Table 2. Factors affecting diversity within species at 4-fold degenerate sites for unpreferred codons using intervals with conserved recombination rate.

Factor Tested	df	Deviance	Residual df	Residual Dev.	F	p Value
Null			26	57.009
Gene density	1	2.3190	25	54.690	2.2948	0.147171
Mutation	1	12.7343	24	41.955	12.6013	0.002289*
Recombination	1	6.1877	23	35.768	6.1231	0.023521*
GC	1	11.1854	22	24.582	11.0685	0.003751*
Gene Density×Mutation	1	2.0720	21	22.510	2.0504	0.169304
Gene Density×Recombination	1	2.8041	20	19.706	2.7748	0.113065
Mutation×GC	1	0.5488	19	19.157	0.5430	0.470669
Recombination×GC	1	0.0007	18	19.156	0.0007	0.978599

A generalized linear model with quasibinomial distribution for the fine-scale intervals on chromosome 2 with conserved recombination rates between D. pseudoobscura Flagstaff, D. pseudoobscura Pikes Peak, and D. miranda after correction for the global modifier. This model illustrates the relationship of average pairwise D. pseudoobscura diversity for 4-fold degenerate sites of unpreferred codons to various factors. Windows that were nonsignificant when analyzed with a rare events logistic regression and had an Odds Ratio between (0.62 to 1.615) across maps were considered “conserved.” For this analysis, the “neutral mutation rate” was set as the average pairwise D. lowei–D. persimilis divergence at 4-fold degenerate sites for unpreferred codons. For consistency between models, if an interaction term was significant in any of the models (see Tables 3, S9, S10, and S11), it was kept in all. Results from uncondensed chromosome 2, XL, and XR exhibit similar relationships (Tables S9, S10, and S11). An asterisk indicates significance at an α of 0.05.