portside of the ship, just forward of the propel- 

 ler and below the fishing racks (Figure 3). 

 Four experienced fishermen were in the racks 

 during fishing and were usually the same four 

 men during each cruise. Fishing terminated 

 when bait was exhausted or when the school 

 sounded, stopped biting, or bit too slowly to 

 justify further expenditure of live bait. 



Baiting 



Shad captured in Wahiawa Reservoir by bait 

 seine were transported in portable tanks by 

 truck to Kewalo Basin, Honolulu, and accli- 

 mated to 100% seawater in 34,850-liter (9,200- 

 gal) holding tanks before being placed in the 

 baitwells of the Gilbert . Nehu were captured 

 by bait seine in Pearl Harbor or Kaneohe Bay, 

 Oahu, and placed directly into the baitwells. 

 Nehu were "rested" overnight in the baitwells 

 prior to departure for the fishing grounds. 

 Shad were acclimated to seawater for 36 to 72 

 hr before being placed in the Gilbert's baitwells 

 several hours prior to departure. The shad 

 and nehu used as bait ranged from about 40 to 

 65 mm. 



The unit of bait measure in the Hawaiian 

 skipjack tuna fishery is the "bucket," so named 

 because the fishermen use stainless steel buck- 

 ets to transfer the bait from the bait seine to 

 the baitwells. A bucket is equal to about 3.2 kg 

 of baitfish. However, a bucket of nehu contains 

 more individuals than a bucket of shad because 

 shad are more deep-bodied than the nehu. In 

 general, it was possible to carry a greater 

 number of buckets of shad than nehu in the 

 same sized baitwell because of the greater 

 hardiness of shad. 



Recording of Data 



On Gilbert cruises 105 and 106 the catch data 

 were recorded by the observer on deck using a 

 hand tally, a stopwatch, and a wristwatch to 

 time the start and end of fishing periods and 

 other events. On cruises 108, 109, and 110, 

 data collecting was automated by the use of an 

 event recorder and a camera to automatically 

 photograph skipjack in the fishing area through 

 the windows of the underwater observation 

 chamber. The catch of tuna and other infor- 

 mation concerning fishing were graphically re- 

 corded on a 20-channel event recorder. The 

 event recorder received data from a small 

 hand-held control panel operated by the scien- 

 tist on deck and also data from the automati- 

 cally run camera in the stern observation 



chamber. One channel recorded the time each 

 photograph was taken. Photographs were taken 

 to determine if either bait resulted in a greater 

 concentration of tuna near the hooks. A photo- 

 graph was taken every 10 sec with 16-mm 

 color film in a cinecamera operated in a single 

 frame mode. Shutter speed was 0.03 sec. The 

 film was analyzed by projecting the sequence, 

 one frame at a time, and counting the number 

 of skipjack tuna present in each frame. Only 

 those frames taken during the time fish were 

 actually caught were used to calculate the con- 

 centration of skipjack near the hooks. 



Behavioral Observations 



Observations on the swimming behavior of 

 shad and nehu and their reaction to predation 

 by skipjack tuna were made during fishing on 

 Gilbert cruises 105, 106, 108, 109, and 110, and 

 also during some preliminary trials during 

 1967 on Gilbert cruises 99, 100, and 102. 

 These observations were made from the stern 

 underwater observation chamber and also by 

 scuba divers. The scientist in the stern cham- 

 ber made continuous voice tape recordings of 

 his visual observations of shad, nehu, and skip- 

 jack tuna. The divers observed bait behavior 

 while swimming in the water adjacent to the 

 stern with the ship stopped. On several occa- 

 sions the divers observed shad and nehu from 

 an underwater sea sled while it was towed from 

 the stern of the Gilbert . 



Statistical Methods 



The response of skipjack tuna schools to the 

 pole-and-line fishing method is highly variable 

 and the frequency of various measures of re- 

 sponse is not normally distributed (Yuen, 1959). 

 Nonparametric statistical tests were used to 

 analyze the data because they do not require a 

 normal distribution. The Wilcoxon matched- 

 pairs, signed-ranks test and the Mann-Whitney 

 U-test (Siegel, 1956) were used. 



RESULTS 

 Experimental Fishing 



Alternating shad and nehu among schools.— Re- 

 sults of experimental fishing on cruises 109 and 

 110 are summarized in Table 2. These results 

 are presented first because they account for 

 76% of the schools fished. The basic catch data 

 for each school are given in Table 3. Inspection 

 of Table 2 shows that shad compared favorably 

 with nehu as a live bait, and this is confirmed 



