ARGONAUTIDAE 399 



The mandibles are unlike those of A. mollis (figured by Joubin, 1895, fig., p. 16) in 

 that there is no marked sub-rostral notch. The radula resembles that of mollis in general 

 (Joubin, lac. cit. pi. v, fig. 11) but difli'ers in sundry details. The shape of the first and 

 third laterals are in particular different. These two species differ markedly irova pacificus 

 (Sasaki, 1929, p. 18 \A . pelagicus^ in error) in having bicuspid admedians. 



There was no trace of the hectocotylus, but exploration of the mantle cavity revealed 

 the presence of a penis with a long diverticle. 



Remarks. Verrill (1880, p. 393) included in his definition of Alloposus the words 

 "mantle united firmly to the head by a ventral and two lateral commissures." Hoyle 

 (1886, p. 72) adopted the genus for a "North Atlantic" fragment without questioning 

 the definition. Joubin (1895, 1900 and 1920) does not discuss the latter, though he used 

 the name for various North-east Atlantic fragments. In 1904 (p. 9) Hoyle in dis- 

 cussing the relationship of Bolitaena (= Allopositia) microcotyla does not allude to 

 infundibulo-pallial sutures. 



In 1902, however, Ijima (in Ijima and Ikeda, 1902, p. 87, footnote) commented on 

 Verrill's description and said that the median infundibulo-pallial suture seems to be 

 simply a part of the ventro-median septum (median pallial adductor) which is extended 

 far forwards and passes under but does not join the ventral edge of the funnel. He 

 also explains away Verrill's "lateral longitudinal commissures." 



This feature is not discussed by Berry (1914, p. 286) in his description of the Pacific 

 form of A. mollis, nor does he include it in his family and generic definition. 



Lastly it is ignored by Sasaki (1929, p. 17) in his description of A. pacificiis. Naef 

 (1923, figs, on p. 727 and p. 731) describes and figures this infundibulo-pallial suture, but 

 I think purely on the strength of Verrill's account 1 



It must be admitted that there might be room for confusion here. Thus, we might 

 have grounds for suggesting [a) that Verrill was not likely to be mistaken, {b) that the 

 subsequent forms, most of them obviously fragmentary, identified as Alloposus by 

 Hoyle, Joubin and Ijima^, etc., were not referable to that genus. In any case no one 

 seems to have troubled to examine the type of A. mollis. What grounds have we then 

 for accepting the subsequently described forms as referable to Alloposus and foi 

 modifying Verrill's description in an important feature, viz. the siphono-pallial suture? 



Actually there is singularly little common ground in the description of Verrill, on the 

 one hand, and of the later writers on the other. Thus Verrill figures the hectocotylus, 

 but this organ is certainly unknown in Berry's specimen and those of the Japanese 

 writers. Joubin and Sasaki figure the radula of the East Atlantic and Japanese {pacificus) 

 forms, but it is not known in Verrill's or Hoyle's specimens. Verrill did not describe 

 the funnel-organ. The common factors are the gelatinous tissues, suckers tending 

 to be partly uniserial, wide mantle aperture, deep web and funnel reaching beyond 

 the eyes. We cannot infer from Berry's female specimen if this form has the remarkable 



1 Ijima makes the observation on his A. mollis that "there are two buttons at the siphon base, fitting into 

 grooves on the inner surface of the mantle." Sasaki does not mention this Decapod-hke trait characteristic 

 of the Argonautidae in his review. There is no "stud and socket" adhesive organ in A. hardyi. 



