Reproductive studies in invasive plants are necessary for an understanding of their potential to establish and spread in foreign environments. Elaeagnus umbellata Thunb. (autumn olive) is an invasive woody shrub that flowers early in the spring and is often noted for its abundant fruit set. This study examined the reproductive biology of E. umbellata in Illinois, where it is highly invasive. Hand-pollination experiments were performed to determine the breeding system of E. umbellata, and floral visitors were collected to determine its pollinators. Experiments showed that E. umbellata is a predominantly outcrossing species with a self-incompatible breeding system. However, individual variation was detected in several reproductive characteristics. Pollen tube analyses revealed that a small percentage of individuals allow successful self-pollen tube growth, and self-fruit set resulting from automatic self-pollination (autogamy) was relatively high in a few plants. Automatic self-pollination is possible because the male and female parts of flowers mature sychronously, but the likelihood of autogamy may vary among individuals due to variability in the spatial separation of male and female parts (herkogamy). Variability in the incompatibility system and the level of herkogamy may impact the outcrossing rates and reproductive success of individuals. The majority of floral visitors to E. umbellata were generalist pollinators. Frequently visiting bees included small and large species such as native Andrena spp., Augochlorella aurata, Bombus spp., Ceratina calcarata, Xylocopa virginica, and the introduced Apis mellifera. Bombylius major (large bee fly) and the moth Mythimna unipuncta (armyworm) were also frequent visitors. Most of the above insect taxa are pollinators of E. umbellata based on analysis of pollen on insect bodies. E. umbellata is likely to achieve its abundant fruit set where these common pollinators and other E. umbellata are present. However, in my study sites, many individuals experienced low fruit set on branches that were open to pollinator visitation, suggesting pollen limitation may be common in some years and at certain sites. The discovery of autogamous individuals demonstrates that some E. umbellata individuals may be able to establish and spread even when mates or pollinators are limiting.

<p>Due to overpopulation and resource-poor habitat structure, deer threaten the<a> future of oak and other browse-sensitive species in hardwood forests. </a>Appropriate tools must be used to ensure desirable, diverse, and ecologically stable regeneration of future forests and the sustainability of native plant communities. We performed two experiments and a review to examine the effectiveness of available methods for managing browse of hardwood seedlings and to discover how these interact with each other and other silvicultural methods. First, we examined how fencing interacts with controlled-release fertilization, seed source (genetically select and non-select), and site type (afforested and reforested sites) to enhance the regeneration of planted northern red oak (<i>Quercus rubra </i>L.), white oak (<i>Quercus alba</i>), black cherry (<i>Prunus serotina</i>), and black walnut (<i>Juglans nigra</i>) at five sites in Indiana. Fencing proved to be the greatest determinant of seedling growth, survival, and quality. Fertilizer enhanced the early growth of white oak and black cherry, though for black cherry this occurred only inside fences. Select seed sources grew better and showed greater quality; however, the survival of select seedlings was limited by deer browse in absence of fences. Trees at afforested sites had lower survival if left non-fenced. Secondly, we also investigated how fencing and invasive shrub removal affected natural regeneration, species richness, and ground-layer plant cover under closed-canopy forests. Honeysuckle (<i>Lonicera maackii</i>) removal had a variable effect depending on species and site. Positive effects were most common for shade-intolerant species, while negative effects occurred for a few shade-tolerant species at some sites. Deer fencing had a positive effect on cherry and hackberry seedling density, and a negative effect on elm seedling density. Honeysuckle and deer fencing interacted antagonistically in some instances. Fencing without honeysuckle removal resulted in lower elm abundance and herbaceous-layer cover. In the densest invasions, leaving honeysuckle intact resulted in a complete lack of recruitment into the sapling layer. Our experiment suggests that invasive shrub removal and fencing be done together. Finally, we synthesized the existing literature on browse management options for hardwood regeneration to evaluate their relative effectiveness. Fences, tree shelters, repellents, facilitation by neighboring plants, deer population control, timber harvest, and slash all had positive effects on height growth of regenerating seedlings under deer browse pressure. Fences were more effective at reducing browse than repellents, while fertilizers increased browse and had no effects on growth. </p>