8 "TERRA NOVA" EXPEDITION. 



still logical aud simple, becomes inadequate to support the weighty conclusions that 

 must be based on it. 



At this point Prof. Bouvier attributes to Prof. Carpenter and myself an opinion 

 that I, at least, do not hold. He writes : " Au surplus si, comme le pensent 

 M. Carpenter et M. Caiman, la paire de pattes posterieures est une paire surajoutee 

 dans les types decapodes, les orifices sexuels des Pentapycnon devraient se trouver a la 

 meme place que chez les Pycnoyonum, a savoir sur les pattes de la quatrieme paire, 

 alors qu'ils sont situes sur la cinquieme." He proceeds to argue that the somite which 

 has disappeared in the octopodous forms is not the fifth but the fourth, on the ground 

 that the dorsal tubercle corresponding to this somite in Pentapycnon persists in 

 Pycnogonum although the somite itself has disappeared. Clearly, however, this 

 evidence might be read in another way. Instead of assuming a transference of the 

 dorsal tubercle from the penultimate somite of Pentapycnon to the last somite of 

 Pycnogonum, we might take the fourth pedigerous somite as homologous in the two 

 genera, and assume a transference of the genital apertures from the fifth somite to 

 the fourth. As a matter of fact, however, there is no evidence at all for the existence 

 of individual homologies between the somites of the two genera. Bateson pointed 

 out long ago the fallacy of the assumption that in variation the individuality of each 

 member of a meristic series is always respected. In writing of " an additional pair 

 of legs " I had not in mind any particular one of the five pairs. There is nothing 

 to prevent us from regarding the series of somites as having been remodelled as a 

 whole in passing from one genus to the other. 



In support of the contention that " the constancy in the number of somites and 

 appendages throughout the comparatively wide range of structure presented by the 

 eight-legged Pycnogons strongly suggests that this is the deep-seated and, so to speak, 

 ' normal ' plan of structure of the group " from which the ten-legged condition is a 

 secondary departure, I called attention to the parallel case of Polyartemia among the 

 Branchiopod Crustacea. Polyartemia difli'ers from the normal type of the Order 

 Anostraca, to which it belongs, in having nineteen instead of eleven pre-genital trunk 

 somites ; and since the number appears to be constantly ten or eleven in the other 

 Orders of Branchiopoda (excluding the abbreviated Cladocera), there seems to be good 

 ground for suggesting that the increased number in this case is due to secondary 

 specialisation. Prof. Bouvier quotes against me the authority of Dr. E. von Daday 

 (1910, p. 411), who considers Polyartemia to be the most primitive of the Anostraca. 

 I find nothing in Daday 's discussion of the question to lead me to change my opinion. 

 He makes no mention of the position of the genital opening in comparing the 

 Anostraca with the other Orders of Branchiopoda ; and his reference to the supposed 

 persistence of a vestige of the mandibular palp in Polyartemia overlooks entirely the 

 fact that the palp is in all cases present in the nauplius. 



It would be easy to multiply parallel instances from other groups of the animal 

 kingdom, but, as Bouvier reminds us, " ii ne convient pas d'etendre a un groupe les 



