Premium
SPATIAL POPULATION GENETIC STRUCTURE IN TRILLIUM GRANDIFLORUM : THE ROLES OF DISPERSAL, MATING, HISTORY, AND SELECTION
Author(s) -
Kalisz Susan,
Nason John D.,
Hanzawa Frances M.,
Tonsor Stephen J.
Publication year - 2001
Publication title -
evolution
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.84
H-Index - 199
eISSN - 1558-5646
pISSN - 0014-3820
DOI - 10.1111/j.0014-3820.2001.tb00675.x
Subject(s) - biology , biological dispersal , selection (genetic algorithm) , evolutionary biology , mating , population , zoology , demography , machine learning , sociology , computer science
.— The roles of the various potential ecological and evolutionary causes of spatial population genetic structure (SPGS) cannot in general be inferred from the extant structure alone. However, a stage‐specific analysis can provide clues as to the causes of SPGS. We conducted a stage‐specific SPGS analysis of a mapped population of about 2000 Trillium grandiflorum (Liliaceae), a long‐lived perennial herb. We compared SPGS for juvenile (J), nonreproductive (NR), and reproductive (R) stages. Fisher's exact test showed that genotypes had Hardy‐Weinberg frequencies at all loci and stage classes. Allele frequencies did not differ between stages. Bootstrapped 99% confidence intervals (99%CI) indicate that F ‐statistic values are indistinguishable from zero, (except for a slightly negative F 1T for the R stage). Spatial autocorrelation was used to calculate f , the average kinship coefficient between individuals within distance intervals. Null hypothesis 99%CIs for f were constructed by repeatedly randomizing genotypic locations. Significant positive fine‐scale genetic structure was detected in the R and NR stages, but not in the J stage. This structure was most pronounced in the R stage, and declined by about half in each remaining stage: near‐neighbor f = 0.122, 0.065, 0.027, for R, NR, and J, respectively. For R and NR, the near‐neighbor f lies outside the null hypothesis 99%CI, indicating kinship at approximately the level of half‐sibs and first cousins, respectively. We also simulated the expected SPGS of juveniles post dispersal, based on measured R‐stage SPGS, the mating system, and measured pollen and seed dispersal properties. This provides a null hypothesis expectation (as a 99%CI) for the J‐stage correlogram, against which to test the likelihood that post‐dispersal events have influenced J‐stage SPGS. The actual J correlogram lies within the null hypothesis 99%CI for the shortest distance interval and nearly all other distance intervals indicating that the observed low recruitment, random mating and seed dispersal patterns are sufficient to account for the disappearance of SPSG between the R and the J stages. The observed increase in SPGS between J and R stages has two potential explanations: history and local selection. The observed low total allelic diversity is consistent with a past bottleneck: a possible historical explanation. Only a longitudinal stage‐specific study of SPGS structure can distinguish between historical events and local selection as causes of increased structure with increasing life history stage.