Impacts of introduced leaf-galling insects on reproduction and seedling survival of Myoporum sandwicense, a native Hawaiian tree

Abstract

Insect herbivores released from biotic and abiotic controls of their native environment can have severe negative impacts on plant reproduction and survival in their introduced range. On Hawaiʻi Island, a recently introduced leaf-galling thrips species (Klambothrips myopori) has infested populations of an abundant native tree called naio (Myoporum sandwicense) causing widespread gall damage and foliage dieback. Mature trees show signs of infestation and have disappeared in some areas, yet the extent to which infestation affects naio reproduction has been unknown. Within two naio populations recently invaded by thrips, one in mesic forest and one in dry forest, I counted flowers and fruits and assessed gall damage and foliage dieback monthly for one year for naio trees with zero (0%), low (<33%), moderate (33%- 66%), and high (>66%) initial gall damage. Gall damage was defined as the percentage of gall-deformed young leaf area. Foliage dieback was defined as the percentage of necrotic leaf tissue compared to total young leaf area. At these same sites, gall damage, foliage dieback and height were recorded for naturally occurring naio seedlings over one year to determine seedling survivorship. I found that naio reproduction decreased, particularly for trees with moderate and high initial gall damage, regardless of site. Reproduction also declined drastically for trees with zero to low initial gall damage at the dry site. I used generalized linear mixed models (glmm) to identify variables that best explained observed patterns in naio reproduction, including thrips-induced gall damage and foliage dieback, tree basal area, precipitation, temperature and humidity. Results from glmms indicated that tree foliage dieback, branch foliage dieback and branch death (precipitated by thrips gall damage) were the most significant variables for explaining naio reproductive decline over time. Thrips’ gall damage and foliage dieback increased for trees with zero to low initial gall damage at the mesic site, while gall damage remained extremely low and foliage dieback was mostly low for trees with zero to low initial gall damage at the dry site. Gall damage and foliage dieback increased for trees with moderate initial gall damage at both mesic and dry sites, while gall damage and foliage dieback were high and then declined for trees with high initial foliage damage at both sites. Naio seedling survival was 34% at the mesic site and 88% at the dry site, but did not appear to be strongly related to thrips damage. At a third experimental site, I used pesticide to exclude thrips and evaluated reproductive differences in treated versus untreated naio trees of low and high initial gall damage classes. Reproduction decreased for all trees at the experimental site, and no significant difference was found in naio reproduction between treated and untreated trees. In the experiment, gall damage and foliage dieback increased for trees with low initial gall damage. Gall damage remained high and foliage dieback decreased for trees with high initial gall damage, but there was no difference in gall damage or foliage dieback between treated and untreated trees. Because I did not see a clear difference in gall damage and foliage dieback between treated and untreated trees, I think it is unlikely that the pesticide treatment was effective in excluding thrips. These results, taken as a whole across the three sites, indicate that introduced thrips have a notable negative impact on naio reproduction, and other factors are also negatively influencing naio reproduction and should be investigated. Also, further study is needed to understand dynamics of naio seedling recruitment and survival. These studies suggest that without management action, the potential for naio to replace itself through reproduction will decline.