I argued above that variation in predation risk among seasons and times of day may affect pollinators' activity. This in turn could have selected for changes in flowering times. For example, selection for early activity by bumble bees (driven by lesser activity of conopid parasitoids) may have selected for earlier blooming of bumble bee flowers. Similarly, the dusk activity of moths may have selected for flowers that open and secrete nectar at dusk.
If predation rates do vary among plant species, this factor may influence floral choices by pollinators, in addition to other factors such as the quality and quantity of reward. Furthermore, differential predation rates may have contributed to specialization and flower constancy by pollinators. These behavioral patterns allow higher rates of intraspecific pollen transfer than would be possible with generalist, inconstant flower visitors.
Various floral traits may reduce predation rates on pollinators. Although it is likely that the effect of many of these traits on predation is incidental, it is conceivable that some evolved due to their effect on lower predation. For example, young Acacia flowers appear to emit a volatile chemical signal that deters ant-guards, allowing pollinating bees safe passage (Willmer & Stone 1997). Other possibilities are outlined in the sections below.
Flower shape and size Morse (1981) noted various floral characteristics that affected predation activity by crab spiders (see 'Flower specialization and flower constancy', above). On the large nectarless rose flowers, spiders ambushed at the centers of the flowers near the stamens, which predictably received insect visits. In contrast, on milkweed and goldenrod, the spiders could not occupy a spot guaranteed to receive insect visits. As a result, spiders on rose attacked 68% of the visiting bumble bees, but spiders on milkweed and goldenrods attacked 33% and 43% of the visiting bumble bees, respectively. (Note, however, that the predation rates (i.e., successful attacks) on bumble bees were actually higher on milkweed and goldenrod than on rose.)
The rose flowers, however, had a major disadvantage as hunting grounds for crab spiders. Each nectarless flower attracted insect visitors for only a few hours during the morning because the pollen was mostly removed by noon. Consequently, the spiders had to move every day to a new flower, a procedure that decreased their daily hunting duration compared to the hunting duration on goldenrod and milkweed (Morse 1979, 1981). These are examples for how floral characteristics affect the activity of an ambush predator on different flowers. Have some floral traits, such as a flower's lifespan, size, or color, evolved due to a negative effect on predation?
Nectar availability The tarnished plant bug, Lygus lineolaris, is rarely attacked by the braconid parasitoid (Leiophron pallipes) on Oenothera, Daucus, Amaranthus, and Solidago, but is parasitized at rates of up to 40% on Erigeron species (Price et al. 1980). Apparently, the parasitoids are more attracted to Erigeron flowers, which provide higher quality nectar. Analogous three-level interactions are possible for flower-pollinator-predator interactions. For example, might bumble bees receive more conopid attacks at flowers with nectar accessible to the flies.? Has the evolution of concealed nectar been driven in part due to its negative effect on pollinators' parasitoids?
The celebrated match between long-spurred orchids and long-tongued hawkmoths (Nilsson 1988, 1998) has traditionally been attributed to a coevolutionary race between the flowers and pollinators, although an alternative version assuming that the long-tongued hawkmoths existed before the long-spurred orchids is also feasible. According to either scenario, the plants with longer spurs receive higher rates of pollen transfer to stigmas because the hawkmoths' proboscis bases are more likely to contact the floral sexual organs when the hawkmoths insert their tongues more deeply into the spurs. The open question is what factor(s) selected for the hawkmoths' long tongues. Two answers are feasible: the traditional answer is that longer tongues are associated with greater net rate of energy gain and fitness while feeding on flowers with long spurs. A non-mutually-exclusive alternative suggested recently by Wasserthal (1993, 1997, 1998) is that tongues longer than spurs also allow hawkmoths to oscillate sideways while hovering at flowers, and that this "swing hovering" decreases predation. Field measurements (Inouye 1980) and controlled laboratory experiments (Harder 1983) indicated that bumble bees with longer mouthparts should prefer long-tube flowers, which typically contain more nectar. These studies, however, do not preclude a role for predation risk in the evolution of long nectar-extracting mouthparts in hawkmoths or other taxa. Hence Wasserthal's innovative suggestion deserves critical experimental evaluation.
Was this article helpful?