Dietary composition
To further explore dietary differences between ginger and native forest, samples were assessed by sampling event, defining a site using the block and transect number, rather than the diet of each individual spider. Incidence and Hellinger-transformed relative read abundances were used to calculate beta diversity using ASV, and order-level identities. Beta diversity values were highest between the sites in ginger and sites in native forest when using incidence data, indicating compositional differences (Figure 3). Ordination space overlapped between ginger and native forest sites, most noticeably when using ASVs (Figure 4); using ASVs, high dietary dissimilarities existed across all sites, both within habitat types and between habitat types (Figure 3), which resulted in less clear differentiation. Hellinger transformed read abundance produced a significant difference between ginger sites and native forest sites when using ASVs but not when using orders. This is likely because dietary communities across ginger and native forest sites showed more similarity at the order level while other sites in ginger habitat were extremely different. However, using beta diversity calculated from incidence data, there was a significant difference between ginger and native forest sites across both ASVs and order (PERMANOVA, p-value < 0.05; Figure 4b, 4d). This provides interesting implications for the different ecological patterns detected with and without abundance information. In our case, a high similarity in a handful of sites across both habitat types is detected when using relative read abundances because certain orders are being eaten in similar quantities; without abundance, rarer orders are having more of an influence.
Another interesting finding is that there is high compositional turnover in the diets of spiders within ginger habitat when we look at order-level diversity; this is reflected in the beta-diversity values (Figure 3) and in the NMDS plots, where the polygon encompassing ginger sites occupies a larger portion of the ordination space (Figure 4). The compositional differences in the diets of spiders from ginger sites are due to the higher diversity of prey consumed, with four orders (Hemiptera, Lepidoptera, Diptera and Entomobryomorpha) detected most commonly (Figure 5). The diets of spiders in ginger are not dominated by any one order, with the most common prey (Hemiptera) detected in the diet of 43.6% of ginger spiders followed by Lepidoptera in 42.3% of spiders and Diptera in 37.1% of spiders; the non-consistency of the diet across spiders in ginger is the cause of increased beta diversity and wider polygons. In contrast, spiders in native forest are consuming more similar diets consisting predominantly of Hemiptera, with 72.4% of spider diets containing Hemiptera. The second most common order, Lepidoptera, only occurs in 36.2% of spider diets. Spiders in native forest are not predating Entomobryomorpha (Collembola) or Coleoptera at the same rate as ginger spiders. In fact, Entomobryomorpha was the fourth most common prey group in ginger, detected in 27 spiders (34.6%) while Entomobryomorpha were only detected in 1 spider (1.7%) in native forest sites (Figure 5).
To further understand diets between native forest and ginger forest, we examined phylogenetic beta diversity within the most common orders, Hemiptera and Lepidoptera. Here, we were interested to see if spiders were eating the same taxa within commonly detected orders. Because of low BLAST identification, phylogenetic beta diversity using 16s within orders was used to allow us to detect any finer scale taxonomic similarity that may exist. We find large overlap between the Hemipteran and Lepidopteran prey in both native forest and ginger (see S6 in Supplemental Information). Hemipteran in the family Cicadellidae and Lepidoptera in the families Geometridae and Hesperiidae were most detected across both habitats. In native forest, more spiders were detected consuming family Cicadellidae and family Hesperiidae (See S7 in Supplemental Information. There were three Hemipteran families and six Lepidopteran families detected in ginger that are were found in native forest; however, these were detected in the diets of very few spiders.