animal is not greatly altered. Thus Pleurobranchxa and Titiscanici do not differ materially
from their shell-bearing relatives. But sometimes (presumably when the larva discards
the shell before assuming its adult form) the whole organization undergoes a profound
change. This happens in the Nudibranchs and also in the Pterotrachasidse, which are
evidently specialized Heteropods but have become euthyneurous and opisthobranch.
Hermaphroditism is not known to occur in the Cephalopoda, Scaphopoda, and
Amphineura. Among the Lamellibranchiata it is not infrequent and prevails in the
whole suborder Anatinacea. But its morphological importance is not great in this group,
for, fertilization being always external, the organs affected are very simple. But in all
heimaphrodite Gastropoda cross-fertilization is the rule and a complex apparatus is
necessary. I t might be plausibly argued that the prevalence of hermaphroditism in
forms where the shell is absent or small, is due to the fact that they have a greater power
of expansion and of finding room for extra organs. The few hermaphrodite Streptoneura
partially support this view. They are (1) the Entoconchidee, parasites with no shells (2)
two genera of the Lamellariidse, Marsenina and Oncidiopsis (3) Bathysciadium and
Cocculina, both patelliform. There is also some evidence that Patella is hermaphrodite
(4) Valvata which has a small shell and spreads considerably beyond it when extended
(5) Odostomia. The last named and many Pulmonata have a spiral shell so that this
conformation is not a barrier to the hermaphrodite condition, but it attains its greatest
complexity in shell-less forms. I t might have been supposed that, as the presence of two
sets of sexual organs is in itself a considerable complication of the ordinary molluscan
structure, the double apparatus would at least be as simple as possible. But on the
contrary it shows a surprising power of development which in the more specialized forms
results in luxuriant intricacy. Even the Tectibranchs exhibit ^considerable elaboration,
for they may have a prostate, two spermatothecas, a mucus gland and an albumen gland.
But in the Dorids and Ascoglossa in addition to all this the female branch of the genitalia
bifurcates, to one portion being assigned the work of fertilizing the ova, to the other that
of providing them with envelopes and depositing them as eggs. There are three genital
orifices, and besides all this there may be present in one species an armature on both the
male and female ducts as well as two or more accessory glands of unknown functions. Nor
is it easy to see what is the object of this extraordinary elaboration, since the comparatively
simple hermaphrodite apparatus of Acteon seems to serve its purpose equally well.1
In this and in many other points Nudibranchs show great specialization—as far as
we can see, specialization for its own sake. They are one of the extreme branches of the
molluscan tree, springing from the Tectibranchs but not giving rise themselves to any
further offshoot. Yet a general survey of the extraordinary variety of forms which they
offer does not suggest an effort to produce any culminating type but rather an inclination
to give free play to several evolutionary tendencies and to encourage any variation which
is not disadvantageous. The whole lineage of the Opisthobranchiata exhibits this
proclivity to vary and to deviate from what seems at one point of the development to be
its goal. A partial view of the Gastropoda suggests that the long series of successive
modifications has for its object to encase a soft defenceless creeping animal in a strong
The complexity of the nidamental glands, however, is not surprising, for the fertilized ova in
the egg-ribbons of Nudibranchs may receive either singly or in groups as many as four envelopes.
spiral shell, but concomitantly with this arises an opposite tendency to dispense with the
shell, and the representatives of this tendency, though not in the majority, form a very
considerable proportion of the whole class. And just as we cannot say that the line of
gastropod evolution moves towards one type, so the more specialized line of nudibranch
evolution divagates and produces several types which are not successive developments or
improvements but to some extent alternative or the antitheses of one another. When
such alternative or antithetic arrangements occur, one may be more specialized than the
other, but as a rule both appear to be equally effective, and the less specialized apparatus
may be found in forms which otherwise are highly specialized.
Thus it can hardly be doubted that perfoliate or laminated rhinophores are an instance
of great specialization. No complete parallel is found in any other Gasteropods, and,
though many Bullacea have laminated rhinophorial organs, these take the form of laminae
arranged not on columns but on the flat sides of the head. Perfoliation of the rhinophores
is almost universal among the Holohepatica and the less specialized Cladohepatica (Den-
dronotus, Scyllaea, But among the ASolids it seems to be present or absent indifferently,
and this in all sections, for if we can draw distinctions among ^Eolids, Glilamylla
and Bimatella seem to belong to the less specialized, Molidia and Facelina to the more
specialized section. But in the two pairs mentioned the first genus has simple and the
second perfoliate rhinophores.1 Again, branchiae may be present or absent without, it
would seem, making much difference to the animal’s general economy. Thus in the three
pairs Tritonia and Tritoniella, Pleurophyllidia and Pleuroleura, Tethys and Melibe, the
first genus has special gills, the second none. In the first two pairs, the forms with gills
constitute the more abundant and successful type, but more species of Melibe than of
Tethys are recorded. In the same way a prostate gland may be indifferently present or
absent in similar forms which seem to get on equally well with or without it. Thus it is
present in Anisodoris, absent in Archidoris, two equally flourishing and otherwise similar
sub-genera of Doris. The branchise of the Dorids offer another instance, for they can be
either simply pinnate or more extensively ramified, tripinnate plumes being very common.
Both forms seem equally serviceable, and it is not clear why one should change into the
other, except on the assumption that there is a natural proclivity to variation for its
own sake. I t might be supposed that the simpler structure would be found in the less
specialized forms, but the branchiae of such relatively primitive forms as Tritonia and
Bathydoi'is are not simply pinnate plumes but arborescent tufts. And if we were to put
together all the Dorids with simply pinnate plumes they would form a most heterogeneous
assemblage exhibiting all degrees of specialization.2 In some cases (Doiis verrucosa and
many species of Ghromodoris) simple plumes occur in some varieties and scantily bipinnate
-plumes in others.
In this and many other more important cases it is difficult to say in which direction
1 I t is probable that perfoliate rhinophores without sheaths are a doubtful advantage, because
the delicate organs are exposed to many accidents, but this does not explain why the sheaths are
lost, nor why, when they are lost, the perfoliations are sometimes kept and sometimes not. Also
Doto has sheaths and smooth rhinophores.
2 Simply pinnate plumes are recorded in Doris verr&cosa, Rostanga, Holla, Bphxrodoris, Ghromodoris
and its allies, in most Goniodorididse, but among the Polyceridas only in Polycera and Polycerella.