DeMartini: Annual variations in reproductive traits of Senphus politus 



accurate measurement of egg mass would 

 have been preferable. 



Egg size 



Egg size varied with female body length 

 and period within spawning season, with 

 larger eggs produced by larger females, 

 and all females producing larger eggs 

 earlier in the season (Table 4; Fig. 2). The 

 mean size of eggs did not vary among 

 years, after adjustments for annual dif- 

 ferences in female size and a significant 



Table 4 



Results of two-way ANCOVA testing the effects of female body length (SL, 

 as covariate), study year (spawning season), and bimonthly period within the 

 spawning season on size of ripe ovarian eggs. See Methods for explanation 

 of egg size measurements. Model R 2 (0.510) significant at P<0.001. 



Dependent 

 variable 



Source 



df 



SS 



MS 



Egg diameter 



SL 



Period 



Yr 



Period 



Error 



x Yr 



b 23 



«i 22 



fa. 



CD 



CD 



E 20 

 CO 



b 



O) 



O) 18 



LU 



SMALL' (<135uu SU 



- 165 f SU 



'LARGE' (>165»iSO 



Early Middle Late 



Subseasonal Period 



Figure 2 



Relationship between mean (±1 SE) egg size (diameter of 

 Gilson's-fixed, hydrated oocytes) and bimonthly period within 

 spawning season for female queenfish of three arbitrary body 

 sizes (small < 13.5 cm, medium 13.6-16.5 cm, and large > 16.5 

 cm SL), by sample year. See Tables 4 and 5 for results of 

 ANCOVA testing the effects of subseasonal period (within 

 year) on egg size, with female body lengths evaluated as a 

 covariate. Data were pooled by bimonthly period to increase 

 sample sizes. 



Table 5 



Summary of a posteriori Bonferroni t -tests for identity of year 

 effects detected by ANCOVAs summarized in Tables 3 and 

 4. Means connected by underlines are not significantly dif- 

 ferent at P = 0.05. See Table 2 for values of adjusted means 

 and SEs. 



Variable 



Year contrasts 



In F 1984 < 1985 = 1986 < 1980 < 1979 



Egg diameter 1979 = 1984 = 1980 = 1985 = 1986 

 RGI 1979 = 1984 = 1986 < 1980 = 1985 



period x year interaction (Tables 4,5). The latter in- 

 teraction illustrates that the pattern of subseasonal 

 decline in egg size varied among years (Fig. 2). Ad- 

 justed for female length, mean egg diameter appeared 

 to vary 10% among years (Table 2). This difference in 

 egg diameters, expressed in terms of volume (as 4/3 

 n r 3 , the volume of a sphere), was 35% of the smaller 

 value. 



Condition indices 



Somatic condition varied with body size. Larger 

 females usually were more robust (Table 6), but somatic 

 condition also varied among years; the slopes and in- 

 tercepts of length-weight relations were lower in 1984 

 and 1985 than in the other three years (Table 6). Larger 

 females in particular were less robust in 1984 and 1985 

 (i.e., there was a highly significant In SL x year in- 

 teraction; Table 6), and this invalidated a straightfor- 

 ward interpretation of the effects of body size and year 

 on a summary index of somatic condition. 



As indicated by values consistently greater than one 

 for the exponent "b" in the equation, G = aW b , per- 

 centage gonad-to-somatic weight allocation increased 

 for larger females (Table 2; Fig. 3). However, relative 



