MOUTH-PARTS OP THE PAL^MONID PRAWNS. 45 



follow, from the fact that these, with the member resembling 

 them which probably repi^esents the apical lobe, number six in 

 many Branchiopoda, that the primitive limb possessed only that 

 quota. In certain cases the series is more numerous. Thus in 

 Anostraca it has probably seven members*, the maxillfeof certain 

 decapod larvse (text-figs. 16, 17) certainly possess eight, and the 

 same limb of Cerataspis (text-fig. 15) bears nine. If, as seems 

 likely, the first member is missing in the maxilla of Calanus (text- 

 fig. 13), there is evidence herealso that the series may consist of 

 eight endites and an apical lobe. It may be that the ancestral 

 crustacean possessed even more endites, but there is some reason 

 for believing that it had eight only, since, as will be shown later, 

 that number allows an arrangement which corresponds with the 

 segmentation of the thoracic limb of Malacostraca, and with what 

 was probably the primitive structure of the maxilla both in 

 Decapoda and in Copepoda. If this view be adopted, there may 

 be recognized in the primitive crustacean limb nine successive 

 regions — eight represented by the endites and a ninth consisting 

 of the apical lobe. These regions would be potential segments, 

 since jointing would certainly take place between the endites rather 

 than across them, as, in point of fact, it is seen to have done 

 in Triarthrus (text-fig. 6), in the larval maxilla of Penaeus (text- 

 fig. 16), and in ISTotostraca (text-fig. 3). In the latter there may 

 be found, immediately behind the attachment of each endite, a 

 distinct articulation. In the case of the distal members of the 

 series, the articulation extends only as far as the longitudinal scle- 

 rite by which the part of the axis that bears the subapical lobe, 

 flabellum, and epipodite is separated from the rest. Just behind 

 the flabellum, however, a very shai-p jointing extends I'ight 

 across the limb, and another less marked articulation I'uns from 

 the base of the epipodite to the inner margin, which it reaches 

 just above the attachment of the blunt-ended proximal endite. 

 Yet another articulation, starting from that just mentioned, 

 joins the inner edge of the limb behind the base of the first of 

 the large, pointed endites, thus cutting ofT a roughly triangular 

 segment which in Apus, but hardly in Lep>iclurus. projects as a 

 small lobe like an endite t. 



The proximal endite — the gnathobase — of the primitive limb 

 may be judged from the evidence afforded by the Branchiopoda 

 and Triarthrus to have differed in shape from the othei'S, been 

 directed towards the mid-ventral line of the body, and served for 

 the manipulation of food. Probably it had a stout, blunt end, 

 set Avith rows of strong, short bristles (text-fig. 27). The 

 remaining endites were, most likely, subsimilar to one another. 



5. It is natural to attempt to derive this limb from the parapo- 

 dium of an Annelid ancestor. If the limb was uniramous, we must 

 seek for analogies among uniramous parapodia, such, for instance, 

 as those of the Syllidfe or the Eunicidte. In the latter family 



* Sec footnote to p. i8. 

 t 8t-'e footnote, p. 48. 



