36 PROCEEDINGS OF THE NATIONAL MUSEUM vol.76 



the specimen. D. " 'protoculus " does in all the above considered re- 

 spects, particularly with regard to the positions of upper SAO and 

 Pol, agree very well with specimen No. 2146 B. O. C, differing how- 

 ever, by the arrangement of the SAO, which in the latter specimen 

 are found in a gently curved, although somewhat unequally spaced 

 series continuous with the last VO, not in an obtuse angle as in the 

 type of D. " protoculus ". 



The type specimen of D. '•'■nanus " differs from those of D. " theta " 

 and '"'' protoculus'''' in having only 5+5 AO (not 6 AO posteriores). 

 According to Gilbert 1908, p. 225, the VLO should be "half way 

 between lateral line and the base of ventrals " ; in the author's 

 opinion, however, they now seem distinctly closer to the lateral line. 

 PLO somewhat nearer to the pectoral base than to the lateral line. 

 Upper SAO and Pol much nearer to lateral line than to the ventral 

 series of photophores. SAO in an obtuse angle. First AO anterior 

 elevated nearly to the level of the second SAO. Pre well separated 

 from the AO posteriores. In all these respects D. nanus should thus 

 be perfectly concordant with D. mollis Taaning, as already suggested 

 by Taaning when he introduced the latter species. As in the case 

 of the specimens in the Bingham Oceanographic Collection, how- 

 ever, the length of the maxillaries does not show sufficient difference 

 from the measurements of the other specimens such as for instance 

 the type specimen of D. protocuhis (see the table, p. 33) to make the 

 form taxonomically recognizable as a separate species on the basis of 

 this character. 



On the basis of these observations on the type specimens of D. 

 theta, D. protoculus, and D nanus the author can only feel confirmed 

 in the opinion that the subdivision of the rafinesquei-li^e, forms into 

 entirely separate species according to the definitions rendered by 

 Taaning (1918 and 1928) is absolutely impracticable, although such 

 subdivision may possibly be of great value for differentiating 

 ecological races or local subspecific forms within restricted oceano- 

 graphical areas. 



The accurate dimensions of the luminous scale at PLO and the 

 exact natural position of the posterior suborbital organ can not now 

 be made out wdth reliability in any of the above considered type 

 specimens. 



KEY TO THE SPECIES OF DIAPHUS WITH TWO SEPARATE ANTORBITAL ORGANS (ONE UPPER AND 

 ONE LOWER) AND ONE DISTINCT SDPRAORBITAL ORGAN ON EACH SIDE 



The discovery that a distinct and very well developed supraorbital 

 organ, in addition to the upper antorbital, is present on each side not 

 only in D. adenomus Gilbert, 1908, and D. anteorMtalis Gilbert, 1913,=^^ 



=' See Parr, 1928, p. 119, key to the genus Diaphna, Division VI. 



