LOUGH: TEMPERATURE-SALINITY EFFECTS ON BIVALVE LARVAE 



coefficient, R^, defined as the sum of squares due 

 to regression divided by the total sum of squares 

 corrected for the mean. It is often stated as a 

 percentage, lOOR^. The larger R^ is, the better 

 the fitted equation explains the variation in the 

 data. Values ofi?^ can be compared at each stage 

 of the regression program. A ^-test also is made 

 indicating the equality of the individual regres- 

 sion coefficients to zero and their level of 

 significance. 



The calculated regression coefficients from a 

 particular equation were fitted by computer to a 

 full quadratic equation in temperature and 

 salinity in order to print a contour diagram of 

 the response surface. The computer program was 

 instructed to print 20% contour intervals, wide 

 enough to exclude the approximate ± 10% experi- 

 mental error reported by the authors. Tempera- 

 ture and salinity scales on all plots were set to 

 range from to 40 in order to facilitate response 

 comparison and to allow the overall form of the 

 surface to be visualized. Contours extrapolated 

 beyond the experimental data are given as dotted 

 lines. 



Analysis of covariance methods, as used in 

 Lough and Gonor {1973a, b), were used to test 

 the significance of the difference between the 

 estimated polynomials for early and late larval 

 survival and between late survival and growth. 



RESULTS 



Crassostrea virginica 



Davis and Calabrese (1964) first reared the 

 larvae for 2 days at six levels of temperature 

 and nine levels of salinity to study the effect 

 of these factors on early development, or the 

 period from fertilization to approximately the 

 veliger stage. To learn what effect these same 

 combinations of temperature and salinity had 

 on late larval development, larvae were initially 

 reared from eggs for 2 days at normal seawater 

 conditions (24.0°C, 27.5%o) and then transferred 

 at the veliger stage to the experimental condi- 

 tions. 



Tables of the multiple regression analyses are 

 given in the Appendix and will not be referred 

 to in this section. Survival to 2 days of develop- 

 ment was affected most by the linear and quad- 

 ratic effects of salinity and the linear effect of 

 temperature. Maximum survival of the 2-day-old 

 larvae (80% survival contour) was estimated to 



occur at temperature and salinity conditions be- 

 tween 19° and 30.5°C and 19 and 30%o (Figure 1), 

 which is in good agreement with the experimental 

 results. 



The analysis of survival of 10-day-old larvae, 

 after 8 days of rearing at experimental conditions, 

 indicated that the linear and quadratic effects 

 of temperature and the quadratic effect of salinity 

 significantly affected survival. Maximum survival 

 after 8 days (60% survival contour) was estimated 

 to occur above 21°C and between 8 and 30.5"/oo 

 (Figure 2). The 10-day-old larvae showed a 

 tolerance to much higher temperature and a 

 wider salinity range than the 2-day-old larvae. 

 Analysis of covariance showed a significant dif- 

 ference (1% level) between the 2- and 8-day 

 survival polynomials further substantiating that 

 the range of temperatures and salinities tolerated 

 by the late veliger larvae were significantly 

 different than that of the early embryos. 



Growth of the larvae during 8 days was af- 

 fected most by the interacting effect of tempera- 

 ture and salinity and the quadratic effect of 

 salinity. Maximum growth (100% response con- 

 tour) was estimated to occur at temperatures 

 and salinities above 19°C and 33%o (Figure 3). 



There was a significant difference (1% level) 

 between the polynomials estimated for 8-day 



40 



30 



UJ 



cr 



H 

 < 

 K 

 UJ 



20 



lO- 



10 20 30 



SALINITY ( %o ) 



40 



Figure 1. — Response surface estimation of percent survival 

 of Crassostrea virginica larvae after 2 days of develop- 

 ment at experimental temperature and salinity combinations 

 given in Davis and Calabrese (1964). 



87 



