Table 2.

Statistics for the effects of sampling date, body length (mm) and their interaction on disease parameters

A. Likelihood of infection (Fig. 2c)	Estimate	Standard error	T value	Pr (> |t|)
Constant	− 1.843	0.449	− 4.106	4.5 × 10−5
Sampling date (28 June 2015)	0.432	0.799	0.541	0.589
Body length	0.592	0.176	3.364	0.00081
Sampling date : body length	0.886	0.351	2.521	0.012

B. Time taken to show infection (Fig. 3)	Estimate	Standard error	T value	Pr (> |t|)
Constant	21.491	1.816	11.834	< 2 × 10−16
Sampling date (28 June 2015)	7.830	2.046	3.827	0.00015
Body length	− 4.098	0.690	− 5.937	5.6 × 10−9
Sampling date : body length	− 4.318	0.775	− 5.572	4.2 × 10−8

C. Total number of offspring produced prior to castration (Fig. 4a)	Estimate	Standard error	Z value	Pr (> |z|)
Constant	2.013	1.908	1.055	0.291
Sampling date (28 June 2015)	3.431	2.284	1.502	0.133
Body length	− 0.158	0.735	− 0.215	0.829
Sampling date : body length	− 2.428	0.932	− 2.605	0.0092

D. Probability of producing a clutch prior to castration (Fig. 4b)	Estimate	Standard error	T value	Pr (> |t|)
Constant	5.618	1.273	4.412	1.4 × 10−5
Sampling date (28 June 2015)	− 0.358	2.177	− 0.164	0.869
Body length	− 1.434	0.463	− 3.098	0.0021
Sampling date : body length	− 2.412	1.095	− 2.202	0.0283

A. Quasibinomial regression model fitted to analyse the likelihood of infection (0/1) (null deviance = 904.88, df = 715, residual deviance = 674.57, df = 712). B. General linear model (GLM) for the time (in days) taken to show infection (R-squared = 0.58, F_3,478 = 220.1, p-value ≤ 2×10⁻¹⁶). C. Negative binomial generalised model for total number of offspring produced prior to parasitic castration (null deviance = 126.012, df = 226, residual deviance = 96.052, df = 223). D. Logistic regression model with quasibinomial distribution for the probability of producing at least one clutch prior to parasitic castration (null deviance = 456.58, df = 332, residual deviance = 170.43, df = 329). P-values in italic are statistically significant at the level α = 0.05