BRAY: INFLUENCE OF WATER CURRENTS AND ZOOPLANKTON DENSITIES 



permanent anchoring point to a buoy within 1 m of 

 the surface. At middepth (6 m), I attached a 

 transect line and swam four transects at a constant 

 speed, one each to the north, east, south, and west. 

 Without moving my head, I counted all blacksmith 

 within my field of view. Fish were tallied in 

 separate columns on a slate according to their size 

 as estimated by eye: juveniles (<125 mm TL), 

 halfgrown fish (125-150 mm), and adults (>150 

 mm). These classes refer to sizes and not neces- 

 sarily to stages of sexual maturity. A complete 

 survey consisted of 16 transects: 4 at each of the 

 four sites on the reef. During the first 26 surveys 

 (9 December 1975-2 November 1976), the length of 

 the transect line was 30 m, so a total of 480 m were 

 traversed each survey. For the remaining 13 

 surveys (9 November 1976-23 July 1977), the 

 length of the transects was doubled to 60 m each 

 (960 m/survey) to see if large aggregations of 

 blacksmith occurred beyond the areas that were 

 initially sampled. I did not conduct surveys when 

 visibility was <2 m. 



I also examined several oceanographic variables 

 at each site on the reef. Water visibility was 

 measured during each transect as the distance at 

 which I could easily discern a fishiike silhouette 

 attached to the line. Water velocities were mea- 

 sured several times at each site by timing the 

 movement of small particles. And surface, mid- 

 water, and bottom water temperatures were taken 

 with a small dial thermometer. 



Plankton Sampling 



Once the movements of blacksmith were deter- 

 mined, I made replicate zooplankton tows at 

 known sheltering sites near the excurrent end of 

 the reef, and known foraging sites at the incurrent 

 end. Incurrent samples were collected at the 

 margin of the kelp bed, over the sand bottom that 

 surrounds the reef. Excurrent samples were taken 

 within the bed, above rocky areas that provide 

 shelter for large numbers of blacksmith at night. 

 The exact location of the sample sites depended on 

 the direction of the water currents (Figure 1). I 

 specifically avoided sampling on days when the 

 current velocity was negligible, when there were 

 obvious eddies, or when the current flow was not 

 along the east-west axis of the reef. 



Plankton were collected between 1000 and 

 1400 h with a 0.5 m diameter 0.333 mm mesh 

 net pushed by a diver. A TSK^ flowmeter, fitted 

 across the net opening, measured the filtered 



volume of water. I randomized (by coin flip) my 

 choice of which end to sample first, thereby 

 restricting such variables as collection time, net 

 clogging, diver fatigue, etc., to random error. Each 

 tow was double oblique, going from the surface to a 

 depth of 6 m, then back to the surface. The diver 

 swam a haphazard pattern through the kelp bed 

 and avoided sampling within 1 m of a kelp plant or 

 the bottom. All samples were immediately fixed in 

 5% buffered Formalin. The time interval between 

 the first and last tows in a collection ranged from 

 1.5 to 4.5 h. 



In the laboratory, samples were split with a 

 Folsum plankton splitter: one-half was used for 

 weighing and the other half was used for counting. 

 For dry weights, samples were filtered (vacuum 

 pressure = 725 mm Hg) onto preweighed GF/C 

 filters, and dried at 60° C to a constant weight. For 

 counting, samples were split two more times, then 

 subsampled with three 10 ml aliquots drawn with 

 a Stemple pipette. The plankton were counted 

 under a dissecting microscope and sorted into 

 broad taxonomic categories. Weights and counts 

 were standardized by conversion to amounts per 

 cubic meter of water sampled. 



I analyzed the data in two ways to compare 

 densities of zooplankton between incurrent and 

 excurrent ends of the reef. First, I compared the 

 individual incurrent and excurrent samples with- 

 in each collection by Mann-Whitney C/-tests to 

 look for significant differences in densities be- 

 tween the reef ends. Second, I compared mean 

 densities between incurrent and excurrent sam- 

 ples of each collection, and tested for incurrent- 

 excurrent differences in these means among the 

 eight collections with Wilcoxon's signed-ranks 

 tests; thus, each collection was a paired (incurrent 

 versus excurrent) observation. 



Foraging Experiments 



To see if blacksmith near the incurrent end 

 consume more prey, I compared gut contents of 

 fish that foraged near the incurrent kelp margin 

 and excurrent shelter sites. I did not compare free- 

 living adults because they were relatively rare 

 near the excurrent end and their major activity 

 might not have been foraging. Instead, I placed 

 individuals in cages located at both ends of the 

 reef. Cages were constructed ofalxlxlm 



^ Reference to trade names does not imply endorsement by the 

 National Marine Fisheries Service, NOAA. 



831 



