Influence of each host species on cross-species transmission
Some host species were more central in the networks of overlapping epidemics (Figure 4), suggesting they might be more important for the cross-species transmission of particular parasites. The eigenvector centrality values depended on the host species and parasite species. Only three parasite species, P. ramosa, B. paedophthorum, andS. cienkowskii , were sufficiently common to create epidemic overlap networks large enough to calculate centrality values. Of those parasites, cross-species transmission networks of P. ramosashowed some hosts (D. dentifera and D. retrocurva ) were more central in the network (Fig. 4A, Kruskal–Wallis test, p = 0.01), suggesting they may be more important for driving cross-species transmission. For B. paedophthorum , host species also varied in their centrality (Fig. 4B, Kruskal–Wallis test, p = 0.01). However, forS. cienkowskii , host species had indistinguishable centralities (Fig. 4C, Kruskal–Wallis test, p = 0.46). Thus, although each parasite (P. ramosa, B. paedophthorum, and S. cienkowskii ) commonly infected multiple host species, the cross-species transmission dynamics might be different because host centrality patterns differ across parasite species; more specifically, in some cases (S. cienkowskii ) the hosts all seem equally important, while in others (e.g., P. ramosa ) only some hosts seem important for cross-species transmission.
D. dentifera and D. retrocurva consistently had the highest centrality values (Figure 4), suggesting they are more important for cross-species transmission of certain parasite species. On the other hand, the host species that were least central tended to vary depending on parasite species (note letters in Fig. 4, and see Table S1 for a full list of significant pairwise comparisons). D. parvula , D. pulicaria , and Ceriodaphnia were all less important for transmitting at least one parasite species. Besides D. dentiferaand D. retrocurva , no other host species had significantly higher centrality than another. Thus, these two host species might be of particular interest for future work on cross-species transmission in these communities.
Broadly speaking, the three common multihost parasites—P. ramosa, B. paedophthorum, and S. cienkowskii — appear to have different patterns of cross-species transmission. For P. ramosa , two host species (D. dentifera and D. retrocurva ) were likely most important for cross-species transmission because they had higher centrality values on average (Figure 4A). For B. paedophthorum only one host species (D. dentifera ) had higher centrality values on average, while two different species (compared toP. ramosa ) had lower centrality values (Figure 4B). However, withS. cienkowskii , there were no differences across host species, suggesting that host species identity is not an important driver of patterns of cross-species transmission.