Discussion
By examining the behavior of an agricultural pest in a remote, non-crop setting, we can gain a better understanding of the ecological, behavioral, and physiological plasticity of the insect. First, the high infestation rates (eggs/g fruit) observed in the forested locations suggest these areas are highly suitable to D. suzukiiestablishment. Distribution models for agricultural pests are trained on occurrence data at a regional or global scale, however oftentimes, the available data are collected in a non-random manner. For instance,D. suzukii sampling in the United States has mostly occurred in and around susceptible cropping areas. As demonstrated with these data, a common criticism of presence-only models is that they do not adequately extrapolate to novel areas (Elith and Leathwick 2009, Roach et al. 2017). Models trained on D. suzukii occurrence data from North and South America performed worse in this ground truthing exercise than the model trained on a global data set, suggesting improvements could result from more diverse sampling schemes.
Second, wild blackberries were as or more susceptible to D. suzukii oviposition at all ripeness stages than the cultivated blackberries in this study. In an evolutionary sense, cultivated crops are thought to be more exploitable by insect pests than wild relatives due to human-mediated plant domestication selecting against plant defensive traits (Chen et al. 2015, Whitehead et al. 2017). For instance, bitter-tasting secondary metabolites that deter insect feeding are greatly reduced in domesticated plant species (Wink 1988). Generally speaking, domesticated fruits are also much larger than their wild ancestors, and frugivores tend to prefer larger fruits, lending support to this plant domestication-reduced defense hypothesis. Indeed, femaleD. suzukii laid more eggs into cultivated blueberries than wild ones (Rodriguez-Saona et al. 2019). While we also saw the greatest eggs per berry in cultivated ripe and purple fruit, there were significantly more eggs in wild berries after mass was taken into account. Although we do not know how larval competition affects survivability to adulthood in these natural areas, laboratory studies have shown high D. suzukii egg and larval densities can lower mean survivorship, however host quality mediates this effect (Hardin et al. 2015).
A number of factors contribute to oviposition site selection in herbivores, including previous experience, host condition, competition, and predator avoidance (Jaenike 1978, Futuyma and Peterson 1985, Papaj and Prokopy 1989, Carrasco et al. 2015). The higher oviposition we observed in under-ripe wild berries may be, in part, related to shorter ripening times. Wild blackberries are about a quarter the size of the cultivated ‘Ouachita’ variety sampled here, and exhibited a swifter progression from the blush to ripe stage during collection. ThatD. suzukii can and do develop on a wide variety of host plants and even non-host plants suggests that larval nutritional needs are plastic (Jaramillo et al. 2015, Young et al. 2018, Little et al. 2020).Drosophila suzukii avoid laying eggs in overripe fruit, presumably to avoid interspecific competition, so it is reasonable to expect a large spillover effect into under-ripe berries if a significant determining factor for fitness is competition. The tradeoff to laying eggs into less-ripe fruit may be small if the appropriate nutrients are gathered as the berry ripens during the same length of time as larvae develop. This strategy might not translate to oviposition in cultivated blackberries because the fruit stay in each ripeness stage for a longer period (see Swoboda-Bhattarai and Burrack 2017). In a sense, thatD. suzukii oviposit more readily into wild fruit suggests that non-domesticated blackberries are more exploitable than their cultivated counterpart.
Another consequence of fruit domestication is a change in fruit ripening windows. Cultivated crops are selected to produce fruit where the majority will ripen around the same time to reduce harvesting labor cost, leaving less diversity among ripeness stages for female oviposition selection (Heiser 1988). In cultivated berries, the number of eggs per berry at each ripeness stage over the season were consistently different from each other, suggesting that oviposition was additive at each stage; the number of eggs per berry increased as a function of time and ripeness. Contrastingly, there was no such pattern in the wild berry samples, which is more indicative of a simultaneous rather than sequential infestation.
Finally, we wanted to assess whether any of the observed difference in oviposition among the cultivation types resulted from host preference. In a direct preference comparison, females preferred laying eggs in cultivated fruit when exposed to a single berry of each type, which may indicate a size or surface area preference given the physical disparity between the two fruit types. However, when offered an equal weight, corresponding to a single cultivated versus several wild berries, females laid more eggs in wild fruit. Long-range perception in oviposition site selection relies heavily on visual cues; at a distance, clusters of berries may appear as a single, large fruit, which could explain why our lab preference results differed when the set-up changed. The observed correlation between apparent fruit size and preference agrees with other visual research on D. suzukii attraction and its effects on oviposition behavior (Rice et al. 2016). Additional experiments that eliminate visual cues will be needed to further assess any preference among cultivated and wild blackberry fruits.
The pattern and timing of infestation we observed in wild-growing berries from natural habitats in the eastern United States is consistent with research in Hawaii, Europe and Japan that trapped adult D. suzukii in montane habitats (Ometto et al. 2013, Mueller 2015, Santioemma et al. 2019). In Japan, while most fruit crops were grown below 600m, the majority of D. suzukii adults were trapped at higher elevations (Ometto et al. 2013). The present study supports this idea, finding a variety of susceptible host plants and high levels of infestation in the most abundant resource, wild blackberry. We observed that once fruit began to ripen, sometime between the green and blush stage, the berries were exploited for oviposition by D. suzukiifemales.
Given this information about the presence of established populations ofD. suzukii in remote, montane regions of the southern Appalachian Mountains, several new research questions arise. Within the wooded landscape, D. suzukii may be affecting the local food web by utilizing wild blackberries upstream of other organisms. What effect does this invasive pest have on other blackberry feeders such as birds, bears, or other invertebrates? In terms of agroecosystem impact, do these types of forest populations serve as a potential source for regional migration into crop habitats? No seasonal migration pattern has been established for D. suzukii , though marked adults have been caught at distances in excess of four times their flight capacity (Tait et al. 2018, Wong et al. 2018). The full extent of current and future impacts of D. suzukii in croplands and beyond has yet to be fully realized.