Figure l. — Location of field sampling 

 stations for speckled sanddabs in King 

 Harbor, Redondo Beach, Calif 



PACIFIC OCEAN 



300 m 



Methods 



We collected adult speckled sanddabs between 

 August 1975 and January 1976 with a 3-m otter 

 trawl. The fish were transported to the laboratory 

 in aerated seawater and acclimated to a range of 

 normally occurring temperatures (10.0°-19.7°C) 

 according to the methods of Ehrlich et al. (1979). 

 Prior to acclimating the speckled sanddabs, we 

 removed the gill isopod Liuonica vulgaris indi- 

 vidually with forceps. During holding and accli- 

 mation, we fed the fish to satiation daily with live 

 and frozen Artemia salina. The behavioral re- 

 sponses of the fish to temperature were studied 

 using a 3.6-m long horizontal gradient and 

 employing the techniques of Ehrlich et al. (1978). 

 Each experiment lasted for 7-8 h with observa- 

 tions every 15 min. We shifted isotherm positions 

 during each experiment to separate selection of 

 temperature from preference for a given position 

 within the experimental chamber. 



Speckled sanddab abundance and distribution 

 were studied using timed diver transects at six 

 stations (Figure 1). Two divers swimming side by 



side for 5 min traversed each 6-m wide transect. 

 They recorded the species and number of indi- 

 viduals observed in the same area. The transects 

 by each pair of divers were run in duplicate on a 

 monthly basis at each station from September 

 1974 through February 1976 and quarterly there- 

 after. In the analyses, we used the largest number 

 of individual fish counted by either diver, but the 

 average count of the two independent observa- 

 tions was used for estimates of large groups of 

 fishes. The divers recorded the temperature at 

 least twice during each transect, with thermome- 

 ters readable to 0.5°C. 



Results and Discussion 



We examined the effects of acclimation temper- 

 ature, size, and sex of speckled sanddabs on their 

 temperature selection during 11 experiments (Ta- 

 ble 1). The presence of some skewed temperature- 

 specific frequency distributions (Table 1) pre- 

 cluded comparison of the results with parametric 

 statistics. We tested these distributions for homo- 

 geneity using a Kruskal-Wallis test (Steel and Tor- 



TABLE 1. — Temperatures selected by speckled sanddabs in laboratory experiments. 



P<0.05. 



868 



