Ecological and genetic examinations of reproduction in a tropical epiphytic orchid across a fragmented forest habitat

Tropical forests are rapidly being dissected into small isolated fragments. However, little is known about the impacts of fragmentation on plant populations remaining in fragments. I examined effects of fragmentation on pollination, reproductive ecology and genetic structure of an endemic epiphytic orchid, Catasetum viridiflavum, on islands created during the construction of the Panamá Canal on and nearby continuous forest sites. I found that pollinators are equally available in fragments and continuous forest. Male and female reproductive success were not significantly different between island and mainland sites in 1997. Across three years, however, fruit set does significantly differ between island and mainland sites. Plants with larger inflorescences had higher reproductive success regardless of site. ^ C. viridiflavum produces large sexually dimorphic flowers and can switch sex between reproductive episodes. I found that current size, light environment and previous year\u27s reproductive effort contribute to the likelihood of flowering. Different factors influence gender determination temporally. Overall, size in the current year had a large effect. In 1997, previous reproductive output influenced the likelihood of a plant flowering female. ^ Genetic structure on mainland and forest fragments were similar, which may indicate that similar balances between gene flow, drift and selection were occurring. Alternatively, sufficient time may not have passed to show signs of isolation in the genetic structure. Stage class structure, breeding populations and spatial location are all important components for describing the impacts of fragmentation on gene flow in this orchid. Ecological studies of pollinator behavior are consistent with a low level of genetic differentiation among spatially isolated sites. Surprisingly for a dioecious orchid, populations have a large deficiency of heterozygotes relative to Hardy-Weinberg expectations. The combination of little among population differentiation and high inbreeding coefficients within sites might be the result of bi-parental inbreeding among plants followed by seed dispersal among islands. The combination of ecological and genetic studies is important because it allowed me to distinguish the relative importance of pollinator behavior, breeding populations, and seed dispersal on the genetic structure in this fragmented habitat.