Table 2. — Times of first spawning in a captive population of 

 Polydactylus sexfilis on nights in 1975 and 1977. Where a single 

 time is given, the egg collector was examined continuously; in 

 other cases, the collector was examined at intervals of 5-15 min. 

 Numbers in parentheses indicate times of peak fish activity, 

 presumed to be the exact time of spawning. 



Disc 



ussion 



Data from a tide gage located at Coconut Island 

 were available for 1974 and showed that spawning 

 nearly always took place on the outgoing tide 

 (Figure 3). Although tide gage data were not 

 available for subsequent years, tides predicted 

 from tide tables showed patterns similar to the 

 1974 tide gage data. For 1975, 1976, and 1977, the 

 time of spawning (i. e., 2030-2130 h) was compared 

 with predicted tides and again found to occur 

 mostly during the outgoing tide. Spawning occur- 

 red on the outgoing tide in 739r of the spawning 

 nights during 1974-77. 



Spawning Beha\ ior 



Observations of spawning behavior were made 

 initially by watching bioluminescence caused by 

 the fish's movement; later, direct observations 

 were made by shining lights on the water at the 

 time of spawning. The level of activity of the fish 

 gradually increased beginning around 2015 h and 

 culminated in the spawning act as determined by 

 the appearance of eggs in the centrifugal pump 

 samples. Occasionally the fish broke the surface of 

 the water during the period of increased activity. 

 During courtship the fish swam rapidly around 

 the net in a circular manner in groups of two or 

 three. They appeared to be chasing one another, 

 and often one fish would contact another from be- 

 hind, either dorsally or ventrally, with snout. 

 Spawning appeared to take place between pairs 

 rather than among larger groups offish. Increased 

 activity usually continued for 20 or 30 min after 

 eggs were first noted. 



The captive population of P. sc.xftlis was clearly 

 spawning with a well-defined lunar rhythm. 

 Other evidence implies that this is a natural be- 

 havior for this species. Lowell (1971) set gill nets 

 weekly in certain shoal areas of Oahu from April 

 to August 1970 and reported that exceptionally 

 large catches of P. se.xfilis per effort occurred "af- 

 ter the full moon and continuing until the last 

 quarter (1 week duration)," and because of the 

 stage of gonadal development among fish in such 

 catches, he termed them "spawning runs." In July 

 1970, female fish caught 3 days before the last 

 quarter all had well-developed ovaries, but fish 

 caught 4 days after the last quarter had spent 

 ovaries. Fishermen seem to have been aware of the 

 habits of P. scxfilis; for a long time: Hosaka 

 (1944:117) stated, "Moon light nights are best for 

 moi (= P. sexfilis) fishing, and this is especially 

 true when the moon is in the last quarter phase." 



Spawning at the time of the last quarter phase of 

 the moon appears to be rare among fishes. Of the 

 50 lunar spawners which Johannes (1978) listed, 

 only two besides P. sexfilis spawned on the last 

 quarter; both of these are species of Amphiprion, 

 and one spawned on the first as well as the last 

 quarter. Since the various species covered in 

 Johannes' list occurred in different geographic lo- 

 cations, the variations in spawning days, taken 

 together with variations in spawning times, could 

 reflect local adaptations such as would occur if egg 

 or larval survival were related to tides or currents. 



The coincidence of spawning in P. sexfilis with 

 the outgoing tide indicates that the remarkably 

 precise timing of spawning may act as a 

 mechanism for offshore dispersal of eggs and lar- 

 vae. Lowell (1971) noted that there was a strong, 

 oceanward current at his sampling site during 

 falling tide, when he estimated spawning occur- 

 red, and results of ichthyoplankton surveys indi- 

 cated that P. sexfilis eggs and larvae are not found 

 in inshore waters in Hawaii (Leis and Miller 1976; 

 Miller et al.»; Watson and Leis»). Johannes ( 1978) 



^Miller, J. M., W. Watson, and J, M. Leis. 1973. Larval 

 fishes. In S. V. Smith (editor), Atlas of Kaneohe Bay, a reef 

 ecosystem under stress, p. 101-105. Univ. Hawaii Sea Grant 

 Tech. Rep. 72-1. Sea Grant College Program, University of 

 Hawaii, Honolulu, HI 96822. 



^Watson, W., and J. M. Leis. 1974. Ichthyoplankton of 

 Kaneohe Bay, Hawaii: a one-year study of the fish eggs and 

 larvae. Univ. Hawaii Sea Grant Tech. Rep. 75-1, 178 p. Sea 

 Grant College Program, Universitv of Hawaii, Honolulu, HI 

 96822. 



903 



