BRAY: INFLUENCE OF WATER CURRENTS AND ZOOPLANKTON DENSITIES 



individual measurements {F-max test, P< 0.001), 

 and for mean velocities among all 27 surveys 

 (variances were: 41.10 incurrent, 16.05 excurrent; 

 P<0.05). These variations were presumably 

 caused by local turbulence produced by kelp and 

 rocky prominences. 



Current velocities were consistently greater at 

 the incurrent reef end. When water flowed from 

 the east, mean incurrent and excurrent velocities 

 were 11.0 cm/s and 4.6 cms (Wilcoxon's signed- 

 ranks test, P< 0.001). When currents were re- 

 versed, incurrent and excurrent velocities were 

 4.9 and 3.6 cm/s (P>0.25); these differences 

 probably would be significant with a larger sam- 

 ple size. 



Water temperatures at the surface and mid- 

 water averaged 15.6° and 15.3° C, and both ranged 

 from 13° C in April to 20° C in October. Bottom 

 temperatures showed a similar pattern, but aver- 

 aged approximately 2° C lower. Temperatures 

 taken in 21 surveys when the current flowed from 

 the east did not differ significantly between reef 

 ends at any of the three depths ( ^tests for paired 

 samples, P>0.25). I recorded temperatures on 

 only two of the four surveys when currents flowed 

 from the west. Surface and midwater tempera- 

 tures both averaged around 15° C while bottom 

 temperatures averaged 13° C. I could not detect 

 differences in temperatures between reef ends. 



Role of Water Currents 



The patterns in water visibility simplified my 

 analysis of the fish counts. I treated each of the 27 

 surveys as a pair of samples, one from the incur- 

 rent and the other from the excurrent reef end. 

 Each member of a pair consisted of a total count of 

 blacksmith from the eight transects taken at one 

 end of the reef (Figure 1). Differences within each 

 pair were then analyzed among surveys with one- 

 tailed nonparametric tests, with the hypothesis 

 that counts of blacksmith are greater at the 

 incurrent reef end. The similarity in water visibil- 

 ity between reef ends automatically standardized 

 the counts for the volume of water that was 

 sampled. Actually, the test of the hypothesis was 

 conservative because visibility (hence the volume 

 of water sampled) was significantly lower at the 

 incurrent end in 6 of 27 surveys. I also calculated 

 for each survey the proportion of blacksmith 

 counted at each end of the reef. Proportions were 

 arc sine transformed for computation of 95% 

 confidence intervals (CI). I checked the preci- 



sion of my surveys on two occEisions by having 

 another diver swim abreast of me and count 

 blacksmith independently. Our total counts of 

 adult blacksmith were similar: 103 (RNB) versus 

 112 (partner), and 254 (RNB) versus 273 (part- 

 ner). Moreover, the surveys seemed to confirm 

 our subjective impressions on the relative abun- 

 dance of adult blacksmith and other fishes in 

 the midwater. 



Blacksmith were the most abundant fish in 

 the midwater and were recorded on every sur- 

 vey. Other abundant fishes in the midwater were 

 kelp bass, Paralabrax clathratus, and senorita, 

 Oxyjulis californica (Table 1). 



Adult blacksmith invariably aggregated at the 

 incurrent end (Table 2). In each of the 27 surveys, 

 more adults were counted at the incurrent than at 

 the excurrent end (Wilcoxon's signed-ranks test, 

 P< 0.001). Doubling the length of transects had 

 little effect on the counts of adults at the excurrent 

 end, but resulted in a 3- to 4-fold increase in counts 

 of adults at the incurrent end. This was because 

 large numbers of adults gathered farther upcur- 

 rent, beyond the area covered in the short tran- 

 sects. The average proportion of adults at the east 

 end of the reef was 0.99 (95% CI = 0.92-1.0) when 

 it was the incurrent end, and 0.09 (95% CI = 

 0-0.32) when it was the excurrent end. 



Adults respond quickly to changes in current 

 direction. On one occasion around midday, the 

 current reversed during a survey and the adults 



Table l. — Relative abundance (percentage of total individuals) 

 and frequency of occurrence (percent of total surveys* of fishes 

 counted in midwater transects at Naples Reef, southern Cali- 

 fornia. Species are listed in order of decreasing abundance. 



833 



