294 



BULLETIN 100, UNITED STATES NATIONAL MUSEUM. 



Nor are the relative numerical proportions of the two classes to 

 each other any more significant. With the evidence at his command 

 Vanhoffen had every reason to use the number of otoliths to each 

 otocyst as a systematic character. But facts afforded by the present 

 specimens show that it can not be given so much weight. Thus in 

 the larger Philippine example I have found otocysts with 1, 2, 3, 

 and 4 otoliths, usually two large and the others small; and in the 

 smaller one, instances with 1, 2, and 3. The only character which 

 stands the test of critical examination, as I have previously pointed 

 out, but which Vanhoffen (1911) has not employed at all, is the 

 relation between the number of otocysts and that of tentacular organs, 

 both large and rudimentary. The recorded numbers are as follows: 



virens Bigelow 



virens Maas 



Philippine specimen 



rribcngha Agassiz and Mayor 



mbengha Vanhoffen 



carolinae Mayer 



Otocysts. 



32 

 32 

 38 



32 

 4S 

 64 



We have every reason to assume that the numerical conditions in 

 the Atlantic carolinae described by Mayer (1910) are fairly con- 

 stant in the adult, for he has found it very abundant on two occa- 

 sions at Charleston, South Carolina, and has also taken it at the 

 Tortugas. The table, then, shows that in all the Indo-Pacific speci- 

 mens, irrespective of locality or of exact stage of development, there 

 are many more tentacular structures than otocysts, often twice as 

 many. On the other hand, in all the Atlantic examples of which we 

 have any account the number of otocysts is as great as that of the 

 tentacular structures. We have here something tangible. 



It is, of course, possible that further studies may reveal speci- 

 mens connecting the two types; but the same possibility is present 

 in the case of every difference which could be used us a specific 

 character. We have no right to assume that this will happen. On 

 the contrary, when we find that all the evidence yet available points 

 to discontinuity of the two forms, and when we find that the 

 Atlantic — that is " carolinae " type has never been found in the 

 Indo-Pacific, although the genus has been recorded there from several 

 widely separated localities, there is nothing to do but to make it the 

 basis for specific diagnosis. 



