Ja. Johnston et al., Intimately linked or hardly speaking? The relationship between genotype and environmental gradients in a Louisiana Iris hybrid population, MOL ECOL, 10(3), 2001, pp. 673-681
Several models of hybrid zone evolution predict the same spatial patterns o
f genotypic distribution whether or not structuring is due to environment-d
ependent or -independent selection. In this study, we tested for evidence o
f environment-dependent selection in an Iris fulva x Iris brevicaulis hybri
d population by examining the distribution of genotypes in relation to envi
ronmental gradients. We selected 201 Louisiana Iris plants from within a kn
own hybrid population (80 m x 80 m) and placed them in four different genot
ypic classes (I. fulva, I. fulva-like hybrid, I. brevicaulis-like hybrid an
d I. brevicaulis) based on seven species-specific random amplified polymorp
hic DNA (RAPD) markers and two chloroplast DNA haplotypes. Environmental va
riables were then measured. These variables included percentage cover by tr
ee canopy, elevation from the high water mark, soil pH and percentage soil
organic matter. Each variable was sampled for all 201 plants. Canonical dis
criminant analysis (CDA) was used to infer the environmental factors most s
trongly associated with the different genotypic groups. Slight differences
in elevation (-0.5 m to +0.4 m) were important for distinguishing habitat d
istributions described by CDA, even though there were no statistical differ
ences between mean elevations alone. I. brevicaulis occurred in a broad ran
ge of habitats, while I. fulva had a narrower distribution. Of all the poss
ible combinations, I. fulva-like hybrids and I. brevicaulis-like hybrids oc
curred in the most distinct habitat types relative to one another. Each hyb
rid class was not significantly different from its closest parent with rega
rd to habitat occupied, but was statistically unique from its more distant
parental species. Within the hybrid genotypes, most, but not all, RAPD loci
were individually correlated with environmental variables. This study sugg
ests that, at a very fine spatial scale, environment-dependent selection co
ntributed to the genetic structuring of this hybrid zone.