The basis for measuring differentiation among subpopulations is discussed and a number of conditions formulated that are desirable for an appropriate measure. These conditions imply that each subpopulation is characterized by the difference in genetic composition between it and its complement. A direct method of determining this difference is described and shown to result from a known measure of genetic distance between populations. The weighted average of the genetic distances between subpopulations and their complements constitutes a measure of differentiation among subpopulations which fulfills all of the desirable conditions and has the additional advantage that its values are directly interpretable. This measure (δ) is equally applicable to gene (δ ge ), gametic (δ ga ) or genotypic (δ go ) frequencies, which guarantees an unequivocal multilocus treatment, provided that the sets of genetic entities to which the frequencies refer are properly defined. The general relationship δ ge ≤ δ ga ≤ δ go is consistent with the principle that increasing complexity of organization of genetic material results in increased opportunity for differentiation. It is demonstrated that Wright's F st (G ST in Nei's notation), which is often used to measure subpopulation differentiation, meets some but not all of the conditions formulated for a desirable measure.

No abstract

Abstract:In southern pines of the United States, resistance to attack by southern pine beetle, Dendroctonus frontalis Zimmermann, is believed to principally involve flow of oleoresin to beetle attack sites. Both environmental and genetic factors are known to affect the quantity of oleoresin flow in loblolly pine, Pinus taeda L., but little is known about the genetic contribution to phenotypic variation in this trait. Here we report estimates of genetic variation in oleoresin flow and growth traits for a population of this species. Oleoresin yield, total height, and diameter were measured on 10-and 11-year-old trees from an experimental test in Santa Rosa County, Florida. Trees were from 72 full-sib families produced by mating 48 parents according to a disconnected partial diallel mating design. Resin yield was determined from breast-height samples collected at two times: once in the summer of 1999 when latewood was being produced (summer resin flow), and once in the spring of 2000 during earlywood formation (spring resin flow). All traits studied were found to be highly genetically variable and to have much greater additive than dominance variance. Estimates of narrow-sense heritability for spring and summer resin flow were in the moderate range and are comparable to values obtained for the growth traits. Additive genetic correlations between oleoresin yield and the growth traits were positive and moderately high, suggesting that directional selection to improve growth in loblolly pine will also result in increased production of oleoresin.