3 | Results
The total marked and recaptured fish, infection parameters, and infection patterns of Salmincola cf. markewitschi during the mark-recapture session are summarised in Figure 2 and Supplement 1.
From June 2020 to July 2020, the hypothesised model constructed by piecewise SEM fitted our datasets well (Fisher’s C = 7.22, p = 0.13; Figure 4a). Both body condition and parasite numbers in the previous month had negative effects on parasite numbers and body condition in the next month, respectively (Figure 4a). Density had no significant effect on body condition (Figure 4a). Fish with higher body conditions exhibited higher growth rates, and growth rates were also affected by body size (Figure 4a).
From July 2020 to October 2020, the model also fitted our datasets (Fisher’s C = 7.12, p = 0.13; Figure 4b). Individual fish that exhibited poor conditions were likely to gain further parasite infections in the post month, but opposite causality was not detected (Figure 4b). As in the previous period, the growth rate was also affected by host body condition and body size in the previous month (Figure 4b). Fish that showed higher growth rates had higher body conditions in the next month (Figure 4b).
The model that included the interaction term of body condition and parasite numbers was selected as the best model showing the lowest AIC value (Table 1, Figure 5), with heavily infected hosts (i.e. fish infected by more than six copepods) exhibiting apparently lower survival rates. The best model predicting capture probability (cp ) included only FL (Table 1).