Field infestation
Both ripeness stage and cultivation type had a significant effect on the number of eggs per berry (Fig. 1a; ripeness stage*cultivation type: p < 0.0001, F3, 1951 = 69.72, GLIMMIX). In both years, cultivated ripe and purple berries carried more eggs than wild berries at the same ripeness stage, while infestation in red, blush and green berries were more similar. However, cultivated berries on average were 3-4-fold the weight of wild ones, and after accounting for weight, wild berries contained more eggs per gram than cultivated berries at all ripeness stages (Fig. 1b; ripeness stage*cultivation type: p < 0.0001, F3,1786 = 9.08).
There was a three-way interaction effect of ripeness stage, week, and cultivation type on infestation per gram of fruit over time (ripeness stage*week*cultivation type: p = 0.01, F7,1759 = 2.52, GLIMMIX). Cultivated berries appeared to maintain a consistent infestation pattern throughout the season, with ripe fruit containing the most eggs and the blush stage (least ripe stage collected) having the fewest eggs (Fig. 2a). In contrast, the correlation between infestation and ripeness stage in wild fruit is much less clear (Fig. 2b), even though ripeness overall was a significant factor (ripeness stage: p < 0.0001, F4,1759 = 91.81). We sampled two habitat types (woods and roadside) for wild-growing berries, and unexpectedly the average infestation between the two types were similar (Supp. Table S2).
Focusing only on the 2018 wild fruit samples collected at different elevations, effects from elevation and ripeness stage were each significant (elevation: p < 0.0001, F4,851 = 6.21; ripeness stage: p < 0.0001, F4,851 = 40.48, GLIMMIX), but their interaction was not, suggesting that all elevations were infested to a similar degree (elevation*ripeness stage: p = 0.35, F12,851 = 1.11). There is evidence for differential timing of infestation, as the three-way interaction among all variables was significant (Fig. 3; elevation*ripeness stage*week: p = 0.002, F13,851 = 2.49). Lower elevations had available fruit for collection two weeks before higher elevations. In general, the only significant differences between ripeness stages within each elevation was between the blush and ripe stage, otherwise mean infestations were not significantly different.
To examine differences in the pattern of infestation among sample types, we calculated the percentage of berries that contained at least one egg per sample group. Ripe and purple wild berries and ripe cultivated berries had above a 95% mean infestation across all timepoints (Table 1). Cultivated and wild fruits of the same ripeness stage (cultivation type: p = 0.0625, F1, 55 = 3.61) were not significantly different from each other, although green wild and blush fruits of both types were significantly less infested overall than those at the ripe stage (cultivation type*ripeness stage: p < 0.0001, F3, 55 = 8.62). A weekly breakdown shows near 100% infestation in wild ripe and purple fruits throughout the sampling period, but more variable infestation in fruit at earlier ripeness stages (Supp. Fig. S1). Sampling other wild-growing fruits found near wild blackberry canes revealed a range of infestation patterns (Table S3). Plant species phylogenetically close to known D. suzukiihost plants were more likely to be infested than those more distant phylogenetically.