(4) In retreating from the light, the fish 

 swims in a zigzag course, the legs of which 

 become progressive!}' more nearly perpen- 

 dicular to the direction of the light, and 

 the fish eventually returns to it ; or the legs 

 become progressively more oblique and 

 the fish escapes into the darkened zone. 



(5) After a short stay in the dark zone, the 

 fish repeats the entire procedure. 



Many of these actions appear to be manifesta- 

 tions of the "dorsal light reaction" (Frankel and 

 Gunn, 1961) in which a fish orients itself in a posi- 

 tion so that its dorsal surface is more or less per- 

 pendicular to the direction of the light rays. 

 Woodhead and Woodhead (1955) have described 

 such a phenomenon in herring larvae. Under nat- 

 ural lighting, where the predominant direction of 

 the light is downward, the herring can orient to the 

 rays in its nonnal swimming {wsture, descending 

 when tlie light is too strong, and rising wheii it 

 decreases. Wlien a single light source is close to 

 the surface or beneath it, phototaxis can take place 

 with no departure from the normal swiiruning 

 post/ure, but dorsal orientation to the light (except 

 when the fish is direx:.tly beneath it) requii-es pos- 

 tures that interfere vnth normal swimming — 

 hence, the apparent disorientation and difficulty in 

 escaping the influence, of the light. 



VARIATION IN SUSCEPTIBILITY TO 

 LIGHT ATTRACTION 



Whatever may be the effect of external condi- 

 tions in modifying the response of herring to light, 

 individual differences among the fish lead to wide 

 variation in this response even under identical ex- 

 ternal conditions. This variability is ~ illustrated 

 by five experiments. 



FREQUENCY DISTRIBUTION OF THE RESPONSE 

 TO LIGHT OF INDIVIDUAL FISH 



Individual fish vary in behavior as well as physi- 

 cal characteristics. If this variation were reason- 

 ably close to a nonnal distribution, the statistical 

 treatment used in this study would be more likely 

 to be valid than if the variation was not normally 

 distributed. The following experiment was made 

 to examine the variation among individaial fish. 



Exiaeriment 1. Ninety herring, selected at ran- 

 dom, were obsei-ved individually to determine the 

 degree to which they were attracted to an imder- 



water light of medium intensity ; 30 were observed 

 at 6° C, 30 at 12° C, and 30 at 16° C. Scoring was 

 based on the amount of time spent in the illumi- 

 nated half of the tank. The mean scores of 16.2, 

 17.2, and 17.8 at 6°, 12°, and 16° C, respectively, 

 were not significantly different (F=0.92, P>0.1). 

 The three frequency distributions each showed a 

 reasonable tendency toward normality (fig. 2). 

 Wlien all the scores were combined, this tendency 

 was more pronoimced (fig. 2, dashed line). 



CHANGES IN SUSCEPTIBILITY TO LIGHT 

 ATTRACTION 



If herring differ from individual to individual 

 in their susceptibility to light attraction, do indi- 

 viduals vary in themselves over a period of time? 



Experiment 2. To answer this question, 25 her- 

 ring were fin clipped to identify individuals and 

 tested and scored as in the preceding exj^eriment. 



6-89 9-119 12-149 15-179 18-209 21-239 24-269 



MINUTES IN LIGHT ZONE 



Figure 2. — Frequency distribution of scores (experiment 

 1) based on the time in minutes out of a ix>ss4ble 30 

 minute.s spent by individual herring in the Illuminated 

 half of a tank. Solid lines, distribution of scores at each 

 of thrtK' temperatures; dashed line, distribution of all 

 scores, regardless of temperature. 



FAOTORS INFLUENCING ATTRACTION OP ATLANTTIC HERRING WO ARfTIFICIAL UGHTS 



77 



