Home | CV | Research | Teaching | Publications | Cactus Bees | Cactus growth data | Personal


Dissertation Abstract:

Although bees are herbivorous insects, they are seldom studied as such. My dissertation research characterizes the interactions between a guild of solitary cactus-specialist bees and the cacti they visit, examining these bees both as herbivores and as pollinators.

I first examined the reproductive biology of two species of Ferocactus as background for the pollination study. Both F. cylindraceus and F. wislizeni were obligate outcrossers, and neither species was pollen-limited. Fruit set was high in both species, and reproductive output was limited by architectural constraints.


Whereas most other herbivorous insects are antagonists of the plants they feed on, bees are usually regarded as mutualists (pollinators) of their host plants. However, not all floral visitors are pollinators, and pollen-foraging specialists have been postulated by some to be particularly antagonistic to the plants they visit. To test whether cactus-specialist bees are actually mutualists of the cacti they visit, I examined both the quantity and quality components of pollinator effectiveness of bees visiting and F. wislizeni . Despite the generalized morphology of the flowers of these plants, there were very few floral visitors other than three species of cactus-specialist bees. Flowers of both species were pollinated almost exclusively by cactus-specialist bees, primarily Diadasia rinconis.


I also tested the pollen preferences of four species of cactus bees. Bees were offered flowers from their normal host cactus in which the stamens had been removed and replaced with novel pollens. Novel cactus pollens were accepted to some degree by all four species, but Diadasia rinconis and D. opuntiae accepted more novel non-cactus pollens than did either Lithurge apicalis or Idiomelissodes duplocincta. Both species of Diadasia showed significant acceptance of Sphaeralcea (Malvaceae) pollen, thought to be the ancestral host plant for the genus Diadasia. Caged, naive D. rinconis bees did not initiate nesting when presented with Sphaeralcea flowers alone; nesting began immediately when cactus flowers were offered. Once nesting was underway, most D. rinconis bees switched to Sphaeralcea if cactus flowers were removed. In both D. rinconis and D. opuntiae, there were striking differences in pollen preferences among individuals.


Summary of Research Experience and Interests

My research interests lie at the interface between herbivorous insects and their host plants. I am currently studying the interactions between pollen-specialist bees and the flowering plants they visit. This work combines traditional pollination biology with the relatively unusual perspective of viewing bees as herbivorous insects. In particular, I am studying the pollinator effectiveness, foraging behavior, and host preferences of several taxa of cactus-specialist bees. My research uses field-based ecological experiments to study the "how" of pollen specialization (host recognition, degree of specialization), and to determine whether these bees evince behavioral flexibility that would allow them to switch to other plants if cacti were unavailable. My work also seeks to locate these interactions along the mutualism-antagonism continuum, by characterizing the costs and benefits of pollen-specialist bees to the plants they visit.

1. Pollen specificity in cactus-specialist bees

In the deserts of the southwest, the flowers of many cacti are visited by a guild of relatively unrelated bees that all specialize on cactus pollen. To test the host-specificity of these cactus bees, I placed novel pollen into cactus flowers, and observed whether these bees would accept these novel pollens. I used pollens from both native and non-native plants, including plants that are used by close relatives of cactus bees. I addressed the following questions:

1. Do all cactus bees show the same strictness in specialization?
2. Is there behavioral variation within populations of each species of bee that would allow
host shifts onto non-cactus hosts?
3. Do cactus bees recognize their host plant by recognizing the flowers (in which case they
might collect any pollen found in a cactus flower), or do they recognize the pollen itself?
(these two are not mutually exclusive).


I found that most non-cactus pollens were rejected by cactus bees, suggesting that host acceptance takes place at the level of the pollen itself, not the flower. Rejection often took place after landing, and hence may involve chemosensory and/or tactile cues, in addition to the visual and olfactory cues the bees can use during flight. In particular, I tested the pollen specificity and preference of naive and experienced Diadasia rinconis (Hymenoptera: Apidae), free-flying and in captivity, in choice and no-choice experiments. I found that naive bees that had never seen cactus flowers, and that were provided with other flowers, did not initiate nesting (and therefore pollen-foraging) until cactus flowers were introduced. However, once they were nesting with cactus pollen, they did switch to globe-mallow for pollen if cactus flowers were denied them. Experienced individuals of both D. rinconis and D. opuntiae accepted a few non-cactus pollens, including globe-mallow pollen. Individuals of Lithurge apicalis (Hymenoptera: Megachilidae) that were tested rejected all non-cactus cacti. Free-flying experienced individuals of Idiomelissodes duplocincta (Hymenoptera: Apidae) accepted a few non-cactus pollen, showing more behavioral flexibility than L. apicalis, but not as much as D. rinconis and D. opuntiae .

My results suggest that cactus-specialist bees in the genus Diadasia may have the behavioral flexibility to switch to other plants if cacti are not available, but that not all cactus bees may be so flexible. They also suggest that the history of host-use in the clades to which cactus bees belong may play a role: globe-mallow is thought to be the ancestral host plant for the genus Diadasia , and globe-mallow pollen was accepted by Diadasia cactus bees but not by other cactus bees.

2. Pollination biology of barrel cacti

Because host-plant specialization is widespread in both bees and other herbivorous insects, it might suggest that specialization per se is adaptive for herbivores. However, selective forces on bees may be quite different than those that affect other herbivores. Bees are thought to have a beneficial effect on the plants they visit through their dispersal of pollen, so plants should not need to defend themselves against bees, the way they do against other herbivorous insects. On the other hand, many flower visitors are neutral or even harmful in their effect on the plant (e.g., removing pollen or nectar without pollinating the flower). Using the technique of "single-pollinator visits," I investigated the effect of cactus-specialist bees, and other flower visitors, on two sister species of barrel cacti, Ferocactus cylindraceus and F. wislizeni, over a 3-year period. Despite the fact that many insect-pollinated plants are visited by a wide variety of visitors, both specialist and generalist, I found that these two plants are visited primarily by three species of cactus-specialist bees, and that these specialist bees are also the best pollinators. Visits by specialists accounted for 100% of the seed production in F. cylindraceus , and 90% in F. wislizeni. I also found that visits by female bees led to greater seed production than visits by male bees, and that visits for nectar alone resulted in the greatest seed production. My results suggest that the relationship between cactus bees and cacti is primarily mutualistic (in contrast to other herbivorous insects).

3. Plant reproduction, growth, and demography

As part of my investigation into the pollination biology of barrel cacti, I conducted hand-pollinations to determine the breeding system of the plants. I also measured plant size and reproductive output over a 5-year period, to elucidate the role of plant size in reproduction, to determine whether there is a tradeoff between growth and reproduction, and to identify those factors that impact reproductive output. I found that, despite the fact that they are sister-species, they differ significantly in their flowering phenology, seeds per fruit and seed mass. In both species, as in many plants, plant size had a positive effect on reproductive effort and on overall fecundity. Also in both species, a minority of individuals had the ability to self; most were obligate outcrossers. Even in selfers, insect visitation was required to set fruit. They were interfertile and may be hybridizing in the wild. Fruit set (excluding flowers and fruits damaged by herbivores) was extremely high in both species (90-95%). Overall fecundity was also high, around 14,000 seeds/plant/year for F.cylindraceus , and 22,000 for F.wislizeni . These data allowed me to place the results from my single-pollinator visits into the overall context of the normal reproduction of these plants.

I am also a member of a team conducting a long-term study of the growth, reproduction and demography of a disjunct population of Echinocactus horizonthalonius, which has been ongoing since 1995. We have found that these cacti are extremely slow-growing, and despite regular reproduction, are experiencing very low recruitment.

4. Solitary bee and wasp nesting and foraging behavior

The reproductive success of solitary bees and wasps depends in part on their ability to locate suitable nesting sites. During my undergraduate work at U. C. Berkeley, I conducted a study of the nest-substrate preferences of twig-nesting bees and wasps. These insects burrow into dead, pithy stems, and dig nesting cells and galleries for their offspring. I offered them twigs from a range of plant species, and in different sizes, to determine what their nesting requirement were. I found that Pemphredon wasps preferred large, woody stems, possibly because they dig complex, branching burrows. Ceratina acantha bees, which dig a simple linear tunnel, preferred woody twigs but had no size preference.

I have also studied the phenology and activity of several nesting aggregations of two cactus-specialist bees, Diadasia rinconis and D. opuntiae , in and around Tucson in 1998-2000. My data provide supporting evidence for the partial bivoltinism of D. rinconis that has been postulated by previous workers.


| Home | CV | Research | Teaching | Science Writing | Cactus Bees |


Last Updated: 30 October 2001
© copyright 2001 Margrit E. McIntosh