ation within populations; rather, the majority of variation, 87%, is among populations ( Hong, Hipkins, and Strauss 1993). Similarly, 96% of the variation revealed by mtDNA was among populations ( Strauss, Hong, and Hlpkins 1993). Thus, the degrees of differentiation of populations are heterogeneous among the allozymes, cpDNA, and mtDNA, and the relative degrees of differentiation are not consistent with the potential gene flow of the sets of markers.
The degrees of allozyme, cpDNA, and mtDNA differentiation appear to fit expectations based on gene flow in lodgepole pine ( Dong and Wagner 1993, 1994; Wheeler and Guries 1982). Allozyme differentiation in lodgepole pine is typical of conifers; values of Fst estimated from allozymes rarely exceeded 6% ( Wheeler and Guries 1982) and were generally less than 10% in conifers ( Hamrick and Godt 1990). Large proportions of the mtDNA variation within lodgepole pine were among populations within subspecies (Fst = 66%) and among subspecies (Fst = 31%). CpDNA variation was high within populations, with little differentiation--less than 5%--among populations. Thus, the differentiation of populations was strong, for mtDNA, but slight for both allozymes and cpDNA.
Allozymes and DNA markers demonstrate contrasting patterns of genetic variation in limber pine, Pinus flexilis ( Latta and Mitton 1997). Allozvme variation was used to describe variation within and among populations from the lower tree line ( 1650 in) to the upper tree line ( 3350 m) in Colorado ( Schuster, Alles, and Mitton 1989). The average for these loci was 0.02, indicating little differentiation among sites and gene flow on the order of ten migrants between populations per generation. But this estimate of gene flow did not seem to be possible, for the pollination phenology of limber pine varied dramatically with elevation. Most sites along the elevational transect that differed by 400 in or more did not have overlapping pollination periods and therefore could not exchange genes during a single generation. A survey of RAPD markers yielded two patterns of variation. Approximately two-thirds of the markers had a pattern similalar to that of the allozyme markers, with values of Fst very close to zero. But approximately one- third of the markers had values of Fst near 0.50, far higher than values for the other RAPD markers and the allozymes.
An allele's electrophoretic mobility at a protein polymorphism is related to its frequency, with rare alleles tending to migrate eitheir more quickly or more slowly than the common alleles do.
The loci coding for proteins are often assumed to be a random sample of the genome, but it is not clear that this assumption is valid. Certainly, some groups of proteins are more genetically variable than other groups. Regulatory enzymes appear to be more variable than nonregulatory enzymes, and enzymes that work on many substrates appear to be more variable than enzymes that utilize a single substrate. Genetic variation tends to decrease from monomer to dimer to tetramer as the steric restrictions on the molecule increase. The number of alleles at a locus and the heterozygosity at a locus tend to increase with the subunit size of the protein.
For both morphological characters and single-gene polymorphisms, the divergence among populations increases with the average variability within populations. This gen-