8 PROCEEDINGS OF THE NATIONAL MUSEUM vol. 124 



0.677, and the number of lower jaw teeth, 0.686, and standard length 

 were significant above the 99.5 percent level. The regression equation 

 for the upper jaw was Y=0. 62769 X-f- 152.491, and for the lower, 

 Y=0.43656 X+95.149, indicating thus that number of teeth does 

 increase with increase in standard length. 



Distribution. — The known distribution of A. fuscus is given in 

 figure 1. Literature records were included only if we were reasonably 

 certain of the author's identification. The overall distribution, East 

 Africa to Tahiti, is not an uncommon one for marine shorefishes of 

 the Indo-West Pacific. Apparent absences of A. f. fuscus from Indian 

 Ocean island groups such as the Seychelles are unexpected but may 

 well be real. (J. E. Bohlke, et al., recently spent several months 

 collecting fishes in the Seychelles Islands and failed to obtain speci- 

 mens in habitats that would be expected to harbor A. fuscus). In the 

 Pacific, A. f. holomelas appears to be restricted primarily to the 

 southern island chain. The northern island chain (Marianas, Marshall, 

 Phoenix, and Line Islands) have been collected frequently in recent 

 years, and one would expect A. f. holomelas to have been collected 

 there if present. (A. fuscus was reported from Saipan by Fowler, 1945, 

 but this report was based on a misidentification.) Inasmuch as A. f. 

 holomelas has not been collected otherwise from the northern island 

 chain in recent years, we have some doubts as to the validity of the 

 locality record for the two specimens from Boston Island, Marshall 

 Islands. These are old specimens that were originally in the Musuem 

 Godeffroy collections. Perhaps support for our doubts is to be found 

 in the high total dorsal element counts of the two specimens, outside 

 the range for Solomon Islands and New Hebrides specimens (nearest 

 localities to Boston Island for which there are records). It is also 

 possible, however, that the Boston Island specimens represent a newly 

 established or relict population (if the species were more widely 

 distributed in the past) that has tended to diverge as a result of 

 isolation. 



Much of the Pacific distribution of A. f. holomelas is peripheral in 

 the sense discussed by Springer (1967). 



Habitat. — Springer collected Atrosalarias abundantly at One Tree 

 Island, Great Barrier Reef. At that locality the species occurs at 

 depths of less than a meter up to about three meters. It inhabits 

 living or dead coral, most abundantly on the leeward side of the 

 island. Relatively few specimens were taken on the windward side 

 of that island. During the day at least, A. f. holomelas remained well 

 hidden within the interstices of the coral, occasionally darting out 

 and moving short distances to other positions in the coral. That the 

 species is almost entirely black may indicate it is primarily a noctur- 



