The nectar of many plant species contains defensive compounds that have been hypothesized to benefit plants through a variety of mechanisms. Nectar attracts both pollinating and non-pollinating floral visitors, and plants may benefit from nectar traits that deter floral predators without also deterring pollinators. We experimentally manipulated gelsemine, the principal alkaloid of Gelsemium sempervirens, in nectar to determine its effect on pollinator visitation, nectar robber visitation, and male and female plant reproduction. We experimentally augmented or diluted nectar alkaloids on plants in a field array in separate experiments over two years. We found that nectar robbers and most pollinators probed fewer flowers and spent less time per flower on plants with high compared to low nectar alkaloids. High alkaloids decreased the donation of fluorescent dye, an analog of pollen used to estimate male plant reproduction, to neighboring plants by one-third to one-half. However, nectar alkaloids did not affect female plant reproduction, measured as pollen receipt, fruit set, seed set, and seed mass. The lack of a consistent effect of nectar alkaloids on female reproduction might indicate that neither nectar robber nor pollinator visits affected female reproduction. Alternatively, the lack of an effect on female reproduction may represent a balance between the costs and benefits of altered nectar robber and pollinator behavior. Taken together, our results suggest that secondary compounds in nectar provide more costs than benefits to plant fitness.