LARSON and DeMARTINI: DISTRIBUTION OF FISHES IN KELP FOREST 



dent species (kelp bass and California sheephead) 

 that contributed most to biomass estimates for fall 

 1979 were not consistently larger or smaller, if dif- 

 ferent at all, at SOK during fall 1 980 and 1 98 1 . Hence 

 we feel that our estimates for fall 1979 are typical for 

 SOK during this season. Furthermore, while species 

 such as kelp bass and sheephead were most abun- 

 dant at SOK-D during fall 1979, this was not always 

 true in 1980 and 1981; the site of greater abundance 

 switched between SOK-U and SOK-D for many 

 species over the period of 1979-81 (DeMartini et al. 

 footnote 6). Thus we also conclude that apparent dif- 

 ferences between SOK-U and SOK-D during fall 

 1979, although perhaps statistically real, are not 

 meaningful for our general characterization of stand- 

 ing stock at SOK. For this reason, we have provided 

 data for the areas separately as brackets for our 

 estimates of conditions in the San Onofre kelp bed in 

 general, and do not specifically attribute the greater 

 abundance of fishes at SOK-D to greater numerical 

 density of giant kelp plants > 1 m tall. 



The surprising aspect of our standing-stock 

 estimates is that they are as large or larger than those 

 of Quast (1968b) in nearshore areas of greater bot- 

 tom relief. Subtracting elasmobranchs, "nonresi- 

 dent" teleosts, and cryptic bottom species, his 

 estimates of standing stock at two sites near San 

 Diego were about 366 kg/ha for Del Mar and 299 kg/ 

 ha for Bathtub Rock. Thus, even though our areas at 

 San Onofre lacked many individuals of such great 

 contributors to biomass at Quast's sites as opaleye, 

 blacksmith, kelp rockfish, and garibaldi, our brack- 

 eted values of biomass were of the same order to 

 nearly twice Quast's estimates. Below, we examine 

 three possible reasons for this perceived disparity: 

 Bias due to sampling methods, bias due to the times 

 and places sampled, and the possibility that there 

 really was a relatively large standing stock of fishes at 

 San Onofre. 



Our sampling methods may have led to over- 

 estimates, or Quast's (1968b) to underestimates, of 

 standing stock. Quast's quantitative collection at Del 

 Mar lacked a wall net, so some fish may have escaped. 

 Although he used transect densities for three of the 

 abundant species in his corrected estimates, his tran- 

 sect method of counting fish to the limits of visibility 

 may have led to reduced searching efficiency (as dis- 

 cussed above). It is less likely that we counted fish in a 

 larger volume than we think. We may have inflated 

 our estimates of integrated abundance by sampling 

 the bottom stratum on different days than the water- 

 column strata, so that the same individuals could 

 have figured into average density in more than one 

 stratum as distributions changed from day to day. 



Such errors would have been most serious in the cos- 

 mopolitan species, and perhaps in large bottom 

 species (like California sheephead) that also 

 occurred in the water column. However, even in our 3 

 m stratum, the average numbers of senorita and 

 white seaperch per transect (uncorrected for 

 visibility) were greater than similar averages 

 obtained by Ebelingetal. (1980a, b) in cinetransects 

 off Santa Barbara, implying that these species really 

 were abundant during the fall at San Onofre. For kelp 

 bass, the average standing stock above the bottom 

 was 48 ± 13 (SE) kg/ha at SOK-U and 148+40 at 

 SOK-D. These values are large fractions of our total 

 respective estimates of about 69 and 183 kg/ha. 

 Similarly, our estimates of sheephead biomass on the 

 bottom alone were 23 ± 5 kg/ha at SOK-U and 75 ± 5 

 kg/ha at SOK-D, compared with our total estimates 

 of about 26 and 112 kg/ha at the respective areas. We 

 conclude that, while sampling problems may have 

 contributed some bias to both our estimates and 

 those of Quast's, much of the difference between 

 Quast's estimates and ours is real, and fish really 

 were relatively more abundant in the areas we sam- 

 pled at SOK during the fall. 



Our selection of sampling times and places could 

 have led to estimates that are somewhat unrep- 

 resentative of conditions in general at San Onofre. 

 Seasonal factors might be involved for some of our 

 "resident" species. Dense concentrations of some 

 fishes (notably white seaperch) may be atypically 

 high at SOK and perhaps other kelp beds during the 

 fall, when these areas are used for breeding. Many 

 species of fish can be found in kelp beds all year, but 

 their abundances might nevertheless fluctuate 

 greatly as individuals move among areas within kelp 

 beds, between different kelp beds, and perhaps be- 

 tween different nearshore habitats. We feel that our 

 samples accurately characterize the standing stock 

 of fishes at San Onofre kelp in the fall, but cannot 

 extend our observations to other seasons. 



Horizontal patchiness in the distribution of fish may 

 also have affected our estimates. Our kelp-forest 

 sampling areas were near the offshore edge of a large 

 area of surface canopy, and fish often were quite 

 dense at the actual edge of the kelp forest. Limbaugh 

 (1955), Quast (1968a), Feder et al. (1974), Hobson 

 and Chess (1976), Bray (1981), and others have dis- 

 cussed this "edge effect". Although many of our tran- 

 sects did not (by chance) sample the edge of the bed, 

 the averages we calculated nonetheless may have 

 overestimated the density of some species through- 

 out the entire bed. However, our estimates of fish 

 density at the particular study areas should be 

 relatively unbiased. Quast's (1968b) Del Mar collec- 



51 



