430 Bulletin American Museum of Natural History. [Vol. XIX, 



edge of the block bearing the fish ; the part remaining meas- 

 ures 77 mm., and would have overlapped somewhat the 

 caudal rays. 



Other specimens of the genus found at Hajula are referred 

 to N . hottcB. Some of these are quite certainly such, but 

 others may belong to N. longispinus; being either quite 

 young individuals, or having had the elongated dorsal ray 

 broken off. If the form here called N. longispinus is really 

 N. bottcB, it appears strange that more individuals have not 

 been found at Hakel, Davis's specimen being the only one 

 found there possessing a greatly elongated dorsal. 



Microcoelia dayi, sp. nov. 

 Plate XXXI, Figure 3; Plate XXXIV, Figure i. 



There is in the collection a considerable number of speci- 

 mens of what must be regarded as a distinct species of the 

 genus Microccelia. Of these. No. 4525a (3692) is taken as 

 the type (PI. xxxiv, Fig. i). The total length of this is 98 

 mm.; to the base of the caudal fin 75 mm. All portions of 

 the body are represented, although in places the scales are 

 flaked off, leaving only their imprint on the fine matrix. The 

 length of the head, including the opercular apparatus equals 

 23 mm. The depth of the body, greatest between the pec- 

 toral and the ventral fins, is 28 mm.; but other specimens 

 seem to indicate that this is not natural. From the shoulder 

 girdle to the base of the caudal fin is 53 mm. There are 42 

 vertebra behind the pectoral arch, of which 24 belong to the 

 caudal region. Altogether there are 49 vertebras. Each ver- 

 tebral centrum is strengthened by about 5 longitudinal ridges 

 on each side. The jaws are tightly closed, so that the teeth 

 cannot be observed. The gape is evidently large and con- 

 siderably oblique, the length of the lower jaw being 15 mm. 

 The orbit is large; the diameter of the eye, as shown by the 

 sclerotic bones, is 7 mm. There are 25 or 26 rays in the dorsal 

 fin, including 5 short, undivided ones in front. I count 22 

 supporting interneurals. In the anal fin there appear to be 



