that P. rhinorhynchus uses its fangs not to feed, 

 but rather to defend its territory. 



Plagiotremus goslinei hovers above the reef 

 during only part of the day. Much of the time it 

 occupies abandoned mollusk and worm tubes on 

 the rocks, and these retreats also serve as resting 

 places at night. In the eastern Pacific, P. azalea 

 uses similar tubes in the same way (Hobson, 

 1968a, 1969). 



CONCLUSION.— P/a^to^remus goslinei is a 

 diurnal predator that feeds on mucus and dermal 

 tissue of larger fishes. 



General Remarks on Combtooth Blennies 



The combtooth blennies are generally regarded 

 as diurnal. For example, Starck and Davis ( 1966) 

 did not see members of the family, known to be 

 present, during many night observations on 

 Florida reefs, and Randall (1967) reported the 

 group to be diurnal in the West Indies. 



Although food habits remain unknown or un- 

 certain for most combtooth blennies, reportedly 

 many feed by scraping filamentous algae and de- 

 tritus from rocks. These items predominated in 

 the diet of all four blenniid species that Randall 



( 1967 ) examined in the West Indies, and in all five 

 studied by Hiatt and Strasburg ( 1960) in the Mar- 

 shall Islands. In Kona, this mode of feeding occurs 

 in Cirripectus variolosus, hni Exallias brevis may 

 be exceptional in feeding mostly on the tissue and 

 mucus of scleractinian corals. The significance of 

 coral mucus as food of £. breuis may relate to the 

 significance of fish mucus as food for blennies of 

 the genus Plagiotremus. Bbhlke and Chaplin 



(1968) suggested that at least some combtooth 

 blennies which scrape algae from rocks may gain 

 most of their nourishment from small organisms 

 living on or around the algae. Clearly, much about 

 blenniid feeding remains unknown. Because these 

 small fishes scrape their food from various sub- 

 strata, their gut contents are difficult to analyze. 

 One can easily see that species of Plagiotremus 

 have a mode of feeding that differs from those of 

 other blenniids, because their manner of taking 

 food is uniquely spectacular. In comparison, dif- 

 ferences distinguishing the feeding modes of other 

 combtooth blennies are relatively subtle. 



Family Acanthuridae: surgeonfishes 



The surgeonfishes are the predominant fishes 

 over most Hawaiian inshore reefs, but this report 



1000 



FISHERY BULLETIN: VOL. 72, NO. 4 



treats only the two species that feed on zooplank- 

 ton in the water column. The habits of these two 

 were only superficially touched on by Jones 

 (1968), who provided a thorough treatment of the 

 many species occurring in Kona that take their 

 food directly from the substratum (see general 

 remarks on surgeonfishes, below). 



Acanthurus thompsoni (Fowler) 



Acanthurus thompsoni (Figure 38) swims in 

 stationary aggregations in the water column 

 above the reef in several locations along the outer 

 drop-off, 20 to 30 m deep. Often mixed with this 

 surgeonfish in these groups are several other 

 species, especially Chromis verater, C. ovalis, and 

 Naso hexacanthus. At nightfall, A. thompsoni 

 descends to the reef below where, inactive but 

 alert, it remains under cover until morning. 



Fourteen individuals (141: 128-185 mm) were 

 speared at different times of day and night. All six 

 that were taken from crevices during the hour 

 before daybreak had empty stomachs, whereas, all 

 seven collected from aggregations in the water 

 column at various times during afternoons had 

 full stomachs, including fresh material. Finally, 

 one solitary individual speared during midafter- 

 noon close among the coral in about 6 m of water, 

 approximately 200 m from the nearest feeding 

 aggregation, had its stomach empty. The seven 

 individuals with material in their stomachs con- 

 tained the items listed in Table 60. 



The data show a strong trend in the diet toward 

 relatively large, semitransparent, and often 

 gelatinous prey. Some planktivorous fishes from 

 other families feed heavily on one or another of 

 these prey, as does the pomace ntrid Chromis vera- 

 ter, which feeds heavily on larvaceans (see species 

 account, above). But in none of these others is the 

 diet similarly dominated by an array of such prey. 

 However, the sparse information on the food hab- 

 its of A. thompsoni given by other authors does 

 not show this trend. Gosline and Brock (1960) 

 reported only mollusk eggs and copepods, whereas 

 Jones (1968) noted copepods, crab zoea, crab 

 megalops, and mysids. But these reports did not 

 indicate how many specimens were examined, nor 

 the relative proportion of each type of prey in the 

 diet. Most important, they did not indicate how 

 much of the gut contents remained unidentified. 

 The major food items that I found in A. thompsoni 

 are types quickly rendered unidentifiable by di- 

 gestion, and thus easily missed if the sample is not 

 fresh. 



