72 DK. G. 0. BOURNB ON THE UANINID^ : 
the second and first maxillipeds. It is to direct this current under the edge 
of the mouth frame into the branchial chamber that the exopod o£ the second 
maxilliped is elongated and modified by being channelled along its ventral 
surface. The modification is carried furthest in Notosceles (fig. 44) and 
Rardna (fig. 42), two forms in which the accessory antennary modifications 
are higlily developed, and in these the exopod has lost its ilagellum. But in 
JYoiopus, in which genus the antennse are most highly specialised for respira- 
tory purposes, the exopod in question (fig. 32) retains a large flagellum, 
and the flagellum is present but small in Lyreidus (fig. 52) and Raninoides 
(fig. 49). This may seem a trivial character, but it gains importance when 
comparison is made with the same appendage in the Leucosiidse. In Philyra 
Icevis for example (fig. 54) the exopod of the second maxilliped is of simple 
shape, shorter than the endopod and convex ventrally. In this species, as in 
all the Leucosiidse, there is a definite exostegal canal covered in below by the 
exopod of the maxilliped : the second maxilliped, therefore, does not lie in 
the course of the incurrent stream and is not modified. I have already 
(p. 54) called attention to the diiference between the first maxillipeds of 
the Raninidse and those of other Oxystomatous crabs. Indeed, as I have 
already indicated when dealing with the endophragmal skeleton, the anterior 
thoracic sterna, and other structures, the Leucosiidse differ in a hundred ways 
from the Raninidse, and I regard the comparison with Lyreidus as wholly 
illusory. The exostegal canal of the Leucosiidse is most probably derived 
from the shallow groove in the pterygostome occupying the same position 
in Cyclodorii^pe. It would be interesting and not wholly irrelevant to 
attempt to trace out the various lines of descent in the families of Oxysto- 
mata other than the Raninidpe, and I have collected much evidence on this 
subject based on a study of the endophragmal skeleton and other anatomical 
features not taken into account by M. Edwards and Bouvier. But it would 
almost double the length of this memoir if I were to make the attempt. 
I have already written enough, and have, I hope, proved my main thesis that 
the Raninidse cannot have descended, by way of the Dorippidas and the 
Dromiacea, from the Macrura. They have originated independenth^ from 
the last named, and must therefore be separated from other Oxystomes as a 
separate tribe Grymnopleura. I have endeavoured to show that the numerous 
brachyuran features exhibited by this tribe must be explained by their having 
adopted the same habits as other crabs, and that such modifications as the 
redaction of the abdomen ; the loss of the uropods ; the adherence of the 
branchiostegite to the thoracic epimera (exhibited, however, in a quite 
peculiar manner in the Gymnopleura) ; the expansion of the ischium and 
nierus of the third maxillipeds; the broadening of the front, with which 
is correlated the formation of orbital cavities and the reduction of the 
rostrum ; the great development of the muscle-cavities of the thorax, — are all 
characters of high survival value in decapods which have acquired tlie habit 
