Y

<>, constant, 5/2

1.76***

0.65***

a Reward indicates whether sucrose solution was present (Y) in flowers tested or not (N).

b Rewards equal in both color types (=); rewards not equal in both color types (<>); rewards the same within each reward type (constant); one or both color types had variable rewards (variable). Values in the reward schedule column indicate amount of sucrose added; two values are given if color types have different reward levels. c The Greenwood & Elton (1979) model was fitted to data in all cases to test for selection (see text). The degree of frequency dependence (b) and its significance (* p <0.05,

** p <0.01, *** p <0.001) are given; a b value <1 indicates preference for the rare morph; b >1 indicates common-morph preference. The value and significance of frequency-independence (V) is indicated the same way. Frequency independence increases with departure from unity.

stronger learned responses to patterns of empty and full flowers (Perez & Waddington 1996; Waddington, this volume). Neither reward schedule (Smithson 1995) nor variability (Table 12.1, Fz§4 = 0.38, NS) significantly affected the strength of frequency dependence, although that strength tended to decrease with increasing variability. Other experiments testing the effects of density and total reward also found no significant effect on the strength of frequency dependence (Smithson & Macnair 1996,1997a). However, where neither color morph contained reward, behavioral pat-

terns were reversed, and bees showed significant rare-morph preference (Smithson & Macnair 1997b).

Visitation order was correlated with frequency dependence: when bees displayed common-morph preference, they visited the same colors sequentially, but when they displayed rare-morph preference, they tended to visit morphs disassortatively (Smithson & Macnair 1996, 1997a, b). Frequency-dependent behavior increased with experience, being weak in the first 50 flowers visited, developing to its maximum extent over 50-100 flowers visited, and changing little thereafter; however, the level of assortative or disassortative visitation remained constant over all flower visits (Smithson & Macnair 1996).

Bumble bees forage with unequivocal frequency-dependence on arrays of simple artificial flowers that vary in color. Variation in flower size produced no FDS (A. Smithson unpublished data), but other floral traits such as shape or fragrance have not been tested for a frequency-dependent response, nor have other pollinator types been tested.

Was this article helpful?

0 0

Post a comment