F = 0.28

F = 0.27

Ackerman etal. 1997


" Reward is either present (Y) or absent (N).

b The numbers of replicates are given as the total number, with the number of different frequencies considered in parentheses. c Where possible, frequency dependence and frequency independence are tested as for Table 12.1 using the Greenwood & Elton (1979) model. Where numbers of replicates were too low, frequency independence gives the result of a x2 test for heterogeneity and frequency dependence the result of testing deviations of expected values for rank correlation with morph frequency (see text for discussion). In one case, the authors' F-statistics are given.

For the two studies with larger sample sizes, researchers found significant preferences for the common corolla colors in Ipomoea purpurea, and overall preference for one corolla color over another in both I. purpurea and Cirsium palustre (Table 12.2). For the two studies with smaller sample sizes, there is no indication of non-random pollinator visitation for Clarkia gracilis, but non-randomness is suggested for Delphinium nelsonii. There was no common-morph advantage in either data set; indeed, the marginally significant pattern for D. nelsonii suggests negative rather than positive frequency dependence. However, in this species one of the four data points shows a particularly large deviation from expected values accounts for most of the correlation; conclusions from such small sample sizes must be tentative. Using meta-analysis (Kirby 1993), I combined all results for the rewarding species to test the hypothesis of positive frequency dependence. The result did not support the hypothesis (average effect size = 0.17, t = 0.42, p = 0.71); the experiments were significantly heterogeneous (heterogeneity xz = 11.32, p = 0.01).

One unrewarding species has been studied with respect to frequency-dependent pollinator visitation - an epiphytic orchid, Tolumnia variegata. This species produces no nectar but relies on deception (without mimicry) of naive pollinators for pollen transfer. Such species are termed "non-model deceptive" (Dafni 1986). Tolumnia variegata is polymorphic for the presence or absence of scent production. By manipulating morph frequencies in adjacent trees, Ackerman et al. (1997) found that pollinators showed neither overall preference for a morph nor frequency-dependent selection in favor of the rare morph.

Overall, a survey of the literature does reveal significant common-morph preference in at least one rewarding plant species, but this is weak; predictions are not upheld in three other rewarding species. For an unrewarding species, pollinators did not show significant preference for either morph type. Of course, these studies involved different species of animals.

Was this article helpful?

0 0

Post a comment