FISHERY BULLETIN: VOL. 70, NO. 3 



repeated in the Gulf of California, would gen- 

 erate data much like those presented in this re- 

 port. Although the faunas of the two areas have 

 only a few species in common, there are many 

 generic parallels. Genera common to both areas 

 include: Bodianus, Thalassoma, Halichoeres, 

 and Hemipteronotus (labrids) ; Abudefduf, 

 Chromis, and Pomacentrus (pomacentrids) ; 

 Chaetodon and Holacanthus (chaetodontids) ; 

 Acanthurtis (acanthurids) ; Batistes and Sujf la- 

 men (balistids) ; Holocentrus and Myripristis 

 (holocentrids) ; Apogon (apogonids) ; and 

 many others. Congeners from these two widely 

 separated areas behave similarly. 



In the Gulf, one can readily relate the distan- 

 ces small fishes are from shelter at different 

 times during the transition period with the con- 

 current attacks by large piscivores on the school- 

 ing fishes; when this piscivorous activity peaks, 

 most of the small nonschooling fishes are under 

 cover. As I pointed out (Hobson, 1968, p. 84), 

 ". . . during those periods of morning and eve- 

 ning twilight, when schooling fishes are most 

 heavily exploited, neither diurnal nor nocturnal 

 plankton feeders (mostly Chromis, Abudefduf, 

 Myripristis and Apogon) are in midwater." 



Significantly, the reef fishes in Kona all are 

 close to shelter at the time, relative to sunrise or 

 sunset, that schooling fishes in the Gulf of Cal- 

 ifornia are under heaviest attack (Figures "14 

 and *15). At the same time that predation up- 

 surges in the Gulf, the water column is vacated 

 in Kona, both by nocturnal fishes during early 

 morning twilight and by diurnal fishes during 

 early evening twilight. Furthermore, the water 

 column is not then reoccupied in Kona, either 



' When preying on schooling flatiron herring, Hareng- 

 ula thrissina, in shallow water in the Gulf of California, 

 the cabrilla, Mycteroperca rosacea attacks from below. 

 Its charge carries the cabrilla up through the herring 

 school and momentarily out of the water (see Hobson, 

 1968). Because the herring schools are close to the 

 beach throughout the day and into twilight, an observer 

 on the beach can obtain an index of predatory activity 

 by noting the incidence of attacks — seen as cabrilla 

 jumping out of the water amid the herring. These data 

 were collected on seven consecutive evenings and were 

 presented earlier (Hobson, 1968, Figure 18). 



In Figure 14, I have pooled these data, and present 

 the mean values in time relative to sunset (when pre- 

 sented in Hobson (1968) these data were broken down 

 by days, and shown against time of day). The graph 

 also reflects the fact that discrete attacks involve vary- 



o 



z 



3 



-3 



< 



a. 

 m 



< 



z 



=5 



o 





BEGIN KONA OUIET PERIOD 



n. 



ID 20 



TIME RELATIVE TO SUNSET (min) 



— 1 

 30 



Figure 14. — Time of peak predation (Mycteroperca ro- 

 sacea when preying on schooling Harengula thrissina) in 

 the Gulf of California relative to the evening interim, 

 or "quiet," period in Kona. Duration of twilight is 

 similar in the two areas as the difference in latitude is 

 only about 4°. Stippled area of the bars represents the 

 frequency of attacks involving one M. rosacea; white 

 area of bars, the frequency of attacks involving two to 

 seven M. rosacea; hatched areas of bars, the frequency 

 of attacks involving eight or more M. rosacea charging 

 simultaneously. Also plotted is the mean (A) and the 

 range of times, relative to sunset, that diurnal fishes 

 vacated the water column in Kona. (For further in- 

 formation regarding this figure, see text footnote 3.) 



ing numbers of cabrilla charging simultaneously. Un- 

 fortunately, the attacks by cabrilla on herring do not 

 provide an index of predation beyond about 20 min after 

 sunset; at this time, with the cabrilla and other pred- 

 ators still highly active, the herring migrated offshore. 

 Thus the sudden drop in attacks shown in the figure re- 

 flects the herring's departure, not a decline in predation. 

 The predators continued to attack other schooling fishes 

 in the area, mostly pomadasyids, for a short while long- 

 er, and these prey schools did not leave the inshore waters 

 until the predators had withdrawn (Hobson, 1968). In 

 any event, because the attacks by cabrilla on herring 

 provided an index only for the initial upsurge in preda- 

 tory activity, it is meaningless to relate these data to 

 the latter stages of the Kona quiet period. 



* Data on predation by cabrilla on flatiron herring 

 were collected on three consecutive mornings (see Hob- 

 son, 1968). Unlike the evening situation in the Gulf of 

 California, where the herring leave the inshore waters 

 while the cabrilla and other predators are still highly 

 active, in the morning the herring arrive inshore before 

 the predators have begun feeding. Therefore, the at- 

 tacks by cabrilla on flatiron herring in the morning pro- 

 vide an index of predatory activity spanning all of the 

 twilight period, and thus can be related to the entire 

 interim, or "quiet," period in Kona. 



736 



