• O‘Connell, Lisa
  • Eckert, Christopher G.


It is widely thought that many apomictic angiosperms exhibit variation in sexual versus apomictic seed production, yet there has been little investigation of the extent to which apomixis varies among or within natural populations. In dioecious Antennaria parlinii, previous reports of wide variation in both population sex ratios and the capacity of individual plants for autonomous seed production indicate that sexuality varies extensively both among and within populations. We used a combination of population surveys and pollination experiments to test this hypothesis. A survey of seed production in 15 populations containing only females and nine populations containing both sexes revealed that the absence of males was not associated with reduced seed set in female populations ( ) compared with mixed‐sex populations (0.69). In mixed‐sex populations, seed production appeared strongly dependent on pollination. Seed set of females in mixed‐sex populations correlated positively with both the total number of inflorescences in a population ( ) and the frequency of male inflorescences ( ), whereas seed set was independent of inflorescence number in female populations. In addition, the seed set of individual plants within two mixed‐sex populations correlated negatively with distance to the nearest male inflorescence. Excluding pollinators from inflorescences almost entirely prevented seed set in two mixed‐sex populations but had no effect on seed production in two female populations. In a pollinator‐free greenhouse, plants from nine mixed‐sex populations only set seed when hand pollinated, whereas pollination had no effect on seed production in plants from 13 female populations. Taken together, these results indicate that populations of A. parlinii are either predominantly sexual or predominantly apomictic. Partial apomixis and/or polymorphism for sexuality within populations appear to be uncommon in this species. Strong differentiation for sexuality is in accord with theoretical models that predict disruptive selection for apomixis when pollination is not required for seed maturation.


Plots were surveyed