270 DISCOVERY REPORTS 



The subgenus Spinoserolis, which corresponds to group (i) above, contains the two 

 species S. latifrons and S. beddardi. 



The subgenus Serolis, containing the majority of species, and corresponding to 

 group (iii) above, is further sub-divided into five sections (called groups in Nordenstam's 

 paper). If these sections are compared with the key (p. 278) of the present paper, it will 

 be seen that they correspond very closely to the main subdivisions shown. 



Section I contains the single species S. gracilis and corresponds with B. I. A. of the 

 key. 



Section II, containing the species S. paradoxa, S. schythei and S. polaris, agrees with 

 the division B. I. B. 



Section IV comprises the species S. gaudichaiidii, S. co/ivexa and S. plana, and corre- 

 sponds to the section B. I. G. II. a. 2, to which group also belongs the species S. laevis. 



Section V includes the single species S. carinata. The formation of this group depends 

 on the statement that "the dorsal sutures of all the coxal plates are lacking on all the 

 pereion segments". This statement, however, is inaccurate, for as in the majority of 

 Serolids the terga of the first three free thoracic somites are separated from their 

 respective coxal plates by sutures. Consequently S. carinata should be included in 

 Section III. 



Section III contains the remaining species of the subgenus Serolis: S. irilobitoides, 

 S. septemcarinata, S. antarctica, S. neaera, S. bronileyana, S. polita, S. meridionalis, 

 S. glacialis, S. gerlachei and S. exigua. In the key this section is represented by the 

 subdivision B. I. C, excluding the subdivision B. I. C. II. a. 2; and includes, in addition 

 to the species mentioned by Nordenstam, S. cortiuta and the new species ^S'. kempi and 

 S. elliptica, as well as the transferred species S. carinata. 



Nordenstam places the remaining species in his two subgenera Homoserolis and 

 Heteroserolis. The former is characterized by the fact that "the tergum of the sixth 

 (seventh actual) thoracic somite is coalesced with the first abdominal segment so that the 

 suture-line between this segment and the abdomen has been efi^aced in the middle ". In 

 the latter the central portion of the tergum of the sixth (seventh actual) is entirely 

 missing. 



The definitions of these two sub-genera correspond with those of my groups ii a and 

 ii c respectively, but Nordenstam has no subgenus corresponding to group ii b above, 

 in which the terga of the sixth and seventh thoracic somites are fused with each other 

 and with that of the first abdominal segment so that the suture-lines between these 

 segments have been effaced in the middle. 



The omission of this subdivision has resulted in a certain amount of confusion in 

 Nordenstam's classification, as he has placed one member of this group, S. bouvieri, in 

 his subgenus Homoserolis, and three others, S. australiensis, S. longicaiidata and 

 S. elongata, in the subgenus Heteroserolis. Thus, Homoserolis and Heteroserolis contain 

 certain species which have characters in common, and which do not agree with the 

 diagnostic characters of either subgenus. Since Nordenstam has based his classification 

 on the degree of the reduction of, and coalescence between, the last three thoracic 



