They were then transferred to another tank and 

 left undisturbed for 1 week. After this interval 

 they were again scored as befoi-e. (Two of these 

 fish died in the interval, so that data from only 

 23 were complete.) 



If the degre« of response is a fixed characteristic, 

 the response should be the same on the second oc- 

 casion as on the first, or if a change has occurred, it 

 should affect all the fish more or less similarly, so 

 that a high degree of correlation should be found 

 between the scores in the first test and the scores 

 in the second. Actually, some fish increased their 

 scores and some decreased them (table 2) ; some 

 changed from a positive response (more than half 

 the time spent in the lighted zone) to a negative 

 one. The correlation coefficient of R = 0.55 was vei-y 

 weak, although it was barely significant. This 

 result can be interpreted to mean that although 

 individuals tend to maintain a certain inherent re- 

 sponse characteristic, this characteristic may vary 

 substantially with time. 



Table 2. — Scores (number of miinulps spent in iHuminated 

 half of tank) for indwidual herring in two 30-mmute 

 tests, 1 week apart (experiment 2) 



First score (by rank) 



Score 1 

 week later 



Minutes Minutes 



25.27 16.18 



24.17 ..__ ia06 



23.18. 16.45 



21.90 19.32 



21.02 19.35 



20.67 20.58 



20.03 14.15 



19.03.. 18.08 



18.98 10.70 



18.62 15.43 



18.13 14.53 



18.08..... 14.80 



17.77. 19.37 



17.73 18.05 



15.82..... 16.78 



15.60 17.98 



14.82 20.85 



14.17. 8.46 



14.08 13.88 



13.05 9 97 



12.48..... ::::::: 7.35 



12.12... 13.20 



12.08 11.03 



EFFECT OF CONDITION ON SUSCEPTIBILITY TO 

 LIGHT ATTRACTION 



There may be many possible reasons for the 

 variability in response among individual fish. 

 Three reasons which readily suggest themselves 

 are differences in sex, physical condition, and age. I 

 did not attempt to determine the effect of sex ex- 

 perimentally because of the need for economy in 

 use of specimens; determination of sex requires 

 killing the fish. Evidence that strong sex differ- 

 ences occur, however, does not show in the fre- 



78 



quency distribution of responses. If males and 

 females differed greatly in response, one would 

 expect some indication of bimodality to the 

 distribution. 



Experiment 3. To determine the effect of condi- 

 tion, 32 herring were tested individually at 15 to 

 17° C, and the time spent in the illuminated zone 

 was recorded for each 30-minute trial. The 

 medium-brilliance, subsurface illumination was 

 reversed end for end every 5 minutes according to 

 the routine procedure. Half of the herring were in 

 excellent physical condition and half were starved 

 and emaciated. The mean scores of 18.1 for the fish 

 in good condition and 18.9 for those in poor condi- 

 tion were not significantly different (F = 0.22^. 



EFFECT OF AGE ON SUSCEPTIBILITY TO LIGHT 

 ATTRACTION 



Two experiments were done to determine 

 whether any differences in attraction to light could 

 be attributed to the age of the herring; the first 

 of these dealt with individual fish, the second with 

 groups of fish. 



Experiment 4. Nineteen herring in age group O 

 and an equal number in age group I were tested 

 individually according to the same procedure used 

 in experiment 3. The mean scores of 20.4 for the 

 0-group fish and 19.5 for the I-gi'oup fish were 

 not significantly different (t = 0.59). 



Experiment 5. Herring of age group O and age 

 group I were tested in groups of 10 individuals 

 according to the second routine procedure de- 

 scribed under "Methods." The scores were based 

 on the number of fish counted each minute for 30 

 minutes in the illuminated half of the tank, which 

 was reversed end for end every 5 minutes. The 

 mean scores for 10 trials with each age group were 

 120 for the O-group and 132 for the I-group fish ; 

 the difference was not significant (F<1.0). 



EFFECTS OF EXTERNAL VARIABLES ON 

 THE ATTRACTION OF HERRING TO 

 LIGHT 



EFFECT OF TEMPERATURE 



The experiments in the preceding section dealt 

 primarily with variables inherent in the fish them- 

 selves. In this and following sections the experi- 

 ments deal chiefly with external variables. 

 Evidence from experiment 1 indicated that tem- 

 perature had little effect on the attraction to light 

 of individual fish. The experiments in the present 



U.S. FISH AND WILDLIFE SERVICE 



.J 



