In addition to obtaining quantitative data In the manner described above, 

 the observer also recorded his visual Impression of the strength of the response 

 in one of five categories Indicated as "-", an apparent repulsion; "0", no res- 

 ponse; "X", a weak positive response; "XX", a moderate positive response; and 

 "XXX", a strong positive response. 



(.iuantltatlve data were obtained only with the tunny. Yellowfln were also 

 present In the pond, but no attempt was made to measure their reactions because of 

 their erratic behavior. It might be noted, however, that when they did enter the 

 attraction area during an experiment, their response was similar to that of the 

 tunny except that It was generally less pronounced for a given stimulus. 



Response to Extracts of "Natural" Food Substances 



Although the tunny In Pond No. 5 had been fed exclusively on mainland 

 squid for about 2 months prior to July 1, 1953, before their diet was changed 

 they showed a positive response to extracts of skipjack viscera (#362, but not 

 #357, 358, 363, 368), skipjack flesh (#359, 360, etc., but not #367), skipjack 

 blood (#383), dolphin ovary (#377), anchovy (#384, but not #396), hammerhead shark 

 flesh (#393, but not #373), mainland shrimp (#374, 375), mainland squid (#370, 376, 

 etc.), and marine plankton (#387). The two tunny In the concrete tank, which had 

 been fed exclusively on tuna flesh, showed a positive response to extracts of skip- 

 jack flesh (#378, 389), anchovy (#385), mainland shrimp (#382, 386), mainland 

 squid (#380), but not to an extract of hammerhead shark flesh (#392). 



In general, the response In the pond was weak and extremely variable as 

 compared with that obtained with extracts of tuna flesh, viscera, etc. In the sum- 

 mer of 1952. To produce responses with tuna flesh extract It was necessary to use 

 about five times as much material as was used In the summer of 1952 (about 100 g. 

 as compared with about 25 g. of flesh) . However, the amount used was not as much 

 as was necessary during the spring of 1953 (about 350 g.), when the two surviving 

 tunny were particularly unresponsive (Tester e_t al^. 1954). 



The fact that the squid-fed tunny did respond to such a wide variety of 

 extracted food substances (even to plankton in one experiment) indicated that they 

 were not conditioned to the smell or taste of the particular food which was being 

 fed. However, at times, it seemed that the response of the pond tunny to squid 

 extract was more pronounced than that to extracts of other substances. 



It was decided to undertake a series of experiments involving three test 

 substances (extracts of squid, skipjack flesh, and shrimp), each to be tested on 

 each of three days per week (Tuesday, Wednesday, and Thursday), the tests to run 

 for 3 consecutive weeks. Prior to and during each successive test period the food 

 was to consist respectively of squid, skipjack flesh, or shrimp. Each week one 

 of these foods was to be fed to the fish at 4 p.m. each day from Thursday (after 

 the Thursday tests were made) to the following V/ednesday. The order of testing 

 was to change each day, so that no two tests would appear in the same order on 

 successive days of each week (Latin square design) . 



Two departures from the plan were made. Due to an error, the same order 

 of testing was used In two of the three days in Week 1. A fourth substance was 

 added to the three primary tests, namely, a blend (equal parts) of squid, skipjack 

 flesh, and shrimp extract. Thus two tests were made in the morning and two In the 

 afternoon of each day. The series either started or ended with the blend, which 

 was either preserved or not preserved with sodlvun bisulphite. A summary of the 

 results is given in appendix I (#398 to 409; #414 to 425; and #435 to 447). It 

 will be noted that In each test the same method of preparation and the same amoiint 

 of material was used (100 g. extracted with 2,000 ml. water). 



