Alison K. Brody

The Dual Role of Floral Traits: Pollinator Attraction and Plant Defense

Plants are under siege from a diversity of enemies that consume both leaf and floral parts. Plants resist damage to leaves in a variety of ways, and we now have a rich literature documenting how plants defend themselves against herbivore attack. In con- trast, the mechanisms by which plants resist enemies that consume floral parts or resources are much less known, even though damage to floral tissue usually has tighter links to plant fitness than damage to leaf tissue. Many plants experience nectar robbing, whereby floral visitors remove nectar from flowers, often without pollinating. Nectar robbers can reduce plant fitness to degrees comparable to, or even surpassing, reduction by herbivores. How- ever, because nectar attracts both pollinators and nectar robbers, plants face a dilemma in defending against nectar robbers without also deterring pollinators. Here, we extend the conceptual framework of resistance to herbivores to include resistance to nectar robbers, focusing on nectar traits. We review published data and find that an array of nectar traits may deter robbers without deterring pollinators. Although resistance traits against robbers have been broadly identified, the costs and benefits of these traits in terms of plant fitness remain poorly understood. We present data showing that a nectar trait (dilute nectar) might directly, as well as indirectly, benefit plant fitness by deterring nectar-robbing bumble bees of Ipomopsis aggregata without deterring hummingbird pollinators. However, the magni- tude of any plant fitness benefit will depend on the degree to which plants are pollen- vs. resource-limited in a given year. The results of our work offer both conceptual and empirical insight into how plants cope with attack by nonpollinating floral visitors through a relatively unexplored trait, nectar.

Spatial Pattern Enhances Ecosystem Functioning in an African Savanna

Termites indirectly enhance plant and animal productivity near their mounds, and the uniform spatial patterning of these mounds enhances the overall productivity of the entire landscape.

Nectar robbing in Ipomopsis aggregata : effects on pollinator behavior and plant fitness

Abstract Hummingbirds foraging in alpine meadows of central Colorado, United States, face a heterogeneous distribution of nectar rewards. This study investigated how variability in nectar resources caused by nectar-robbing bumblebees affected the foraging behavior of hummingbird pollinators and, subsequently, the reproductive success of a host plant (Ipomopsis aggregata). We presented hummingbirds with experimental arrays of I. aggregata and measured hummingbird foraging behavior as a function of known levels of nectar robbing. Hummingbirds visited significantly fewer plants with heavy nectar robbing (over 80% of available flowers robbed) and visited fewer flowers on those plants. These changes in hummingbird foraging behavior resulted in decreased percent fruit set as well as decreased total seed set in heavily robbed plants. These results indicate that hummingbird avoidance of nectar-robbed plants and flowers reduces plant fitness components. In addition, our results suggest that the mutualisms between pollinators and host plants may be affected by other species, such as nectar robbers.

Flowering phenology and compensation for herbivory in Ipomopsis aggregata

The mechanisms and circumstances that affect a plants ability to tolerate herbivory are subjects of ongoing interest and investigation. Phenological differences, and the timing of flowering with respect to pollinators and pre-dispersal seed predators, may provide one mechanism underlying variable responses of plants to herbivore damage. The subalpine wildflower, Ipomopsis aggregata, grows across a wide range of elevations and, because phenology varies with elevation, phenological delays associated with elevation may affect the ability of I. aggregata to compensate for or tolerate browsing. Thus, we examined the response of I. aggregata to herbivory across an elevation gradient and addressed the interactions among phenological delays imposed by damage, elevation, pre-dispersal seed predation and pollination, on I. aggregatas compensatory response. Among high and low elevation populations in areas near the Rocky Mountain Biological Laboratory (RMBL) in Gothic, Colorado, we compared the responses of naturally browsed, artificially browsed (clipped), and unbrowsed (control) plants of I. aggregata. We compared responses in the date of initiation of flowering, timing of peak bloom, floral display, nectar production and sugar concentration, oviposition and fruit destruction by the pre-dispersal seed predator Hylemya sp. (Anthomyiidae), fruit production, and aboveground biomass production. Clipping had the greatest effect on reproductive success and clipped plants at high elevation exhibited the lowest tolerance for herbivory. The effects of browsing appear to be mediated by flowering phenology, and both browsing and elevation delayed flowering phenology. Time needed for regrowth delays flowering, and thus affects the overlap with seed predators and pollinators. As a result of delayed flowering, naturally browsed and clipped plants incurred lower rates of seed predation. In the absence of seed predation, plants would exhibit a lower tolerance to herbivory since naturally and artificially browsed plants had fewer fruits destroyed by Hylemya larvae. We provide additional evidence that, for populations near the RMBL, clipping and natural browsing do not have the same effect on I. aggregata plants. This may be due to the selection of larger plants by herbivores. Although under some conditions plants may tolerate browsing, in areas where the growing season is short a phenological delay imposed by damage is likely to significantly reduce plant fitness. Identifying the mechanisms that allow plants to tolerate herbivore damage will help to develop a general framework for understanding the role of tolerance in plant population and community dynamics, as well as plant-herbivore interactions.

Crowding and a plant's ability to defend itself against herbivores and diseases

Populations of spider mites were reduced on cotton seedlings that had been damaged as cotyledons by a brief bout of feeding by mites. However, this induced resistance was found only for seedlings grown in relatively uncrowded conditions. The strength of induced resistance against both spider mites and Verticillium wilt diminished as plant density increased. Results were qualitatively similar when differences in foliage quantity were factored out. This suggests that induced resistance in cotton does not result from a passive removal of plant tissue but is caused by an active response. Crowded plants may be unable to express the response to damage. If this result is general, then preventing induced resistance against herbivores and diseases may represent an important and overlooked consequence and mechanism of plant competition.

Termites create spatial structure and govern ecosystem function by affecting N2 fixation in an East African savanna

The mechanisms by which even the clearest of keystone or dominant species exert community-wide effects are only partially understood in most ecosystems. This is especially true when a species or guild influences community-wide interactions via changes in the abiotic landscape. Using stable isotope analyses, we show that subterranean termites in an East African savanna strongly influence a key ecosystem process: atmospheric nitrogen fixation by a monodominant tree species and its bacterial symbionts. Specifically, we applied the 15N natural abundance method in combination with other biogeochemical analyses to assess levels of nitrogen fixation by Acacia drepanolobium and its effects on co-occurring grasses and forbs in areas near and far from mounds and where ungulates were or were not excluded. We find that termites exert far stronger effects than do herbivores on nitrogen fixation. The percentage of nitrogen derived from fixation in Acacia drepanolobium trees is higher (55-80%) away from mounds vs. near mounds (40-50%). Mound soils have higher levels of plant available nitrogen, and Acacia drepanolobium may preferentially utilize soil-based nitrogen sources in lieu of fixed nitrogen when these sources are readily available near termite mounds. At the scale of the landscape, our models predict that termite/soil derived nitrogen sources influence >50% of the Acacia drepanolobium trees in our system. Further, the spatial extent of these effects combine with the spacing of termite mounds to create highly regular patterning in nitrogen fixation rates, resulting in marked habitat heterogeneity in an otherwise uniform landscape. In summary, we show that termite-associated effects on nitrogen processes are not only stronger than those of more apparent large herbivores in the same system, but also occur in a highly regular spatial pattern, potentially adding to their importance as drivers of community and ecosystem structure.

MUTUALISM AS RECIPROCAL EXPLOITATION: AFRICAN PLANT‐ANTS DEFEND FOLIAR BUT NOT REPRODUCTIVE STRUCTURES

The foundation of many plant-ant mutualisms is ant protection of plants from herbivores in exchange for food and/or shelter. While the role of symbiotic ants in protecting plants from stem- and leaf-feeding herbivores has been intensively studied, the relationship between ant defense and measures of plant fitness has seldom been quantified. We studied ant aggression, damage by herbivores and seed predators, and fruit production among Acacia drepanolobium trees occupied by four different acacia-ant species in an East African savanna. Levels of ant aggression in response to experimental disturbance differed strongly among the four species. All four ant species recruited more strongly to new leaf growth on host plants following disturbance, while recruitment to developing fruits was on average an order of magnitude lower. Host plants occupied by more aggressive ant species suffered significantly less vegetative damage from leaf-feeding insects, stem-boring beetles, and vertebrate browsers than host plants occupied by less aggressive ant species. However, there were no differences among fruiting host plants occupied by different ant species in levels of seed predation by bruchid seed predators. Fruit production on host trees was significantly correlated with tree stem diameter but not with the identity of resident ants. Our results demonstrate that defense of host plants may differ substantially among ant species and between vegetative and reproductive structures and that fruit production is not necessarily correlated with high levels of aggression by resident ants.

Host-Plant-Mediated Interactions between a Generalist Folivore and its Tachinid Parasitoid

We investigated the consequences of feeding on different species of host-plants on the performance of a generalist lepidopteran larva (Platyprepia virginalis (Bvd.) Arctiidae) and its interactions with a tachinid fly parasitoid (Thelairia bryanti Curran). Growth rate of P. virginalis was not different for larvae reared individually on either poison hemlock, Conium maculatum L., bush lupine, Lupinus arboreus Sims, or fiddle neck, Amsinckia menziesii (Lehm.) under laboratory conditions. 3. Further, larvae caged in the field on either hemlock or lupine did not differ in developmental rates or pupal mass. In 1 of 2 years survivorship was higher for moth larvae caged on hemlock

Termites, vertebrate herbivores, and the fruiting success of Acacia drepanolobium

In African savannas, vertebrate herbivores are often identified as key determinants of plant growth, survivorship, and reproduction. However, plant reproduction is likely to be the product of responses to a suite of abiotic and biotic factors, including nutrient availability and interactions with antagonists and mutualists. In a relatively simple system, we examined the role of termites (which act as ecosystem engineers--modifying physical habitat and creating islands of high soil fertility), vertebrate herbivores, and symbiotic ants, on the fruiting success of a dominant plant, Acacia drepanolobium, in East African savannas. Using observational data, large-scale experimental manipulations, and analysis of foliar N, we found that Acacia drepanolobium trees growing at the edge of termite mounds were more likely to reproduce than those growing farther away, in off-mound soils. Although vertebrate herbivores preferentially used termite mounds as demonstrated by dung deposits, long-term exclusion of mammalian grazers did not significantly reduce A. drepanolobium fruit production. Leaf N was significantly greater in trees growing next to mounds than in those growing farther away, and this pattern was unaffected by exclusion of vertebrates. Thus, soil enrichment by termites, rather than through dung and urine deposition by large herbivores, is of primary importance to fruit production near mounds. Across all mound-herbivore treatment combinations, trees that harbored Crematogaster sjostedti were more likely to fruit than those that harbored one of the other three ant species. Although C. sjostedti is less aggressive than the other ants, it tends to inhabit large, old trees near termite mounds which are more likely to fruit than smaller ones. Termites play a key role in generating patches of nutrient-rich habitat important to the reproductive success of A. drepanolobium in East African savannas. Enhanced nutrient acquisition from termite mounds appears to allow plants to tolerate herbivory and the reduced defense by a relatively ineffective ant partner. Our results underscore the importance of simultaneously examining top-down and bottom-up effects to understand those factors most important to plant reproductive success.

Lack of a tradeoff between constitutive and induced defenses among varieties of cotton

Selecting crop plants for high yields has been associated with reduced levels of resistance against herbivores. This raises the concern that varieties selected for high levels of constitutive resistance may have reduced levels of induced defenses. Induced resistance has been hypothesized to evolve in plants when constitutive defenses are costly and not always beneficial. In this study we examined the relationship between constitutive resistance against spider mites and induced resistance against mites and verticillium wilt in ten varieties of cultivated cotton known to possess different levels of constitutive resistance. Levels of constitutive resistance against spider mites were estimated as the number of mites that were found on undamaged plants. Levels of induced resistance against mites were estimated by calculating the percent reduction in mite populations on plants that had been previously damaged relative to undamaged controls. Induced resistance against verticillium wilt was assayed by comparing the number of mitedamaged plants that recovered from exposure to the disease versus the number of undamaged control plants that recovered. We found no evidence for a negative correlation between levels of constitutive and induced resistance against mites. Similarly, we found no evidence of tradeoffs between constitutive resistance to mites and induced resistance against verticillium wilt. Selecting plants for high levels of constitutive defenses should not necessarily bring along with it the unwanted consequence of lowered levels of induced resistance.