B. R. Dickerson, K. W. Brinck, M. F. Willson, P. Bentzen, and T. P. Quinn. 2005. Relative importance of salmon body size and arrival time at breeding grounds to reproductive success. Ecology 86:347–352.

Appendix B. Statistical tables for models used in the main text.

Table B1.  Linear regression demonstrating the variation in fecundity explained by fish length and individual egg mass. 

   Notes: Length and total egg mass were log-transformed (base 10) to account for allometric growth.  Year was tested as a categorical predictor, both individually and for an interaction with length and was not found significant (F_3,143 = 0.76, P = 0.52 and F_3,143 = 0.32, P = 0.82) so was not included in the model. Sample size consisted of 20 females sacrificed prior to spawning in 1997, 50 in 1998, 49 in 1999, and 36 in 2000.

Table B2.  ANOVA tables, parameter estimates, and standard error for general linear models used to analyze male pink salmon reproductive success. 

   Notes: Number of offspring produced included many cases of zero and appeared to approximately follow a Poisson distribution, hence these counts were square-root transformed to achieve approximate normality (Zar 1996: Eq. 13.2).  Fish length, hump depth, and snout length in millimeters were log-transformed (base 10) to account for allometric growth. Year is a factor variable distinguishing between the two years for which offspring numbers were available; Y = 0 in 1997, Y = 1 in 1999. Mode of death is a factor variable indicating whether a male died of senescence (= 0) or bear predation (= 1). Dominance is the male’s mean dominance score, and entry represents the number of days the male entered the spawning ground, counting from the start of the run.

Table B3.  ANOVA tables, parameter estimates, and standard error for general linear models used to analyze female pink salmon reproductive success. 

   Notes: Number of offspring produced included many cases of zero and appeared to approximately follow a Poisson distribution, hence these counts were square-root transformed to achieve approximate normality (Zar 1996: Eq. 13.2).  Fish length, hump depth, and snout length in millimeters were log-transformed (base 10) to account for allometric growth. Year is a factor variable distinguishing between the two years for which offspring numbers were available; Y = 0 in 1997, Y = 1 in 1999. Mode of death is a factor variable indicating whether a female died of senescence (= 0) or bear predation (= 1). Fecundity was estimated with the model described in Table B1, and individual egg mass was the mean size, in grams, of a female’s egg.

LITERATURE CITED

Zar, J. H.  1996.  Biostatistical analysis. Third edition.  Prentice Hall, Upper Saddle River,  New Jersey, USA.