| Populations of a species are commonly observed to differ from one another in traits thought to be associated with fitness. Experimental work in many species indicates these differences often have a genetic basis. What is generally unknown, however, is the genetic architecture of population divergence: Do patterns of trait covariation change between populations; how many factors are responsible for phenotypic divergence; to what extent do factors underlying divergent phenotypes fail to interact properly in a hybrid genetic background? In this thesis I examine the genetic architecture of population divergence in the moss Ceratodon purpureus.; Natural populations of C. purpureus differ morphologically, and common garden experiments indicate many of these differences have a genetic basis. In the first chapter, I describe molecular population genetic analyses of three loci from a global sample aimed at understanding broad demographic patterns in the species as a whole. Overall, I found no evidence for genealogical structure, suggesting recurrent migration among populations. In loci showing low divergence among geographic regions, I found evidence of positive selection, suggesting that globally adaptive alleles sweep through even geographically disjunct populations.; In the second chapter I examine the genetic architecture of life history variation in male and female gametophytes from two populations, focusing on genetic correlations within and between the sexes. The structure of genetic correlations was most divergent between the sexes in one of the populations, and between females of the two populations. Collectively these results suggest that the genetic architecture of life history variation between these populations is labile in spite of ongoing gene flow.; In the third chapter I use quantitative trait locus (QTL) mapping to dissect the genetic architecture of differences between two geographically distant populations. The populations were little differentiated at early developmental stages, but half of the progeny from an interpopulation cross showed abnormally slow growth, suggesting hybrid breakdown. I identified four major QTLs underlying the low-fitness phenotype. In total, these results suggest that populations can diverge in trait means and patterns of covariances in spite of ongoing gene flow. Moreover, these results suggest the raw material for reproductive isolation may be segregation within species. |
