24 



Fishery Bulletin 91 (1), 1993 



fers were discontinued by day 15, when larvae were 

 approximately 4-6 mm long. Dry food supplements 

 were provided for larger fish, and Artemia densities 

 were diminished so that juveniles eventually subsisted 

 entirely on dry food. These and other details of rearing 

 followed Holt etal. (1990). 



Experimental design and protocol 



Three sizes of red drum were studied. Mean (±SD) 

 total lengths for the small, medium, and large size- 

 classes were 8.7 (±0.9), 22.8 (±2.3), and 34.0 (±2.6) 

 mm. The lower acclimation temperature (21°C) is well 

 below optimum for red drum larvae (Holt et al. 1981), 

 so those reared at that temperature grew more slowly 

 and exhibited higher mortality rates during the first 

 week than did the 26° C fish. Trials were conducted at 

 both nominal temperatures, yielding four treatments. 

 Trials on fish placed in water of the same temperature 

 as their rearing tank (acclimation temperature) are 

 termed 'high' (26°C) or 'low' (21° C) controls. Trials at 

 a temperature differing from the acclimation tempera- 

 ture are referred to as 'upward' (21° to 26° C) or 'down- 

 ward' (26° to 21° C) transfers. 



For each trial, fish were transferred individually from 

 the rearing tank to separate transparent experimental 

 arenas and left undisturbed for the duration of the 

 observations. Arenas were rectangular from above, with 

 sides in a ratio of 5:3. Arena sizes were scaled such 

 that the longer dimension of the rectangular surface 

 area was 6 to 8 times the average total length of the 

 fish. Water depth was 4.6 to 7.5 times the greatest 

 body depth of the fish (3.5 to 7.5 times depth with fins 

 expanded). Small fish were pipetted individually, while 

 larger fish were released from 100 mL beakers con- 

 taining a single fish in 50 mL of water from the rear- 

 ing tank. 



Behavior was recorded on videotape through a cam- 

 era mounted above the arenas. The recorder was acti- 

 vated from about 1 min prior to transfer until 20 min 

 after transfer, then for 5-min periods at intervals in- 

 creasing from 15 to 60 min. In all, behavior was quan- 

 tified during 19 5-min observation periods beginning 

 5, 10, 15, 30, 45, 60, 75, 90, 105, 120, 135, 165, 195, 

 225, 255, 315, 375, 435, and 495 min after transfer. 

 Each of the four treatments was applied to six fish. 



Temperatures in rearing tanks were controlled by 

 balancing air temperature with submersed aquarium 

 heaters, and were slightly above nominal levels. Mean 

 acclimation temperatures during the 10 d preceding 

 each trial were 21.3-21.7° C and 26.1-26.5°C. Tem- 

 perature in the experimental arenas was maintained 

 by controlling room air temperature. This sufficed for 

 trials with medium and large fish; however, arenas for 

 small fish were placed in a larger water bath to stabi- 



lize their temperature. During the 8h trials, the arena 

 temperatures were 20.7-21. 5°C and 26.1-26.5°C. By 

 design, control fish were to experience no temperature 

 change, and upward and downward transfers were to 

 experience ±5.0° C. Actual differences between rearing 

 tank and trial temperatures were -0.8° to +0.2° C for 

 controls, +4.5° to +5.0° C for upward transfers, and 

 -4.7° to -5.7° C for downward transfers. 



In quantifying behaviors relevant to foraging, we 

 recognized the dichotomy of searching techniques in 

 fishes that travel in search of food and locate prey 

 visually. Some species actively search while swimming 

 and are called 'sweep' searchers (Laing 1938, O'Brien 

 et al. 1986). Others search during brief periods when 

 swimming is interrupted (described by Janssen 1982, 

 and termed 'saltatory' searchers by O'Brien et al. 1989). 

 Both types of active, visual foraging have been sug- 

 gested for young fishes. For example, larval Atlantic 

 herring Clupea harengus are thought to be sweep 

 searchers (Rosenthal 1969, Rosenthal & Hempel 1970), 

 while larval white crappie Pomoxis annularis exhibit 

 saltatory searching behavior (Browman & O'Brien 

 1992). Bell (1990) suggested that the saltatory style is 

 more generally employed by teleosts. 



We examined three measures of foraging behavior: 

 activity, pause frequency, and mean pause duration. 

 Activity, the total amount of time spent swimming 

 during each 5-min period, is a measure of foraging 

 effort for sweep searchers, since they search new ter- 

 ritory while swimming. Saltatory searchers use peri- 

 ods of inactivity (pauses) for finding food. Therefore, 

 the frequency of pauses and their duration per obser- 

 vation period are indices of the time spent scanning 

 the environment. 



A BASIC computer program (available from the se- 

 nior author) was written to act as an event recorder. 

 Data were obtained by replaying the video tapes at 

 normal speed and making the appropriate keystroke 

 each time the subject's behavior changed (swimming, 

 pausing). The program recorded intervals between key- 

 strokes from which all variables were calculated. All 

 observations from every trial ( 19 time-periods x 6 rep- 

 licates x 4 treatments x 3 sizes = 1368 5-min obser- 

 vation periods) were made by the junior author. When 

 a selection of observation periods was reanalyzed, the 

 differences in activity averaged 4.1% (extremes 0.1- 

 9.2%) of the combined mean. 



Data analysis 



Since observations within each temperature treatment 

 and size-class followed the same individuals over time, 

 repeated-measures multivariate analysis of variance 

 (MANOVA) was used to identify effects on spontane- 

 ous behavior due to fish size, temperature, and time 



