable
to disuse and a high degree of concentration of development. This is another of the cases
of extreme concentration of development seen so markedly in the development of the
oyster.

The gills of the young oyster, Pl. xxrv, figs. 1-2, and fig. 6, consist of two rows of fila-
mentous appendages, one on either side. They are fixed to the body mass anteriorly,
but are free posteriorly, and are capable of considerable dorsal and ventral rotation.
Hight club-shaped filaments were counted in either row at this early stage. Hach fila-
memt is ciliated and perforated by a tube which expands dorsally into a canal connect-
ing the tubes of the several filaments. The filaments are entirely free and separate from
one another excepting on their dorsal border and are not bent back upon themselves on
their ventral margins as they are in later stages. Each filament represents a hollow bag,
or pouch, in which the blood is aerated by water kept in constant motion by the active
cilia. This stage of development of the gills is exactly comparable to that figured in the
young of Mytilus by Lacaze-Duthiers (41). The condition is also comparable to that de-
scribed by Mitzukuri in the adult of Nucula excepting that in that genus a pair of filif-
erous branchiz exist on either side. After tracing the further development to the adult
condition in Ostrea it will be seen that the gills of the other Pelecypoda described are
only modifications of this simple early type.

V. OstreaA: DEVELOPMENT OF THE SoFT Parts.

THE SPAT.

At the close of the period just described, the young oyster enters on stages of growth
both in the soft and hard parts which are rapidly approaching the conditions character-
istic of the family to which it belongs. They are stages distinctly post-embryonic and
I shall consider them as the nepionic or babyhood stages."

The young oyster, Pl. xxiv, figs. 1-2, growing ona glass slide, was kept alive in a

1 For a discussion of the nomenclature of stages vide section ILL.

304 ROBERT TRACY JACKSON ON THE

beaker of sea water for four days. When twenty-four hours older than the figures cited,
a good deal of spat growth had taken place; the velum had disappeared and the anterior
adductor was difficult to discern, but I think still existed in a position much nearer the
hinge margin than in figs. 1-2; the muscle had therefore revolved dorsally. The pos-
terior adductor had not changed much in its position; but the palps had revolved dorsally.
The animal was yellowish-green as on the previous day, and the mantle border was
wide and actively extended and retracted. The action of the parts was more clearly seen
when viewed from the left side through the glass to which it was attached than from the
right upper side. From this view many observations were made. ‘The gill filaments at
the end of twenty-four hours were not yet joined to one another on their ventral mar-
gins as in later stages.

When the same specimen was forty-eight honrs older than Pl]. xxtv, figs. 1-2, no an-
terior adductor muscle was seen. The posterior adductor from its earlier position had
moved ventrally so that it Jay near the border of the prodissoconch valves. The palps
had revolved dorsally so that they were situated near the hinge line, but were directed
ventrally; a fact to be discussed later. I did not see the radial muscles of the mantle
visible in later stages, but they doubtless existed; no marginal tentacles were yet devel-
oped. The rhythmical pulsating body, noticed on the first day in the umbo of the left
valve, had disappeared or become hidden in the increasing liver follicles. No heart
was discovered on the dorsal border of the adductor, where it is later found. The liver
was greenish brown, not light green, as at first.

Pl. xxiv, fig. 5, is from the same specimen three days older than figs. 1-2. The spat
shell has grown considerably, the posterior adductor has moved to a more ventral posi-
tion than it occupied forty-eight hours after attachment, and the palps have moved far-
ther dorsally. The gill filaments are now connected on their ventral margins by cross
connecting bars, and this is the beginning of the intricate system of cross connecting
network, characteristic of the adult ostrean gill. The filaments in young and adult
Pecten are joined by interlocking but separable cilia, similarly to those of Mytilus as
described by Peck, but the connection in Ostrea is a complete organic union brought
about by the concrescence of fleshy processes from the margins of adjoining filaments.
The cross bars begin to be developed between the filaments in the anterior portion of the
gill and from that region are formed posteriorly in a serial manner between successive
filaments. In oysters up to 2.2 mm. in height only a single series of cross bars exist
and those are on the ventral margin as in Pl. xxrvy, fig. 3. The ends of the gill fila-
ments in the spat, fig. 3, are recurved and joined by concrescence at their tips with the
recurved filament tips of the opposed gill-lamella. This is understood by comparing the
diagrams of the gills of developing Ostrea of this stage, Pl. xxrv, fig. 7, with those of an
adult Pecten, Pl. xxrv, fig. 12. The contiguous surfaces of the direct and reflected
portion of the filaments are not organically connected at this early period, so that each
gill-lamella represents a hollow cavity bordered by its filamentous walls and a condition
similar to this has been observed in adult Perna. The two gills are of about equal size
at this period. JTorst states that in young Ostrea edulis the gills are filamentous, jomed
only at the bases and tips; and he compares this condition with Lacaze-Duthiers’ (41)
well-known observations on Mytilus.

PHYLOGENY OF THE PELECYPODA. 305

We will now consider an oyster, Pl. xxtv, fig. 4, considerably older than our last fig-
ure. Ryder (63) published a figure and description of an oyster somewhat younger
than this and yet older than my PI. xxtv, fig. 3. He compared the anatomy of his fig-
ure with that of Professor Huxley’s figure (my fig. 25, p. 300) and demonstrated serial
connections between these and the adult. Pl. xxry, fig. 4, illustrates a young oyster of
the spat stage growing on a glass slide, on August 6, 1888. The single adductor muscle
occupies a sub-central position similar to that of the adult. The palps have revolved
dorsally and assume a position intermediate between that shown in Pl. xxtv, fig. 3, and
that characteristic of the adult, Pl. xxv, fig. 12. They differ from the adult in that the
outer pair of palps forms a hood above the mouth which opens downwards and not as
directly forwards as in the adult. Both these differences are discussed by Ryder (63).
Four palps exist, but they are marked by few (five) furrows on their surface, instead of
a great number as in the adult. Ryder observes this difference, and says that though
few in the young he has counted over a hundred furrows on one side of the lower palp
of an adult.

The heart is situated in a pericardial cavity above the adductor muscle, as in adults.
From ninety to one hundred and ten heart-beats per minute were counted. This is a very
high number, as Professor Ryder (64) notes that in the adult the heart-beats probably
do not exceed twenty per minute. <A large pulsating artery, v, passes to the mantle
border where it is confluent with the posterior tips of the gills, and lesser ramifying ar-
teries bifurcating frequently are spread over the mantle in every direction. These are not
shown in the figure as they would confuse other details. The blood corpuscles travers-
ing the arteries may be readily seen through the thin shell as indicated by arrows and,
when the heart is beating rapidly, the blood may be seen returning to the heart through
the canals on the dorsal border of the gills. It passes out from these canals through the
foramina ¢ and d, and also from between foramen d and a point nearer the anterior tips
of gills. At d the afferent and efferent vessels are connected and though most of the
blood from the gills passes into the heart, some corpuscles may be seen joining the ef-
ferent current and flowing into the large artery, v, which supplies the mantle border.

This specimen growing on glass afforded excellent opportunity to study the gills.
The filaments of the two gills, A, A,’ Pl. xxrv, fig. 4, are connected by several rows of
cross bars, instead of only a single row on the ventral margin. (Compare with PI.
xxty, fig. 5.) An oyster, 6 mm. high, was observed to have five cross filaments in the
larger gills. The two first gills originated simultaneously as far as known. They be-
gan as tubular disconnected bags, as described; by the concrescence of lateral fleshy pro-
cesses and of the inwardly grown recurved tips of the filaments, their present form, which
is diagrammatically shown in Pl. xxry, fig. 8, has been built up. But this accounts for
only two gills and there are four in the adult; the development of the last two gills has
not been previously studied, though Ryder (63) states that in very young spat only two
are present. When first studied on the sixth of August, the young oyster, Pl. xxry,

e. 4, besides the two larger gills, A and A’, had a third very short gill, B’,on the right
side; but no corresponding, fourth short gill on the left side. The gill B’ was continu-
ous along the margin of the adjoining gill A’ for its whole extent. It originated from
the same base as did A’ for the outer border of A’ was first formed, the inner border be-

306 ROBERT TRACY JACKSON ON THE

ing the reflected portion and of later growth. This will be better understood by com-
paring the diagrams of the gills of Ostrea, figs. 6-9, with those of Anomia, fig. 11, and
Pecten, Pl. xxtv, fig. 12, where the origin and relation of separate gills and gill folds
are perfectly clear. The newly formed gill B' was composed of the same number of
filaments as was the well developed gill A’. In dissecting similar specimens it was
found that the tips of the filaments of the gill B’ were already folded back upon them-
selves, but they are reflected outwardly instead of reflected inwardly (toward the median
line of the body) as is the case with the filaments of the first formed gills (again com-
pare with the diagram of gills of Pecten and Anomia). After an interval of two days,
a fourth gill, B, shown diagrammatically in Pl. xxrv, fig 8, appeared on the left side on
the basal margin of A. It began as a minute bud-like process and grew as already
described for gill Bb’.

In other specimens up to 4.75 mm. in height, a third gill, B’, had begun to develop on
the right upper side as in Pl. xxtv, fig. 4, but no fourth gill had yet appeared on the left
lower side. Oysters of 1 em. in height had all four gills developed, but in dispropor-
tionate degrees. The right outermost was short, the right innermost long, the left in-
nermost longer still, and the left outermost (the last formed) very short, as in the
diagram, Pl. xxtv, fig. 8.

First, it is noteworthy that those gills, which are to form the right and left outermost
gills of the adult, appear much later than the two innermost gills. Secondly, it is a re-
markable and interesting fact that a considerable time elapses between the origination of
the right outermost gill B’ and the left outermost gill B. Stoliczka, in his “Cretaceous
Pelecypoda,” on page 2, says, “I have examined one or two animals of almost every fam-
ily of the Pelecypoda, and in all I found that there are at least two gill-lamelle present
on either side,—never a single one, though they are often unequal and one of them some-
times, as in the Tellinidee, nearly obsolete.”! The late appearance of the third and fourth
gills in Ostrea may have phylogenetic significance, but the irregularity of their appearance
is probably due to other causes, perhaps due in some way to the pressure exerted on the
parts as the oyster lies on the left side in its adopted habit of fixation.

Viewing our specimen from the left side through the glass to which attached, it was
found that the dorsal margins of the several gill-lamellee were not superimposed; the
united borders of gill B’ and the outer lamella of gill A’ were near the middle line of
the valves. The reflected borders of A’ and A united to form a margin nearer the ven-
tral edge of the shell than the previous gill margin. The outer lamella of A and the new
gill B, when formed, made a united border still nearer the edge of the shell. The result of
this want of juxtaposition was that, when viewed from the right side, fig. 4, Pl. xxrv, the
gill A looked much larger than A’, because it extended nearer to the margin of the
shell. In degree this was untrue, though as a matter of fact A is somewhat larger
than A’. Dr. Pelsener, in his report on Lamellibranchiata in the Challenger Reports,
gives a very clear figure showing the condition and relation to one another of the gill
filaments in a large number of genera.

The course of the alimentary tract has been traced in the young oyster by Ryder (63),

1 According to Dall (16), a single gill-lamella, composed briefly discusses the relations of the gills and classification
of separated filaments, exists on either side in Dimya. He on this basis in Bull. Mus. Comp. Zool., Vol. xvut, 1889.

PHYLOGENY OF THE PELECYPODA. 307

but it may not be amiss to repeat the observations as my specimens, growing on transparent
glass, were very clear. In Pl. xxtvy, fig. 4, the dotted outlines of the alimentary canal are
those portions not visible from the right side, but seen when the oyster was viewed from
the left side through the glass microscope slide to which it was attached. From the
upper part of the hood-shaped palps a short cesophagus, @, leads directly into the stom-
ach, st. The stomach, embedded in greenish-brown liver, has a circuitous outline espe-
cially on its ventral border. The intestine leads from the posterior end of the stomach
and extends in a nearly direct line to a point half way past the adductor muscle. It then
takes a sharp bend and, returning on itself, curves and follows the dorsal outline of the
body; again turning when near the umbos, it follows the outline of the body and passes
to the left of the eesophagus. It is then produced posteriorly in a curving line, passing to
the left at the two points where it crosses the earlier portion of its own track and emerges
from the liver mass just above the heart. It passes outside of the heart and wrapping
itself around the adductor muscle terminates in close relation to that muscle. This
course of the alimentary canal is identical with that figured by Ryder (60), in an adult,
excepting that in the adult the cesophagus is directed more anteriorly than in the spat.
The fzeces, of a peculiar, nearly horse-shoe shape,' pass from the intestine in small par-
ticles and are carried from the cloacal space by the excurrent water from the gills
through the opening, s, Pl. xxrv, fig. 4, formed by a sinuosity of the border of the right
valve. All water from the body passes out through this opening, as indicated by arrows,
and the mantle is retracted to allow of its passage.

A special excurrent passage is characteristic of younger and older spat than the one
described (PI. xxv, figs. 5-6); but no trace of it has been found in the adult. The sinu-
osity of the right valve is apparently due to the constantly retracted condition of the
right mantle border at this point. It is constantly retracted to admit of the passage of
the excurrent water and being so retracted, as a necessary result, the shell is not built
at this area as rapidly as at other places where the mantle is fully extended. It might
be considered that the peculiar excurrent opening was due to the mechanical conditions
under which the oyster lives, an opening being necessary on account of the close rela-
tion of the shell to the object of fixation, and it is possible that this is the case; but
againstit may be urged the fact that no such character is found in Anomia which is sub-
ject to similar conditions. I think that this excurrent passage and the correlated sinu-
osity of the right valve are of hereditary significance. Examining the fossils, we find asim-
ilar sinuosity of the left and right valve characteristic in marked degree of the genus
Grypheea; it is also found frequently in Exogyra and in many fossil members of the Avic-
ulidee, from which last group the Ostreadz were doubtless evolved.”

As in the case presented, many peculiarities of the form of shells can doubtless be
elucidated by a study of the soft parts and life-habits of the animals. The byssal notch,
ridges of large siphonated forms as the Myide, and siphonate processes of Gasteropods
are examples of such characters, many of which are specially studied in this paper.

1The horse-shoe shape is the form of the cross sec-
tion of the feces and owes its origin to the peculiar form

forms in the line of evolution of the group. I do not
think this militates against what is above stated. Ostrea

of the intestine Ryder (60).
*In section vit, itis stated that Gryphwa and Exogyra are
probably aberrant members of the Ostreadi, not typical

has an excurrent sinuosity of the valves similar to that
found in Gryphea and Exogyra, in all, inherited (if at all)
from far-off ancestral groups.

308 ROBERT TRACY JACKSON ON THE

The mantle borders of PI. xxtv, fig. 4, are free throughout their extent, with the excep-
tion of a small area near the umbos, where the borders of the two folds are confluent.
The margins are thickly set with tentacles which are slightly pigmented with black.
When undisturbed and quietly feeding, the mantle stretches out to the margin of the shell
on allsides and even protrudes considerably beyond it. It moves actively, being constantly
retracted and extended by the radial muscles shown in my figure and described by
Ryder (63). The radial muscles are best seen from the left side when the spat is
viewed through the glass to which attached, as the lower valve is thinner and less pig-
mented than is the upper right valve. The marginal tentacles are constantly in motion,
being continually elongated and retracted. When elongated they are used with a tenta-
tive, feeler-like motion, and if irritated are capable of extreme prolongation. The gills
and palps move frequently and suddenly. The adductor muscle in PI. xxtvy, fig. 4, is
seen to be composed of two parts, as is that of the adult, Pl. xxv, fig. 12, and a condition
of the adductor similar to that of Ostrea is seen in Perna.

During later stages of growth the soft parts rapidly assume the form, structure and
position which they occupy in the adult, Pl. xxv, fig. 12. The mouth parts are revolved
towards the umbos and the anus and adductor are revolved in the opposite direction,
towards the free ends of the valves. The gills increase in complexity by the develop-
ment of cross partitions within the gill chambers, between the contiguous faces of the
direct and reflected portions of each gill lamella. These partitions form walls joining
gill filaments and separated from succeeding walls by eight or ten filaments not so joined.
Their existence gives rise to the folds characteristic of the adult oyster’s gill. Large
water tubes are thus formed, each tube bordered by eight or ten gill filaments on either
side with their network of cross bars.

An epitome will now be given of the changes by which the adductor muscles of the
young oyster become transformed into those of the adult and the relation of these changes
to what exists in other adult Pelecypods is discussed. The embryo first passes through
an early single-muscled stage, fig. 25, p. 300, in which the single adductor is the equiv-
alent of the anterior adductor of dimyarians. This stage, as stated I believe, is a typi-
‘al early stage of all Peleeypod embryos. Later, the oyster has two adductors, Pl. xxiv,
figs. 1-2, entitling it to ancestral kinship with dimyarian Pelecypoda. Finally the ante-
rior adductor is lost, and the remaining posterior muscle is alone retained.

In the embryonic oyster, fig. 23, p. 299, the mouth and anus are closely approximated
on the ventral aspect of the body. Passing to figs. 24 and 25, pp. 299-300, the anterior
adductor muscle has developed, and the anus has revolved dorsally. In Pl. xxrv, figs.
1-2, the posterior adductor has developed on the ventral side of the intestine, which
wraps around that organ as in later stages, and in all adult Peleeypods. In the last fig-
ures the velum, interposed between the anterior adductor muscle and the mouth, prevents
the mouth revolving dorsally as it otherwise might have done. Passing to Pl. xxtv,
fig. 3, the mouth has revolved dorsally similarly to the anus, and lies nearer the umbos
than in previous stages. However, as noted by Ryder (63) in his studies, it still opens
downward and not as directly forward as in the adult. In Pl. xxtv, fig. 4, the mouth
lies still nearer to the umbos of the valves and from this stage to the adult, Pl. xxv,

1 Vide p. 280, and section xy.

PHYLOGENY OF THE PELECYPODA. 309

fiz. 12, where it immediately underlies the umbos, and opens toward them is merely a
series of mechanical stages. Correlatively with the dorsal revolution of the mouth the
anterior adductor muscle has disappeared in the growing oyster.

In the dimyarian stage of the young oyster, Pl. xxrv, figs. 1-2, the relative position of
the axis of the body to the hinge axis of the shell’ resembles that of adult typical dim-
yarian Pelecypoda,’ Pl. xxv, fig. 9, where the hinge is dorsal, free edges of the valves
ventral, the month and anus lying at either end of the equal valves, so that the antero-
posterior axis’ is nearly parallel to the hinge axis. In the adult oyster, Pl. xxv, fig. 12,
this relation is different. The mouth lies close up under the hinge line, marking this as
the anterior end of the shell. The free ends of the valves are at the posterior extrem-
ity, the gills lie on the ventral side, and the intestine on the dorsal.t. The antero-pos-
terior axis passes nearly through the umbos and centre of the valves and at an angle of
about ninety degrees from the hinge axis.

To bring about the adult condition the antero-posterior axis has therefore revolved
through an angle of about 45° from the position it had in the dimyarian stage. In this
revolution the anterior adductor has disappeared and the posterior adductor has become
sub-central, thus occupying the functionally most effectual position. In other monomy-
arians, as Pecten and Anomia, Pl. xxrx, figs. 1-2, the mouth closely underlies the umbos,
which are in the median plane of the shell; thus the relations of the antero-posterior
axis to the shell are similar to those of Ostrea.

Tn heteromyarians, as Mytilus, Modiola, Pl. xxv, fig. 10, Perna, Pl. xxv, fig. 11, and Avi-
cula, a transitional series may be studied in which the relations of the axes are changed
from what exists in the typical dimyarians, such as Mya, Pl. xxv, fig. 9, to what we find
in the single-muscled group, as Ostrea, Pl. xxv, fig. 12.

In the clam, Mya arenaria, fig. 9, the adductor muscles are at either end of the lon-
ger axis of the shell, the mouth lies close behind the anterior adductor, the anus over-
lies the posterior muscle and the umbos lie dorsally. The antero-posterior axis is nearly
parallel to the hinge axis. It is seen that in Modiola, Pl. xxv, fig. 10, the mouth and an-
terior adductor have revolved dorsally, so that they lie nearer to the umbos than in the
clam. Similarly the posterior adductor and anus have revolved in the opposite direction,
and occupy a position much farther removed from the hinge line thanin the clam. The
antero-posterior axis lies at an angle of about 25° from the hinge axis. In Perna ephip-
pium, Pl. xxv, fig. 11, the mouth lies close up under the umbos, but the umbos are not
in the median plane of the shell. No anterior adductor was found, but it probably ex-
ists in closely related species. The posterior adductor and anus are still further re-

rior end of the body has already been pointed out by sev-
eral authors, and Hyatt (28) figures an oyster drawn over
a clam to show their relations. Ryder (63), on page 787,
notes the rotation of nearly ninety degrees that must take

1The hinge axis refers to an ideal line drawn through
the hinge area, and coinciding with the axis of motion of
the valves, as shown in Pl. xxv, figs. 9-12, inclusive.

*I say resembles because the mouth does not closely un-

derlie the anterior adductor muscle, as the velum inter-
venes, and therefore the comparison is not exact, though
to all intents and purposes the comparison is perfectly
correct.

®The antero-posterior axis, as shown in Pl. xxv, figs.
9-12, is considered as passing through the mouth and mid-
dle of the posterior adductor muscle and nearly or quite
coinciding with the termination of the intestine.

‘The fact that the wnbos in the oyster are at the ante-
MEMOIRS BOSTON SOC, NAT. HIST., VOL Iv. 42,

place from my fig. 25, p. 300, after Huxley, to bring about
the adult relation of parts.

°Mr. Purdie, of New Zealand, notes that Mytilus latus has
no anterior adductor, while two otherwise closely related
species MW. edulis, and M. magellanicus, have anterior adduc-
tors. In M. latus the loss of the anterior muscle is ac-
companied by a movement of the one existing muscle
inwards, nearer the central plane of the valves, where its
mechanical action is more effectual.

310 ROBERT TRACY JACKSON ON THE

moved from the hinge line than was the case with Modiola, and the antero-posterior axis
lies at an angle of about 50° from the hinge axis.

A revolution of the axes similar to that traced in oysters, and in the above serial groups,
may be seen in Mulleria lobata, Pl. xxvi, figs. 5-7, a widely separate genus. Accord-
ing to Adams and other authors, it has two muscles when young; but, when fully grown,
one muscle. The anterior muscle has disappeared, and the posterior, which is retained,
lies in the middle lower portion of the shell, in a position closely similar to that of an oys-
ter. This case is particularly interesting, as the change takes place much later than in
the oyster; for the two-muscled stage is not an embryonic one, but is of a much later pe-
riod of development. In Aitheria, a near ally, the two muscles are retained throughout
life, in a partially revolved position. Tridacna is another genus having but a single ad-
ductor muscle in the adult, and that muscle is the posterior. According to Woodward’s
figure the single muscle is in a sub-central position and the posterior retractor of the foot
occupies its common point of insertion close to the dorsal aspect of the adductor. The
mouth and palps have revolved dorsally so that they lie close up under the umbos as in
Ostrea, marking this as the anterior end of the valves, and a large byssus by which the
animal is attached is protruded through a notch in the valves immediately below the
umbos. The antero-posterior axis lies nearly perpendicularly through the umbos and
middle of the valves. The revolved position of the axes of the soft parts in Tridacna
is not accompanied by any marked changes in the form of the shell excepting the posi-
tion of the byssal notch. Allies of 'Tridacna are typical dimyarians and it is doubtless
dimyarian when young.

The theory of the revolution of the axes explains the changed position of the axes
and adductor muscles on purely mechanical grounds. It is found in widely separate
genera of Pelecypoda that the revolving of the axes is accompanied by a reduction and
final loss of the anterior adductor as it moves into a position close to the hinge area
where its mechanical action is slightly if at all effectual in closing the valves. At
the same time the single posterior adductor increases in size as it moves into a sub-cen-
tral position where its action is most effectual. The revolution of the axes in Pelecypods,
relatively to the axis (hinge) of motion of the valves, is therefore seen to be closely cor-
related with the reduction and final loss of the anterior adductor, together with the in-
crease and final exclusive retention of the posterior. This theory substantially as here
presented was published in my preliminary paper. While that paper was in press, a paper
was published by Dr. Benjamin Sharp in which he presents a mechanical theory for the
loss of the anterior and retention of the posterior adductor muscle in Peleeypoda, which
is quite similar to that of the revolution of the axes.

It isa very noticeable fact that oyster spats attached to various foreign bodies do
not lie with their axes arranged in any special direction. Different individuals attached
to the same stone or shell, may have their axes at every angle of deviation, in the plane
of the surface of attachment, as compared with one another. I have never seen them ar-
ranged with any reference to the resistance of prevailing currents. On the other hand,
Anomias and Crepidulas are sensitive to prevailing currents. Large numbers of Anomias,
attached to a single block, will very commonly be found with their axes parallel and
headed one way (see discussion on development of Anomia, section x1), and Crepi-

PHYLOGENY OF THE PELECYPODA. Sila

dulas, Pl. xx, fig. 22, may be found piled on one another to the height of several inches,
each individual with its axes in the same vertical plane. These comparisons go to prove
what was claimed earlier, that calcareous fixation of the oyster’s lower valve took place
when it became attached. It is almost certain thatif any form of prehensile attachment
had existed for an appreciable time, we should at least occasionally find evidence of the
arrangement of the shells in the line of least resistance to currents.

VI. OstreaA. DEVELOPMENT OF THE SHELL.

In the study of the soft parts it is shown that the shell of the oyster develops from a
primitive structure, the preconchylian gland, which is found not only in other genera of
Pelecypoda, but in other classes of mollusca as well. According to Horst, a simple
unpaired-cuticular membrane, my fig. 22, p. 297, the product of the cells of the shell-
gland precedes the paired calcareous shell.

The two valves of an adult oyster are together homologous with the single valve of
adult Cephalous molluses (see p. 281). In both, the shell originated from similar struct-
ures, one as a single and the other as a paired organ. The adult shell of Cephalous mol-
lusca is termed a conch. Therefore, in view of its homology and strikingly different
double character, I have suggested (36) the name dissoconch! for the shell of an adult
oyster and similarly the same terminology is applicable to the shell of other Pelecypoda.

The protoconch of Professor Owen, in Cephalopods, figs. 1-8, p. 292 and Pl. xxm,
figs. 15-16, is the early cup-shaped shell which precedes the conch, or true shell. A spi-
rally twisted protoconch in Gasteropods, Pl. xx, figs. 19, 21, 23, and a saddle-shaped
periconch (protoconch) in Scaphopods, Pl. xxm, figs. 13-14, may be also found as an
embryonic shell preceding the conch or true shell of these groups (see section mr). It
will be seen from the study of Ostrea that the true shell, that which characterizes the
adult, originates with the introduction of the spat stage, Pl. xxrv, figs. 19-21 and PI.
xxv, figs. 1-7. The earlier shell, Pl. xxtv, figs. 17-18, is strikingly different. It is the
completed shell of the embryonic stages, and as shown in previous studies, it has devel-
oped from structures closely similar to those found in Cephalous molluscs. The embryonic
shell of Ostrea, therefore, is distinctly the homologue of the protoconch and periconch.
Tn the oyster, however, this shell is not single- but double-valved and, therefore, deserves
a distinct name. As it precedes the dissoconch or true shell, I have suggested (36) the
name prodissoconch? or early, double shell, for the first-formed shell of the oyster. A
prodissoconch or embryonic shell, strikingly different from the succeeding dissoconch, is
figured in this paper in many widely separate genera of Pelecypoda (see Pls. xxrv to
xxx and cuts in the text) and it may be safely considered as characteristic of the class.

The shell of oysters, at different periods of growth, presents striking dissimilarities.
There are four well-marked stages of growth: (1) phylembryo, early embryonic stage ;
equivalvular, with a straight hinge line; (2) completed prodissoconch, symmetrical fry
stage, with a curved hinge line and high umbos developed; (3) nepionic (spat) stage,

"Atoods, double ; Kozy, Shell. # []oé, before.

2See footnote, p. 292, and discussions in section on
Classification of Stages of Growth and Decline.

312 ROBERT TRACY JACKSON ON THE

with a flat, lower left valve and convex, upper right valve; (4) adult, flat, upper right
valve and concave, lower left valve.

The early prodissoconch or phylembryonic stage of Ostrea edulis is equivalvular, nearly
discoidal and flattened into a straight line at the hinge area. On the hinge line, minute
interlocking teeth are visible when viewed from the dorsal side. Embryonic shells of
Ostrea edulis of this stage are represented in Pl. xxty, figs. 13-16. They were removed
from the gills of the parent and sent out as preparations by a London dealer in micro-
scopical supplies. The age is similar to fig. 25, p. 300. The valves are saucer-shaped,
and are marked by concentric lines of growth closely approximated dorsally, but farther
apart ventrally and anteriorly, so that the centres of calcification are already pushed pos
teriorly by the growth that has taken place, as seen in figs. 13 and 15, viewed from the
left and right sides respectively. By polarized light the shell gives the characteristic
lines of crystallized carbonate of lime. A straight hinge line seems to be characteristic of
the phylembryonic stage of Pelecypoda before the development of umbos (see section
xv). As examples, I would refer to embryos of Ostrea, Pl. xx1v, figs. 13-16, Anodon,
Cardium, Modiolaria and Montacuta, Pl. xx, figs. 7-11, also to Hatschek’s figs. 19-21 of
developing Teredo and Ryder’s (62) figure of an embryonic Ostrea virginiana.

The valves of the phylembryonic stage of Ostrea edulis, Pl. xxtv, figs. 13-16, have
scarcely any umbos developed, whereas our next stage of shell development has promi-
nent umbos. The completed prodissoconch of Ostrea virginiana, Pl. xxiv, figs. 17-18,
has deeply concave valves which are nearly equal. The lower left valve is, however,
somewhat larger and deeper than the upper right valve, fig. 18. The umbos are highly
developed and are always inclined upward at a high angle, as the shell is invariably at-
tached by the extreme border of the left valve. The umbos are directed posteriorly as
shown in the studies of the soft parts, Pl. xxrv, figs. 1-2, and they invariably point to
the left of the observer viewing them from above, a necessary consequence of their be-
ing attached by the left valve.’’ (For further discussion of the fixation of the oyster by
the left valve, see section tv.) Coincident with the development of high umbos, the
early straight hinge line, seen in the phylembryonic stage, is lost and is superseded by
one of a gentle curvature. The prodissoconch is the completed shell of the embryonic
free-swimming veliger period, and the young oyster becomes attached
permanently at the close of this period, the fixation consisting of the
soldering of the ventral margin of the left valve to the object of sup-
port.

The completed prodissoconch of Ostrea edulis, fig. 26, differs from
= S that of our species, according to Horst. The umbos are less devel-

Fig. 26. Completed oped than in our species, but they are directed posteriorly. Teeth,
prodissoconch of Ostrea a]so, are mentioned on the hinge line at this stage and earlier by both
edulis, attached by bor- ‘ i k
der of left valve in an Horst and Lacaze-Duthiers, whereas in our species no teeth have
almost upright position een observed by either Ryder (63) or myself. The histological
CaP ea structure of the shell, as described by Huxley and Horst, is similar
to that of our species. Specimens of Ostrea edulis kindly sent to me by Dr. Horst show
unmistakably the teeth and other features described.

1 Professor Ryder informs me by letter that he has sought for umbos pointing to the right.
never seen an exception to this nor have I, though I have

PHYLOGENY OF THE PELECYPODA. 313

The prodissoconch of Ostrea virginiana is homogeneous and laminar in arrangement;
“not prismatic” as in the immediately succeeding stages of spat growth as described and
figured by Ryder (62). It is composed of lime infiltrating an amorphous matrix of con-
chyolin, as shown by treatment with acid. One of the marked characters of the prodis-
soconch is the uniformity of shape and size found in different individuals, it not having the
ostrean tendency to variability noticeable in later stages. The height of the fully de-
veloped prodissoconch is about 545 of an inch. ‘The left attached valve turned over and
viewed from the lower side is seen to be exactly like the right free valve in appearance.
It is not flattened nor are there other indications to mark where it was attached, so that
this attachment is evidently of a very superficial and delicate nature.

A prodissoconch identical, or very closely similar, to that found in Ostrea virginiana has
been found in many species of the Ostreadz. Ryder (62) has observed it in four species:
Ostrea equestris, Say, from Florida; O. conchaphila, Calitornia;. O. cucullata, Viti Isl-
ands, Pacific; and O. peruviana, Peru. I have observed it in undetermined oyster
spats, from Cuba, and from the chamber of a Nautilus pompilius, from the Pacific. I
also found ostrean prodissoconchs in the following: the spats of tertiary oysters from
Carolina; Gryphea vesicularis, Lam., Cretaceous, found in both Kuropean and American
specimens; G’. calceola, var. nebrascensis, M. & H., Pl. xxtv, figs. 23-24, from the Juras-
sic of Wyoming Territory, specimen number, 1881, U. 8. National Museum; Hxogyra
conica, Cretaceous, Europe; #7. columba, Lam., Cretaceous, Europe; 2. arietina, Creta-
ceous, America; and /H. costata, Say, Cretaceous, America. Of this last species abun-
dant very young specimens, which I collected at Delaware City, Delaware, showed the
typical early prodissoconch.

From the number and wide distribution of the species, it cannot be doubted that the
prodissoconch found in Ostrea virginiana is characteristic of the Ostreade. The teeth
noted above in the prodissoconch of O. edulis is the most striking difference that has been
found in the prodissoconch of the family.

There is, in the prodissoconch and spat stages of the oyster, an organic conchyolin at-
tachment of the shell itself, thus described by Ryder (62): “The cementing material
seems to be the organic matrix of the shell which forms a perceptible layer on the out-
side of the valves, and which constitutes the epidermis or periostracum of the oyster.”
In the prodissoconch, as well as spat stage, the cement is so firm that, as Ryder says
(62), the shell may be broken before it can be removed. To prove that this is an organic
cement, if we put a drop of water on a dead, young shell and allow it to remain a short
time, it can be readily picked up on the point of a knife. Quite large spat may be re-
moved when dead (7. e., the organic matter inert) by soaking or boiling. Treatment
with acid, on the other hand, does not loosen the hold of the shell. Quantities of young
spat die during the summer, yet few dead shells are found on cultch, because soon after
death the organic cement decays, is dissolved, and the shell falls off. Professor Huxley
notes that the Huropean oyster separates naturally from unfavorable objects of support;
and that the separation takes place naturally at an early stage can be seen markedly in
many fossil members of the family. Conversely, the calcareous plug of Anomia, which is
fixed to the object of support by a purely calcareous union, remains indefinitely after the
death and separation of the individual to which it belonged. All the evidence is there-

314 ROBERT TRACY JACKSON ON THE

fore in favor of the assumption that the cement by which the oyster is attached is wholly
organic.

With the introduction of the spat stage, a fundamental histological difference arises
abruptly in the structure of the shell of Ostrea virginiana as in O. edulis, fig. 27. As
stated above, the prodissoconch consists of lime infiltrating an amorphous matrix of con-
chyolin, but in the initial spat stage the first layer of shell of the right valve is deposited in a
“tessellated or prismatic” manner in a horny matrix, as described and figured by Professor
Ryder (62, 64). Soon after the formation of the spat stage, the subnacreous, white por-
cellanous layer begins to be deposited on the yellowish-brown, prismatic layer, first mak-
ing its appearance as irregular blotches in the centre of the valves. These two layers con-
tinue throughout the rest of the life of the oyster, and together, though in disproportionate
degree, build up the massive adult shell. The prismatic layer always remains thin, as may
be readily seen by sections, or by treating an adult with acid, when the remaining conchy-
olin of the prismatic layer will separate from the overlying subnacreous layer like a dis-
sected skin.

It is a remarkable fact that in
young Ostrea virginiana, the
prismatic structure is typical of
the upper right valve alone; the
lower valve presenting a fine
granular structure, as in O. edu-
lis, fig. 29. In later life in Ostrea
virginiana and some other spe-
cies, a very thin layer of pris-
matic tissue may be found in the
left valve though it is for the
most part eroded away, but the
same layer is thicker and there-

Fig. 27. Young Ostrea edulis. The prodissoconch is composed of ho- fore better preserved in the right

ae » early dissocone ris jc cellular tissue.' The a 3
mogeneous and the early dissoconch of prismatic cellular tissue The valve. The prisms ae the right
gills are filamentous.

Fic. 28. Fragment of border of upper valve, enlarged to show the valve of the adult oyster are not

prismatic structure. as perpendicular to the surface
Fig. 29. Fragment of lower valve (after Horst). pager lea young, fic. o7. "There-
fore in horizontal section the outline of the prisms presents oblique rather than hexag-
onal faces. A similar obliquity of the prisms may be seen in Solemya mediterranea.
The nepionic or infant stages of molluscs begin according to the classification of
stages adopted with the formation of what Professor Owen called the apex of the conch,
or true shell. It is shown that the spat marks the beginning of the dissoconch (true
shell) in Ostrea; therefore the spat is considered as the nepionic period.’ In my pre-
liminary paper I said the spat of Ostrea virginiana is divisible into five stages which were
described at some length. I also said that they are more or less clearly marked in every

1The relations of structure of the prodissoconch and dis- ered as the ‘‘silphologic” period. In a footnote to this
soconch are the same in Ostrea virginiana, also in Perna, paper Professor Hyatt has given up the use of the term
Pecten and Anomia, as described in the text. silphologic, and has supplied in place of it the term nepi-

2In my preliminary paper this same period was consid- onic (vide p. 290).

PHYLOGENY OF THE PELECYPODA. 315

regular growing, unworn, young specimen. There are often well defined stages indi-
cated as arrests of growth in the early growth of the spat; but I must retract the state-
ment that there are typically five and that they are always more or less clearly marked in
regular, unworn, young specimens. Frequently no indication of later spat stages are
present, and they cannot be considered as constant. T’o remodel my conclusions and
statement: The spat stage in Ostrea virginiana, as a whole, is frequently marked by
stages of growth. Stages are characteristic of the earliest spat growth and as many as
five have been found, but often no more than two or three are indicated. To study these
stages specimens should be sought growing in protected places, as the wear of sand
and waves destroys their demarcation rapidly. Some spat found in the inside whorls of
dead Sycotopus shells and others, secured on glass in drain pipes and on inverted flower
pots, retained the markings clearly.

The spat growth begins along the ventral margin of the prodissoconch valves
Pl. xxty, figs. 18-21 and Pl. xxv, figs. 3-4, and throughout life the most rapid growth
is in this direction. The growth rapidly pushes up the sides of the valves, and between
the hinge areas of the latter, Pl. xxv, fig. 3, lifting them apart and separating them by
interstitial growth. <A little to the right of the umbo of the lower, left prodissoconch
valve, as seen from above, in the initial stages of spat growth, is a notch which indicates
the beginning of the cartilage pit of the lower valve. This is seen clearly in PI. xxtv, fig. 19.
It is necessary to describe the left and right valves separately. The upper, right valve,
Pl. xxry, fig. 20, is continuous with the prodissoconch valve and follows the curve of that
stage though in a lessening degree. It spreads out laterally on the hinge line in a de-
scending curve, and on the margin overhangs the lower valve by a considerable amount,
Pl. xxrvy, fig. 21, as it continues to do while the shell is flatly related to the object of
fixation. The same overhanging of the upper valve may be seen in young Pectens,
Pl. xxvin, and Anomias, Pl. xxrx. The lower left valve of the young oyster, Pl. xxrv,
fig. 21, starts as does the upper, following the curve of the early valve. When this curve
has been followed for an extremely brief period, the valve suddenly flattens and becomes
closely related to the surface of attachment. The result is, that a slight groove runs
around the border of the spat valve, beyond which it is abruptly continuous with the flat
later growth as mentioned by Ryder. I have, however, to note a peculiar and significant
exception. Sometimes, instead of becoming flat and closely related to the object of at-
tachment, the spat valve is continuously curved throughout its early growth, and not
at all flattened. It then resembles the upper valve in shape. This curved shape is
held, as far as observed, only on rough objects of attachment. Seeking for such mate-
rial on rough stones and shells, one will always find some valves more or less curved in
their earliest stages. This is interesting, as showing the modifying influences of enyi-
ronment on a very young stage, where heredity might be supposed to fashion the form
with overpowering force. Further notes on this character are given when discussing
Exogyra, section vil.

When the spat shell has attained the height of 0.8 mm. to 1.2 mm., frequently a well
defined stage or period of growth is marked by an interruption in the lines of growth as
shown in Pl. xxv, figs. 1-2. This indicates what is considered as the close of the first
nepionic stage. In the shell growth succeeding the first nepionic stage the upper valve,

316 ROBERT TRACY JACKSON ON THE

fig. 1, is in form a continuation of the first stage. The lower valve spreads out over the
object of attachment by wing-like extensions of its anterior margins, and makes a shoulder
on its anterior border by its increased growth, where it comes in contact with the outer
limit of the first stage, fig. 2, and similar shoulders mark the periods of growth of later
stages in the lower valve, Pl. xxv, fig. 6. The lower valve may remain curved, as in the
first nepionic stage, fora limited extent. When the spat has attained the height of from
about 1.4 mm. to 2.4 mm., another stage is frequently marked by an interruption in the shell
growth as shown in Pl. xxv, figs. 1-2, which illustrate the first and second stages of the
nepionic or spat growth. After the second stage, as described, other stages may be found
in the spat growth. They are marked as arrests in the regularity of shell growth and in
the lower valve shoulders are formed by their contact with earlier stages as shown on
the hinge border in PI. xxv, fig. 2. In all, five stages have been observed; but the num-
ber and the later stages cannot be considered as characteristic as the first two stages are.

After the close of the second spat stage the shell is still even and rounded in outline,
as the convolutions and distortions common in oysters do not often begin till growth has
considerably advanced. The lower valve spreads out over the object of attachment, and
leaf-like layers of shell deposition may be seen spreading beyond the limits of the hinge
line, as growth continues, as shown in Pl. xxrv, fig. 19, and Pl. xxv, fig. 5. The upper
valve in regular growing specimens frequently departs from the form which is ob-
served in the earlier growth, Pl. xxtv, fig. 20, and spreading out laterally becomes more
wing-like on the margin than in any earlier stages, Pl. xxv, figs. 5 and 7. These wing-
like productions of the spat growth were pointed out and figured by Ryder (60). In
describing such stages he says: they are“ . . . . handsomely rounded in shape, not
unlike a scallop in form, which they simulate still more by the lobes which they develop
on the shell on either side of the hinge.” This scallop-like form, shown in Pl. xxv,
fig. 7, is quite likely due to the irritating action of a foreign body on the extended mantle
border of the left valve, or it may be due to the opportunity to make more secure the
fixation by deposition of lime in this direction. In spat growth the free ends of the
valves often change the direction of their growth to the left or right of the median line.
By this change the umbos may be revolved so that they often point in the opposite di-
rection from that which they would have otherwise assumed, as noted by Ryder (65).
A twist in the direction of growth as described is characteristic of the fossil genera, Ex-
ogyra and Gryphzea, in which the variation is always to the left of the median line viewed
from above. The young oyster when it twists to the left resembles very much a young
Exogyra; but when the twist occurs in the opposite direction, or to the right, as fre-
quently occurs, this close comparison cannot be made. The right valve of the spat has
a sinuosity, sz, Pl. xxv, fig. 5, which is related to the excurrent flow of effete water as
described, p. 307.

Throughout the period of spat growth the young oyster shell has an approach to the
rotundity characteristic of the early spat stages. The attached valve is flat and the free
convex, but when the oyster projects beyond the object of fixation this relative form of
the valves is changed and the condition characteristic of the adult, Pl. xxv, fig. 8, is
brought about. The attached valve becomes deeply concave or spoon-shaped and the
free yalye conversely becomes flat, as pointed out by Brooks and Ryder. At the close

PHYLOGENY OF THE PELECYPODA. Sli7/

of the spat period if the shell does not project freely from the object of fixation, then the
relative shape of the valves characteristic of the spat is retained, the attached valve still
spreading over the support, is flat, and the free valve is convex. In such forms the true
spat period is marked as a smooth rounded convex area at the umbos.

In young Ostrea virginiana spat, as noted by Ryder (60), bands of purple color radi-
ate from the hinge line, widening toward the free edges of the valves; but these colora-
tion bands are in the prismatic layer and are therefore limited to the upper right valve.
In none of the oysters growing on glass, have I seen color bands in the lower valve,
which is nearly or quite white; in later growth, however, faint traces of color bands may
be occasionally seen in the left valve. I found spat in the living chamber of a Nautilus
Pompilius, when unpacked on arrival from Singapore, and these are similar to the spat
of O. virginiana, showing similar prodissoconchs and nepionic stages, but no coloration
bands, the upper valve being yellowish-white. The early spat growth of Ostrea edulis,
fig. 27, p. 314, is very similar in form to that of O. virginiana, Pl. xxrv, fig. 20, but I
have not had specimens for a comparison of later stages.

VII. Gryeuma, Exoayra AND AdULT OsTREA, WITH A CONSIDERATION OF THE
MECHANICAL ORIGIN OF THE OsTREAN Form OF SHELL.

The genus Gryphea is described as free or attached in the young, yet I have not
been able to find a species in which it is free at'a very young stage. All the species
studied have a flat area in the lower valve and a curved area in the upper valve at the
umbos, corresponding to the condition found in the attached spat stage of the oyster.
G. arcuata of the Lias is commonly considered as free, yet [ have found well preserved
specimens, P]. xxrv, fig. 22, in which the lower valve is flat at the apex, where it was
attached when young, and correspondingly the upper valve has a small convex area at the
umbo, as noted by Sowerby. Flat areas of attachment have also been found at the umbos
of two other highly arcuate species, namely, G'. calceola, var. nebrascensis, M. & H., PI.
XXxrv, figs. 23-24, from the Jurassic of this country, and G'. pitchert, Morton, from the
Cretaceous of Texas.

The area of attachment in many species of Gryphea is uniform in size, indicating that
the shell dehisced at a very definite period from the object of attachment. In fact, highly
arcuate forms must have separated early, as the involute form could not have been de-
veloped unless the shell was free. Other species were attached for a very variable period,
as is seen by the differing size of the flat area of the lower valve and corresponding
curved area of the upper. This is well shown in any large collection of the European
form of Gryphea vesicularis from the Cretaceous, where the arcuateness of the adult
varies inversely as the length of period of attachment.

Exogyra, also, is said to have been free or attached; but here the case is somewhat
more complex. Take, for example, the typical 2. costata, Say, of the Cretaceous. Spec-
imens are found in which a flat area at the apex of lower valve distinctly shows them to
have been attached when young, but rarely are adults found with the object of early fixa-
tion still clinging to them, showing that as in Gryphzea and many oysters, they separated
naturally. Other specimens are found which, upon most careful examination, show no

MEMOIRS BOSTON SOC. NAT. HIST., VOL. IV. 15

318 ROBERT TRACY JACKSON ON THE

flattened area on the lower valve. I have noted that Ostrea may in its early nepionic
period be attached, and yet not conform to the object, or- become flattened at all, so thatif
it could have separated at the close of this period, the adult would show no sign of early
fixation. This is what apparently took place in specimens of Exogyra which show no
flat area. In such there is no proof that they were not fixed during the young stages,
but merely that they dehisced before the shell conformed to the outline of the object of
support. Some individuals may be attached for a considerable extent, and a specimen in
the American Museum in New York has an area of attachment measuring 11 em. in its
greatest length. Thisspecies is also very variable in the degree of costation. Aged speci-
mens, as noted and figured by Whitfield, often lose the ridges and become quite smooth.
I have seen many specimens in which the lower valve was first smooth, then costate, and
later smooth again, or the reverse of this order. In a large collection one may find every
gradation between the highly costate type, known as Haogyra costata, and the smooth
type without costations known as #. ponderosa, Roemer. ‘The relations of these spec-
ies is about the same as that which exists between Ostrea virginiana, Lister, and O.
borealis, Lam., which Verrill unites under the former species, as he says they may easily
be connected by all sorts of intermediate forms.

In the Ostreadz there is a striking peculiarity-of the adult which has not been con-
sidered in its full significance, so far as Tam aware. The two valves are unequal, one
being concave and the other flat; but they are not only unequal, they are very dissimilar,
as different as if they belonged to distinct species in what would be considered typical
Jorms. This character of the shell is due, I believe, to the mechanical conditions of di-
rect cemented fixation acting upon a Pelecypod shell, and is of great value in consider-
ing the systematic relations of attached groupsas I shall attempt to show in the follow-
ing pages.

As examples amongst the Ostreade, showing dissimilarity of the two valves, we may
take Ostrea edulis of Europe, O. compressirostris, from the Hocene of this country, and
Exogyra costata, from the Cretaceous. In these species the lower valve is plicated,
and the upper smooth, and regular-growing, without plications. The examples could be
multiplied extensively, but these suffice to give my meaning. In consulting Coquand’s
finely illustrated monograph of Cretaceous oysters, examples of this difference of the two
valves will be found in many species. This is a highly interesting feature, as it is a case
of inherited or acquired characteristics. finding very different expression in the twovalves
of a group belonging to a class typically equivalculur. As might be expected, variations
from the rule are occasional as in Ostrea marshii, O. larva, ete., where the two valves are
more closely alike. Dissimilarity is nevertheless the rule. The upper valve may bear
characteristic features not found in the lower valve. As an example we may mention
Gryphea vesicularis of the Cretaceous, in which the upper valve in both European and
American specimens is marked by radiating depressed lines. Similar radiating lines -
may be found in G. dilatata from the Jurassic, and indications of them have been ob-
served in other species. These lines are a feature which would probably be of value in
tracing the serial connections of the several species possessing them.

The differenees in the two valves may be observed in species of Spondyli, especially
those in which the right, attachcd valve is very ostreaform, while the free, left valve is

PHYLOGENY OF THE PELECYPODA. 319

less modified, as shown in Spondylus hippuritarum, VOrb., Pl. xxvi, fig. 1. We would
also refer the reader to M. @’Orbigny’s plates £50 to 461 of Cretaceous Spondyli, es-
pecial attention being called to his figures of S. histrix, Goldf., in which the attached
valve is rough and ostreaform, whereas the free valve is evenly plicated and smoother.
Striking differences in the features of the two valves may be seen in the Rudiste as
shown in Radiolites fleuriausa, VOrb., Pl. xxvt, figs. 11-12, in which the lower at-
tached valve is irregularly cornucopia-shaped, whereas the upper free valve is compara-~
tively regular growing, flattish and plicated. The young of this species and several
species of Hippurites studied resemble an Exogyra (compare the fizures cited with the
adult and young of Caprotina semistriata, Pl. xxvt, figs. 10 and 20). Because the young
of these aberrant forms resemble an Exogyra, I do not wish to imply any genetic connec-
tion with that group. Istate later, p. 322, the conviction that the ostrean form is in-
duced by the condition of fixation, and the similarity to an Exogyra in the young of the
Rudistz and Chamidz is believed to be due to the similar conditions of a fixed cemented
habit common to both groups, and not as expressing anyaffinities. This is a point which
must be understood, as in this paper high importance is attached to the form of young
shells in tracing genetic relations.

Aberrant forms of the Chamidee are shells which show in a high degree differences
in the attached and free valves. Caprinula boissyi, VOrb., Pl. xxvt, fig. 8, from the
Cretaceous of Europe is a remarkable form; the attached valve is deeply conical with
perpendicular ruge ,whereas the free valve is spirally twisted, without plications and re-
sembles a typical Diceras. Caprotina semistriata, VOrb., Pl. xxvt, fig. 10, is another
form not less remarkable in the striking dissimilarity of the two valves. In this species the
attached valve is arcuately conical with perpendicular rugee and showing at its base a
spirally twisted tip as in Radiolites, Pl. xxv1, fig. 12. The free valve is of a simple con-
cave form, not at all aberrant and not presenting the rug seen in the attached valve.
A very young specimen of Caprotina semistriata, Pl. xxv1, fig. 20, was found attached
to the base of a colony of this species in a specimen in the Museum of Comparative
Zoblogy. This young Caprotina resembles an Exogyra, taking for comparison a young
Ef. costata or an adult . spiralis, Goldf. The attached valve is highly twisted and the
free valve convex showing little twisting as in Hxogyras. The specimen was too much
worn to show lines of growth, but in the figure they are inserted where I think they
would appear if not eroded, in order to render the form and its relations to the adult
clearer. That they are fairly represented will be evident upon a comparison with similarly
twisted shells where they arepreserved. As the attached valve in Caprotina, Pl. xxvt,
fig. 20, grows from its early twisted condition to the deeply conical condition found in
the adult, Pl. xxvz, fig. 10, the lines of growth gradually and progressively assume a
more and more horizontal position until in the adult they are nearly or quite horizontal
and the attached valve has lost the twisted condition characteristic of the younger stages.
The outgrowth from an early twisted condition of the attached valve to an adult deeply
conical condition, as just described, doubtless took place in the development of Caprinula
boissyt, Pl. xxv1, fig. 8, and the change may be seen in Monopleura marcida, White, PI.
xxvI, fig. 9. In Dr. White’s paper (74) the same feature and the modification of at-
tached shells are very well shown in many species.

Our examples of differences in the two valves amongst attached Pelecypods might

320 ROBERT TRACY JACKSON ON THE

be carried further and for such inquiry we would refer to published works on conchology
and paleontology. Believing that the inequivalvularity of the above genera is due to
the conditions of environment acting on a bivalve shell, I would refer to species of at-
tached Brachiopods in which is found further proof of the correctness of the conclusion.
Davidsonella rugosa, Pl. xxvt, figs. 14-15, the Cretaceous, is soldered to an object of
support by the lower valve, and here, as in the above cases of Pelecypods, the attached
valve, which is rugose and concaye, is strikingly different from the upper valve which is
flattish and smooth. The same feature is shown in species of Strophalosia recently dis-
covered and shown me by Dr. Beecher. Productus proboscideus of the Carboniferous
has a tubular highly modified attached valve and a nearly normal free valve’.

In the various genera considered, the free valve is that which is least modified and
nearest resembles the form existing in both valves of less aberrant related genera. It is the
attached, cemented valve, whichis highly modified and so much so as by itself in some
cases to be with difficulty recognizable as belonging to the Pelecypods.

In the shell of Ostrea virginiana, Pl. xxvt, fig. 13, there are lenticular cavities in the
hard subnacreous portion which are filled with a soft, white, chalky deposit. These cay-
ities are also noticeable in many fossil oysters and Gryphzas. In Exogyras they are com-
monly very large and abundant, especially near the umbos where shell growth takes
place rapidly and to a great thickness. xogyra costata shows these cavities well, and
they are highly developed in Hxogyra aquila, Goldf. and E. couloni, VOrb. from the
Cretaceous of Europe. Not infrequently the cavities in living and fossil shells will be
found filled with mud, and in such cases it is evident that the shell roofing was formed
to include the mud; being caused by the irritating action of that substance on the over-
lying mantle. Specimens of Mya arenaria, which have included mud in this fashion
are frequent, although this shell never builds true camerated structure. The shallow
chambered structure of the oyster was noticed by Professor Owen (50, 51) and was
compared by him to similar conditions in the shell of Spondylus, Autheria and some
univalves. Professor Woodward (77) compared the structure of the oyster also with
camerated structure in species of Teredo and Caprinella which he figures. This last
genus is remarkable for the regularity and size of its chambers which truly resemble
analogous structures in Cephalopods. Owen (50) notes the fact that Spondylus and
other bivalves, which have camerated structure, are cemented to extraneous bodies by the
shell. I have not made an extensive study of this subject; but what has been made,
leads to the conclusion, that this statement may be safely reversed and we may say,
that bivalves which are cemented by one valve to extraneous bodies have a tendency to pro-
duce a camerated structure of the shell which is not found in free forms. Thin shelled, ce-
mented Pelecypoda, such as Dimya, quite likely would not show cameration, but amongst
other groups it is general, whereas no free lococomotive forms of Pelecypods, however
thick the shells, have been found to possess such structure.’

In the developing shell of Vermetus radicula, fig. 9,
p. 294, I have observed in well-preserved specimens that
the shell is evenly spiral and regular growing while young
and unattached. As soon as attached, however, the ir-
regular loosely spiral features characteristic of the adult
are introduced and are retained throughout subsequent
growth.

2It is important to notice that fossil corals of the Tet-
racorolla type have a highly developed camerated struct-
ure. Sosimilar is theirappearance in their regular cham-
bering that they may often be compared to chambered
Cephalopods especially of the Cyrtoceras type which they
simulate by their arcuate form. They also bear often a
striking resemblance both in external and internal form to

PHYLOGENY OF THE PELECYPODA. 321

In adult Ostree virginiana, the left valve is concave, and the right flat, when grow-
ing naturally, the left valve being lowermost; and Ryder notes that the same relative
shape is held by oysters which grow vertically in crowded beds. Oysters frequently
attach themselves to the lower side of objects, so that the left valve is uppermost, and
right valve lowermost, the reverse of the common position, Pl. xxv, fig. 8, and here
again the normal relative shape of the valves is maintained, the left valye concave, the
right flat. The shell of the oyster is so variable that we should expect it to be especially
susceptible to varying conditions of environment; yet these strikingly different positions
produce no change in the resulting form. Whatever the cause of the typical form of
the group, the changed relations brought about by inversion, or other positions assumed
by the growing valves, are insufficient to modify greatly the form of a modern individ-
ual, where hereditary or other influences bind the shell within the limits of its typical
form.

Professor Hyatt (29) has published the view that the attached supported valve of fixed
bivalve shells is the larger, the element of gravity being the active factor according to
his ideas, and he ascribes the form of the oyster and similar shells to such conditions of
environment. My conclusions on the form of attached Pelecypod shells, although some-
what similar, differ from those of Professor Hyatt. The conclusions, as presented in
these pages, are that the form of fixed Pelecypoda is due to the conditon of cemented fix-
ation; whereas his assume that the form of mequivalvular bivalves is due to support,
the shell being free or not as the case may be. The two views are therefore based on
different grounds of consideration, which should be borne in mind in a comparison of
our studies. :

The family of the Ostreadz is characterized throughout by being attached, at least
when young, by the left valve, as shown in PI. xxv, fig. 8, and Pl. xxv1, fig. 13. The
left attached valve is typically, deeply concave, of rough, irregular growth and com-
paratively thick. The right free valve, on the other hand, is flat or flattish, thinner,
and less irregular. To perceive the truth of this statement it is necessary to look over
a collection of the Ostreadze, or in default of it some of the illustrated monographs, as
Coquand’s or White’s. It will then be seen that all typical forms are referable to this
description in a more or less marked degree. I say more or less marked, because some
species, especially among the Alectryonia division, have right (free) valves which are
comparatively like the left valves; but on careful study one finds that it is only a matter
of degree and that they are actually less concave and generally less irregular than are
the left attached valves.

The systematic position of the family to which the oyster belongs has long been a per-
plexing problem to the naturalist; the asymmetry and irregularity of the valves has
rendered them a group difficult to characterize and of uncertain systematic position.
The late Mr. Tryon in his work on Conchology says: “'Uhe union of the Ostreidee

the attached valve of Hippurites. Culceola sandalina, an eration strongly recalling the similar structure induced in
operculated coral, has a deep thick attached valve and Pelecypods by the condition of cemented fixation. Wood-
flattish thinner free valve; bearing thus considerable ward (77) suggested that septa and other forms of cam-
analogy to the effects produced in bivalve molluses by eration in Molluses are induced by the mechanical necessity
cementation. Beatricea, a fossil of uncertain affinities, of cutting off the rear unoccupied portions of the chamber
has well-marked camerated structure. In all these at- of a shell as the animal is urged forward by the growth

tached forms there is an irregularity of growth and cam- taking place on the free edges of the living chamber.

o22, ROBERT TRACY JACKSON ON THE

and Pectinidse, as proposed by the authors of the ‘History of British Mollusca,’ has not
proved satisfactory. The genus Ostrea stands quite alone, and distinct from all the Pec-
tinide in the structure of its gills, which are like those of Avicula, and by resting on
its left valve. The shell also is more nacreous than that of the scallops.” Another
leading naturalist, Dr. C. A. White, in a recently published monograph of the Ostreade
of North America, says: “To what primary cause this asymmetry among the Ostreidee
is due, it is, with the present limitation of our knowledge, impossible to say; but it is
certainly a characteristic of the whole family, including all its genera and its fossil as
well as living forms.”

My studies of the Ostreadae, and other Pelecypoda as well, have brought me to a con-
clusion in regard to the origin of the peculiar forms we have here to deal with, which
is that the ostrean form is due to the conditions of a direct cemented fixation, acting
upon a Pelecypod shell. Given such conditions, and a closely similar form is the result
in widely separated genera of this class, as I shall attempt to show. The form which is
claimed as the result of the conditions of direct, cemented fixation is a concave attached
valve; a flatter free valve, commonly much thinner than the attached valve; an irregu-
larity and asymmetry. of growth tending to the displacement of characters normally
found in near allies of the subject under consideration, and as a general thing a camer-
ated structure of the shell. The fullest modification in this line of variation is the
production of a shell in which the attached valve is cup-shaped, conical, or sub-cylin-
drical as seen markedly in the Chamide and Rudistee. In this group all the transitions
may be traced from a simple ostreaform or exogyriform shell to the most highly modified
conical type. A further result of the modified condition induced by fixation and which
I consider perhaps the most interesting of all, is the fact that the attached valve is the
most highly modified; the free valve being the least modified and retaining more of the
ancestral characteristics, enabling one through it to trace the genetic connections of the
group, evidences of which may be completely eradicated from the attached valve. If the
ostrean form is due to the conditions of direct fixation, then all or nearly all Pelecypoda
which have one valve cemented to a foreign body for a longer or shorter time should be
ostreaform or modified in this line of variation.’ First, fixed forms will be considered as

1The Rudistz are conical or cup-shaped Pelecypods with
a superficially marked radial symmetry. So striking is
the radial feature that they have been classed with the
corals or Cirrhipeds, and the term radial is combined fre-
quently in generic and specific names of the group.
Barretia monolifera, described by Woodward (78), is
highly radial, and infoldings of its thin walls closely
resemble the radial septa of corals. In other animals
which are permanently attached by calcareous fixation as
corals, some worms, and Brachiopods, Cirrhipeds and
others we find closely comparable forms in which radial
symmetry is a marked feature. When similar forms are
built up on different lines of descent, there is strong evi-
dence that common forces acting on all alike have in-
duced the resulting form.

I suggest the following as a possible explanation of
the cause of these markedly similar forms, and would
say that I have a great deal of evidence on the question
but it must be reserved for a future publication. The

equal impact of moving water on all sides of an attached
growing organism it seems would cause an equal effort
of resistance on all sides, and therefore induce an equal
growth on all sides, thus producing a form circular in sec-
tion at any one horizon and sub-conical, cup-shaped or glo-
bose in its entirety as are all the attached forms which we
are considering. To strengthen the walls of a round or-
ganism, the wall might be thickened by solid accumulation,
by vesicular formation or by perpendicular partitions ar-
ranged at right angles to the supported wall. Such me-
chanical supports are characteristic of attached Coelenter-
ates, Rudistz, Cirrhipeds, and some Vermes. As all sides
of the periphery of an attached organism are equally ex-
posed to food-supply, danger, etc., the organs, as tentacles,
nerve-centres and eyes, would gradually tend to become
situated at all points on the periphery, or radially. It is
well known that the external parts of an animal are more
easily modified than the deeper-seated parts. It is also
known that the modification of deeper-seated parts may be

PHYLOGENY OF THE PELECYPODA. 393

proof of this fact and then special cases will be considered which show the relation of
cause and effect.

The genus Spondylus is attached by the right valve.
flatter left and a highly developed camerated structure.
spinous, uneven, and the lower valve is often highly ostreaform as shown in Spondylus
hippuritarum, P|. xxv1, fig. 1. The group is called spy oysters from the typical spiny
growth and ostrean form. In the Challenger Report, a species is described as Spondylus
ostreoides, Smith, so named, on account of the close similarity to a small oyster, the
features of comparison being marked in both valves. Plicatula, a genus nearly related
to Spondylus and attached by the right valve, is also highly ostreaform.
as is the case with the Ostreadz, becomes attached at the close of the prodissoconch
stage, as shown in Pl. xxvu, fig. 5, and the ostrean characters are assumed at that
period of development.

The Chamidz as a group are characterized by ostrean features, though departures
in degree rather than in kind are seen in several genera. The genus Chama has typi-
cally a deep, irregular, thick, attached valve and flatter, thinner, free valve. A Terti-
ary species is named Chama grypheata, on account of its close likeness to the genus
Gryphza. Chama constitutes one of the best proofs that the condition of fixation
causes the form of the valves, because some species are attached indifferently by the
right or left valve, and in both cases the resulting form is the same, a deep attached
and a flatter free valve. Chamostrea albida, P|. xxvt, fig. 2, which is attached by the
right valve, is almost indistinguishable in external characteristics from many forms of
Exogyra. The aberrant forms of the Chamide, as Caprinula, fig. 8, Monopleura, fig. 9,
and Caprotina, fig. 10, Pl. xxv1, also similar allied genera, are forms which may be con-
sidered as extremes of variation due to fixation. The attached valves are deep and
mostly cornucopia-shaped, often highly irregular and asymmetrical, whereas the free
valves, though frequently modified, are yet not different from what is often met with
in genera of unattached Pelecypods. In these peculiar genera it is the free less modi-
fied valves and extreme young to which we must turn in studying the relationships of
the groups. In occasional species the attached valve is not deeper than the free valve,
which is the larger and deeper of the two. Such forms are exceptional, as in all other

It has a concave right valve,
The growth is irregular,

This genus,

progressively produced from without inwards (example: adult. A similar line of comparison may be made in the
development of internal from external skeletons in fishes), development of corals. Cirrhipeds in the shell show a
or centripetally as I would call it. To return, if we have very complete radial symmetry especially seen in Coronula.

an attached animal that is round, with some organs ar-
ranged on the periphery (a very common case), then fur-
ther modification or development of such organs or parts
would tend to be preduced centripetally or radially, for the
centripetal variation of a round form equally exposed to
modification on all sides, would, as a mechanical necessity,
be radial variation. Inthe Hydrozoa there is strong proof
of the correctness of this view. Passing from Photohydra
to Hydra, to the hydroid stage of Aurelia, then to the free
medusoid stage of Aurelia and other Discophores, we find
progressively a more and more perfect radial symmetry
built up centripetally. Passing to free swimming Hydro-
zoa, as the Ctenophores, radial symmetry is eradicated,
and an almost complete bilateral symmetry adopted by the

Many permanently attached forms of bilateral classes of
animals assume a considerable degree of radial symmetry
and, conversely, many free locomotive genera of radiated
animals assume a considerable degree of bilateral symme-
try. As bilaterality is a feature induced and progressively
built up by the conditions of free locomotion, it is believed
that radiality also is induced and progressively built up
by the conditions of permanent fixation with its attendant
influences of environment.

Many animals are attached by more or less fleshy proc-
esses from the body, as Brachiopods by a peduncle and
Pelecypods by a byssus, but these do not adopt a round
form, as do those groups which are attached by a more
permanent calcareous fixation.

324 ROBERT TRACY JACKSON ON THE

fixed Pelecypods that I know of, the normal condition of a deeper fixed valve is main-
tained. However, the relative size of the fixed and free valves is a factor, which may be
subject to variations without impairing the general conclusions as to the effects of fixation.

The Rudistz is a group having an enormously developed attached valve of highly
irregular growth and a flattish free valve. The free valve may be of a form similar to
that found in unattached normal Pelecypods, while the fixed valve is most irregular as
shown in Radiolites fleuriausa, Pl. xxvi, figs. 11-12. Commonly, however, the free valve
of the Rudistze is radially plicated in the adult and differs widely from any normal form
of Pelecypod shell. This extreme variation marked in both valves is accounted for as a

‘ase in which the conditions of fixation have produced the extremes of modification
(see note, p. 322), affecting the free as well as the fixed valve, instead of the latter
principally, as is the normal case.

Dimya is a genus of rare and little-known Pelecypods. The shell is irregular os-
treaform and attached by the beak of the right valve which is deeper and larger than the
other. We fail to see any important characters of anatomy (described in D. argentea by
Dall) which would connect it closely with Ostrea to which it has been compared by
writers. From my view, whatever the affinities, its ostrean form of shellis due to the
conditions of fixation.

The Mulleriaceze comprising the genera Mulleria, Adtheria and Bartlettia is a divis-
ion of the Unionidee, which in the adult condition resemble oysters. Mulleria lobata, Fer.,
is so remarkably like an oyster that it has been called the fresh-water oyster. In the
Monomyarian adult, Pl. xxv1, fig. 5, the shell is rough and irregular, with a deep attached
and flattish free valve,and a specimen in the Museum of Comparative Zodlogy is indistin-
guishable in shape from forms commonly found in Ostrea virginiana. The shell as in
AMitheria, also has a finely camerated or blistered structure. The young shell of Mulleria,
Pl. xxvi, figs. 6-7, is Anodon-shaped, equivalvular and dimyarian as described by au-
thors. In this genus we have the important feature of an Anodon-shaped shell, losing
its symmetry and becoming irregular and ostreaform as soon as attached.

Hinnites is an important genus in our considerations. In the young it is free and pec-
teniform, but in the adult is attached by the right valve, is irregular and more or less os-
treaform. So close is the likeness to an oyster that in the synonymy of the genus it has
been named Ostrea and Ostracites. JZinnites cortesi, Defr., Pl. xxvi, figs. 3-4, from
the Tertiaries, illustrates clearly the characters found in the young and adult shell. We
will consider the two valves separately. In the young nealogic period the right valve,
fig. 3, is purely pecteniform. It has well developed ears, a deep byssal sinus, and an
evenly plicated shell which at this stage is nearly or quite equivalvular. At a definitely
marked period, as shown in the figure, the valve becomes attached to a foreign body by a
cementing fixation of the shell. With this change im condition, a most marked change in
form immediately takes place. The shell grows irregularly, losing the evenness and in
considerable degree even the existence of the plications marking its nealogie period.
While the early nealogic shell was but slightly concave and pecteniform the adult fixed
valve is deeply concave and highly ostreaform. ‘The byssal notch, which is a marked
feature of the nealogic stage, is filled up and completely wiped out of existence by the
irregular ostreaform growth. The left free valve, Pl. xxvi, fig. 4, in the nealogie stage,

PHYLOGENY OF THE PELECYPODA. 325

is even and regularly pecteniform like the right valve. When the right valve becomes
attached, the left valve affected somewhat by the same conditions loses much of the
regularity and evenness of marking characterizing its nealogic period, but it never be-
comes as irregular in growth and ostreaform as does the right (attached) valve. Hin-
nites is marked therefore by a byssated pecteniform stage followed by an attached os-
treaform stage. The transitions described in Hinnites cortesii, I have also observed
in the living species H. gigantea, Gray, and H. sinuosus, Gmelin.

Ostrea we have shown becomes attached at the close of the prodissoconch period, and
the same is true of Plicatula (see section x1); as soon as attached both these genera, as
in Hinnites and Mulleria, assume the irregular ostrean form of shell, losing entirely the
symmetry in the valves of the early unattached period.

Pernostrea, a Jurassic genus, is of interest in these studies. According to Stoliezka
the shell is “rounded or oval, solid, more or less tumid, inequivalve, the left valve being
in adult specimens attached; structure lamellar, resembling that of Perna See 0%
Again Stoliczka says: “This genus forms a connecting link between Melina (Perna)
and Ostrea, differing from the former especially by its sessile habitat, absence of a bys-
sal sinus and strongly excavated muscular scar; from the latter by the presence of sep-
arate ligamental grooves. Hxternally Pernostrea is barely distinguishable from Ostrea.’’
From our point of view, the ostrean form is due to the condition of fixation. Therefore
Pernostrea is like Ostrea because it is cemented to foreign bodies. I have shown that
in Hinnites, Pl. xxv1, figs. 3-4, the byssal notch disappears as soon as the shell becomes
soldered to an object of fixation, as the byssal attachment at that period is abandoned
and the shell assumes an irregular ostrean growth. Therefore the disappearance of a
byssal notch in Pernostrea is another character fairly attributable to the condition of
cemented fixation. In the species of Perna in Pl. xxv1, figs. 16-18, we find that from the
typical form of this genus havinga long hinge area and many cartilage pits, fig. 16, a se-
ries may be traced to a form with few cartilage pits and wing-like productions of the
shell, fig. 18. If we should carry this reduction of the cartilage pits still further till only
one remained while still retaining the wing-like productions of the shell we should have a
form similar to the nearly allied genus Malleus. Still more to the point is the fact that
in studying young Perna I have found that at first only one cartilage pit exists, which is
triangular in outline and oblique in position as in Meleagrina, Malleus and Ostrea. It
seems that there is no difficulty in a single cartilage pit being derived from many ear-
tilage pits; so, that other points being considered, Pernostrea may very properly be con-
sidered as intermediate in form between Perna and Ostrea. As Pernostrea is known
only in the Jurassic, and true oysters are known from older formations, it is best to con-
sider Pernostrea not as in the line of ancestry of the Ostreadx, but as a separate branch
from Perna parallel to the branch on which the Ostreadz evolved as discussed at the
close of this section and in section xvr.

In looking back to the earliest known oysters, we find that they are very typical, in-
equivalvular ostrean forms. M. Barrande’s Preostrea bohemica, Barr., a form which
seems referable to the oysters, is based on isolated valves attached by the umbo of appar-
ently the left valve. His Pl. 111, fig. 2, is an upper free valve and bears at the extreme
tip a rounded, very marked early stage which resembles the ostrean prodissoconch as

MEMOIRS BOSTON SOC. NAT. HISYT., VOL. IV. 44

326 ROBERT TRACY JACKSON ON THE

described and figured in this paper. Whether this comparison can be properly made
‘an only be ascertained by a study of the specimen. Preeostrea is from the formation,
H. 2 and 3, Systéme Silurien, a period equivalent to a portion of the Upper Silurian.
If a true Ostrea, itis the earliest appearance of the group yet recorded. Professor L. de
KKoninck figures a form from the Carboniferous as Ostrea nohillisima. His figures re-
semble a Perna, and are very much the form we might expect if that genus, or a similar one
(as Perna is not known before the Trias), should have become attached. Ostrea pater-
cula, Winchell, from the American Carboniferous, appears to be a true Ostrea and has
the typical form of the genus.

As it is shown that the ostrean form is due to the conditions of fixation, these early
and questionable species have to be looked at sceptically, as any Pelecypod adopting a
fixed habitat might assume a form which in isolated specimens could well be taken for
an oyster. The paucity of the Permian and Triassic formations gives but little evi-
dence concerning the early history of the Ostreadze, but in the Jurassic the family is well
developed, and from that time on there are abundant species and individuals for study.

In the early Jurassic, Exogyra and Gryphiea are developed as well as typical Ostreas,
and the question comes up as to which is to be considered the typical, least modified
form of the family. I show (below) that Ostrea, the type of the family, is connected
with Perna, or Perna-like forms, by important characters of anatomy and shell structure.
It has been shown that the attached valve is that which is most modified in all Pelecypoda
which solder one valve to a foreign body. Therefore, the same line of reasoning inay
be followed, and as Ostrea is near to Perna, Exogyra and Gryphza should be consid-
ered as extremes of variation in the Ostreadee. They are extremes of the Ostreadz as
Caprinula, Pl. xxv1, fig. 8, and Monopleura, PI. 7d., fig. 9, are extremes of the Chamide,
and in the same line of variation. The almost simultaneous appearance of the three
genera of the Ostreade in marked abundance in the early Jurassic is probably due to
the recognized law of quick development of new types of animals, together with the fact
that the Trias and Permian are formations bearing few fossils, so that what Ostreade
did exist in those periods have been but fragmentarily preserved.

After our studies in previous chapters and the present discussion of shell form we are
in a position to consider the probable ancestry of the Ostreade. The anatomy of Os-
trea bears a close similarity to that of Perna. The gills of the two genera bear a close
resemblance.’ The palps, heart, adductor muscle, position and termination of the in-

Dissecting Perna ephippium I find that the dorsal united behind the foot. Probably he refers to the adult.

border of each pair of gills presents the cross connecting
lattice-like meshes characteristic of Ostrea. In some
specimens the two pairs of gills are separate from one an-
other throughout their extent, as represented in Pl. xxtv,
fig. 10, whereas in other specimens the two median gills
are connected by concrescence at their dorsal border thus
uniting the two pairs of gills, as in Ostrea. The degree of
concrescence varied in different specimens which possessed
it; the gills being united for their whole extent or only
posteriorly. That such a difference as this should exist in
what is commonly considered a fundamental character is
peculiar and should be studied on fresh specimens (mine
being alcoholic). Ina young Meleagrina, I find the two
pairs of gills separate, though Woodward says they are

In an Avicula also, [ find the two pairs of gills separated
throughout their extent. In both Avicula and Meleagrina
the characteristic lattice-like meshes connect the la-
mellke of either pair of gills on their dorsal border as in
Perna and Ostrea. The anatomy of Meleagrina and Avi-
cula as far as ascertained from my rather limited material
is as near to that of Ostrea as is the anatomy of Perna,
and it may be that Ostrea is a direct branch from Avicula
rather than from Perna Ihave adopted the latter view
because the shell of Perna approaches nearer to Ostrea
than does the shell of Avicula, and because in Pernostrea,
an unquestionable branch from Perna, we find a form al-
most identical with Ostrea.

PHYLOGENY OF THE PELECYPODA. 327

testine and mantle border are also closely alike in Perna and Ostrea. In Ostrea the foot
is wanting, but probable remnants of it are preserved as the pedal muscle mentioned by
Dall, pd, Pl. xxv, fig. 12. The prodissoconch of Ostrea is strikingly like that of Perna.
Tn both genera it is an oval, nearly or quite equivalvular shell with umbos directed pos-
teriorly. In both it is composed of homogeneous lime, and shows fine concentric lines
of growth. The succeeding dissoconch in both Ostrea and Perna is strikingly different
from the prodissoconch. In Perna it is composed of an external laminar prismatic cell-
ular layer with internal nacreous layers. In the oyster the right valve has an external
thin laminar prismatic layer, the bulk of the valve being composed of sub-nacreous tis-
sue. The left valve, on the other hand, in the young, is composed wholly of sub-nacre-
ous lime, not having the typical prismatic layer (vide p. 314). This absence or extreme
reduction of the prismatic layer in the left valve of the oyster may be due to the mod-
ifying influence of fixation or to other causes, for in young Pectens a prismatic layer
is found in the right valve although wanting in the left.

Having seen some important likenesses of the two genera, Perna and Ostrea, let
us examine the differences and see if they can be accounted for. Ostrea is attached;
this habit is acquired by many Peleeypods in widely separate groups. The loss of the
foot and accompanying byssal notch in Ostrea is easily explained as the result of fixa-
tion, for such a reduction and disappearance are described in attached Hinnites, Pl. xxv1,
figs. 3-4, Spondylus, Pl. xxvu, fig. 4, and Pernostrea. The shell of Ostrea is inequivalvu-
lar, of rough, irregular growth, especially marked in the lower valve, which is deeply con-
cave; the sub-nacreous tissue is characterized by a shallow camerated structure. This
form and the cameration we have shown are the result of the conditions of fixation acting
on attached Pelecypod shells. Ostrea has but one cartilage pit in the middle of the line
of ligamental tissue connecting the valves, whereas Perna has many pits situated on
such a line of tissue, Pl. xxv1, figs. 16-18; but this character is variable in Perna and
in the young only a single pit exists, figs. 50,31, p. 329. The nepionic period, as shown
in many genera in this paper, is of value in tracing relationships which are with difficulty
or not at all traceable in the adult. Therefore we naturally turn to the nepionic period
of Ostrea. I have not been able to trace connections between the nepionic stage of Os-
trea and similar periods in Perna, but it is to be remembered that Ostrea, even at this
early period, is already attached and highly modified, so that close comparisons are not
to be expected. All differences seem referable to details, which are explicable on estab-
lished bases of argument. We fecl justified, therefore, in supposing that Ostrea is a
close ally of Perna, the differences between the two genera being largely explained by
the effects of the changed conditions of environment acting on the genus Ostrea. Os-
trea probably descended directly from Perna or from a close common ancestor of the
two genera, and by soldering its shell to a foreign body at the close of tie prodisso-
conch period has by this means eradicated from the shell, features which might otherwise
render the young referable to Perna in form. .

VII. Perna, AvicuLA AND NEAR ALLIES.
Seeking on masses of Perna ephippium, L., preserved in alcohol, a number of young
8 1 , S 5
specimens were secured, which had the embryonic shell well preserved. The young were

328 ROBERT TRACY JACKSON ON THE

attached by a byssus as are adults, and it was due to this fact that they were collected
at the same time with the adults.

The prodissoconch of Perna, as shown in PI. xxvuz, figs. 1-3, and Pl. xxvut, fig. 11, is
very closely similar to that characteristic of Ostrea, the umbos are directed posteriorly
and the histological structure is homogeneous and laminar. The succeeding dissoconch is
sharply marked off from the prodissoconch by the shape of the shell and by the fact, as in
Ostrea, that it is composed of lime deposited in a prismatic manner in aconchyolin matrix.
These two characters are features as clearly marked and as striking in the young of
Perna as they have been shown to be in the young of the oyster, fig. 27, p. 314. The
byssal notch of Perna originates in the initial stages of the dissoconch, as shown in PI.
xxvil, figs. 2-5, viewed from the right side. It is a fact to note in our comparative
studies that there is no indication of a byssal notch in the prodissoconch valves of Perna;
and the same is true of the prodissoconchs of Spondylus, Pl. xxvu, fig. 4, Avicula, fig.
33, p- 330, and Pecten, Pl. xxvu, figs. 8-9, all tending to prove that the embryos lived a
free locomotive existence as did Ostrea up to the close of the prodissoconch period and
crawled, if at all, on the ventral border of the valves. As viewed from the left upper
side the young shell of Perna, Pl. xxvu, fig. 1, is auriculate, but shows no indications
of the byssal notch; as in the lower right valve, Pl. xxvu, figs. 2-3, the same being true
of young Pectens, Pl. xxvut, figs. 9-12. This is attributed to the fact that it is the
right valve which is most affected by the presence of the byssus, for that organ in moor-
ing the animal to its habitat passes over the edge of the right valve, as the animal lies
on the right side.

~The degree in which a special notch exists for the extrusion of the byssus and foot
appears to bear a direct relation to the position of the axes of the body as compared
with the hinge axis of the shell (relation of axes to the shell, see p. 309). In byssated
Pelecypods which have the anterioy portion of the body closely underlying the hinge
area, a notch is always developed. As examples, may be cited Perna, Pecten, Avicula
and Tridacna. On the other hand, byssated Pelecypods in which the antero-posterior
axis is more nearly parallel to the hinge axis, commonly there is only a slight byssal
notch, or none at all. Such we find to be the case in Mytilus, Modiola, Pinna and most
Arcas. In studying these genera it is evident that when the foot is close to the hinge
axis a notch in the valves greatly aids in its easy protrusion; as if no notch existed the
valves would have to open very widely to permit of its passage. Young Pecten, how-
ever, in virtue of the deep notch, protrudes its foot and crawls freely, while lying on the
right side (see studies of Pecten), although the valves are almost closed. Genera in which
the foot is far removed from the hinge line can protrude it, even if of considerable size,
by a comparatively slight gaping of the valves, and a permanent byssal attachment
seldom induces a considerable notch. Those in which the foot is far removed from the
hinge line mostly crawl on the ventral border, with the valves raised perpendicularly, not
as in. Pecten and similar genera where crawling is effected by dragging the shell still lying
on its right side, the foot being extended through a special notch in the right valve.
From studies of Pecten I am led to believe that a notch is a feature caused by the ex-
tension of the foot quite as readily as by the existence of a byssus, although either or-
gan would affect shell growth by its permanent existence, or activity.

The shell of Perna in the nepionic stage has not yet acquired adult characteristics.

PHYLOGENY OF THE PELECYPODA. 329

Viewed from the left side, Pl. xxvu, fig. 1, it possesses a rounded wing anteriorly and
posteriorly the wing is rounded but not produced beyond the borders of the valves.
The hinge of this early period of growth, fig. 30, is markedly different from the adult.
The cartilage pit is single, acutely triangular in outline, situated obliquely (as in Mal-
leus, Pl. xxv, fig. 19),
and extends downward
from the hinge area of
the prodissoconch valve
as in Ostrea, Pl. xxtv,
fig. 19. As the shell

grows, fig. 31, new

cartilage pits originate eds pa es ‘ Sieh Se NES oie
on the hinge line each gm ee. ae \ eee x
being triangular at first, “4 { 3] Pe Pt bs \ v\ "

but not as acutely so as
is the first formed pit. Bie: 40-—woung Perna INTIS right valve; p, prodissoconch,, 1, single triangular
= a = cartilage pit; c, cardinal and ¢, lateral teeth. X 90 diam.
Hinge teeth exist in FiG. 31.—Perna ephippium, older specimen than fig. 30, showing several triangular
young Perna (figs. 30-_ cartilage pits and two lateral teeth. Lettering as in fig. 30. 24 diam. (Drawn by the
31), although wanting a
in the adult, Pl. xxv, fig. 16. Two cardinal teeth in the right valve and one in the left
are situated directly beneath the umbos. A slight posterior lateral tooth exists as well
in either valve. The teeth correspond with similar teeth in a specimen of Avicula sterna
with which they were compared and they are found in specimens of Perna ephippium
up to a centimeter or more in height. These teeth are apparently directly inherited
from Avicula, which is in all probability the immediate ancestor of Perna. During later
growth new cartilage pits originate on the hinge line, and the triangular outline char-
acteristic of the developing pits is lost, giving place to the perpendicular form of pits
characteristic of the adult, Pl. xxvi, fig. 16. The hinge teeth also disappear during
growth, and the form of shell characteristic of the species is rapidly built up.

The anatomy of adult Perna resembles that of Ostrea as already considered, section v,
and the anatomy of the prodissoconch period of Perna is probably closely like that de-
scribed in the oyster, p. 300, Pl. xxrv, figs. 1-2. At least one difference, however, ex-
ists, and that is the presence of a well-developed foot. The foot of Perna as indicated
by the position of the notch in the succeeding dissoconch growth (PI. xxvu, figs. 2-3)
occupies an antero-ventral position in the prodissoconch shell, and a similar position was
occupied by the foot in the young prodissoconch period of Avicula, fig. 33, p. 330, Pecten,
Pl. xxvu, figs. 8-9, and Anomia, Pl. xxrx, figs. 6-7. The position of the foot may be
considered as affording evidence concerning the anatomy of the prodissoconch stage, as
it doubtless closely underlies the anterior adductor muscle, its normal position in dimya-
rian Pelecypods, for Perna was unquestionably dimyarian at that stage. The foot of
adult Perna is small compared with that found in Pelecypods where it is used actively in
locomotion; it bears a deep longitudinal distal cleft as in Pecten, and from a second
deeper proximal cleft, the byssus originates. The prodissoconch of Perna in shell feat-
ures and implied anatomical parts is referable in origin to the Nuculoid radical, which

330 ROBERT TRACY JACKSON ON THE

is also represented by the prodissoconch in Ostrea, Avicula, Pecten, ete. (see full dis-
cussion, section Xv).

In studying fossil Pelecypods we find many species of Palseozoic Aviculoids which
are directly comparable to the nepionic stage of Perna, Pl. xxvu, fig. 1. Some species
of Leioptera and Leptodesma, as figured by Hall in the Paleontology of New York,
Vol. v, part 1, are very similar to young Perna. The most striking resemblance, how-
ever, is found in species of Rhombopteria (gen. noy. described in section xvi) from the
Silurian, as figured by Barrande. Rhombopteria glabra, sp. Mist. (see Barrande, PI.
228), is almost identical in form with young Perna. The form of shell found in the adult
of species of Rhombopteria is shadowed in the stages of development of many genera
of the Aviculide and their allies as Avicula, fig. 32, p. 330; Pecten, Pl. xxvii, fig. 9,
ete. From these close comparisons, Rhombopteria is selected as the radical from which
these genera are evolved as discussed in section xvi, and represented graphically in the
table at the end of that section. The prodissoconch of Perna in common with that of
Ostrea, Avicula, Pecten, ete., is the representative in the ontogeny of the individual of

FIG. 32. FIG. 33. Fig. 3t-

Fic. 32.—Young Avicula sterna, showing, p, the prodissoconch and succeeding nepionic stage, viewed from the left side.

Fic. 33.—Same specimen viewed from the right side, showing the byssal sinus. Figs. 32-33 are drawn from the um-
bonal tip of a young specimen by following the lines of growth. X 38 diam.

Fic. 34.—Young Avicula sterna, showing a much later growth than figs. 32-34, X 6.3 diam. (Drawn by the author.)
the Nuculoid radical from which the whole group evolved (section xvr), and the suc-
ceeding nepionic stages are referable in origin to species of Rhombopteria the next step
up in the phylogeny of the group. The prodissoconch, also the teeth, single triangular
cartilage pit and form of the shell in the nepionic stage of Perna show how closely
stages in growth may be compared to ancestral kindred and how valuable such stages
are in tracing the phylogenetic history of groups.

The developing shell of Avicula presents characters of interest in themselves and of
great value in our phylogenetic studies. Through the kindness of Prof. A. E. Verrill,*
I recently received some specimens of young Avicula sterna, Gould, from Panama. The
specimens were growing thickly on the branches of Gorgonias, to which they were at-
tached by a byssal fixation. A well defined prodissoconch exists at the umbos of the
valves, figs. 32-33, and the right valve of the prodissoconch is somewhat larger than
the left valve. The umbos of the prodissoconch are directed posteriorly; it is composed
of homogeneous lime and shows fine concentric lines of growth. The prodissoconch of

1} am indebted to Professor Verrillof Yale University, stitution for many valuable specimens of developing Pel-

for other specimens besides the Aviculas mentioned, and ecypods; some of which are described in this paper.
Lam also indebted to Dr. C. E. Beecher of the same in-

PHYLOGENY OF THE PELECYPODA. aril

Avicula is closely comparable to the same stage of Ostrea and Perna as described, and
the anatomical features of this stage of Avicula are doubtless similar to those of the
genera mentioned.

The early dissoconch growth of Avicula is sharply defined from the prodissoconch by
its changed form and histological structure. Whereas the prodissoconch is composed of
homogeneous lime, the dissoconch externally is composed of prismatic cellular tissue,
with an inner layer of nacreous tissue (compare with Ostrea, fig. 27, p. 314). The deli-
cate spinose productions of the epidermis are made up of prismatic tissue like the rest of
the superficial layer of the shell. In form the nepionic stage of the dissoconch is sub-
rhomboidal, fig. 32; the hinge line is produced beyond the borders of the prodissoconch
in a straight line and anteriorly is terminated by a slightly auriculated wing; posteriorly
the wing is slightly rounded at the hinge line, and is not produced beyond the free bor-
ders of the valves as in later stages. In the right valve, fig. 33, the byssal sinus origi-
nates with the introduction of the nepionic stage and is progressively produced to the
free border of the valves. At this stage, the developing shell of Avicula is closely like
the developing shell of Pecten at the same stage, Pl. xxvii, figs. 9-10 and Perna, PI.
xXxvil, fig. 1. It also approaches near to the form of the young of many living and fos-
sil allied genera, and in its left valve, fig. 32, bears a close resemblance to the adult of
Rhombopteria, the primitive radical of the Aviculidze (see section xv1). During the
later growth of Avicula sterna, fig. 34, the posterior wing is gradually produced in an
increasingly sharper posterior extension as indicated by the lines of growth in the fig-
ure, and soon the full characters of the adult are introduced. The stage represented
in fig. 34, is compared in section xvt, with the ancient genus Leptodesma, which this
stage probably represents in the ontogeny of Avicula. In the youngest specimen of
Avicula observed, the hinge teeth are similar to those of the adult, and the cartilage pit
is triangular, oblique and produced downwards from the hinge line of the prodissoconch
as in Perna, fig. 50, p. 329. Similar characters to those found in Avicula sterna are
shown in a specimen of Avicula atlantica, sp.? sent me by Dr. C. E. Beecher.

The prodissoconch of Avicula as in Perna and Ostrea, represents in development the
Nuculoid radical of the whole group (discussion, section xv). The nepionic stage which
is the next stage in the ontogeny of the individual corresponds closely with, and is the
representative of Rhombopteria, which is the second step upward in the phylogeny of
the group (see table, section xv1). The third stage in Avicula in which we find the
posterior wing produced and the anterior wing auriculate, fig. 34, bears a close resem-
blance to typical species of the genus Leptodesma which is considered as the third step
in the phylogeny of the group (serial steps and discussion, see section xvr). Thus
Ayicula in its development represents very fully the phylogeny of the genus, as also
seen in Perna and very clearly in Pecten, section x.

A specimen of young Meleagrina glabra, Gould, from New Zealand, No. 584, Yale
University Museum Collection, has a prodissoconch and succeeding nepionic dissoconch
similar to that of Avicula sterna. One difference, however, exists, in that the young
Meleagrina like the adult has no cardinal or lateral teeth as in Avicula.

Specimens of Monotis tenuicostata, from the Jurassic of the Black Hills of Dakota,
were loaned me by Dr. Beecher for study. The prodissoconch is preserved in both
valves. It has the umbos directed posteriorly, and resembles the prodissoconch of Ayic-

302 ROBERT TRACY JACKSON ON THE

ula as far as can be ascertained in its partially eroded condition. The byssal sinus of
the right valve originates in the initial stage of the dissoconch growth, as in Avicula.

In Pseudomonotis echinata, sp., Sow., I have been able to observe the condition of the
young in specimens from the Cornbrash of Wilts, England, No. 10737, Museum of Com-
parative Zodlogy. A well-preserved prodissoconch like that of Avicula, exists at the
umbo of either valve. The nepionic stage of the left valve, which seems to be the more
normal, least modified valve in this species, is sub-rhomboidal in form and resembles the
same stage of Avicula, fig. 32, p. 330. In the right, flattish valve, the byssal sinus orig-
inates with the introduction of the nepionic stage as in Avicula, ete.

In some well-preserved specimens of right valves of Oxytoma inequivalvis, sp. Sow.,
from the middle Lias, Pl. xxvu, fig. 7, I found a well-preserved prodissoconch. As in
related genera the umbos of the prodissoconch are directed posteriorly, and the byssal
sinus originates with the introduction of the nepionic stage on the ventral border of the
prodissoconch as shown in the figure.

In specimens of Cassianella from the Triassic, a prodissoconch has been found similar
to that of Avicula. It is worth noting that Cassianella is from the oldest geological
formation from which I have so far succeeded
in finding a Pelecypod with the prodissoconch
still intact. Dr. Beecher sent me recently
specimens of Cassianella decussata, Miinst,
which show the feature mentioned, fig. 35.
The prodissoconch is ovate with umbos di-

Hic. 3b. Caaidneligmicciseetns Bin OF leEL. valve: rected posteriorly, and it is sharply defined
showing, p, prodissoconch, and early dissoconch from the succeeding dissoconch, as in Avicula
a te ate es ae Museum. X30 and allied genera. The nepionic stage of

Cassianella is similar to the same stage of
Avicula (compare tip of valve of fig. 35, with the whole valve of fig. 32, p. 330). The
prodissoconch has also been observed in Cassianella grypheata, Minst, from the Triassic.

In specimens of Vulsella rugosa, Lam., kindly given me by Professor Verrill, I have
found stages of interest. At the umbos of a young individ-
ual, fig. 386, a well-defined prodissoconch exists similar to
that of Avicula. The prodissoconch is oval with umbos di-
rected posteriorly. It is composed of homogeneous lime,
and bears fine concentric lines of growth. The nepionic
stage of the dissoconch, as in related genera, is sharply de-
fined from the prodissoconch by its changed form and histo-
logical structure. The shell of the nepionic and later stages
is composed on the outside of prismatic cellular tissue, with Vihela hanke die een
nacreous tissue in the deeper seated parts. The form of the jage pit. p, prodissoconch,succeeded
nepionic stage is sub-rhomboidal, bearing a certain resem-_ Py theearly nepionic growth. X 38
blance to the same stage of Avicula, fig. 32, p. 330; but the “"™ (Prawn Py the author.)
wings are almost lost sight of in Vulsella. During later growth the hinge line becomes
twisted out of the straight line and commonly, for a brief extent at least, curves around
the ligamental area as shown in fig. 36.

The genera Malleus and Pinna might be considered in this section, but I have not

PHYLOGENY OF THE PELECYPODA. 333

succeeded in getting very young specimens of either genus. The observations which
have been made on the development of these genera are given in section XvI.

IX. Pecten. ANATOMY AND HAsits.

Material collected at Buzzards Bay, in 1888 afforded an opportunity to study Pecten
irradians, Lam., fig. 37, which is common on the southern shores of Cape Cod. Through
the kindness of Dr. Dall, I have also been able to study the young of several other species
from the collections of the Smithsonian Institution at Washington. The consideration
of the anatomy and life habits of Pecten irradians is taken up in this section and a study
of the shell of that and the other species studied is described in section x.

On the fourth of August, in examining the glass plates inmy drain-pipe traps (described
in section Ir), a number of extremely young Pectens were found. These specimens
proved of great interest on account of the shell form
and structure andalso on account of anatomical feat-
ures and the habits which I had the opportunity of
observing, as they were kept alive for several days.
No very young specimens were found on the sand-
bar immediately around the drain pipes although
diligently searched for, and the traps were a most
invaluable assistance in catching extremely young
molluscs which were easily seen on plates of glass by
allowing the light to strike them obliquely.

The specimens found on the fourth of August
varied slightly in size as shown in Pl. xxvin, figs.
9-12. <A few of the same age were found on the glass
the next day; but later none of these youngest stages — Fig. 37.—Adult Pecten irradians, Lam. Life
were seen. The habits of these were alike in all ‘i (ftom Binney’s Gould.)
the individuals observed and the anatomical features were the same as far as traced.

The young Pectens, Pl. xxvut, figs. 1, 9, 12, were free, crawling on the glass. No
byssus was observed until a considerably later stage and it is safe to say that this early
period has not yet developed the habit of byssal fixation. It is a free period, preceding
the byssal period. The young scallops were extremely active crawlers with a highly de-
veloped foot, which is capable of extreme prolongation, Pl. xxvim, fig. 1, and is often
produced beyond the edges of the valves to a length equal to the whole height of the
shell. While lying on the right valve, the foot is extended through the notch close un-
der the hinge line and is attached by a sucker-like action to the glass, shortening its
length. It thus drags the shell along on its right side. The motion is very rapid and
the little creature quickly passes over a considerable distance. Shells were frequently
observed to clap asin the adult; but they did not swim by this means and it is quite pos-
sible that they were not capable of doing so, as special features of the mantle which
serve the adult in swimming were not yet developed. The foot is so mobile and exten-
sible that it may be twisted and turned in any direction at the will of the animal.

Older scallops and the adults lie on the right valve, and are rarely found lying on

MEMOIRS BOSTON SOC. NAT. HIST., VOL IV. 45

334 ROBERT TRACY JACKSON ON THE

the left side. This latter position, is due probably to an accidental overturning, which
the scallop alters at an early opportunity. Professor Hyatt informs me in a letter that
he has examined over three hundred specimens of Pecten irradians, all of which lie
habitually on the right valve. The upper, left valve very commonly is more or less over-
grown with alge, but I have never observed them on the lower right valve. This habit
of lying on the right valve is characteristic of many related genera as Perna, Spondylus,
Plicatula, Hinnites, and Anomia. Patten (53) in his studies of Pecten Jacobus, says
that this species which has a deeply concave right valve and a flat left valve, lies habit-
ually on the right side, and if turned on the left side in an aquarium, soon rights itself.
The studies of young Pecten irradians were interesting on this point. I turned them
several times on the left valve intentionally and in every case they almost immediately
extended the foot with which they laid hold of the glass, and with a sudden jerk righted
themselves, showing that they were uneasy in the reversed position. While actively
crawling, the heart action is very rapid as may be seen through the translucent valves.

The foot of Pecten is marked by a deep groove in its distal portion and by means of
this groove is capable of laying hold of objects of support by a distinctly sucker-like ac-
tion. By this means they are capable of suspending themselves and I have inverted a
glass slide on which a young Pecten was crawling and seen it hang perpendicularly sus-
pended by the foot. This only applies to the very young, as older individuals have not
the power in suflicient degree to support their weight. The crawling action may, how-
ever, be seen in older specimens more easily than in the very young, as they are large
enough to be observed without the aid of the microscope.

The foot of a young Pecten has a well developed posterior retractor muscle, pr, Pl.
xxviul, fig. 1, on the left side. It takes origin in the proximal portion of the foot and is
attached to the left valve on the dorsal border of the adductor muscle. A dissection of
a Pecten of the same age as that of Pl. xxvut, fig. 2, showed the posterior retractor in-
serted behind the adductor in its normal position. An anterior retractor was also observed
attached to the shell in the umbonal region. It is an interesting fact that in P. irradians
there is only one posterior retractor of the foot, and that one, the left retractor, whereas
normally there are two posterior retractors in Pelecypods, one of which is inserted in either
valve on the dorsal border of the posterior adductor muscle. This important peculiarity
of the absence of a right retractor is apparently due to the fact that Pecten habitually
lies on the right valve. When lying in this position it is obvious that the left retractor is
in the direct line of traction, while crawling with the foot extended at an angle so as to
reach the ground; whereas the right retractor, if it exists, would not be in the line of
traction and has disappeared probably from relative disuse and atrophy.’ In adult Pecten
magellanicus both posterior retractor muscles are wanting, and it is stated as character-
istic of the genus that the posterior retractors are absent or present only on the left
side. Another feature due to this changed position in the mechanism of parts is the
relative situation of the cleft in the foot. In Pelecypods which have a cleft in the foot
and also crawl on the ventral border of the valves, the cleft is in the ventral border of

1 A similar disappearance of the right retractor muscle
exists in Anomia, Pl. xxrx, figs. 1-2. Lacaze-Duthiers has

cribes the loss of the right retractors to disuse and at-
rophy brought about by the peculiar position of the mus-
demonstrated it in adult Anomia ephippium and he there as- cles and their relations to the byssal plug.

- PHYLOGENY OF THE PELECYPODA. 335

the foot as seen in Mytilus, Modiola and Nucula. On the other hand, in Pecten, PI.
XxviH, fig. 1, Anomia, Pl]. xxrx, fig. 2, and their allies which lie on the right valve, the
cleft in the foot is in the right side, this being the side apposed to the ground in crawl-
ing. Whether the position of the cleft has changed, or the apparently changed position
of the cleft is brought about by a twist in the foot, is not known.

Young Pecten irradians differs from the adult in many features of anatomy and habit.
I will first consider the condition in the young and will call attention to the differences
when discussing older stages. The mantle border of the young, Pl. xxvut, fig. 1, is a
simple thickening of the edge as shown more highly magnified in Pl. xxvut, fig. 6. It
has not yet acquired the complex structure of the mantle border of the adult. The eyes
are arranged on the border, one in each plicated fold of the shell. As the plications in
the two valves alternate, so the eyes alternate on the two mantle borders, each retaining
a position in the furrow of its apposed valve. The tentacles at this stage alternate with
the eyes, as shown in the figures; they are highly extensible, sensitive, prehensile, and
are covered with papillose projections. When crawling, the young scallops use the
tentacles actively and they constantly lay hold of the glass by their sensitive tips.
Whether they assist in crawling by their prehensile activity is not known; but it is possible.
They will wind the tips around a body put in their way, such as a needle or pencil point,
and I have seen very young Pectens suspended from the glass sides of a beaker by the
clinging power of extended tentacles, the animal being otherwise completely inactive.

At a little later date, a few Pectens were found considerably older than those thus far
considered. The eyes of this stage, Pl. xxvu, fig. 2, are still single in each plication of
the shell, no more having yet appeared; the active tentacles, however, differ from the
earlier stage, fig. 1, in that there are now two alternating with each eye. The mantle
border has as yet no perpendicular wall or guard tentacles (described later) but presents
one striking feature. When at rest the two borders are protruded beyond the edge of
the shell in a short tubular form at the base of the dorsal’ ear as shown in Pl. xxvin,
fig. 2. At this point excurrent action takes the direction indicated by the arrow and a
rapid but intermittent outflow may be traced by means of the contained particles of ef-
fete matter. This extension of the mantle is at that time a pseudo-siphon; it is com-
posed of delicate tissue and is retracted and extended exactly as may be observed in the
anal siphon of young clams (vide section xtv), but it differs from a true siphon in that
the walls are not joined by the concrescence of their apposed borders. No siphonal ex-
tension of the mantle was seen in Pectens younger or older than that here described, but
in the adult it is important to notice that all excurrent action takes place at this same
area of the shell. Dr. Dall informs me that he has seen a similar condition of siphonal
extension of the mantle border in Pectens dredged from great depths.

It is an interesting fact in young Ostrea and Pecten that excurrent action takes

1Tn considering the relative position of the axes of the and the gills lie on the ventral border. The ear possessing

body to the shellit must be borne in mind that as in Ostrea
(vide p. 309) the axes of Pecten are different from those of
typical dimyarians (Pl. xxv, fig. 8). The position of the
palps and foot in Pecten indicates the umbos as the anterior
end of the shell, the free ends of the valves are posterior,

the byssal notch through which the foot extends is there-
fore the ventral ear, not the anterior ear as commonly con-
sidered; and the ear on the opposite side is correlatively
the dorsal not the posterior ear.

33 ROBERT TRACY JACKSON ON THE

place at one definite area. These genera have mantle borders separate around the en-
tire margin of the shell; yet the outflow is not in any direction as might be supposed,
but is limited to one direction.

The gills of Pecten of the age of Pl. xxvuu, fig. 2, consist of four sets of filaments all
equally developed, but there is none of the irregularity of development of the separate’
gill lamellee described in the oyster. The inner pair of gill filaments are reflected in-
wards and the outer pair are reflected outwards as in the adult, Pl. xxrv, fig. 12. The
reflected borders are about half the total length of the direct borders whereas in the adult
the reflected borders are proportionately longer. Thiscondition in the young was doubt-
less preceded by a condition in which the filaments were simple, straight processes with-
out reflected borders as described in young Ostrea and Anomia. ‘The gills are ciliated
but have not yet the ciliated interfilamental knobs or processes existing in the adult.
There are no septa joining the direct and reflected portion of the filaments, character-
istic of the adult, observed at this early stage.

When a Pecten of the age of Pl. xxvi, fig. 2, had been kept in confinement for a few
days it became attached by a byssus and from my observations it is probable that, up to
this period, the young lead a free, crawling existence without any fixation. With the in-
introduction of the byssal period, as it may be called, the young Pecten undergoes a
marked change in habits. Previously an active crawler, it now becomes most persistent in
its byssal fixation, and leads a sedentary life. The animal lies on the right valve; the bys-
sus is extended through the byssal notch as a number of delicate fasciated threads flat-
tened into a disc at the point of contact with the object of fixation. The specimen if
kept in a beaker of water frequently attached itself high up on the perpendicular sides of
the glass to which position it had previously crawled. It occasionally broke loose volun-
tarily, or was intentionally detached and in a short time, commonly a few minutes, made
a new attachment, showing that fixation is a feature of this condition. It was in the
specimen of this early stage, when attached by a byssus, that the pseudosiphonal protru-
sion of the mantle was observed. When the animal is attached, the tentacles are more
or less fully extended, the foot may be protruded or withdrawn, and feeding is active as
indicated by feecal discharges in the excurrent water.

Pectens considerably older than those already discussed, measuring 1-3 em. in height,
are instructive. At this age the byssal habit of fixation still exists; in fact, specimens
measuring slightly over 8 cm. have been found still attached, but I have never observed
such a condition in adults. Pectens 1-3 em. high, though retaining the byssal habit,
are not as constantly attached and are more active, than the younger stage just consid-
ered. Two hundred or more individuals of this size were seen on the bar and perhaps
half were attached. They were not observed in the act of crawling but they are capable
of swimming actively as later described. On account of the size of the specimens it
was easy to observe the process of formation of the byssal fixation. In observing it,
the scallops should be confined in a dish as every movement is then clearly perceptible.
Lying on the right valve, the foot is extended on the surface of the dish, the flattened
distal portion taking a firm hold as if about to crawl. This position is maintained for a
moment or two and then the foot is withdrawn within the body, by the motion of retrac-

PHYLOGENY OF THE PELECYPODA. Bay

tion it draws out, or spins, the byssal thread, which the creature had fixed to the surface
of the dish while the foot was laid closely against it. Soon the foot is again extended,
pressed flatly against the dish and another byssal thread isspun. The second byssal thread
is always attached at a point a little removed from the point of fixation of the first
thread; sometimes the two are separated by a distance of two or more millimeters. Ad-
ditional threads may be spun; but three was a common number with specimens in con-
finement. Those on the bar, especially the larger individuals, frequently spun a large
number of threads in the byssus. The byssal gland is situated in a proximal cleft-like
depression in the foot separate from the more distal cleft-like depression which serves
the animal in crawling, so that between the two there is a slight interspace without a
cleft. Frequently when forming the byssus, the foot may be arched up in this interspace,
Pl. xxviu, fig. 7; the hold being maintained by the tip of the foot and at the same time
the byssal cleft being pressed closely against the glass, so as to make the fixation of the
byssal thread. While spinning the byssus the scallop is preoccupied and pays little at-
tention to pricks or stimuli which at other times would meet with immediate response.

The attached scallop commonly remains fixed for a considerable time. If disturbed,
however, or at will without apparent cause, it may break the byssal attachment by clap-
ping its valves. It then commonly swiins about, soon renewing, or not, the attachment
as the impulse moves it. If one break the byssus loose it is torn out from the foot and
left clinging to the sides of the dish. On examining separated threads we find that the
proximal portions of the several threads are generally coalesced into one band, the distal
portions being separately attached to the glass by expanded disc-like extremity. The
threads may measure a centimeter or more in length but are more often shorter. In at-
tempting to preserve specimens of Pecten in alcohol with the byssus intact it was found
that they immediately detached themselves. Success was attained however, by first
paralyzing them with cocaine, adding the solution gradually to a small volume of sea
water.

The mantle of Pectens 2-3 em. high has acquired adult characters. In young stages,
as shown, the eyes are situated one in each plication of the shell and a single tentacle
alternates with each eye, Pl. xxv, figs. 1 and 6. Later the position of the eyes being
the same, two tentacles alternate with each eye, Pl. xxvii, fig. 2. In the age under con-
sideration, as in adults, the eyes are for the most part situated in the plications of the
shell, but in addition to one eye in each plication, new eyes have arisen between those
first formed. The new eyes, however, do not appear at regular intervals as did the first-
formed eyes. In each mantle border may be counted (as noted by Verrill) about forty
eyes, and between them are a large number of tentacles instead of only two as in the
earlier stage. These marginal tentacles, like those in the young, PI. xxvrn, figs. 1 and
6, are papillose, extensible and active, frequently winding about a small object in a pre-
hensile manner.

Besides an increase in the number of eyes and tentacles, the mantle border of Pectens,
of the age under consideration, has developed new features not seen in earlier stages.
The border is produced perpendicularly in a thick marginal wall or curtain which has
important functions. This wall is highly developed along the plicated borders of the
valves but is reduced in thickness, and height as the base of the angle forming the ears,

308 ROBERT TRACY JACKSON ON THE

fig. 37, p. 333, is reached. It is richly pigmented on its outer face and marked by per-
pendicular folds and ridges. * Just below the free, inner margin of the mantle wall it is
thickly set, with small equidistantly placed, arcuate tentacles, diagrammatically shown in
PI. xxviwt, fig. 4, which may be called guard tentacles from their functional use. These
tentacles are yellowish, not papillose, incapable of extension and move only in one plane.
When the scallop is at rest the guard tentacles stand upright. In young individuals the
tips of the tentacles of either mantle wall come in apposition, so that a grating or guard
is formed between the separated mantle walls. In virtue of this grating no large body
can pass into the gill chamber, while the animal is feeding with valves open, without
striking the guard tentacles. Collision with these tentacles commonly causes the an-
imal to shut its valves quickly. I say commonly, for irritation sometimes fails to pro-
duce the expected result. When swimming, the guard tentacles are turned outward and
laid flatly against the outer border of the mantle wall. In adults the valves frequently
are so widely separated that the tips of the guard tentacles are not in apposition, and
the tentacles are less sensitive to stimuli than they are in young scallops of two or three
centimeters in height. The guard tentacles are arranged equidistantly on the borders
of the mantle walls from the base of the ventral to the base of the dorsal ear. The re-
maining portions of the mantle walls which are relatively slight and low, bear no guard
tentacles. It is to be noticed, that the regions free from guard tentacles are the regions
marked by the extension of the foot when young on one side and on the other marked
by the excurrent flow of effete water, as well as by the flow of water when swimming.

The mantle borders of young and adults are separated for the greater part of their
extent, being confluent on the hinge line and for a slight distance toward the byssal notch,
on the ventral border. Dorsally they are confluent to the point marked as the inner
angle of the ear, fig. 37, p. 333, being a somewhat greater distance than that of the ven-
tral confluence. This relative confluence is correlated with the functional needs of the
animal, as excurrent action takes place dorsally over such an area that it is not affected
by the partial union of the mantle borders, whereas ventrally the foot actively swings up
to the hinge line, requiring and maintaining a greater separation of the borders to that
point. The mantle border isvery broad and strong, and numerous, well-developed, radial
muscles which bifurcate and anastomose freely serve to retract it. The muscles are more
highly developed in Pecten than in any other Pelecypod studied, which fact is probably
correlated with the active swimming habit.

It is stated that Pecten swims by clapping the valves, but how the motion is effected
and the direction in which the animal moves has not been studied so far as I am aware
and seems to be commonly misunderstood. The swimming habit of Pecten is one of
the most striking features of this genus, and it is accomplished with such energy and
rapidity that the creature attains great speed. In swimming, as when at rest, the left
valve is always uppermost, and the plane of the edges of the valves is inclined to the
surface of the water at an angle of about 45°.!

It is best to study the swimming in young Pectens some 3 em. high, as at that age
it is more easily seen than in adults, and does not differ from what may be observed in

The Pectens in which the swimming habit was studied angle given and some other details might vary slightly if
were confined in small aquaria, and it is possible that the swimming freely in the open water.

PHYLOGENY OF THE PELECYPODA. 339

them. Lying on the bottom with tentacles extended, the scallop suddenly folds the
guard tentacles back so that they lie closely against the outer border of the perpendicu-
lar mantle wall. The valves are then closed by a quick action of the adductor muscle
and water is forcibly expelled. The first water expelled is driven out posteriorly in the
direction of the arrow a, Pl. xxvut, fig. 8, and if this were the only or the main direction
in which a current is expelled, the animal would by impact of water be impelled in the
opposite direction or anteriorly; but the act of swimming is more complicated than this
would indicate. When the valves have closed to a slight extent the borders of the two
thick, perpendicular mantle walls come in contact and then no more water passes out as
indicated by the arrow a, but instead, during further closure of the valves, it is forcibly
ejected from the lower border of one ear, where the mantle wall is low and thin, as in-
dicated by the arrow 6, Pl. xxvim, fig. 8. The water expelled at the point 6 is the
most forcible current and probably of the greatest volume; by its means the creature
is impelled in the direction of the arrow, c. The valves open quickly and clap again.
The second time as before the first water is driven out posteriorly; but when the mantle
walls come in contact, the direction of the excurrent water is again changed and is
forced out from the lower border of one ear in the direction of the arrow d, Pl. xxvutt,
fig. 8a; being the strongest current, it impels the animal in the direction of the arrow
e. This striking difference is noticeable, viz.: that at successive claps the water is driven
out from alternate ears, first on one side and then on the other. The resultant action of
the several currents and successive claps, illustrated in Pl. xxvin, figs. 8-8 a, is therefore
to drive the animal in the direction of the free borders of the valves or posteriorly. It
is due to the alternate expulsion of the water first from one ear and then from the other
as described, that the animal presents a succession of zigzag jerks in swimming. ‘The
direction of the current alternately to the two ears appears to be voluntary, as a seal-
lop can scuttle over the bottom of a dish in a sidelong direction by successively expell-
ing the water at each clap from one and the same ear. The action of the first current
of water expelled posteriorly, before the mantle walls come in contact, gives the animal
an upward jerk and it is in virtue of this jerk combined with the momentum in a poste-
rior direction that it maintains its position on the surface of the water and also the high
angle to the surface which it presents inswimming. The current driven out posteriorly
in the initial closure of the valves is so powerful that water may be squirted by adults
to the height of five inches or more from the surface by this action.

In studying the shell we find a feature correlated with the swimming habit in the in-
complete closure of the valves on the eared areas; so that when the free borders of the
valves are in immediate contact there is still sufficient space for excurrent action at these
areas, and the valves close completely at each clap, for there is in swimming a sharp
clicking noise which could be caused only by the striking together of the valves. Seal-
lops are said to migrate in bands, being seen swimming in large numbers at some sea-
sons of the year. They are among the most active of all Pelecypods and a careful study
of their habits would doubtless prove fruitful and highly interesting.

As stated, in young scallops, the outflowing water which has served the function of
aeration is passed out from the shell at a definite point on the dorsal border as indicated
by the arrow in Pl. xxv, fig. 2. This is equally marked in character and more notice-

340 ROBERT TRACY JACKSON ON THE

able in the adult. In such the water is driven out so forcibly by the rapid ciliated action
of the gills that a rapid gyrating current was imparted to the whole mass of water in a
large dish or preserve jar in which the animal was. Further, several times on turning a
scallop over, so that the excurrent action took place in the opposite direction, almost in-
stantly the whole mass of water responded and moved in the reversed course impelled
by the changed direction of the current. The fseces of Pecten may be seen escaping in
pellets, the form of which in section is that of a clover-leaf, and they are produced in
more or less drawn-out ropy masses, Pl. xxvin, fig. 5; in one instance observed, meas-
uring 14cm. inlength. Ryder (60) observed that the feces of the oyster have a distinct
almost horse-shoe form, due to the shape of the intestinal tract and in insects the form
of the feeces is often a specific characteristic. It is possible that a study of this feature
in molluscs might show similar results.

The adult of Pecten irradians is a free swimmer. No indications of a byssal fixation
were seen at this age, and the foot is a highly atrophied organ. I have never seen the
foot extended beyond the margin of the valves in the adult, although it is so active in the
young, and it may then fairly be considered a retrograde and probably functionless organ.
There is, therefore, a series of periods of habits in the life of Pecten. In the prodisso-
conch stage of Pecten (discussion later, section x), it is either a veliger swimmer, or
crawls on the ventral border of the valves; in the early nepionic stage (Pl. xxvu1, figs.
1, 9, 10) it is an active crawler, lying on the right valve; in late nepionic and nealogic
stages, it is attached by a byssus more or less constantly, but also has the power of us-
ing the foot; in the adult, it is a free swimmer and the habits of crawling and byssal fixa-
tion seen in the young are lost. Some species, as Chlamys (Pecten) varius, are said to
retain the byssal habit of fixation throughout life.

Frequently Pectens show an extensive loss or injury of the mantle. One adult speci-
men observed had the left lobe on the ventral side entirely wanting, and the correspond-
ing lobe on the right side was destitute of the eyes, tentacles and pigment. The
damaged lobe was bereft of sensation, the creature showing no response when it was
pinched or shadowed; but the animal was apparently uninjured otherwise. Such injury
is probably due to the attacks of predatory animals, as when lying with its valves wide
open it is an easy prey. Some indications pointed toward disease, and the loss of the
mantle may be due to that cause. That scallops suffer a severe loss or injury to the man-
tle and recover, is amply proven by frequent specimens of mended shells. ‘The margin
of the shell is not injured by the predatory attacks (if such they are), but further shell
growth is prevented at the injured point by the loss of the mantle. When one mantle
border is injured, the other repairs the damage which would be caused by local want of
shell growth, by curving the shell deposition of the uninjured side rapidly inwards, thus
obviating the deficiency of the injured area. This ingrowth is probably induced by the
lack of resisting support on the part of the damaged border. Adults which have been
thus damaged in youth show that in time the injured mantle border is renewed, and nor-
mal shell deposition takes place, the early damaged period being indicated by hollow-
like, uneven depressions in the shell surface. In the oyster I have never seen injuries
to the mantle in living animals, and no evidence of serious injuries, later repaired, are
noticeable in a study of shell growth. The oyster is more sensitive to touch than the

PHYLOGENY OF THE PELECYPODA. 341

scallop, the valves are less widely separable, when open, and at time of danger stay shut,
instead of opening almost immediately after closure, as with scallop shells; therefore, the
oyster is less liable to attack, at least from certain classes of enemies, than is the scallop.
. A surprising want of sensibility is noticeable in adult Pectens. It is often possible
to insert a quite large body into the shell before it closes; but the young are more sen-
sitive. Frequently pricking the mantle border with a needle produces little effect in
adults. Both adults and young are sensitive to light, so that passing the hand or even
a pencil in front of the valves they will quickly close, as described in Pecten jacobeus
by Patten. They open almost immediately, however, and frequently repeated stimuli soon
lose their effect.

Pecten trradians is hermaphroditic, the reproductive organ is of large size, and the
two sexual functions are distributed regionally. The gland as a whole is roundly tri-
angular, extending between the gills on the ventral border of the adductor muscle, while
anteriorly it passes into and is lost in the liver mass. The gland is composed of two dis-
tinctly marked portions or areas. The ventral area is pinkish white and consists whol-
ly of the male or sperm-bearing portion; the dorsal area is brownish-yellow in color and
consists of the female or egg-bearing portion of the gland. The divisions of the organ
lie parallel, but anteriorly the’whole mass is male, while the female portion occupies the
larger part of the posterior portion of the organ. The whole gland is more or less pig-
mented with black, especially on its ventral keel. Abundant active spermatozoa were
found in the pinkish or male portion, with well developed heads and vibratile tails; but no
eges, it being solely spermatic. In the yellowish-brown dorsal portion were found de-
veloping and ripe eggs. The eggs are rounded or pear-shaped, granular, show nuclei and
resemble the egg of an oyster, in size and appearance. A. few spermatozoa were found
in the female portion, but their presence was probably due to accidental introduction
while examining the contents of the gland, as it is almost impossible to avoid introduc-
ing a few under such circumstances, and we may fairly suppose that the female portion
also produces but one element. The sexual gland of Pecten irradians, it will be seen, is
very similar to that of P. glaber, as described by Lacaze-Duthiers (40).

The gills of adult Pecten trradians are equal in length and differ much from the gills
of the oyster. The filaments of the gills are reflected on themselves to such a length
that the reflected portion equals about two-thirds the length of the direct portion as shown
diagrammatically in Pl. xxtv, fig. 12. The reflected borders are free, not joined by con-
erescence with the reflected borders of opposing gill filaments, or with the mantle lobes,
as in Ostrea. The filaments are joined to one another by the interlocking of tuft-like,
ciliated nodes, situated at intervals. About 25-30 nodes are borne on each filament,
and the nodes of contiguous filaments are in apposition. A similar structure exists in
Mytilus, and is described by Peck. This connection does not exert a very strong bind-
ing influence, so that when alive the filaments of Pecten are often separated, and in alco-
holic specimens are notably disconnected. The ciliated nodes serve to keep the filaments
together, but quite as much to keep them separated, by an interspace to allow of the free
passage of water. In the adult Pecten, at intervals of every 11-15 filaments, a delicate
connecting septum or wall is developed between the direct and reflected border of the
filaments, Pl. xxv, fig. 38. These walls strengthen the reflected borders and. also

MEMOIRS BOSTON SOC. NAT. HIST., VOL. IV. 46

€

342 ROBERT TRACY JACKSON ON THE

apparently direct currents of water in channels as it flows'through the grate-like gill.
When the reflected border of the filament ceases to cover the direct border, this wall is
built into a peculiar plicated ridge or blade, b/ (fig. cited), from the direct border of the
filament, and continues in this blade-like fashion to the point of origin of the filament.
The gills are suspended by fasciated membranes which connect the bases of the two
pairs with the adductor muscle, with the visceral mass anteriorly, and with the mantle
lobes posteriorly. Similar fasciated membranes also exist in Anomia glabra, Pl. xx1x,

figs. 1-2.

X. Precren. THE SHELL.

Characters are found in the young developing shell of Pecten trradians and other
species which are widely different from those peculiar to the adults and a consideration
of these features will now be taken up.

The first specimens of young Pecten irradians obtained were found in the drain-pipe
traps at Buzzards Bay on the fourth of August. The youngest of these specimens is
shown in Pl. xxvitr, fig. 9, viewed from the left upper side, while fig. 10 is the same
individual viewed from the right side on which the animal habitually lies. Pl. xxvu, fig.
9, is a similar specimen viewed from the umbonal area of the right side, more enlarged.
The prodissoconch, p, in the several figures is as sharply marked off from the succeed-
ing dissoconch as in Ostrea, Perna and Avicula, Pls. xxiv, xxv and xxvit. It is sep-
arated from the dissoconch as in the genera mentioned, by its form, histological struct-
ure and inferred anatomical features. The umbos of the prodissoconch of Pecten are
directed posteriorly ; but they are not as pronounced and arcuate as in Ostrea and Perna.
The structure of the prodissoconch is homogeneous and laminar, showing fine concentric
lines of growth. There is not the slightest indication of the byssal notch in the prodis-
soconch valves; but it originates in the initial stages of growth of the succeeding disso-
conch, PI. xxvir, fig. 9 and Pl. xxvin, figs. 10 and 13. As was argued when discussing
Perna and Avicula (p. 329), the position of the byssal sinus in its relation to the pro-
dissoconch shell indicates that the foot of that period of development occupied a position
on the free ventral borders of the valves, its normal position in dimyarian Pelecypods.
The prodissoconch stage of Pecten was doubtless dimyarian, a supposition confirmed by
the shape of the shell, the position of the foot and the affinities of Pecten to Avicula
and Ostrea, in which last genus two adductor muscles are shown to exist at that stage,
Pl. xxiv, figs. 1-2.

The habits of the prodissoconch stage of Pecten, we know only from analogy and
the structure indicated by the hard parts; but as it differs from the Ostrean prodisso-
conch in possessing a foot, it may be at this period, a veliger swimmer as is Ostrea or
a purely crawling form. If the latter, it doubtless crawls on the ventral border of the
valves as is the habit in Pelecypods in which the foot is extended from between the ven-
tral border of the valves opposite the hinge line. Only very young Pectens show the
prodissoconch still intact, as it is not retained in this genus as long as in the sedentary
oysters, Aviculas and Pernas. The prodissoconch of Pecten as in Ostrea, Perna, Avic-
ula, ete., is traceable in origin to the Nuculoid radical, from which the Aviculidee and
their allies were derived (discussion, section xv, and table, section xvr).

PHYLOGENY OF THE PELECYPODA. 343

The succeeding dissoconch growth in Pecten is clearly and sharply separated off by
its changed form and histological structure. Whereas the prodissoconch shell is homo-
geneous and laminar in structure, the succeeding dissoconch of Pecten irradians, in its
early stages, Pl. xxviu, figs. 9-12, has a thin layer of prismatic cellular tissue in the
right valve, the same feature having been observed in Ostrea, fig. 27, p. 314; also in An-
omia, as shown in later studies. The early stages of the dissoconch growth of Chlamys
(Pecten) islandicus, Chemn.,. Pecten dislocatus, Say, P. magellanicus, Gmelin and P.
monotimeris, Con., also show marked prismatic cellular structure in the right valve. The
prismatie stricture of these several species does not occur in patches here and there,
but extends over the whole early dissoconch shell in a thin but clearly marked layer.
This is a most interesting fact as adult Pecten trradians does not exhibit any evidence
of prismatic structure, though it is well developed in the young. Professor Carpenter
in his first paper on shell structure, in 1844, says that he has seen traces of cellular
structure in the Pectinidz; but believes that it is not to be considered as constant or pe-
culiarly characteristic of the group. In his later paper in 1848, he states that he has
since found prismatic cellular structure, thin, but beautifully preserved in Pecten nobilis,
which he figures. This is the only case I find mentioned in which prismatic structure
is known in Pecten. Dr. Dall (16), however, in recent studies of deep-sea molluses
has observed prismatic cellular structure in many species of Amusium, and in the allied
genera Pseudamusium and Propreamusium. In these cases the prismatic structure,
though often limited in extent, was not confined to the right valve, as in my observa-
tions of young Pectens. Dall notes that this structure is especially characteristic of
abyssal species of this group, in which also the shell is extremely thin and fragile. My
observation of prismatic structure in the young of five species studied, proves that this
feature which is not a characteristic of the group may yet be found in the young. The
Aviculidz, in which prismatic structure is highly characteristic, are in the line of ances-
try of the Pectinide. The occurrence of prismatic structure in young Pectens is there-
fore a good example of the principle of acceleration of development which is that
ancestral characters may be found in the young which are wanting in adult individuals.

Although prismatic structure is characteristic of the right valve in young Pecten irra-
dians, the left valve in this and the other species studied has no prismatic layer. In
P. irvadians, the left valve in its early nepionic stage is marked by a peculiar pitted
structure. The degree of pitting varies in different individuals, but in none was it as
marked a feature as in young P. dislocatus, where it is highly accentuated.

In Pecten the shell of the first true dissoconch or nepionic period, Pl. xxvin, figs. 9-
10, is markedly different from succeeding stages, as well as from the early completed
prodissoconch stage. The hinge is produced as a straight line and the borders of the
valves descend in a curve from the extreme limits of the hinge line without possessing
the ears which are characteristic of the adult in all species of the genus. The ventral
border of the left valve, fig. 9, has no byssal notch but near the hinge area is produced
in a lobe-like curve, the lobe projecting beyond the border of the byssal sinus of the right
valve as shown in fig. 10. There are no plications of the shell at this stage although
these originate early and are characteristic of the adult. The features of the ventral
lobe, and the absence of ears and plications, give the shell as shown in the figures cited

344 ROBERT TRACY JACKSON ON THE

a totally different aspect from that of later stages and also different from any form
known in other species of the genus.

The right valve of young Pecten irradians, on which it rests, Pl. xxvit, fig. 10, is
marked by a deep byssal notch, originating as stated in the initial stages of the disso-
conch shell as indicated by the lines of growth shown in the enlarged figure of the hinge
area in Pl. xxvur, fig. 9. The border of the notch on its lower side is marked by a well
defined tooth-like process, Pl. xxviu, fig. 10. In later grewth, Pl. xxvuit, fig. 13, more
tooth-like processes develop and they are characteristic of young specimens up to four
or five centimeters in height but none exist in the adult shell. The left upper valve
overhangs the right lower valve by a slight extent along the free margins, fig. 10, as in
young Ostrea and Anomia. The color of the young shell in both valves is of a lght
yellowish-brown.

In the consideration of Perna (p. 328) it is shown that when the foot lies close up
to the hinge area in Pelecypods a special notch exists for its extrusion. The notch is
required because if not existent the valves would have to open very widely to allow of
the passage of the foot at a position so close to the point of union of the valves. Per-
haps an equal reason for the need of the notch is the fact that such forms lie on one
valve and drag the shell in crawling by extending the foot over the edge of the lower
valve. They do not creep on the free borders of the valves as do ordinary dimyarians
in which the foot is extended from a point far removed from the hinge line by a slight
gaping of the valves, without the aid of any special sinus. As shown on p. 333, Pecten
irradians, when very young; Pl. xxvin, fig. 1, has no byssal fixation, the animal leading
an active crawling existence. The notch may therefore be considered as a foot notch
rather than a byssal notch at this period. The existence of a byssus or the constant
extension of an active foot would equally affect shell growth by exerting pressure on
the mantle border, thus causing its local retraction and retarding shell deposition at the
point of extrusion, so that it is not necessary to imply the existence of a byssus because
a notch exists at an area where the byssus would be produced if it existed.t| This has
important bearing on fossil forms which are considered as byssated or attached, if a
notch exists in this region; whereas they may with equal reason be considered, as far
as the notch is concerned, as free forms which crawled while lying on one side with the
foot extended through a special notch produced by the existence of such a habit. This
mechanical explanation of the formation of the notch in Pecten shows what close re-
lations may be traced between the form of hard parts and the anatomy and habits of
the animal.

A view of the hinge area and byssal notch of a Pecten of the same age as the above
considered specimen viewed from the right side is shown in PI. xxvn, fig. 9. The re-
lation of the prodissoconch to the dissoconch and byssal notch is clearly indicated and
it is noticeable that the left umbo protrudes beyond the right. This is due to the greater
convexity of the left valve, as is also noticeable in Chlamys islandica, Pl. xxvur, fig. 8.

The early nepionic stages of Pecten magellanicus, P. dislocatus (traced by following

‘Tn this connection a comparison may be made of the Pl. xxix, fig. 6. The early condition of Pecten does much
early nepionic shell of Pecten, Pl. xxvin, fig. 10, viewed to elucidate the somewhat similar form of shell growth of
from the right side with a similar view of young Anomia, young Anomia.

PHYLOGENY OF THE PELECYPODA. 3d45

the lines of growth) P. teste and Chlamys islandica, Pl. xxvui, fig. 14, agree with the -
same stage of P. irradians, Pl. xxvii, figs. 9-10, in the form of valves. This form
may therefore be considered as typical of the first stage of growth of the dissoconch
shell of the group. In studying genera of the paleozoic Aviculide, we find genera that
agree closely in appearance with the young shell of Pecten; especially when it is viewed
from the left side as in P]. xxvut, fig. 9. The genera Actinoptera, Leiopteria and Lepto-
dosma, as figured in Hall’s Palzeontology of New York, are very comparable to this early
stage of Pecten. A still nearer comparison may be made in species of the Silurian ge-
nus Rhombopteria (gen. noy.) fig. 51, section xv1, which are closely comparable to young
Pecten in all the species observed. Rhombopteria is selected as the primitive radical of
the Aviculidx and their allies (see discussion and table, section xvi), and it is one of the
best proofs of the correctness of that view that the nepionic stage of Pecten so closely
corresponds with that early type. In the studies of Avicula and Perna, the young of
those genera, fig. 32, p. 330, and Pl. xxvur, fig. 1, is compared closely to the same an-
cient genera to which I have just compared young Pecten. A comparison of young Pec-
ten, Pl. xxvu, fig. 9, with the similar period of growth of young Avicula and Perna,
in the figures cited, shows that at this early age the three genera are closely similar.
Differences in the angles of contour exist, especially in the hinge line; but they are
slight in degree, and the important phylogenetic conclusion may be made that these
genera in the nepionic stage are so alike that they are referable in their inherited form
to the same ancestral types. Thus we see that the form as well as the prismatic struct-
ure of the shell of young developing Pectens may be traced to ancestral types from
which the genus was evolved.

A stage of shell growth in Pecten irradians, P|. xxvin, figs. 12-13, later than that just
considered, shows the beginning of features characteristic of the family and species to
which the young scallop belongs. Plications similar to those of the adult begin to be
marked in either valve and ears to be developed, as shown in fig. 12, viewed from the left
side, and fig. 13, viewed from the right side. The byssal notch is strongly marked, as
in the earlier stage, and it is to be observed that the notch is much deeper relatively in
this early, active, creeping stage than in the adult, where the use of the foot as a creeping”
organ is lost. The base of the byssal notch, fig. 13, shows four marked, tooth-like pro-
cesses, of which one only was present in Pl. xxvii, fig. 10. These teeth may prove of
value in tracing relationships; but their function, if any exist, is not understood. The
plications of the shell are clearly defined and originate suddenly, indicating a stage by
their introduction. The plications are like those of the adult, fig. 37, p. 333, and they
are the same which are retained throughout life, no new plications, originating by bifur-
cation or interposition, in this species. The left upper valve does not extend beyond the
right valve on its free margin as in the earlier stage, fig. 10.

A later period of growth in Pecten trradians is shown in Pl. xxviu; fig. 2. The ears
are more accentuated than in the previous figures, though they do not bear to the whole
shell the same relative size and shape, as in the adult, fig. 37, p. 333. The hinge line of
this stage, fig. 2, is much shorter than in the earlier stage, Pl. xxv, figs. 12-13, but
much longer relatively to the breadth of the valves than in the adult. In other features,
excepting the depth of the byssal notch, this age does not differ essentially from the adult

346 ROBERT TRACY JACKSON ON THE

in the form of the shell. Differences in the anatomy and habits, however, exist, as dis-
cussed in previous pages.

J have shown that the early, nepionic shell growth of Pecten, Pl. xxviu, figs. 9-10, has
a form referable in origin to the adults of ancient members of the Aviculide, and I be-
lieve that the form of the later shell growth is also of phylogenetic significance. The
hinge line of Pecten irradians, in Pl. xxvu, figs. 12-13, is relatively long as compared
with the condition found in the adult of this or other species of Pectens, and correlative-
ly the ears are very slightly developed as compared with the adult. The same condition
of relative great length of hinge line and want of development of the ears is seen in sim-
ilar stages of development of Chlamys (Pecten) islandica, Pl. xxvin, fig. 14, Pecten
magellanicus and P. dislocatus; therefore it may be considered as typical of developing
Pectens. In Hall’s Palzeontology, Vol. v, Part 1, we find forms of Devonian Pectinidz
under the genus Pterinopecten which agree closely with the form of shells here figured
in developing modern Pectens. The agreement is in the relatively great length of the
hinge line and slight production of ears, which features characterize Pterinopecten and
separate it from the nearly allied genus Aviculopecten. Pterinopecten is the least re-
moved from the Aviculidz, and is the simplest known form of the Pectinids. The
young of Pterinopecten resembles the adult of Rhombopteria and the adult resembles
the young of Aviculopecten and Pecten (see section xvr). It forms, therefore, a strong
link in our phylogenetic series, and is placed as the first departure from the Aviculide
in the Pecten line of variation (see discussion and genealogical table, section xv1).

The paleeozoic genus Aviculopecten, as amended by Hall, differs from Pterimopecten
in a greater production of the ears and in a relatively shorter hinge line. Turning to the
developing shell of Pecten trradians, we find that the later growth, Pl. xxv, fig. 2,
differs from the earlier growth, Pl. ¢d., figs. 12-13, as Aviculopecten differs from Pter-
inopecten. The same stage also differs from the adult fig. 37, p. 333, as Aviculopecten
differs from Pecten. This later stage is therefore traceable to Aviculopecten as its an-
cestral representative, and also the position of Aviculopecten is clearly indicated in its
relation to other genera (see discussion and table, section Xv1).

‘The prodissoconch of Pecten as stated (p. 342) is referable in origin to the Nuculoid
radical, represented in the development of the Aviculidee and their allies as illustrated
by typical genera. It is further shown that in the development of a modern Pecten we
find in the first stages of dissoconch growth a form of shell, Pl. xxviun, fig. 9, presenting
characters which make it referable in ancestral origin to Rhombopteria a member of the
true Aviculide, later sueceeded by a growth, Pl. ¢d., fig. 12, bearing marked features
referable in origin to an ancestral genus Ptermopecten, which is transitional between
the Aviculide and the Pectinide. Still later a stage exists Pl. ¢d., fig. 2, which is ref-
erable in its inherited form to Aviculopecten, and finally, the true Pecten features char-
acteristic of the adult, fig. 37, p. 333, are established. The geological sequence of
these several groups is in the order indicated by the development of Pecten. We have,
therefore, a clear case of the ontogeny of an-individual illustrating the phylogeny of a
group (see table, section xv1).

As further substantiation of my position that the Pectinids were evolved from the
Aviculide, it may be noted that in the Challenger report, Vol. xu, part xxv, Professor

PHYLOGENY OF THE PELECYPODA. 347

Smith figures a deep-sea species, which he calls Pecten aviculoides, on account of the
close resemblance which it bears to Avicula, although he considers it a Pecten. Again,
a fossil from the Lias shales, as Professor Carpenter notes, was by one author named
Avicula and by another Pecten, the mixture of characters being such as to sanction its
being placed in either genus, according to the value attributed to different features.
Carpenter affirms that it is a true Pecten because of a coarse shell structure like Pecten
and an absence of prisms found in Avicula. Both these forms may properly be consid-
ered as representing in the adult a combination of types which I have shown is traceable
in early stages of the developing young.

‘Whereas very young specimens of Pecten tirradians are yellowish-brown, older young
and adults are commonly bluish-gray, the coloration often being less or wanting in the
furrows of the upper, left valve. The lower valve may be of the same color as the up-
per, but is generally less pigmented, especially in young individuals, and specimens are
frequent in which the upper valve bemg dark blue-gray, the lower is nearly or quite
white. This difference bears a direct relation to the amount of light to which the two
valves are exposed, the paler being the lower valve, which is constantly turned away
from the light. In studying other species of Pecten, many of which are highly colored,
we find that the same condition exists, the lower right valve being white, or of a paler
color than the left upper valve, which may be brilliantly colored: red, brown or yellow.
An example of a striking contrast is Amusium (Pecten) japonicum, in which the upper
valve is rich red, the lower, pure white. In the Spondyli, which are attached by the
right valve, the lower valve also is white, or paler than the upper left valve, which is
turned toward the light and is often highly colored. Fischer notes the difference in
color of the valves in Spondyli, and states that an abyssal species of Spondylus has both
valves blanched. This relation of the coloration to the amount of light received by the
two valves in bivalves is more or less noticeable in the lower and upper side of univalyes
and is briefly considered by Tryon. Dr. Dall (16) considers the allied subject of the
coloration, surface marking and thinness of the shells of abyssal species, in his interest-
ing conclusions published in the Report of the Blake Mollusca.

Pecten magellanicus Gmelin (P. tenwicostatus, Mighels). A number of young speci-
mens of this large species measuring from 2-6 mm. in height were dredged at several
stations by the U.S. Fish Commission Steamer Albatross. They are entered as No. 62307
in the Smithsonian Institution Catalogue. Sharply-defined but quite small prodisso-
conchs, like those figured in Pecten irradians, mark the umbos of all well-preserved spec-
imens. Fig. 38, p. 348 represents one of the smaller specimens. The early dissoconch
as indicated by the lines of growth is similar in form to Pecten irradians, Pl. xxvint,
figs. 9-10, no ears having developed and there is a deep byssal notch in the right valve
although it is slightly indicated in the adult. The byssal notch of the right valve, how-
ever, bears no teeth as in the other species described. The hinge line equals nearly the
whole width of the shell as in Pecten irradians, Pl. xxvun, fig. 12; but, as in that species,
differs in this character from the adult. The hinge area on either side of the cartilage
pit bears a wide hinge plate, sinuous in outline and marked by numerous flutings. The
flutings are nearly perpendicular to the hinge line, but are slightly inclined toward the
umbos. Although characteristic of the young, this hinge plate is wanting in the adult.

348 ROBERT TRACY JACKSON ON THE

A similar hinge plate exists in Pecten monotimeris, Con., and it is described by Dall (16)
in the young or adult of several species of Pecten, and in Pseudamusium and Dimya.
This fluted plate, Dall considers as a form of dentition augmenting the strength of the
hinge line (on the hinge of Pelecypods, etc., Am. Jour. Sc. Dec. 1889).

The right valve of Pecten magellanicus has a thin layer of well defined prismatic
structure, fig. 39, and the prisms are larger than those observed in other species of Pec-
ten. They are peculiar in being for the most part almost quadrangular, and are arranged
with extreme regularity parallel to the lines of growth; but where slight fractures have
induced irregular shell growth, they are polygonal as in other species of Pecten. The
right valve in the young specimens studied seldom shows radiating strize but they some-
times exist slightly emphasized in the last-formed portion of the shell, as series of dis-

Fic. 0. @onnected knobs. The ventral ear of the right valve,

Fag ANN pi on the contrary, bears coarse costz crossed by imbrica-

SS \ ting lines of growth. In the adult, the radiating ridges
of the right valve are much flatter and less accentuated
than in the left upper valve. The left valve, fig. 38,
after the earliest nepionic stages, which are smooth, is
marked by a stage of concentric and radiating ridges
both sharply defined and dividing the shell surface into
sunken squares. Knobs or bosses ornament the inter-
section of the ridges. The structure seen here is a com-
mon form of surface ornamentation among fossil Avi-
culidee. This peculiar marking of the early stage is
es Feit See eee very characteristic although limited in extent and ceases
plate byssal sinus, and marking of thelett Suddenly being succeeded by fine somewhat sinuous
valve characteristic of the period of de- radiating ridges which are retained thereafter, and
velopment. 18 diam. toe
Fra. 39.—Peclen magellanicus, prismatic Ve characteristic of the adult.
structure of right valve. X 90 diam. Some adult specimens of Pecten magellanicus, from
Fig. 40.—Peclen disiocaius, prismatic the Georges’ Banks, are interesting for a comparison
structure of right valve. X 90 diam. 5 ’ 5
Original specimens in Smithsonian In- with the observations on anatomical features of P. irra-
ae a (drawn by J- 1 dians. The species is well fitted for study on account
of the size, adult specimens measuring 15 em. or more
in height. The adductor muscle is double and the two portions are strongly accent-
uated. The borders of the two portions of the muscle are free on their limits and at
the point of insertion in the shell produce two well-marked impressions instead of a
single slight one as in Pecten irradians. There is no posterior retractor muscle of the
foot in this species; but it quite likely exists in the young, as that bears evidence in its
shell, fig. 38, of being an active crawler. The foot is marked by two clefts, the distal
one for crawling, the proximal one the seat of the byssal gland as in Pecten irradians.
The mantle as in that species, has a thick marginal perpendicular wall, decreasing at the
eared areas and thickly set with guard tentacles, extensible tentacles and eyes, indicat-
ing similar habits to those deseribed in Pecten trradians. The gills, intestine, and re-
productive body are also the same as may be observed in Pecten ¢rradians.
Chlamys (Pecten ) islandica, Chemn. sp., specimen No. 27294, Smithsonian Institution

\
-—

Fic. 39.

PHYLOGENY OF THE PELECYPODA. 349

catalogue, is a young individual 3 mm. in height. This specimen is figured in Pl. xxvit,
fig. 14, as viewed from the left side and Pl. xxvu, fig. 8, is the umbonal area more mag-
nified, viewed from the right side. The prodissoconch is sharply defined and the um-
bos are directed slightly posteriorly. It bears concentric lines of growth, no indication
of a byssal sinus and is homogeneous in structure. The succeeding dissoconch is yel-
lowish-brown and in its early stages has a well-defined, prismatic layer in the right
valve. In the oldest period of this young specimen no prismatic layer exists, the struct-
ure being lost thus early in shell growth. The left valve has a punctate structure as
mentioned in some other species. The early nepionic stage as discerned by following the
lines of growth is seen to be similar in form to the nepionic stage of Pecten irradians,
Pl. xxvi, fig. 9. The left valve is marked by a deep byssal notch, Pl. xxvu, fig. 8, the
lower border of which bears teeth similar to those of Pecten irradians and these teeth
exist also in the adult of Chlamys islandica. At the close of the smooth, early nepionic
stage of C. islandica, we find radiating plications similar to those of the adult originat-
ing at a definitely marked area, Pl. xxvin, fig. 14. New plications develop in succeed-
ing growth between those first formed, instead of the first formed being the only ones
built as in Pecten irradians. The hinge line is very long proportionately to the width
of the shell as compared with the adult condition; the same feature being characteristic
of the young of all the species described. The ears in this early stage are nearly equal,
whereas in the adult the ventral (anterior) ear is characteristically much larger than the
opposite or dorsal ear.

Pecten dislocatus, Say. A number of young specimens of this species are in the collec-
tions of the Smithsonian Institution, under the catalogue number 62278. They were
dredged off the coast of the Carolinas at several stations. The specimens vary from 3 to
13 mm. in height. All the youngest show a prodissoconch like that of Pecten irra-
dians, P|. xxvit, figs. 9-10. The prodissoconch bears fine concentric lines of growth,
has no indication of a byssal notch and is not prismatic; also the umbos are directed pos-
teriorly as in all this group. The right valve has in early nepionic stages a well-defined
prismatic layer, fig. 40, p. 348, the prisms being of about the same size as those in P.
irradians and Chlamys islandica, but smaller than those in Pecten magellanicus. Con-
versely, the left valve has no prismatic layer; but it has a very characteristic pitted struct-
ure in its early stages. The pits are not deep, but numerous and very closely related
to one another. They disappear with the close of the smooth nepionic stage. A similar
structure is described in some other species; but in none, unless P. teste, is it as notice-
able and clearly marked as in this species.

The early nepionic stage of Pecten dislocatus, as ascertained by following the lines of
growth, which are particularly well defined in this species, is seen to be of the same form
as that of Pecten trradians, P|. xxvu, figs. 9-10. No ears exist at this period and the
hinge line in length equals the entire width of the shell. A deep byssal notch exists in
the right valve originating with the close of the prodissoconch stage. In later growth
plications in the shell originate suddenly as in Pecten irradians, Pl. xxvun, fig. 12, caus-
ing by their introduction a marked stage as in that species. Hars also develop with suc-
ceeding growth and the hinge line gradually loses its great length proportionately to the

MEMOIRS BOSTON SOC. NAT. HIST., VOL. IV. 47

350 ROBERT TRACY JACKSON ON THE

width of the shell. The base of the byssal notch bears a number of teeth like those of
Pecten irradians and Chlamys islandica, but they are wanting in the adult.

Pecten teste. Yn the Jeffreys collection, now in the Smithsonian Institution, there are
two very young specimens of this species, catalogued as No. 62607. I sawthem in Wash-
ington, but did not borrow them as I did specimens of the above described species. From
a drawing made at the time (previous to my studies of other young Pectens) the upper left
yalve exactly agrees in form and is of about the same size as Pecten trradians, P|. xxvut,
fig.9. A prodissoconch like that of P.irradians and other species exists in my drawing
of the specimen. The early dissoconch shell has no ears and is comparable to ancient
forms of the Aviculidz as discussed in considering the same period of development of
Pecten irradians, with which it agrees in form as far as my drawing indicates. Minute,
punctate structure of the dissoconch was observed in this species similar to that noted
in Pecten dislocatus; but I cannot say whether prismatic structure exists although im all
probability it does.

Specimens of Pecten monotimeris, Con., recently received from California, have a sim-
ilar prodissoconch and nepionic stages to those described in P. irradians, as ascertained
by following the lines of growth. In the nepionic stage, the right valve has prismatic cel-
lular structure and the left valve punctate structure as in P. irradians, but as in that
species these features are early lost.

XI. GENERA ALLIED TO PECTEN.

Of the genera here considered I have had only limited opportunities to study the young ;
for, as conchologists commonly pay little attention to the young of molluscs, it is a diffi-
cult matter to obtain early stages of genera not living in our own waters.

Hinnites, Pl. xxv1, figs. 3-4,is a genus closely related to Pecten, but as it has been
described in section vu, I will give only brief supplementary notes at this point. It
is shown, p. 324, that HHinnites cortesi is pecteniform and regular when young; later it
solders one valve to a foreign body and with this change in habit assumes an irregularity
of growth rendering its valves, especially the attached one, comparable to Ostrea in form.
The form of the valves is ascribed to the condition of cemented fixation. The relation
of Hinnites to Pecten I believe is directly comparable to the relation of Ostrea to Perna;
the difference being that Hinnites in the nepionic and nealogic stages is free or byssated
and then retains the ancestral Pecten features. It does not become attached and assume
the ostrean irregular growth (wiping out early characters) until a comparatively late
stage of development, Pl. xxvi, fig. 3. On the other hand, Ostrea is attached at the
close of the prodissoconch period and at once assumes the ostrean form, by this means
eradicating features which might otherwise render the early dissoconch growth compara-
ble to the ancestral Perna or Perna-like form. The age at which fixation takes place
varies somewhat in individuals, as may be seen in series of specimens of [Hinnites gi-
ganteus. That species is instructive as showing how, in the development of the individ-
ual, the imbrications of the lines of growth in the nealogic pecteniform stage gradually
become more and more produced, until in the adult they form spinous productions of the

PHYLOGENY OF THE PELECYPODA. 351

valves. The spines of the left (free) valve are comparatively sharp and acicular, while
those of the right (attached) valve are flatter and spathulate ; though both forms of spines
may be serially traced to the imbrications of the young, where they are alike in both valves.

Spondylus as a genus of attached Pelecypoda naturally draws attention in these studies.
It is cemented by one valve and its irregular, inequivalvular shell, Pl. xxv1, fig. 1, has
been ascribed to the conditions of cemented fixation (vide pp. 318 and 323). The cam-
erated structure of Spondylus has been also considered and I wish here only to describe
the young.

In the collections of the Boston Society of Natural History are specimens of Spondylus
asperrimus, Sby., from the Pacific Islands, bearing the catalogue number 8494. The up-
per left valves of the specimens are produced on the hinge line as slightly developed ears
and have a decidedly pecteniform aspect, as is characteristic of this valve in living and
fossil Spondyli. The lower right valves which are or have been attached, are composed
of highly irregular, concentrically rugose and spinous shell growth, as is also characteris-
tic of species of this genus. One of the specimens of a right valve, however, is very well
preserved up to the initial stages of growth. At the umbo of this well-preserved speci-
men of Spondylus asperrimus, P|. xxvir, fig. 4, is a prodissoconch similar to that described
in species of Pecten. The prodissoconch is rounded or oval in outline and is clearly marked
off from the succeeding shell growth.’ It is not sufficiently well-preserved to show lines
of growth; but it has no byssal notch which exists, however, in the first stages of the
succeeding dissoconch growth, as in Pecten, Pl. xxvn, fig. 9. The first nepionicstage of
shell growth in Spondylus, as shown in the figure, is pecteniform. It hasa long hinge line
relatively to the width of the shell at that age, and a deep byssal sinus. As in the pro-
dissoconch most of the lines of growth of this early stage have been worn off. At the
close of the pecteniform stage the animal became soldered to a foreign body and the suc-
ceeding shell deposition, affected by the new condition of cemented fixation, has assumed
the irregular and rugosely distorted growth characteristic of young Spondyli. In the
first stages of the irregular growth, the byssal notch is soldered over, and eradicated in a
similar fashion to that shown in Hinnites, Pl. xxv1, fig. 8. A prodissoconch and nepi-
onic stage similar to that described in Spondylus asperrimus, has been observed also in
specimens of S. croceus, Chemn. and S. variegatus, Chemn. sp.?, at the New York State
museum; and in S. ducalis, Chemn. and S. varians, Sow., at the Peabody Academy of
Science, in Salem.

The above description of the development of the shell of Spondylus is directly com-
parable to what has been described in the shell of Hinnites. Both valves before fixa-
tion are pecteniform in these genera; as soon as attached the lower valve loses this early
existent, ancestral feature and assumes the irregularity characteristic of attached Pele-
eypoda with of course the peculiar order of irregularity normal to Spondylus or to
Hinnites. As the lower valve becomes irregular and distorted, the upper valve in both
genera, acting in sympathy, becomes irregular also; but, as previously claimed, the dis-

1A specimen ofa left valve of a young Spondylus, in the valve. It is rounded, with slightly developed umbos. Its
Smithsonian Institution collections, catalogue No. 62242, limits are marked off from the succeeding dissoconch as in
shows a prodissoconch like that described above in a right Pl. xxvui, fig. 4, of a right valve of 8. asperrinus.

302 ROBERT TRACY JACKSON ON THE

tortion of the upper free valve is seldom as great as the distortion of the lower cemented
valve. Specimens of Spondylus imperialis, Chemn., in the Society’s collections, show
the remarkable condition of being perfectly equivalvular. Both valves have ears on
the hinge line and though spinous, both are equally so. The right valve has no triangu-
lar hinge area or talon and the umbos of the two valves are closely related. Studying
these shells, I found no evidence that they had been attached during any period of their
growth. With this freedom from fixation we haye, therefore, an equivalvular shell which
is a wide departure from the typical condition of the genus.

The conclusion arrived at that the free valve of attached, cemented Pelecypods is
least modified and therefore retains more of the ancestral characteristics, which are nearly
or quite eradicated from the cemented valve (section v1), thus finds strong support in
Spondylus. In many figures of fossil Spondylus, published by MM. d’Orbigny and Pictet,
(also see Pl. xxv, fig. 1), we find striking dissimilarity in the two valves, the attached
one being always the most modified. The free valve which is least modified retains in a
marked degree, especially in the young, characters which are traceable to the Pectini-
dx, from which the Spondyli have doubtless evolved. Ihave not dissected a Spondy-
lus; but from published notes gather that this genus is closely connected with Pecten by
anatomical features.

Plicatula is a genus placed next in kinship to Spondylus, the anatomy being similar
according to authors. Some very young shells of Plicatula marginata, Say, Pl. xxvu,
fig. 5, from the Miocene Tertiaries of North Carolina present features of interest. A
distinct prodissoconch, similar in form to that of Pecten and Spondylus, exists. It is
rounded with slightly developed umbos which are directed posteriorly. The specimens,
as shown in the figure, bear a considerable resemblance to young oysters of a similar
age; and as it may be thought that a mistake has been made in the identification, I may
say that they show the typical teeth of Plicatula, also series of specimens have been traced
from the very young to the mature condition. The prodissoconch umbos in specimens of
Plicatula, when attached to the object of fixation, Pl. xxvny, fig. 5, point to the right of
the observer viewing them from above, whereas in Ostrea, Pl. xxv, figs. 1-5, they point
to the left, which is due to the fact that Plicatula is attached by the right valve, where-
as Ostrea is attached by the left. In both genera, the prodissoconch umbos are directed
posteriorly relatively to the anatomy which is known directly or by analogy to have ex-
isted at the close of the prodissoconch period.

The early nepionic stage of Plicatula, Pl. xxvu, fig. 5, is very similar in appearance
to a young oyster, as stated. The lower right valve is flat and closely related to the
object of fixation, while the upper left valve is convex.' No byssal notch is discover-
able in the earliest stages of dissoconch growth, and this feature, together with the
form of the valves, indicates that Plicatula became permanently attached at the close of

1It has been observed by authors, that the upper valve
of fixed Pelecypods, such as Ostrea and Anomia, also uni-
valves, as Crepidula, frequently simulates the outlines of
the object of support, ridges, knobs, etc., on the foreign
body, being perfectly reproduced in the free valve. The
most striking case of this, I have seen, is in a finely pre-

served specimen of Plicatula marginata, from the Tertia-
ries, in which the free valve reproduces with beautiful
exactness the structure of a Bryozo6n oyer which the at-
tached valve was growing. Not only a general copy is
made, but each cell is outlined in the upper valve of the
Plicatula.

PHYLOGENY OF THE PELECYPODA. 300

the prodissoconch stage, as is conclusively shown to be the case with oysters. Adult
Plicatula marginata, is deeply plicated; but the plications, as in plicated oysters, do not
originate until the shell growth extends beyond the limits of the object of fixation.

Plicatula, taken in connection with the preceding genera, is a valuable proof of the
correctness of the conclusion that cementation destroys indications in the shells of Pel-
ecypoda of their ancestral relationship. Pecten, Hinnites, Spondylus and Plicatula may
be serially connected in the order indicated by a study of the shell, and from what I can
gather, the anatomical features would present a similar series. Using some license so
as to express it graphically, one may say: solder a Pecten when partly grown and the
succeeding growth is irregular, forming the genus Hinnites, Pl. xxv1, figs. 3-4; solder a
Pecten at a much younger period and the succeeding growth is irregular, forming the
genus Spondylus, Pl. xxvu, fig. 4; solder a Pecten at the close of the prodissoconch
stage, before any dissoconch growth has taken place, and the succeeding growth is ir-
regular, forming the genus Plicatula, Pl. xxvu, fig. 5. In all, the irregularity of growth
commences with the introduction of the cemented period. Spondylus is probably a branch
from Pecten through the intermediation of Hinnites, which unquestionably is closely
akin to Pecten. So near are Spondylus and Hinnites, that Stoliczka says some fossil
forms, first considered as Spondyli, have since been transferred to Hinnites.

Brief studies of Lima haye been made. Lima has a perpendicular mantle wall like
that of Pecten which is shown to be so actively functional in swimming, and it probably
subserves the same purpose in Lima, as the swimming of that genus, described by Dr.
Jeffreys (see Tryon), is performed similarly to that of Pecten. This difference exists,
however; Lima swims with the plane of the edges of valves perpendicular, whereas in
Pecten they are horizontal while swimming. From the shape of the mantle walls and
shell, also from the description of the swimming movement, it appears that in Lima the
excurrent flow of water which propels the animal all takes place through one eared area,
the ventral, instead of alternately from either ear as in Pecten. In the species of Lima
examined, there were ho guard tentacles on the free borders of the mantle walls as in Pec-
ten, Pl. xxvu, fig. 4, but Jeffreys’ description of Lima includes such organs. The margi-
nal tentacles at the base of the perpendicular wall are thick and numerous. Dr. Jeffreys
describes them as twining actively and presenting a tenacious prehensibility, grasping the
fingers when approached toward them. This prehensibility I have described in Pecten
though it is there less marked in degree. The gills of Lima have the reflected portion of
the filaments equal in length to the direct portion, and the tip of the reflected portion
of each lamina is again reflected, presenting an abbreviated additional lamina.’ The gills
of Lima differ from those of Pecten in that the separate filaments are joined by bars
which form a direct concrescence of the several filaments instead of by the interlocking
of ciliated processes, as in Pecten; and in this character Lima agrees with Ostrea.

A young Lima elliptica, in the Jeffreys’ collection at the Smithsonian Institution, shows
a well-marked prodissoconch. In a drawing which I made at that museum, Pl. xxv,
fig. 6, the prodissoconch is sharply marked off from the succeeding dissoconch, and it

‘In Pecten irradians, the tip of the reflected filaments Lacaze-Duthiers also observed such a brief reflection in
is bent back on itself slightly, Pl. xxrv, fig. 12, indicating the outer filaments of Anomia.

the initial stages of such an abbreviated additional lamina.

354 ROBERT TRACY JACKSON ON THE

has well-developed umbos. The succeeding dissoconch is limaform, but no prismatic
structure was cbserved as in young Pectens; it may, however, exist in well-preserved
specimens.

XI. Anomia. ANATOMY AND HABsItTs.

Anomiz freely attached themselves to the glass slides in my drain-pipe traps and
therefore I had an opportunity to study the young under this favorable condition with
transmitted light. The species is Anomia glabra, Verrill, figs. 41-42, which is abun-
dant in the shoal waters of Buzzards Bay, often covering shells and stones as thickly as
do oysters.

Anomiz haye not in the same degree as oysters the power of resisting drying which
is so characteristic of that mollusc, and rarely set at points which are long exposed at
low tide. They grow rapidly and some speci-
mens attached to glass slides measure 25 mm. in
height, the result of the summer’s growth. These
almost equal the adult in size as the latter rarely
exceed 40 mm. Examining Anomiz where a num-
ber are attached to a single rock or shell, it will
be found that the longer axes of the specimens in
most cases are nearly parallel, the reverse of the
condition of oyster spats, where the axes point
at every angle in the plane of the surface of at-

Fig. 41. Fic. 42.
Fig. 41. Adult Anomia glabra viewed from the tachment. This parallel arrangement in Anomia

right, lower side; natural size. seems to bear a relation to the direction of strong-

Fig. 42. The same in profile view. (From Bin-

Tove onTe.) est currents, the animals arranging themselves

during their early locomotive stage in the plane
of least resistance and afterwards becoming permanently fixed in this position. On the
other hand, oysters which have no crawling locomotive stage, pay no attention to prevail-
ing currents in their position of fixation. A marked case of animals arranging them-
selves with regard to currents was observed in Crepidula fornicata, Pl. xxi, fig. 22.
Large numbers of colonies were found at Buzzards Bay like the figure in which all the
individuals were headed one way. This arrangement in one plane does not always hold
in Crepidulas as they may be arranged at different angles; in such cases we may rea-
sonably suppose that no strong prevailing current existed in the location where the ani-
mals grew.

Anomia was formerly considered as related to Brachiopods, but such a connection has
long been disproven. Forbes and Hanley and Lacaze-Duthiers compared Anomia to
Pecten; but notwithstanding this, a. close relationship of the genus to Ostrea has been
as a probability quite generally entertained. This assumption appears founded on a
certain similarity of the shell of Anomia to an oyster, especially a young oyster. Anomia
belongs to the same broad group as the oysters, as proven by features of anatomical
and shell structure; but it is more nearly related to Pecten than to Ostrea. Previous
studies have shown that the form of the shell in Peleeypods is often highly modified by

PHYLOGENY OF THE PELECYPODA. 355

special conditions, as the cementation of a valve, or by anatomical features and habits of
the individual. I think therefore that the peculiarity of the shell in Anomia may be con-
sidered as an adaptation to its environment’ and not due to any genetic connection of the
group. On this question we have the support of the conclusions of Professor Lacaze-
Duthiers, who in his well-known paper on Anomia ephippiwm, L., says:“ . . .
Yossicule de ’Anomie est un byssus, et toutes les anomalies sont la conséquence de la
position de animal sur la cété droit, et de la soudure du byssus aux corps étrangers.”
His studies were anatomical, based on the ambiguous adult and it is my pleasure as a
result of studies of the young, to be able to show conclusively the correctness of some
of the deductions of the eminent French investigator.

The youngest specimens found alive on glass slides were about the size of Pl. xxrx,
figs. 5-6. At this age and up to a considerably later period of growth, the young are
freely locomotive, crawling actively at intervals by prolongation of the highly developed
foot. The lines of growth in fig. 6 indicate that, when very young, the byssus was on
the extreme margin of the valves; but in this figure the byssal sinus has already become
sub-central in position and partially enclosed by the growth of the right valve. At this
age the byssal fixation is only maintained a part of the time, or perhaps more correctly
speaking the byssal attachment may be said to be voluntarily and frequently abandoned.
The byssus of this early period is composed of aseries of bundles of threads which flatten
against the surface of the glass as seen from the right side through the glass to which it
is attached. The byssal hold at this age is an act of volition on the part of the animal,
as if paralyzed with cocaine they may be easily washed off a glass slide by a stream from
a pipette, whereas in the normal condition they cannot be dislodged by any force short
of scraping. The cocaine treatment is a good method to employ in collecting very young
Anomiz attached to shells, rocks, etc., because if scraped off, the lower valve, which, in
this species, is of extreme tenuity, is almost certain to be destroyed. When older the
byssal attachment becomes permanent by the calcification of its tip.

A young Anomia viewed from the left side is shown in Pl. xxrx, fig. 1,and the reverse
or right side of the same individual seen through the glass to which attached is shown
in fig. 2. The relations of these views to one other and to the adult described by La-
caze-Duthiers are of much interest. The single adductor muscle of the valves, ad,
figs. 1-2 (also in the diagram PI. xxrx, fig. 3), is situated a short distance below the
heart, 2. It is very small relatively to the size normally found in Pelecypods, a fact hav-
ing important bearings. This is the only muscle indicated in the left valve which passes
to the right valve or which finds its counterpart in that valve and the same condition ex-
ists in the adult. A large byssal muscle, bm, fig. 1, inserted in the left valve passes to
the foot and byssus, the calcified attachment of which is shown at dbs, fig. 2. The byssal
muscle at its insertion in the left valve has a peculiar sinuosity in its posterior face
which is not understood. A posterior retractor of the foot, pr, figs. 1 and 3, is attached
to the left valve a short distance from the byssal muscle and close to, but separated from,
the adductor. It is somewhat remarkable that the posterior retractor lies on the ventral
border of the adductor, as the typical position of the posterior retractor in Pelecypods

1In Anomia, the shell itself is not cemented to foreign is by means of the calcification of the byssal tip, and is
bodies as in the ostrean forms considered; the attachment not to be confounded with the cementation of one valve.

356 ROBERT TRACY JACKSON ON THE

is on the dorsal border of the adductor, as in Mya and Modiola, Pl. xxv, figs. 9-10. The
fibres of the posterior retractor pass forward to their insertion in the foot and merge
with the byssal muscle as distinctly seen from the left side. An anterior retractor of
the foot, ar, figs. 1-3, is inserted in the left valve near the umbonal region. Whereas in
the young the adductor, posterior retractor of the foot and byssal muscles are distinct
and separate, in the adult they are rendered ambiguous by being in close contact with
one another, although making three separate scars in the left valve of the shell. The
true nature of the muscles of the adult, as described by Lacaze-Duthiers, is therefore
seen to be clearly apparent in the young.

In young Anomia attached to glass slides where the action of parts may be observed
through the very translucent valves, it is noticeable what an active part the byssal and
retractor muscles of the foot take in pulling down or closing the upper, left valve, on ac-
count of being attached distally by means of a calcareous union to the object of support.
This is rendered clear by comparing Pl. xxrx, figs. 1-2 with the diagram fig. 3. By the
action of these muscles the upper left valve is brought in close contact with the object of
support, so that it conforms to the surface as does the shell of a Crepidula, or limpet, under
somewhat comparable conditions of fixation. In correlation with this changed function
of the retractor muscles of the foot, and byssal muscle in Anomia glabra, it is important
to notice that the adductor muscle is very small and the lower right valve extremely thin,
a mere veil of tissue in the young, so that it would be almost functionless for protection,
if the young could he exposed to attack from this side, which however is not the case.
It is thus evident that the byssal muscle and the anterior and posterior retractors of the
foot in the young are together pseudo-functionally an adductor in their action and as a
correlation a reduction of the actual adductor and lower valve has taken place. This con-
dition of affairs is evidently brought about by the position of the parts relatively to the
surroundings and is ascribed to them. Lacaze-Duthiers ascribes the absence of retrac-
tor muscles of the byssus and foot on the right side to atrophy and loss from disuse.
The great size and amalgamation of the byssal muscle and posterior retractor he as-
cribes to the changed directions of the fibres and the functional activity in the changed
position, perpendicular to the object of leverage. He may have intended the compari-
son of the muscles to an adductor which I make, but it is not so understood from his
text.

The young Anomia opens its shell by the lifting of the upper valve alone and it has
considerable forward and backward as well as lateral motion due to its peculiar method
of attachment. When quietly feeding, the anterior retractor of the foot, ar, Pl. Xxrx,
figs. 1 and 3, may shorten and this muscle being inserted at the byssal region of calca-
reous fixation the animal responds by moying backwards; conversely, if the posterior re-
tractor, pr, figs. 1 and 3, shortens, the shell is moved forwards, as will be understood
by considering the diagram, fig. 3. In the adult there is little or none of this forward
and backward motion, as the byssal plug is closely surrounded by shell growth and also
the position of the posterior retractor is changed so that it could hardly bring about
such a movement.

The foot in adult Anomia glabra is a small reduced organ; but in the young it is very
large and active, Pl. xxrx, figs. 1,2 and 5. The base of the foot is marked by a deep

PHYLOGENY OF THE PELECYPODA. 357

cleft up to its extreme distal portion and the very young, fig. 5, in crawling extends and
flattens it against the object of support in a prehensile disc-like fashion, as described in
Pecten irradians. After the animal becomes permanently fixed, figs. 1 and 2, the foot
was not observed to extend beyond the margins of the valves; but it was constantly moved
within the mantle walls in a sinuous manner. The foot is of large volume embracing
the byssus at its point of attachment to the glass slide, fig. 2, and the open cleft of the
foot makes a V-shaped space immediately in front of the byssus; but soon the walls of
the cleft are in contact. The byssus is thus seen to originate from a cleft in the base of
the foot as in ordinary byssated Pelecypods.

The upper proximal face of the calcareous plug of adult Anomia is marked by lines
parallel to the longer axis of the shell and similar lines are noticeable in the basal portion
of the plug, bs, where attached to the glass, Pl. xxrx, fig. 2. The lines are made up of
parallel series of amalgamated, calcareous, discoid points seen in the fore part of the
byssal attachment in the figure. The discoid points are directly comparable to the discoid
points of fixation of separate byssal threads observable in Pecten, Mytilus, ete., the differ-
ence being that in Anomia the several points are calcified. In the young we have there-
fore the fullest substantiation of Lacaze-Duthiers’ homology of the plug of Anomia with
the byssus of other Pelecypoda; for we can actually see the byssus coming from the cleft
in the foot and the separate byssal threads amalgamating to form the calcareous plug.

The palps of Anomia, Pl. xxrx, figs. 1-2, originating close to the umbos pass to the
right of the anterior retractor muscle and to the left of the foot and terminate in close re-
lation to the anterior limits of the gills. They are marked by plicated ridges on their pos-
terior borders. The coils of the intestine were not followed, being hidden in the dark
brown liver follicles which cover the greater part of the body mass. The liver presents
two regions of densest portion separated by a lighter portion in which there are few folli-
cles as shown in the figures. The rectum passes over the dorsal border of the adductor
muscle and is produced to a considerable extent into the cloacal chamber, the anus, a, being
free, figs. 1-2, and in the adult a similar condition exists. A highly developed cecum,
¢, fig. 2, as in the adult, is prolonged in a curved line from the body mass past the ven-
tral border of the adductor and to the right of the gills ending blindly in the cloacal cham-
ber. It is clear in the young as it is not complicated by the existence of the reproduc-
tive gland found in the adult. 3

The heart, /, consists of a ventricle and probably two auricles as‘in the adult,! but the
heart in the young is surrounded by a pericardium, whereas in the adult Lacaze-Duthiers
says that this is wanting. He expresses the belief that the pericardial walls may have be-
come united with the walls of the heart proper and in view of its early existence this is a
reasonable explanation of its disappearance. On the left side a large vessel, v, fig. 1, passes
from the heart in front of the byssal muscle and to the gills of the left side; another vessel,
v, passes between the adductor muscle and retractor of the foot. These are similar to
those of the adult after Lacaze-Duthiers. Other vessels ramify in the mantle as described
in young oysters. The pulsations of the heart are variable. In fig. 1 were counted eighty

1T express this with a degree of caution, as I did not parts being so coincident with his figures and description,
actually observe in the young the second auricle described its presence is assumed ; as he says it is with difficulty seen
in the left side of the adult by Lacaze-Duthiers ; but other in the adult.

MEMOIRS BOSTON SOC. NAT. HIST., VOL. IV. 45

358 ROBERT TRACY JACKSON ON THE

to eighty-four heart-beats per minute in several counts. At other times in the same
specimen forty-four, forty-two and twenty-one beats were counted; but these were much
below the normal. In a younger specimen sixty beats were recorded at several counts.
Watching the heart it may be seen to stop for a considerable time and then begin again
first beating slowly ; but soon getting up to its normal rapidity. The cessation of pulsa-
tions may be induced by disturbing the animal and the pulsations are most regular and
active while the Anomia is quietly feeding with the mantle extended.

The gills in very young Anomias the age of P]. xx1x, fig. 5, are simple, straight filaments
not joined to one another in any way and the tips not yet recurved upon themselves.
(Compare with young oyster gills, Pl. xxrv, fig.6.) Dissecting specimens of the age of
Pl. xxrx, fig. 1, it was found that the inner pair of gill plates had their filaments reflected
dorsally and the filaments of the opposed lamellze were joined to one another by the con-
crescence of their dorsal borders. The two outer gills, on the other hand, had their fila-
ments reflected upon themselves; but the recurved part equalled only about half the
length of the direct part. The same relation of filaments exists in the adult gill diagram-
matically shown in Pl. xxtv, fig. 11. The individual gill filaments in specimens of the age
of fig. 1, as also in the adult, are joined to one another by concrescence at the tips of the
dorsally reflected borders; but they are not joined to one another at other points. Further
the filaments have not the nodular ciliated processes characteristic of the gills of Pecten.
Therefore, the gills of Anomia have filaments which are individually free to a somewhat
remarkable extent. This is interesting as showing an almost primitive condition in one
organ when other organs are so modified as to render Anomia one of the most highly
specialized of the Pelecypoda. The gills are connected with the body by two membranes,
s, Pl. xxrx, figs. 1-2, which have been described (Tryon) as falciform. These membranes
are attached to the adductor muscle, to the body anteriorly from that point and to the
mantle wall posteriorly to the adductor. The posterior attachment is not shown in La-
caze-Duthiers’ figures; but it is a marked feature in the young of Anomia glabra. A
large foramen exists in each of these membranes. The foramen of the left membrane, fig.
1, is the larger and my attention was attracted to it by seeing the extensible and restless
foot prolonged across the body and thrust through this hole. In fact it was a favorite act
with this individual at least, for after noticing it I frequently found the foot thrust through
the foramen. The foramen in the right gill membrane, fig. 2, is smaller than that in the left.
The membranes extend in a free curving line to the posterior extremity of the gills. The
gills are capable of considerable motion and frequently in the young may be seen pro-
truded beyond the margins of the valves and even beyond the borders of the extended
mantle lobes as well. The length and width of the several gills is different as pointed out
by Lacaze-Duthiers. In comparing Pl. xxrx, figs. 1-2, itis seen that the gills on the left
side are much longer and broader than on the right side on which the animal rests, a fea-
ture which, as Lacaze-Duthiers said, is doubtless connected with the conditions of fixation.

The mantle of Anomia presents peculiarities which are necessitated by the abnormal
position of the byssus. ‘The mantle lobes are free except for a short distance on the liga-
mental area (observed in an adult) where the lobes of the two sides are confluent. The
right lobe has a deep re-entrant fold where it surrounds the byssus, shown in Lacaze-
Duthiers’ figures and slightly indicated in Pl]. xx1x, fig. 2. The margins are fringed

PHYLOGENY OF THE PELECYPODA. 359

with active extensible tentacles which are in constant motion. The tentacles of the young
are pigmented with black in scattered spots and bands of color which are arranged at
right angles to the longer axis of the tentacles. I did not observe radial muscles in the
mantle of Anomia similar to those described in Ostrea and Pecten, and if they exist as
they probably do, they are inconspicuous.

XII. Anomra. THE SHELL.

Professor Morse published a brief account of some extremely young shells of Anomia
in which he showed that an early, probably “dimyarian” shell precedes the characteris-
tic Anomian form of shell. He also showed the initial stages of the byssal notch. The
specimens studied, judging from his description and a comparison with my own obsery-
ations, were the young of was Anomia aculeata, Gmelin.

In Anomia glabra an ovally rounded, nearly equivalvular prodissoconch exists at the
umbos of perfectly preserved young specimens, PI. xxrx, figs. 4-6. The prodissoconch
in this species is opaque, of homogeneous structure, bearing fine concentric lines of
growth, and the umbos which are nearly central are directed slightly toward the poste-
rior end of the shell." In this position of the umbos my observations differ from those
of Professor Morse as he says the umbos are nearer the anterior margin of the shell.
The left prodissoconch valve of Anomia glabra has no indication of a byssal notch but
the right valve has a notch on its ventral border as shown by the lines of growth in PI.
XXIx, fig. 6. It is therefore seen that at this extremely early period Anomia is already
capable of forming a byssal attachment, and crawled on the right side with the foot ex-
tended at this area. The shell evidently lay on one side as observed by Morse. A
byssal notch in the prodissoconch is a point in which Anomia differs from the byssated
Pelecypods which have been so far discussed. In Perna, Avicula, Pecten and Spondy-
lus, Pl. xxvii and xxvii, a byssal notch is indicated in the first stages of the dissoconch
growth; but no trace of a notch exists in the prodissoconch.

The dissoconch growth of the left valve of Anomia glabra in its early stages, Pl. xxrx,
fig. 4, is at first nearly round, but slightly larger on the ventral side. It is thin, translucent,
has a highly polished surface and bears numerous, delicate lines of growth, but shows no
prismatic structure. An older specimen of a left valve, PI. xxrx, fig. 5, is less rounded, its
faster ventral growth having produced a bulging in this direction which seems to be char-
acteristic of specimens of this age. Whereas in fig. 4,the prodissoconch is on the anterior
border of the valve and projects slightly beyond it, in fig. 5, shell growth is taking place
in front of the prodissoconch, which is the primitive umbo of the valve, and in succeeding
growth this feature is carried still farther as seen in fig. 8, so that in the adult the ini-
tial umbo is pushed in quite far from its original position on the anterior border of the
valves. As growth continues from the stage fig. 5, the shell becomes more and more
rounded, rapidly assuming the form characteristic of the adult. The young is so highly
polished that the shell well deserves the popular name silver-shell which it bears.

The right valve at a very early period, Pl. xxrx, fig. 6, differs much from the adult

1The umbos are directed posteriorly so slightly in Ano- Pecten, etc., in which the umbos of the prodissoconch are
mia that it is not a marked feature as in Ostrea, Perna, more inclined toward the posterior region of the shell.

360 ROBERT TRACY JACKSON ON THE

and is less highly modified. The initial stages of dissoconch growth commence on the
free border of the prodissoconch valve; but succeeding shell deposition is extended pos-
eriorly and then ventrally more and more, as may be seen by following the lines of
growth, figs. 6-7, until finally a deep almost enclosed sinus is formed around the byssus,
as shown in the figure. At this stage it is desirable to compare the right valve of Ano-
mia, fig. 6, with my figure of the first stages of dissoconch growth of Pecten, Pl. xxvmt,
fig. 10. In both, a deep and somewhat comparable notch is formed by the existence in
one case of a byssus and in the other of a highly active foot; both organs affecting the
form of the valve by causing a local retraction of the mantle and thus putting an obstacle
in the way of shell deposition at that area. We see in the comparisons of these figures
the verification of the statement made by Forbes and Hanley nearly fifty years ago that
“the perforation in one of the valves of Anomia is chiefly a greater extension of the
auricular sinus in Pecten.” In later growth the ventral border of the valve surrounds
the byssal notch in an increasing degree until soon it is completely enclosed. The lower
valve is peculiar in that the mantle border of the umbonal region deposits shell around
the prodissoconch, enclosing it by degrees and partially covering its umbonal portion
as shown in PI. xx1x, figs. 6-7. This is due to the fact that the primitive umbos recede
from the shell border, being pushed back by the peripheral growth of the valve as seen
viewed from the left side in Pl. xxrx, figs. 4,5 and 8. When the byssus in young An-
omia is nearly or quite surrounded by the growth of the right valve as in PI. xxrx, fig. 2,
the foramen is very small. In the adult, fig. 41, p. 354, on the other hand, the foramen
of the right valve is many times larger and to accomplish this increased size of the hole
an extensive resorption of the shell has taken place. During this resorption, lines of
growth which were continuous when young become separated and may be traced on
either side of the foramen, as in the figure cited.

The lower right valve in the young does not equal the size of left upper valve which
overhangs its border by a considerable amount, Pl. xxrx, figs. 6 and 2; it is also of ex-
cessive thinness and so delicate that great care is required in handling not to injure it.
The right valve shows the most beautiful, prismatic structure of which it is entirely com-
posed in early stages, and it is so transparent that the organs can be seen as through
a glass film. The extreme reduction of the lower valve in young Anomia glabra is ev-
idently correlated with its conditions of environment. The byssal and retractor muscles
are functionally active in drawing down the upper valve, and when that valve is closed
the lower valve is completely protected. The reduction of the true adductor muscle as
described and the thinness of the right valve are doubtless correlated; for, on account of
its delicacy, the strong action of a normal-sized adductor would certainly break the valve.
At this early period the right valve truly seems like a disappearing organ. In later
growth the right valve increases in thickness and frequently extends beyond the object
of support, but it does not acquire the thickness and solidity of the left valve.

T have stated that the right valve in young Anomia glabra is wholly composed of
prismatic cellular tissue. This structure predominates throughout life in this valve, and
in the adult the right valve entirely consists of prismatic cellular tissue with the excep-
tion of a thin layer of whiter porcellaneous tissue which surrounds the byssal foramen
and point of insertion of the adductor muscle. The left valve, on the other hand, does

PHYLOGENY OF THE PELECYPODA. é 361

not have a prismatic layer at any period in the life of the individual; but is sub-nacreous
excepting for a small area around the point of insertion of the muscles, where as in the
right valve there is a deposition of porcellaneous tissue. This difference in the structure
of the two valves is a remarkable feature and occurs in related genera, for I have shown
that a thin prismatic layer exists in the right valve of young Pecten and in the right
valve of young Ostrea but not in the left valve of either genus in early stages.

T have not found any published observations of prismatic structure in_Anomia glabra,
but Carpenter, in his second paper (1848), on shell structure, stated that a thin external
prismatic layer exists in Anomia ephippium; but is early worn off. In this connection,
an interesting point is recorded in one of Dr. C. A. White’s papers,’ that among Tertiary
fossils it is rare to find a right valve of Anomia although left valves are abundant. He
ascribes the loss of right valves to their fragility owing to their prismatic composition.
On the other hand, a number of species of Huropean Tertiary Anomias in the collections
of the Museum of Comparative Zodlogy
have both valves equally preserved. A very
thin layer of prismatic tissue exists in the
right valve of Anomia ephippium, as ob-
served in specimens kindly sent me by Dr.
Dall; but it does not exist in the left valve of
that species, both valves being composed
wholly or for the most part of sub-nacreous
tissue, with the addition of porcellaneous tis-
sue around the muscular insertions as in _An-
omia glabra. The difference in the structure
of the right valve of Anomia glabra makes
a good specific distinction, which separates
that species clearly from _A. ephippium, which =
it approaches very nearly in external form. Fic. 43. Young Anomia aculeata, tip of left valve,

Young specimens of Anomia aculeata, Pp, “Prodissoconch, Buceeceeds ie ee senlas and tate
Gmelin, se 43, avedollectedisity Cohasset eae ee eda Cohasset. X 60 diam. (Drawn
and Beverly, Mass. A yellowish prodisso-
conch exists at the umbos. The left valve of the prodissoconch has a slight sinuosity
which corresponds with the comparatively deeper byssal notch in the right valve, whereas
no sinuosity of the left valve and a shallower notch in the right exists in the prodisso-
conch of Anomia glabra, Pl. xxrx, figs. 4-8. This feature is shown in fig. 45, but is
more marked in degree in younger specimens. The prodissoconch of A. aculeata is
ovally rounded in outline with fine concentric lines of growth and umbos which are nearly
central but directed slightly posteriorly as in A. glabra. The succeeding dissoconch
growth of Anomia aculeata is chalky white, of somewhat irregular growth; and not pol-
ished or as translucent as is A. glabra. The left valve of the dissoconch does not per-
petuate the sinuosity of the prodissoconch valve. This valve is smooth for a brief period
and then the plications and spines characteristic of the species are commonly introduced.
I say commonly, for associated specimens are frequently wanting in the spinous charac-
ter, although unquestionably belonging to the same species. The right valve of the dis-

YN

\\as
Timi

} Non-marine Fossil Mollusca of North America, Washington, 18838, p. 16.

362 ROBERT TRACY JACKSON ON THE

soconch of A. aculeata has a growth similar to that described in A. glabra; but is chalky
white and much thicker than in that species. The right valve is wholly or for the most
part composed of prismatic cellular structure whereas that tissue is wanting in the left
valve as in A. glabra. My observations of the young and adult lead me fully to concur
with Professor Verrill who says that he believes this species is perfectly distinct from
Anomia glabra.

In regard to the affinities of Anomia, it is closely related to Placuna and many inter-
mediate forms may be traced as noted by authors. The anatomy of Placuna is essentially
like that of Anomia; but the mantle border has a curtain like Pecten (Tryon). The
shell of Placuna is composed of fine nacreous layers, but in the young I have found
traces of prismatic cellular tissue in the right valve. This tissue was very thin and
limited in extent on account of being worn away, but existed in patches. Woodward (79)
mentions that, when young, Placuna has a byssal sinus in the right valve and I have found
young specimens which show an enclosed byssal plug like that of Anomia, as described
in section xvi. Anomia, anatomically and in shell structure, is much more nearly allied
to Pecten than to Ostrea. It is probable that Anomia was derived from the Pectinids
from the Amusium or Hemipecten group, for the latter resemble Anomia in having thin
nacreous shells. In considering Hemipecten, Zittel remarks that Jurassic species of Pla-
cunopsis are probably confounded with Hemipecten, thus inferring a close similarity in
the two genera. Placuna was probably derived from the early Anomiadz as all evi-
dence points toward an ancestry in that group. The geological appearance of these
groups accord with their inferred zodlogical affinities as I have described them. That
such a highly modified form of shell as Anomia could have been derived from the Pee-
tens is more easily believed when we see how highly modified are Pedum, Hinnites and
Spondylus which are more direct derivatives of Pecten.

XIV. Srupies oF A FEW OTHER GENERA.

The genera so far studied are all linked in one great group of the Aviculidz and their
allies (see table, section xvr), by characters of anatomy and shell structure which connect
the several members. The genera thus referred to are Ostrea, Gryphzea, Exogyra, Perna,
Avicula, Monotis, Pseudomonotis, Cassianella, Vulselia, Pecten, Spondylus, Plicatula,
Lima, Anomia and Placuna. I now wish to consider the young of some genera belong-
ing to other groups of Pelecypoda distinctly removed from those previously considered
and not always closely connected with one another. Figures of the adults of most
of the species described in this section are to be found in the works of Verrill and
Gould.

The groups, previously studied, have been characterized by possessing prodissoconchs
of homogeneous laminar structure, but not prismatic, and with umbos directed posteri-
orly. The succeeding dissoconchs have been shown to be markedly different in form
from the prodissoconch, and in the earliest stages at least are characterized by well-de-
fined, prismatic, cellular structure, either in both valves or the right valve.’ In all gen-

1 The change from homogeneous to prismatic structure do show it, that the absence of the feature in the few
was not observed in the fossils described or in Spondylus specimens observed need not be considered in this connec-
and Lima, but these genera are so united with groups that tion.

PHYLOGENY OF THE PELECYPODA. 363

era now taken up, excepting the Unionid, two marked features of difference occur as
compared with those previously studied. The prodissoconchs are directed more or less an-
teriorly, and the succeeding dissoconchs are not characterized by prismatic cellular
structure.

The Pelecypoda is a class in which it is very difficult to find binding characters on
which to make systematic divisions. The direction of the prodissoconch umbos and the
condition of prismatic structure in the succeeding dissoconch, I am aware, is not wholly
confined to the Aviculids and their allies, but these features are characteristic of the
whole of that group and are not together characteristic of any other group so far as
known, excepting, possibly, the Unionidz. A thin, but well-defined, prismatic layer ex-
ists in Pandora rostrata, Lam.; a prismatic layer also exists in the deeper-seated portion
of the periostracum of Solemya mediterranea, Lam., and it has been observed in limited
patches in young Hnsis viridis, Say. Carpenter observed the prismatic layer in the two
former genera and mentions finding traces of it in other genera, but he was impressed
with the fact that it is a marked feature in the group of the Aviculidze and allies.

Mytilus edulis, L. This molluse belongs to one of the oldest families of the Pelecy-
pods extending back to early paleeozoic times. It is therefore interesting to know the
conditions existent in the young shell. In Pl. xxrx, fig. 9, it is shown that a well-
defined prodissoconch, p, exists which is yellowish-brown and marked by fine concentric
lines of growth. The form of the prodissoconch is evenly rounded and it has well de-
veloped umbos which are directed anteriorly. The succeeding dissoconch growth of
Mytilus is sharply marked off from the prodissoconch stage by its changed form and
color. In form the dissoconch has already acquired features characteristic of the group
to which it belongs, though not of the genus itself. The hinge line is produced poste-
riorly in an arched line and anteriorly the lines of growth are curved inwards toward
the umbonal area, but they are not crowded closely as in the adult or any species of the
genus. The shell at this age reminds one strongly of forms seen commonly in the
Modiomorphas, paleeozoic kindred of the Mytilidze. In color the dissoconch is distinct
from the prodissoconch. Whereas the latter as stated is yellowish-brown, the dissoconch
is characterized by the violet color characteristic of the adult. The transition from brown
to violet color in Mytilus is as sharply defined as is the change from homogeneous lime
in the prodissoconch of Ostrea to the prismatic structure of the succeeding dissoconch,
fig. 27, p. 314. Specimens of Mytilus edulis of the light horn colored variety of course
do not show this change to violet with the introduction of the dissoconch period. No
trace of prismatic cellular tissue has been found in the shell of Mytilus or other member of
this family at any period of growth, but a well defined epidermis exists.’ Lacaze-Duthiers
(41) published an anatomical figure of young Mytilus edulis comparable to the prodisso-
conch stage and he notes the striking difference between the shell of that age and the
adult. Lovén also published an anatomical figure of developing Mytilus which compares
favorably in its shell form with Pl. xxrx, fig. 9.

An older specimen of Mytilus edulis, Pl. xxrx, fig. 10, shows no important differences
from the younger specimen, fig. 9. The prodissoconch valves on account of the increased
arcuateness of the umbos are not seen in side view as well as in the earlier stage, a

1prismatic structure is described in the inner layers of Dreissena, but this genus has not been studied.

364 ROBERT TRACY JACKSON ON THE

necessary feature of developing bivalves. Instead of being rounded in appearance as
in fig. 9, the prodissoconch valves appear flattened from perspective. The prodissoconch
of fig. 10 differs from that of the earlier stage in being chalk-white, the yellowish-brown
organic appearance of the earlier condition having disappeared from the wearing away
of the epidermis. ‘The same is true of older specimens as long as the prodissoconch re-
mains intact, and it may often be found in specimens up to 4-5 mm. in length. It then ap-
pears as a little white cap at the umbo of either valve. The nepionic stage of fig. 10 is
similar to that of fig. 9, except that the lines of growth anteriorly are more crowded inward
toward the umbo, in this feature making a nearer approach to the condition characteristic
of the adult. Verrill notes collecting abundant young specimens of Mytilus edulis on the
12th of April, and I have found them in July, September and December, so that this spe-
cies evidently has wide limits in its breeding season.

Modiola hamatus, Verrill. This species occurs on our coast from Vineyard Sound
to Mexico. A number of young and adult specimens were collected at Martha’s Vine-
yard and Long Island Sound, also specimens were obtained from a mass of Perna
ephippium collected at Cuba. The young Modiola, has a well defined prodissoconch, p,
Pl. xxrx, fig. 11, which is yellowish-brown and with umbos directed anteriorly as in
Mytilus. The prodissoconch of Modiola appears flatter, less rounded than the prodis-
soconch of Mytilus, Pl. xxrx, fig. 9; but this is due in part at least to perspective as
noted in the older specimen of Mytilus, Pl. xxrx, fig. 10. The hinge line of Modiola is
produced posteriorly in a curving line and anteriorly the lines of growth approach the
umbos in a sweeping curve, fig. 11. The form of this stage is not widely separated from
that found in some species of the genus, but it more nearly approaches the form of the
Modiomorphas, palzeozoic allies of the Modiolas. An older specimen of Modiola hama-
tus, Pl. xxx, fig. 1, shows the introduction of the plications which are characteristic of
the species. ‘The plications originate suddenly, marking a distinct stage by their intro-
duction.

Crenella decussata, Montagu. Young specimens of this species, loaned me by the
Smithsonian Institution, show distinct prodissoconchs, p, as shown in Pl. xxx, fig. 3, an
umbonal view of aspecimen. ‘The prodissoconchs have rather highly developed umbos
which are directed anteriorly and show fine concentric lines of growth. The succeeding
dissoconch in its initial stages of growth is already marked by radiating plications as
well as concentric lines of growth which are coarser than those of the prodissoconch.

The Mytilide are classed with the Aviculidze, as heteromyarians, the connection being
based on the condition of unequal adductor muscles. This basis of classification, as pre-
viously considered (Introduction), is not deemed one of high, systematic importance and
is insufficient to connect otherwise differing groups. The striking differences in the pro-
dissoconch and nepionic stages of the Mytilidze and Aviculide are sufficient I think to
separate these groups, and the Mytilidee should be put in a group distinct from the Avic-
ulidee and their allies, which I have shown are all bound together by important features
as one group.

Argina ( Arca) pexata, Verrill. An extremely small specimen, Pl. xxx, fig. 2, was found
in a drain-pipe trap at Buzzards Bay. The specimen was dead, but the shell was perfectly
preserved. The prodissoconch is ovally rounded, very sharply defined; and has well devel-
oped umbos which are nearly central. The prodissoconch shows delicate concentric lines

PHYLOGENY OF THE PELECYPODA. 365

of growth, and is very yellow compared with the succeeding dissoconch growth which is
pure lime white, thus indicating a somewhat more marked conchyolin basis of the prodis-
soconch. The succeeding dissoconch has the typical hinge line characteristic of the
Arcas. Its first stages of growth are smooth and marked by nearly concentric lines of
growth; but with no plications; well defined plications exist, however, in the later growth.
Taking the shell as a whole, the height (measuring from the free border of valves to
the hinge line) is much greater posteriorly than anteriorly. In the middle of the ventral
portion a distinct depression exists on the border of the valves and also in the preceding
lines of growth. The early dissoconch before the introduction of plications differs much
in appearance from the adult and bears a strong resemblance to fossil members of the
Arcidxe. Especially does it resemble species of Macrodon from the Devonian as figured
by Hall. The teeth on the distal limits of the hinge line of young Argina are oblique like
those of Macrodon, and between these oblique teeth are the perpendicular Arca-like teeth ;
but at an earlier stage of growth the perpendicular teeth were apparently wanting, and
then the teeth of young Argina would have been closely comparable to those of Macrodon.
Area ectomata, Dall, from abyssal depths has all the teeth oblique like those on the distal
limits of the hinge line in young Argina and like those of the early and later species of
Macrodon.

The stage represented by the whole shell of Pl. xxx, fig. 2, differs in form from the
earlier Macrodon stage and also from the adult, and is comparable to Area, the type-
genus of the family. Transitional series of specimens of Argina pexata older than Pl.
xxx, fig. 2, show in constantly increasing degree the oblong-ovate form and shaggy ep-
idermis characteristic of the adult. In the ontogeny of Argina, therefore, we find that
there are four distinct stages which represent the ancestral forms from which the genus
was evolved: (1) the prodissoconch representing a primitive ancestral radical, at present
unknown; (2) the Macrodon stage, comparable to palzozoic forms of the family; (3) the
Arca stage, comparable to the type genus of the family and, finally (4), the Argina stage
in which the features of the genus and species are introduced and fixed. A prodisso-
conch and nepionic stage similar to that described in Argina pexata has been found in
a young specimen of Arca barbata, L., kindly given me by Professor Verrill, and in
some young Arcas of an undetermined species from Japan. The nepionic stage of Par-
allelopipedum tortuosum, Lama., is also similar to the same stage of Argina, the tortuous
character of the adult shell not having been acquired at this early period of growth.

It is an important fact that members of the Arcide have a prodissoconch, because as
one of the oldest living groups of Pelecypods it shows the widespread existence in time
as well as genera of a distinctly marked early embryonic shell. Species of so many and
such widely separated families possessing a prodissoconch in the young is strongest proof
that it is characteristic of the whole class of Pelecypods. We also see that the nepionic
stages in members of this family can be made use of in tracing genetic relations, as in
the Aviculidse and their allies.

The Unionide are naturally turned to in studies of this nature as they are characterized
by a peculiar early stage called the glochidium. Although the glochidium is known by
itself, I do not think the relation of this early shell to the adult has been previously studied.
Through the kind intervention of Dr. Charles E. Beecher, I recently received some ma~-

MEMOIRS BOSTON SOC. NAT. HIST., VOL IV. 49

366 ROBERT TRACY JACKSON ON THE

terial which was collected by that gentleman, from Mr. William B. Marshall of the New
York State Museum. The specimens embrace glochidial stages from the gill cavities of
the parent and adults with perfectly preserved umbos, so that the early embryonic shell is
still intact, and thus affords the opportunity of a comparison with later stages.

Unio cariosus, Say. The glochidium stage of this species, fig. 44, p. 367, is sub-quad-
rate with a straight hinge line and rounded angles on the ventral border. The shell has a
double line around its border, the inner line representing the limits of an earlier stage of
growth. A similar double line exists in the glochidium of all Unionide observed, or which
T have seen figured, and Dr. Lea makes the same observation. On the middle of the ven-
tral border of the valves is a very abbreviated spur which is not, however, spinous as in
Anodon. The single adductor muscle is situated near the hinge line and a little nearer
the anterior than the posterior border of the shell.’ The shell is granulated on its surface
and presents no traces of prismatic structure, the same feature being found in the glo-
chidium of all the Unionidsze observed. The glochidial shell of U. cariosus and other
species of the genus is whitish and much smaller than is the glochidium in Margaritana
and Anodon. The glochidium of Unio nasutus, Say, is similar to U. cariosusas described.
Dr. Lea figures the glochidium from the gills of the parent in twenty-five species of Unio.
Twenty are similar to U. cariosus in form with some little variation in contour. Of the
five remaining species three are sub-triangular resembling some species of Anodon; but
two have the ventral angles produced instead of rounded and each angle of either valve
bears a small spur.

Unio pressus, Lea. A specimen 47 mm. in length has the umbos well preserved and
still bears the early glochidial shell. This embryonic shell is sub-triangular and has the
same form as is figured by Lea ina specimen from the gill chambers of the parent of this
species. It is immediately succeeded by the nepionic stage of the dissoconch which has
the prismatic structure and concentrically rugose growth which is characteristic of the
young shel] in Unionid and is often figured at the umbos of well preserved specimens
(see Lea and other authors). The glochidial shell in its relation to the dissoconch there-
fore fills the requirements of the prodissoconch and may be compared with that stage as
described in many genera in this paper. The Unios above described were collected in
Albany County, New York, by C. E. Beecher.

In Margaritana undulata, Lea, the glochidium, fig. 45, p. 567, is sub-triangular, but
the anterior end of the embryo is considerably the larger. The hinge line is straight and
the valves at this area are raised in little humps indicating abbreviated umbos, not shown
in the figure. The angle at the ventral border of the shell is quite sharp; it is posterior
to a median line drawn perpendicularly through the shell and bears a denticulate spur on
either valve similarly to Anodon. In other features the shell is like the glochidial stage
of Unio. A young specimen of MZ. undulata 13 mm. long, fig. 48, has the umbos well
preserved and they retain the glochidium or prodissoconch intact. The hinge line of the
prodissoconch is parallel to the hinge line of the succeeding dissoconch growth and the
larger end of the embryonic shell is directed toward the anterior end of the shell. The
prodissoconch has the granulated surface characteristic of the glochidia of this and other

) The glochidial shell of all or nearly all Unionide is a perpendicularly through the hinge line and middle of the
little larger on one side than on the other of a line drawn ventral border of the valves.

PHYLOGENY OF THE PELECYPODA. 367

species of the family; but the dissoconch has an external layer of prismatic tissue. Cam-
era drawings of this specimen show that the prodissoconch is of the same size and form
as are the glochidia when removed from the gills of the parent as shown in the figures,
and the same observation holds in regard to the size and form of glochidia and com-
pleted prodissoconchs in every species in which I was able tomake the comparison. The
glochidial shell of Margaritana marginata, Lea, is similar to that of AZ. undulata. Lea
figures the glochidium of five other species of this genus. Of these, three are similar to
the glochidium of M/. undulata; but two are less triangular and more rounded on the
ventral border. A specimen of Margaritana margaratifera, which is 55 mm. long has
the umbos well preserved and they bear at the tip the embryonic shell. As in the pre-
vious species the prodissoconch of this species is sharply marked off from the succeeding
dissoconch, and in a camera drawing it coincides in form and size, with the glochidium
from the gills of the parent. Specimens of Margaritana deltoidea, 19 mm. long, have
the prodissoconch well preserved and it presents the same features in relation to the suc-
ceeding dissoconch as in the above-mentioned genera. The specimens of Margaritana
undulata studied were collected in Albany County, New York, and those of M. mar-
garitana and M. deltoidea at Ann Arbor, Michigan, by C. H. Beecher.

Fig. 44.—Unio cariosus, glochidium from the gills of the parent.

Fig. 45.—Margaritana undulata, glochidium from the gills of the parent.

Fig. 46.—Anodon cygnea, glochidium from the gills of the parent.

Fig. 47.—Anodon fluviatilis, glochidium from the gills of the parent.

Fic. 48.—Margaritana undulata, umbonal tip of a young shell, showing the prodissoconch, p, still in place and the
nepionic stages of the succeeding dissoconch.

Fic. 49.—Anodon fluviatilis, umbonal tip of a young shell showing the prodissoconch, 9, still in place and the nepionic
stages of the succeeding dissoconch.

All the figures (except fig. 46) are of left valves. The arrows are directed toward the anterior end of the shell; s, glo-
chidial spurs; h, hinge. All the figures X 50 diam. (Drawn by the author.)

In Anodon the glochidium is better known than the same stage of other genera of this
family, as it has been more often studied by embryologists. Schmidt studied it during the
period of parasitic life which it leads on fishes, and good figures of this stage are given by
that author. The glochidium of Anodon fluviatilis, Lea, fig. 47, is sub-triangular in

368 ROBERT TRACY JACKSON ON THE

outline but is less acutely angular on the ventral border than is this stage of Margaritana.
The hinge line is straight; the anterior end of the shell is somewhat the larger and spi-
nous spurs exist at the ventral angle of either valve, Pl. xxm, figs. 7-8. A double line
surrounds the limits of each valve and the surface of the valves is granulated, the same
characters existing in the glochidium of Unio and Margaritana. ‘The inner line is dis-
tinct from the edge of the shell and represents an earlier stage of growth. In A. cygnea,
fig. 46, p. 367, the double line represents the thickness of the edges of the shell, and the
inner line representing an earlier stage as in fig. 47, is not seen in this view. The glochi-
dium of Anodon differs from that of the other genera in the species observed in being of
a brown color. A glochidium similar to that described in Anodon fluviatilis I have ob-
served in glochidia from the gills of the parent of three other species, namely, A. wndu-
lata, A. cylindracea and A. cygnea, fig. 46. There is great constancy in size and
appearance of the glochidia in the different species of Anodon observed, the uniformity
being greater than in the other genera of the family.

A well preserved specimen of Anodon fluviatilis 25 mm. in length, fig. 49, p. 367, has
the embryonic shell or prodissoconch still in place. The prodissoconch ina camera draw-
ing of this specimen coincides in size and form with the glochidia from the gills of the
parent, no increased growth being observed. The larger end of the embryonic shell is
directed anteriorly and the hinge line is parallel to the hinge line of the succeeding dis-
soconch. The dissoconch has the rugose fluting characteristic of the young (frequently
figured at the umbos in well preserved specimens by Lea and others) and presents well
defined prismatic structure. The glochidia of Anodon fluviatilis described and the older
specimen of the same species as alsothe glochidia of A. undulata and A. cylindracea were
collected in Albany County, New York, by C. E. Beecher. The glochidia of A. eygnea
are from Belfast, Ireland, and were loaned to me by the Smithsonian Institution.

The fact that the larger end of the prodissoconch is directed anteriorly in all the species
of Margaritana and Anodon in which it was still found in place at the umbos ofthe valves,
is convincing proof of the correctness of Schmidt’s observation that the larger end of the
embryo of Anodon is the anterior end. It is a remarkable fact that the completed pro-
dissoconch in the various species observed has the same form and size as does the glo-
chidial stage of the same species when removed from the gills of the parent. We know
from the observations of Balfour and Schmidt that after quitting the gills of the parent
the embryos of Anodon (and probably, from analogy, other genera of the family as well)
lead a parasitic existence in the gills, fins, ete., of fishes. During this period of parasit-
ism considerable anatomical changes take place, yet these changes are not accompanied
by any noticeable modification or increase in size of the embryonic shell.

Tt is seen that the glochidial shells of the genera of the Unionide described are sub-tri-
angular or subquadrate; but always with a straight hinge line. But a single adductor
muscle exists and other anatomical features are in a very early condition of development.
This stage is directly the equivalent of what I have described as the phylembryonic stage
(p. 291) of development in Pelecypods and it may be compared to the embryos of Mo-
diola, Cardium and Montacuta, PI. xxi, figs. 9-11, also of Ostrea, fig. 25, p. 300, and Pl.
xxty, figs. 18-15. The anatomy of the glochidium from the gills of parents is less ad-
vanced than is the same (phylembryonic) stage in other genera, as shown in the figures
cited; butthis is explained as a peculiarity due to the specialized development of this type.

PHYLOGENY OF THE PELECYPODA. 369

It is a striking fact that the embryonic shell of the completed prodissoconch period of the
Unionide has not changed in form or size from the condition it had when ready to leave
the gills of the parents. This fact and the correlated fact that umbos have not devel-
oped, although they exist in the prodissoconch of the twenty-six other genera described
in this paper, are proof of the extreme reduction of the embryonic shell. The completed
prodissoconch of the Unionidee, therefore, has not progressed in the features of the shell
beyond the phylembryonic stage, which is found only in early embryonic stages of other
genera. This high reduction is probably to be explained by the condition of parasitism
which exists at this period in the development of the embryo. While the prodissoconch
of the Unionide is in all cases granulated on the surface and shows no prismatic tissue,
the succeeding dissoconch is characteristically prismatic. This condition is comparable
to what I have described in the Aviculide and their allies, but differs from what I have
found in other genera.

What the affinities of the Unionide are is a puzzle, for they are not clearly related to
any other family. The peculiar form, the spurs and absence of umbos in the prodisso-
conch are special features, apparently due to the peculiar environment of the developing
embryo. Although it cannot be supposed to fully represent in the ontogeny of the in-
dividualany ancient ancestral form, yet certain data can be gathered from this early shell.
As the prodissoconch of the Unionide has no prismatic tissue, it is doubtless the repre-
sentative of an early ancestor which was without that structure (the same comparison that
is made in the prodissoconch of the Aviculide). The radical ancestor, of which this
prodissoconch is the inherited representative, probably had umbos directed posteriorly,
for though the prodissoconch of the Unionidz has no umbos, the more rapid growth on
the anterior border would have pushed umbos posteriorly had they existed. That this
ancestor was dimyarian might be conjectured; but is fully implied by the fact that the
completed prodissoconch has two adductor muscles, for Balfour’s and Schmidt’s figures
of a mature glochidium of Anodon froma fish have two adductors. The nepionie stages
of the Unionid, though peculiar, do not so far as I am aware indicate kinship with early
predecessors of the group.

Spherium securis, Adams. This species was collected in a culvert at Medford, Mass.,
in November. The adults were crowded with large embryos, and as many as fourteen
were taken from the gills and mantle cavity of the parent. The embryos, Pl. xxx, fig.
4, show no traces of a velum as far as discerned, but have a well-developed foot and
crawl actively when removed from the parent. The foot hasa prominent otocyst as fig-
ured in Cyclas by Ziegler and other investigators ; but no byssal gland was observed as
described in that genus. Two adductor muscles, palps and filamentous gills exist. The
mantle border is produced posteriorly in a single siphon, fig. 4, but is free throughout
the rest of its exposed margin. In oldér specimens and the adult, two siphons exist,
an excurrent and an incurrent. This was not seen in the young of the age under con-
sideration, but as the point was not in mind when studying the specimens, it is possible
that two might be found. If not, it is an interesting character, for the adult of Pisidium
has but one and the existence of one in young Sphzrium would appear to be of genetic
significance. The shell of Sphzrium, Pl. xxx, fig. 4, represents an incompleted prodis-
soconch stage. In Pl. xxx, fig. 5, also from the gills of an adult, a completed prodisso-

370 ROBERT TRACY JACKSON ON THE

conch, p, exists and is succeeded by considerable dissoconch growth. The prodissoconch
is globose and almost circular, slight umbos being developed. When young, fig. 4, it is
very translucent of a light horn color, but in older specimens, fig. 5, the shell is thick-
ened by shell deposition, and it is darker and more opaque. The prodissoconch in Sphe-
rium is of gigantic size compared with the prodissoconch found in other Pelecypods de-
scribed and figured in this paper. Pl. xxx, fig. 5a, is a young specimen magnified only
ten times in order to show the large size of this early shell. By definition, the prodisso-
conch in Peleeypods is the shell of the completed embryonic period, and it is natural that
in Sphxrium we should have a large prodissoconch because of the relative great size of
the embryo in that genus.

Besides in Spherium securis, a similar strongly marked prodissoconch has been ob-
served in S. partumeium, Say, S. rhomboideum, Adams, and 8S. transversum. ‘The
specimens of these species observed are in the choice Beecher collection of fresh water
shells, in the New York State Museum.

Echinochama (Chama) arcinella, L. Ina specimen sent me by Dr. Beecher, a pro-
dissoconch was found intact. It is ovally rounded with umbos directed anteriorly. “The
nepionic stage is pink in color, is ovately drawn out, and truncated posteriorly. Con-
centric ridges mark the nepionic period and as pointed out by Fischer, this early stage
resembles a Venerupis in form. The shell of the nepionic stage is equivalvular as ob-
served in several specimens in the collections of the Boston Society of Natural History.
As soon, however, as the right valve becomes attached to a foreign body (which occurs
when the shell is about 3mm. in length), both valves immediately change their form and
assume the irregular growth characteristic of the Chamas, and soon the spines and pit-
like depressions characteristic of the species are introduced. We find in this case a full
substantiation of our views on the modifying effects of the conditions of fixation on
shell form as fully discussed in section vim, p. 322.

Petricola pholadiformis, Lam. At Nantucket in July, 1889, a large number of young
specimens of this species were collected, varying from very small to 12 mm. in length.
The largest individuals had the adult characters fully developed. The specimens were
burrowing in shallow Mytilus beds which covered rocks between tide marks. In the
youngest specimens found, Pl. xxx, fig. 6, also in some older individuals, a distinct pro-
dissoconch, p, exists at the umbos. It is not widely different from the succeeding disso-
conch in form, but is sub-ovately rounded and appears like a cap on account of its
whiteness compared with the succeeding grayer dissoconch. 'The whiteness of the pro-
dissoconch is noted in several genera in this section, and as described in Mytilus it is
doubtless due to the wearing away of the organic epidermis of the valves and consequent
exposure of the chalky-white lime basis. The early dissoconch of Petricola, Pl. xxx, fig.
6, is strikingly different from the adult. It is oblong, anteriorly roundly compressed and
posteriorly slightly truncated. The lines of growth indicate an earlier more oyal form.
In this early stage the ligament, /, projects from the hinge line as in the later stage, PI.
xxx, fig. 7, and in the adult; but no plications of the shell exist until considerably later.
The specimens, although taken from their burrows in the Mytilus bed, proved to be
active crawlers having a highly developed foot, fig. 6. The siphons of this early stage
fig. 4, are separate for the entire length produced beyond the limits of the shell. The

PHYLOGENY OF THE PELECYPODA. 371

incurrent siphon as in later periods is the longer of the two and the two tubesare widely
divergcnt as in the figure. The excurrent siphon is not retracted and extended periodi-
cally as described in Mya, Pl. xxx, fig. 13, and no tentacles surround the base of either
siphon tube in Petricola. The distal tips of the tubes in Petricola, fig. 6, are fringed by
by a few simple unbranched tentacles.

An older specimen of Petricola pholadiformis than that just considered is shown in
Pl. xxx, fig. 7. The umbos are too arcuately incurved to show the prodissoconch if
it were existent, and as a fact, it is not, being eroded away before this period is reached.
The younger portion as indicated by the lines of growth is oval and marked by coarse
concentric lines. A distinct stage is indicated in the oval portion of this specimen and
similar stages are common, but variable in individuals. The umbos are progressively
pushed anteriorly during the younger growth. Rugose plications originate suddenly, the
earliest stages having none. From the period of the introduction of plications, the shell
rapidly assumes features characteristic of the adult, differing principally in not being so
elongate, differing also of course in that the rugose plications of the adult are produced
nearly to what then becomes the umbo of the shell. The siphon tubes of Petricola of
the age of fig. 7, are long and are capable of being extended to a length considerably
greater than the total length of the shell. They are separate throughout their length
beyond the shell borders as in the younger specimen, fig. 6. In the adult, on the con-
trary, Verrill says they are united for about a quarter of their length. In specimens of
the age of fig. 7, the branchial siphon, Pl. xxx, fig. 8, is surrounded at its tip by tenta-
cles which branch laterally, intermediate with which are some simple unbranched tenta-
cles. The dorsal excurrent siphon of the same and of similar specimens on the contrary
has no branching tentacles, all being simple as in fig. 6, This condition of the orifices
is similar to that of the adult described by Verrill, excepting that he says some branch-
ing tentacles exist at the orifice of the excurrent siphon though fewer than at the orifice
of the branchial siphon.

Venus mercenaria, L. Isecured a number of young shells of this species in the drain-
pipe traps at Buzzards Bay. The youngest specimen, Pl. xxx, fig. 9, has a sharply
defined prodissoconch which is yellowish-white, covered with a delicate epidermis and
bears fine concentric lines of growth. It is separated from the succeeding dissoconch
growth by the changes then introduced in the marking of the shell, angles of curvature
and color, which though covered with a thin epidermis is nearly chalk-white. The first
stages of the dissoconch are marked by the series of concentric ridges and furrows char-
acteristic of young shells of this species, but which are wanting in the adult. This
feature is shown at the umbos of Dr. Gould’s figure of the species. Ridges and furrows
like those found in young Venus mercenaria are characteristic of the adult in some spe-
cies of the genus, as Venus rugosa, Chem., of the West Indies. The same feature is shown
in a high degree in the closely allied genus Chione, and its occurrence in young Venus
mercenaria is evidently of genetic significance. In its earliest growth, Pl. xxx, fig. 9,
the dissoconch shows only faint traces of a lunule (which was unfortunately over-
looked when the figure was drawn) but it is a marked feature in the older specimen,
Pl. xxx, fig. 10, and is characteristic of the adult.

Tottenia gemma, Perkins. A few young specimens were collected in August at Buz-

372 ROBERT TRACY JACKSON ON THE

zards Bay. The prodissoconch is oval, rounded and not sharply defined from the sue-
ceeding dissoconch, as both have nearly the same shape and the same peculiar violet
color. The prodissoconch is marked by a fine pitted structure which partially obscures
the lines of growth. The dissoconch bears closely crowded lines of growth and is vio-
let or amethyst color over its whole surface. Inthe adult, on the other hand, the shell is
mostly white or rosy, the amethyst color being limited to the umbonal and posterior por-
tion of the shell. Gould says this species is commonly taken for the fry of Venus mer-
cenaria on account of the color of the tip (young). This is an unnecessary mistake as
the lunule, elevated ridges and white color of young individuals of that species (Pl. xxx,
figs..9-10) are wanting both in the young and adult of Tottenia.

Scrobicularia tenuis, Pl. xxx, fig. 11, illustrates a specimen from the collections of
the Smithsonian Institution. It has a distinct prodissoconch which is not, however, as
sharply defined as in many genera because the succeeding dissoconch is so like it in form
that no marked contrast is produced by the introduction of that later stage. The pro-
dissoconch shows concentric lines of growth and its limits are defined by a varix of
growth indicating the close of the stage.

Saxicava arctica, Desh. Young specimens were collected on the roots of Laminaria
at Cohasset and Chelsea Beach in September, 1889. A very sharply defined prodisso-
conch exists at the umbos, fig. 50. It is ovate with umbos directed anteriorly and shows
fine concentric lines of growth. The
succeeding dissoconch is oblong, coy-
ered with a delicate epidermis, as is the
prodissoconch and bears comparatively
coarse strongly defined lines of growth.
The ligament is external and forms a
prominent feature of this early stage as
well as in the adult. Two ridges which
extend posteriorly from the limits of the

Fie. 50.—Young Saxicava arctica, showing oval rounded pro-
dissoconch and succeeding spinous dissoconch. The ligament is prodissoconch valves are ornamented

external and prominent. X 38 diam. (Drawn by J. H. Emerton.) with spines. The spines vary in size
and number, but persist in specimens up to four or five millimeters in length after which
they were not observed. Their existence is noted in the young of Saxicaya in Binney’s
Gould.

Mya arenaria, L. A number of young living individuals of this species were found
actively crawling on the glass slides in the drain-pipe traps at Buzzards Bay during
August, 1888; the specimens varied from 0.6 mm. to 5 mm. inlength. At Chelsea Beach,
Mass., in September, 1889, many clams were also collected, from very young to those hay-
ing the adult characteristics fully developed. The specimens varied from 1.5 mm. in
length upwards, and instead of being free crawling, as were those collected at Buzzards
Bay, they were for the most part burrowing in a bed of dead eel grass where the roots
and mud formed a comparatively solid mass. The burial was superficial and so abundant
were the little clams that a handful of the material might contain fifty specimens more or
less.

In very young Mya, PI. xxx, fig. 15, a remarkable condition of the siphon tubes exists.

PHYLOGENY OF THE PELECYPODA. 373

The tubes are divided from the base beyond the shell margins and are not coalesced at
any point visible when fully extended. The relation between this condition and that
found in the adult is of much interest and was fully illustrated in the specimens which I
studied alive. The excurrent tube, Pl. xxx, fig. 15, is composed of most delicate trans-
parent tissue. It is prolonged ina proboscis-like fashion and is frequently extended and
retracted with a jerky irregular motion. Water is expelled intermittently, and between
the expulsions the siphon tube is somewhat retracted. The tube has a small contracted
orifice and a tentacle was noticed at its base. More tentacles at this region doubtless ex-
isted, for many are found in the next stage, but they were not observed. The incurrent
branchial siphon tube, fig. 15, is much shorter than the excurrent tube. It has thicker
walls, bears a circlet of simple tentacles at its orifice and does not take part in the fre-
quent extensions and retractions characteristic of the excurrent tube. The siphon of an
older specimen measuring about 7 mm. in length is represented in PL. xxx, fig. 14. Short
termini of the excurrent and incurrent siphon tubes similar to those of fig. 13 exist at the
tip; but between these and the shell, a massive coalesced growth of the tubes has de-
veloped by interstitial growth of the extended mantle border. Thus the portions of the
siphons which are found in fig. 13 are now pushed out to the terminus of what may be
called the massive siphon. The massive siphon is pigmented in dots, especially near the
terminus which approaches nearest to the light, and it is covered by a delicate epi-
dermal layer as in the adult. Tentacles fringe the free, distal border of the massive tube
around its entire circumference, and a line of black pigment closely underlies the tentacles.
The excurrent terminus, as in fig. 15, is composed of delicate tissue, is extensible and has a
small contracted orifice which is plain, bearing no tentacles. Under higher magnification
small dots regularly placed were observed on the walls of the excurrent terminus. The
incurrent siphon terminus, as in fig. 13, is shorter than the excurrent. It is composed of
thicker tissue and is constantly extended, not intermittently extended and retracted
as is the excurrent terminus. Tentacles surround the orifice of the incurrent siphon and
between the tentacles is black pigment bordered by orange; the tentacles also are them-
selves slightly pigmented. The pedal orifice of the mantle border in specimens of this
age bears short papillose tentacles.

The development of the siphon from the condition just described, Pl. xxx, fig. 14, to
the adult. condition is merely a series of mechanical changes. The massive siphon elon-
gates proportionately until, as Verrill says, in the adult it can be extended to the length
of a foot or more. The termini of the tubes described in the young exist in the adult,
but are so reduced in proportion to the entire siphon that they are hardly noticeable.
The incurrent terminus is so short that it appears, as [ have seen it described, that two
rows of tentacles exist at this opening, whereas the fact is that one row surrounds the
tip of the massive siphon and the other the free border of the incurrent siphon terminus
as in fig. 14. The true condition of the parts in the adult is clearly figured by Morse, in
his “ First Book of Zodlogy.”

A condition of the siphons, similar to that of Pl. xxx, fig. 14, exists in adult Mya trun-
cata, as figured by Forbes and Hanley, and a condition similar to that of the earlier stage
of Mya arenaria, Pl. xxx, fig. 13, is found in the adults of several genera, some closely al-
lied to Mya and some not. Most important to us is the fact that this early condition of

MEMOIRS BOSTON SOC, NAT. HIST., VOL. IV, 50

BYE: ROBERT TRACY JACKSON ON THE

the siphon exactly corresponds to what is found in the adult of Corbula (see Fischer, fig.
855). Corbula presents in the adult many features found in the early development of
Mya and as it is the most ancient member of the Myidx, being known from the Trias
up, these facts are features of genetic importance in Mya. The great development of
the siphon in Mya arenaria is most distinctly a correlation of structure induced by its
habit of deep burial, and in the development of the individual we find that with increas-
ing age as the individual burrows deeper the siphon elongates.

Professor Ryder (63), in a paper recently published, announces the discovery of a bys-
sated condition in young J/ya arenaria. Specimens studied at Buzzards Bay did not show
this feature, although kept alive for several days; the clams from Chelsea Beach, however,
did show it. Thad them in glass dishes and many became attached to the sides and bottom.
The byssus is produced from the ventral border of the foot at about the position of its point
of extrusion from the shell in Pl. xxx, fig. 13. The animal attached itself as observed
by a few (1-4) colorless threads of great delicacy. The threads were not expanded in
disc-like distal termini as in the byssus of Pecten and Mytilus, but were glued to the
glass for a considerable part of their length. Sharp angles in the line of the attached
portion often existed, and separate threads often coalesced proximally. Byssal fixation
occurred in specimens varying from 2.7 mm. to 9 mm. in length, and it was evidently a
strong habit, for when detached, in a few minutes the fixation was again effected. A
byssal cleft is characteristic of Corbula and Sphenia, near allies of Mya, according to
authors, and Forbes and Hanley state that a byssal cleft exists in the foot of Mya trun-
cata. In avery young specimen of Mya arenaria studied alive, the gills were filament-
ous, ciliated and expanded at their tips as shown in the gills of young Ostrea, Pl. xxrv,
fig. 3. I did not determine certainly whether cross-bars united the tips or not.

A young specimen of Mya arenaria, Pl. xxx, fig. 12, shows the prodissoconch, p,
clearly. The prodissoconch is rounded in outline, with umbos directed anteriorly. It
is glistening and translucent, covered by a delicate epidermis and marked by fine con-
centric lines of growth. The prodissoconch may be retained in well preserved speci-
mens, much older than fig. 12, even in those which are several millimetres in length; but
commonly it is lost early from erosion. ‘The prodissoconch is not so sharply marked off
from the succeeding growth in Mya as in many genera, because the dissoconch growth
has no striking features which originate with its introduction. It is, however, perfectly
clear when preserved.

Professor Ryder (63) in his paper on Mya arenaria, published January, 1889, says
that he did not find an embryonic shellin young clams such as is characteristic of young
oysters. Since that time I have sent him specimens showing the embryonic or larval
shell (prodissoconch) of Mya, and he acknowledges their existence in a letter, saying:
“The specimens which you have been good enough to send show the larval shell very
distinctly indeed, much more plainly than my own materials, in fact.”

The dissoconch growth of Mya arenaria is covered with a delicate epidermis and is
much whiter than the prodissoconch. I did not find any trace of prismatic cellular tis-
sue inthe young shell of Mya. The early dissoconch growth, PI. xxx, fig. 12, is rounded
in outline anteriorly, but truncated posteriorly and marked by fine concentric¢ lines of
growth. An older specimen of a clam is seen in PI. xxx, fig. 13. The prodissoconch

PHYLOGENY OF THE PELECYPOBA. 375

of this later stage, on account of the arcuateness of the umbos, is not seen so clearly in
side view as in the younger stage. The shell differs much from the adult in form, being
less rounded anteriorly and less pointed posteriorly, also the valves gape less proportion-
ately than in the adult. An angular ridge runs from the umbos posteriorly to the ven-
tral border of the valves, emphasizing a flattened area above it. The features described
are found in older young specimens, but gradually fade out and in successive growth
give place to the characteristics of the adult.

The young shell of Mya arenaria bears a marked resemblance to some species of
Corbula (leaving out of consideration the inequality of valves in many species of that
genus) ; especially does it resemble C. contracta, Say, of our coast. The anatomy of
young Mya also bears a close resemblance to the adult of Corbula (see Fischer’s fig.
855) in possessing a similar siphon, Pl. xxx, fig. 13, papillose tentacles on the mantle
border at the point of extrusion of the foot and a byssated foot. As Corbula is the old-
est known genus of the Myidx, being found from the Triassic upwards, and. as it pre-
sents similar features to those found in the young of Mya, it is probably in the line of
ancestry of the Myas and is a near representative of the form from which they have
descended.

XV. On THE GENETIC RELATIONS OF THE PRODISSOCONCH.”

In the development of Pelecypoda at an early embryonic stage, which I have described
as the phylembryo stage, the two valves of the shell are sub-circular or saucer-shaped ;
but flattened into a straight line at the hinge area where they are joined by a ligament.
This stage is characteristic of many developing Pelecypods (PI. xxii, figs. 7-11; PI.
XxIVv, figs. 13-16 and fig. 25, p. 300)* and apparently represents a primitive ancestral
condition common to the whole class. Embryology shows that the Pelecypod shell orig-
inates from a single shell gland and that the two valves doubtless arose from the split-
ting on the median line of a primitive univalvular ancestral form. If that ancient
univalyular radical had a saddle-shaped shell as in young Dentalium, Pl. xxun, fig. 13,
or a cup-shaped shell as exists in very early stages of cephalous mollusca,* the first re-
sult of the introduction of a median hinge would be to straighten the shell on the hinge
line. This is a simple problem in mechanics; for, if one tries to break by flexion a piece
of metal which is saddle-shaped or cup-shaped, the metal will tend to form a straight

1 Since this paper was in press, a prodissoconch has been 3Tt is also described in Teredo by Hatschek. Excep-

discovered in a number of genera not considered in the
foregoing pages. An oyally rounded prodissoconch has
been found in Nucula proxima and in Tellina. In both
genera the umbos of the prodissoconch are directed pos-
teriorly as in the adults. A prodissoconch has also been
discovered in Chione, Cardita, Cardium, Mactra and a
number of undetermined genera belonging to other families
of Pelecypods. In these latter the umbos of the prodisso-
conchs are directed anteriorly, and the succeeding disso-
conch is destitute of prismatic cellular tissue.

2This section and the next one, xvi, have been con-
siderably altered and added to since this paper was offered
as athesis at Harvard University on May 1, 1889. The ad-
ditions are not indicated excepting by this note.

tions to the rule occur as in Pisidium, where according
to Lankester (43) the early embryonic valves are triangu-
lar. I think such cases may be considered exceptions in
which the straight hinge period is skipped by acceleration
of development.

4The shell of Gasteropods, as I have observed in stud-
ies of Sycotopus, is cup-shaped before it acquires the
spiral form of the completed protoconch, and a similar
condition is shown in Lankester’s (43) figures of Aplysia.
The shell of Pteropods and Cephalopods is also at first
cup-shaped as is obyious in studying the changes that
must take place in the building up of the completed pro-
toconch, Pl. xxim, figs. 15-18.

376 ROBERT TRACY JACKSON ON THE

line on the axis of flexion. A directly parallel case is the condition existent in bivalvu-
lar crustaceans. These originate by the introduction of a hinge in the median line of
the carapace of a normal crustacean. The ancient Ostracoda,—Leperditia, Aristozoe,
ete., have a straight hinge line and sub circular valves which are united on the hinge
fie by a ligament. Thus, the resulting form of the bivalvular shell in these two entirely
distinct classes is strikingly similar.

The next stage in the development of the Pelecypod shell is the loss of the early
straight hinge line which is superseded by a curving of the hinge line and a centraliza-
tion of connecting tissue and teeth of the two valves in the median line of the hinge.’
The necessary form of the valves of an equilateral bivalve, freely crawling on the ven-
tral border, is curved in section. Therefore, more or less arcuate umbos would develop
during successive growth. This condition of growth would bring the shell to the form
of the completed prodissoconch stage which is the representative of ancestral type rad-
icals in the ontogeny of the individual. A close parallel of this later stage of shell
growth is again to be found in bivalvular crustaceans. Some genera, as Hstheria, have
a curved hinge line, more or Jess arcuate umbos, and bear a close resemblance to Pele-
eypods. In this consideration of the mechanical origin of the Pelecypod shell, especial
attention should be called to a highly interesting sand important paper by Dr. Dall,’ in

which he demonstrates in a most striking manner the mechanical relations of the liga-
ment and cartilage to the hinge in Balser oils and the progressive, mechanical devel-
opment of systems of hinge teeth in this class.

The form of the shell up to the close of the prodissoconch stage seems to be the nat-
ural outcome of the bivalvular condition. I think the muscles which close the valves
may be demonstrated to be the necessary consequent of the mechanical conditions of a
bivalve shell. The anterior adductor muscle in Pelecypods is reduced and finally dis-
appears when it comes into such position that its functional activity is lost, e. g., Ostrea,
Pecten, Mulleria, etc. (section vy). The posterior adductor also may disappear through
disuse. In Aspergillum, where the two valves have united so as to form a simple tube,’
and therefore, of course, the action of the adductor muscles is lost, the posterior adduc-
tor has disappeared and the anterior is reduced to a few disconnected shreds (Fischer).
In the phylembryo stage of Pelecypods, fig. 25, p. 300, Pl. xxin, figs. 7-11, the valves
are closed by a single adductor muscle, the simplest condition mechanically possible to
effect the desired end. This adductor does not seem to be homologous with any muscle
in other classes of molluscs, and it probably developed as a necessary consequence of the

1A straight hinge line is characteristic of many Pelecy-
pods and it isnot easy to perceive what causes led to the

of as if it were something separate and distinct from the
shell, whereas it is a modified condition of the shell and

adoption of a curved one in the primitive radicals of the should be so considered. In the young of Aspergillum ey-

class. That it was adopted, however, is predicated by the
form of the completed prodissoconch as described in these
pages. See Plates xx1v-xxx. The prodissoconch of the
Unionide has a straight hinge line, the reasons for which
are discussed on p. 368.

* On the hinge of Pelecypods and its development, with
an attempt toward a better subdivision of the group. By
Wm. H. Dall, Am. Journ. Sci., Vol. xxxvin, pp. 445-462.
Published since this paper was in press.

° The tube of this and similar genera is commonly spoken

ery step can be traced between the typical bivalvular stage
and the period when the two valves merge into one anoth-
er by concrescence and form the truly univalvular condi-
tion characteristic of the genus. This concrescence of
the apposed borders of the two valves in Aspergillum is
paralleled by the conerescence of the two sides of the em-
bryonic saddle-shaped shell of Dentalium, Pl. xxi, fig. 14,
which unite to form the tubular shell of the adult, as de-
scribed by Lacaze-Duthiers (42).

PHYLOGENY OF THE PELECYPODA. a HA

conditions of the case.!. In support of this view, the bivalvular crustaceans may again be
cited as they have an analogous adductor muscle, developed, of course, on an entirely
different line of descent, but under closely similar mechanical conditions. In the com-
pleted prodissoconch of all Pelecypods there are two adductor muscles, Pl. xxrv, figs.
1-2, Pl. xxx, fig. 4,” thus arriving at what may be called in this respect the typical con-
dition of the class. During later life the anterior, the posterior, or both adductors, may
be retained, reduced or lost, according as the persistence or changes in correlated feat-
ures of anatomy retain in use or bring into disuse the muscles in question.

The preceding studies show the existence of an embryonic shell to which I have given
the name of prodissoconch (p. 311) in many widely separated genera of Pelecypoda.
We will now consider this early shell and see if any close connection can be traced be-
tween it and ancestral forms of which it is the representative in early stages of develop-
ment.

Two classes of prodissoconchs and succeeding shell structure are described in the
preceding pages. In the first division the prodissoconch has umbos directed posteriorly
and is not prismatic, but is succeeded by a dissoconch having an external layer of pris-
matic cellular tissue which is more or less developed, but exists at least in the early
nepionic stages of one valve. These features are characteristic of the Aviculidx, Os-
tread, Pectinidee and Anomiade, as shown by representative genera. The second di-
vision is less well defined; but with one exception can be separated from the first. The
prodissoconch of this second division has umbos which are directed more or less ante-
riorly. It has no prismatic cellular layer and. is succeeded by a dissoconch which has no
prismatic layer. Such features are characteristic of genera representing the families My-
tilidee, Arcidze, Chamide, Petricolidze, Cycladidx, V eneridze, Glycimeridze, Scrobicularidee
and Myide. The young of the Unionide as described (page 305) does not come
under this second section but in some features approaches the first. Only a part of the
families of Pelecypods are included under these two divisions. Whether other families
will come under one or the other, or yet new divisions, must be proven by future inves-
tigations (see note, p. 375).

The first of the divisions above considered embraces a group of families (the Avicu-
lide and allies) intimately bound together as I have attempted to show in this paper.
It will probably prove to be a group of ordinal value and J shall limit myself to this
group in considering the derivation of the prodissoconch. ‘The prodissoconch of the sec-
ond division has doubtless a different ancestral kinship or perhaps several lines of ances-
tral kinship as the case may be.

The features by which a possible connection may be traced to the radical ances-
tor from which the prodissoconch form is inherited includes features of shell structure
and anatomy, for we know both in a measure from the results of the studies in this pa-

1The mantle of molluses often has a highly developed 2Also figures by Lovén, Schmidt and other investigators
muscular system, and it seems probable that the adductors of marine and fresh-water Pelecypoda The fact of the
of the Pelecypod arose from a modification of mantle two adductors is also attested by the form of the prodis-
muscles. A somewhat similar modification of these mus- soconch in the genera figured on Plates XxvI-xxx inclu-
cles is seen in the retractor muscles of the siphon, charac- Sive.

teristic of many Pelecypods, as the clam, Pl. xxv, fig 9.

378 ROBERT TRACY JACKSON ON THE

per. These two lines of inquiry linked with the proper paleontological sequence are
found to point directly to a form which fills the necessary requirements of such an ances-
tor. Data for the inquiry. Zhe shell. In Avicula, Perna, Ostrea, Pecten, Anomia
and allies a prodissoconch exists in which the structure is homogeneous and laminar,
sub-nacreous, not prismatic, though succeeded by a dissoconch having well-defined pris-
matic structure. The umbos of the prodissoconchs are directed posteriorly, and the
hinge line is curved. The only genus in which we know the prodissoconch before any dis-
soconch growth has taken place, so to enable us to determine features of the hinge line,
at this stage is Ostrea. In Ostrea edulis, fig. 26, p. 312, the prodissoconch has teeth
at right angles to the hinge line, and these teeth closely resemble the teeth of Nucula,
when a shell of that genus is viewed from the umbonal region as in the figure cited. In
O. virginiana, however, no teeth were observed. The anatomy as illustrated by the va-
rious genera yields points of great synthetic value. The prodissoconch has two ad-
ductor muscles as proven in Ostrea, Pl. xxtv, figs. 1-2, and inferred in the other genera
from the form of the prodissoconch valves and the kinship to Ostrea (see studies of
Perna, Avicula, Pecten and Anomia). In the prodissoconch stage of Ostrea the gills
are simple unpaired filaments and they are evidently simple filaments at first in other
genera as indicated in our studies of Anomia and Pecten.'| The mantle border of the
prodissoconch is free around its entire margins except at the hinge line in Ostrea, and
bears no tentacles or siphons. A similar condition of freedom exists in the youngest
stages we know of related genera, and the development of tentacles and eyes described
in Pecten indicates that at a very early period the border was without them in that ge-
nus. <A foot evidently exists in the typical prodissoconch, for though wanting in Os-
trea,” it is present and highly developed at the close of the prodissoconch period in the
young of many related genera (Avicula, Perna, Pecten, Anomia, ete.), as I have shown.
The foot where known is marked by a proximal byssal cleft and a distal cleft which lat-
ter is flattened out when crawling, giving the foot a dise-like, clinging power (see Pec-
ten and Anomia, sections 1x and xIr).

With the above described prodissoconch features as known characters we have to seek
for possible kinship with ancient forms. These characters are found to be combined in
a remarkable degree in the ancient genus Nucula which, though living to-day, extends
back to very early fossiliferous formations, fifty species (Tryon) being known from the
paleozoic. Nucula is one of the few genera known which has the umbos of the shell
directed toward the posterior end of the body. The shell of Nucula is oval, nacreous,
not prismatic! and has a curved hinge line with teeth at right angles to its axis. A tri-
angular, central cartilage pit is a characteristic feature of the Nuculide, and though it
is not definitely known to exist in the prodissoconch it may exist at that stage in some
genera, and is a marked feature of the succeeding dissoconch in all the ascendent series

1A simple, filamentous condition of the gills as first de-
scribed by Lacaze-Duthiers in Mytilus is considered by
embryologists as characteristic of all developing Pelecy-
poda.

2 The absence of the foot in Ostrea is fully discussed in
previous pages.

°’The only genera outside the Nuculidé, in which the um-

bos are directed posteriorly as far as I have found from
the conchological manuals, are Anatina, Donax and Tel-
lina, the two latter belonging to one family.

4Traces of prismatic cellular structure are noted in Nu-
cula by Carpenter, but he equally notes it in many other
genera of Pelecypods, as sometimes seen in slight patches
though not a feature of the shell as in Avicula, etc.

PHYLOGENY OF THE PELECYPODA. 379

‘of the Aviculidee and their allies. Nucula is dimyarian, an adductor muscle existing at
either end of the equal valves. The gills are in two pairs, composed of simple, disunited
filaments! (Mitzukuri). The mantle borders are plain, without tentacles or a siphon, and
free around the margins. The foot is marked by a longitudinal cleft; accordingly it can
be flattened into a circular dise (Mitzukuri) and it has been observed crawling up the
sides of a glass of sea water (Tryon) indicating considerable clinging capacity of the
foot. All these characters coinciding so closely with the features of anatomy and shell
structure known in the prodissoconch, together with the fact that Nucula is known in
very early fossiliferous formations combine to render it extremely probable that Nucula
or a Nuculoid form is the type we are seeking as the early ancestral radical repre-
sented by the completed prodissoconch in the development of Avicula, Perna, Ostrea,
Pecten, Anomia and their allies.

The fact that Nuculais found in the Lower Silurian formations and still lives up to the
present day without sufficient changes in its shell structure to make these extremely
separable forms in time generically separable, argues strongly for a very high antiquity
for this genus. It was so firmly established in its earliest known forms that it must at
that time have been an ancient genus. It may therefore well have been the primitive
ancestral radical toward which the prodissoconch of the genera under consideration
points as the archetype, still indicated in the development of a modern individual.

I have, therefore, from the results of these studies been able to trace, to Nucula or a
Nuculoid form as a probable type-ancestor, the prodissoconch which I have found to be
characteristic of developing Avicula, Perna, Ostrea, Pecten, Anomia and their allies. In
the next section (xv1) the relationship existing between these several genera and their
allies is traced as far as I am able to do so with my present understanding of the groups.

XVI. GENEALOGICAL CONNECTION OF THE AVICULID® AND THEIR ALLIEs.

It is shown in section xv that the prodissoconch, which is characteristic of the Avic-
ulidse and their allies, as represented by typical genera, can be traced to Nucula or a
Nuculoid form as a probable ancestral radical of the whole group. In this section later
stages of shell growth are considered in their phylogenetic bearing and a tabular view is
given which represents the genealogical connection of the several genera as it is under-
stood from the results of these studies. Frequent reference is made to the preceding
text for a full consideration of points only briefly alluded to, and many genera not pre-
viously studied are introduced.

The early nepionic stage of Avicula, figs. 32-33, p. 330, Perna, Pl. xxvtt, fig. 1, and
Pecten, Pl. xxvm, fig. 9, is closely alike in many species, indicating a common line of
ancestry for these genera. The points of resemblance are as follows:—The body of the
shell is sub-rhomboidal, oblique, the hinge line straight; the posterior wing is rounded
and is not produced beyond the margin of the body of the shell. Anteriorly the straight
hinge extends beyond the limits of the prodissoconch, and is terminated by a moderate
auricle as in Avicula and Pecten, or the auricle may descend directly from the umbonal
area without any anterior extension of the hinge line as in Perna. In the several genera

'Nucula has been considered a primitive prototypal Pel- plicity of the gills, whichare paralleled in the early stages
ecypod by several investigators on account of the sim- of many genera of Pelecypeds.

380 ROBERT TRACY JACKSON ON THE

the ligament joins the valves on the hinge line and a triangular cartilage pit extends
inwards from the limits of the prodissoconch valve as shown in young Perna, fig. 30,
p- 329. Teeth may exist on the hinge line as in Avicula and Perna, or not, as in Pecten.
The right valve in the young of these genera has a byssal sinus originating in the ini-
tial stages of the dissoconch growth on the border of the prodissoconch valve, thence
progressively produced to the free border of the valves, fig. 53, p. 330; Pl. xxvu, figs.
2, 3s 0,85 Shanda l x xvii aie et 0!

The form of the young shell as above described is found to be closely similar to that
characteristic of the adults of many Avyiculoids from the Lower and Upper Silurian for-
mations. Hspecially is it comparable to early Aviculoids figured by Barrande. These
early forms differ widely from Avicula and require one or more distinct generic names.
The Devonian Aviculoids have been subdivided generically by Professor Hall, but none
of his genera suflice to cover the types which are here referred to. I would, therefore,
introduce a new genus— Rhombopteria.

Rhombopteria, gen. nov.!| A group of Lower and Upper Silurian Aviculoids, in which
the shell is sub-rhomboidal, oblique, with umbos in the anterior half. The hinge line is
straight, less than the length of shell, and is commonly produced anteriorly as well as
posteriorly from the umbos. The borders of the valves posteriorly descend in a curved
line from the terminus of the hinge without having any posterior extension of the wing.
Anteriorly the borders of the valves descend from the hinge, forming an auricle close to
the hinge line. In some species the auricle may be in close proximity to the umbos with-
out any anterior extension of the hinge line, as in Rhombopteria ( Av.) glabra, sp. Munst.
Two adductor muscles and lateral teeth exist in closely related forms and probably are
characteristic of this genus; but the features are not definitely known. It is quite pos-
sible, also, that a cartilage pit exists, although it also is not known, as the internal char-
acters of the genus have not been ascertained. Type— Rhombopteria ( Av.) mira, sp,
Barr., fig. 51. For other figures of this species see Barrande’s Systeme Silurien de la
Bohéme, Vol. VI, Pl. 226. Besides the type several species figured
by M. Barrande come under the genus Rhombopteria, among which
should be mentioned: /?. pseudomira, sp. Barr.; R. cognata, sp. Barr.;
R. scala, sp. Barr.; R. patricia, sp. Barr. and R&R. glabra, sp. Munst, all
of which are figured on Plates 226-228 inclusive.

Rhombopteria is the nearest approach to a primitive Aviculoid that
ae ieee eee has been found in studying fossil species. Not only can the nepionic
E (upper Silurian), stages of modern Avicula, Pecten and Perna be closely compared to
ae ees this type, but camera drawings of the young of Aviculopecten, Lepto-

"desma and other paleeozoic and mesozoic allied genera bear a close re-
semblance to Rhombopteria as well. The American Lower Silurian species, Avicula?
trentonensis, Con., A.? gregaria, Ulrich, and A.? demissa, Con., are near allies of Rhom-
bopteria, especially the species trentonensis. In Aviculoids as a rule, the younger the
specimen or the older the geological formation from which a specimen is obtained the
nearer does it approach to the features of the genus Rhombopteria. In the comparisons
of young Avicula, Perna, Pecten, etc., to Rhombopteria, it will be observed that there

1 “popfos, a rhomb, and zt<pov, a wing.

PHYLOGENY OF THE PELECYPODA. 38]

is a discrepancy in that the right valve of these genera has a well defined byssal sinus,
whereas no byssal sinus exists in any of the figures of Rhombopteria cited. Dr. Beecher
writes me that young Avicula? demissa has a byssal sinus and this species isa near ap-
proach to Rhombopteria, especially when young, but differs in having the posterior wing
produced in the adult. It is quite possible that species of Rhombopteria may be found
which havea byssal sinus; but if no such species should be found, it would not seriously
affect our comparison of young Aviculas, ete., to this early type. The byssal sinus acquired
in later periods of the development of a group might well be pushed back to the nepionic
stages in the ontogeny of an individual by acceleration of development, also the exist-
ence of the character is quite variable in some genera, as in Aviculopecten.

The genetic relations expressed in the following pages and table must be looked at as
a first attempt to classify the various groups and genera in their serial connections, for
such I believe it is. Changes may and probably will be necessary in parts, but as a
whole I believe it represents a close approximation to the true phylogenetic sequence of
the genera considered.

The dimyarian prodissoconch characteristic of the early stages of this whole group is
the representative in development of a dimyarian Nuculoid radical as discussed in sec-
tion xy. Therefore a Nuculoid form makes a basis from which to start our phylogenetic
series.

These studies tend to show that the primitive Aviculoid ancestor came off directly from
the Nuculoid radical because the first nepionie stages of Avicula, Leptodesma, Perna,
Pecten, Aviculopecten, ete., correspond closely with the earliest and simplest known
Aviculoid. If intermediate forms existed between the Nuculoid radical and the earliest
known Aviculoid, traces of it are not indicated in the shell, so that such possible transi-
tional steps cannot be ascertained with our present knowledge. Therefore Rhombop-
teria, as the most primitive known Aviculoid, is considered the next step upward in the
phylogeny of the group.

From this point up the phylogeny of the various groups is on divergent lines as ex-
pressed in the tabular view, and in the following pages we shall consider the groups in
succession as indicated in the table, p. 391.

Leptodesma,! Hall, is a genus embracing a group of Devonian Aviculoids characterized
by an oblique body. The posterior wing is more or less produced beyond the borders
of the valves; anteriorly the shell is acute or nasute. The hinge is narrow with a slen-
der, lateral tooth. The posterior adductor muscle is large, subcentral, but the anterior
is not figured by Hall and is therefore presumably very reduced. The genus forms a
natural passage from Rhombopteria to the true Aviculas. As Professor Hall says, the
species present a remarkable assemblage of forms. At one extreme the shell has a well-
developed posterior wing, acute anterior border and byssal sinus closely resembling Avi-
cula. From this extreme it passes“ . . . through various phases in form, proportions,
ete., until the wing becomes neariy obsolete, the byssal sinus obscure and the anterior
end rounded. . . . ” This other extreme brings the form very close to our genus

1¥or a consideration of this and other palzeozoic genera York, Vol. v, Parts 1m. I am also recently indebted to
I am greatly indebted to Professor Hall’s fine work on that venerable palzeontologist for the opportunity to study
Pevonian Lamellibranchs in the Natural History of New his collections and types at Albany.

MEMOIRS BOSTON SOC. NAT. HIST., VOL. IV. dL

382 ROBERT TRACY JACKSON ON THE

Rhombopteria. Leptodesma orus, Hall, is a species with a well-developed posterior
wing but a camera drawing of the young is rhomboidal and bears much resemblance to
Rhombopteria. The same comparison may be made in a degree by noting the differ-
ences in the younger and older specimens figured by Hall in several species as L. math-
ert and LZ. jason and it may also be observed by following the lines of growth in many
of his figures. Leptodesma is therefore considered the next step in the ascendent phy-
logenetic series. .

Avicula, the representative of the family, is clearly derived from Leptodesma-like forms.
Using Avicula hirundo, L., the type species of the genus, as our basis of comparison
we find that Avicula differs from Leptodesma in having a highly produced posterior wing,
a deep byssal notch in the right valve, and small cardinal teeth. A more or less well-
developed triangular cartilage pit exists in all Aviculas. In species with a narrow hinge
line, as A. sterna, the pit is only prominent in the young; but in others, as in the sub-
genus Meleagrina, it is a marked feature of the adults.’ In the ontogeny of the individ-
ual we find that Avicula at the prodissoconch stage is referable in ancestry to the Nu-
culoid radical of the group. In the early nepionic stage, fig. 32, p. 330, it is referable
to Rhombopteria; in the later nepionic stage, fig. 33, p. 330, a considerable wing has de-
veloped and the form approaches near to the winged Leptodesmas. Later still the full
Avicula features are acquired. Thus the ontogeny of the individual represents very fully
the phylogeny of the genus. The first appearance of true Avicula in the geological
horizon is uncertain as many early types called Avicula certainly do not properly belong
to that genus and the limitations of the genus in the past have not been fully ascer-
tained. It probably is truly represented as early as the Devonian or Carboniferous.
From Avicula many genera and sub-genera are derived, a part of which only we are
prepared to discuss. (Further consideration of Avicula, see p. 330.)

Pseudomonotis. The young, as observed in P. echinata, has a prodissoconch and nepi-
onic stage like the same period of Avicula, figs. 32-34, p. 330. Later, the hinge line
becomes relatively shorter and very moderate abbreviated wings are differentiated. Mo-
notis occurs later in time and has features closely comparable to Pseudomonotis, there-
fore it is considered provisionally as derived from that genus. Daonella and Halobia
are genera which are sub-rhomboidal in form and present a high reduction of Aviculoid
characters. They are unquestionably side-issues off the ascendent line of the group and
should probbly be associated with Monotis or Pseudomonotis, but as I have not had speci-
mens for study they are not further considered. (Further description of Monotis and
Pseudomonotis, see pp. 331, 332.)

Oxytoma, a genus near to Avicula, has a short hinge line with moderately developed
ears and a slight obliquity of the body of the shell. It seems to be a variation from
Avicula, in the same line as Monotis and Pseudomonotis. Some features of the young
are shown in Pl. xxvu, fig. 7. (Further description of Oxytoma, see p. 332.)

Cassianella is highly inequivalvular; the young has a prodissoconch and nepionic
stages like the young of Avicula, as shown in C. decussata, fig. 35, p. 332. The hinge
line is of moderate length, the wings are moderately produced and cardinal teeth are

‘A cartilage pit is characteristic of the Pecten division teria and Leptodesma. In allit may be inherited from the
of this group and it is probably characteristic of Rhombop- Nuculoid ancestor in which it is a characteristic feature.

PHYLOGENY OF THE PELECYPODA. 383

well developed. Provisionally, Cassianella is considered a branch from Avicula, in
the same line as Pseudomonotis, ete. (Further description of Cassianella, see p. 332.)

Vulsella, when young, fig. 36, p. 332, bears a considerable resemblance to young Avic-
ula, fig. 32, p. 330, as it has a similar prodissoconch and nepionic rhombic form. La-
ter, the resemblance is lost, as wings are wanting, the hinge line is highly reduced and
often produced in a curve around the deeply-sunken cartilage pit. Vulsella is evidently
a side-issue from the Avicula stock and provisionally is placed on the same branch as
Malleus.

Malleus is a genus closely allied to Avicula. The main features of difference are the
great development of prismatic cellular tissue, loss of teeth, deep cartilage pit and con-
torted form. In the young it resembles Avicula, as pointed out by Woodward. It
should be observed, however, that it is the young of Avicula, fig. 32, p. 330, which the
young Malleus most nearly resembles, as shown at the umbonal region or early growth
of fig. 52. The irregular, contorted growth of adult Malleus, Pl. xxvun, fig. 19, and
also the relative great development of prismatic
cellular tissue and extreme reduction in area and
thickness of nacreous tissue, are features of the
adult which point toward the probability that Mal-
leus is a degenerate or retrograde genus.

On another line of descent we find Meleagrina.
When young it resembles young Avicula, fig. 52,
p. 330, but it has then no teeth. The cartilage pit
descends from the limits of the prodissoconch as
in Avicula and Perna, fig. 30, p. 329. With in-
creasing age the oblique form of the body of the

Fic. 52.—Malleus vulgaris, L. Tip of left valve Shell decreases and the characteristic rhombic form
showing the form of the young. X 13 diam. of the adult progressively increases. The rhom-
poe bic form characteristic of the adults of this and
some other genera is not necessarily to be considered as an inheritance from Rhombop-
teria. It is rather an approach to a form normal to the group and which is frequently
repeated in genera not closely akin. (Further description of Meleagrina, see p. 331.)

Perna is anunquestionable branch from Avicula. The nepionicstage of Perna, Pl. xxvit,
fig. 1, is similar to the same stage of Avicula, fig. 32, p. 330, the main difference being that
the anterior auricle descends directly from the prodissoconch without an anterior exten-
sion of the hinge line. At this age but a single triangular cartilage pit exists, fig. 30,
p- 329, and cardinal and lateral teeth like those of Avicula underlie the hinge line.
Later, new cartilage pits develop on the hinge line, fig. 31, p. 329, and in successive
growth they assume the perpendicular form characteristic of the pits of the genus, PI.
xxvi, fig. 16. The number of pits is variable, but twenty or more exist in some fossil
species. Wings may or may not develop in Perna, Pl. xxv1, figs. 16-18, but the body
of the shellis produced obliquely or vertically as in Avicula or Malleus, Pl. xxvr, fig. 19.
In no species are teeth known in the adult and they seem to be purely hereditary char-
acters limited to early stages of development. The microscopic shell structure of Perna
is identical with that of Avicula, and the anatomy of the two genera is alike as far as

384 ROBERT TRACY JACKSON ON THE

observed. The numerous and peculiar cartilage pits, loss of hinge teeth and form of the
shell point out clearly that Perna is a side branch, off the ascendent line of the Aviculan
series. In these studies Perna is considered the main stem from which important
branches are given off. (Further description of Perna, see pp. 326, 327.)

Bakewellia appears to be a branch from Avicula,in the same line of variation as Perna,
on account of the series of cartilage pits. The form of adult Bakewellia is very similar
to the young of Perna, Pl. xxvu, fig. 1, so that it might appear to be in the line of ancestry
of Perna. An objection to this assumption is the high development of both adductor
muscles, whereas neither Avicula nor Perna has more than a slight development of the
anterior adductor. A dimyarian condition has been shown to be characteristic of the
prodissoconch stage of Ostrea, Pecten, Avicula, ete., and in Bakewellia, we merely have to
assume that the early anterior adductor was retained instead of abandoned in the ontogeny
of the individual. A similar case of the retention of the anterior adductor is seen in
Dimya, which has two adductors in the adult though belonging to a typical monomya-
rian group. The form of Bakewellia is sub-rhomboidal, oblique and is quite comparable
to Rhombopteria. The cartilage pits, primitive form of the valves and large anterior
adductor muscle, point to the conclusion that Bakewellia is a specialized and partially
retrogressive genus.

Gervillia has important characters of form and cartilage pits connecting it with Perna
from which genus it was probably derived. Hornesia, an allied genus, is apparently a
modified branch from Gervillia differing principally in its inequivalvular form.

TInoceramus was probably derived from Perna, through the interme diation of Gervyillia.
The forms of the genus which are rhomboidal and the most equivalvular, as Jnoceramus
crispt, Mant., are to be considered the more normal, least modified forms of the genus as
they approach nearest to the condition found in ancestral genera. ‘The oval and pli-
cated genus Actinoceramus is a modified branch from Inoceramus. Volviceramus is a
still more modified, oval and highly involute genus probably derived from Actinocera-
mus.

Pernostrea from the Jurassic is a descendant of Perna which attached itself by the
umbo of the left valve and the consequent conditions of fixation modified the shell, in-
ducing an ostrean form. The shell has a series of cartilage pits like Perna and a form
which closely resembles an oyster. (Further description of Pernostrea, see p. 325.)

From the studies on the development of the oyster, sections Tv—vil, we have come to
consider Ostrea as close akin to and derived from Perna (pp. 326-327). The points of
likeness are features in the anatomy of the soft parts, the prodissoconch and the struct-
ure of the right, free, least modified valve. The points of difference are the absence of
the foot, the highly modified attached valve and the single cartilage pit of Ostrea. The
first two features of difference are paralleled in Pernostrea and are ascribed to the effects
of the condition of cemented fixation (section vit). The oyster is a highly modified
form and the reduction of the several cartilage pits of Perna to a single one in Ostrea
is considered as part of the specialization of the genus. When young, Perna also has
but a single cartilage pit like the oyster; compare fig. 30, p. 8329 with Pl. xxtvy, fig. 19.
Ostrea is the least modified member of the Ostreadze, as it nearest approaches to Perna
from which it was apparently derived. Gryphza is a more highly modified genus and

PHYLOGENY OF THE PELECYPODA. 385

Exogyra is the most highly modified member of the family. Therefore they are con-
sidered as derived from Ostrea and are in the ascendent series of this peculiar group.
The marked sinuosity in the left valve of Gryphza and Exogyra is compared to a similar
sinuosity in young Ostrea, Pl. xxv, figs. 5-6, and is ascribed to the local retraction of
the mantle border at that area where excurrent action takes place (p. 307).

Exogyra is compared to the adult of species of Chama (the base of the Chamide) and
to the young of Caprinula, Pl. xxv, fig. 20, Hippurites and Radiolites (extremes of the
same group) which in external form it closely resembles. ‘Tne form of all these genera
is believed to be built up as the result of cemented fixation with its accompanying in-
fluences of environment. When a Pelecypod becomes attached by cemented fixation
the first form’induced by the new conditions is a form like Ostrea. A further modifica-
tion in this line of variation is a form like Exogyra. The next step in the same line of
variation isa form like Caprotina or Caprinula, Pl. xxv1, figs. 10 and 8. Finally the last
step in the same line of variation is a form like Hippurites or Radiolites which are coni-
cal or cup-shaped with a considerable degree of superficial radial symmetry (see section
vil, and especially foot-note, p. 322).

Alectryonia which is strongly plicated in both valves and adherent by recurved clasp-
ing spinous processes, is undoubtedly a side issue
from Ostrea and not in the ascendent line of the
family.

I have not had opportunity to study the Pinnidee
satisfactorily as attempts to obtain very young spec-
imens have so far failed. The following is offered
as the probable genetic relations of the family. Pa-
lwopinna (Hall) in form is not far removed from! OT of vos net fo
some species of Leptodesma and it can be naturally cat. No. 2327. X4 diam. (Drawn by the au-
derived from that genus by reducing the posterior ‘°™)
extension of the wing and increasing the length of the antero-posterior axis. From Pa-
lzopinna or Aviculopinna to Pinna, is a natural step introduced by an increased acute-
ness anteriorly and a greater curvature of the lines of growth posteriorly. In support of
this view we find that young Pinnas (20mm. more or less in length), fig. 53, are obtuse
anteriorly, and posteriorly have an obtuse abbreviated extension of the hinge line, like
Aviculopinna. These features gradually fade out in succeeding growth and give place
to the acute anterior and roundly curved posterior border of the valves which is charac-
teristic of the genus. The Pinnidz should, I believe, be considered a degradational
group because of the great development of prismatic cellular tissue and loss of nacreous
tissue (a weakening character), loss of teeth on the hinge line, and relatively large de-
velopment of the anterior adductor muscle (a larval character, highly reduced or want-
ing in the ascendent series of related groups).

Atrina is a peculiar genus of the Pinnidee of twisted irregular growth. In the later
growth of the shell there is a direct fusion of the two valves both on the dorsal and ven-
tral border though in less degree on the latter. In a specimen of A. saccata sp., L. in
my collection, 114 mm. in length, the fusion dorsally occupies a length of about 50 mm.

Ss

aD IE T =o
Sa NWT
i ] /

j ; / fii | HM/ ih}
Liiyy

“uM

386 ROBERT TRACY JACKSON ON THE

and in this area the lines of growth pass directly across from one valve to the other
without any indication of the hinge line. The nacreous tissue is highly reduced in ex-
tent. The shell has become in a physiological sense a flattened tubular univalve which
may be compared in a measure in its resultant form to an Aspergillum (vide note 3, p.
376). In the young of Atrina the two valves are perfectly separable and have not yet
acquired the twisted abnormal features characteristic of the adult. Atrina should be con-
sidered a very retrograde genus as it is weak and abnormal to a high degree.

Trichites, another genus derived from Pinna, is peculiar for its excessive development
of prismatic cellular tissue which may reach an inch or more in thickness. It is an ex-
treme modification of the Pinnidz of highly irregular growth and without hinge teeth.
The great development of prismatic cellular tissue in the Pinnidz disproves the state-
ment commonly made in text-books that this structure is only formed by the edge of the
mantle. The formation of prismatic tissue commonly is limited to that portion of the
shell built by the edge of the mantle, as in Avicula, Ostrea, Anodon, etc., the deeper-seated
portions of the mantle in such cases building nacreous or sub-nacreous tissue. In Pinna,
Atrina, Trichites and Malleus, however, the deeper seated parts of the mantle as well as
distal edges must build prismatic tissue or this structure would not attain the thickness that
it does, and further we may actually see in fresh specimens that the nacreous structure, as
in Atrina and Malleus, is limited to a very small area near the umbonal portion of the shell.

Pterinea is closely related to Aviculoid types and yet is not in the ascendent line of
Avicula proper. The shell is nasute anteriorly and the wing may or may not be pro-
duced posteriorly. The ligamental area of the valves is striated longitudinally (compare
with Meleagrina); but no cartilage pit is differentiated, although this is characteristic
(either as a single or several pits) of Avicula and all its various descendants as represent-
ed in the accompanying table. The young of Pterinea as shown by Goldfuss' in P. levis
is rhombic in form and bears a very close resemblance in contour to Rhombopteria, fig. 51,
p- 380. Pterinea is considered as a side branch, probably derived from Rhombopteria.

Actinopteria (Hall) and Ptychopteria (Hall) are genera described by Professor Hall as
differing from Pterinea in the absence of a strong striated igamental area and in the ab-
sence of cardinal and lateral teeth. Following his figures and descriptions, it is seen that
Actinopteria more nearly approaches to the characters of Pterinea than does Ptychop-
teria; therefore, in our table they are arranged in that serial order. Both genera appear
to be on a side branch from Pterinea; not ina direct line of ascent because they are char-
acterized by a loss, not an accentuation or modification of Pterinean features.

Pterinopecten (Hall),a Devonian genus of the Aviculopecten group, is characterized by
a long hinge line, with ears not well defined, being simple expansions or extensions of the
hinge line. The various species are frequently described by Hall as sub-rhomboidal, espec-
ially when young, and he mentions that some species resemble the young of Aviculopecten
and Lyriopecten. Pterinopecten dignatus, Hall, P. conspectus, Hall, and similar species
bear a close resemblance to the young of Pecten irradians, Pl. xxvut, figs. 12-13. They
resemble the figures cited in both valves as the right valve has a deep byssal sinus. The
young of Pterinopecten is similar to the adult of Rhombopteria as I have observed in P.

1 Petrefacta Germanie, Pl. exrx, fig. la.

PHYLOGENY OF THE PELECYPODA. 397

intermedius, Hall, P. vertumnus, Hall, and P. conspectus, Hall. Pterinopecten is nat-
urally derived from Rhombopteria by the introduction of slight ears on the hinge line.
It is the simplest known form of the Pectenide and is traceable in the early stages of
development of Aviculopecten and Pecten; therefore, I shall consider Pterinopecten as
directly descended from Rhombopteria and as the base of the Pectinidz from which all
later forms were derived. (Further discussion of Pterinopecten, see p. 346.)

Aviculopecten (McCoy), as amended by Hall, includes forms with a hinge usually
shorter than the transverse diameter of the shell and with both ears well defined. A caru-
lage pit is described by Hall in several species of this genus, also in Lyriopecten and other
allies. A deep byssal sinus occurs in the right valve of many species of Aviculopecten
and diverging furrows exist on the hinge line in some species similar to those found in
certain recent Pectens as P. maximus. A finely preserved specimen of Aviculopecten
Jasciculatus, Hall, shows that when young (3 mm. in height) the shell was rhomboidal
and in its general contour, resembled Rhombopteria. During later growth a long hinge
line and moderately developed ears develop, making the shell at this stage comparable to
Pterinopecten; later still, the characters of a shorter hinge line and pronounced ears give
the shell the characters of the genus Aviculopecten. Similar observations have been
made on the development of Aviculopecten princeps, Hall, and a few other species. Avi-
culopecten, when young, resembles Ptermopecten and when adult resembles the young
of Pecten, Pl. xxvin, fig. 2. It is, therefore, unquestionably intermediate between these
two types. (Further discussion of Aviculopecten, see p. 346.)

From Aviculopecten we pass to Pecten in the direct line of ascent. Pecten has a rel-
atively short hinge line, both ears well defined, but the ventral (anterior) a little larger;
either a slight or no byssal sinus exists in the ear of the right valve. The ligament and
cartilage are as in Aviculopecten. In the development of Pecten irradians and other
species (sections Ix-x) we have shown a remarkable series of phylogenetic characters
which may be briefly alluded to. The prodissoconch, p, of Pecten is dimyarian, PI.
XXvu, figs. 8-9, Pl. xxviu, figs. 9-10, and is traced to the Nuculoid radical of the whole
group (see table). The first nepionic stage, Pl. xxvutt, fig. 9,is rhomboidal, oblique, and
is traced to Rhombopteria; the later nepionic stage, Pl. xxvin, figs. 12-13 has a long
hinge line, and ears which are simple expansions of the hinge margin; these features
and the deep byssal sinus are traceable to Pterinopecten. The nealogic stage of Pec-
ten, Pl. xxv, fig. 2, has a relatively shorter hinge and well developed ears, and this
stage is traceable to Aviculopecten. Only when the individual has advanced consid-
erably in age are the full characters of the adult, fig. 37, p. 333, acquired. The first
appearance of Pecten geologically is doubtful as the limitations of the genus among pa-
lzeozoic fossils has not been fully worked out. All of Professor Hall’s Devonian Pectin-
idee are, however, referable to other genera and we will provisionally consider Pecten as
dating from the Carboniferous. From Pecten as a base many genera have sprung as
will be presently considered.

Lyriopecten (Hall) differs from Aviculopecten in possessing a short hinge and a small
anterior ear. It also differs from Pecten in this last character and it is considered as a
side branch from Aviculopecten. In support of this view we find that a specimen of
Lyriopecten orbiculatus Wall, which had a short hinge in the adult, in the young (as ascer-

388 ROBERT TRACY JACKSON ON THE

tained by following the lines of growth) had a long hinge line, with a comparatively slight
development of the ears.

Crenipecten, Hall, is a genus like Aviculopecten in form, but with crenulations on the
hinge margins. Pernopecten, Winchell, has a large triangular cartilage pit and a crenu-
lated hinge plate. The crenulations in these genera are described as cartilage pits. A
crenulated hinge plate is found in the young of Pecten tenuicostatus, fig. 38, p. 348, and
P. monotimeris. Tt has also been found more or less developed in the young of several
recent species of Pecten by Dall (16). He considers the crenulations as an order of denti-
tion! and the crenulations in the fossil genera cited are probably of similar significance.
As evidence against the view of considering them as cartilage pits, it should be observed
that in Perna, etc., where numerous cartilage pits are typical, the primitive pit as well as
succeeding ones assumes the perpendicular sides characteristic of the pits of the genus,
figs. 30-31, p. 329, and Pl. xxv1, fig. 16. Crenipecten and Pernopecten are considered as
side issues from Aviculopecten.

Lima has many features of anatomy and shell form linking it with Pecten; it is, there-
fore, presumably a side issue from the early members of this genus. (Further deserip-
tion of Lima, see p. 353.)

The genus Chlamys (Bolton), characterized by the great development of the ventral
(anterior) ear and correlated reduction of the dorsal ear, is an undoubted branch from
Pecten. When young, Pl. xxvumt, fig. 14, both ears are nearly equally developed as in
Pecten. Chlamys is a swing in one direction as Lyriopecten is a swing in the other di-
rection, for the latter has the anterior ear very reduced and the posterior ear largely de-
veloped. (Further description of Chlamys, : see p. 348.)

Vola (Klein) is anear ally of Pecten; in the adult it is highly inequivalvalar, the right
valve is deeply concave, while the left rales is flat or even slightly concave. When young,
however, it is nearly or quite equivalvular. In a finely preserved specimen from Japan,
a prodissoconch and nepionic stages were observed in both valves, similar to those char-
acteristic of Pecten irradians, Pl. xxvuiu, figs. 9-12.

Amusium (Schumacher) (see Dall) is characterized by radiating, internal ribs, a short
hinge line, nearly equal ears and a high reduction or complete absence of the byssal notch.
The ears of the right valve may be produced beyond the limits of those of the left, on the
hinge border, as in Amusium japonicum, thus making an abbreviated approach to the
condition characteristic of the genus Entolium, Meek. Commonly a pair of ridges di-
verge from the hinge line and terminate in tubercular processes close to the auricular
sinus of either ear. Dall has named them auricular crurze. Crure similar to those of
Amusium may be traced in most species of Pecten, though they are often very reduced,
the terminal tubercle at the auricular sinus alone existing. The function of the crurze
is not understood. Dr. Dall, in his studies of deep-sea Amusium, notes the existence
of prismatic structure in several species, in which, however, it was not limited to the right
valves, as shown in our studies of young Pecten (section x). Amusium is considered
a side issue from Pecten, rather than in the line of ascent of the group.

Anomia is doubtless a side issue from Pecten as discussed (section x1m). In the
young there is, as in Pecten, a deep byssal sinus in the right valve, Pl. xxrx, fig. 6.

1 See reference to a paper by W. H. Dall, foot-note, p. 376.

PHYLOGENY OF THE PELECYPODA. 389

Later the byssal sinus becomes enclosed to form the byssal foramen as described (fig.
41, p. 354). The peculiarity of the shell and many features of anatomy are clearly at-
tributable to the special conditions of environment of the animal (full discussion, sec-
tions XII—XxII1).

Hnigma is an extreme modification in the Anomia line of variation. The right valve
is very attenuated, the umbos are far removed from the hinge line, and the byssal fora-
men is sub-central. A specimen of d7xigma enigmatica, Chemn., in my collection is ob-
liquely oval, closely resembling the form found in young Anomia, Pl. xxrx, fig. 5. The
left (upper) valve from the umbo to the periphery of the shell shows the fusion of the
lines of growth that takes place at that area as in young Anomia, Pl. xxrx, fig. 8.
Mnigma, therefore, though with full probability a descendant of Anomia, in some char-
acters resembles the nepionic period of that genus.

Placunanomia differs from Anomia in the calcareous union of the borders of the
byssal foramen anteriorly; it also differs in the development of diverging lamelle from
the hinge border. By the encroachment of the borders of the byssal foramen the byssus
in later life tends to become solidly fixed in the lower valve. In Placunanomia macro-
chisma, Desh., the byssal foramen is open in specimens two inches in height and is only
partially closed in specimens somewhat over four inches in height, as far as observed in
the few specimens at command. In the single specimen of Placunanomia rudis, Brod.,
observed, the byssal foramen is entirely closed. Placunanomia is off the direct line of
ascent in the Anomia order of modification, and is transitional between Anomia on one
side and Carolia and Placuna on the other, as demonstrated by Fischer.

Carolia is a genus which seems clearly to be intermediate between Placunanomia and
Placuna, as demonstrated by Fischer. When young, Carolia has a deep byssal sinus like
adult Anomia. Later the sinus becomes progressively deeper like Alnigma, then the bor-
ders of the notch become joined on the anterior border like Placunanomia and finally the
byssal foramen is cut off by shelly growth, when the shell approaches to the condition of
Placuna.

Placuna in its anatomy is essentially like Anomia but the mantle border has a cur-
tain, like Pecten, as shown by Woodward. The structure of the shell is also very simi-
lar to Anomia. In the New York State Museum I observed an inclosed byssal plug in
young specimens of Placuna sella. ‘The plug occupies the base of a foramen which is
filled up by calcareous growth similarly to Anomia; but in the specimens observed the
plug was already cut off from the body of the animal by a calcareous secretion that had
taken place on the proximal side of that area. A similar condition to that observed in
Placuna sella is described and figured by Fischer in young P. papyracea. The crure
of Placuna, though existing only in the right valve of the adult (the left valve having
corresponding depressions), are equally developed in both valves of the young. The
crure of Placuna seem to be a modification and posterior extension of the knob-like car-
tilage support found in the right valve of Anomia. (Further discussion of Placuna, see
p- 362.)

Dimya, on account of some features of anatomy and shell structure, is placed by
Dr. Dall near to Pecten. The shell being ostreaform, he considered that the genus lay be-
tween Pecten and Ostrea. I have shown I think conclusively that the ostrean form of

MEMOIRS BOSTON SOC. NAT. HIST., VOL. IV. 52

390 JACKSON ON THE PELECYPODA.

shell is induced by the conditions of fixation in any attached Pelecypod; therefore Dimya
should be considered as a highly modified branch from Pecten, which is ostreaform be-
sause attached. It is shown in this paper that the prodissoconch of Ostrea, Pecten, etc.,
is dimyarian, therefore we may fairly infer that the existence of two adductor muscles in
adult Dimya is due to the retention of both muscles which are doubtless existent in the pro-
dissoconch stage of that genus. It is to be noted also that the gills of Dimya are in an
archaic condition of simplicity as described by Dall. It is shown in the table of genera,
p- 391, that Pecten and Ostrea are very widely separated genera, therefore no one genus
can properly be considered as intermediate between these two types. Dimya is consid-
ered as a side-issue from Pecten, possessing many archaic and modified features which
render it a highly retrogressive or degradational form. (Further description of Dimya,
see pp. 306, 524.)

Pedum is evidently a descendant of Pecten. In the nepionic period, Pedum is purely
pecteniform in both valves, not yet having acquired the peculiarities of the genus.
Later, the byssal sinus is enclosed so as to become, in a measure, foraminal, and the shell
is apparently closely related to the object of byssal attachment. The right valve is the most
highly modified, although both valves are modified, and in the adult condition bear little
resemblance to Pecten. The peculiar form of the shell of Pedum, it may be reasonably
inferred, is due to the habit of close byssal fixation.

Hinnites, as already discussed, is an important genus in our conclusions. Up to the
nealogic period, it is free or attached by a byssus and pecteniform, Pl. xxv1, figs. 3-4.
It then becomes permanently attached by soldering the right valve to an object of sup-
port and with this change in habit we find that the irregular ostreaform growth is
introduced as fully discussed (section vit). Hinnites is an undoubted descendant from
Pecten and owes the peculiarity of its form to its adopted habit of fixation. (Further
description of Hinnites, see pp. 324, 350.)

Spondylus, like Hinnites in its nepionic stage, is pecteniform, Pl. xxvn, fig. 4; then
becoming attached at this early stage, the irregular growth characteristic of the genus
is introduced and the shell becomes spondyliform. The right attached valve is the most
highly modified, and the left valve is the least modified and nearest approaches to Pecten
in form (discussed in sections vi and x1). Spondylus is considered as descended from
Hinnites and is more highly modified than that genus because it earlier adopts the con-
dition of fixation. The degree to which a talon is developed in Spondylus varies, being
least developed in species which are most nearly equivalvular (p. 852). The peculiar in-
terlocking teeth of Spondylus are believed to be a modification of the auricular crure
characteristic of most Pectens and especially Amusium. The several teeth of Spondylus
in the young seem directly traceable to the knob-like processes of the crure existing at
the auricular termination of the crurz in Pecten. In Spondylus apparently by the short-
ening of the crure, the distal knobs of the same are brought in close proximity to the
hinge line, where they are further modified so as to produce interlocking, articulating
processes. If this assumption in regard to the teeth of Spondylus is correct, they are
teeth of different morphological nature from ordinary cardinal teeth and are to be com-
pared to those only as to mechanical equivalents. (Further consideration of Spondylus,
see pp. 318, 323 and 351.)

GENEALOGICAL TABLE OF THE AVICULIDA AND THEIR ALLIES.

(391)

Volviceramus (Cret.)

| Meleagrina (Jura, )

Actinoceramus (Cret.) Exogyra (Jura.)

| |

TInoceramus (Trias. ) Gryphea (Jura.)

Hornesia (Trias. )

Gervillia (Trias. ) Ostrea (Trias. ) J

|

Bakewellia (Perm. ) Perna (Trias. ) |

Trichites (Jura. )

|

—Pinna (Dev.) |
|
Ptychopteria (Dey.) Ayiculopinna (Carb.)

| Paleopinna (Dey.)
Actinopteria (Dey.) |

a Pterinea (Low. Sil.)

ae

Alectryonia (Trias. )

Atrina (Rec.)

capo Leptodesma (Dey.)

Vulsella (Tert.) Malleus (Ree. )
| |

Monotis (Trias. )
|

Pseudomonotis (Dev.) Oxytoma (Trias.) Cassianella (Trias. )
Avicula (Dev.?) | | |

Placuna (Rec.)

Carolia (Tert.) Chlamys (Trias.) Vola (Cret.)
| Aenigma (Rec.)

7 |
Anomia (Jura. ) Spondylus (Trias. )

|
ates (Jura. ) . Hinnites (Trias. )

Pedum (Rec. )

Dimya (Tert.)

Lima (Carb.)
| Pecten (Carb.?) |

Pernopecten (Carb. ) ;

|
Crenipecten (Dey.) Lyriopecter (Dey.)

[aa sae Aviculopecten oj
|

Pterinopecten (Dev.)
ees ae (Low. and Up. Bi ae aie De )
Nuculoid Radical (Low. Sil.)

392 ROBERT TRACY JACKSON ON THE

Plicatula is an undoubted ally of Spondylus, as usually considered. As we have
shown, the shell of Plicatula becomes attached at the close of the prodissoconch stage,
and at once assumes the ostrean form, thus eradicating features of the nepionic stage
which (as in Spondylus and Hinnites) might render that stage comparable to Pecten.
Plicatula is considered the farthest removed from Pecten of its group, for in it the ostrean
form is earliest acquired and most completely emphasized in the shell. (Further descrip-
tion of Plicatula, see p. 352.)

The table on the preceding page represents graphically the genetic relations expressed
at various places in this paper and especially in the present section. When genera are
considered as derived from other genera and in the ascendent line of the group, they are
placed directly over one another, as in the case of Gryphxa and Ostrea. When a genus
is derived from another, but is not considered as in the ascendent line of the series then
it is placed as a side branch, as Alectryonia from Ostrea. Broken lines are used where
the affinities expressed appear to be open to doubt. With each genus is expressed in
brackets its earliest geological appearance, the data for these statements being for the
most gathered from Zittel’s Paleontology. The following abbreviations are used: Low.
Sil., Lower Silurian; Up. Sil., Upper Silurian; Dey., Devonian; Carb., Carboniferous;
Perm., Permian; Trias., Triassic; Jura., Jurassic; Cret., Cretaceous; Tert., Tertiary;
Rec., Recent.

ERRATA.

Page 281, note, the prodissoconch is described in a few genera not included in this list. For such additions see sections
VIII, X and XIV.

Page 291, paragraph 5, line 5, for Apysia, read Aplysia.

Page 291, paragraph 6, attention should have been called to Dr. W. Branco’s monograph: Beitrage zur Entwickelungs-
geschichte der fossilen Cephalopoden. Palzontographica, B. 27, 1880-81.

In section IV, at several points the oyster is spoken ef as showing concentration of development, or as a concentrated
type. In these cases for concentration, or concentrated, read acceleration, or accelerated, as that word is more in accord with
the principle involved.

Page 314, line 14, for overlyiny, read underlying.

Page 3380, line 5, for Leioptera, read Leiopteria.

Page 333, line 8, from bottom of page, for shortening its length. It, read then shortening its length, it.

Page 334, paragraph 3, line 13, for exists, read were existent.

Page 335, paragraph 3, line 10, for at that time, read therefore.

Page 336, paragraph 3, line 8, for specimen if kept, read specimen kept.

Page 336, paragraph 3, line 12, for condition, read stage.

Page 337, paragraph 2, line 4, for Jf one break the byssus loose it, read Jf the scallop breaks loose, the byssus.

Page 3838, paragraph 2, line 6, for over, read at.

Page 345, line 6, for Actinoptera, read Actinopteria.

Page 361, description of cut, for 60, read 30.

Page 366, line 3, for adults, read young shells.

Page 370, paragraph 3, line 12, for ViJZread VII.

PHYLOGENY OF THE PELECYPODA. 393

XVII. LIST OF PAPERS QUOTED OR REFERRED TO IN THE TEXT.

In the text, numbers in parentheses referring to this list are appended to references when there is need of it, as when there
are several references from one author in this list, or when with few references, there might be ambiguity as to which
one is referred to. When but one reference is given in the list, or the reference intended is clear from the context, no
number is appended in the text.

28.

Apams, H. & A. Genera of Mollusca. London, 1858.

Batrour, F. M. Comparative Embryology. London, 1880.

BaRRANDE, J. Systéme Silurien de la Bohéme.

Bouran, L. Recherche sur la Fissurella. Archiv. de Zool. Exp. et Gen., 2& serie, T. ur, suppl.

Bronn, H. S. Die Klassen und Ordnungen des Thierreichs. 1862-66.

Brooks, W. K. Embryology of the Fresh Water Mussels. Proc. Am. Assoc. Ady. Sc., 1875.

The affinities of the Mollusca and Molluscoidea. Proc. Bost. Soc. Nat. Hist., Vol. xvi, 1876.

— Development of Fresh Water Pulmonates. Johns Hopkins Uniy., Biol. Lab. Studies, 1879.

Development of the Oyster. Johns Hopkins Uniy., Biol. Lab. Studies, 1, 1879-80.

Butscuu, O. Entwickelungsgeschichtliche Beitrage. Zeitschr. fiir wissensch. Zool., B. xx1x, 1877.

CarpPreNnTER, W. B. On the Microscopical Structure of Shells. Reports, British Assoc. 1844, 1848.

CHALLENGER Reports, Vol. xm, Part xxxv, Report on the Lamellibranchiata, by E. A. Smith.

Vol. xxvu, Part Lxxty. Anatomy of Deep Sea Mollusca, by Paul Pelsener.

Corr, E. D. Origin of the Fittest. New York, 1887.

Coquanp, H. Monographie du Genre Ostrea, Terrain Crétacé. 1869.

Dati, W. H. Report on the Mollusca (from the Reports of the Dredgings of the Blake Expedition). Bull. Mus.
Comp. Zodél. 1886. Dall; see also page 376.

FIscHerR, Paut. Manuel de Conchyliologie. Paris, 1881.

Forses & Hantry. A History of British Mollusca. London, 1853.

Goutp, A. A. Edited by W. G. Binney. Invertebrata of Massachusetts. Boston, 1870.

Haut, James. Natural History of New York, Palzontology. (See p. 381, note.)

Hatscuexk, B. Ueber Entwickelungsgeschichte der Teredo. Arbeiten aus dem Zool. Inst. Wien, Vol. mr, 1881.

Entwickelung der Trochophora yon Eupomatus uncinatus. Arbeit. aus dem Zool. Inst. Wien. T. v1, 1886.

Horst, R. Contribution to our Knowledge of the Development of the Oyster, O. edulis. Bull. U. S. Fish Comm.,
Vol. o, 1882. Translation.

The Development of the Oyster. Rep. U. 5. Fish Comm. for 1884. Washington, 1886. Translation.

Huxtery, T. H. Anatomy of Invertebrated Animals. New York, 1878.

Oysters and the Oyster Question. Eng. Ill. Mag., Oct. and Noy., 1883.

Hyarr, AteHeus. Fossil Cephalopods of the Museum of Comparative Zodlogy, Embryology. Bull. Mus. Comp. Zo6l.
Vol. m1. B

Guides for Science Teaching. The Oyster, Clam and other Common Mollusks. Boston, 1880.

Transformations of Planorbis at Steinheim, with Remarks onthe Effects of Gravity upon the Forms of Shells

and Animals. Proc. Am. Assoc. Ady. Sce., Vol. xxrx, 1880.

Larval Theory of the Origin of Tissue. Proc. Bost. Soc. Nat. Hist., Vol. xxm, 1884. See also Proc. Bost.

Soc. Nat. Hist., Vol. x, 1866, pp. 302-303.

Parallelism between the Individual and Order in Tetrabranchiate Cephalopods. Mem. Bost. Soc. Nat. Hist.,

Vol. 1, 1867.

The Evolution of the Cephalopoda. Science, Vol. m1, Nos. 52-53, 1884.

Values in Classification of the Stages of Growth and Decline, with Propositions for a new Nomenclature. Proc.

Bost. Soc. Nat. Hist., Vol. xxm, 1888.

Genesis of the Arietide. Smithsonian Contributions to Knowledge. Washington, 1889.

JacKSON, Ropert T. Catching Fixed Forms of Animal Life on Transparent Media for Study. Science, Vol. x1,
No. 275, 1888.

The Development of the Oyster with Remarks on Allied Genera. Proc. Boston Soc. Nat. Hist., Vol. xxi,

1888.

Kine, W. Monograph of the Permian Fossils of England. Palzontographical Society, 1850.

DE Koninck, L. Animaux Fossiles Carbonifére, Belgium. 1842-44, Supplt.

Lacaze-DuTuiers, H. Mémoire sur lOrganisation de VAnomie. Ann. des Sc. Nat., rve serie, T. m, 1854.

Mémoire sur les Organes Genitaux des Acéphales Lamellibranches. Ann. des Sc. Nat., Ive serie, T. 1m, 1854.

Mémoire sur le Développement des Branchies des Mollusques Acéphales Lamellibranches. Ann. des Sc. Nat.,

Ive serie, T. v, 1855.

Histoire de ?Organisation et du Développement du Dentale. Ann. des Sc. Nat., rve serie, T. v1, 1856. Con-
tinued in Vol. wu, 1857.

LaNKESTER, E, Ray. Contributions to the Developmental History of the Mollusca. Phil. Trans. Vol. 165, Part 1,
1875.

394. ROBERT TRACY JACKSON ON THE

44,

45.
46.

Doran a
SSan

a
—

62.

79.

Lankester, E. Ray. Mollusca. Encyclopedia Britannica, Ninth Edition, 1883.

Lea, Isaac. Observations on the Genus Unio, Vol. vi. Philadelphia, 1857.

Lovén, S. Bidrag till Kannedomen om utvecklingen af Mollusca Acephala Lamellibranchiata. Kongl. Vetenskaps
Akademiens Handlinger. Stockholm, 1848. Translated into German in Archiy fiir Naturgeschichte, Vol. xv, 1, 1849.
A German reprint with plates was also published by Professor Lovén, in 1879.

Mirzukunri, K. Structure and Significance of some Aberrant Forms of Lamellibranchiate Gills. Quart. Journ. Mic.
Sc., 1881. Reprinted in Johns Hopkins Biol. Lab. Studies, Vol. m, 1883.

Morse, Epwarp S. On the Relations of Anomia. Proc. Bost. Soc. Nat. Hist., Vol. xv, 1870-71.

p’ORBIGNY, A. Paléontologie Francaise, T. 3 et T. 4. 1843-47.

OweEN, RicHarp. Structure of the Shell in the Water Clam (Spondylus varius). Proc. Zool. Soc., Vol. v, 1837.

On the Relative Positions to their Constructors of the Chambered Shells of Cephalopods. Proc. Zool. Soc.,

1878.

Patren, W. The Embryology of Patella. Arbeit. Zool. Inst. Wien, B. v1, 1886.

Eyes of Molluscs and Arthropods. Mittheilung d. Zool. Station zur Neapel, B. v1, 1885-6.

Peck, R. Horman. The Structure of the Lamellibranchiate Gill. Quart. Journ. Mic. Se., 1877.

Picter, F. J. Mollusques Fossiles de Genéve.

Purpir, ALEX. The Anatomy of the Common Mussels. Colonial Museum and Geol. Survey Dept., Wellington, New
Zealand, 1887.

Ras, C. Ueber die Entwickelungsgeschichte der Malermuschel. Jenaische Zeitschr., Vol. x, 1876.

Rerye, L. A. Conchologica Iconica, London.

REICHEL, L. Ueber das Byssusorgan der Lamellibranchiaten. Zool. Anz., x, pp. 488-490.

Ryper, Jonn A. An Account of Experiments in Oyster Culture and Observations Relating thereto. Rep. Maryland
Fish Comm., 1881.

The Microscopic Sexual Characteristics of the American, Portuguese and Common Edible Oyster of Europe

Compared. Bull. U. 8S. Fish Comm., Vol. um, 1882.

On the Mode of Fixation of the Fry of the Oyster. Bull. U. 8. Fish Comm., Vol. m, 1882.

On the Metamorphosis and Post-larval Stages of Development of the Oyster. Rep. U.S. Fish Comm., for

1882. Washington, 1884.

A Sketch of the Life-history of the Oyster. Published in Dr. C. A White’s Monograph (this list, No. 73).

A Contribution to the Life-history of the Oyster, in ‘‘The Fisheries and Fishery Industry of the United

States.” Section 1, 1884, pp. 711-758.

The Byssus of the Young of the Common Clam (Mya arenaria, L.). Am. Nat., Jan., 1889.

Suarp, Benjamin. Remarks on the Phylogeny of the Lamellibranchiata. Proc. Acad. Nat. Sc. Phil., 1888.

Scumipt, FERDINAND. Beitrag zur Kentniss der Postembryonalen Entwickelung der Najaden. Archiy fiir Naturge-
schichte, L1, B. 1, 1885.

SowrrsBy, G. B. The Genera of Recent and Fossil Shells. London, 1821-26.

Sroxiczka, Ferp. Cretaceous Pelecypoda of Southern India. Mem. Geol. Survey of India, 1871.

Tryon, GrorGr W., JR. Structural and Systematic Conchology. Philadelphia, 1882.

Verrritt, A. E. Invertebrate Animals of Vineyard Sound, etc. Rep. U. 8. Fish. Comm. for 1871-72.

Wuirr, C. A. A Review of the Fossil Ostreide of North America. Rep. U. 8. Geol. Survey, Iv, 1884.

On Mesozoic Fossils. Bull. U. S. Geol. Survey, 1884.

WHI?FIELD, Roperr P. Brachiopoda and Lamellibranchiata of the Raritan Clays and Greensand Marls of New Jer-
sey. Monogr. U. 8. Geol. Survey, 1885.

WinsLow, Francis. An Account of an Experiment in Artificially Fertilizing the Ova of the European Oyster.
Rep. (—) Maryland Fish Comm., 1881.

Woopwarp, HeNry. On the Structure of Camerated Shells. Pop. Sc. Rev., Vol. x1, April, 1872.

Woopwarp, 8. P. Some Account of Barretia. The Geologist, Oct , 1882.

Manual of the Mollusca. Fourth edition, by Ralph Tate, London, 1880.

ZIEGLER, H. ERNsT. Die Entwicklung von Cyclas cornea, Lam. Zeitschr. fiir wissensch. Zool., Vol. x11, 1885.

ZirvEL, Kart A. Traité de Paléontologie. Translation by Dr. Charles Barrois, 1887.

XVIII. EXPLANATION OF PLATES.

The figures which illustrate this paper were drawn by Mr. J. H. Emerton or the author, with a few exceptions which
are noted in the descriptions. All figures of microscopic material were either drawn directly from specimens with the aid
of a Zeiss camera lucida, or were redrawn from camera drawings. The locality and date of collecting specimens are in-
dicated in the description of figures excepting cases where those facts are given in the text. The magnification of
figures and the principal references to the pages in the text where the figures are described or discussed are given as an
aid to readers. Specimens figured in the plates or text, unless otherwise mentioned , are in the collection of the author.
I have to thank Mr. Emerton for the pains he has taken in drawing the figures which largely illustrate this paper.

PHYLOGENY OF THE PELECYPODA. 395

Besides the plates the following figures illustrate the text:
Figs. 1-8. Orthoceras elegans and O. politum, protoconch and apex of conch, by kind-

ness of Professor Hyatt, : B ¢ p. 292
Fig. 9. Vermetus radicula, adult, .- 6 G : ¢ p. 294
Figs. 10-18. Ostrea edulis, embryology, é ¢ 5 : 5 : & 5 A 5 ; 290.
Figs. 19-22, © %€ ot US Geta. 4 Ayia ee ae eee aie 297.
Figs. 23-24. “ ee eee ee 299.
Fig. 25. Ce a OOS ae “ : : : 5 “ ¢ C 300.

Fig. 26. Ostrea edulis, prodissoconch, . : > ; .
Figs. 27-29. Ostrea edulis, the spat, : : “ " o 5

314.
Figs. 30-31. Perna ephippium, young, é © 6 ° 5 : 329.
Figs. 32-34. Avicula sterna, se . B A 330.
Fig. 35. Cassianella decussata, young, . 5 ° 5 5 . 332.
Fig. 36.. Vulsella rugosa, young, .- : 5 5 5 ° - 7 5 5 5 332.

Fig. 37. Pecten irradians, adult, ;
Figs. 38-39. Pecten magellanicus, young, 9 2

348.
Fig. 40. Pecten dislocatus, shell structure of young, 5 : 2 348.
Figs. 41-42. Anomia glabra, adult, c f 354.
Fig. 43. Anomia aculeata, young, : c 5 5 - : ° : 361.
Fig. 44. Unio cariosus, glochidium, é a 6 A ° 0 : A 367.
Fig. 45. Margaritana undulata, glochidium, A : : 7 4 367.
Fig. 46. Anodon cygnea, glochidium, 5 5 5 c 5 3 - 3 5 : é 367.
Fig. 47. Anodon fluviatilis, glochidium, ° 5 ° : 2 5 : . 367.
Fig. 48. Margaritana undulata, young, . 367.
Fig. 49. Anodon Jluviatilis, young, 5 5 > 5 5 . 367.

Fig. 50. Saxicava rugosa, young, 6
Fig. 51. Rhombopteria mira, adult, : 5 5 5 5

SUVEP PPP PUPP PPP EVP P PSP Py ES
[J\)
roe
a

380.
Fig. 52. Malleus vulyaris young, . 5 6 > 5 3 5 383.
Fig. 53. Pinna sp. indet., young, 4 A 5 é : 5 . : 5 5 > 385.

Figs. 9, 37, 41 and 42 are electrotypes furnished by the kindness of Professor A. E. Verrill. Figs. 10 to 24 and 26 to 29
inclusive are electrotypes furnished by the kindness of Colonel MacDonald, United States Commissioner of Fisheries.

PLATE XXIII.

Fig. 1. Embryo of Patella in the morula or mesembryo stage. Fig. la is an optical section of a similar stage (after
Patten). p. 291.

Fig. 2. Embryo of Paludina, gastrula, or metembryo stage: g, gastrula invagination; m, mesoderm cells (after
Butschli). p. 291.

Figs. 3-4. Embryo of Patella, trochosphere or neoembryo stage viewed from the dorsal side. In the middle of this
region the shell gland appears. Fig. 4 is a similar specimen viewed from the opposite or ventral side: b, blastopore (after
Patten). pp. 288, 291.

Fig. 5. Embryo of Pisidium pusilum, typembryo stage: s, shell gland; m, mouth; a, anus; f, foot. p. 291.
Fig. 6. Embryo of Pleurobranchidium, typembryo stage: s, shell gland and beginning of shell; v, velum; m, mouth;
J, foot. p. 291. (Figs. 5-6, after Lankester.)

Figs. 7-8. Glochidium of Anodon fluviatilis, from the gill cavities of the parent: ad, adductor muscle. This period is
referable to the phylembryo stage. The shell has a straight hinge line, and the remarkable spurs, sp, on its ventral border.
Fig. 8 is a different view of a similar specimen. Medford, November, 1888. XX 75 diam. pp. 301, 312, 367.

Fig. 9. Developing Modiolaria marmorata, phylembryo stage: ad, anterior adductor muscle, which alone exists at this
stage; m, mouth; s, shell, which has a straight hinge line.

Fig. 10. Developing Cardium exiguum, phylembryo stage: ad, anterior adductor, which alone exists at this stage;
v, velum; m, mouth; #, intestine; a, anus; s, shell, which has a straight hinge line.

Fig. 11. Developing Montacuta ferruginosa, phylembryo stage: ad, anterior adductor muscle, which alone exists at this
stage; v, velum; m, mouth; a, anus; f, foot; 7, 7, 7, retractor muscles of the velum; s, shell, which has a straight hinge
line. (Figs. 9-11 after Lovén.) Figs. 9-11, pp. 280, 289, 291, 312, 375.

Fig. 12. Adult Slava jibrosa, showing a very marked change in form of the shell: 2, nealogic stage; e, ephebolic stage
(after Barrande), life size. pp. 293, 294.

Fig. 13. Developing Dentalium, phylembryo stage: v, velum; s, saddle-shaped shell. pp. 289, 291.

Fig. 14. Developing Dentalium, older than fig. 13. The completed periconch stage, comparable to the period of the
completed protoconch in Gasteropods, ete.: v, velum; s, shell. (Figs. 13-14, after Lacaze-Duthiers.) pp. 289, 293.

Fig. 15. Young Goniatites compressus : p, ovoid protoconch, succeeded by true conch.' pp. 291, 311, 375.

1From Hyatt (27) after Sandberger.

396 ROBERT TRACY JACKSON ON THE

Fig. 16. Young Goniatites canaliculatus, var. gracilis: p, globular protoconch, succeeded by true conch.' pp. 291, 311.

Fig. 17. Young Pteropod: p, ovoid protoconch (after Bronn). p. 291.

Fig. 18. Tentaculites acuarius, Richter: p, protoconch (after Fischer). p. 291.

Fig. 19. Developing Sycotopus canaliculatus, completed protoconch stage: h, heart; f, foot; v, velum. The shell
shows distinct lines of growth (after camera drawing bythe author). X 24 diam., pp. 289, 291, 375.

Fig. 20. Developing Sycotopus canaliculatus, a period much later than the close of the protoconch stage yet still re-
taining a velum, v, which is enormously developed, and has a ramifying muscular system, shown in the figure: f, foot; ¢,
tentacles; e, eyes and m, mouth. X 12 diam., p. 289. Figs. 19-20 from egg cases, Buzzards Bay, September, 1889 (after
camera drawing by the author).

Fig. 21. Young Crepidula fornicata: p, spirally twisted protoconch, succeeded by the sub-conical nepionic growth.
Buzzards Bay. X 24 diam., pp. 292, 293.

Fig. 22. Crepidula fornicata, a series of adult specimens growing on one another and with axes arranged in relation to
currents. An oyster is attached to the base of the colony. Buzzards Bay. Life size, p. 354.

Fig. 23. Crepidula fornicata, much younger than fig. 21: p, the protoconch, which is sharply defined, is succeeded by the
early nepionic growth. Nantucket, July, 1889. 24 diam., pp. 292, 311.

Fig. 24. Young Acme testudinalis var. alveus. Interesting as showing instead of a protoconch, a plug-like cicatrix, c,
formed where the protoconch dehisced. The cicatrix corresponds in form with the base of a typical protoconch. (Com-
pare with fig. 23). X87 diam., pp. 292, 293.

Figs. 7, 8, 17, 18, 23 and 24 drawn by the author; all others drawn by Mr. J. H. Emerton.

PLATE XXIV.

Fig. 1. Embryonic Ostrea virginiana, growing ona plate of glass from a drain-pipe trap. It is the completed prodis-
soconch stage and dimyarian: a. ad, anterior adductor muscle; p. ad, posterior adductor; g, gills, which consist of simple
filamentous processes (compare with fig. 6); v, velum; pl, palps; a, dnus; m, mantle border, which is broad and bears
no tentacles. The shell is wholly prodissoconch growth, no spat growth having taken place, and is cemented to the glass
by the ventral border of the left valve. The early straight hinge line similar to that of Ostrea edulis, fig. 25, p. 300, is
lost and superseded by a curved one. X 120 diam., pp. 282, 300, 308, 378.

Fig. 2. Same specimen as fig. 1, viewed from the left side through the glass to which attached; lettering as in fig. 1.

Fig. 3. Ostrea, the same specimen as figs. 1 and 2, but three days older: pi, palps; 7, rectum; g, gills, which at this
stage are composed of filaments connected at the tips by cross connecting bars; m, mantle border, broad and not yet
possessing marginal tentacles; p, the prodissoconch, which equals nearly half the entire height of the shell. Redrawn
from camera drawing by the author. p. 304.

Fig. 4. Ostrea virginiana, young spat growing on glass: ad, adduetor muscle; pl, palps. The body is permeated with
liver follicles; g, gills. Three gill lamellex are in view, A, A’, B’, for an understanding of which see fig. 8. In the early
growth of the right, outer gill B/, no corresponding gill B exists on the left side; s¢, stomach; int, intestine; a, anus;
h, heart; v, large blood vessel passing to mantle border; c,d, foramina through which afferent blood passes to heart;
m, mantle border fringed with tentacles; rm, radial muscles which retract the mantle; s, a sinuosity in the upper right
valve, exists at this point; the mantle is constantly retracted and all excurrent action takes place through this sinuosity
as indicated by arrows; p, prodissoconch. Redrawn from camera drawing by the author. X 12 diam., pp. 305, 307.

Fig. 5. Gill plate from a young oyster of age of fig. 4, showing filaments joined by cross connecting bars. p. 305.

Figs. 6-12. Diagrams of gills of developing Ostrea and adult Perna, Anomia and Pecten. The filaments in all are let-
tered the same so as to show the relation of the several diagrams: A, direct filament of inner gill of left side; a, reflected
border of same filament; A’, direct filament of inner gill of right side; a’, reflected border of same filament; B, direct
filament of outer gill on left side; b, reflected border of same filament; B! direct filament of outer gill of right side; b/,
reflected border of same filament.

Fig. 6. Gills of young Ostrea virginiana, age of figs. 1-2, this plate, pp. 301, 303.

Fig. 7. Gills of Ostrea, later stage, age of fig. 3, this plate, p. 304.

Fig. 8. Gills of Ostrea, still later stage, age of fig. 4, this plate, p. 305.

Fig. 9. Gills of adult Ostrea, see Pl. xxv, fig. 12, p. 306.

Fig. 10. Gills of adult Perna ephippium, see Pl. xxv, fig. 11. pp. 304, 326.

Fig. 11. Gills of adult Anomia glabra, p. 358.

Fig. 12. Gills of adult Pecten irradians, pp. 336, 341.

Figs. 13-16. Ostrea edulis, early phylembryonic stage of shell growth; 13, viewed from the right side; 14, from the
ventral border; 15, from the left side; and 16, in end view. Compare with fig. 25, p. 300, an anatomical figure of about
the same age. The shell has a straight hinge line and only a slight indication of umbos. From the gill chamber of the
parent. X 120 diam., pp. 299, 312, 375.

Fig. 17. Ostrea virginiana, completed prodissoconch stage, viewed from the upper right side, the specimen being al-
ready cemented to an object of support: 7, right valve; J, left valve. Compare with fig. 26, p. 312. X 87 diam., pp. 312, 378.

Fig. 18. A specimen slightly older than fig. 17, viewed from the anterior end: 7, left valve; r, right valve; s, spat
growth, which has begun along the margin of the prodissoconch valves. X 87 diam., pp. 312, 315.

Fig. 19. Young Ostrea virginiana. Tip of left valve from inside whorls of a dead Sycotopus, where it was preserved from

1From Hyatt (27) atter Sundberger.

PHYLOGENY OF THE PELECYPODA. 397

all eroding action: p, prodissoconch; s, spat growth; 1, cartilage pit; f, flange-like extension of margin of the shell ex-
tending over the object of fixation, ), border of shell proper, exclusive of flange. 87 diam., pp. 315, 329.

Fig. 20. Young Ostrea virginiana, growing on glass, viewed from right side, showing p, left and right valves of
prodissoconch, and s, the first stages of spat growth. X 24 diam., p. 315.

Fig. 21. Same specimen as fig. 20, viewed from the left side through the glass to which attached : p, prodissoconch ;
8, Spat growth. p. 315. Figures 17 to 21 are of specimens from Buzzards Bay.

Fig. 22. Gryphea arcuata, umbos of shell showing flat area of early fixation in left valve and corresponding convex
area of right. Specimen from Lower Lias; in Museum of Comparative Zodlogy. XX 11 diam., p. 317.

Fig. 23. Gryphea calceola, var. nebrascensis, showing a flat area of early attachment in this highly arcuate species and
a prodissoconch, p. Jurassic of Wyoming, specimen No. 1881, U.S. National Museum. X 11 diam., pp. 313, 317.

Fig. 24. Same specimen as fig. 23, viewed from in front.

Figs. 3-4 and 22-24 drawn by Mr. J. H. Emerton; figs. 17-18 drawn by Mr. F. W. Cobb; all other figures drawn by
the author.

PLATE XXV.

Fig. 1. Young Ostrea virginiana, growing on glass, viewed from the right side: p, left and right valves of prodisso-
conch; 1s, 2s, limits of first and second spat or nepionic stages.

Fig. 2. Same specimen as fig. 1, viewed from the left side through the glass to which attached: p, left valve of prodis_
soconch; Is, 2s, limits of first and second spat or nepionic stages; 7, margin of left valve, beyond which the right
(upper) valve extends. Figs. 1 and 2, X 24 diam., p. 315.

Fig. 3 Young Ostrea virginiana, viewed from the hinge area: p, prodissoconch; lv, left valve; rv, right valve; s, early
spat growth. The left valve has not yet flattened to the object of attachment. X 120 diam , p. 315. F

Fig. 4. Young Ostrea virginiana viewed from right side, the specimen being attached to a shell: p, prodissoconch,
succeeded by s, the early spat or nepionic growth. X 98 diam., pp. 311, 315

Fig. 5. Young Ostrea virginiana separated froma glass slide on which it had grown and viewed from right upper side :
p, prodissoconch; 1s, 2s, 3s, 4s, four spat or nepionic stages of growth clearly indicated in this specimen; s?, a sinuosity in
the right valve where excurrent action takes place. Compare with left view of same specimen, fig. 6. X 8 diam., pp.
307, 316.

Fig. 6. Same specimen as fig. 5, but viewed from the left side, by which it was attached to the glass before remoyal,
The left valve is seen to overhang the right on its free borders; at s?, the excurrent sinuosity of the right valve is very
noticeable in this figure, and four nepionic stages are marked in this valve as in the right valve. p. 316. Figs. 1-6, inclu-
sive, are of specimens from Buzzards Bay.

Fig. 7. Young Ostrea virginiana from Provincetown, Massachusetts, showing nepionic or spat stages of growth of
the right valve (ls, 2s, 3s, 4s, 5s). The first two stages were drawn from a younger specimen, as they were worn off
in the specimen at hand. X 4 diam. p. 316.

Fig. 8. Ostrea virginiana. Two adult individuals fixed to the same object of attachment. The left hand specimen is
growing in the normal position with the right valve uppermost, and the left valve lowermost. The right hand specimen
is growing in a reversed position with the right valve lowermost, and the left valve, uppermost; in both the same rela-
tive form of the valves is maintained. Buzzards Bay. X } diam., pp. 316, 321.

Figs. 9-12. Revolution of the axes, showing the relations of the axes of the body to the shell in typical Pelecypods,
and the serial mechanical relation of the same. Lettering in the figures: ap. ax, antero-posterior axis, drawn through the
mouth and middle of the posterior adductor muscle; h. az, hinge axis, the axis of motion of the valves; a. ad, anterior
adductor muscle; p. ad, posterior adductor; m, mouth; pl, palps; a, anus; g, gills; pd, pedal muscles; f, foot; 5, byssus;
h, heart.

Fig. 9. Mya arenaria, p. 309.

Fig. 10. Modiola plicatula, p. 309.

Fig. 11. Perna ephippium, pp. 309, 329.

Fig. 12. Ostrea virginiana, pp. 305, 308, 327.

Fig. 8, drawn by the kindness of Mrs. J. W. Elliot; all other figures drawn by the author.

PLATE XXVI.

Fig. 1. Spondylus hippuritarum, having a highly modified, deep, attached valve and less modified free valve. European
Cretaceous (after @Orbigny). pp. 319, 323, 351.

Fig. 2. Chamostrea albida, showing an ostreaform, cemented valve and flatter free valve (after H. & A, Adams). p. 323.

Fig. 3. Hinnites cortesi, right valve, showing an early nealogic, regular pecteniform stage, possessing a byssal sinus,
which is succeeded, when the shell becomes attached to a foreign body, by an irregular ostreaform growth, possessing no
byssal sinus, and losing almost entirely the features characteristic of early stages. pp. 282, 293, 294, 324, 350 and 390.

Fig. 4. Hinnites cortesi, left valve, showing an early regular pecteniform shell, as does the right valve, fig. 3. When the

MEMOIRS BOSTON SOC. NAT. HIST., VOL IV. 53

398 ROBERT TRACY JACKSON ON THE

shell becomes attached, the left valve, acting in sympathy with the right, becomes somewhat irregular in growth; but is
less modified than the right by the new condition of cemented fixation. Life size, p. 324. Both specimens from the Eu-
ropean Tertiary, in the Museum of Comparative Zodlogy. A

Fig. 5. Mulleria lobata, showing adult oyster-like form and at the tip of umbos a regular Anodon-shaped shell preced-
ing the attached stage (after H. & A. Adams). pp. 310, 324.

Figs. 6-7. Mulleria lobata. Tips of the valves magnified to show the early stage, viewed from the dorsal and left side
(after Bronn). pp. 310, 324.

Fig. 8. Caprinula boissyi, having a deep, conical, rugose, attached valve and an arcuate, plain, free valve. European
Cretaceous. pp. 282, 319, 323. ,

Fig. 9. Monopleura marcida, showing a highly modified attached valye and a normal free valve. American Cretaceous
(after White (74), slightly modified). pp. 319, 323. -

Fig. 10. Caprotina semistriata, showing a deep, rugose attached valve and shallow, plain, free valve. European Creta-
ceous. Compare with fig. 20, a young specimen of this species. pp. 319, 323.

Figs. 11-12. Radiolites fleuriausa, having a highly modified, cornucopia-shaped, attached valye, fig. 12, and a less modi-
tied flattish free valve, fig. 11. European Cretaceous, pp. 319, 324 (figs. 8 and 10-12, after d’Orbigny).

Fig. 13. Ostrea virginiana, section showing the deep, concave, cemented valve, and flat free valve: /, lacune in the
sub-nacreous structure are filled with chalky lime; ad, adductor muscle, which makes a purple sear at its area of insertion
in the valves. ‘The black line indicates the position of the adductor at different periods of growth, and is very noticeable
in the section, being marked by a continuous purple line. An instructive section to show action of adductor muscle, liga-
ment, growth of shell and change of base of the muscle in successive growth. Buzzards Bay. Life size, p. 320.

Figs. 14-15. Davidsonella rugosa, sp. VOrb., showing rugose, lower, attached valve (fig. 14) and a flat free valve (fig.
15), indicating that cementation in Brachiopods produces a similar result to that in Pelecypods. European Cretaceous
(after VOrbigny). p. 320.

Fig. 16. Perna ephippium, adult, pp. 327, 329, 383.

Fig. 17. Perna isogonum, p. 327.

Fig. 18. Perna vespertilio, p. 327.

Fig. 19. Malleus albus, adult. pp. 327, 383 (figs. 16-19 after Reeve, reduced).

Fig. 20. Young Caprotina semistriata, from the Cretaceous. The lines of growth are introduced according to judgment
to render the parts clear, as the specimen is too much eroded to show them. Compare with the adult, fig. 10, this plate.
Specimen No. 2902 in collections of the Museum of Comparative Zodlogy. X 24 diam. pp. 319, 385.

Fig. 20 drawn by the author; all others drawn by Mr. J. H. Emerton.

PLATE XXVIL.

Fig. 1. Young Perna ephippium, viewed from left side, showing, p, prodissoconch and succeeding dissoconch growth.
Compare with Pecten, Pl. xxvut, fig. 9 and Avicula, fig. 52, p. 330. The form of the shell is strikingly different from the
adult, Pl. xxv1, fig. 16. Specimen from Museum of Comparative Zodlogy. X 50 diam. (another view of this specimen,
see Pl. xxvin, fig. 11). pp. 328, 379, 383.

Fig. 2. Young Perna ephippium, viewed from the right side, showing, p, prodissoconch and succeeding dissoconch
growth. The byssal notch originates at the close of the prodissoconch period in the initial stages of dissoconch growth.
The figure shows the margin of left valve extending beyond the right in the byssal area and several threads of the byssus.
Specimen from Boston Society of Natural History Collections. Cat. No. 2989. 24 diam., p. 328.

Fig. 3. Perna ephippiwm, same specimen and view as fig. 2, showing the prodissoconch, p, and the early stages of
growth of the right dissoconch valve. It is seen in figs. 2 and 3, that as in Pecten the left valve in early stages is larger
than the right, and overhangs it at the umbos. X 87 diam., p. 828.

Fig. 4. Young Spondylus asperrimus, tip of right attached valve: p, prodissoconch; the succeeding early dissoconch
growth is pecteniform with a deep byssal notch; as soon as attached the growth becomes irregular, the pecteniform shape
and byssal notch being eradicated. Compare with Hinnites, Pl. xxvii, flg. 3. Specimen in Boston Society of Natural His-
tory, Cat. No. 8494. X50 diam., pp. 351, 390.

Fig. 5. Young Plicatula marginata, viewed from the left side, showing a prodissoconch, p, and succeeding dissoconch
growth. Specimen in Museum of Comparative Zodlogy. % 50 diam., pp. 325, 352, 392.

Fig. 6. Young Lima elliptica, showing, p, prodissoconch, succeeded by the dissoconch growth. Specimen in Smith-
sonian Institution. From drawing made when at that Museum. Magnified, pp. 353, 388.

Fig. 7. Oxytoma tnequivalvis, Sow., umbonal region of right valve showing p, prodissoconch, succeeded by a dissoconch
which has a byssal notch in its initial stages of growth, anda triangular cartilage pit. Compare with Pecten, several figs.,
Pl. xxv, and Avicula, fig. 33, p. 330. Specimen from the Middle Lias in the Museum of Comparative Zodlogy, Cat. No.
9222. X50 diam., pp. 332, 382.

Fig. 8. Chlamys islandica, hinge area of same specimen as Pl. xxvum, fig. 14, viewed from the right side, showing, p,
prodissoconch, and the byssal notch originating in the initial stages of dissoconch growth. Compare with Pecten trradians
Pl. xxvint, figs. 10 and 13. The left valve is larger than the right. X 87 diam., p. 344.

Fig. 9. Pecten irradians. Umbonal view from left side of a similar specimen to that in Pl. xxvin, fig. 10, more enlarged,
showing p, prodissoconch, and the byssal notch originating at the close of the prodissoconch stage. X 87 diam., p. 342.

Figs. 4 and 6 drawn by the author; all others drawn by Mr. J. H. Emerton.

PHYLOGENY OF THE PELECYPODA. 399

PLATE XXVIII.

Fig. 1. Young Pecten irradians, viewed from the left side: ad, adductor muscle; g, gills; 7, intestine; a, anus; h,
heart; f, foot; p7, posterior retractor of the foot; m, mantle border bearing eyes, e; and alternating tentacles, ¢. The
shell bears a prodissoconch, p, and is of about the age of fig. 12, this plate. Redrawn from camera drawing by the author.
X 40 diam., p. 333.

Fig. 2. Pecten irradians, older than any other young Pecten figured. Shows an active extensible foot; the mantle bor-
der bears a single eye in each plication of the shell and alternating with each eye two tentacles (tentacles are omitted in the
right mantle lobe for the sake of clearness). Under the dorsal ear a special protrusion of the mantle border (pseudosi-
phon) exists, and at this point all excurrent flow takes place as indicated by the arrow. The shell shows no prodissoconch,
but the smooth early nepionic stage is indicated. The plications are as in the adult, fig. 87, p. 333, in number and form.
The hinge line is relatively long, and but slight ears are developed. X 16 diam., pp. 335, 345, 387.

Fig. 3. Pecten irradians, diagram of a gill filament of adult: @, direct border; a’, reflected border; s, septum connect-
ing direct and reflected border of about every fifteenth filament; b7, blade commencing when septum diminishes. Compare
with Pl. xxiv, fig. 12. p. 341.

Fig. 4. Pecten, diagrammatic section of adult, to show relations of mantle wall and tentacles: s, shell; lig, ligament;
ad, adductor; m, mantle; mw, mantle wall; e, eye; t, marginal tentacles; y¢, guard tentacles. p. 338.

Fig. 5. Feces of Pecten, seen in section and longitudinally. p. 340.

Fig. 6. Pecten, mantle border of fig. 1 enlarged, showing cup-shaped eyes alternating with single papillose tentacles ;
no mantle wall or guard tentacles are yet developed on the mantle border. pp. 335, 337.

Fig. 7. Young Pecten, diagram of foot. In forming the byssal attachment it may lay hold of a foreign object at its tip
and at the point of the byssal cleft, arching up between those points as in the figure. p. 337.

Figs. 8, 8a. Pecten, diagram to illustrate the swimming. Arrows indicate the several currents of water ejected in swim-
ming as described in the text. p. 339.

Fig. 9. Young Pecten irradians, viewed from left side: p, prodissoconch, succeeded by the early nepionic stage of the
dissoconch. The form of the shell is distinct and differs from all later stages. Compare with Perna, Pl. xxvut, fig. 1, and
Avicula, fig. 32, p. 330. X50 diam., pp. 333, 342, 379, 387.

Fig. 10. Pecten, same specimen as fig. 9, but viewed from the right side. The byssal notch originates with the initial
stages of dissoconch growth, it is deep, and has a marginal tooth. The left valve overhangs the right on its free border.
Compare with Pl. xxvu, fig. 3, and Pl. xxrx, fig. 6. pp. 328, 342.

Fig. 11. Perna ephippium, same specimen as Pl. xxvut, fig. 1, viewed from umbonal region: 7 v, left valve; which is
slightly larger than rv, the right valve. The wnbos of the prodissoconch, p, are directed posteriorly. X 50 diam., p. 328.

Fig. 12. Pecten irradians, older than figs. 9-10, viewed from the left side: p, prodissoconch, succeeded by smooth
early nepionic growth which gives place to a plicated later growth. The hinge line is proportionately longer than in earlier
stages, figs. 9-10, and is also relatively longer than in the adult, fig. 37, p. 333. Very slight indications of ears exist at this
stage. X40 diam., pp. 345, 387.

Fig. 13. Pecten, same specimen as fig. 12, viewed from the right side. Description of shell as in fig. 12. A deep bys-
sal notch exists in this valve, originating at the close of the prodissoconch period. Four teeth exist on the border of the
notch instead of one as in fig. 10. At this age the left valve has ceased to be produced beyond the border of the right on
on its free margins, as in fig. 10. The shell as viewed from both sides differs markedly from earlier and later stages in form.
p. 345.

All the specimens of Pecten irradians tigured were collected at Buzzards Bay, August, 1888.

Fig. 14. Young Chlamys islandica, viewed from left side, showing, p, prodissoconch, and a smooth early nepionic
growth, succeeded by a plicated later growth. The plications are like those of the adult and mark a stage by their intro-
duction. The hinge line is longer relatively to the width of the shell at this stage than in the adult as noted in Pecten ir-
radians, and has slightly developed ears. Specimen No. 27294, Smithsonian Institution Catalogue. 23 diam., pp. 349 and
388. An enlarged umbonal view of this specimen seen from the right side, vide Pl. xxvmy, fig. 8.

Figs. 3-8 drawn by the author; all others drawn by Mr. J. H. Emerton.

PLATE XXIX.

Fig. 1. Young Anomia glabra, growing on glass, viewed from the left side: ad, adductor muscle; p 7, posterior re-
tractor and @ 7, anterior retractor muscles of the foot; 7, foot; pl, palps; a, anus; b, body permeated with liver fol-
licles; g, gills. Two gill lamelle are seen from this side, A and 6 (compare with diagram, Pl. xxtv, fig. 11); s, sheath
supporting gills; h, heart; v, v, blood vessels; m, mantle border; p, prodissoconch. p. 355.

Fig. 2. Anomia, same specimen as fig. 1, viewed from the right side through the glass to which attached: ad, adduc-
tor muscle; @7, anterior retractor muscle of the foot, attached to the left upper valve, but seen through the glass; bs,
calcifying byssal tip composed of separate discs of lime which unite to form parallel lines, as shown in the figure; f, foot;
pl, palps, showing slightly in this view; y, gills of which three lamellie, a, A’, b', are visible from this side (compare with
Pl. xxtv, fig. 11); s, sheath which supports the gills; h, heart, a ventricle and auricle are visible from this side; @, anus;
c, cecum, highly produced; m, mantle border; rv, border of right valve which by a sinuosity encloses the byssus; lv
border of left valve. Figs. 1 and 2 magnified about 20 diam. Redrawn from camera drawings by the author. p. 355.

Fig. 3. Diagrammatic cross-section of fig. 1, to show action of the muscles; lettering as in fig. 1. p. 355.

,

400 JACKSON ON THE PELECYPODA.

Fig. 4. Anomia glabra, very young, viewed from the left side: p, prodissoconch, which is on the border of the valve,
not yet being enclosed by the dissoconch growth. X 50 diam., p. 359).

Fig. 5. Young Anomia glabra, viewed from the left side: p, the -prodissoconch, is not yet surrounded by dissoconch
growth. Compare with figs. 4and 8. The dissoconch is sub-oyal and an extensible foot, f, exists. X50 diam., pp. 355, 389.

Fig. 6. Anomia, same specimen as fig. 5, viewed from right side; p, the prodissoconch, shows indications of a byssal
sinus in the lines of growth. The dissoconch is extremely thin, surrounds the byssal notch and anteriorly has grown
around the prodissoconch covering its umbo. The method of growth is indicated by the lines of growth. pp. 359, 388.
Compare with Pecten, Pl. xxvim, fig. 10.

Fig. 7. Young Anomia glabra, wnbonal region of right valve: p, the prodissoconch, shows indications of a byssal
notch which is accentuated in the succeeding dissoconch growth; the latter finally grows around the byssus so as to en-
close it. Compare with fig. 6, this plate, also with Pl. xxvn, fig. 3, and Pl. xxvumt, fig. 10. X 50 diam., p. 360.

Fig. 8. Anomia, tip of valve, older than figs. 4 and 5 and in the same view. The dissoconch growth has now enclosed
the prodissoconch, p, and the borders of the opposing sides form a marked area by their ingrowth at the umbo. X 50
diam., pp. 359, 389. Specimens of Anomiz, figs. 1-8, were collected at Buzzards Bay, August, 1888.

Fig. 9. Mytilus edulis, very young: p, the prodissoconch, succeeded by early nepionic growth. Beach Bluff, Mass.
X 50 diam., p. 363. 3

Fig. 10. Mytilus edulis, older than fig. 9; p, prodissoconch, succeeded by nepionic growth. Nantucket. X 47 diam.

Fig. 11. Young Modiola hamatus: p, prodissoconch, succeeded by a smooth nepionic stage. Later, the smooth ne-
pionic stage is succeeded by the plicated growth typical of the species as shown in Pl. xxx, fig. 1. Specimen on oyster
shells from Baltimore, collected at Martha’s Vineyard. X 50 diam., p. 364.

Figs. 3, 5. 6, 9, 10 and 11 drawn by the author; all others by Mr. J. H. Emerton.

PLATE XXX.

Fig. lL. Modiola hamatus, older than Pl. xxrx, fig. 11: p, prodissoconch, succeeded by an early plain nepionic stage,
which is again succeeded by a plicated later stage. Attached to a cluster of Pernas in Museum of Comparative Zoology.
X 50 diam., p. 364.

Fig. 2. Argina pexuta, very young, showing a prodissoconch, p, and succeeding dissoconch growth. Buzzards Bay, Sep-
tember, 1888. X 87 diam., p. 364.

Fig. 3. Orenella decussata: p, prodissoconch, succeeded by the early dissoconch growth. Specimen from Smithsonian
Institution. X 23 diam., p. 364.

Fig. 4. Spherium securis: embryo from gills of the parent. The whole shell of this specimen represents the prodisso-
conch period: a. ad, anterior adductor; p. ad, posterior adductor; g, gill filaments; p7, palps; f, foot; of, otocyst; st, si-
phon. Medford, Mass., Noy. 1888. X about 60 diam., p. 369.

Fig. 5. Spherium securis, a little older than fig. 4, viewed from left side : p, prodissoconch, succeeded by early dissoconch
growth. X about 25 diam., p. 369.

Fig. 5a. Spheriumysecuris, a little older than figure 5, viewed from right side: p, prodissoconch. X8 diam., p. 370.

Fig. 6. Young Petricolapholadiformis, showing p, prodissoconch and succeeding nepionic stage which is strikingly differ-
ent from later stages: f, foot; m, mantle border; s?, siphons, both with simple, unbranched tentacles. X 47 diam., p. 370.

Fig. 7. Petricola pholadiformis, older than fig. 6, shows early nepionic smooth stage succeeded by the later characteristic
plicated and elongatejfeatures typical of the adult. X 8 diam., p“371. Figs. 6-7, of specimens from Nantucket, July, 1889.

Fig. 8. Petricola, tip of branchial siphon of a specimen of about the age of fig. 7, showing the arborescently branching
tentacles which surround the orifice. X19 diam., p. 371.

Fig. 9. Young Venus mercenaria : p, prodissoconch, succeeded by early dissoconch growth. X 87 diam., p. 371.

Fie. 10. Venus mercenaria, an older specimen than fig. 9, showing a prodissoconch p, and succeeding nepionic growth.
>< 40 diam., p. 871. Figs. 9-10 from Buzzards Bay. ;

Fig. 11. Scrobicularia tenuis: p, prodissoconch succeeded by early dissoconch growth. The prodissoconch is not so
sharply defined as in many genera, because the early dissoconch is not characterized by markedly differing features. Speci-
men from Smithsonian Institution. 50 diam., p. 372.

Fig. 12. Mya arenaria, very young: p, prodissoconch, and well defined succeeding dissoconch. Buzzards Bay, 1888.
X 87 diam., p. 374.

Fig. 13. Mya arenaria, older than fig. 12: a prodissoconch, p, less clearly defined than when younger, exists at the
umbos. The dissoconch growth has acquired characteristics of the genus in form. <A large active foot exists. The
siphon consists of an extensible excurrent and a short non-extensible incurrent tube. Buzzards Bay. X 20 diam., p. 372.

Fig. 14. Siphon of a young clam older than fig. 13. The massive siphon is fringed with tentacles. The incurrent
tube is also bordered by tentacles, but the excurrent tube bears none; it is thin walled and retractile: s, border of shell
from which the siphon extrudes. Chelsea Beach. X10 diam., p. 373.

Figs. 4 to 8 inclusive, and fig. 14 drawn by the author; all others drawn by Mr. J. H. Emerton.

E XXIII.

PLATE

Memoirs Hoston Soc, Nat. Hist,, Val. IV.

He vals ;
Laat see

:

oO

ig ©) ve
K ee, 2) Ht

Se

JACKSON, PHYLE@GENY OF THE PELECYFPODA.

a

Memoirs Baston Soc, Nat, Hist,, Vol, IW.

ane
(uulu(

\\)
oe

iS

i eS

LUX CO., ENGRAVERS & PRINTERS, BOSTON

JACKSON, PHYLOGENY OF TH!

BE PELECYPODA,

PLATE XXIV.

a ———

Memoirs Boston Sac, Nat, Hist, Vol. IW.

LUX CO., ENGRAVERS & PRINTERS, BOSTON

JACKSON, PHYLOGENY OF THE PELECYFPODA,

PLATE XXVY.

Memoirs Hoston Sac, Nat. Hist,, Vol. IW, PLATE XXXVI,

x ich’
8 SP tye.

Se

AN
Pe
‘ Be im

i

LUX CO., ENGRAVERS & PRINTERS, BOSTON.

JACKSON, PHYLOGENY OF THE PELECYFPODA,

PELEBECYFODA,

GENY OF THE

ON, PHYLO

JACKS

el

Memoirs Boston Soc, Nat. Hist., Vol. IY. PLATE XXVIII,

Bays
iti

; -

\ WW
\
1

)

A fy
i)

¢

Hy

HY} 5)

M) )
vf

\ CA MY Sans
= Mi 3S
, ol
CON
\

N
Be

LUX CO., ENGRAVERS & PRINTERS, BOSTON.

JACKSON, PHYLOGENY OF THE PELECYPOODA.

Memoirs Boston Soc. Nat Hist., Vol. IW. PLATE XXIxX.

Qa :

))
Ab

2 lille
= Jay

ENN

@
A’
A Pai =
/
4
Y4,
WA

LUX CO. ENGRAVERS & PRINTERS, ROSTON.

JACKSON, PHYLOGENY OF THE PELECYFPODA,

Memoirs Boston Sac, Nat Hist., Wol. IW. PLATE Xxx,

JALKSON, PHYLOGENY OF THE PELEGY POOR,

Sales will be sold anya sets. Pilate "bers; $ othe public
Vols. XII-XXIV per volume, cloth, $5. 50 to- ‘members, ‘A féw complete : se s will be ae at spe

iy $4.50 to the public. q
J OURNAL. ‘BVO. Series now ended.
Vol. I, part 2 $2.00; part 3, $1. 50. Vol. V, part 1, $1.75. es
* TI, $6.50; Vol. II, part 1, $2.00; part 2, $2.00; parts “VI, $6.00; Vol. 6, part. , 81. 75; part 2
3-4, $3.00. j $1.75; part 4, $1. te
© TV, part 2, $1.50; part 3, $1.50; part 4, $1.50. “ VII, part 2, $1.75; part 3, $1.50; ‘pare 4, a 25

MEMOIRS. 4to. Being anew Series of the Journal.
Vols. 1, I. and III, each with many plates, $14.00 to mem- | Subscription per quarter volume,-averaging 130 ream
bers, $16.00 to the public. to members, $4.00 to the nde Ki
Vol. IV is now being issued in separate articles as printed. ready. ; os
SEPARATE ARTICLES, EXTRACTED FROM THE MNMOIRS, VOLS. I TO IV.
Volcanic Manifestations in New England. By Wm. T. | The Early Types of Insects. By 8. H.. Seudde

Brigham. pp. 28. 50 cents. 35 cents. ~ re
Early stages of Terebratulina, By Edwards. Morse. .Two | Paleozoic Cockroaches. By 8. H. Scntace
Steel Plates. pp. 12. $1.50. Plates. $8.00., iy
Osteology and Myology of Didelphys virginiana. By Elliot | New Hydroids from Chesapeake Bay. By a: F
Coues. With an appendix on the Brain. By Jef- 8. Three Plates. 75 cents. ;
fries Wyman. pp. 114. $2.50. Carboniferous Myriapods.
Development of Limulus Polyphemus. By A. 8. Packard, Plates. $2.00. .
Jr. Three Steel Plates. pp. 48. $2.00. Embryology of Teleosts. By J. S. iinestey and Hf Bf Conn.
Description of Balenoptera musculus. By Thomas Dwight, pp. 80. Three Plates. $1.75. ‘
Jr. Two Plates. pp. 28. $1.50. Carboniferous Hexapod Insects of Great Britain. By S Hees
Fossil Myriapods from Nova Scotia. By S. H. Seudder. Scudder. pp. 22. One Plate. 70 cents.
pp. 9. 35 cents. Embryo'ogy of (canthus and Teleas. By Howard ayo oe
Note additionelle aun Mémoire intitulé “Volcanic Manifesta- pp. 56. Eight Plates (two colored) $2.50. pee
tions in New England.” Par M. Albert Lancaster. | New Carboniferous “Myriapods, Species of Mylacris. By
pp. 9. 35 cents. S. H. Sendder. pp. 25. Two Plates. $1.25.
Embryology of Terebratulina. By E. S. Morse. Two Steel | Notes on the Peeping Frog. By M. H. Hinckley. Pp.
Plates. pp. 16. $1.50. : One colored Plate. 75 cents. asim
List of the Birds of Western Mexico, Oy George N. Law- | Affinities of Paleozoic Hexapods. Winged Tusebes” from a “6
rence; with Notes on their Habits, etc., by A. J paleontological point of view. By S. ie Seud Ti 4
Grayson. pp. 50. $1.25. 40 pp. Four Plates. $2.00.
Recent changes of Level on the Coast of Maine. By N.S. | Life History of the Hydro-meduse. By W. K. Brook
: Shaler. pp. 20. 40 cents. pp. 72. Hight plates. $3.00. >
The Species of Pamphila. By Samuel H. Scudder. pp. 13. | The oldest known Insect-Larva.
Two Plates (one colored). $2.00. By S. IL. Seudder. pp. 55. Four Plates.
Antiquity of Caverns and Cavern Life of the Ohio Valley. | The Significance of Bone Structure. By Thomas Dwi;
By N.S. Shaler. pp. 9. One Plate. 40 cents. pp. 15. Three Plates. $1.25. _ es
Monograph of the Tabanida, Pt. 1. By C. R. Osten Sacken. | Development of the Ostrich Fern. By Douglas Hou on
pp. 83. 65 cents. Pt. pp. 59. $1.00. Campbell. pp. 36. Four Plates. $1.50.
Revision of North American Porifere, Pt. 1. By A. Hyatt, | Introduction and Spread of Pieris rapae in North Amer
pp. 10. Qne Plate. 75 cents. “Pt. 1. pp. 74: 1860-1885. By S. H. Seudder. pp. ‘16. ae Ma
Three plates. $2.00. 50 cents.
Gynandromorphism in Lepidoptera. By A.S. Packard, Jr., | North American Geraniacer. By W. Trelease. pp. 4
Structure and Transformations of Eumeeus Atala. Four Plates. $1.25. is
By 8S. H. Scudder. pp.11. One Plate. 75 cts. The Taconic of Georgia. By Jules Marcou. pp. 27. ”
Appendix, Index and Title-page to Vol. II. pp.17. 85cts. | The Entomophthoree of the United States.
\Distomum crassicolle: with Notes on the Classification of Thaxter. pp. 68. Hight Plates.
Worms. By C. S. Minot. pp. 20. One Plate. 75 cts. | The Flora of the Kurile Islands. By K, Miyabe.
One plate.

OCCASIONAL PAPERS. i cae
Tn ENTOMOLOGICAL CORRESPONDENCK OF Dr. T. W. Harris. Edited by 8. H. Sendder. 8vo. Portrait an
4 Plates. 1869. Cloth, $4.50. “See
Tue Srrprrs or THe Unitep Srares. By N. M. Hnnrz. Edited by Edward Burgess, with Notes by James H.
> Emerton. 8vo. 21 Plates. 1875. Cloth, $3.50; Paper, $3.00. i ena
© GroLoGy or Massacnusnrrs. By W. 0. Crospy. 8vo, 266 pp. Colored Map and 5 Plates. Cloth, $3.50;
aper, $3.00. e =

MEMORIAL VOLUME. Les ¥
y ANNIVERSARY MEMOIRS, containing a history of the Society and a series of memoirs, 4to, 635 page
traits, and 31 plates. Price, $15. The HISTORY of the Society is also poe ee separately. with per:
ete. Price, $5700. = mia

Address J. WALTER a eee Secretary. Gi aes

+i a.) 7

a

= Ly, = a WN 5 7 = iy, 2
> : Yd 2 > aye x 2 G>
= Pe - = a eo Was
on’ Pas 2) m S\N 2) = ice =
z AN z aa Ss z a z ih /
= o é =
NLILSNI NVINOSHLIWS S3IYVYUSIT LIBRARIES SMITHSONIAN INSTITOTION: Nother ts 0) MIN lange arcane ais
Ty,* - 5 : 2 \g : ue ;
tig a 7) a n nA ~~ Ws wo iH oO
Yl. = fe) Be fe) = Mw Io ae fe)
iMG = Z. Si. < E N\S Z, E S
a 3 a es z = os 5 a Zz |

RARIES SMITHSONIAN INSTITUTION NOILALILSNI_NVINOSHLINS S3INVYGIT LIBRARIES SMITHSONIAN INSTITUTION

= S 7) = o as

tw Zz ul a uw a w a

= = = 4 = 4

oc 25 oc = = = = mo

< Al < C WN Ee S = e

Be 7 : 2 N's = a. =
= 9° = Zz on af 2 a .

= a ej Fe
LALILSNI NVINOSHLINS S3IYVUGI7_ LIBRARI ES_ SMITHSONIAN INSTITUTION NOLLNLILSNI_NVINOSHLIWS S318V¥ aI
A oe: z = ro) i= ro) = ro)

o eas & o E a = Oo =

> aYY. 5 a > re =) ee =)

> Wg," 2 > = > = > E

Pe] = a Ee 2 = 2 =

= D Ps 2 m 2 m 2

x 4 = ; wn = top) 5

wn s 2 7) =

RARIES SMITHSONIAN INSTITUTION NOMALLESNIL NVINOSHINNS sailed a bi CUBR ARTES eon ia

= ee < < = a Vo =
Se Z2SN 3 Vs 2 BQN 3
BG S Na SD. 4 Lf m2) QD, 72) VW “ e 2) A
MQ. 2 END 2 Uy * 2 Ee ™\ 2 =
= = Mee 5 z a We a a
LALILSNI NVINOSHLIWS Sa 1YWwy g ne LIBRARI Fae a eee ee ee ee ee ee INES Lae 1yYVuyal

= va <p) > ee sa te

a NSS a ” uw yA 2S 2 2 2p) Ke a

= YY NO a = e 7, bb = # = WN Ss 2s

3S WSN OC a <YY Y Vl = < = SKS <

S VAN = e a Gry S cc Ss WN ac

= YQ o = ao “fff 5 {oa} S : oO
fe) — 2) Fy % ai Sa |
. 3 z =i 2 4
-RARI ES’ SMITHSONIAN” INSTITUTION — NOILALILSNI NVINOSHLIWS >33 luVvud he LIBRARI ES_ SMITHSONIAN ee
Zz Zz a _

oO ty, == {e) wo S) = 2 y w

- Yay = = 2 E s t Yip >

3) the Gi. = > = > = Y fU kh >

& YF 4 = 2 = 5 = Oki =

= tus yee = = = oe ey fee ar

Tihie Cie Es = 2) m = m = m

= 7) = eh ms no = o

LALILSNI NVINOSHLIWS Saluvyaia LIBRARIES SMITHSONIAN INSTITUTION NOILMLILSNI_ NVINOSHLINS S31Yvual
\

IWS
a.

‘ACS
AS

Bes

SMITHSONIAN
NVINOSHLIWS
SMITHSONIAN
* . Zp
NVINOSHLIWS
SMITHSONIAN
NVINOSHLIWS

NVINOSHLIWS

WX

\e SS
SS

SMITHSONIAN

RARIES SMITHSONIAN INSTITUTION NOILALILSNI_NVINOSHLINS S31YVYAIT LIBRARIES SMITHSONIAN INSTITUTION

no = no Zz a S on z
Ww Dy, WwW 7) w 5 WwW ra
= OG," : = z = e ae
< AD fae < = < ~ e A
Cae. GY fe = a = ES = a =|
mo’ = roa) ry = = : =
5 2 5 2 a 2 5 2 :
ANLILSNI_NVINOSHLIWS | $3 luvugd ee BRARIES > SMITHSONIAN INSTITUTION Ba Ure Le aS ee lyVvud
= x 5 ~ re) = fs) = Oo
o w = @ = w =
z = 2 5 “2 5 x 5
> rad > E74 Z = = a =
= = - Gy = =
ae E = UE yy - eS
rm . is 3 “Gf » Z B Z
ie ae 2 z z _ 2
SRARI Be oN SOE. NOLES NI Nea TlNs ies lyvud mee BRARI ES SMITHSONIAN
wet we ene = =
2c 2 2 @. £ a 2 Sas s
Xs 5 ESN 5 = 3 SAN 2 3
RK Sw ® Ws a a By a We a a
KO 2 EM 2 c 2 E WS 2 c
E z 2 =) eee = ame ts
- a 2 4 wo . .
ILNLILSNI NVINOSHLINS S3INVYSIT LIBRARIES SMITHSONIAN INSTITUTION NOILOLILSNI_NVINOSHLIWS | Saluvid
a < es
Z Se » a w & iz a =
ZW = = = 2 “i 2
ERR = : < : : z =
ENN = = cc a
ie 2 e : © :
2 ey Zz Py oa = Fa ee)

mae wTt ie rk Ait A ml taieworIIrT

se NN
SN
“ bh.

RIES SMITHSONIAN |

saiuvad
INSTITUTI

Sialuvud

NSTITUTION NOILNILILSNI

INSTITUTI
~

NY
SaluVUE

Ss YAS

NVINOSHLINS S3IYVYSIT LIBRARIES SMITHSONIAN INSTITUTION

INSTITUTI
SAIYVUE
INSTITUT!

r?
NYINOSHIINGS SaAIlINWN Ga

w ras 2)
g 2s g z a ae = <
= 7 = z ® Ho
3 2 Wy 2 . 8 z 5 We: 3 N
a 8 2 WW B “2 g Z Sh 3 2
Z e wz. E 2. Ee NY2 E AS
ae ee ie 5 z ie ae =
JLSNI_NVINOSHLINS S3IYVUGIT LIBRARIES SMITHSONIAN INSTITUTION NOILALILSNI_NVINOSHLINS Saluvy aly
| ie = 2 if é.

a ul # us 7% u! o & “ 3

4 = 4 =f =

_ ao — a eS [oa = ce =

= < as < PS < <a VSS < =

5 : = e s e 5 AA's :

: 5 g = S = SOs :

a!
RIES SMITHSONIAN INSTITUTION NOILNLILSNI NVINOSHLINS S3IYVYGIT_LIBRARIES SMITHSONIAN INSTITUTION

Zz = z Es be z & z = z

2 ps = o se 2 @ = o c

= re > aD & x = a 5) a 2

Ea me Ee a = es Bs a r

Ee cy = ou = 2 - ee F

2 @ i Z m 2 m Z rm :

— _ (op) = —- =

LILSNI NVINOSHLINS S3IUVYGIT LIBRARIES SMITHSONIAN INSTITUTION NOILNLILSNI NVINOSHLINS S3IYVUEIT

z g SG 2 Be a ee z g

Z 3 2S 2S 5 e = *

fe) a57) D Ke K = fo) WS : BG ro) ccam C

2 Bt 2 Me 2 YR 8 z By

E 2 = Se = SS Zz = 2 ce

= > = > S > = > P

” = ” ~ rae 7) =i Fe 79) rs G

\RIES SMITHSONIAN _ INSTITUTION NOILMLILSNI_NVINOSHLIWS S3IYVYGIT_ LIBRARIES SMITHSONIAN _ INSTITUTION

Ww 2 w Z a. is = a z

e 2g & 2 Ep a ga |

x. = oc a WN a x. 79 Oe = c

<>, a < a x < a < 4, Lf = :

oc = a = Qh a = in hg C4 = c

a We 3& S aL No ue 3 a W@ 3

st ¥ — = -

LILSNI_NVINOSHLINS S3IYVYSIT LIBRARIES SMITHSONIAN INSTITUTION NOILNLILSNI_NVINOSHLINS S3IYVYEIT

i PS z is z bss z Es Z r

= fe) = re) = re) = ° ;

Q = wo = ow = oO = C

» = 5 = al 5 2 5

> E > = > — > — 2

2 - 2 me, pie = = - :

. Z m 2° m 2 m SO 8
= 2) = on = — c

RIES SMITHSONIAN INSTITUTION NOILNLILSNI RWINOSHEWS (Sala eaiiq eC UE RANTES SMITNSONIAN INSTT OU ON
= “* wn a ; x za ¥
= =s)> = < = < = 2%
= > x =I 2 =| z " % = > aN
YAY So a io} IE ro) “(URES OO rT ANS
z E YQ = 7 = N&’ 2 E XS :

— — zi = — 2 _ NS _ = \ S$ :
HIISNI_NVINOSHLINS S31YWVYaIT_ LIBRARIES SMITHSONIAN INSTITUTION NOILNLILSNI_NVINOSHLINS S31YVYa17.

& u! = by, & us a uw

c _ < 2 <x = < Cc Ww ANS <= ;

su a = land = ao o sy SS a -

fo) a = a. = oO = SSXsyr” :

g = 2 5 2 = ee <= idler

= :

ARIES _ SMITHSONIAN INSTITUTION NOILLILSNI NVINOSHLIWS S3IYVYGIT_ LIBRARIES SMITHSONIAN INSTITUTION
es = ;

| fe) = ° = fs) S < ow
re a2Q = o a wo = 2@ :

5. 2 = x i 2 5 2 !

EY = [— > E > Ee : >

= es ras 2 = a — 2 i

n” m w = w” = nn = ‘

| z j on z a Fe n Z ra)
IILSNI_NVINOSHLINS S3INVYAIT_LIBRARIES SMITHSONIAN INSTITUTION NOILNLILSNI_ NVINOSHLINS, S31UVU aI)
yy = Sec te Ze at = < = *
mS = eh = > S = ‘S = “ y Sead
( ; = = ; = = z \) SS = Va ¢ A = x en ty

2 zB 3 z 3 5 PMf 3 bY

: 2 = 2 : $i * 2 Gy

a 2 a . 3 3 z 7 2 |

“. =
ARIES SMITHSONIAN _ INSTITUTION NOILMLILSNI NVINOSHLINS S31YVYEIT_ LIBRARIES SMITHSONIAN _ INSTITUTION
<2 Ks uJ = 2”

- = 2 = AN = = = z |
= cc = : WN oc = oc = |
= oO = XS [20] = ao =
{e) — fo) = ro) = (e)

— big > : 1 > oly =

Tee Oe Cans
hee ges ih eh 1G 464 SE LES Dye ees
ferein s 9 4 at sand PRUE ET a rs ONE ee Ce ate ALS ae q~ ¥ 20.0% 860 ate wes Feed &
eine Crimes 6 gl te wi Cede UNO SEY Cs nee vie ete eyes ete oye Wes / es
Pests bedeh ae sta eset ae ae AE a edn aed va ES eh be : ‘ pret . eee * ~¥
PU ee Se a é mene Stee teste is y Ky 4 Fa
$e ET 8h oth a CNN oes inks we i AMAT A * eae
Me ee TL cea D ts y Pa
athe 8 a S pers ‘ fh

iad en I idea re pirrh Ber Tan ee

A etodhaiebiee 46 08 TSA ts So) RRR ais an ka N
Eine oon aed te ea eae) A
Pe See fain eed ete a

We ES ee ain)
‘ee ee Sheen

Asie

roe ee

by ee hear ere ass
Hine ope beeen pre tary s ‘an Meee “an af vinta mies f
Hee tute diets acto Bae ha ee aay r
TEE Ee aac Oe a Cig tere
Scat did nav he eve ne ad ita Ah ae ord f thee
ede ae Wien AN Hebe oO pene cack « Ee ag has ny
icine ets SR eyteee ae : 7) wet $ : so
ene Aiuen naa: rt yee ited » tas : ‘=
Bi canmmnae ieee ‘ eee 7 nhmoll qQE811.J14 1890 oy
si Teeiiee cele moat te k t Phylogeny of the Pelec ‘yee
i ypods j ‘ae
ei
tes,

meee
et eu

Piper CO Re ay |
rate er a Ty a

Hd woth vt ce aife ram tomvmnrs anak
arty ny Aa eae
std aya ane Oo Se

ne rr ed
Aare oar rc eC ET
Neath eet ON km
Foun

hd’
sarida att

ore #
Tn de Cb te
ter)

Viidnneiu eee
Tole dom ed em mE

Pere ae er
ahh 2 ea el Te

ieldeath en esd a
aie

iaaea eae SO
nebo ne eat
oe ahaa ie Ot
BRANES

ek
natstebeseehat eke
Veookeh ee eae WIM a
Gon eee Dae ot
perro Te es

Me

wi Matatetauttl mrantcts te
ont

fren rr
Crater a7) “eh @s

a
Op ait t
ACL ARME Ne me He a

ere rrr eas

rors Sr
Oe red we tba lie ore
oer rte cea
Leh ae oY
"ine
BNO net

ne

Wevaere ee

ree ee

eM Owen &
€

Cys
Cee bhatt

scagutncnitisced sti
PLP REE Se eel itacea ene
SAAN hel fa land chicos aia ctesaen fates,
‘hea ah eis Gane eisai: pie ‘)
eineerievemneal eye gs AR une ees
Perivart yee penetra

rinks
ieee ae

eu teate:

yar ete
et he ree et erent
perry ere

ry as

Ta bevonsed era abies
erect at enema at a?

Got nal ae Cohen dhe
ofa k on thasnenh 24
sae dae gabe

Fd tet se eater ee
oh ku Timed 2 aRs on
Sreby Teame rts

wren poder ery
ea hide emer vented
Noaen ined)

se ee reat epee rts

“ Bh a etter fs

EE eee todig aie em tna sxe 4s

% sha ets ey A
hae

Thi elec te ey ae
Ane on

pips Vi aot
*

aliyia anteg tensed nest 2 es
Bre pany Sree erry iy

Rae ein oe y
Wale ab isn 43

Ore seen Me mignd «6 Te
pactowee th tah

ed tiny OE a aE
paletinn = 404"

OO wees Fh rere e
ae

AAS LAU TAS

PPR at: ‘ aA ar a at
Aesth EEE A NAL 2
." ath) TORT aS .
EY SS iy ‘ ,
ear . 4

Moss
Neat

WORE Prete carany oie
Rete OP POT a

Ss.
Shek eres
ATRL Soesutth aia
PRTC TNs Cane Sie Se

Sete hi ie

. Sunt OE et

puto en tas Pare cor

CN Ee

RA TUN tee
Feb ks ie

Ree

SMe wha
sargase
PRR rr)

weet ews s
oes

{<a
ENS Smee wea
apes

“praise egy ta

Pr ieee a

Svea ste
oes

Wing ARR oye

Sm Wee arta Sete ab Pee H
OVI A Ses Soe ones Sits See sagan
Ph men Ea he

PeererG ines
Sess octet Soe ONS
Huey
Nnaarese tvs
SS a

Pa AS Gs tie?
spate eee rie aR ca ee at

Semcerenegc ucts retain ge
Sein * Rutan aia ahargs. AMCU Te
Meecuveats Sree peices ae aes ore : ; ti
Wi tya moss usm = wa trove ee 5 f :
Rey ei Peete es ‘ ‘
Taek hits i wit , A \
SNe iii easy 8 i '
sey betd 7 aN eck) Se ee '
eh eh hee CVE Na wtb a teresa cae
ae THETORE REST RE Cee ea
RoHS * Veen e eka
sea Ny tateaa te beva te ‘ 2
ate as ithe i 7
ost sore tear | ; . ove
wats pa Seah, Srey ne} ‘ aia ’
Shaper ae teers Sey z Ma Mar ate. oysep GFE : an
Par arene: 2k cerarynasen ree Gone OE iA i ot SLL hn Vase ‘ veal ‘ 3 Me
ree ues ncacs eat rumy ite : ; Meletoes , : ‘ t bs
base PONCE Ter TL th ooh tha Pal oa wats Pa, r ‘ Hot ae
ratty Hee SSNS Ae EN tae seas AY oe tier NEC ORR Sth ei RA SAE FE
Stag Tisene setae erasure ae oA SRY iter Roa Waa Toticetea Sig dine gece Pee a eats t é " By
aisieiere res aed acase te eaten ifest win : ‘ f
Sh siswat Am wren) ¥ ; d 3

ape wanes, Rea

were eH
HteteNsa etree eee ANS Aru pravecaren sl ewer
sie ccarepurese lk rerucanarg erieUasUN 2 bse cgrcaucnaiie we Perici
Sot were it See ernwraty Ty 4 P } *
at ote Petia erin rineonn a 5 ‘Lk
as ‘ L ‘ oT ERA A '
Tein peop zoe ey nt ESE : nateneees Para Ss
Lee LS eae ea geen wm UNCUT Gee Ak OR Tw rant ;
beri SMRSUAM ie ea ouiee OES Sige a bay oven ee tit Ve G or.
nee Sie ts USAT Re wa Ra meme PEE wh ge Ne LE Ue are Pate ak he ; i *
sayrune aenceteses Sys ee tk bangs axe ug Lee eeg Wed LOMO os
henna tick oat he eeu Rees sc Spe thew, went Fa TA ' es ; t
~ Sa mail Ate sen “ype Yer iret Sct Roe Ta tastes LAE . ft
idurey tse Tare rage ayer #4 et BNRISS ee edie eure er aste AAA SA ih ;
carseat PEL Tabak Hahei SA Te paxil : ro * :
sepa Cee 4 Ay * Ley F
ih weoas SSNs Reino i 4 hay x
om senate ¢ pea suri Get Ve eer Ca a “3

eg WS

ety Perey Men
mae at

aes Saks

Aaa
bh
raresge REO
zi} Meson a AOC mTOR)
He s ass apa oe teak a netey
Taye ee Cr eat Sain vee mu Pare ae
eke Katt meine

resttarvencese
eraser =

i
SRO UNS

ay
a Ny

PANS

onner arin
ieee at

SET T Py Prete

ites

AS ES he nie Ogle
wari artes cr
a ig tem ees CAL ALA
arene Soe ices
wk Am
Pa! eee CRAG , a
SEN a
Cin err) & : il
HG: acts a a
\var ats ae ? i;
i
a

o
saeriuantage tess
aa Sees

hit

a

a

seek

oar eh
Chan
ee ese

aa ah

WAN Ee:

heey +

ne we Ae Pevsvap)
ey Se ee ay ek