treated profitably only if the characters of all the exotic genera are analysed, a task which
is hardly possible at present, and certainly does not fall within the scope of the present
work. Of the British species which have been called Doi'is, the majority belong to the
Pseudodorididge, and we have only eleven species of true Cryptobranchiatse. Five of these
belong to the genus Doris, of which the type is Doris verrucosa. The remaining six represent
a genus apiece. Aldisa has a tuberculate back, no labial armature, and denticulate teeth of
unusual shape. Gadlina has a granulate back, a labial armature, a radula of denticulate
teeth, with a median tooth occupying the rhachis. Bostanga is covered with minute
hispid tubercles, and has simply pinnate branchiae, a labial armature, and a radula containing
two kinds of teeth. Those in the inner part of each half row are stout and
hamate; those in the outer part are very long, thin, and bifid at the tips. Jomnna is
nearly allied to the tropical Kentrodoris. The back is granulate: the branchiae form a
cup. The anterior margin of the foot is very deeply grooved and the upper lamina is
developed into two ample lobes. The teeth are simply hamate, but the five outermost are
lono- and thin. The other genera, Geitodoris and Aporodoris, offer greater difficulties, and
cannot be regarded as certain. The first has a labial armature; the outer teeth of the
radula are flat, and lie as if they were arranged in bundles. But Doris testudinaria
Tiar a somewhat similar arrangement. Aporodoris millegrana is discussed on pp. 106—107»
and illustrates the difficulty one may have in referring a Dorid to its proper genus.
There remain three smaller groups, Doridopsidge, Phyllidiidge, and Corambidge. The
first two are confined to the warmer seas, and have neither jaws nor teeth, but in their
place a suctorial tube. Except for this peculiarity Doridopsis is hardly distinguishable from
the Dorididas, and might be made a subdivision of them.
The Phyllidiidge are superficially very unlike the Doridopsidge. They are leathery
animals elongate-oval in shape, and often ornamented with brilliantly coloured tubercles.
The branchiae are represented by a line of lamellae running round the body under the mantle
edge. The rhinophores are lamellated and retractile.^ The vent is usually dorsal:
rarely (Fryeria) terminal and under the mantle-margin.
The Corambidae are a very distinct group, but seem to be small in size and numbers.
They are flat doridiform animals with retractile rhinophores, but the gills are represented
by a few simply-pinnate lamellae set at the end of the body under the mantle, oh either
side of the vent. The digestive apparatus appears to be as in the Pseudodorididae. There
is an ingluvies buccalis, a labial armature (though there is some discrepancy in the
accounts of its shape), and a narrow radula (4 + 1. 0. 1 + 4). Only one spermato-
theca has been found. This group seems to bear the same relation to the Phanero-
branchiatae as the Phyllidiidae bear to the Cryptobranchiatae. Each anthobranchiate
class has parallel to it forms with the branchiae under the mantle-margin.
The Claflohepatica, as here understood, form a large group exhibiting considerable
differences oi: appearance and structure, for it includes, the .ZEolidioidea and most of the
Tritonioidea of Pelseneer, as well as the debatable ascoglossan families. All these forms
agree in having the liver divided and generally ramified.1 Very frequently it is contained
1 The singular Pseudovermis paradoxus described by Kowalevsky is apparently a degenerate
A3olid. I t has lost its cerata, and the fiver being obliged to occupy the body-cavity becomes a
compact mass.
wholly in the dorsal papillae. Occasionally (Bornella, Scyllsea, Fiona, Tethys, Pleurophylli-
diidse) there are special branchial membranes of various kinds, but as a rule respiration
is conducted by dorsal appendages containing diverticula of the liver, or when these are
absent, by the whole surface of the body (Phylliroidee, Elysiidge, Limapontiidse). In the
digestive apparatus jaws are nearly always present (except in Hedyle, Tethys, and the
Ascoglossa). The radula is rarely broad (Pleurophyllidiidge, most Janidge, Scyllaea, and
Lomanotus), very often it is reduced to three teeth in a transverse row (Goryphella, Galvina,
etc.), and still more frequently to a single longitudinal series of teeth (all Ascoglossa,
most iEolids). All Cladohepatica have a median tooth1 and it never disappears as in
the Holohepatica. The vent is usually (but with some exceptions) on the right side. The
hermaphrodite gland is usually an independent mass and not a layer over the liver. As
a rule there is only one spermatotheca, but there are two in many Ascoglossa and
none at all in Phylliroe. The genital ducts appear to be diaulic in many genera. But it
is rash to generalize on this point, for it is hard to determine the character of ducts and
openings in small preserved specimens, and many species and genera of -vEolids are
known only by the examination of such specimens. In some species the spermatotheca
is bifid, and though the external duct of the mucus gland (which corresponds to a
uterus) is single, it is often divided by a fold so that the arrangement is practically
triaulic. The arrangement in AElolidia papillosa has already been discussed.
The Cladohepatica show considerably greater variety in shape and structure than the
Holohepatica, and may be divided into twenty-two families as follows:
T ribe I I . Cladohepatica.
Family 1. Doridoeididge (Doridoeides).
2. Pleurophyllidiidas (Pleuropliyllidia, Linguella, Pleuroleura, etc.).
8. Hedylidge (Hedyle).
4. Dironidge (Dirona).
5. Dendronotidge (Dendronotus, Gampaspe).
6. Scyllaeidge (Scyllsea, Grosslandia).
7. Bornellidge (Bornella).
8. Tethymelibidge (Tethys, Melibe).
9. Lomanotidge (Lomanotus, HancocJda).
10. Phylliroidge (Phylliroe, Gtilopsis, Gephalopyge, etc.).
11. Janidge (Antiopella, Janolus, Proctonotus, etc.).
12. Notgeolidiidge (Notseolidia).
13. JEolididge (numerous genera).2
14. Grlaucidge (Glaucus).
15. Fionidge (Fiona).
16. Heroidge (Hero).
17. Dotonidse (Doto).
18. Myrrhinidae (Myrrhine).
, 1 But it is very difficult to see in Lomanotus, though present.
3 I t is highly probable that Galma should be excluded from the ^Eolidiid® and made the type
of a separate family.