Parasitism
61 spiders had reads from entomopathogenic fungi or parasitic wasps (Figure 8); these likely are from parasitized prey eaten by the spiders, rather than parasitism of the spider itself. However, the latter cannot be ruled out. Using both number of ASVs associated with parasites in individual spiders and relative read abundances, there is a significant difference between ginger and native forest sites (Welch t-test; p-value < 0.005). 49 of the 61 spiders had sequences from parasitic wasps, covering 29 ASVs. All but one of the parasites were identified as non-native. 31 spiders in ginger had parasitic wasp reads compared to 18 in native forest. This results in a significant difference, albeit weak, between invaded habitat and native forest (Welch t-test; p-value < 0.05). Braconid wasps, predominantly Rhopalophorusand Cotesia, were most common among forest types. This was followed by ichneumonid wasps, predominantly Ichneumon . Three families of parasitic wasps were detected only in ginger sites; however, these were detected in very few specimens.
Entomopathogenic fungi were identified in 22 spiders in ginger sites and only 1 spider in native forest. There was a highly significant difference between sites (Welch t-test; p-value < 0.001).Beauveria was identified in 16 spiders in ginger, followed byOphiocordyceps in 7 spiders. Most spiders (20 of 22) had a single fungal type. However, Beauveria and Ophiocordyceps were found co-occurring in two spiders in ginger forest. The single spider in native forest was identified with Gibellula , a known arachnid parasitic fungus. This genus was not detected in ginger sites, although it falls in the same family as Beauveria . Infections of spiders with Beauveria have been reported, but this is uncommon (Evans & Samson 1987).