HOBSON: ACTIVITY OF HAWAIIAN REEF FISHES 



a number of different nights, and even these did 

 not seem to have a long-standing attachment to 

 such locations. More important, I was unable 

 to recognize attachments to particular locations 

 in most of the diurnal species that occupy resting 

 spots when inactive. There are scattered re- 

 ports of reef fishes returning to established rest- 

 ing spots: Winn and Bardach (1960), for ex- 

 ample, noted that a certain species of parrot- 

 fish in Bermuda has "home caves," and Starck 

 and Davis (1966) suspected the same to be true 

 of two parrotfish species in Florida. Neverthe- 

 less, pending accumulation of additional data 

 from a variety of fish groups the question on a 

 broad scale remains unresolved. 



Nocturnal fishes seeking cover in the morning 

 were not seen in conflict over resting spots. A 

 major general behavioral difference between di- 

 urnal and nocturnal fishes may account for this. 

 Unlike diurnal fishes, which generally are sol- 

 itary in their resting places, many nocturnal spe- 

 cies, including Myripristis spp., Holocentrus 

 spp., and Priacanthus cruentatus, aggregate in 

 their daytime shelters. 



Increased tendency to aggregate in locations 

 more visible to the observer. — The general im- 

 pression of increased activity during the periods 

 of twilight defined above stems mainly from the 

 many diurnal fishes that congregate one to sever- 

 al meters above certain parts of the reef at these 

 times. These fishes, which concentrate over 

 areas where at least most of them seek shelter, 

 mill about actively, but without recognized di- 

 rection. Although surgeonfishes predominate 

 in these aggregations, species from many fam- 

 ilies of diurnal fishes are represented. However, 

 it can be questioned whether or not this pheno- 

 menon does in fact involve increased activity. 

 At least to some extent the fishes only appear 

 to be more active. At this time, compared to 

 other times of day, they are more concentrated 

 and swim in locations that are more visible to the 

 observer. Mostly these are gregarious species, 

 but the groups in which they swim during mid- 

 day are less visible than the twilight assembla- 

 ges, being smaller, more loosely organized, and 

 occurring closer to the substrate at many dif- 

 ferent places on the reef. 



THE INFLUENCE OF PREDATORS 

 ON TWILIGHT ACTIVITY 



I have emphasized that the twilight activity 

 in reef fishes proceeds in an established, well- 

 defined sequence. Now I suggest that this well- 

 ordered series of events is shaped by the threat 

 of crepuscular predators. 



In the Gulf of California (Hobson, 1965, 1968) 

 and certain parts of the tropical Atlantic (Starck 

 and Davis, 1966), large piscivorous fishes are 

 primarily crepuscular. The increased vulnera- 

 bility of smaller free-swimming fishes during 

 twilight, especially those in schools, has been 

 discussed (Hobson, 1968) . Mechanisms that re- 

 duce the threat from predators during daylight 

 and darkness apparently are less eflfective dur- 

 ing the transition between these two major seg- 

 ments of the diel cycle. Thus attacks by large 

 piscivores on schooling clupeids, pomadasyid^, 

 and small carangids in the Gulf of California be- 

 come increasingly frequent at about the time of 

 sunset, and peak about 20 min later. After 

 peaking, the predation ceases, and the piscivores 

 withdraw. Most of the schooling prey, which 

 are nocturnal predators, then migrate to their 

 offshore feeding grounds (Hobson, 1968). Be- 

 cause this activity is a dominant feature of the 

 reef situation in the Gulf of California, one can 

 relate it to the concurrent actions there of such 

 smaller nonschooling fishes as labrids, pomacen- 

 trids, chaetodontids, acanthurids, and balistids. 

 The pattern of cover-seeking in these fishes, and 

 the subsequent emergence of such nocturnal 

 forms as holocentrids and apogonids, follows 

 much the same pattern in the Gulf as that of 

 their close relatives in Kona, described in this 

 report. Observations on the transition events 

 in the Gulf of California were not detailed to 

 the extent of those in Kona; nevertheless, many 

 of the same phenomena were reported from the 

 Gulf. For example, referring to evening obser- 

 vations I pointed out that labrids are among the 

 first of the diurnal fishes there to seek cover and 

 that the emerging Apogon mix with some of the 

 diurnal fishes close among the rocks as the latter 

 are taking cover (Hobson, 1965, 1968). I am 

 confident that the techniques used in Kona, if 



735 



