.6

MEMOIES

OF THJi

CAKNEGIE MUSEUM.

VOL. IV. NO. 6.

A MONOGRAPH OF THE NAJADES OF PENNSYLVANIA.
By Arnold E. Ortmann, Ph.D.

GENERAL INTRODUCTION.

The writer began to collect Najades in Pennsylvania as early as 1904. Ad-
ditional material was incidentally secured in 1905 and 1906, while he was hunting
for crawfishes. In 1907 the collecting of mussels was systematically begun, and
the particularly favorable seasons of 1908 and 1910, and in part also of 1909,
permitted the taking of large numbers of species and specimens. Not only the
shells but also the soft parts were collected, and the writer endeavored to find
both males and females, the latter in the sterile as well as in the gravid state.
In this he generally was successful, and a great quantity of material representing
the soft parts was preserved in alcohol, partly with the shell, partly without, but
always with the proper records. In this way a collection of the Pennsylvanian
Najades, both of shells and soft parts, has been brought together, which certainly
is superior to any other local collection; in fact, the writer believes that he has
at his command a larger assemblage of the soft parts than is found in all other
collections combined.

The chief purpose of this work was to secure the material for a systematic
monograph of the Najades of our state. A preliminary list of the species found
in Pennsylvania has been previously published (Ortmann, 1909), but since then,
further information has been obtained with respect to the systematic and zoogeo-
graphical side of the subject, and in the present monograph a final account will be
given. In addition, the writer has investigated the anatomical structure of the
mussels, and very important structural features, chiefly in the gills, have been
discovered, which had been in part entirely overlooked by previous authors. The
knowledge of the anatomical details thus obtained has enabled the writer to revise

279

280 MEMOIRS OF THE CARNEGIE MUSEUM

the system of the North American Najades, and to put it upon a more secure
foundation. A preliminary account of this work has also been published (Ortmann,
1910).

The most prominent advance in our ideas as to the systematic arrangement
of the fresh-water mussels is marked by the publication of Simpson's paper on the
Najades (Simpson, 1900). We cannot but admire this work, for the general prin-
ciples of classification are correctly recognized by Simpson, and the soundness of
his views has been amply confirmed by our own investigations. Yet Simpson
suffered under two serious disadvantages : he did not possess enough material repre-
senting the soft parts to settle all systematic questions, and did not make any
microscopical investigation of the anatomical structures. This latter shortcoming
has been in part remedied by Sterki (Sterki,' 1903). In the present monograph
additional observations will be furnished, and with regard to the Pennsylvanian
species, it may be said that the soft parts of all of them, with very few exceptions,
have been at hand, and the}' are now sufficiently well known to enable us to assign
each species to its proper place in the system. Of course in the case of the species
and genera not found in Pennsylvania, this reference remains to be made; but
after the main points of view have been indicated, this will be a comparatively
easy task. The fact, however, should be emphasized that there is now no excuse
whatever for collectors of the Najades preserving only the hard parts, and throwing
away the soft parts. The soft parts form an integral portion of the animal, and
should be preserved and examined together with the hard parts, and this is most
imperative in the case of those genera, of which our present knowledge is more or
less defective.

The greater portion of the material at hand, forming part of the collections
belonging to the Section of Invertebrate Zoology in the Carnegie Museum, was
secured by the writer in person. The art of collecting mussels has to be learned,
because, while it is easy enough to pick them up, it can only be done after a thorough
knowledge of the proper places and seasons has been acquired. The Najades
inhabiting our rivers, creeks, ponds, and lakes, can only be obtained alive by going
after them into the water, but as soon as their whereabouts is discovered, it is
easy enough to get them. Their discovery cannot be accomplished under all con-
ditions. Part of the season our creeks and rivers have too much water, or the
water is too muddy to locate the mussels. This is the case chiefly in early spring,
when the snow melts, and in the first half of summer, the time of the copious
summer rains.

In Pennsylvania the best season is generally from August to October, when

ORTMANN: A MONOGRAPH OF THE NAJADES OF PENNSYLVANIA 281

the water courses are low and clear; but very often there is also a season of low
water from the end of April to the end of May. In smaller creeks, where the
water is never verjr deep, any dry period of about a week's duration will bring
about favorable conditions, but in the larger rivers successful collecting is generally
possible only late in summer and in autumn. The exceptionally dry summers of
1908, 1910, and partly also of 1909, afforded unusual and unexpected facilities
for making collections. Collecting in the lakes (Lake Erie for instance) does not
depend so much on the season, but it depends a good deal on the weather. Quiet
days, when there are no wind and surf, should be selected.

Hand-picking is the best way of securing specimens. Devices like the so-
called "clam-dredge" of the mussel fishermen of the Mississippi and Ohio are useless
in our rivers on account of the rough and rocky character of their bottoms. Yet
in Lake Erie upon sandy and sandy-muddy bottoms I have used a similar apparatus
to great advantage. This is a modification of the common "clam dredge," and
I am especially indebted to Dr. W. J. Holland for suggestions in constructing

Fio. 1. The Holland Clam-dredge.

this device. It consists (See fig. 1) of a piece of gas-pipe with two wheels at the
ends, to which four-pronged hooks are attached. The hooks are made of wire,
with their stems rather long (six to ten inches), and they are tied to the gas pipe
with very short strings: this prevents the tangling up of the whole mass, which is
a great nuisance, when, as in the usual clam dredge, short hooks with long strings
are used. With this apparatus, "the Holland clam-dredge," I obtained shells in
Presque Isle Bay, Lake Erie, at depths of from six to fifteen feet.

In 1910 I had special advantages for collecting Unionidoz in Lake Erie, by
being offered the opportunity to go out with the "sand-sucker" of the Erie Sand
and Gravel Co. I thus obtained specimens at depths of from ten to fifteen feet.

282 MEMOIRS OF THE CARNEGIE MUSEUM

I cannot close these introductory remarks without making public acknowledg-
ment of the encouragement and valuable help furnished to me at all stages of my
work by Dr. W. J. Holland, the Director of the Carnegie Museum. First of all,
the study of the Najades forms part of Dr. Holland's general scheme and his
policy of studying our local fauna (See Holland, W. J., in the Proceedings of the
American Association of Museums, Vol. Ill, 1909, p. 12) and thus the general idea
of this work, as well as my previous work on the crawfishes (See Memoirs of the
Carnegie Museum, II, 1906), is largely due to his suggestion. In addition, he
supplied me with the necessary means for carrying on my collecting, and the chief
item in such work is the expense. As to this, I can only express my great gratifica-
tion and sincere thanks for the liberal allotment of funds made to me. Finally,
Dr. Holland furnished me with all the paraphernalia for the work of the laboratory,
including everything that belongs to its literary and technical side. It goes without
saying that he always followed the progress of my studies with the closest attention,
helped me along in difficult cases, and was ever ready with suggestions and advice.

The wealth of the subject matter to be treated makes it necessary to divide
this memoir into parts. Since it is absolutely essential to lay a proper foundation
for the systematic part, and since the anatomical studies furnish this foundation,
it is obvious, that the latter should be given first. In the present publication
therefore I shall take up questions of anatomy and their bearing upon the system
of the North American Najades. I reserve for a future publication the description
of the various species and their zoogeography.

PART I. ANATOMICAL INVESTIGATIONS.

The anatomy of the European Unionidce appears to be rather well known,
at least after the fashion in which they are treated in the textbooks. (I have con-
sulted those of Cooke, 1895, and of Taylor, 1894 and 1900.) We have also several
accounts of the anatomy of American forms (for instance, Brooks, 1890). But
looking more attentively at these, we find that comparatively few genera and
species have been studied. This is not astonishing in the case of Europe, where
only a few types are found. With reference to the American shells, however, it
is very strange that we possess descriptions, more or less satisfactory, only of
Unio and Anodonta, and only of eastern species of these genera. Further, we
shall see that even in these some of the most important parts of the anatomy have
been neglected.

With regard to the numerous other genera found in the great interior basin,
we possess only scanty notes on the gross anatomy. These have been supplied

ortmann: a monograph of THE NAJADES OF PENNSYLVANIA 283

almost exclusively by Lea and Simpson. The latter in his great "Synopsis" (1900)
gives generalized accounts of the soft parts, chiefly in the diagnoses of the higher
groups, genera, and their subdivisions; but we often remain in doubt whether a
particular species has been examined by Simpson or not. Simpson furnished de-
scriptions of the soft parts of species for another publication (See Baker, 1898),
but in this case it is to be observed that he sometimes described the soft parts of
a species which he did not examine, or rather that he inferred some of the char-
acters from other species, which he believed to be allied. Quadrula coccinea is an
instance. The description of the marsupium is here positively incorrect and was
introduced apparently in the belief that this species, being supposedly a Quadrula,
should have the marsupium of this genus.

Nevertheless, Simpson's observations published in the "Synopsis," are of the
highest value, for he correctly recognized the most essential part of the anatomy,
namely the marsupium. In all Unionidce (as in the Najades in general), the eggs
are carried in the gills, where they develop into larvae, known as glochidia. This
is a well known fact; and it was also long ago known (chiefly through Lea) that
the marsupium is formed in the different species by different gills, or parts of gills.
But this character had never been employed in making a systematic arrangement,
and the credit for recognizing the importance of the marsupium, and for using
it successfully in the division of the Unionidaj into genera and in the grouping of
the latter, distinctly belongs to Simpson.

But Simpson studied only the general features of the marsupium, and thus
his groups are not in all cases without objection. Further in a number of species
he did not succeed in ascertaining the shape of the marsupium, and thus was able
to place them only tentatively. As has been said above, several well supported
corrections have been introduced by Sterki (1903), who paid attention chiefly to
the shape of the glochidium, together with that of the marsupium.

The present writer made it his chief purpose to ascertain the shape of the
marsupium in all species which came under his observation, and to this end to
obtain gravid females. In this undertaking he was generally successful, and, with
very few exceptions, gravid females of all Pennsylvanian species were secured.
But when he took up the closer anatomical study, he made a very important
discovery, namely that in all our Unionida? the anatomical structure of the gills,
which serve as marsupia, is permanently differentiated. It was known and empha-
sized by Sterki that this is externally the case in a certain group of genera (Lamp-
silis-typc), while it was maintained that in the rest no such differentiation was
visible in the female, when not gravid. But this latter assumption is incorrect.

284 MEMOIRS OF THE CARNEGIE MUSEUM

There is always in the female an anatomical difference in the marsupial gills, and
this difference may be detected even without a microscopical investigation, although
the latter is generally advisable in order to render the structure absolutely clear.
In addition the writer has found, chiefly along the edge of the mantle, a
number of structural differences which are also important, and will be described
below.

A. General Structure of the Gills.

The primary elements of the gills are the gill-filaments (gill-tentacles of Brooks)
— cylindrical or somewhat compressed tubes, which arise in two longitudinal rows
from a common base (ctenidium) between the foot and the mantle of the mussel,
and hang down into the branchial chamber. These filaments are doubled back
upon themselves, the outer row outwardly, the inner inwardly. The filaments
of each row being connected with each other (See below), there are thus two gills
on each side of the animal, an inner and an outer one, each consisting of two laminoe
(lamelke) composed of filaments, the laminae including between them a space. The
two laminae are distinguished as the primary and the secondary limb of the gill,
the former being the descending adaxial part, the latter the reflected and ascending
part. The latter part is the outer lamina in the outer, and the inner lamina in
the inner gill.

On the inside of each lamina, toward the cavity of the gill, there develops
a peculiar tissue, called interlamellar outgrowth (Taylor, p. 177). This connects
the gill-filaments of each lamella, leaving more or less regularly placed open holes
(ostia), by which the inner cavity of the gill communicates with the outside. This
interlamellar tissue forms a rather large part of the gill, in fact the lamella itself
appears composed chiefly of this tissue, while the original gill-filaments form only
its outer layer, separated from one another by the intertentacular (Brooks) or inter-
filamentar grooves. The interlamellar tissue is permeated by lacunar blood-vessels.
The filaments develop along their sides chitinous rods for the support of the gill.
Two such rods belong to each filament, but these rods are not continuous, but
interrupted, consisting of longer or shorter, disjointed pieces.

The filaments and their chitinous rods run about vertically, that is to say,
from the base toward the edge of the gill; a slightly fan-shaped arrangement is
sometimes indicated, corresponding to the curved edge of the gill, upon which
they stand vertically.

The interlamellar cavity of each gill is not simple, but is divided. There
are "interlamellar partitions" (Brooks) or "interlamellar junctions" (Taylor), which
connect the two laminae of the gill. In most cases, these connections are developed

ORTMANN: A MONOGRAPH OF THE NAJADES OP PENNSYLVANIA

285

as septa, dividing the inner cavity of the gill. They run about parallel to the gill-
filaments, that is to say, vertically toward the edge of the gill, and consist of a
tissue similar to that of the interlamellar outgrowth, with differences which will
be described below. Most of these septa run continuously all the way from the
base to the edge of the gill, while others are shorter, and are intercalated at various
distances from the base, but each septum is always continued to the edge. Thus
the interlamellar cavity is divided into a number of canals, running from base to
edge, which are called water-tubes. The size of these water-tubes is variable, and
differs considerably according to sex (See below).

There is one exception among the Najades known to me from North America,
which does not show this arrangement, Margaritana margaritifera.1 Here the gills
are also composed of filaments and possess an interlamellar outgrowth, which,
however, is rather slightly and irregularly developed. But there are no septa dividing
the cavity of the gills. Instead of this, the two
laminae are connected by more or less cylindrical
or irregular thickenings of the interlamellar out-
growth, which are rather heavy and solid, and are
placed quite irregularly. Sometimes these con-
nections are a little elongated in the direction
of the filaments, but they never are continuous
septa, but rather patches of tissue, and if any reg-
ularity is observable, it is a somewhat diagonal
or oblique arrangement. Toward the edge of the
gill, and in its hindmost part a vertical arrange-
ment is indicated, but even here there are no com-
plete septa, and consequently no water-tubes (See
fig. 2).

This peculiar development of the interlam-
ellar junctions in Margaritana is in my opinion very important. While the pres-
ence of water-tubes is a character common to all our other Najades, the absence
of this feature in Margaritana is very remarkable, and I have no doubt that this
character, together with the otherwise slight development of the interlamellar out-
growth, which gives to the primary gill-filaments a greater significance in the
whole make-up of the gill, points to a more primitive stage in the development
of the gill in Margaritana, than in any other mussel.

The secondary limb of each gill (outer in the outer, inner in the inner gill)

'In quoting species, I shall, for the present, use Simpson's nomenclature.

Via. 2. Margaritana margaritifera (L.);

right frills spread out and seen from below,
showing arrangement of interlamellar connec-
tions.

286 MEMOIRS OF THE CARNEGIE MUSEUM

extends upward as far as the base of the primary limb. Its distal margin in the
outer gill is fused with the mantle, and thus a small space is cut off from the bran-
chial chamber, which appears as a longitudinal canal along the base of the outer
gill — the suprabranchial or cloacal canal of the outer gill, into which the water-
tubes of this gill open. In the case of the inner gill, the secondary (inner) lamella
is always connected anteriorly with the wall of the visceral mass (abdominal sac)
for a shorter or longer distance; behind this connection it is sometimes free as
far as the posterior end of the foot; and behind this point, the two inner laminae
of the right and left inner gill are connected with each other up to the posterior
end. Thus the inner gill also has a suprabranchial or cloacal canal, which is double
(right and left) anteriorly, lying on either side of the foot and abdominal sac,
but becomes united back of the foot into one cavity. In many cases this canal is
not entirely closed, but communicates with the branchial chamber through a slit
along the sides of the abdominal sac.

There are a number of variations in the development of the characters men-
tioned in the last paragraph.

The outer lamina of the outer gill is united with the mantle up to the posterior
end of the gill in practically all species known to me, the only exception being
in the case of Margaritana marcjaritifera. Here (Compare fig. 6) the outer lamella
is free posteriorly for a considerable distance, and the posterior end of the gills
assumes in consequence a very peculiar shape (See below, under "diaphragm").

As has been stated, the connection of the inner limb of the inner gill with
the abdominal sac is variable. In the most primitive condition, such a connection
is present only at the anterior end of the gill for a short distance, and in this case
the inner lamina is free and leaves a slit-like communication open between the
suprabranchial canal and the branchial chamber. In other forms, this slit is more
or less, often entirely, closed; and then the whole inner lamina of the inner gill
is connected, first to the abdominal sac, and farther behind, back of the hind end
of the foot with that of the gill of the other side, and thus the suprabranchial
canal of the inner gill is entirely separated from the branchial chamber.

Long ago Louis Agassiz called attention to these differences of structure. Later
on their value was doubted by Lea as well as by Simpson, and Sterki also does not
consider them as of much systematic significance. I have investigated my material
with reference to this question, and find that the connection of the inner lamina
of the inner gill with the abdominal sac is indeed variable, and that sometimes it
is differently developed in the same individual on the right and left side. But it
is only in certain forms that this variability occurs. In others this character is quite

ORTMANN: A MONOGRAPH OF THE NAJADES OF PENNSYLVANIA 287

constant. Thus in the species placed by Simpson in the genera Qimdrnla, Pleuro-
bema, Unto, Margaritana, Symphynota, Anodontoides, and Anodonta, I always found
the inner lamina of the inner gill entirely free from the abdominal sac, with the
exception of its anterior end. This is the more remarkable, since both Lea and
Simpson (See Simpson, 1900, p. 766, footnote 2) affirm that in Quadrula mulliplicata,
a species which stands near our Q. undulata, the inner gills are sometimes free,
sometimes connected. I have not seen that particular species; but of Q. undulata
I have examined some thirty specimens, and have never seen a connected inner
gill. Although the number of specimens investigated in this case seems small,
the fact that I never found a connected inner gill in the genera named above, is, I
think, rather significant, for I have looked over about four hundred individuals
belonging to these genera, and it would be very strange, if variations occur here
at all, that I by a curious chance never came across a single one.

Further, I found the inner lamina of the inner gill always almost entirely
free in the following species of other genera: Alasmidonta heterodon (fifteen speci-
mens investigated) Lampsilis parva (three specimens). In Obliquaria reflexa (four
specimens) the gill is also free, but for a shorter distance, a little more than half
of the abdominal sac. The same is the case in Plagiola securis (six specimens),
where the length of the free edge is somewhat variable (in one case the slit was
very short, but only on one side!).

On the other hand there are forms, in which the inner limb of the inner gill
is always entirely united to the abdominal sac. This is the case in the genus
Alasmidonta (excluding A. heterodon), in Obovaria ellipsis, circulus, and retusa, in
the genus Truncilla, and in • many species of the genus Lampsilis (gracilis, alata,
recta, orbiculata, ligamentina, radiata, luteola).

In all those not hitherto mentioned, there is variation in this respect. This
is most evident in the genera Strophitus and Ptychobranchus. In fact in the latter
genus I have observed all conditions, from an inner lamina almost entirely free
to an inner lamina entirely connected. In Strophitus it varies from free for about
half the length of the abdominal sac to entirely connected. In Plagiola elegans
(two specimens) the open slit was in one case short, about one fourth of the length
of the abdominal sac, and in the other there was only a small hole at the posterior
end of the foot. In Lampsilis iris, nasuta, niultiradiata, cariosa, ventricosa, and
ovata, the inner gill is entirely connected, or has a small hole at the posterior ex-
tremity of the foot; and in Micromya fabalis (four specimens), the slit is short,
from about one fourth to one half of the length of the abdomen.

The closing of the slit proceeds always in an antero-posterior direction, and

288

MEMOIRS OF THE CARNEGIE MUSEUM

if there is only a small opening left, this is never placed anteriorly, but always at
the posterior end of the foot.

According to these observations, I must attribute to this character a certain
limited systematic value. There are a number of forms, in which the primitive

condition of an unconnected, free inner lamina of
the inner gill is always found. There are other
forms, which belong to the more advanced types,
in which this lamina is always entirely, or almost
entirely, connected with the abdominal sac. Be-
tween these stand others, in which all transitional
conditions are found, and in some of these an in-
termediate condition seems to be characteristic,
while in others this character is variable.

As has been mentioned above, posteriorly to
the abdominal sac and foot, the inner lamina? of
the two inner gills unite in the median line of the
body (See fig. 3). By this union, together with
the connection of the outer laminae of the outer
gills with the mantle, a complete separation of the
branchial chamber from the posterior part of the
suprabranchial canals (cloacal chamber) is effected
by a septum or diaphragm, which forms a hori-
zontal division, from which the gills hang down
(See fig. 4). At the posterior end the four supra-
branchial canals of the two gills of the right and
left side are united into one cavity, the cloacal
chamber. This is brought about by the fact that
the gills extend posteriorly beyond their attachment to the body at their base, or
rather, that this attachment (ctenidium) does not extend to the posterior end of
the body (See fig. 5). This attachment, the common base of the two primary
limbs of the gills on each side, runs backward to a point, which generally lies
about midway between the posterior end of the foot and the posterior margin of
the mantle; beyond this point the inner limbs of the two gills on each side remain
connected with each other, but are no longer attached to the body, and the
two suprabranchial canals on each side are fused (See figs. 3 and 5). Thus of the
four suprabranchial canals present in the anterior part of the body, each of which
belongs to one gill, the two median canals are united first immediately behind the

Fig. 3. Anodonta grandis gigantea (Lea)
Left half of body, seen from below, with gills
removed; showing attachment of gills. — pip,
palpus; ct, ctenidium (base of gills); a, line
attachment of outer lamina of outer gill to
mantle; 6, attachment of inner lamina of in-
ner gill to abdominal sac; 61, free edge of the
latter; 6s, line of connection of the two inner
laminae of right and left gill; o, sexual and
renal orifices.

ORTMANN: A MONOGRAPH OF THE NAJADES OF PENNSYLVANIA

289

posterior end of the foot; and a little bit farther back, the two outer canals (be
longing to the outer gills) open into and unite with this cloacal chamber.2

In a posterior view (See fig. 4) the diaphragm separating
the branchial and cloacal openings forms a rather short, sim-
ple, horizontal bridge, from which the gills hang down.
Sometimes there is a slight median incision, when the two in-
ner laminae of the inner gills are not fully united to their
ends, but according to my investigations this has no system-
atic value.

We are to note an exception to the normal structure, and
this is again Margaritana margaritifera. Here (See fig. 6) the
diaphragm represents an entirely different aspect, which is
due to the fact that, as we have seen above, the outer lam-
inae of the outer gills are not connected with the mantle up to
their ends. Thus on the one hand the outer attachment of
the diaphragm is not seen at the outer edge of the mantle, and the outer laminae
of the outer gills are free for a considerable distance, which causes the whole dia-

Fiu. 4. Anodonia grandit

giyantea (Lea). Posterior
view, showing branchial,
anal, and supra-anal open-
ings, diaphragm, and pos-
terior ends of gills.

Fio. 5. Anodonta grandis ffigantea (Lea). Posterior end of Fin. 0. Margaritarui margaritifera (L.). Posterior view,
body, seen from below, with the diaphragm cut open in the showing branchial and anal openings, and the peculiar, in-
median line, to show cloacal chamber. — 0, openings of the su- complete diaphragm. No supra-anal is separated from the
prabranchial canals of inner gills; ?', openings of the supra- anal
branchial canals of outer gills.

phragm to appear as if it were tapering to a median point; and, on the other
hand, the cloacal opening (anal siphon) is much larger than usual.

2This condition is not clearly represented in Brooks' (1800, p. 275, fig. 143) diagrammatic figure. The openings of the
suprabranchial canals of the outer gills into the common posterior cloacal cavity are not shown. Their position would be
indicated by the two nervous strands starting from the parietosplanohnic ganglia and curving backward toward the posterior
end of the gills.

290 MEMOIRS OF THE CARNEGIE MUSEUM

B. Structural Features of the Finer Anatomy of the Gills.
1. The Water-tubes and the Septa dividing them.

The following description of the water-tubes necessarily leaves out of con-
sideration the peculiar form, Margaritana margaritifera, for this has no water-tubes.

We have seen above that the water-tubes are formed by the interlamellar
junctions, which run from near the base of the gill to its edge, and are generally
developed in the shape of continuous septa, which do not differ in their tissue
from that of the interlamellar outgrowth, that is to say, in their normal develop-
ment their tissue is lacunar, often containing blood-vessels (See PL LXXXVI,
figs. 1, 5, 6a, 7a, 10a, 12a, 13, 14a, and PL LXXXVII, figs. 1, 4, 8).

This simplest structure is always found in both gills of the male, and there is
no further differentiation of it in this sex, with the exception that toward the
base of the gills the septa become sometimes a little more independent, being
wider on account of the greater distance of the two laminae from each other. The
water-tubes formed by these septa are rather wide in the longitudinal direction
of the gill, since the septa are distant from each other. The distances are variable,
and often, chiefly so in the genera Alasmidonta, Strophitus, Symphynota, Anodon-
toides, and Anodonta, they are rather large near the base of the gill, where also the
septa become irregular, running for a short distance diagonally, and being some-
times somewhat reticulate. In the genera just named there are often as many as
forty and even more filaments between two septa. The average is from about
fifteen to thirty; in other genera, the average is from about ten to twenty or
twenty-five, although there is occasionally a larger or smaller number present.
We may say, that normally there are more than fifteen filaments between two
septa.

The identical structure is found in those gills of the female, which do not
serve as marsupia, that is to say, in the inner gill, when the outer gill alone is
used as a marsupium, and in the inner and the anterior part of the outer gill,
when only the posterior part of the outer gill is so used (See PL LXXXVI, figs.
6a, 7a, 10a, 11a, 12a, 14a; PL LXXXVII, fig. 5). In those forms (genus Quadrula),
where both gills serve as marsupia, this structure is not found at all in the female.
(See PL LXXXVI, figs. 2, 4.)

In the marsupial gills of the female important structural modifications take
place, and these modifications, which we shall describe presently, are quite marked,
and invariably characterize the female sex. Consequently it is possible not only
to distinguish a sterile female (with the marsupium not charged) from a male, but
also to discover the general character of the marsupium in a sterile female.

ORTMANN: A MONOGRAPH OF THE NAJADES OF PENNSYLVANIA 291

In the sterile female the septa separating the water-tubes in the marsupial gill
(See PI. LXXXVI, figs. 2, 4, 76, 106, 146; PL LXXXVII, figs. 2, 6, 7, 9) are much
more crowded than in the male. There are, on an average, only from about five
to ten filaments between two septa, although in some cases this number may increase
to twenty. But in every species this number is considerably smaller in the female
than in the male. This difference is very striking, and may be detected by the
naked eye on holding up the gill toward the light. When I was hunting for sterile
females among my material I always used this method in discovering the sex, and
the much more crowded septa in the gill examined always indicated the female sex.
I never made a mistake, and a specimen selected as a sterile female according to
this test always proved to belong to this sex, when a closer microscopical examina-
tion was made.

In consequence of the more crowded condition of the septa the water-tubes
have a different cross-section and are not elongated in the longitudinal direc-
tion of the gills, as those of the male, but are either more rounded or square (See
PL LXXXVI, figs. 10, 14; PL LXXXVII, figs. 2, 9), or even elongated in the
transverse direction (See PL LXXXVI, fig. 7).

But not only does the more crowded character of the septa indicate a marsupial
gill, but the septa themselves have a different structure. While in the male and the
non-marsupial gill of the female the septa are not much different from the inter-
lamellar outgrowth, in the marsupial gill they are considerably altered. They
are more strongly developed, appear more as independent structures, and their tissue
is more compact, without lacuna? or blood-vessels. Generally, they are thicker,
and also longer in the direction from lamina to lamina, and, what is their most
striking character, their epithelial layer is quite different. While in the male type,
described above, the epithelium, which faces the water-tubes, is a simple layer, in
the marsupial gill of the sterile female it is folded and thrown up into wrinkles,
which often assume considerable proportions (Compare PL LXXXVI, figs. 1 and
2, 7a and 76; and PL LXXXVII, figs. 1 and 2, & and 9).

There is no question that this peculiar structure of the septa of the marsupial
gills is an adaptation to their function. When the marsupium in the breeding
season becomes charged with eggs, these gills naturally must swell, and they
actually do so, as is well known. The swelling can be only in one direction, namely
transversely to the longitudinal plane of the gill, that is, in the direction from
lamina to lamina. This means, that the two laminae of the gill are forced apart,
and since they are connected by the septa, the latter, if they are not to be torn,
must be built accordingly. The description and the figures given here of the

292 MEMOIRS OF THE CARNEGIE MUSEUM

structure of these septa show that the latter is highly favorable to stretching,
and while they stretch out, they become thinner, and the folds of the epithelium
become smoothened and flattened (Compare PL LXXXVI, figs. 2 and 3, 8 and
9, 14 and 15; and PL LXXXVII, figs. 2 and 3). After the discharge of the
mature embryos (glochidia) the gill and its septa assume again their normal
shape.

This structural feature of the marsupial gill of the female is met with in all
our Unionidce, and in some it is the only one found. To the latter belong the
following genera of Simpson: Quadrula, Pleurobema, Unio, and Tritogonia. But
among these we observe a division into two groups. In some both gills, inner
and outer, have the structure of marsupial gills, and become charged with eggs
in the breeding season (See PL LXXXVI, figs. 2, 3, 4); in others it is only the
outer gill which serves this purpose (See PL LXXXVI, figs. 6, 7, 10, 11, 12, 14).
According to my investigations, in the following species both gills are marsupial:
Quadrula subrotunda (Lea); Q. kirtlandiana (Lea); Q. rubiginosa (Lea); Q. trigona
(Lea); Q. pustulosa (Lea); Q. lachrymosa (Lea) (according to material from
Kansas); Q. metanevra (Rafinesque) ; Q. cylindrica (Say); Q. undulata (Barnes);
Q. hippopcea (Lea); Tritogonia tuberculata (Barnes).

Only the outer gills serve as marsupia in : Quadrula tuberculata (Rafinesque) ;
Q. pyramidata (Lea); Q. obliqua (Lamarck); Q. coccinea (Conrad); Q. cooperiana
(Lea); Pleurobema cesopus (Green); P. clava (Lamarck); Unio crassidens Lamarck;
U. gibbosus Barnes; U. complanatus (Dillwyn); U. productus Conrad.

The bearing of these facts upon the systematic arrangement will be discussed
later.

There are further structural modifications of the water-tubes found in an associa-
tion of forms of Unionidce, which apparently compose a natural group, and have
been already recognized as such by Sterki. This group is composed of the genera
Anodonta, Anodontoides, Symphynota, Strophitus, and Alasmidonta of Simpson's
"Synopsis." Of these I have investigated the following species: Anodonta cataracta
Say; A. grandis Say; A. imbecillis Say; Anodontoides ferussacianus (Lea) (and
var. subcylindraceus (Lea)); Symphynota compressa Lea; S. viridis (Conrad); S.
costata (Rafinesque); S. complanata (Barnes); Strophitus edentulus (Say); Alas-
midonta undulata (Say); Alasmidonta marginata (Say) (and the var. varicosa
(Lamarck)); A. heterodon (Lea).3

3The existence of the structures described in the following paragraphs has been denied by Lefevre and Curtis (1910).
As I have pointed out, however, (Ortmann, 1910a), their own figures show them. These lateral water-tubes cannot be blood-
vessels, for which they apparently have been taken by these two authors, and this is so easily seen (even macroscopically)
that we do not need to discuss it any further.

ORTMANN: A MONOGRAPH OF THE NAJADES OF PENNSYLVANIA 293

In all these only the outer gill serves as a marsupium, and during the breeding
season, and only during this time, each water-tube is divided into three tubes running
from the base to the edge of the gill, the division being effected by two secondary
septa, which grow out at right angles from the primary septa (or interlamellar
junctions). Two of these secondary water-tubes are located toward the faces
(lamina?) of the gill, while the third one, which is much larger than the lateral
ones, occupies the middle, and is separated from the two laminae by the lateral
tubes. Only this central tube contains ova and embryos, and properly should be
called an ovisac (See the cross-sections, PL LXXXVI, figs. 8, 9, 116, 16). Thus
we see that, while in the forms discussed above the whole water-tube assumes the
function of an ovisac, in the present cases only a part of the original water-tube
discharges this function.

In addition, these ovisacs of the Anodonta-group are also closed above, at the
base of the gill, by a similar tissue growing out from the septa at their lower ends
(See PI. LXXXVIII, figs. 3a, 11). As far as I can see, this tissue, which forms the
ovisacs here, is furnished by, and forms part of, the epithelial layer of the septa.

This structure apparently is an adaptation to the lengthened breeding season
and the extreme swelling of the marsupium, which is observed in these forms, and
serves for the aeration of the mass of embryos. The swelling of the marsupium
during the breeding season is very great here, much greater than in any of the
forms belonging to the Quadrula-group. Thus it might become difficult to furnish
the necessary oxygen to the developing eggs in the ordinary way, by the minute
ostia perforating the lamina?, and the cutting off of a part of each water-tube close
to each lamina must have the effect of keeping up a lively current of water within
the gill, between the lamina? and the central ovisac.

As has been said, this threefold division of the tubes is present only during
the breeding season. I have examined numerous sterile females, in which this struc-
ture is not seen (PI. LXXXVI, fig. 15); in others it is indicated (PI LXXXVI,
figs. 76, 126) by a stronger fold (or group of folds) of the epithelial layer of the
septa, near their insertion into the lamina?. In specimens where the eggs begin to
go into the gills, this structure is sometimes not yet developed, but it appears soon,
and the epithelial folds, which form the secondary septa within the water-tubes,
begin to grow into the lumen of the water- tubes, and the folds of the opposing
faces of the two septa finally unite in the middle. The point of union (cross-
section of the line of union) is often distinctly seen in sections. After the dis-
charge of the glochidia these structures disappear rapidly. I have seen them still
present in some specimens, from which the embryos had recently been discharged

294 MEMOIRS OF THE CARNEGIE MUSEUM

(PI. LXXXVI, fig. 126) ; in others they were greatly reduced, and the septa had
assumed their normal shape; but in specimens of this group, where the marsupium
was well filled with eggs or embryos, I have always seen them, although it is some-
times hard to obtain good slides, especially when the embryos have reached the
glochidium-stage, since at that time the making of sections is rather difficult.

There is one form among those above enumerated, in which the development
of this structure goes a step further. This is Strophitus. It was known long ago
that this genus is characterized by a peculiar arrangement of the ova in the gills,
they forming short, horizontal rows, which run transversely from lamina to lamina.
According to my investigations this arrangement is brought about by further out-
growths of the epithelial layers of the septa, which fill the spaces between two septa,
or rather only the middle part, the ovisac (See PI. LXXXVIII, figs. 7, 8, 9), and
thus the simple ovisac of Anodonta and other genera is here divided into a number
of smaller, secondary ovisacs, running transversely across the gill, each of which
contains a short, more or less cylindrical mass of eggs or embryos. These masses
(placentulse) will be discussed later. Also in Strophitus these structures are not
present in sterile females, and after the discharge of the glochidia they soon dis-
appear. The gradual development is seen in the vertical longitudinal sections of
the marsupial gill on PL LXXXVIII, figs. 6 to 9.

All other forms of Unionidcv, which have not thus far been mentioned, have
the simple structure of the marsupial gills, as described in the case of Quadrula,
etc., but there are other differentiating structural features on the edges of the
gills, which must be compared with the more primitive forms.

2. Differentiating Structures on the edge of the gills.

There is a difference in the structure of the edge of the inner and the outer gill,
which is found in all our ATajades, including Margaritana, and of this I shall first
speak.

Making vertical cross-sections through the gills of a male (See PL LXXXVII,
figs. 11, 12, 15; PL LXXXVIII, fig. 12), we observe that the edge of the outer
gill is simply rounded off, while that of the inner gill possesses a peculiar longitudinal
furrow along it, which may also be distinctly seen macroscopically. As to the
meaning of this furrow, I can not make any suggestion, and am only able to state
the fact of its presence. By the presence of this furrow an inner may always be
distinguished from an outer gill.

The same difference is found in the inner and outer gill of the female (See
PL LXXXVII, fig. 13), and the conformation of the edge of the gill is practically

ORTMANN: A MONOGRAPH OF THE NAJADES OF PENNSYLVANIA 295

identical with that of the male in all sterile individuals. There is no change in
this matter in the gravid females of the genera Unio, Pleurobema, and Quadrula.
The result is that the edge of the marsupial gill does not participate in the swelling
of the gill. The greatest swelling in these genera occurs near the base or in the
middle of the gill, while it decreases toward the edge, which always remains rather
sharp, and this sharp edge of the marsupial gill is a very important macroscopical
character in the species of these genera. A microscopical examination shows that
there is no difference of the structure of the edge from that of the sterile female
(Compare PI. LXXXVII, figs. 13, 14, 16; PI. LXXXVIII, fig. 1).

In our other Unionida? it is different, and we distinguish two types of develop-
ment, though in all cases the differentiation concerns only the edge of the outer,
or marsupial gill.

The first type is found in the Anodonta-group (in those forms, which have
the water- tubes divided into three tubes, as described above) . It has already been
said that in these forms the marsupium swells to an extreme degree, and this
immense swelling also affects the edge. The two laminae of the gill draw apart
at the edge, and this strain would cause a rupture at this point, if it were not for
the fact that there is here an extraordinary thickness of tissue, which is capable
of stretching, and thus keeps the edge of the gill closed (See PI. LXXXVIII, figs.
26, 36, 5, 10). There are no filaments, or chitinous rods, or ostia visible in this
elastic tissue.

This transverse stretching of the edge of the gill takes place in various degrees,
and is well observed only in marsupia, which are fully distended. Even in fully
charged marsupia it is often absent in certain parts, chiefly so toward the anterior
and posterior ends of the edge of the gill, which then remains simple, but appears
as a kind of ridge or cord upon the swollen marsupium. In the middle of the gill,
however, this stretching takes place regularly, and is sometimes accompanied by a
slight bulging out of the edge, beyond the original margin of the gill (PL LXXXVIII,
figs. 36, 10). In other cases such bulging is hardly noticeable, and the edge of
the gill does not appear rounded off, but rather truncated (PI. LXXXVIII, figs.
26, 5) . The peculiarities of this structure may be easily seen with the naked eye
or with an ordinary magnifying glass, and the most striking effect is that the edge
of the marsupium in these forms does not appear sharp, as in the [/mo-group, but
blunt, rounded off, or truncated. The single water-tubes (or ovisacs) are not dis-
tinctly marked off externally along the edge of the marsupium.

The other type of the specialization occurring on the edge of the marsupium
is found in the Lampsilis-gvoup, to which belong the following genera and species
of Simpson's "Synopsis," females of which I have investigated:

296 MEMOIRS OF THE CARNEGIE MUSEUM

Truncilla triquetra Rafinesque; T. perplexa rangiana (Lea); Micromya fabalis
(Lea); Lampsilis ventricosa (Barnes); L. ovata (Say); L. cariosa (Say); L. mul-
tiradiata (Lea); L. radiata (Gmelin); L. luteola (Lamarck); L. ligamentina (La-
marck); L. orbiculata (Hildreth); L. recta (Lamarck); L. nasuta (Say); L. iris
(Lea); L. parva (Barnes); L. alata (Say); L. gracilis (Barnes); Obovaria retusa
(Lamarck) ; 0. circulus (Lea) ; 0. ellipsis (Lea) ; Plagiola securis (Lea) ; Obliquaria
reflexa Rafinesque; Ptychobranchus phaseolus (Hildreth).

This type resembles the one just described in so far as the two lamina? of
the gill yield at the edge to the tendency of the marsupium to swell. The swelling
of the gill itself is generally not carried to the same excessive degree as in the
^4noctonta-group, although it is sometimes quite considerable; in its stead, the chief
expansion takes place at the edge itself, and in this case beyond it. The thickened
tissue is not simply stretched out in a transverse direction, but it bulges out,
thus protruding beyond the original edge of the gill (See PI. LXXXVIII, fig. 14).
The latter is always clearly indicated by a line at which the filaments and their
chitinous rods stop; and this line may be seen with the naked eye or with a low
power lens. Beyond this line the gill sometimes projects to a considerable distance,
and this projection, into which the septa are continued, acquires a peculiar appear-
ance. It does not project and bulge out uniformly, but the septa on the inside,
acting as stays, check the bulging at their lines of insertion, and thus the edge of
the marsupium assumes a beaded appearance, each bead representing a water-tube,
being separated from the adjoining bead by a kind of constriction or sulcus, repre-
senting the septum. This protruding, beaded structure of the marsupial edge is
very characteristic of these forms, and is distinctly visible to the naked eye.
Apparently Simpson's diagnostic character, that here the ovisacs are distinctly
"marked out by a sulcus," refers to the edge of the marsupium; for upon the
lateral faces of the marsupium there are no such distinct sulci, and there is hardly
any difference in the appearance of these faces from what is seen in the other
genera, which, according to Simpson, have no such „sulci. However, when the
water-tubes are very large, as for instance in Lampsilis alata and gracilis, the
sulci are a little more distinct on the lateral faces.

In the genus Ptychobranchus (PI. LXXXVIII, figs. 13, 14), the bulging out
of the edge takes place all along the gill, which, in addition, is peculiar^ folded,
as is well known. In the rest of the genera the bulging is localized, and it is alwaj^s
the posterior section of the outer gill which shows this structure, sometimes to a
greater, sometimes to a lesser extent. The fact that the marsupium is restricted
in these genera, is a character emphasized by Simpson. There is some differentia-

ortmann: a monograph of THE NAJADES OF PENNSYLVANIA 297

tion in this, for instance in Obliguaria it is only a small part of the gill immediately
behind the middle which forms the marsupium. But this is, in my opinion, rather
unimportant, for in the typical forms of Lampsilis, Obovaria, and Plagiola very
often the outmost posterior end of the outer gill is not used as marsupium. But
it is without exception a rule that the marsupial part always projects beyond the
original edge of the gill (as marked by the filaments); and this is also noticed,
although in a lesser degree, in the sterile female. When the gills are not charged,
of course, the protruding, outbulging tissue is much contracted; yet the mass of
this tissue is so considerable, that it is even then clearly seen, and it keeps the
laminae of the gill permanently apart at the edge (See PI. LXXXVIII, figs. 13, 15,
17, 21), so that the beaded appearance is also preserved. By this character a
female of these forms may always be recognized even when sterile, and further
the marsupial edge may generally be distinguished by its color. Often it appears
simply opaque whitish, in distinction from the water-soaked, transparent tissue
of the rest of the gill, but in other cases a peculiar dark pigment develops, chiefly
on the inner side of this tissue in the epithelial layers (See PL LXXXVIII, figs.
15, 17, 18, 19, 20).

The part of the outer gill of the female which is not used as marsupium has
the normal structure, that is to say, it corresponds essentially to the structure of
the edge of the outer gill of the male. In sterile females it simply forms the
anterior continuation of the marsupial part, and is not marked off, except by the
lesser elevation of the edge (PI. LXXXVII, figs. 6, 7, 10). In the gravid female,
however, the marsupium swells principally at the edge, and it also increases its
longitudinal (antero-posterior) dimension. Thus it is forced to push past the
anterior part of the gill, forming a fold with it (See PI. LXXXVII, fig. 5). This
fold probably is analogous to the numerous folds seen in the genus Ptychobranchus
(PI. LXXXVII, fig. 3), at any rate it serves the same purpose. Sometimes, when
also the hindmost part of the outer gill does not take part in the formation of the
marsupium, a similar small fold is also seen posteriorly, but this is rather insignifi-
cant and variable. At the anterior fold, the marsupium pushes forward on the
inside of the anterior, non-marsupial part of the gill.

C. The ova and the larvae (glochidia).4
1. Arrangement of the ova and embryos in the marsupium.
As is well known, the sexual orifices are situated on each side of the abdominal
sac, and open into the anterior part of the suprabranchial canal of the inner gill.

'Wherever I use the words "ova" or "onus," as distinguished from "gloohidia," I mean (lie eggs and their subse-
quent stages of development, before the glochidial shell is formed.

298 MEMOIRS OF THE CARNEGIE MUSEUM

From this point the way is open for the eggs all along the base of the inner gill
backward to the common cloacal chamber and forward again into the suprabranchial
canal of the outer gill. From the suprabranchial canals at the bases of the gills
the ova reach the water-tubes in the gills. I have repeatedly found specimens, in
which this process was going on, and where both suprabranchial canals were more
or less filled with ova. (Instances are Ptychobranchus phaseolus, Strophitus eden-
tulus, Alasmidonta undulata, Quadrula undulata.)

After the eggs have reached the water-tubes (ovisacs), they are firmly packed
together there, until the marsupium is filled to its utmost capacity. In no
case have I observed that the ova become connected in any way with the tissue
of the gills; on the contrary, they always remain free from them, but generally are
connected with each other, so that the contents of each water-tube or ovisac form
a more or less solid mass, which in its shape conforms to the shape of the ovisac.
The connection of the eggs between themselves is merely a "sticking together,"
which sometimes is rather slight, but produces in other cases a rather firm mutual
adhesion. It is accomplished by the outer egg-membranes, which apparently have
the property of freely imbibing water, and thus of becoming gelatinous.5 For
these gelatinous masses containing the eggs we may use the term placentae, intro-
duced first by Sterki with reference to the genus Strophitus. (As will be seen later,
I call the placenta; of Strophitus placentidaz.) These placentae may be more or
less compact (See PL LXXXVI, fig. 3; PI. LXXXVII, fig. 3; PL LXXXVIII, fig.
14) ; they are well developed in Quadrula subrotunda, kirtlandiana, and rubiginosa,
where they are rather cylindrical, since the water-tubes do not expand very much
and remain subcylindrical in the charged marsupium. In Ptychobranchus the
placentas are also very solid, and there is in the outer layer a peculiar brown stain
developed, when the glochidia are formed. The placenta} are subcylindrical here,
but somewhat club-shaped, on account of the swollen distal end. In all these
forms, it is easy to take the placentae out of the gill whole, without injury.

In Quadrula metanevra, undulata, coccinea, in the species of Pleurobema and
Unio, in all forms of the Lampsilis-type (except Ptychobranchus), the water-tubes
are more or less stretched out transversely, so that their lumen becomes compressed,
and consequently also the placenta} have a similar shape, leaf-like, lanceolate, or
elliptical. Here the mutual adhesion of the eggs is not so firm, so that they come
apart more easily, and it is not feasible to take the placentae out entire. In some
of these forms (Quadnda undulata, Unio, and the Lanipsilis-species) , the adhesion

'There may be a BecretioD furnished by the epithelium of the watei-tubes. This question is reserved for further his-
tological investigation.

ORTMANN: A MONOGRAPH OF THE NAJADES OF PENNSYLVANIA 299

is noticed only in the eggs, while later on the glochidia are more or less free. In
Quadrula coccinea and Pleurobema, however, the placentae are more solid, and remain
so until they are discharged (See below) .

In Strophitus, as we have mentioned, peculiar conditions prevail. The ovisac
(inner part of a water-tube) is subdivided into smaller partitions, running cross-
wise to the gill; and in each of these compartments, which are almost cylindrical,
a small number of ova are located which again stick together rather firmly
and can easily be taken out whole. These "placentae" of Sterki, consequently,
are not entirely homologous to the placentae of the other genera, since in the latter
each ovisac contains only one placenta, while in Strophitus there are many of them
in one ovisac. The eggs of Strophitus form a single, irregular row in each placentula,
as T should prefer to call them here, and there are from about two to eight or ten
(generally seven) in each placentula.6

In the genera Anodonta, Anodontoides, Symphynota, and Alasmidonta, no pla-
centae appear to be present. The eggs as well as the glochidia fill the ovisacs in
immense numbers, and seem to be entirely free. At any rate I never succeeded in
isolating the egg-mass from an ovisac, but invariably, as soon as an ovisac is
injured, the eggs or glochidia flow out freely, without sticking together, although
in some instances in some of my sections of the genus Symphynota, a placenta-like
cohesion is here and there indicated in the younger eggs.

2. The Glochidia.

I am not prepared to say where the fertilization of the eggs takes place. It
must take place somewhere between their issue from the genital orifice, and their
final deposition in the gills. It may take place in the suprabranchial canals, or
in the gills. At any rate as soon as the marsupium is well filled cleavage begins,
and the eggs develop within the gill into the young larval shell, known as glochidium.
The development from the egg to the glochidium is rather rapid, while the glochi-
dium may remain unchanged a long time in the marsupium before it is discharged.

We generally find in individual shells a rather uniform stage of development
of the embryos, and, as a rule, when there are eggs, there are onty eggs, and when
there are glochidia, we find only these. Nevertheless, as already observed by
Sterki (1898, p. 19) in Cyprogenia, there are sometimes cases, where all stages from
the egg to the fully developed glochidium are found in one and the same marsupium.

6I have evidence in some of my slides that not all eggs in a placentula of StropKittu develop into glochidia, but that
some become abortive. It may possibly be that the latter furnish material for the placentula!. In other species also I
have seen what looks like abortive eggs (see Lampsilis alata, PI. LXXXIX, fig. 18). This matter should be investigated
more closely. Sterki (1898, p. 19) already mentions it.

300 MEMOIRS OF^THE CARNEGIE MUSEUM

I have seen the identical condition in Anodontoides and Symphynota, with young
eggs at the base of the gills, glochidia near the edge, and with the intermediate
stages in the middle.

The shape of the glochidium has been studied repeatedly by different writers,
but the credit of using it successfully for the systematic arrangement of the genera
belongs to Sterki. The glochidium of Anodonta possibly is best known. It is
rather large, subtriangular, with a spine at the tip of each valve (PL LXXXIX,
fig. 13). Yet this shape cannot be regarded as typical, and it represents a specializa-
tion. As Sterki points out, it is found only in that group of genera, which I have
repeatedly called the Anodonta-gr oup, and it is one of the diagnostic features of
this group.

I have seen this glochidium in the following forms: Strophitus edentulus
(Say); Anodonta grandis Say; A. imbecillis Say (PL LXXXIX, fig. 13);
Anodontoides ferussacianus (Lea) (PL LXXXIX, fig. 12); Symphynota compressa
Lea (PL LXXXIX, fig. 10); S. viridis (Conrad); S. costata (Rafinesque) ; S.
complanata (Barnes) (PL LXXXIX, fig. 11); Alasmidonta undulata (Say) (PL
LXXXIX, fig. 9); A. marginata (Say); A. marginata varicosa (Lamarck); A.
heterodon (Lea) (PL LXXXIX, fig. 8).

In all these the glochidia are rather uniform in size as well as in shape. Gen-
erally they are rather high and pointed, but sometimes (Symphynota compressa
and viridis) they are wider and less pointed. We do not need to go into detail
here, since most of them have been already figured by Lea (Observ. VI and XIII).7
The glochidia of Alasmidonta undulata and Anodontoides ferussacianus are here
figured for the first time (PL LXXXIX, figs. 9 and 12).

Lea, who is the only writer, who has furnished drawings of the glochidia,
has observed the same type in the following additional species: Margaritana del-
toidea (= Alasmidonta calceola) (See Observ. VI); Margaritana spillmanni (= Stro-
phitus spillmanni) (See Observ. XIII). They also belong to the same group. On
the other hand, Lea says of Margaritana triangulata (= Alasmidonta triangulata)
(See Observ. VI) from Georgia, that he could not observe any hooks, and the same
is the case in Margaritana hildrethiana (= Hemilastena ambigua, see Observ. VI),
Anodonta ferussaciana (= Anodontoides ferussaciana ibid.), and in Anodonta ovata
(= grandis, see Observ. X). All these species should have hooks, if Sterki's theory,
that this is a fundamental character of the glochidia of this group, is correct.

'A number of Lea's figures are rather unsatisfactory, because they are not drawn to scale. Thus, for instance, the
glochidium of Plagiola securis (Observ. VI, pi. 5, fig. 6) does not show its real character, being too small in comparison with
those of other species (also the gaping margins are not shown) . The glochidia on our Plate LXXXIX are all from photographs
taken under the same enlargement, and thus show the actual comparative proportions.

ortmann: a monograph of THE NAJADES OF PENNSYLVANIA 301

And indeed I have seen the hooks in the two last-named species. Lea himself
in the case of Hemilastcna suggests that the glochidia he has seen are too young,
and that the hooks might develop later. And this is apparently true, for I have
observed in a number of species, that young glochidia have no hooks, while they
are present in older ones. In the following cases this was most evident: specimens
of Alasmidonta marginata, collected Sept. 2, 1907; of Alasmidonta marginata varicosa,
collected Aug. 13, 1908; of Symphynota viridis, collected Sept. 6, 1909; of Symphy-
nota costata, collected Aug. 31, 1906, and Sept. 21, 1907; of Symphynota complanata,
collected Sept. 14, 1909; and of Strophitus edentulus, collected Aug. 24, 1909. All
had no hooks, and we see that they all were collected in the months of August
and September, the beginning of the breeding season. In other specimens of the
identical species, collected generally a littler later in the season or in the spring,
hooks (spines) were present.

Thus we have every reason to believe that Sterki's opinion that this kind of
glochidium is common to, and characteristic of , the genera Strophitus, Alasmidonta,
and Anodonta, is essentially correct, and that it should be enlarged so as to embrace
all forms belonging to the "Anodonta-group."

A second type of glochidium presents a somewhat rectangular shape, more or
less dilated ventrally (celt-shaped), with two spines or hooks, one at each of the
ventral corners of the shell; further the lateral (anterior and posterior) margins
of this shell are here not in contact all along their edges, as usual, but are gaping.
In Pennsylvanian species, I have only observed this type in Lampsilis alata (Say)
(See PI. LXXXIX, fig. 18), and in addition in L. Icevissima (Lea), from Kansas.
Lea describes it (Observ. X and XIII) from the same species and from L. purpurata
(Lamarck) .

These species stand in Simpson's system in the subgenus Proptera of Lampsilis.
This has induced Sterki to believe that this shape of glochidium is common to all
forms belonging to Proptera, and to elevate it, on that account, to generic rank.

Although I agree with Sterki in making Proptera a genus by itself, I would
not place all species listed under Simpson's subgenus into it, for we have at least
one species, which has a glochidium of a different shape. This is Lampsilis {Prop-
tera) gracilis (Barnes) (See PI. LXXXIX, fig. 19). Here the glochidia are in
shape similar to those of the third type, to be described presently, but they differ
considerably from them in size, for they are much smaller. In fact, the glochidia
of this species are by far the smallest I know among our Unionidce. The natural
consequence is that we are compelled to create a new genus (Paraptera) for it.

In the shape of the glochidia Lampsilis alata (and the forms allied to it)

302 MEMOIRS OF THE CARNEGIE MUSEUM

stands rather isolated, but the gaping margins are found at least in one other,
species, Plagiola securis (Lea) (See PI. LXXXIX, fig. 17). 8 Here the general shape
is similar to the third type, but it is rather distinctly spatulate and quite large.
I think that the glochidium of Plagiola securis is the only one which might be
compared with that of Lampsilis alata, and it is rather significant that the only
glochidium, which permits such a comparison, is found in a group, which, as we
shall see later, is rather closely allied to Proptera.

The third type of glochidium is found in all the rest of the genera of Simpson
investigated by myself. Here the shell is semicircular or semielliptical, the ventral
margin evenly rounded, and no spines are present. With the exception of Plagiola
securis the margins shut completely. There are minor differences in size and shape;
as has been mentioned, those of Paraptera gracilis are very small (See PI. LXXXIX,
fig. 19). Of those of the genera Quadrula, Pleurobema, Unio, Ptychobranchus, and
Obovaria, we may say, that they are generally rather small, while those of Lampsilis
and Truncilla are somewhat larger.

The following is a list of the Pennsylvanian species, in which I have seen
glochidiaof this type: Truncilla triquetra Rafinesque (PI. LXXXIX, fig. 24); Trun-
cilla perplexa rangiana (Lea) ; Lampsilis ventricosa (Barnes) (PI. LXXXIX, fig. 23) ;
L. ovata (Say); L. cariosa (Say); L. multiradiala (Lea); L. luteola (Lamarck); L. liga-
mentina (Lamarck) (PI. LXXXIX, fig. 16) ; L. orbiculata (Hildreth) (PI. LXXXIX,
fig. 22); L. recta (Lamarck) (PI. LXXXIX, fig. 21); L. nasuta (Say); L. iris (Lea)
(PL LXXXIX, fig. 20); L. gracilis (Barnes) (PI. LXXXIX, fig. 19); Obovaria cir-
culus (Lea) (PI. LXXIXX, fig. 15); Plagiola securis (Lea) (PI. LXXXIX, fig. 17);
Ptychobranchus phaseolus (Hildreth) (PI. LXXXIX, fig. 14); Unio crassidens La-
marck (PI. LXXXIX, fig. 6); Unio gibbosus Barnes (PL LXXXIX, fig. 7); Pleuro-
bema clava (Lamarck) (PL LXXXIX, fig. 5); Quadrula undulata (Barnes) (PL
LXXXIX, fig. 3) ; Q. coccinea (Conrad) (PL LXXXIX, fig. 4) ; Q. rubiginosa (Lea)
(PL LXXXIX, fig. 2) ; Q. subrotunda (Lea) ; Q. kirtlandiana (Lea) (PL LXXXIX,
fig. 1).

Other Pennsylvanian species, the glochidia of which are known to be of this
type (See Lea, Observ. VII and X), but have not been seen by myself, are: Lamp-
silis radiata (Gmelin), L. parva (Barnes), and Obovaria retusa (Lamarck).

3. The breeding season.
Lea (Observ., II, III, VII, X) had already paid attention to the time of
breeding of the Unionidce. Sterki (1898 and 1903) has pointed out that there are

8See footnote on page 300.

ortmann: a monograph of THE NAJADES OF PENNSYLVANIA 303

important differences between the different forms with regard to the time, when
the marsupium is filled, and the present writer has devoted a special article to
this topic as far as it concerns the Pcnnsylvanian species (Nautilus, 1909), which
confirms and supplements Sterki's observations. Additional facts have been fur-
nished by Conner (1909, p. 112).

Since the publication of my paper in 1909, I have made a large number of
further observations, which generally serve to support my previous conclusions,
and the microscopical investigation of the contents of the marsupia of soft parts
at hand has added a number of more important facts. These new observations
will be reserved for the systematic part, and will be treated under each species.
Here I shall give only a condensed statement of the results obtained by Sterki
and myself, and a few considerations as to their systematic value.

Among our Najades, we have in our latitude two types, summer-breeders and
winter-breeders. Those belonging to the first class, have the marsupium filled only
in summer, that is to say, in the period between the end of April and the beginning
of August. Of course this indicates the extremes; for in the case of single species
and single individuals the breeding season is much shorter; and the month of
June, and the beginning of July is generally the time when most of these forms are
gravid. The development from the egg to the glochidium takes place within a
very short time (a few weeks), and, after the glochidia are formed, they are dis-
charged immediately. The consequence is that individuals with glochidia are
rather rarely found.

It is a remarkable fact, first pointed out by Sterki, that summer-breeders
are found only in a restricted group of genera; and, as we shall see, this physio-
logical character is always found in the forms belonging to this group. It is com-
posed of the genera: Quadrula, Pleurobema, and Unio of Simpson, and, as majr
be seen from the description of their soft parts given above, these, according to
their morphological characters, form a group by themselves. Likewise the genus
Margaritana, which stands by itself in the system, seems to be a summer-breeder;
at all events, Conner (1909, p. 112) found gravid females in June and August,
but not in November and January, and I failed to find them in the beginning of
May, the beginning of June, and on August 4. (At the latter date only one speci-
men was taken.) In Europe, Margaritana is known to breed in July and August
(See Harms, 1907, p. 818).

All other genera are winter-breeders. In these the marsupia fill some time
in summer, from July to September. The development of the eggs also takes
place rapidly, and generally the glochidia are fully developed in September or

304 MEMOIRS OF THE CARNEGIE MUSEUM

October. However, they are not, as a rule, then discharged, but are carried through
the winter in the marsupium, and are not set free till warmer weather sets in again,
in April, May, or June of the next year.

Among the winter-breeders we have in some cases the breeding season still
more extended, so that in summer (June, July) the end of one season overlaps
with the beginning of the next (of course, not in the same individual), and in
such forms gravid females may be found all the year round.

In order to understand these conditions, we must inquire what are their es-
sential features. Sterki used the terms summer- and winter-breeders, and he cer-
tainly was right so far as it concerns the species in our latitude.9 But in my
opinion, the most important difference of the two groups is, that in the one case
the breeding season is short, and the glochidia are discharged immediately, while
in the other case it is long, and the glochidia are kept in the marsupium for a long
time. This would be well expressed if we call the summer-breeders tachytictic
forms, and the winter-breeders bradytictic forms.10 It seems to me, that the former
condition is the primitive and original one, while the latter is apparently a special
adaptation to the climatic conditions of higher latitudes. From this point of view
we gain a rather satisfactory interpretation of the raison d'etre of the two types,
and also ascertain a very significant correlation between these physiological char-
acters and certain morphological features.

If a short breeding season is the more primitive condition, we can understand
perfectly well, that those Najades, which are tachytictic, have also a comparatively
primitive and simple morphological structure in their marsupium. This is actually
the case in Margaritana as well as in Quadrida, Pleurobema, and Unio. On the
other hand, in the forms with a long breeding season we should expect morpho-
logical adaptations conforming to it. And this is actually the case. In the Ano-
donta-group we have seen, that there exists a very complex structure of the mar-
supium, the purpose of which apparent^ is, to furnish a proper water circulation
for the embryos enclosed in the marsupium (See above, p. 293), and in all members
of this group the embryos stay a long time in the marsupium. In the members of
the Lampsilis-group, which also are bradytictic, the ^Inodcmta-structure is not pres-
ent. But we have become acquainted above with a peculiar bulging out of the
marsupium, which apparently serves the same purpose, namely to bring the
glochidia into closer contact with the water, since the membrane which covers
this outbulging marsupium is very thin. But in addition, I shall describe below

9I know cases of exotic forms, which behave like summer-breeders, but do not breed in "summer."
10From rax^s quick, /SpaSiJs slow, and ruw«6i breeding.

ORTMANN: A MONOGRAPH OF THE NAJADES OF PENNSYLVANIA 305

certain other structures (flaps and papillae of the edge of the mantle), which most
emphatically serve for the aeration of the marsupium, and thus are to be regarded
as adaptations to a long breeding season.

Having thus correlated physiological function with anatomical and morpho-
logical structures, we may rest assured, that we have discovered an essential prin-
ciple in the development of the Najades, and we may say with all confidence that
a systematic arrangement, which is founded upon such structures, which we are
able to understand, must be the correct one. Sterki's discovery of summer- and
winter-breeders, although originally not fully understood, thus becomes the most
important step in our knowledge of the system of the Najades.

4. The discharge of the Glochidia.

As may be inferred from the above, the discharging of the glochidia must
take place at two different seasons: in the bradytictic forms in spring or early
summer (April to June),11 in the tachytictic forms from the end of June to August,
While this fact is beyond doubt, the manner of discharge remains yet to be in-
vestigated, and for some time this proved a serious puzzle to me. But now I
am satisfied that my observations are complete enough to answer the question.

Lea repeatedly observed the discharge of the glochidia and placentae, but he
only mentions that they come out of the shell at the posterior end (in Lampsilis
radiata, L. parva, Strophitus edentulus (Observ. X), and Unio complanatus (Observ.
II), and he further quotes (Observ. Ill) a similar fact as observed by Kirtland in
Quadrula cylindrica. As we shall see, these observations are probably correct,
and represent the normal conditions.

On the other hand, Lea (Observ. II, p. 52) reports that in a gravid female of
Anodonta fluviatilis (= cataracta) he made the young come "from the orifice at
the inferior part" (= edge) of the marsupium, by the pressure of his fingers upon
the sides of the marsupium. I consider this not normal, but an artificial effect,
in fact, there are no such "orifices."

Other authors do not treat of the discharge of the glochidia, except that
Simpson (1900, p. 616) incidentally expresses the opinion, that in Strophitus the
placentulse are discharged through the walls of the gills, which, as I believe, is
entirely wrong.

"Sometimes it is postponed to the beginning of July, chiefly so in Lake Erie. Lea lias repeatedly and in various
species observed discharges during winter. These always took place, however, in the case of specimens kept in captivity
under artificial conditions, which probably induced an abnormally early discharge. Since the glochidia of bradytictic forms
are fully developed in the beginning of the winter, artificial conditions (temperature for instance) might start the discharge
at any time during the winter.

306 MEMOIRS OF THE CARNEGIE MUSEUM

Considering the anatomical structure of the gills, and the fact, that the embryos
are contained in the water-tubes, we should expect, a priori, that the discharge
would take place by the natural channels. This would be by the way of the
suprabranchial canals into the cloacal chamber, and thence outward through the
anal opening of the mantle. And indeed, I have collected evidence for this natural
discharge in a number of cases, which furnish positive proof of it. They are the
following: In cutting up specimens of Quadrula subrotunda, on June 22 and 24,
1909, I repeatedly observed placentse coming out of the anal opening. Of
course, this might be due to rough handling, but I have never seen them coming
out at any other place, unless I forced them out by so strong a pressure, that
the walls of the gills burst open. On June 24, 1909, I collected a good number
of specimens of Quadrida subrotunda and coccinea, which I took home alive,
and kept over night in a bucket with water. The next morning I saw quite
a number of placentse floating in the water and was able to distinguish easily those
belonging to the first species, which are subcylindrical and pink, from those belong-
ing to the second, which are flattened (leaf-shaped or lanceolate) and white.
On June 23, 1910, I took home a number of gravid Quadrida coccinea, and
watched the discharge of the placentse. They distinctly came out of the anal
opening. On June 24, I opened one of these specimens, and saw the placentse
slowly moving out from the water-tubes through the suprabranchial canal to the
anal siphon. While I was holding this specimen in my hand (with the shell opened
and the left half of the mantle thrown back), four placentse moved out. The
coming out was not by a jerk or a squirt, but was slow and gradual, as if carried
by a steady current of water. A second specimen, opened shortly after this,
exhibited the same phenomenon. On July 25, 1910, two specimens of Pleurobema
wsopus were found in the act of discharging placentse. One was seen in the water,
with a number of pink placentse lying behind the posterior end of the shell, which
were greedily devoured by a number of minnows. The other was taken home, and
the discharge was observed at leisure in a basin of water. The pink placentse came
slowly out of the anal opening, often in pairs, and sometimes, when the shell was
squirting water out of the anal opening, a number of placentse were thrown out.
Among my alcoholic material, there are several individuals, which were pre-
served in the act of discharging glochidia in the natural way. The best specimen
is an Anodontoides ferussacianus, collected May 14, 1908, which has the posterior
half of either marsupium empty (discharged), while the anterior half is yet full.
Here I observe, that the suprabranchial canals of the marsupial gills are crowded
with loose glochidia back to the cloacal cavity, and a mass of glochidia is located

ortmann: a monograph of THE NAJADES OF PENNSYLVANIA 307

in the anal opening. The dark brown color of the glochidia contrasts strongly
with the whitish, semitransparent tissue at the base of the gills, and thus the supra-
branchial canals are easily traced by this color. The same phenomenon was seen
in a specimen of Quadrula undulata, collected July 8, 1909, where also numerous
loose glochidia, but of whitish color, were found in the suprabranchial canals of all
four gills. On May 13, 1910, a specimen of Symphynota costata was found, also in
the act of discharging loose, brown glochidia through the suprabranchial canals.
The glochidia were not plentiful in the canals. This specimen was not preserved.
An additional case, corresponding closely to the instances just mentioned, was
observed on July 12, 1910, in Anodonta imbecillis. Finally, I have observed in
alcoholic material of five specimens of Strophitus edentulus collected April 22 and
24, 1908, a similar condition. But here there were whole placentuhe as well as free
glochidia in the outer suprabranchial canals, and in no case were they very
numerous.

Entirely different from these are the following observations :
In a gravid female of Ptychobranchus phaseolus (collected August 31, 190G;
probably unusually delayed in the discharge), there were on the margin of the
right marsupium about half a dozen placenta? protruding through and beyond the
edge, that is to say, in the act of escaping from the marsupium through openings
in the outbulging edge of the gill. This peculiar fact suggested to me the idea,
that the discharge of the glochidia might take place here not by the natural channels
and the anal opening, but by breaking through the edge of the marsupium into
the branchial cavity, and thence through the branchial opening into the surrounding
water. I consequently made a careful search among my material for additional
specimens caught in the same act, and I found indeed such evidence for this "un-
natural" discharge. What I was able to ascertain is that there are actually, at
the time of the discharge, holes breaking through the edge of the marsupium, which
thus furnish an outlet for the glochidia. These holes may be seen macroscopically,
and are also shown quite satisfactorily in some of my slides. The best slides I
have are made from Lampsilis luteola and multiradiata (See PI. LXXXVIII, figs.
18 and 20). But I also possess such from Lampsilis nasuta and Obovaria circulus,
which show this feature, although less distinctly. I remember further, to have
noticed macroscopically these holes in Lampsilis alata, but I did not succeed in
obtaining a satisfactory slide, and I have a slide of Lampsilis gracilis (See PI.
LXXXVIII, fig. 10), which shows, as I believe, the closing up of these holes after
the discharge of the contents of the ovisacs.

In 1910 I found similar instances. A specimen of Ptychobranchus phaseolus,

308 MEMOIRS OF THE CARNEGIE MUSEUM

collected June 27, had most of the placentae discharged, but of the remaining
few, some were protruding from the edge of the gill, exactly as described above.
Females, discharging glochidia, and with openings in the edge of the marsupium,
were seen in the following additional cases: Lampsilis luteola, May 6 and July
7 (on the latter date in several individuals); Lampsilis ventricosa, May 21 and
June 23; Lampsilis multiradiata, June 23.

To all appearance, these holes are present only during the actual discharge
of the placentae or the glochidia, and immediately after it, and close up again
very soon, so that it is difficult in sectioning, to strike just the right point of the
edge of the gill, where they are best visible, and this may account for my failure
to obtain a greater number of good slides. It also should be remembered, that
gravid females in the act of discharging, are naturally scarce among material, which
was not collected with this question in view, and it is indeed astonishing, that I
found so many instances supporting the assumption, that in the Lampsilis-group,
that is to say, in those forms, in which the edge of the marsupium bulges out
beyond the original marginal line of the gill, the glochidia are discharged, not by
way of the natural outlets, but by breaking through the marginal wall of the gill.
If this is so, a very important correlation is suggested : the anatomical peculiarities
of the edge of the marsupium in these forms appear as a direct adaptation to the
way of discharging the glochidia: the edge of the marsupium is built for this purpose.

Nevertheless this phenomenon should be further studied. I have given here
my observations, and I think they are correct and conclusive, for I cannot imagine
that conditions as represented in the figures 18 and 20 on Plate LXXXVIII are
wholly artificial and abnormal.

From the above instances we see, as well as from some of the older observations
of Lea, that in some cases the whole placentae are discharged, in others free glochi-
dia. This is connected with the greater or lesser solidity of the placentae (See
above, p. 298), but its systematic value is not very great; at any rate, each one of
the greater groups contains examples of either method of discharge.

D. Hermaphroditism.

Lea and others have demonstrated that most of our Unionidce are gonochorists,
and the existence of hermaphrodites remained doubtful, till Sterki found them.
In a number of cases, hermaphroditism is occasional and exceptional, according
to Sterki (1898, p. 30), and he names the following instances: Quadrula rubiginosa,
Q. pryamidata, Lampsilis parva. But in another case, that of Anodonta imbecillis
(Nautilus, 12, 1898, p. 87), he found that hermaphroditism is the normal condition.

ortmann: a monograph of THE NAJADES OF PENNSYLVANIA 309

Sterki decided this question by investigating the sexual glands. Since I have
found that the sex of any individual may be determined by an examination of the
gills, it should be possible, to recognize hermaphroditism by the structure of the
latter, for in the case of hermaphroditism no individual should be found exhibiting
the male structure of the gills.

This is indeed true, and I am not only able to confirm Sterki's observation
with reference to Anodonta imbecillis, but in addition I think I have discovered
among the Pennsylvanian species in Symphynota compressa and S. viridis, two
other cases in which hermaphroditism is at least the normal condition. The par-
ticulars are as follows:

Of Anodonta imbecillis, until toward the end of the summer of 1909, I found
only 6 specimens (in May and beginning of June) , every one gravid. On September
14, 1909, I collected a large number (about thirty) of this species in the outlet of
Lake Leboeuf, which again were all gravid with one exception, and the latter
proved, upon anatomical examination, to be also an individual with the female
structure of the gills. At that time, I was aware of the existence of this question,
and was looking out for males, but was unable to find any. In 1910 I collected
only a single individual, but again a gravid specimen.

In the case of Symphynota compressa my observations are not so positive, and
only during 1909 and 1910 did I pay attention to this species with this point in
view. I repeatedly collected specimens, which were not gravid, in May, June,
July, and the beginning of August; but these might have been sterile females as
well as males. I also collected four specimens, not gravid, on October 10, 1907
(together with gravid specimens). Of these I preserved two, believing them to be
males; but upon investigation they proved to possess the female type of gills.
During the month of September of 1908 I collected about twenty-five specimens,
and among them there was not a single one which was not gravid. On August 18,
1909, I found about twenty specimens at one locality, and on September 27, 1909,
about fifteen at another place. On these occasions I was hunting for males, but
was unable to find any. On the latter date I found four very young specimens,
which were not gravid. Three of these were preserved, but they also had the
female gill-structure. In 1910 the result was the same. Of three specimens found
every one had the female structure, or was gravid. However, from Professor
Charles Brookover in Akron, Ohio, I received a specimen of £. compressa coming
from Lake Erie, at Cedar Point, Ohio, which had the male anatomy of the gills.
This is indeed very remarkable in view of the fact, that among all the specimens
collected in Pennsylvania and preserved in alcohol or examined on the spot, there
was not a single male, and that I was altogether unable to find a male in our state.

310 MEMOIRS OF THE CARNEGIE MUSEUM

Of the hermaphroditic nature of Symphynota viridis I feel more confident.
I received thirty-four specimens of this species, collected by Dr. Atkinson on July
11, 1908, and preserved the soft parts of fifteen of them. None was gravid, but
they all had the female gills. In August, 1908, April, August, and September, 1909,
I found this species repeatedly, altogether from fifty to sixty individuals, but all
proved to be gravid, unless they were very young, and even in a number of the
latter (in the second and third year of age) the female gill-structure was recog-
nizable. In August, 1910, I collected altogether over one hundred specimens, and
examined them in the field one by one, and every one had the female gill-structure.
Thus in this case also I hunted in vain for males.

It would be a very strange coincidence, that I should fail to find males only
in these two closely allied species, if they exist. In all other species of Unionidce
I never had any trouble in discovering males, provided there were more at hand
than a few. In fact, it seems that the males generally slightly outnumber the
females. Nevertheless I could not find them in the above cases, and even those
sterile specimens, which I took home as males, turned out to have the female
structure.

I therefore think that the conclusion is justified, that in these two species of
Symphynota no males exist; and that they are, as a rule, hermaphrodites; although
occasionally males may turn up, as is shown by the specimen from Cedar
Point.

I should also add that the Carnegie Museum possesses from a branch of the
Rio Grande at Mercedes, Hidalgo County, Texas, an Anodonta, belonging to the
imbecillis-group, which probably is A. henryana Lea. There are seven specimens,
and all were collected by Dr. Atkinson on May 15, 1907, and every one of them is
gravid. Of course, the number is too small to express a final opinion, yet, according
to my experience with other species, even among seven individuals we should
expect to find both sexes. Thus in this case also it is suggested, that we have to
deal with a hermaphrodite, which is also probable on account of the affinity of
this species with Anodonta imbecillis.

I may possibly be able to give the results of further investigations of this
question in the future.

E. Sex recognition in the young.
Sterki (1898, p. 29) has pointed out, that in young individuals, from two to
four years old, the sexual glands are not developed, and that in those of the Lamp-
silis-type the sexual differences of the shells and the gills are not noticeable; there-

ORTMANN: A MONOGRAPH OF THE NAJADES OF PENNSYLVANIA 311

after, the marsupium begins to form, and with increasing age the number of ovisacs
increases.

Since I have shown, that there are in the females of all species permanent
differentiations of the marsupial gills, the question arises, whether these differences
are observable at an early age, and I am able to report the following cases:

The smallest specimen of any species I ever investigated, in which the sex
was recognizable, was a specimen of Strophitus edentulus, 17 mm. long. Here
(See PL LXXXVI, fig. 10), in a cross-section, the septa of the outer gills were
decidedly more crowded and more solid than those of the inner gills. However,
the peculiar folding of the epithelium of the water-tubes (as shown in older speci-
mens, see PL LXXXVI, fig. 76) was not yet developed. This specimen upon the
outer surface of the shell exhibited one growth-rest; assuming that this was formed
in the first winter, this specimen would have been in its second year.12

Additional evidence, that in some species already in the second year the female
structure of the gills begins to develop, is furnished by the following cases.

Anodonta grandis gigantea Lea. Here I found a young individual, 46 mm.
long, but with only one growth-line, distinctly possessing the female structure
(PL LXXXVI, fig. 14).

Symphynota compressa Lea. A specimen, 44 mm. long, with only one growth-
rest had the female structure fully developed.

Symphynota viridis (Conrad). A number of young ones, of various sizes (18
to 31 mm. long), showed distinct traces of the female structure, and one of them
(31 mm.) was even gravid (PL LXXXVI, fig. 11).

Besides these cases, I did not find the female type of gills in the second year
in any other species, although I examined specimens of Quadrula undulata hip-
pop&a, Unio complanatus, Lampsilis ligamentina, L. multiradiata, L. ventricosa,
and L. ovata.

That specimens in the third year have the female structure, is shown in a
specimen of Symphynota complanata, 35 mm. long, with two growth-lines. Here
the difference of the septa of the outer gill became very striking, when compared
with a male, 54 mm. long, with three growth-lines.

12I am aware of tin' fact that the value of growth-rests for the determination of agfl hai 1 " disputed. It is, however,

beyond doubt that during each winter a growth-line is formed. It i- also established, that growth-lines may form at any
other time, whenever unfavorable conditions prevail. But the latter arc undoubtedly abnormal and irregular. In young
specimens, we may safely assume, that the first few growth-rests, chiefly when they are regular, correspond to the stoppage
of growth in winter. In older ones, it is more difficult to recognize the annual growth-lines, and to distinguish them from
the accidental lines. Yet in well developed specimens they may be told apart. It may incidentally be mentioned here, that
in shells from Lake Eric the growth-lines arc often extremely regular, simply because in the lake accidental disturbance of
the summer growth is very rare, and thus only annual lines are developed.

312 MEMOIRS OF THE CARNEGIE MUSEUM

In Quadrula undulata hippopwa, I found a specimen distinctly of the female
type, which was 36 mm. long, and had two growth-lines; here again the female
structure became very evident, after a comparison was made with a male specimen,
34 mm. long, also possessing two growth-lines.

I have an individual of Lampsilis nasuta, 38 mm. long, with two growth-lines,
which is gravid.

That absolute size is no proper criterion in this question, is shown, for instance,
in the case of Lampsilis iris, of which I have a specimen 33 mm. long (smaller
than in the instance just mentioned) which is gravid, but has at least three distinct
growth-lines, and consequently would be at least in its fourth year.

All other females observed, sterile or gravid, had at least three growth-rests,
and should be considered as being in their fourth year of age, or older.

Thus it becomes apparent, that the female structure, at least in some species,
develops as early as the second year, and that one at least (Symphynota viridis)
may already become sexually mature at this time. In others, the female structure
is distinctly seen in the third year, and some of them (Lampsilis nasuta) may
bring forth progeny. Later, from the fourth year onward, the time of maturity
is reached in most of the species.

F. Differentiating structures of the edge of the mantle.

Simpson in the diagnosis of a number of genera states that the edge of the
mantle is "double." According to my observations, it is double along its whole
extent in all forms. While the outer edge is closely adjacent to the margin of the
hard shell, and is always simple, the inner edge, which is generally parallel and
close to the outer edge, is more or less expansile, and is able to fill and close the
gap between the two halves of the shell, when the adductor muscles relax, that is
to say, when the shell "opens." And further, this inner edge of the mantle is not
uniform in its structure all along its margin, but shows certain differentiating
structures, some of which are systematically valuable.

1. The anal and branchial openings ("siphons").
The part of the edge of the mantle at and near the posterior end of the shell
is modified so as to form what is known as the anal and branchial openings, the
former lying more dorsally, and immediately above the latter. They are formed
by the mutual approach of the edges of the inner mantle at certain points, and
sometimes by their coalescence. The two openings are separated from one another
by the diaphragm, discussed above (See p. 288), composed of the posterior ends of

ortmann: a monograph of THE NAJADES OF PENNSYLVANIA 313

the gills, which serve to draw the mantle together, although, in our species, the
two sides of the latter never unite at this point (Compare fig. 4, on p. 289). The
part above this point is the anal opening, which forms the outlet for the supra-
branchial canals and the cloacal cavity (with the anus). Its upper (dorsal) bound-
ary will be discussed below (see supra-anal opening).

Below the diaphragm is the branchial opening, which leads into the branchial
chamber, into which the gills hang down. Its lower or anterior boundary is
marked by slightly projecting lobes of the inner edge of the mantle (often indicated
only by a curve reaching from posterior to the lower margin), which approach
each other, and beyond which the edges of the inner mantle touch each other, but
without growing together.

Anal and branchial openings generally possess on the inner edge peculiar crenu-
lations, teeth, or papillse, which were already studied by Lea, who took special
pains to describe them accurately, believing that they might be of systematic
value. He lays great stress upon the presence or absence of papillae on the inner
edge of the anal siphon. But according to my observations, this is not an important
feature. There may be papillse, which may be larger or smaller (always smaller
than those of the branchial opening), and they pass gradually into fine teeth,
scallops, or crenulations; or the edge may be practically smooth. Sometimes the
teeth, or crenulations, appear different according to the state of preservation.

In most cases crenulations (wrinkles or small folds), fine teeth, or very fine
papillse are present. A smooth or almost smooth edge is found principally in the
following forms: Quadrula metanevra, Q. cylindrica, Q. tuberculata (Rafinesque),
Anodonta grandis, and A. imbecillis. In Anodonta cataracta I found crenulations
of very variable character. On the other hand more distinct papillse appear in
Pleurobema clava, Unio crassidens, U. gibbosus, U. complanatus, U. productus, and
A nodontoides ferussacianus.

The branchial opening always has papillse, and they offer, although slightly
varying in their development, no remarkable differentiations. They may be simpler
or more complex (arborescent) , and may be larger or smaller.

2. The supra-anal opening.

In most of our species, the upper part of the anal opening is separated from the
anal by a complete union of the inner edges of the mantle, and is called the supra-
anal opening (See fig. 4, on p. 289). This opening communicates under the united
edges of the mantle with the anal opening, and its purpose is rather obscure. I
am inclined to believe, that the connection of the edges of the two halves of the

314 MEMOIRS OF THE CARNEGIE MUSEUM

mantle is the essential feature of this structure, which serves to close and define
the anal siphon dorsally, while the upper opening (supra-anal), which remains
unclosed, is only incidental to this, so to speak a by-product.13

There are a few forms, in which no supra-anal opening is present, that is to
say, where the edges of the mantle do not at all coalesce. In our state, we have
only two such forms, Margaritana margaritifera (See fig. 6, p. 289) and Quadrula
(Rotundaria) tuberculata, and in each of them, we have a large, single, undivided
anal opening. This feature was known previously, and in my opinion is rather
significant, as showing the preservation of a primitive condition.

In all other forms, there is a connection of the margins of the mantle separating
a supra-anal from the anal opening. Yet in a number of species this connection
is very short, and sometimes deciduous; or it may be torn easily, or may be oc-
casionally altogether absent. I have observed specimens with this connection
missing in the following species: Quadrula subrotunda, Q. rubiginosa, Q. coccinea, Q.
obliqua, Q. undulata, Pleurobema ossopus, Unio crassidens, and U. gibbosus. But
on the whole, such specimens are rare and in all of these species I have found
other individuals with this connection present. In some cases it seemed that this
connection was naturally lacking, while in others it had been apparently torn
apart by rough handling, which may happen easily, since in all these cases the
connection is very short.

In other instances the connection is a little longer, but it is generally situated
so that the supra-anal is a good deal longer than the connection, and than the
anal opening. This condition of a rather short connection is found in all species
not mentioned here as possessing a different conformation.

In certain forms the connection of the mantle becomes somewhat longer, so
that it is about as long as the anal opening, and also about as long as the supra-
anal. I have observed this in Alasmidonta undulata, Symphynota complanata, and
Anodontoides ferussacianus.

In still other forms the connection becomes rather long, restricting the supra-
anal a good deal, so that the latter is distinctly shorter than the connection, and
hardly longer than the anal opening. These species are Anodonta grandis (fig. 4,
p. 289), Anodonta cataracta, and Anodonta imbecillis.

Finally, in one case, that of Lanipsilis parva, I found the supra-anal opening
entirely closed. Of this I have only three specimens, and in one I think I have
seen a very small supra-anal opening, while in the other two the inner edges of the

,3In rare and individual cases, there is a double connection. Lea mentions a few such instances, and I have observed
the same thing in Obovaria retusa. I consider this an abnormality.

ORTMANN: A MONOGRAPH OF THE NAJADES OF PENNSYLVANIA 315

mantle were entirely connected as far as the posterior dorsal end. Of course the
outer edges of the mantle remain separated, and if only superficially investigated,
the appearance of a slit-like supra-anal, identical to that of other forms, is imitated.

3. The edge of the mantle in front of the branchial opening.

Anterior to the branchial opening the margin of the mantle is originally with-
out any special differentiation. The two edges are about parallel and close to-
gether to the anterior end of the animal, and do not possess any appendages in
the shape of papilla? and the like. At most the inner edge is slightly crenulated,
scalloped, or wrinkled, and the latter appearance often depends on the state of
preservation (it becomes wrinkled by strong contraction in alcohol). The com-
monest condition is when the papillae of the branchial opening suddenly become
small, assume the shape of crenulations, which soon disappear, and thenceforward
the inner edge appears more or less smooth. The margin of the mantle (the part
under both edges), which is thickened at the branchial opening, thins out gradually,
and the thickening is hardly any longer noticeable at a short distance in front of
the branchial opening.

This condition is seen in all species investigated belonging to the following
genera of Simpson: Quadrula, Pleurobema, Unio, Tritogonia, Alasmidonta, Stro-
phitus, Symphynota, Anodontoides, Anodonta, and Ptychobranchus.

In Margaritana Lea (Observ. VII, PI. 29, fig. 104) has figured strong, papillae-
like crenulations on the edge of the mantle for a considerable distance in front of
the anal opening. He does not give a detailed description of this structure in
the text. In the specimens in my possession I find that the inner edge of the
mantle in this region is strongly wrinkled and crenulated, the crenulations passing
posteriorly into strong papillae, such as are usually found at the branchial opening.
I do not regard this as a special differentiation of structure, but it rather appears
to me to be a forward continuation of the papillae of the branchial opening; so
that we should in this case regard the branchial opening as less distinctly defined
anteriorly. In fact I cannot discern any sharp demarcation of the latter, such
as is seen in other species. The branchial opening in Margaritana is ill defined,
and passes gradually into the anterior margin of the mantle, a condition which
again emphasizes the primitive character of this genus.

There remain a number of genera, which are characterized by the peculiar
structure of the marsupium, and in which it is restricted to the posterior part of
the outer gills, and projects beyond the margin of the latter. These are the genera
Obliquaria, Plagiola, Obovaria, Lampsilis, Micromya, and Tnmcilla. Among these
we meet with interesting special structures on the edge of the mantle.

316 MEMOIRS OF THE CARNEGIE MUSEUM

As has been said, Ptychobranchus, which is also associated with these genera,
has the normal structure. The same holds true of Obliquaria, Plagiola, Obovaria,
and certain species of Simpson's Lampsilis, namely: L. ligamentina, alata, and
gracilis. But the beginning of a differentiation is here indicated. In all these
the thickening of the margin of the mantle is more distinct, and extends farther
forwards, sometimes as far as the middle of the ventral margin, where it disappears
rather suddenly. Along this thickening we observe the inner edge of the mantle
as a distinct keel, which is more or less crenulated or toothed. These teeth, which
are the continuation of the papillae of the branchial opening, are also sometimes
found for a short distance in normal cases, as described above; in the present
genera, however, they extend farther forward, generally as far as the thickening
of the margin of the mantle, and disappear beyond, or give way to mere crenula-
tions. But in no case do these teeth assume the shape of papillae, or appear as the main
feature of this structure. They are rather elements accessory to the edge of the
mantle itself, which is a narrow keel or lamella, projecting from the thickened
margin of the mantle. This lamellar, projecting character is most evident in
Lampsilis alata and gracilis.

This structure of the edge of the mantle is present in both male and female,
but it is generally more distinct in the female, chiefly so far as it refers to
the thickening of the margin of the mantle and the size of the crenulations or
teeth.

While we may say that in these forms the structure of the margin of the mantle
is practically normal, except that its thickening, and the crenulations of the inner
edge are more emphasized; we find in the remaining species of Lampsilis, and in
Micromya and Truncilla, other and more strongly pronounced differentiations, and
there are two types observable within the genus Lampsilis (including Micromya),
while in Truncilla peculiar conditions prevail.

The first type is found in the following species: Lampsilis parva, iris, nasuta,
recta, and Micromya fabalis. In these the thickening of the margin of the mantle
is rather pronounced, and along it runs the inner lamelliform edge, which, however,
is not simply crenulated or toothed, but possesses distinct, well developed, conical,
more or less regular papilla?, which are rather sharply distinguished from the ir-
regular, crowded, or massed papillae of the branchial opening. These papillae run
forward in various extent, and disappear rather suddenly; generally with the dis-
appearance of the thickening of the margin, and give way to the normal crenulations
or the normal smoothness of the anterior edge. In the different species, the
papillae show the following peculiarities:

ORTMANN: A MONOGRAPH OF THE NAJADES OF PENNSYLVANIA 317

Lampsilis parva. In this species, Lea already had described and figured (Ob-
serv. VII, PL 29, fig. 102), what he called a "caruncle" on the edge of the mantle,
just in front of the branchial opening. This is according to him a black, spongy
mass, which, being quite expansible and retractile, may assume various shapes.
In the three gravid specimens before me, I see a group of black, projecting papilla;
standing upon the marginal thickening of the mantle, which extends about as far
as the middle of the ventral margin of the mantle. This group of papillae is rather
short, and resembles indeed a "caruncle," yet does not differ in any essential
respect from the structures, which will be described in the other species belonging
here. It is composed of from four to six larger, and some smaller papillae, and in
front of it the inner edge is smooth.

Micromya fabalis. Lea (Observ. X) calls the edge of the mantle below the
branchical opening "fringed," and Simpson (1900, p. 525) says in the generic
diagnosis of Micromya, that it is "fringed," and in the female is "developed into a
flap with a distinct, toothed ridge inside." In the only (sterile) female of Micromya
fabalis I have seen, the structure of the edge of the mantle is similar to that of
Lampsilis parva, only the group of papillae extends farther forward, but not quite
as far as the thickening of the margin of the mantle. The number of papillae
is greater. There are from eight to ten larger ones, and a few smaller ones, and
they are further apart from each other than in L. parva. In the male the same
structure is present, but the thickening is very slightly developed, and the papillae
are decidedly smaller, and occupy a shorter space on the edge.

Lampsilis iris. The mantle of this species (as novi-eboraci) has also been
described by Lea (Observ. X) as "fringed," and he says that there are large, distant
papillae. This is quite correct. I observe in my specimens, that in the female
there are from four to ten (according to age) very large, conical papillae upon the
inner edge of the mantle in front of the branchial opening. There are a few smaller
papillae between the larger ones, and the latter are rather remote from each other,
and are placed at irregular intervals. This row of papillae extends forwards for a
considerable distance, but not quite as far as the middle of the ventral margin.
It stops very suddenly, and beyond this point the inner edge of the mantle is only
finely crenulated or smooth. In the male, the papillae are much smaller.

Lampsilis nasuta. The female agrees with that of L. iris in the presence of
papillae, but the latter are very small, fine, and crowded, forming a rather regular
row, running forward almost to the middle of the ventral margin, where they dis-
appear rather gradually, and pass into the smooth anterior part of the inner edge
of the mantle. In this species the inner edge in front of the branchial opening is

318 MEMOIRS OF THE CARNEGIE MUSEUM

more lamellar, and the papilla? form merely a fringe on this narrow lamella. In
the male of this species, the papillae are extremely small.

Lampsilis recta. The female is like that of L. nasuta, but the papilla? are
much larger (sometimes with two tips), rather regular, and form a crowded, very
distinct row, which stops quite suddenly at about the middle of the ventral margin.
The largest papilla? stand just at the point where they stop, but the difference in
size is not very great, and the increase in size is slight in the postero-anterior
direction. The strong development of the papillse makes the inner edge of the
mantle, on which they stand, appear as distinctly projecting beyond the irregularly
crenulated or smooth anterior part. In the male, the margin of the mantle is
less thickened, and the papillse are much smaller, yet they are present.

The second type of development of the inner edge of the mantle, found in the
genus Lampsilis, is observed in the following species: L. luteola, radiata, orbiculata,
ventricosa, ovata, multiradiata, and cariosa. Here the specialization is shown not
by a development of papilla?, but the inner edge itself is greatly enlarged. It
forms a rather broad, lamellar keel, which stands upon the thickened margin of the
mantle, possesses great powers of expansion and retraction, and occupies a con-
siderable distance from the branchial opening forwards, as far as to about the
middle of the ventral margin. Here the lamella suddenly stops, but is prolonged
anteriorly into a, free, longer or shorter, ribbon-like flap, which also may be expanded
to a considerable length and again retracted.

This peculiar "flap" has been noticeed by Lea in several species (radiata, luteola,
ventricosa, ovata, multiradiata). He also has observed that there is, in some forms
on the posterior end of the lamellar expansion, right in front of the branchial
opening, a very curious "eye spot" (dark, round spot, upon a light background).

Simpson mentions this character for his section Lampsilis of the genus Lamp-
silis, and says also that in the section Eunjnia it is "sometimes" present. All of
the species of the first section known to me possess it, but not all of the second
one. Apparently Simpson did not attribute much value to this feature, and he
says that it is found only in the female, and during the breeding season. According
to my experience, however, it is a regular character of all species, where it is ob-
served; and is found not only in the female, but also in the male, although in the
latter only in a rudimentary condition. I have been able to make it out in all
males investigated. I consider it a very important systematic character, which
is connected with the aeration of the embryos in the marsupium, and consequently
developed at its best in the female during the breeding season, but it is permanently
present. Probably also in the other species of Lampsilis described above the cor-
responding papilla? have the same physiological function.

ORTMANN: A MONOGRAPH OF THE NAJADES OF PENNSYLVANIA

319

Lea has figured this "flap" of Lampsilis radiata (Observ. II, PL 15, figs. 48
and 49) and of L. ventricosa (Observ. VII, PI. 30, fig. 107), and he calls special
attention in L. radiata to its different shape in different individuals. Indeed in
his two figures (figs. 48 and 49), there is a great difference in so far that in fig. 48
the anterior prolongation of the flap is not seen. But disregarding the possibility
that this figure may be inaccurate,14 the condition shown in fig. 48 may be either
abnormal, or the anterior end of the flap may be
injured, or it may not be fully expanded. (I think
the latter is very likely the case.) Fig. 49 is rather
correct, and also the figure of the flap of L. ventricosa.

I have repeatedly observed this flap in life, and
have seen it in function. I made the best observa-
tions in the case of L. ventricosa and multiradiata.
When the animal is undisturbed, and the posterior
end of the shell is gaping, we notice, in the gravid
female, that the marsupia, which are easily recog-
nized, are pushed outward, so that they project even
a little beyond the shell (already observed by Lea).
They show below and in front of the branchial open-
ing (See fig. 7). The lamellar expansions of the in-
ner edge of the mantle occupy exactly the place, where
the marsupia are visible, and are expanded on both

., i-,ii, •J_-i_ iii c Fia. 7. LampsUu ventricosa (Barnes).

sides of the latter, projecting beyond the margin of , t .

Shell of gravid lemalc, seen Irom above and

the shell and beyond the marsupia, forming a lateral behind, showing: a, anal opening; 6, bran-
wall on each side of the latter. In addition, the free ohiai opening; m, marsupia; mi, mantle;

n r n li 1 i.- and f, mantle flap in front of branchial

flaps are fully expanded, appearing as prolongations

opening.

of the lamellar portions, floating free in the water.

This whole apparatus performs rhythmical, wave-like contractions, beginning at
the posterior end, and ending at the free tips of the flaps. The contractions follow
each other in quick succession (two to three in a second), and must produce a very
lively current of water over the exposed edges of the marsupia. The position of
the shell being generally much tilted, the animal almost "standing upon its head,"
and the shell being directed with its anterior end against the current, the flaps
float horizontally in and with the current, I once (Sept. 15, 1909) had a chance
to depict the flaps in action, standing knee-deep in the water of the Allegheny

"This is suggested by the generally poor character of the figures on this plate, whirl, an extreme^ superficial and un-
satisfactory.

320 MEMOIRS OF THE CARNEGIE MUSEUM.

River, at Warren, Pa. The figure here given (fig. 7) is a reproduction of this
sketch. It represents a shell as seen from above, with its hind portion turned
up. Allowance must be made for the difficult conditions, under which this sketch,
which is rather diagrammatic, was made.

The general shape of the lamella and flap in L. ventricosa, is ribbon-like,
much longer than wide, and the free anterior projection is rather long. It arises
rather suddenly at the posterior end near to and immediately in front of the
branchial opening, forming there an angle, at which the eye spot is situated. The
edge of the ribbon (inner edge of mantle) is, when expanded, rather smooth, but
there are, chiefly toward and at the free anterior lobe, a few irregular teeth, which
give to this lobe a lacerated appearance. These teeth are quite variable. When
contracted, the edge is wrinkled and crenulated, and the free lobe is much shorter,
triangularly pointed or rounded off. Immediately in front of the lobe, the inner
edge has a few small teeth, or it is scalloped, and becomes smooth anteriorly.

This peculiar lamella or flap is also distin-
guished by its color. In L. ventricosa it is gray
on the outside; on the inside it is a beautiful
fig. s. Lampsiiis ventricosa (Barnes), fe- paje orange, with a blackish longitudinal line

male. Right Mantle Flap, seen from inside. /ol c ON „,, , . , ■, . e , ,

(See fig. 8). The thickened margin of the man-
tle between the inner and outer edge is also pale orange, or light brown, with
black mottlings. Together with the eye spot, mentioned above, and with the
whitish marsupium with its black markings on the edge, being visible between
the flaps, the whole aspect of these parts in the living shell in this condition becomes
rather attractive.

In the male the lamella and flap are similar, but smaller in all dimensions, the
colors are less sharply defined, and the free lobe is very small, yet distinctly present.

Practically the same apparatus is seen in L. ovata, multiradiata, and cariosa,
only in the two latter species the teeth of the edge are more numerous, and the
black line is often missing in L. ovata and cariosa. Altogether there is much
variation in the intensity of the color, the light parts shading from deep brown to
light orange. The eye spot is always present in the female, but indistinct in the
male. In alcoholic material, of course, the colors largely fade.

In L. orbiculata the same flap as in L. ventricosa is developed quite typically;
in alcoholic material a distinct black line is seen, and between this line and the edge
the flap is rather dark black. The only difference is that here the teeth along the
edge arc more numerous and more crowded, increasing in size in the postero-
anterior direction. L. orbiculata generally is placed with L. ligamentina in the

ortmann: a monograph of THE NAJADES OF PENNSYLVANIA 321

system, but I think, according to the above observation, that it has no closer
relationship with this species. The latter has no trace of a flap, in fact, it hardly
shows any differentiation on the edge of the mantle. Lampsilis ligamentina is no
Lampsilis at all. It is true, the males of these two species are very similar; but
this is not true of the females. The peculiar shape of the female shell of L. orbiculatu
is quite remarkable, and entirely unlike that of the female L. ligamentina.

Lampsilis lutcola and radiata possess a similar flap, and are practical^ identical
with each other in this respect. Here the edge of the ribbon possesses numerous,
irregular teeth, increasing in size toward the anterior free lobe; and further, there
are distinct teeth on the inner edge of the mantle in front of the free lobe, di-
minishing in size anteriorly till the edge becomes almost smooth. In the male,
the anterior end of the flap is indicated by large teeth, which gradually pass into
the scalloped and smooth anterior edge, and again all parts are smaller in the
male, and less typically developed. On the inside of the flap and lamella there is
a broad band of black color, and there is also an eye spot posteriorly, but this is
hard to see in contracted alcoholic material.

There remains the genus Truncilla to be discussed. In T. triquetra we see
conditions, which somewhat approach the first group of Lampsilis, where there
are papillae in front of the branchial opening. The swelling of the margin of the
mantle is distinct, but does not extend far forward, and the inner edge along the
swelling possesses a number (four to six) of conical papillse, which are not very
large, and are followed farther forward by a number of still smaller papillse di-
minishing in size. In front of these the edge is smooth. But in this species we
see a feature in the female, which, although not very striking, distinctly indicates
the beginning of the most prominent peculiarity of T. perplcxa rangiana, to be
discussed below. The inner edge of the mantle, where it carries the papillse, is
not quite close and parallel to the outer edge, but runs at a certain distance from the
latter, and the space between the two edges has a peculiar spongy structure. This
spongy structure is also developed to a certain degree in the species of Lampsilis,
but it is not so markedly restricted to the posterior part of the margin, just in
front of the branchial opening. In the male of T. triquetra the two edges are
almost parallel and close together, and the papillse of the inner edge are rather
small. While in the female T. triquetra the separation of the inner edge from
the outer is barely indicated, this condition furnishes the most striking feature in
T. perplexa rangiana. Lea already described it (Observ. X), and Simpson (1900, p.
516) mentions it in the generic diagnosis, but without having properly understood
this structure. In the female of T. perplexa rangiana, the inner edge of the niant le is

322 MEMOIRS OF THE CARNEGIE MUSEUM

posteriorly, in front of the branchial opening, widely separated from the outer edge, and
approaches the latter again at about the middle of the ventral margin. In the space
between the two edges, the margin of the mantle is thick, white, and spongy, and of
a peculiar structure. Thus the inner edges cut off in part a separate chamber within
the shell at the posterior lower end of the latter. Along the inner edge in this
region are fine papillae, the largest behind, diminishing anteriorly, and becoming
effaced before the edges come together again. Farther in front, the inner edge
is smooth. In the male of this species, this peculiarity is not found. In this sex
the two edges are almost parallel, and the papillae of the inner are extremely small.
There is no doubt, that this structure also has something to do with the aeration
of the glochidia during the prolonged breeding season. But its particular action
should be further studied.

I have tried to find other features of the soft parts, which might offer dif-
ferentiations available for the general systematic arrangement of the Najades of
our region, but was not successful. Thus, for instance, as Lea has already pointed
out, the palpi show certain differences, but these are rather slight, and of no sys-
tematic value, and it is not necessary to give a detailed description of the latter.

We are now ready, to discuss the general system of our Najades with reference
to the characters described above.

PART II. THE SYSTEM OF THE NORTH AMERICAN NAJADES.
A. Division into Families and Subfamilies.

In the course of the preceding investigation of the soft parts we took occasion
to point out that Margaritana margaritijera is unique among our mussels. In
fact some of the anatomical characters are so unusual, that this species appears
as a stranger in our fauna, differing fundamentally from all other species. The
most striking features are: (1) the absence of water- tubes in the gills; (2) the
peculiar conformation of the diaphragm at the posterior end of the gills. There
is not a single species among those which I have investigated, which shows even
an approach to the structure seen in Margaritana, and the first one of these is
something quite unexpected in so far that it does not conform at all to the type of
gill-structure, as laid down in the text-books.

In addition, the lack of a separate supra-anal opening and the incomplete
demarcation of the anterior boundary of the branchial opening present additional
peculiarities. Further, although I have not seen gravid females, and am un-
acquainted with the marsupium and the larvae, these are known in the European

ORTMANN: A MONOGRAPH OF THE NAJADES OP PENNSYLVANIA 323

Margaritana through the studies of Harms (1907, pp. 817 and 818). The rnarsupium
is formed by all four gills, and the glochidium is apparently of the QwadruZa- type.15
These characters we must regard as expressing a distinctly primitive type of de-
velopment in the case of Margaritana.

All this is sufficient to impress upon Margaritana the stamp of oddity and
singularity, which justifies us in saying that this form stands by itself in the system,
and should not be associated with any other form. This is best expressed by
creating a family for it (Margaritanidce).

Yet there is possibly another species, which may go with it, Margaritana
monodonta (Say). The soft parts of this have been described by Lea (Observ. X),
and as far as the description goes, we may say that it agrees with Margaritana
margaritifera. Lea especially emphasizes the peculiar diaphragm and the absence
of a defined supra-anal opening.

Leaving the genus Margaritana aside and turning our attention to the rest
of our Najades, we find that the most prominent differentiations are found in the
sexual apparatus, and chiefly in the general form and the finer anatomical structure
of the rnarsupium. This part of the anatomy has been used by Simpson for his
system. But his classification only serves in part to bring out the natural affinities.
Summing up what has been discussed above, we are led to distinguish three groups
among our Unionidce according to the structure of the gills.

1. In some the water-tubes of all or of some gills serve, in the gravid female,
as ovisacs (receptacula for the ova and embryos). Their only differentiation is
in the development of the septa separating the ovisacs, which more closely approach
each other, are stronger, and possess a folded epithelium, adapted to the swelling
of the marsupial gill in the breeding season. There are no other modifications,
and the edge of the marsupial gill is not changed, and always remains sharp.
It goes without saying, that this is clearly the most primitive type of the three.
There are no special adaptations to an extended breeding season, and the latter
is short in all these forms, — they are tachytictic. This type is found in Simpson's
genera Unio, Pleurobema, Quadrula, and Tritogonia.

I have given above the essential characters of the soft parts. Of less impor-
tance are the following: marsupial gill little or only moderately swollen in the
breeding season; rnarsupium either formed by all four gills, or only by the outer
gills in their whole length ; supra-anal opening only in rare instances not separated
from the anal, generally well defined by the connection of the margins of the

''Harms (p. 820) describes "rather strong" hooks in the glochidium, but in his figure (p. Ms, tig. 4) there are only
mere traces of them, and they cannot be compared, by any means, with those of the .lnodu;i(<i-glochidium.

324 MEMOIRS OF THE CARNEGIE MUSEUM

mantle, which is rather short, and leaves a supra-anal of considerable length;
branchial opening rather well defined anteriorly; no papillae or special struc-
tures on the edge of the mantle in front of the branchial opening; inner lamina of
inner gills free from the abdominal sac;16 glochidia imbedded in rather well
developed placentae, or placentae poorly developed; glochidia small or of medium
size, semicircular or semielliptical, with rounded ventral margin, and without spines.
The discharge of the glochidia is by the natural channels through the anal opening.
These forms are tachytictic.

With regard to the characters of the shell little can be said, for the shape of
the shell is very variable. Generally the shell in this group is rather heavy, from
more or less rounded to more or less elongate. The colors of the epidermis are
dull or brighter, rarely with bright markings (rays) ; sometimes peculiar, odd shapes
of the shell are met. The sculpture of the beak is variable, concentric, or with indi-
cations of double loops, or pustulous, or even with traces of zig-zag sculpture toward
the disc. Very often the sculpture of the beak is obsolete; hinge always complete,
with well developed teeth, often quite heavy; male and female shell in most cases
absolutely undistinguishable.

Remarks: It is practically impossible to recognize this group by shell char-
acters, although there are certain peculiar types of shell, which are found in this
group alone. Simpson failed entirely to recognize the close affinity of the genera
belonging to this group, and places them in three different divisions of his system,
bringing his Unio and Pleurobema together with the Anodonta-grouY), with which
there is no close relationship whatever, and removing Tritogonia entirely from
them. The genera of Simpson's system belonging here cannot be maintained as
defined by Simpson, and material changes will be necessary as will be shown here-
after.

2. The second group comprises the following genera of Simpson: Strophitus,
Anodonta, Anodontoides, Symphynota, and Alasmidonta.

Here each ovisac of the gravid female is not formed by- a whole water-tube,
but only by a part of it, the middle one, which is separated from two lateral canals
by a folding up of the epithelium of the septa. In addition, the ovisacs are closed
above at the base of the marsupial gill, thus forming a completely closed sac within
each water-tube. In one case (Strophitus) this sac is again divided into secondary
compartments. The edge of the marsupial gill becomes adapted to the swelling
of the gill in the breeding season, and gives way, the two laminae drawing apart;

16This holds good only for the Pennsylvanian species examined by me. In exotic forms belonging to this group, the
inner lamina may be connected with the abdominal sac (See Nautilus, 23, April, 1910, p. 139).

ortmann: a monograph^of THE NAJADES OF PENNSYLVANIA 325

and there is heavy tissue at the edge capable of being stretched out. This causes
a rounded or truncated edge in the charged marsupium. This peculiar structure
of the marsupial gill is developed only in the gravid female, and is absent in the
sterile female.

These characters are apparently connected with the prolonged breeding season,
and the peculiar secondary water-tubes serve for the aeration of the embryos in
the marsupium. The stretching of the edge of the gill is easily understood.

Other characters are the following: marsupial gills always extremely swollen
and marsupium always formed by the outer gills alone in their whole length; supra-
anal opening well separated from the anal; the separating connection of the margins
of the mantle well developed, sometimes short, but normally somewhat longer,
and even quite long, thus leaving only a short slit for the supra-anal; branchial
opening well defined anteriorly; no papillae or other special structures on the edge
of the mantle in front of the branchial opening; inner lamina of inner gills free
from the abdominal sac, more rarely more or less connected with it; glochidia
very remarkable, and furnishing a very important character of this group. They
are generally very large, subtriangular, in rare instances almost semicircular in
shape, and are more or less pointed in the middle of the ventral margin, where
they possess a peculiar spine or hook (this spine is undeveloped in very young
glochidia). The glochidia are sometimes imbedded in well developed placentae
or placentulae. Sometimes the placentae are poorly developed, or the glochidia
appear completely isolated. The discharge is by the natural channels and the anal
opening, and these forms are bradytictic.

The essential characters of this group are found in the shape and structure
of the marsupium, which represent adaptations to the extreme swelling of the
latter, and to the long breeding season, for all forms belonging here are bradytictic.
The formation of lateral water-canals in the marsupium undoubtedly is a device
to provide proper aeration for the glochidia enclosed so long a time in the mar-
supium. Whether the special hooks of the glochidia are connected with a dif-
ferentiation of function from that in the Quadrula-group is yet unknown.

The shell of the forms belonging to this group is never very heavy and solid,
and never distinctly rounded, but it is more or less elongate, rather thin, often
with distinctly colored epidermis (banded and rayed). The sculpture of the beak
is of two types: either more or less distinctly double-looped, or concentric; in the
latter case often very heavy. The hinge-teeth have a distinct tendency to
become obsolete. They are present in some forms, but are then often of peculiar
shape, and in others they exhibit all stages of reduction to complete disappearance.

326 MEMOIRS OF THE CARNEGIE MUSEUM

The male and female shells are generally quite undistinguishable, only in rare
cases is there a slight differentiation.

Remarks: The genera belonging here stand together in Simpson's system, but
associated with others, which do not belong here, and the essential characters of
this group have been entirely overlooked. Sterki has correctly recognized their
natural affinity, relying chiefly upon the remarkable shape of the glochidium. But,
as we have seen, there are additional characters in the gills of the gravid female,
which are common to all these forms, which are so peculiar, and so strongly marked,
that there cannot be the slightest doubt, that this is a natural group.

3. All the remaining genera form a third group, and are more closely allied
to each other than to any one of the groups thus far discussed. This comprises
the following of Simpson's genera: Truncilla, Micromya, Lampsilis, Obovaria, Plagi-
ola, Obliquaria, Cyprogenia, and Ptychobranchus.

They all agree in that here again each ovisac is formed by only one water-tube,
and that this remains simple. But we have here a special structure at the edge
of the gills, which permits a bulging out of the ovisacs beyond the original edge of
the gills, and this causes a projection of the marsupial gills beyond the original
edge, indicated by the ends of the gill-filaments. The swelling of the marsupium
takes place chiefly at and near the edge of the gill, and since near the base no
swelling in a longitudinal direction is possible, the arrangement of the ovisacs
often is fan-like, they being narrower and thinner near the base, and thicker (club-
like) toward the edge. This necessarily causes a folding of the edge of the gill,
which is most pronounced in one genus (Ptychobranchus) , where the whole outer
gill is used as marsupium. In all other genera it is only the posterior section of
the outer gill, which serves as marsupium (sometimes only part of this section);
and here the marsupium is almost kidney-shaped and forms only a single, irregular
fold with the non-marsupial anterior part, and sometimes a smaller fold with a
posterior non-marsupial section of the gill. The edge of the marsupial gill is
peculiarly beaded, the beads corresponding to the protruding ovisacs, the furrows
between them to the septa on the inside of the gill.

The marsupium is generally much swollen, and the ovisacs are rather large,
and, toward the edge, rather sharply defined externally; supra-anal opening always
well separated from the anal; the connection of the mantle generally short or
of medium length, leaving a rather large supra-anal opening; only rarely is the
latter small, or entirely closed; branchial opening well defined in front; anterior
to it there may be no special structures, or such may be present in the shape of
papillae or flaps on the mantle edge; inner lamina of inner gill generally entirely
connected with the abdominal sac, more rarely partly or almost entirely free.

ORTMANN: A MONOGRAPH OF THE NAJADES OF PENNSYLVANIA 327

Glochidia generally of medium size and normally of simple shape, semicircular
or semielliptical, without spines; yet in certain forms peculiar glochidia are found
(See p. 301). Placentae more or less well developed, but often atrophied, so that
the glochidia appear free. The discharge of the glochidia takes place in an un-
usual manner, through the edge of the marsupium.

The essential characters of this group (marsupium and edge of mantle) un-
doubtedly arc connected with the prolonged breeding season (all these forms are
bradytictic), and with the unusual manner of discharging the glochidia. The soft
parts are built according to a special plan, which is adapted to this peculiar dis-
charge, and facilitates the proper aeration of the glochidia during the long breeding
season, but this latter adaptation is different from that seen in the Anodonta-group
serving the same purpose.

Shape of the shell very variable, round and heavy, or more or less elongated
and lighter. Color of epidermis dull or bright. Sculpture of beak generally more
or less double-looped, more rarely concentric, often obsolete. Hinge mostly well
developed, with strong teeth, rarely somewhat reduced. Male and female shells
more or less differentiated, sometimes only very slightly, but in other cases the
differences between the shells of the two sexes are very striking.

Remarks: The essential structure of the marsupium of this group has been
entirely overlooked by Simpson. He, indeed, brings these forms together in his
system, but groups them according to characters, which are rather unimportant.
In addition, his generic divisions are not at all satisfactory, and need complete
revision, as we shall see below. Sterki also has recognized the mutual affinity
of these forms, but on account of the similarity of the glochidia to those of the
genera of the first group (Quadrula, etc.), he brings them too near to the latter.
The structure of the marsupium in this group is again so unique, that it alone
suffices to sharply separate the group from all others.

I do not think there should be any doubt or hesitation in recognizing these
groups as natural. The differences are such and so uniform, that we must regard
them as fundamental. They concern in the first instance the sexual apparatus,
and the development and specialization of the structures, which serve the physio-
logical process of propagation, we are compelled to take as the surest indication
of a phylogenetic progress within the group. One of the most important char-
acters of the fresh-water shells, called Najades, is found in the way they take care
of their young, not discharging the eggs directly into the water, but keeping them
for a time in the gills; and it is only natural that this peculiar feature, after it
had once developed, is further improved, complicated, and specialized.

328 MEMOIRS OF THE CARNEGIE MUSEUM

All modifications of the gills are directly or indirectly correlated with the
sexual apparatus, and besides a number of other characters of the soft parts as well
as the hard parts are connected with it; in fact the sexual apparatus is the most
important feature of the Najades. This is not at all astonishing; we find the same
conditions in so many other groups of animals. Thus it is clear that a system
founded upon these characters must be a natural one, and that it shou'd be fitted
to furnish us the key for the proper understanding of the systematic relations.

Thus we are to regard the above three groups (excluding Margaritana) as so
many natural divisions of the family Unionid^e, which then, of course, should
rank as subfamilies,17) namely: the Unionince, Anodontince, and Lampsilin^e. For
Margaritana it is unquestionably best to create a new family, the Margaritanid^e.

It is evident, even without the knowledge of certain details of the sexual appa-
ratus, that the Margaritanidce form the most primitive group of the Najades. The
character of the gills surely indicates this, and also the character of the branchial and
anal openings, although possibly the shell may not represent the most ancient type.

Of the Unionida1, the Unionince are certainly more primitive than the other
two subfamilies, as is evidenced by the simple character of the structure of the
marsupial gills. The Anodontince and Lampsilina? are more advanced, but they
have advanced in different directions, and each has developed special features
of the sexual apparatus. Generally speaking, the Lampsilinw contain the most
highly advanced types, as is shown by the restriction of the marsupium to a part
of the outer gill, and by the strong expression of the sexual differentiation in the
outer shell. Yet there are forms among the Anodontinoz, which show extremely
complex structures (Strophitus) , unparalleled in any other genus, and the peculiar
glochidia of the Anodontince surely mark a high stage of development. The ar-
rangement and sequence of the families and subfamilies would possibly be the
best as given above; but we must not forget that, while, for instance, the Lamp-
silinoe show relations in the structure of the edge of the gill to the Anodontince,
they more nearly approach the Unionince in the shape of the glochidia. This is,
however, the usual difficulty in all our systematic arrangements.

"Simpson unites all these forms in one subfamily, Unionince, from -which he distinguishes as a second subfamily the
Hyrianw (properly Hyriinw), with the marsupium formed by the inner gills alone, and with radial beak sculpture. The
conception of the latter subfamily on the part of Simpson is entirely wrong. Very little is known about their anatomy, but
I recently have been able to investigate some of the genera, and find that some are really Unionince (in the new sense), while
others rather agree with the family Mutelidw (See Nautilus, 23, April, 1910, p. 139, and January, 1911). With regard to
the sculpture of the beak, serious mistakes have been made. The shape of the marsupium certainly is important, but at pres-
ent in many genera is unknown. The Mtttelidw are reported to possess a lasidium in place of the glochidium, but again the
larvae of many genera of these are entirely unknown. I hope to be able to contribute to the solution of this question in
the near future, on the basis of a splendid collection representing the "Hyriinw" from South America.

ORTMANN: A MONOGRAPH OF THE NAJADES OF PENNSYLVANIA 329

I think that the families and also the three subfamilies of the Unionidce
are geologically rather old, and we cannot say that one descended from the other.
They probably go back to a common ancestral group, which probably stood some-
where between the present Margaritanidce and Unionince. It is not my object, at
present, to go into detail in this respect, and to discuss the phylogeny of the Na-
jades. This will be possible only after a fuller knowledge is at hand of the exotic
forms, and by studying also the fossil Najades. But I may here incidentally
mention that I believe that these main groups probably go back to the Cretaceous,
but the Lampsilince may be a little, but not much, younger. This is also expressed
in the geographical distribution, the Lampsilince alone being restricted to North
America (including Central America), while the others are also found in the Old
World. But I shall leave this question for the present.

B. The Genera.

1. The Margaritanidce consist solely of the genus Margaritana.

2. The Unionince contain the following genera of Simpson: Tritogonia, Unio,
Pleurobema, Quadrula. With regard to Tritogonia, Sterki (Nautilus, 21, 1907, p.
48) was the first to point out that it possesses the marsupium attributed by Simpson
to the genus Quadrula, and my own investigations have convinced me that this
is quite correct, and that Tritogonia is a true Quadrula. The other three genera
fall according to Simpson into two groups. In Unio and Pleurobema only the
outer gills serve as the marsupium, while in Quadrula both outer and inner gills
become charged with ova.

My investigations have shown that there are indeed a number of species, in
which all four gills are marsupial, while there are others, in which this is true only
of the outer gills. This is an absolutely constant character, and I have not seen
a single instance in any species, where there was any variation in this respect.
Thus it is evident, that this character, emphasized by Simpson, is important, and
it is indeed, so far as I can see, the most important one which marks a division
within this subfamily. Thus we must retain the genus Quadrula in the sense that
it comprises those species, in which all four gills serve as marsupium. However,
Simpson placed in the genus Quadrula a number of species, which according to my
observations have the marsupium formed only by the outer gills. This he did,
of course, because he did not know the marsupium. As long as we were unable
to recognize the shape of the marsupium, unless we had gravid females, it was
impossible to assign the proper place to any species, in which the gravid female
was unknown. But since I have shown that the sex of any specimen may be

330 MEMOIRS OF THE CARNEGIE MUSEUM

ascertained by the septa of the gills, and that we are able to make out the character
of the marsupium also from the sterile female, it is now possible to properly place
all species of which females are at hand. Thus I have ascertained, that the
following forms of Simpson are true Quadrulas: Q. subrotunda, kirtlandiana, rubi-
ginosa, trigona, pustulosa, metanevra, cylindrica, undulata, plicata hippopcea, and
Tritogonia tuberculata. Q. lachrymosa also belongs here, judging from specimens
obtained in Kansas and from the middle Ohio River.

Of other forms of Simpson's Quadnda which I have investigated the following
proved to have only the outer gills built to receive eggs: Q. tuberculata, coccinea,
pyramidata, obliqua, cooperiana. With the exception of the first these all fall
under Pleurobema, in Simpson's sense, but it is somewhat different in the case of
Q. tuberculata. Here we have additional characters, which have already been
pointed out by Simpson, and which induced him to place this species in a separate
subgenus, Rotundaria. The chief features are the absence of a supra-anal opening
and the peculiar sculpture of the beak. The structure of the hinge and the color
of the nacre are also rather peculiar, and there are so many distinctive characters
in evidence, that I feel fully justified in elevating the subgenus Rotundaria to the
rank of a genus.

All other forms with only the outer gills serving as marsupium should belong
to the two genera Pleurobema and Unio. There are no differences whatever in
their soft parts, and these two genera can be distinguished only by the characters
of the shell. Looking over Simpson's elaborate diagnoses (they are rather de-
scriptions), it is hard to find well-marked differential characters. The shape of
the shell may be taken as the most important criterion. In Pleurobema it is quite
solid, triangular, rhomboidal, or subovate, more or less distinctly oblique, that is
to say, with the two axes giving the two principal dimensions (height and length)
forming a distinct and sharp angle with each other. This causes the umbonal
region, which is more or less elevated, to assume an anterior location in the shell.
In Unio on the contrary the shell varies from oval to elongate, is straight, with the
axes very nearly vertical to each other, with the result that the umbonal region is
less elevated and less anterior. Further, in Pleurobema the color of the epidermis
is from dark brown to light brown or olive, with prevalence of the lighter shades,
and the color of the nacre is white or silvery, rarely lighter or darker pinkish or
salmon-red, but never purple. In Unio, the epidermis is brown to black, very
rarely lighter, with prevalence of the darker shades, and the nacre may be white,
but more often assumes red to purple tints.18

"As will be shown elsewhere, the North American genus Unio should be separated from the typical European Unio
(Type: pictorum Linnaeus). For the former the name Elliptio RaBnesque 1819 is available.

ortmann: a monograph of THE NAJADES OF PENNSYLVANIA 331

The Pennsylvanian species placed under Unio are unquestionably more closely
allied to each other than to those placed under Pleurobema, but there are other
extralimital species, which cannot be so easily associated with this genus. Further-
more, we have several types included in the genus Pleurobema which are rather
distinct, and which may be entitled to generic rank, unless we prefer to leave all
these forms united in the old genus Unio. The Pennsylvanian material representing
the species of this group is scarcely extensive enough to enable us to settle this ques-
tion. There are a multitude of forms of Unio and Pleurobema in the central and
southern parts of North America, many of which apparently form connecting links.
Before these have been more closely studied, and before it has been positively ascer-
tained that they belong in this group, it is futile to attempt a final classification of
these forms. Thus I retain the main divisions (Unio and Pleurobema) introduced by
Simpson, adding to the latter genus the species removed from Quadrula. With
regard to Q. cooperiana it is absolutely certain that it stands close to another
species placed by Simpson in Pleurobema, P. ossopus, and is still more closely allied
to P. cicatricosum. With regard to Q. coccinea, pyramidata, and obliqua, this is
not so evident, but I think that there are some southern types of Pleurobema,
with which these should be associated.

I assume that Quadrula having four gills serving as marsupium, represents a
more primitive stage than the others, where only the outer gills have this func-
tion and the corresponding structure. It appears, as a matter of course, that the
differentiation between the gills, the inner serving exclusively the function of breath-
ing, and the outer exclusively that of propagation, represents a higher stage of
development than in those cases in which all gills serve both functions; and this is
further supported by the fact that in certain species of Quadrula (subrotunda,
kirtlandiana, rubiginosa) we have types, in which the swelling of the marsupium
in the breeding season is so small, that the water-tubes and their contents remain
subcylindrical, which undoubtedly is a primitive condition. The fact that the
most primitive form, Margaritana, has also all four gills marsupial, further speaks
for our assumption. Yet I cannot suppress the thought that the opposite opinion
has something in its favor, and the evidence for this is furnished by the fundamental
differentiation of the edge of the inner and outer gill, as is shown by my description
of the peculiar furrow along the edge of the inner gill, which is absent in the outer
gill. As stated above, I cannot surmise what the meaning of this feature is. It
ma}r be possible that it represents an old, primitive differentiation of the two gills,
the edge of the outer gill indicating the original restriction of the function of propa-
gation of this gill. Then, of course, we would have to assume that in Quadrula

332 MEMOIRS OF THE CARNEGIE MUSEUM

this function spread secondarily to the inner gill. I must confess that this idea
appears to me as unlikely, since it means a going back to a less specialized stage,
two organs, each of which originally served a different function, now assuming
both. The balance of evidence is, I think, in favor of the assumption made here,
but the question is open to further studj'.

3. The Anodontince. Simpson's genera: Strophitus, Anodonta, Anodontoides,
Symphynota, and Alasmidonta belong here. The definition of these according to
shell characters, as given by Simpson, is rather satisfactory, and I merely suggest
another arrangement of them. There is no question that Strophitus, in spite of
the peculiarity of the marsupium, comes nearer to Alasmidonta, than to any other
genus, which is chiefly shown by the heavy sculpture of the beak, and the tendency
of the inner lamina of the inner gill to become united with the abdominal sac.
On the other hand Symphynota and Anodonta are more closely allied, as is shown
by the sculpture of the beak. Anodontoides is essentially an Anodonta, but with
different sculpture of the beak. We apparently are dealing here with a number of
types, which have developed independently from a common source, and of which
each has preserved one or the other primitive character, while it is more advanced
in other features. Strophitus is extreme in the structure of the marsupium; Ano-
donta is extreme in the reduction of the hinge-teeth. With regard to the hinge-teeth
we observe that we have primitive types in Alasmidonta heterodon, and Symphynota
compressa and viridis; yet they are again peculiarly transformed. The inner lamina
of the inner gills is more primitive (free) in Anodonta, Anodontoides, and Symphy-
nota, while a more advanced stage is observed in Alasmidonta and Strophitus.
Anodonta represents an extreme stage in the development of the supra-anal opening.
The following sequence possibly would be advisable: Alasmidonta, Strophitus, Sym-
phynota, Anodontoides, Anodonta.

4. Lampsilince. The following genera of Simpson's "Synopsis" belong here:
Truncilla, Micromya, Lampsilis, Obovaria, Plagiola, Obliquaria, Cyprogenia, Ptycho-
branchus. Some of them are very well defined, while others stand in need of a
thorough revision.

First of all, Ptychobranchus stands by itself, on account of its very unique
marsupium, the peculiarity of which is fully appreciated by Simpson. Yet, as
we have seen, Ptychobranchus is truly lampsiline, for it has the characteristic struc-
ture of the edge of the marsupium. There is no question that it is the most primi-
tive form in this subfamily, although possibly the present condition of the marsupium
is not the original one.

Among the others we see that the structure of the marsupium is practically

ortmann: a monograph of THE NAJADES OF PENNSYLVANIA 333

alike; there are very few minor differences, but the genera, Obliquaria and Cypro-
genia, stand out more prominently among them.11' They possess a remarkably
small number of ovisacs, which stand rather in the middle of the outer gill, and
the shell also is peculiar in its shape and ornamentation. Cyprogenia is in addition
distinguished by the great prolongation of the marsupium, which forces it to coil
up within the shell. Thus these two genera are well defined.

As has been seen above, there is among the Lampsilince a further differentia-
tion connected with the sexual apparatus, which, in my opinion, is of prime value.
This is the development of special structures on the edge of the mantle in front of
the branchial opening. Although some of these structures have been noticed by
previous writers, no importance has been attached to them.

We have observed, that in the genera Truncilla, Micromya, and in certain
species of Lampsilis such structures are present, while in other species of Lampsilis,
as well as in Obovaria, Plagiola, Obliquaria, Cyprogenia, and Ptychobranchus, they
are absent, or barely indicated. We have seen, that Ptychobranchus, Obliquaria,
and Cyprogenia may be distinguished on other grounds. This leaves certain species
of Lampsilis {ligamentina, alata, gracilis), and the genera Obovaria and Plagiola as
a group, possessing a similar conformation of the edge of the mantle. Plagiola,
Obovaria, and Lampsilis ligamentina are practically identical in the entire structure
of the soft parts, and in the form of the shell they also stand rather close together.
With regard to Lampsilis ligamentina, I do not see any reason why it should not
be placed with Obovaria; in fact it fully agrees with Simpson's diagnosis of Obovaria,
and could be grouped very well with 0. ellipsis. Plagiola can be distinguished
from Obovaria only by the characters of the shell, chief among which are the presence
of a distinct posterior ridge and the peculiar markings of the epidermis. (In ad-
dition, P. securis has a peculiar glochidium, but the writer does not know whether
this glochidium is found in other species of Plagiola.)

Lampsilis alata and gracilis have been placed by Simpson in a separate sub-
genus, Proptera, on account of the peculiarities of the shell (general shape and
character of hinge-teeth). Sterki has proposed to elevate it to the rank of a
genus on account of the odd glochidium of L. alata, assuming, that this shape is
also found in the other species of Proptera. Since I have shown, that the glochidium
of L. gracilis differs fundamentally from that of L. alata, and also from all other
Lampsilince, we are forced to restrict the genus Proptera to those forms, which
have a glochidium like L. alata (purpurata, Icevissima), and should create a new

"Recently, Sept. 24, 1910, 1 found :. gravid female of Cyprogenia irroraia in (he Ohio Rive* a< Portsmouth, Scioto Co.,
Ohio, and had a chance to study its soft parte.

334 MEMOIRS OF THE CARNEGIE MUSEUM

genus for L. gracilis. I propose here the name of Paraptera for it. Paraptera
and Proptera are closely allied in the structure of the soft parts, and agree especially
in the edge of the mantle.20 They also have many similarities in the shell, and
greatly differ in this respect from Obovaria and Plagiola.

The separation of the species would leave in the genus Lampsilis the following
species: parva, iris, nasuta, recta (forming a group with papilla on the edge of the
mantle); and ventricosa, ovata, multiradiata, cariosa, orbiculata (with a flap on the
edge of the mantle). But I also unite with this genus Micromya fabalis, which
does not differ in any essential point from Lampsilis, and possesses on the edge
of the mantle the characteristic papilla? of the first group of Lampsilis.

Lastly we have the genus Truncilla, which is not only characterized by the
soft parts, as we have seen above, but more emphatically by the shape of the shell.
Simpson's division of this genus into subgenera is well supported.

It is clear, that Truncilla and Lampsilis are the most advanced t}rpes of this
subfamily, and that Ptychobranchus in the first line, and then Obliquaria, Obovaria,
and Plagiola are rather primitive. Cyprogenia, Paraptera, and Proptera represent
special modifications of this primitive type.

We thus obtain the following arrangement of the families, subfamilies, and
genera of the Najades found in Pennsylvania. (Compare my Preliminary Report,
1910, p. 114.)

1. Family Margaritanid^e Ortmann (Nautilus, February, 1911).

Diaphragm incomplete, formed by the gills; posteriorly the outer lamina of
the outer gills not connected with the mantle for a considerable distance; anterior
end of the inner gills separated from the palpi by a gap;21 branchial and anal
openings ill-defined, and the latter not closed above; no supra-anal developed;
gills without water-tubes, and with scattered interlamellar connections, which in
certain places form irregular diagonal rows; marsupium formed by all four gills;
larva a small, semicircular glochidium, without distinct hooks; shell elongated;
sculpture of the beak concentric; hinge-teeth imperfect; epidermis blackish.

Genus Margaritana Schumacher.22

20l do not understand Simpson's description (1900, p. 567) of the edge of the mantle of Proptera, and have not seen
anything which could be called a "thickened flap of the outer fold."

"This item should be inserted in the diagnosis on account of the different structure found in the Mutdida (See Nautilus,
January and February, 1911).

'"< v-rtain writers (See Thiele, in Brauer, Suesswasserfauna Deutschlands, Heft 19, 1909, p. 32) consider Margaritana
the type of the genus Unio, and call Unto by the name of Lymnium Oken 1815. This is not correct, as Simpson has already
clearly shown (1900, p. 674, footnote 1).

ORTMANN: A MONOGRAPH OF THE NAJADES OF PENNSYLVANIA 335

2. Family Unionid^e (d'Orbigny), Ortmann (Nautilus, February, 1911).

Diaphragm complete, formed by the gills; posteriorly the outer lamina of the
outer gill connected with the mantle to its posterior end; anterior end of the inner
gills separated from the palpi by a gap; branchial and anal openings sharply
separated from one another by the diaphragm; anal opening very rarely not closed
above and without supra-anal, generally closed and with a supra-anal opening
(which very rarely may be obliterated); gills with water-tubes and distinct, con-
tinuous interlamellar septa, running parallel to the filaments. Marsupium in all
four gills, or only in the outer gills; larva a glochidium. Shell of very variable
shape; sculpture of the beak more or less reduced, of various types, but originally
of the concentric or zig-zag pattern; hinge-teeth perfect or imperfect; epidermis
plain or with color-markings.

1. Subfamily Unionin^e (Swainson) Ortmann.

Inner lamina of inner gills generally free from the abdominal sac (sometimes,
in extralimital forms connected) ; supra-anal opening sometimes not separated from
the anal, normally present, the closed part rather short; branchial opening well-
defined; no papilla? or flaps on edge of mantle in front; marsupium formed by
all four gills, or by the outer gills only; edge of marsupium always sharp and not
distending; water-tubes not divided in the gravid female; glochidium semi-elliptic
or semicircular, without spine; shell generally heavy and solid, rounded to elon-
gated, mostly with dull-colored epidermis; sculpture of the beak generally rather
indistinct, concentric, or pustulous, or with indications of double loops or zig-zag
bars; hinge always complete, with rather strong teeth; generally no difference of
sex shown in the shell.

Genera

a1. All four gills serving as marsupium ; anal and supra-anal opening separated, but separation often deciduous ;
shell of various shapes, always rather heavy, outside smooth or with sculpturings.

Genus Quadrula Rafinesque.
a2. Only the outer gills serving as marsupium.

b1. Supra-anal not separated from the anal opening; sculpture of the beak consisting of numerous, fine,
irregular corrugations; shell rounded; beak-cavities very deep; nacre dark violet or purple.

Genus Rotundaria Rafinesque.
6*. Supra-anal opening separated from the anal ; sculpture of the beak consisting of few, generally concen-
tric ridges; shell rounded or elongated; beak-cavities of medium depth, or shallow.
c1. Shell oblique, ovate, or rhomboidal, with prominent beaks, which are placed rather anteriorly;
color of epidermis brown to light, generally not very dark, often with rays; nacre varying from
light red to white. Genus Pleurobema Rafinesque.

336 MEMOIRS OF THE CARNEGIE MTjSEUM

c2. Shell ovate or elongate; beaks not very prominent, and not much produced; color of epidermis
brown to black, only indistinctly rayed, if at all; nacre varying from white to deep purple, with
the dark shades prevailing. Genus Elliptic) Rafinesque .2

2. Subfamily Anodontin^e Ortmann.

Inner lamina of inner gill free from the abdominal sac or more or less connected
with it, rarely entirely connected; supra-anal opening well separated from the anal,
sometimes the connection of the mantle separating it from the anal is very long,
and the supra-anal is quite short; branchial opening well defined, no papillae or
flaps in front of it on the edge of the mantle; marsupium formed by the outer gills
in their whole length, distending, when charged, and the thickened tissue at the
edge capable of stretching out in a direction transverse to the gill, but not beyond
the edge (or only slightly so); water-tubes in the gravid female divided longi-
tudinally into three tubes, with only the one in the middle used as an ovisac, and
closed at the base of the gill; glochidium semicircular or triangular, with a spine
(hook) in the middle of the ventral margin of each valve; shell generally not
very heavy, often thin, never round, but more or less elongated; color of epidermis
generally bright and with color markings; sculpture of the beak double-looped or
concentric, in the latter case often extremely heavy; hinge rarely complete, and
if so, of peculiar structure; generally there is a distinct tendency toward the re-
duction of the hinge-teeth, and often they are completely absent; sexual differences
in the shell very rarely present.

Genera

a1. Sculpture of the beak concentric, heavy; inner lamina of inner gills having a tendency to become united
with the abdominal sac.

b1. Shell with cardinal teeth; lateral teeth present or wanting; ovisac of each water-tube in the gravid
female not subdivided; placentae poorly developed, and glochidia practically free.

Genus Alasmidonta Say.
b2. Shell with the hinge-teeth obsolete; only a mere vestige of the cardinal teeth remaining; ovisac of each
water-tube subdivided into a number of compartments, each containing a well developed placentula,
running crosswise to the gill; placentula? persistent till they are discharged.

Genus Strophitus Rafinesque.
a2. Sculpture of the beak generally double-looped, if concentric, not heavy; inner lamina of inner gills always
free; ovisacs not subdivided.

61. Shell rather solid, lateral hinge-teeth more or less imperfect; cardinal teeth present, but of peculiar
shape, sculpture of the beak more or less double-looped. Genus Symphynota Lea.

b-. Shell thin; hinge edentulous, or with mere traces of cardinal teeth.

aSee above p. 330, footnote. The European genus Unio Retzius, although similar in anatomical detail, differs in shape
and color of shell, and chiefly in beak sculpture (wavy, zig-zag or double looped) and glochidia. The type species of Ellip-
tic) is U. crassidens Lamarck.

ORTMANN: A MONOGRAPH OF THE NAJADES OF PENNSYLVANIA 337

c1. Sculpture of the beak concentric, not heavy; supra-anal about as long as the united part of the

margin of the mantle, which separates it from the anal opening.

Genus Anodontoides Simpson.
c2. Sculpture of the beak generally double-looped; supra-anal small, widely separated from the anal

opening. Genus Anodonta Lamarck.24

3. Subfamily Lampsilin^e Ortmann.
Inner lamina of inner gills rarely more or less free from the abdominal sac,
generally connected with it throughout; supra-anal opening separated from the
anal, rarely entirely closed; branchial opening well defined; edge of the mantle
in front of the branchial opening smooth to crenulated, or with peculiar papillae
or a flap; marsupium rarely formed by the whole outer gill, generally only by or
within the posterior part of the outer gill; edge of marsupium, when charged,
distending, and bulging out beyond the original edge of the gill, generally assuming
a beaded appearance; water-tubes simple in the gravid female; glochidium semi-
circular or semi-elliptic, without spine, rarely celt-shaped and with two spines;
shell heavy or lighter, rounded, or oval, to elongate; color of epidermis rarely
dull, mostly bright, with color markings; sculpture of the beak generally double-
looped, but often obsolete, more rarely concentric; hinge generally complete, with
well-developed teeth, which only in rare cases show a tendency to become reduced;
sexual differences more or less noticeable in the shell, often very strongly expressed.

Genera

a1. Whole outer gill serving as marsupium, its edge thrown into a number of folds.

Genus Ptychobranchus Simpson.
a2. Marsupium located in the posterior part of the outer gill, and projecting considerably beyond its original
edge.
6l. Inner edge of mantle in front of branchial opening smooth, crenulated, or with few and weak teeth,
but without papilla? or flaps; male and female shell differing only slightly, or hardly at all.
c1. Shell rounded, with surface sculpturings; marsupium consisting of comparatively few ovisacs,
located near the centre of the outer gill.
d1. A row of large knobs on each valve; marsupium projecting beyond the edge of the gill, not
coiled up. Genus Obliquaria Rafinesque.

d2. Sculpture of the surface nodular; marsupium projecting immensely beyond the edge of the
gill, and coiled up spirally. Genus Cyprogenia Agassiz.

c2. Shell rounded, oval, or compressed, without surface sculpture; marsupium large, kidney-shaped,
occupying a large portion of the posterior part of the outer gill.
d1. Shell solid, rounded, ovate or triangular, more or less swollen; if compressed, not winged.

"European writers (See Tliiele in Brauer, Bueeswaaserfauna Dautsohlande, Heft 19, 1909, p. 32) recently use Anadontitet

Bruguiere, 1792, for Anodonta Lamarck 1790. This in entirely wrong. The type of AnodontiUs is A.crispata Bruguiere a
South American shell, I elonging to Global is Gray 1847. Of course, Aiwdontilvs has the priority over Qtiba i», and should be
used in its place.

338 MEMOIRS OF THE CARNEGIE MUSEUM

e1. Shell rounded or ovate, without distinct posterior ridge. Color of epidermis dull, without
color-markings, or simply with rays. Glochidia normal in shape and size.

Genus Obovaria Rafinesque.
e2. Shell ovate or triangular, with a distinct, and often sharp posterior ridge; epidermis
painted with broken rays, forming a pattern of lunate, squarish, or arrow-shaped
marks; glochidia (in one species at least) rather large and with gaping margins.

Genus Plagiola Rafinesque.
d2. Shell not very solid, or thin, ovate, more or less compressed and winged along the dorsal
margin (at least when young).
e1. Glochidia of normal shape, but of unusually small size. Genus Paraptera gen. nov.26
e2. Glochidia celt-shaped, with two spines on each valve, with gaping margins.

Genus Proptera Rafinesque.

b2. Inner edge of mantle in front of the branchial opening, chiefly in the female, differentiated by papillae

or flaps, and sometimes not parallel to the outer edge; male and female shells distinct, generally very

markedly so.

c1. Inner edge of mantle essentially parallel to the outer edge, with papillae or flaps; shell ovate or

elongate, the postbasal region of the female somewhat swollen and more rounded than in the male.

Genus Lampsilis Rafinesque.

c2. Inner edge of mantle in the female in front of the branchial opening not parallel to the outer edge,

but more or less remote from it, with papillae; shell ovate or triangular, the postbasal region of

the female very strongly inflated and projecting, giving to the whole shell an entirely different

shape from that of the male. Genus Truncilla Rafinesque.

BIBLIOGRAPHY.

This list contains only those papers, which have been quoted in the text.
Baker, F. C— The Mollusca of the Chicago Area. Pelecypoda. (Bulletin Chicago

Academy Sciences, 3, Part 1, 1898.)
Brooks, W. K. — Handbook of Invertebrate Zoology, 1890.

Conner, C. H. — Supplementary Notes on the Breeding Season of the Unionidse. (Nau-
tilus, 22, 1909, pp. 111-112.)
Cooke, A. H. — Molluscs, in the Cambridge Natural History, 3, 1895.
Harms, W. — Zur Biologie und Entwicklungsgeschichte der Flussperlmuschel. (Zool-

ogischer Anzeiger, 31, 1907, pp. 814-824.)
Lea, I— Observations on the Genus Unio, Vols. 1-13, 1832-1874. (In the text I cite

only according to volume or plate. The pagination of my copy is entirely at variance

with the references given by Simpson.)
Lefevre, G., and Curtis, W. C. — The Marsupium of the Unionidse. (Biological Bulletin

19, 1910, pp. 31-34.)
Ortmann, A. E. — The Breeding Season of the Unionidge in Pennsylvania. (Nautilus,

22, 1909, pp. 91-95, 99-103.)

"Type P. gracilis (Barnes).

ortmann: a monograph of THE NAJADES OF PENNSYLVANIA 339

Ortmann, A. E. — A Preliminary List of the Unionidse of Western Pennsylvania, with

New Localities for Species from Eastern Pennsylvania. (Annals Carnegie Museum,

5, 1909, pp. 178-210.)
Ortmann, A. E— A New System of the Unionidse. (Nautilus, 23, 1910, pp. 114-120.)
Ortmann, A. E. — The Systematic Position of the Unionid Genus Parreysia. (Nautilus,

23, 1910, pp. 139-142.)
Ortmann, A. E. — The Marsupium of the Anodontinse. (Biological Bulletin, 19, 1910,

p. 217.)
Simpson, C. T. — Synopsis of the Najades, or Pearly Freshwater Mussels. (Proceedings

United States National Museum, 22, 1900, pp. 501-1044.)
Sterki, V. — Some observations on the Genital Organs of Unionidse, with Reference to

Classification. (Nautilus, 12, 1898, pp. 18-32.)
Sterki, V. — Notes on the Unionidse and their Classification. (American Naturalist, 37,

1903, pp. 103-113.)
Taylor, J. W. — Monograph of the Land and Freshwater Mollusca of the British Isles.

Structural and General Volume, 1894-1900.

EXPLANATION OF PLATE LXXXVI.

Horizontal cross-sections through the gills of Unionidce, to show structure of water-
tubes, ovisacs, and septa, in the males, sterile and gravid females.

All photographs taken with Bausch & Lomb 1 inch objective. None of the figures
has been retouched.
Fig. 1. Quadrula subrotunda (Lea) Male. — Allegheny River, Kelly, Armstrong County,

June 22, 1909. Cat. No. 61.4400. Specimen 199.

la. Left inner gill. 16. Left outer gill.
Fig. 2. Quadrula subrotunda (Lea) Female, sterile. — Allegheny River, Kelly, Armstrong

County, Sept. 5, 1907. Cat. No. 61.3083. Specimen 5.

la. Left inner gill. 2a. Left outer gill.
Fig. 3. Quadrula subrotunda (Lea) Female, gravid (eggs). — Allegheny River, Kelly,

Armstrong County, June 22, 1909. Cat. No. 61.4399. Specimen 7.

3a. Right inner gill. 36. Right outer gill.
Fig. 4. Quadnda tuberculata (Barnes). Female, sterile. Mahoning River, Mahoningtown,

Lawrence County, Aug. 4, 1908. Cat. No. 61.3574. Specimen 34.

4a. Right inner gill. 46. Right outer gill.
Fig. 5. Elliptio gibbosus (Barnes) Male. — Little Beaver Creek, Enon Valley, Lawrence

County, May 11, 1907. Cat. No. 61.2186. Specimen 54.

5a. Left inner gill. 56. Left outer gill.
Fig. 6. Elliptio gibbosus Barnes. Female, gravid (eggs). Allegheny River, Templeton,

Armstrong County, Aug. 13, 1907. Cat. No. 61.2963. Specimen 56.

6a. Left inner gill. 66. Left outer gill.
Fig. 7. Strophitus edentulus (Say) Female, sterile. Allegheny River, Templeton, Arm-
strong County, Aug. 4, 1909. Cat. No. 61.4150. Specimen 117.

7a. Left inner gill. 76. Left outer gill.
Fig. 8. Strophitus edentulus (Say) Female, gravid, young (eggs), beginning of period of

gravidity. Conococheague Creek, Greencastle, Franklin, County, Sept. 6, 1909.

Cat. No. 61.4160. Specimen 112.

Left outer gill.
Fig. 9. Strophitus edentulus (Say) Female, gravid (eggs), a little more advanced than fig. 8.

Cush-Cushion Creek, Green Township, Indiana County, July 11, 1908. Cat. No.

61.3627. Specimen 109.

Left outer gill.
Fig. 10. Strophitus edentulus (Say) Female, sterile, very young (shell 17 mm. long).

Crooked Creek, Rosston, Armstrong County, Aug. 3, 1909. Cat. No. 61.4148.

Specimen 230.

10a. Left inner gill. 106. Left outer gill.

340

Memoirs Carnegie Museum, Vol. IV.

Plate LXXXVI.

7a

1b

5a 5b

\ \^5»

7 b

2a

4a

4b

8

10 a 10b

11 a 11 b

iw

6a 6b

12b

13a 13b

14 a 14b

Sections of • in.i.s

ortmann: a monograph of THE NAJADES OF PENNSYLVANIA 341

Fig. 11. Symphynota viridis (Conrad) Probably hermaphrodite, grand, very young (shell

31 mm. long) (with eggs and young glochidia). Great Tonoloway Creek, Thompson

Township, Fulton County, Sept. 5, 1909. Cat, No. 61.4221. Specimen 235.

11a. Left inner gill. lib. Left outer gill.
Fig. 12. Symphynota costata (Rafinesque) Female, grand, discharging posteriorly (glochidia).

Ten Mile Creek, Amity, Washington County, April 22, 1908. Cat. No. 61.3303.

Specimen 82.

12a. Left inner gill. 12b. Left outer gill.
Fig. 13. Anodonta grandis gigantea (Lea) Male, young (shell 37 mm. long). Pond, Har-

marville, Allegheny County, Oct. 10, 1908. Cat, No. 61.3668. Specimen 231.

13a. Left inner gill. 136. Left outer gill.
Fig. 14. Anodonta grandis gigantea (Lea) Female, sterile, young (shell, 46 mm. long). Pond,

Harmarville, Allegheny County, Oct. 10, 1908. Cat. No. 61.3668. Specimen 236.

14a. Right inner gill. 146. Right outer gill.
Fig. 15. Anodonta grandis Say. Female, gravid (eggs), beginning of gravidity. Conneaut

Lake, northeast shore, Crawford County, Aug. 8, 1908. Cat. No. 61.3657. Specimen

203.

Left outer gill.
Fig. 16. Anodonta grandis gigantea (Lea) Female, gravid (glochidia), fully charged. Pond,

Harmarville, Allegheny County, Oct. 10, 1908. Cat, No. 61.3668. Specimen 4.

Left outer gill (one lamina only).

EXPLANATION OF PLATE LXXXVIL

Horizontal (figs. 1-10) and vertical (figs 11-16) cross-sections through gills of Najades,
to show structure of water-tubes, ovisacs, and septa, in the males, sterile and gravid females,
and to show structure of the edge of the gills.

All photographs taken with Bausch & Lomb 1 inch objective. None has been re-
touched.

Fig. 1. Ptychobranchus phaseolus (Hildreth) Male. Loyalhanna River, Idlepark, West-
moreland County, Sept. 21, 1907. Cat, No. 61.3035. Specimen 119.

la. Left inner gill. 16. Left outer gill.
Fig. 2. Ptychobranchus phaseolus (Hildreth) Female, sterile. Raccoon Creek, New Shef-
field, Beaver County, June 30, 1908. Cat. No. 61.3275. Specimen 123.

Left outer gill.
Fig. 3. Ptychobranchus phaseolus (Hildreth) Female, gravid (eggs). Loyalhanna River,

Idlepark, Westmoreland County, Sept. 21, 1907. Cat. No. 61.3035. Specimen 120.

Left outer gill.
Fig. 4. Plagiola securis (Lea) Male. Allegheny River, Kelly, Armstrong County, Sept.

5, 1907. Cat. No. 61.3029. Specimen 127.

4a. Left inner gill. 46. Left outer gill.
Fig. 5. Plagiola securis (Lea) Female, gravid (glochidia). Ohio River, Cooks Ferry,

Beaver County, Sept. 10, 1908. Cat, No. 61.3565. Specimen 129.

Left outer gill.
Fig. 6. Paraptera gracilis (Barnes) Female, sterile. Ohio River, Industrj', Beaver County,

Sept. 8, 1908. Cat. No. 61.3549. Specimen 137.

Left outer gill.
Fig. 7. Lampsilis iris (Lea) Female, sterile. Presque Isle Bay, Lake Erie, Erie, Erie

County, May 22, 1909. Cat. No. 61.4090. Specimen 144.

Left outer gill.
Fig. 8. Lampsilis recta (Lamarck) Male. Cheat River, Cheat Haven, Fayette County,

Sept. 16, 1907. Cat. No. 61.3020. Specimen 149.

8a. Left inner gill. 86. Left outer gill.
Fig. 9. Lampsilis recta (Lamarck) Female, sterile. Cheat River, Cheat Haven, Fayette

County, July 10, 1908. Cat. No. 61.3509. Specimen 246.

Left outer gill.
Fig. 10. Lampsilis ventricosa (Barnes) Female, sterile. Little Beaver Creek, Enon Valley,

Lawrence County, May 11, 1907. Cat. No. 61.2144. Specimen 168.

Left outer gill.
Fig. 11. Margaritana margaritifera (Linnaeus). Indian Run, Rene Mont, Schuylkill

County, May 4, 1909. Cat. No. 61.4272. Specimen 200.

342

Memoirs Carnegie Museum, Vol, IV,

Plate LXXXVII.

-

4a 4b

8a 8b

K

11 a

11 b

&l

Ft

g

V

14 a 14b

(

12a 12b

13a 13b

15a 15b

16 a 16b

Sections of ( lu.i.s

ortmann: a monograph of THE NAJADES OF PENNSYLVANIA 343

11a. Left inner gill. lib. Left outer gill.
Fig. 12. Quadrula undulata (Barnes) Male. Mahoning River, Mahoningtown, Lawrence

County, Sept. 2, 1907. Cat. No. 61.3067. Specimen 17.

12a. Right inner gill. 12b. Right outer gill.
Fig. 13. Quadrula undulata (Barnes) Female, sterile. Mahoning River, Mahoningtown,

Lawrence County, Sept. 2, 1907. Cat. No. 61.3067. Specimen 195.

13a. Left inner gill. 136. Left outer gill.
Fig. 14. Quadrula undulata (Barnes) Female, gravid (eggs). Shenango River, Linesville,

Crawford County, June 19, 1909. Cat. No. 61.4339. Specimen 16.

14a. Right inner gill. 146. Right outer gill.
Fig. 15. Pleurobema cesopus (Green) Male. Allegheny River, Kelly, Armstrong County,

Sept. 5, 1907. Cat. No. 61.3064. Specimen 194.

15a. Left inner gill. 15b. Left outer gill.
Fig. 16. Pleurobema cesopus (Green) Female, gravid (eggs). Allegheny River, Kelly,

Armstrong County, June 22, 1909. Cat. No. 61.4333. Specimen 38.

16a. Left inner gill. 166. Left outer gill.

EXPLANATION OF PLATE LXXXVIII.

Vertical (figs. 1-5, 10, 12-18,20, 21) and longitudinal (figs. 6-9, 11, 19) cross-sections
through the gills of Unionidw, to show structure of ovisacs, septa, and edges of gills in
males, sterile and gravid females.

All photographs taken with Bausch & Lomb 1 inch objective. None has been re-
touched.
Fig. 1. Elliptic gibbosus (Barnes) Female, gravid (eggs). Allegheny River, Templeton,

Armstrong County, Aug. 13, 1907. Cat. No. 61.2963. Specimen 56 (Same as pi.

LXXXVI, fig. 6).

Left outer gill.
Fig. 2. Alasmidonta undulata (Say) Female, gravid (eggs), beginning of gravidity. West

Branch Mahantango Creek, Richfield, Juniata County, July 18, 1908. Cat. No.

61.3713. Specimen 75.

2a. Right inner gill. 26. Right outer gill.
Fig. 3. Alasmidonta undulata (Say) Female, gravid (glochidia) fully charged. Schuylkill

Canal, Manayunk, Philadelphia County, April 24, 1909. Cat. No. 61.4240. Speci-
men 206.

3a. Base of left outer gill. 36. Edge of left outer gill.
Fig. 4. Strophitus edentulus (Say) Female, sterile. Allegheny River, Templeton, Arm-
strong County, Aug. 4, 1909. Cat. No. 61.4150. Specimen 117 (Same as PI. LXXXVI,

fig. 7).

Right outer gill.
Fig. 5. Strophitus edentulus (Say) Female, gravid (glochidia), fully charged. Newmans

Branch, Wadsworth, Butler County, April 17, 1908. Cat. No. 61.3276. Specimen

201.

Left outer gill.
Fig. 6. Strophitus edentulus (Say) Same specimen as fig. 4.

Left outer gill.
Fig. 7. Strophitus edentulus (Say) Female, gravid (eggs), beginning of gravidity. Allegheny

River, Templeton, Armstrong County, Aug. 4, 1909. Cat. No. 61.4150. Specimen

116.

Left outer gill.
Fig. 8. Strophitus edentulus (Say) Female, gravid (glochidia), more advanced. Shenango

River, Linesville, Crawford County, Oct, 10, 1907. Cat. No. 61.3118. Specimen 202.

Left outer gill.
Fig. 9. Strophitus edentulus (Say) Same specimen as fig. 5.

Left outer gill.

344

Memoirs Carnegie Museum, Vol. IV.

2a

2b

/si-*.*:1

6

/

10

3 a

3b

?3k

r

T£-

8

12a

12b

Plate LXXXVIII.

19

15

20

$

^

11

w

17

Sections of Gills

ortmann: a monograph of THE NAJADES OF PENNSYLVANIA 345

Fig. 10. Symphynota costata (Rafinesque) Female, gravid (young glockidia). Loyalhanna

River, Ligonier, Westmoreland County, Aug. 3, 1906. Cat. No. 61.1819. Specimen

205.

Left outer gill.
Fig. 11. Anodonta grandis gigantea (Lea) Female, grand ' (glochidia). Pond, Harmarville,

Allegheny County, Oct. 10, 1908. Cat. No. 61.3668. Specimen 4 (Same as PI.

LXXXVI, fig. 16).

Posterior part of base of right outer gill.
Fig. 12. Ptychobranchus phaseolus (Hildreth) Male. Neshannock Creek, Leesburg,

Mercer County, Oct. 17, 1907. Cat. No. 61.3114. Specimen 183.

12a. Left inner gill. 12b. Left outer gill.
Fig. 13. Ptychobranchus phaseolus (Hildreth) Female, sterile. Presque Isle Bay of Lake

Erie, Erie, Erie County, May 22, 1909. Cat, No. 61.4123. Specimen 214.

Left outer gill.
Fig. 14. Ptychobranchus phaseolus (Hildreth) Female, gravid (eggs). Mahoning River,

Mahoningtown, Lawrence County, Sept. 2, 1907. Cat. No. 61.3032. Specimen 247.

Left outer gill.
Fig. 15. Obovaria ligamentina (Lamarck) Female, sterile. Cheat River, Cheat Haven,

Fayette County, Sept. 16, 1907. Cat. No. 61.3013. Specimen 156.

Right outer gill.
Fig. 16. Paraptera gracilis (Barnes) Female, sterile, but just discharged. Ohio River,

Industry, Beaver County, July 11, 1909. Cat. No. 61.4111. Specimen 210.

Left outer gill.
Fig. 17. Lampsilis luteola (Lamarck) Female, sterile. Little Mahoning Creek, Goodville,

Indiana County, May 25, 1905. Cat. No. 61.1760. Specimen 159.

Left outer gill.
Fig. 18. Lampsilis luteola (Lamarck) Female, gravid, discharging glochidia. Ten-Mile

Creek, Amity, Washington County, April 22, 1908. Cat. No. 61.3243. Specimen 207.

Left outer gill.
Fig. 19. Lampsilis cariosa (Say) Female, sterile. Delaware River, Yardley, Bucks County,

April 24, 1908. Cat. No. 61.3417. Specimen 166.

Left outer gill.
Fig. 20. Lampsilis multiradiata (Lea) Female, gravid, discharging glochidia. Little

Beaver Creek, Darlington, Beaver County, Aug. 9, 1907. Cat. No. 61.2921. Speci-
men 163.

Left outer gill.
Fig. 21. Truncilla perplcxa rangiana (Lea) Female, sterile. Allegheny River, Godfrey,

Armstrong County, Aug. 10, 1909. Cat. No. 61.3978. Specimen 173.

Left outer gill.

EXPLANATION OF PLATE LXXXIX.

Glochidia of Unionidce.

All photographs taken with Bausch & Lomb two-thirds inch objective. In figs. 1,
2, and 4, some of the outlines of the glochidia have been intensified.
Fig. 1. Quadrula subrotunda kirtlandiana (Lea). Mahoning River, Mahoningtown, Law-
rence County, Aug. 2, 1907. Cat, No. 61.2977. Specimen 9.
Fig. 2. Quadrula rubiginosa (Lea) . Crooked Creek, Rosston, Armstrong County, Aug. 3,

1909. Cat. No. 61.4366. Specimen 217.
Fig. 3. Quadrula undulata (Barnes). Pymatuning Creek, Pymatuning Township, Mercer

County, July 8, 1909. Cat. No. 61.4340. Specimen 15.
Fig. 4. Pleurobema obliquum coccineum (Conrad) . Slipperyrock Creek, Wurtemberg,

Lawrence County, July 4, 1909. Cat. No. 61.4385. Specimen 19.
Fig. 5. Pleurobema clava (Lamarck). Neshannock Creek, Eastbrook, Lawrence County,

June 18, 1908. Cat. No. 61.3335. Specimen 40.
Fig. 6. Elliptio crassidens (Lamarck). Allegheny River, Kelly, Armstrong County,

June 22, 1909. Cat. No. 61.4297. Specimen 52.
Fig. 7. Elliptio gibbosus (Barnes). Neshannock Creek, Eastbrook, Lawrence County,

June 18, 1908. Cat. No. 61.3331. Specimen 57.
Fig. 8. Alasmidonta heterodon (Lea). Schuylkill Canal, Manayunk, Philadelphia County,

April 24, 1909. Cat. No. 61.4250. Specimen 70.
Fig. 9. Alasmidonta undulata (Say). Crooked Creek, Tioga, Tioga County, Aug. 19,

1909. Cat. No. 61.4242. Specimen 76.
Fig. 10. Symphynota compressa Lea. Conneaut Outlet, Conneaut Lake, Crawford

County, May 14, 1908. Cat. No. 61.3297. Specimen 90.
Fig. 11. Symphynota complanata (Barnes) . Conneaut Outlet, Conneaut Lake, Crawford

County, May 14, 1908. Cat. No. 61.3315. Specimen 80.
Fig. 12. Anodontoides ferussacianus (Lea) . Shenango River, Linesville, Crawford County,

Oct. 10, 1907. Cat. No. 61.3127. Specimen 93.
Fig. 13. Anodonta imbecillis (Say). Leboeuf Creek, Waterford, Erie County, Sept. 14,

1909. Cat. No. 61.4178. Specimen 102.
Fig. 14. Ptychobranchus phaseolus (Hildreth). Slipperyrock Creek, Rose Point, Lawrence

County, Oct. 23, 1907. Cat. No. 61.3116. Specimen 218.
Fig. 15. Obovaria circulus (Lea). Crooked Creek, Creekside, Indiana County, May 27,

1908. Cat. No. 61.3269. Specimen 133.
Fig. 16. Obovaria ligamentina (Lamarck). Allegheny River, Kelly, Armstrong County,

Oct. 24, 1907. Specimen 157.
Fig. 17. Plagiola securis (Lea). Ohio River, Industry, Beaver County, Sept. 23, 1908.
Cat. No. 61.3567. Specimen 219.

346

Memoirs Carnegie Museum, Vol. IV.

c

o

13 *^

17

V

21

^>

10

t)0

^

11

14^ 3^^-

15

18

19

•-.

,- «»

22

23

»

Plate LXXXIX.

8 O

2 ^

16

£9*

20 ■

•

to *

24

Glochtdia

ORTMANN: A MONOGRAPH OF THE NAJADES OF PENNSYLVANIA 347

Fig. 18. Proptera alata (Say). Ohio River, Industry, Beaver County, Aug. 29,

1908. Cat. No. 61.3538. Specimen 141.
Fig. 19. Paraptera gracilis (Barnes). Ohio River, Industry, Beaver County, July

11, 1909. Cat. No. 61.4111. Specimen 138.
Fig. 20. Lampsilis iris (Lea). Conneaut Creek, West Springfield, Erie County,

May 23, 1909. Cat. No. 61,4091. Specimen 145.
Fig. 21. Lampsilis recta (Lamarck). Allegheny River, Godfrey, Armstrong

County, July 23, 1907. Cat. No. 61.2935. Specimen 151.
Fig. 22. Lampsilis orbiculata (Hildreth). Ohio River, Industry, Beaver County,

Sept. 23, 1908. Cat. No. 61.3500. Specimen 154.
Fig. 23. Lampsilis ventricosa (Barnes). Little Beaver Creek, Enon Valley, Law-
rence County, May 11, 1907. Cat. No. 61.2144. Specimen 169.
Fig. 24. Truncilla triquetra Rafinesque. Allegheny River, Kelly, Armstrong

County, Sept. 27, 1907. Cat. No. 61.2985. Specimen 178.