568 W. H. LONGLEY 



passing, it is worthy of note that there is a marked difference in 

 the ratio of full to fasting specimens in N. apodus and N. griseus 

 taken at 5.00 a.m. Empty stomachs are 3.6 per cent of the 

 total examined in the former, while the proportion rises to 16.7 

 per cent in the latter. This may indicate that in nocturnal fishes 

 irregularity in feeding is correlated with failure to conform strictly 

 to the custom of schooling in closely restricted areas by day. 



There are also obvious lines of cleavage among the diurnal 

 fishes. Some are surface feeders from whose environment the 

 bottom colors are so remote as to be practically non-existent. 

 Others, such as Microspathodon chrysurus and Eupomacentrus 

 fuscus, haunt the coral stacks, swim freely in and out of their 

 interstices, but seldom venture more than a yard or two away. 

 Finally, there are the wandering Labrids, Scarids, etc., which 

 differ from one another both in the horizontal and vertical 

 limits of their range. Knowledge of these distinctions must 

 precede appreciation of the correlation of color with habit which 

 prevails among reef fishes. 



Two lists are displayed in table 3. The first includes species 

 which may be observed moving freely in the mixed environment 

 of the open reef. The second comprises others whose diurnal 

 activity seems definitely centered in the coral stacks. Within 

 the limit of present experience each is complete, and is ap- 

 proximately correct, though later modification may appear 

 desirable. Hence it is instructive to analyze both with reference 

 to the distribution of sandy gray among their members, since the 

 two environments differ widely in the proportion of this color 

 they contain. 



The three blue-gray species differ sufficiently from the others 

 to justify their exclusion from consideration in the present 

 instance; yet to have passed them over without comment might 

 have led to misunderstanding. Taking the remaining records 

 as they are, we may say, then, that the color of three of 21 species 

 in series II repeats that of sandy bottom. But in series I, which 

 is only 50 per cent longer, there are almost five times as many 

 which show evident adaptive color changes under the same con- 

 ditions, or have permanent gray markings. 



