RESULTS:

From the eighty primers surveyed using bulked DNA samples from four populations, nine produced interpopulation band differences (Primer: number of bands -- A19: 1, B09: 2, C10: 2, C14: 5, D12: 1, D16: 5, D20: 1, G5: 7, G12: 4). In observing pairs of bands across all individuals for each primer, no obvious pattern emerged to suggest codominant banding patterns. Since most RAPD bands are inherited as dominant Mendelian loci (Williams et al. 1990), and the RAPD patterns in this study appeared to concur with these results, the banding patterns were interpreted as dominant loci.

Dendrograms made using UPGMA and neighbor-joining showed no clear separation between any of the four study populations. Individuals from each population were dispersed throughout the trees, making it impossible to discern any structure (Figure 1). Dendrograms made with only the El Eden populations (sites one and two) also showed no separation between the forest and swamp sites.

Results from the Fisher's exact test over all loci with the swamp and forest populations at El Eden told a different story. The combined p-value was significant at the 0.005 level with 2n=56 degrees of freedom.

The results of the analysis of the distance matrix generated by MEGA are shown below. The range of distances between individuals of the forest population was very similar to the range of distances between individuals of the forest and the swamp. The range of distances between swamp individuals was lower than either of these values, indicating that there is greater similarity among individuals in this population.

Calculations from distance matrix for El Eden swamp and forest populations

Interaction	Mean value	Standard Error	Range
Forest x Forest	0.4101	0.0048	0.4053 ? 0.4149
Swamp x Swamp	0.3890	0.0047	0.3843 ? 0.3937
Forest x Swamp	0.4125	0.0035	0.4090 ? 0.4160