rie 1960) and detected no significant differences 

 (X^odf = 15.99, P>0.05). The overall mean 

 selected temperature from pooled data was 10.8°C 

 (SD = 3.1°C), and the mode was 9°C; 709^ of all 

 occurrences were in the range of 8°-13°C (Figure 

 2). The frequency distribution, however, was sig- 

 nificantly skewed towards warmer temperatures 

 (Figure 2). 



Brett (1971) showed that the preferred tempera- 

 ture coincided with the optimal temperature for 

 growth of sockeye salmon, Oncorhynchus nerka. 

 Crawshaw (1977) found that physiological re- 

 sponses are often optimized in a zone of efficient 

 operation rather than at a peak, and temperature 

 preference reflects this. We are not aware of any 

 work on the effects of temperature on the physiol- 

 ogy of speckled sanddabs. 



Considering the work of Brett (1971) and Craw- 

 shaw ( 1977), it is not unreasonable to suspect that 

 speckled sanddabs may have a range of tempera- 

 tures (approximately 8°-13°C) for efficient growth. 

 The skewness may partially be due to activity 

 increasing with temperature that could have re- 

 sulted in occasional excursions into a greater 

 number of compartments (temperatures) than at 

 colder temperatures. Ehrlich et al. (1978) also 

 suggested that skewness of a temperature-specific 

 frequency distribution could result from preferred 

 temperatures approaching lethal limits. These 

 limits are not known for speckled sanddabs. De- 

 Witt (1967) suggested that skewness of distribu- 

 tion could result from the regulation of body tem- 



15-1 



(ij 



o 



z 



UJ 



(T lo- 

 ir 



O 



o 

 o 



o 



5- 



■f-f 



T — I — I — I T T T T T T — r — I — 7~~l — I — I — I — I 

 4 5 6 7 8 9 10 II 12 13 14 15 16 17 18 19 202! 

 TEMPERATURE CO 



Figure 2. — Temperature-specific occurrences of speckled 

 sanddabs, based on pooled data from 2,290 fish observations. The 

 frequency distribution was significantly skewed toward warmer 

 temperatures (^, = 0.571, fgsdf = 2.33, 0.01<P<0.025) but was 

 mesokurtic, that is not overly peaked or flat (^2 = 3.078, f 99 df = 

 1.59, 0.1<P<0.2). 



perature by the animals depending on ". . . detec- 

 tion of deviations from a set point of some rate 

 process bearing a direct exponential relation to 

 body temperature. . . . Since physiological activity 

 bears a direct exponential relation to tempera- 

 ture, the corresponding body-temperature dis- 

 tribution for the animals concerned would neces- 

 sarily be negatively skewed." Speckled sanddabs, 

 however, show a positively skewed mesokurtic 

 distribution. A mesokurtic frequency distribution 

 is not overly peaked (leptokurtic) or flat (platykur- 

 tic). Ivlev and Leizerovich (1960) and Ehrlich et al. 

 (1979) used kurtosis to quantify the strength or 

 precision of the preference: leptokurtic distribu- 

 tion indicates greatest preference. 



The peak abundance of speckled sanddabs in 

 King Harbor occurred from winter through spring 

 and early summer when bottom temperatures 

 were low (Figure 3). During the late summer and 

 fall breakdown of stratification, warm bottom 

 water was associated with a pronounced decrease 

 in the number of fish observed. Stephens et al. 

 (1974) found a similar decrease in abundance of 

 speckled sanddabs during destratification in Los 

 Angeles Harbor. Figure 3 also shows that the 

 lower winter and early spring temperatures dur- 



20-1 



15- 



o 



o 



UJ 



outer 

 O-o stations 



r30 



r25 

 r20 

 H5 . 

 rio' 



; ( 



^5 



T I I I r^ 



I I I I I I I I 



^ 



en 



SONDJFMAMJJASONDJFMAMJJASONDJFM 



SONDJFMAMJJASONDJFMAMJJASONDJFM 



1974 ' 



1975 1976 



TIME (months) 



'1977 



Figure 3. — Bottom temperatures and number of speckled 

 sanddabs per transect at inner and outer stations in King Har- 

 bor, September 1974 to March 1977. 



869 



