/ 



FISHERY BULLETIN: VOL. 70, NO. 3 



morning twilight on 18 occasions. Generally 

 over 2 hr were spent on station during each 

 observation period, although the transition 

 events detailed in this report were concentrated 

 into a span of about 1 hr, as discussed below. 

 Before the data presented in this report were 

 collected, general observations in the study area 

 at these times of day had provided a broad un- 

 derstanding of the twilight transition. Also 

 contributing significantly to this understanding 

 were extensive observations made during twi- 

 light in other seas, especially in the Gulf of Cal- 

 ifornia (Hobson, 1965, 1968), where related 

 fishes have similar behavior. With this back- 

 ground and using scuba, I was able to recognize 

 and record what I knew to be significant events 

 and to note the time and other associated char- 

 acteristics. Additionally, some data from ob- 

 servations made at various times of day and 

 night during other segments of the program in 

 Kona are pertinent to the present report and 

 thus used here. Furthermore, interpretation 

 of the twilight activities benefited significantly 

 from these other phases of the overall Kona 

 program. 



Throughout all twilight observation periods, 

 Lloyd D. Richards, stationed on the sea floor 

 nearby, recorded incident light in foot-candles 

 each minute with a photometer facing the wa- 

 ter's surface. In this report, all records of light 

 levels during specific events refer to these read- 

 ings. To minimize instrument error, two pho- 

 tometers of the same model (Weston Ranger 9, 

 universal exposure meter. Model 348)° were al- 

 ternated. The accuracy of the absolute values 

 obtained with these inexpensive instruments is 

 not considered critical; rather, the primary con- 

 cern is that the relative values are consistent, 

 thus permitting comparisons of behavior rela- 

 tive to time and light between one observation 

 period and another. 



The time of specific events are presented rel- 

 ative to time of sunrise or sunset, as calculated 

 for the longitude and latitude of the study area 

 using the Nautical Almanac, U.S. Naval Ob- 

 servatory. An event occurring at about the 



* Reference to trade names does not imply endorse- 

 ment by the National Marine Fisheries Service. 



same time each day relative to sunrise or sunset, 

 will, in time of day, vary by as much as several 

 hours over the course of the year. This fact 

 reduced any bias that might have resulted from 

 subconsciously looking for an event at a spe- 

 cific time. Although I was alert for certain 

 events in sequence, based on familiarity with the 

 general situation, I consciously avoided thinking 

 in terms of time relative to sunrise or sunset 

 during the observations. This was not difficult 

 to do. I did not calculate time of sunrise or sun- 

 set until after a period of observation, and neith- 

 er could have been directly observed, even had 

 I not been underwater at the time; in the study 

 area a large volcano blocks much of the eastern 

 sky, and the western sky generally is obscured 

 by clouds at the end of the day. Furthermore, 

 usually a week or more elapsed between obser- 

 vations of morning or evening twilight, so that 

 the time of sunrise and sunset changed substan- 

 tially from one set of observations to the next. 

 Finally, I did not translate the data from time of 

 day to time relative to sunrise or sunset until 

 the program in Hawaii was completed. 



Obviously all events judged significant were 

 not witnessed during each observation period. 

 Many were evident only in certain locations, and 

 several occurred at about the same time. Con- 

 sequently, the number of times that data were 

 recorded for each event varies widely. 



Names used for the fishes generally follow 

 those given by Gosline and Brock (1960); ex- 

 ceptions occur where more recent taxonomic 

 studies have indicated changes. Most of the 

 fishes are referred to some broad feeding cate- 

 gory, such as being a herbivore, a plankton feed- 

 er, etc. These general designations are based on 

 detailed study of food habits to be reported else- 

 where (Hobson, in preparation). 



In parentheses following the first mention of 

 each fish species are given a mean and range of 

 standard lengths, for example (235: 173-284 

 mm). These figures represent the sizes of spe- 

 cimens of that species that were collected for 

 the food habit analysis (Hobson, in prepara- 

 tion). In collecting these specimens an attempt 

 was made to sample individuals of sizes that 

 showed behavior of adults. Hence these figures 

 provide a good measure of the relative sizes of 



716 



