Novel bioresources for studies of Brassica oleracea: identification of a kale MYB transcription factor responsible for glucosinolate production

Plants belonging to the Brassicaceae family exhibit species‐specific profiles of glucosinolates (GSLs), a class of defence compounds against pathogens and insects. GSLs also exhibit various human health–promoting properties. Among them, glucoraphanin (aliphatic 4‐methylsulphinylbutyl GSL) has attracted the most attention because it hydrolyses to form a potent anticancer compound. Increased interest in developing commercial varieties of Brassicaceae crops with desirable GSL profiles has led to attempts to identify genes that are potentially valuable for controlling GSL biosynthesis. However, little attention has been focused on genes of kale (Brassica oleracea var. acephala). In this study, we established full‐length kale cDNA libraries containing 59 904 clones, which were used to generate an expressed sequence tag (EST) data set with 119 204 entries. The EST data set clarified genes related to the GSL biosynthesis pathway in kale. We specifically focused on BoMYB29, a homolog of Arabidopsis MYB29/PMG2/HAG3, not only to characterize its function but also to demonstrate its usability as a biological resource. BoMYB29 overexpression in wild‐type Arabidopsis enhanced the expression of aliphatic GSL biosynthetic genes and the accumulation of aliphatic GSLs. When expressed in the myb28myb29 mutant, which exhibited no detectable aliphatic GSLs, BoMYB29 restored the expression of biosynthetic genes and aliphatic GSL accumulation. Interestingly, the ratio of methylsulphinyl GSL content, including glucoraphanin, to that of methylthio GSLs was greatly increased, indicating the suitability of BoMYB29 as a regulator for increasing methylsulphinyl GSL content. Our results indicate that these biological resources can facilitate further identification of genes useful for modifications of GSL profiles and accumulation in kale.     

Brassinosteroids fine-tune secondary and primary sulfur metabolism through BZR1-mediated transcriptional regulation

For adaptation to ever-changing environments,plants have evolved elaborate metabolic systems coupled to a regulatory network for optimal growth and defense. Regulation of plant secondary metabolic pathways such as glucosinolates(GSLs) by defense phytohormones in response to different stresses and nutrient deficiency has been intensively investigated, while how growth-promoting hormone balances plant secondary and primary metabolism has been largely unexplored. Here, we found that growth-promotin...     

Can plants use an entomopathogenic virus as a defense against herbivores?

It is by now well established that plants use various strategies to defend themselves against herbivores. Besides conventional weapons such as spines and stinging hairs and sophisticated chemical defenses, plants can also involve the enemies of the herbivores in their defense. It has been suggested that plants could even use entomopathogens as part of their defense strategies. In this paper, we show that Brassica oleraceae plants that are attacked by Myzus persicae aphids infected with an entomopathogenic parvovirus (M. persicae densovirus) transport the virus through the phloem locally and systematically. Moreover, healthy aphids that fed on the same leaf, but separated from infected aphids were infected via the plant. Hence, this is proof of the principle that plants can be vectors of an insect virus and can possibly use this virus as a defense against herbivores.     

COSTS OF INDUCED RESPONSES AND TOLERANCE TO HERBIVORY IN MALE AND FEMALE FITNESS COMPONENTS OF WILD RADISH

Theory predicts that plant defensive traits are costly due to trade-offs between allocation to defense and growth and reproduction. Most previous studies of costs of plant defense focused on female fitness costs of constitutively expressed defenses. Consideration of alternative plant strategies, such as induced defenses and tolerance to herbivory, and multiple types of costs, including allocation to male reproductive function, may increase our ability to detect costs of plant defense against herbivores. In this study we measured male and female reproductive costs associated with induced responses and tolerance to herbivory in annual wild radish plants (Raphanus raphanistrum). We induced resistance in the plants by subjecting them to herbivory by Pieris rapae caterpillars. We also induced resistance in plants without leaf tissue removal using a natural chemical elicitor, jasmonic acid; in addition, we removed leaf tissue without inducing plant responses using manual clipping. Induced responses included increased concentrations of indole glucosinolates, which are putative defense compounds. Induced responses, in the absence of leaf tissue removal, reduced plant fitness when five fitness components were considered together; costs of induction were individually detected for time to first flower and number of pollen grains produced per flower. In this system, induced responses appear to impose a cost, although this cost may not have been detected had we only quantified the traditionally measured fitness components, growth and seed production. In the absence of induced responses, 50% leaf tissue removal, reduced plant fitness in three out of the five fitness components measured. Induced responses to herbivory and leaf tissue removal had additive effects on plant fitness. Although plant sibships varied greatly (49–136%) in their level of tolerance to herbivory, costs of tolerance were not detected, as we did not find a negative association between the ability to compensate for damage and plant fitness in the absence of damage. We suggest that consideration of alternative plant defense strategies and multiple costs will result in a broader understanding of the evolutionary ecology of plant defense.     

Interactions between plants and herbivores: A review of plant defense

Ecologists have long ignored or underestimated the importance of plant–herbivore interactions owing to the diversities of herbivores, plant defensive strategies and ecological systems. In this review, we briefly discussed the categories of herbivores. Then we reviewed the major types of plant defenses against herbivores. Selective forces of herbivore pressures have led to the evolution of various defensive mechanisms in plants, which can be classified into (i) resistance traits that reduce the amount of damage received, including physical, chemical, and biotic traits; (ii) tolerance mechanisms that decrease the impact of herbivore damage, and (iii) escape strategies that reduce the probability of plants to be found by herbivores. These strategies have been studied at different levels from molecular genetics and genomics, to chemistry and physiology, to community and ecosystem ecology. We summarized the development of the methodology for studying plant defenses against herbivores. Particularly, 24 of those hypotheses and models, which are influential in the international community concerning the relationship between plants and herbivores, including the defensive mimicry hypothesis, the compensatory continuum hypothesis, the slow-growth-high-mortality hypothesis, etc, were introduced and grouped into four categories according to plant defense strategies in the present review. Finally, we also reviewed the research progress of plant–herbivore interactions in China, and discussed the perspectives of studies on plant–herbivore interactions.     

Combined effect of Nitrogen,Phosphorus and Potassium fertilizers on the contents of glucosinolates in rocket salad (Eruca sativa Mill.)

Nitrogen (N), phosphorous (P) and potassium (K) are the most limiting factors in crop production. N often affects the amino acid composition of protein and in turn its nutritional quality. In Brassica plants, abundant supply of N fertilizer decreases the relative proportion of glucosinolates (GSLs), thus reducing the biological and medical values of the vegetables. Hence effort was made to evaluate the influence of different proportions of nutrient solutions containing N–P–K on the GSL profiles of rocket salad (Eruca sativa Mill.). Fifteen desulpho-(DS) GSLs were isolated and identified using liquid chromatography–mass spectrometry (LC/MS) analysis. Rocket salad plants supplied with lesser amount of N, P or higher concentrations of K showed a typical improvement in total GSL contents. In contrast, total GSL levels were less at higher N supply. Furthermore, with N concentrations above 5 mM and K concentrations less than 2.5 mM, the GSL amounts were on average 13.51 and 13.75 μmol/g dry weight (DW), respectively. Aliphatic GSLs predominated in all concentrations of NPK while indolyl GSLs made up marginally less amount of the total compositions. Five and 2 mM N and P possessed much higher levels of several types of aliphatic GSLs than other concentrations, including glucoerucin, glucoraphanin and dimeric 4-mercaptobutyl GSL. From this perspective, it is contended that supply of less N results in enhancing the metabolic pathway for the synthesis of GSLs in rocket salad.     

Two‐tier morpho‐chemical defence tactic in Aethionema via fruit morph plasticity and glucosinolates allocation in diaspores

Fruit dimorphism and the production of glucosinolates (GSLs) are two specific life history traits found in the members of Brassicales, which aid to optimize seed dispersal and defence against antagonists, respectively. We hypothesized that the bipartite dispersal strategy demands a tight control over the production of fruit morphs with expectedly differential allocation of defensive anticipins (GSLs). In dimorphic Aethionema, herbivory by Plutella xylostella at a young stage triggered the production of more dehiscent (seeds released from fruit) than indehiscent fruit morphs (seeds enclosed within persistent pericarp) on the same plant upon maturity. Total GSL concentrations were highest in the mature seeds of dehiscent fruits from Aethionema arabicum and Aethionema saxatile among the different ontogenetic stages of the diaspores. Multivariate analyses of GSL profiles indicated significantly higher concentrations of specific indole GSLs in the diaspores, which require optimal defence after dispersal (i.e., seeds of dehiscent and fruit/pericarp of indehiscent fruit). Bioassays with a potentially coinhabitant fungus, Aspergillus quadrilineatus, support the distinct defensive potential of the diaspores corresponding to their GSL allocation. These findings indicate a two‐tier morpho‐chemical defence tactic of Aethionema via better protected fruit morphs and strategic provision of GSLs that optimize protection to the progeny for survival in nature.     

The effects of qualitative and quantitative variation of aristolochic acids on preference and performance of a generalist herbivore

Nearly all plants possess chemicals that are inferred to play a role in anti‐herbivore defense or resistance. The effects of various chemical defenses can vary among herbivores. Often, plant defensive compounds are examined in broad, inclusive categories, with an emphasis on total quantity, which might ignore qualitative variation in activity. Aristolochic acids are alkaloids characteristic of plants of the genus Aristolochia (Aristolochiaceae). Although aristolochic acids have been documented as effective herbivore deterrents, it remains unknown whether different kinds of aristolochic acid vary in their efficacy as defense against herbivores. We manipulated the aristolochic acid content of artificial diet to examine the effects of four aristolochic acids on larval preference and performance of the generalist herbivore Spodoptera exigua Hübner (Lepidoptera: Noctuidae). Using choice tests, we observed that the four aristolochic acids tested varied in their deterrent effectiveness, with AA‐I having the strongest effect and AA‐II having the weakest effect. No‐choice tests were used to examine larval performance. The effect on performance varied among the aristolochic acids tested. Higher concentrations of aristolochic acid were generally associated with reduced larval developmental rate and larger size at pupation. These results indicate that various forms of aristolochic acid can vary in their effect on herbivores and that simple aggregate measures of total concentration might not reflect the chemical defensive phenotype of the plant.     

Performance of specialist and generalist herbivores feeding on cabbage cultivars is not explained by glucosinolate profiles

Plants display a wide range of chemical defences that may differ in effectiveness against generalist and specialist insect herbivores. Host plant‐specific secondary chemicals such as glucosinolates (GS) in Brassicaceae typically reduce the performance of generalist herbivores, whereas specialists have adaptations to detoxify these compounds. The concentration of glucosinolates may also alter upon herbivory, allowing the plant to tailor its response to specifically affect the performance of the attacking herbivore. We studied the performance of three Lepidoptera species, two specialists [Pieris rapae L. (Pieridae), Plutella xylostella L. (Yponomeutidae)] and one generalist [Mamestra brassicae L. (Noctuidae)], when feeding on eight cultivars of Brassica oleracea L. and a native congener (Brassica nigra L.) and related this to the GS content. We tested the hypotheses (i) that a generalist herbivore is more affected by high GS concentrations, and (ii) that generalist feeding has a stronger effect on GS levels. Although performance of the three herbivores was different on the B. oleracea cultivars, M. brassicae and P. xylostella had a similar ranking order of performance on the eight cultivars. In most of the cultivars, the concentration of indole GS was significantly higher after feeding by P. rapae or M. brassicae than after P. xylostella feeding. As a consequence, the total concentration of GS in the cultivars showed a different ranking order for each herbivore species. The generalist M. brassicae performed equally well as the specialist P. xylostella on cultivars with high concentrations of GS. Our findings suggest that secondary metabolites other than GSs or differences in nutrient levels affect performance of the species studied.     

Evolutionary irony: evidence that ‘defensive’ plant spines act as a proximate cue to attract a mammalian herbivore

Many plants produce structural defenses to deter feeding by herbivores. However, many previous studies testing whether spines are effective at defending against mammalian herbivores have produced equivocal results. These ambiguous results are hypothesized to be due to herbivore counter‐adaptations. We investigated potential counter‐adaptations in a population of white‐throated woodrats Neotoma albigua that specialize on cactus by investigating feeding behavior and preference for cacti varying in spinescence. Neotoma albigula exhibited a unique behavior of clipping cactus spines, which renders these defenses ineffective. Strikingly, these woodrats chose to collect spiny cacti over experimentally de‐spined cacti, demonstrating that spines act as a proximal cue that attracts woodrats. This attraction is likely due to the higher protein and lower fiber content of spiny cacti compared to naturally non‐spiny cacti. Thus, the ‘defensive’ spines of cacti are ineffective against a specialist herbivore and instead serve as an indicator of nutritional quality that promotes herbivory. Our results support the ‘rule‐of‐thumb’ hypothesis of foraging, which states that herbivores forage according to obvious visual cues that are indicative of nutritional content, rather than sampling nutrient composition of plants. We propose that specialist herbivores are unique systems in which to study other counter‐adaptations to structural defenses and ‘rule‐of‐thumb’ foraging strategies.     

Optimal strategy of plant antiherbivore defense: Implications for apparency and resource-availability theories

Plants produce chemicals as methods against animal herbivory. Such chemical defenses are classified into two major categories: (i) quantitative defenses with massive production of indigestible substances; and (ii) qualitative defenses with production of poisonous substances. A mathematical model was developed that identified factors that favored the evolution of quantitative defenses. Selecting an annual plant for simplicity, the allocation of photosynthetic production between growth substances and defense substances was considered. If the plant invests more in defense substances, it can protect itself more efficiently from herbivory but with a reduced growth rate. If it invests more in growth substances, the contrary holds. Using Pontoryagin's maximum principle, the following results were obtained: (i) the plant should conduct quantitative defenses when the growth rate (G), reflecting resource-availability, is low and the growth period (T) is long as well; (ii) if the plant invests in quantitative defenses, the optimal proportion of defense substances (χ^*) should be higher asG is smaller, but it is independent ofT; and (iii) the value of χ^* is not monotone for the effectiveness of defense substance (A), but has a maximum at an intermediate value ofA. Predictions of the model partly supported both Feeny's apparency theory, claiming that apparent plants or their parts for herbivores should quantitatively defend themselves, and Coley's resource-availability theory, claiming that plants with rich resources should invest in growth rather than defense.     

Overexpression of Three Glucosinolate Biosynthesis Genes in Brassica napus Identifies Enhanced Resistance to Sclerotinia sclerotiorum and Botrytis cinerea

Sclerotinia sclerotiorum and Botrytis cinerea are notorious plant pathogenic fungi with an extensive host range including Brassica crops. Glucosinolates (GSLs) are an important group of secondary metabolites characteristic of the Brassicales order, whose degradation products are proving to be increasingly important in plant protection. Enhancing the defense effect of GSL and their associated degradation products is an attractive strategy to strengthen the resistance of plants by transgenic approaches. We generated the lines of Brassica napus with three biosynthesis genes involved in GSL metabolic pathway (BnMAM1, BnCYP83A1 and BnUGT74B1), respectively. We then measured the foliar GSLs of each transgenic lines and inoculated them with S. sclerotiorum and B. cinerea. Compared with the wild type control, over-expressing BnUGT74B1 in B. napus increased the aliphatic and indolic GSL levels by 1.7 and 1.5 folds in leaves respectively; while over-expressing BnMAM1 or BnCYP83A1 resulted in an approximate 1.5-fold higher only in the aliphatic GSL level in leaves. The results of plant inoculation demonstrated that BnUGT74B1-overexpressing lines showed less severe disease symptoms and tissue damage compared with the wild type control, but BnMAM1 or BnCYP83A1-overexpressing lines showed no significant difference in comparison to the controls. These results suggest that the resistance to S. sclerotiorum and B. cinerea in B. napus could be enhanced through tailoring the GSL profiles by transgenic approaches or molecular breeding, which provides useful information to assist plant breeders to design improved breeding strategies.     

Quantifying the response of free-ranging mammalian herbivores to the interplay between plant defense and nutrient concentrations

While trying to achieve their nutritional requirements, foraging herbivores face the costs of plant defenses, such as toxins. Teasing apart the costs and benefits of various chemical constituents in plants is difficult because their chemical defenses and nutrient concentrations often co-vary. We used an approach derived from predator–prey studies to quantitatively compare the foraging response of a free-ranging mammalian herbivore, the swamp wallaby (Wallabia bicolor), through three feeding trials with artificial diets that differed in their concentrations of (1) the terpene 1,8-cineole, (2) primary constituents (including nitrogen and fiber), and (3) both the terpene and the primary constituents. Applying the giving-up density (GUD) framework, we demonstrated that the foraging cost of food patches increases with higher dietary cineole concentration and decreases with higher dietary nutrient concentration. The effect of combined differences in nutrients and cineole concentrations on GUD was interactive, and high nutrient food required more cineole to achieve the same patch value as low nutrient food. Our results indicate that swamp wallabies equate low nutrient, poorly defended food with high nutrient, highly defended food, providing two contrasting diets with similar cost–benefit outcomes. This behavior suggests that equal concentrations of chemical defenses provide nutrient-poor plants with relatively greater protection as nutrient-rich plants. Nutrient-rich plants may therefore face the exacerbated problem of being preferred by herbivores and therefore need to produce more defense compounds to achieve the same level of defense as nutrient-poor plants. Our findings help explain the difference in anti-herbivore strategy of nutrient-poor and rich plants, i.e., tolerance versus defense.     

Is extrafloral nectar production induced by herbivores or ants in a tropical facultative ant–plant mutualism?

Many plants use induced defenses to reduce the costs of antiherbivore defense. These plants invest energy in growth when herbivores are absent but shunt energy to defense when herbivores are present. In contrast, constitutive defenses are expressed continuously regardless of herbivore presence. Induction has been widely documented in temperate plants but has not been reported from tropical plants. Most tropical plants have higher, more constant herbivore pressure than temperate plants. In this situation, it is hypothesized that constitutive defenses rather than induced defense would be favored. Using natural herbivores of four species of Inga saplings on Barro Colorado Island, Panama, herbivore presence was crossed with ant presence to determine their effects on extrafloral nectar production. Analysis of nectar samples revealed that Inga species do not induce nectar production in response to herbivores. This result is not due to an inability of the plants to respond, as the plants in this study increased nectar production in response to light and ant presence. Contrary to most induction experiments with temperate ecosystem plants, these results demonstrate that tropical plants do not induce one type of defense, and they suggest that the most adaptive defense strategies are different for the two ecosystems.     

Induced defense in Nicotiana attenuata (Solanaceae) fruit and flowers

Plants protect themselves against herbivory using a continuum of strategies, ranging from constitutive defenses to intermittent induced responses. Induced defenses may not provide immediate and maximum protection, but could be advantageous when continuous defense is either energetically or ecologically costly. As such, induced defenses in flowers could help defend relatively valuable tissue while keeping reproductive structures accessible and attractive to pollinators. Thus far, no one has demonstrated the efficacy of induced defenses against floral herbivores (florivores) in the field. Here we show that mechanical leaf damage in wild tobacco, Nicotiana attenuata (Solanaceae), reduced both flower and fruit herbivory in the field and that exogenous application of methyl jasmonate, a potent elicitor of induced responses, reduced both leaf and floral damage in natural populations. This result is consistent with a survey of damage in the field, which showed a negative relationship between leaf damage and flower and fruit damage. Although optimal defense theory predicts that induced defenses should be rare in reproductive tissues, owing to their high fitness value, our results suggest otherwise. Induced defenses in leaves and reproductive tissues may allow plants to respond effectively to the concomitant pressures of defending against herbivory and attracting pollinators.     

Parasitoid-plant mutualism: parasitoid attack of herbivore increases plant reproduction

We tested whether a plant's life time seed production is increased by parasitization of herbivores in a tritrophic system, Arabidopsis thaliana (Brassicaceae) plants, Pieris rapae (Lepidoptera: Pieridae) caterpillars and the solitary endoparasitoid Cotesia rubecula (Hymenoptera: Braconidae). We established seed production for intact A. thaliana plants, plants that were mechanically damaged, plants fed upon by parasitized caterpillars and plants fed upon by unparasitized caterpillars. In the first experiment, with ecotype Landsberg (erecta mutant), herbivory by unparasitized P. rapae caterpillars resulted in a strongly reduced seed production compared to undamaged plants. In contrast, damage by P. rapae caterpillars that had been parasitized by C. rubecula did not result in a significant reduction in seed production. For the second experiment with the ecotype Columbia, the results were identical. Plants damaged by unparasitized caterpillars only produced seeds on regrown shoots. Seed production of plants that had been mechanically damaged was statistically similar to that of undamaged plants. Production of the first ripe siliques by plants fed upon by unparasitized caterpillars was delayed by 18–22 days for Landsberg and 9–10 days for Columbia. We conclude that parasitization of P. rapae by C. rubecula potentially confers a considerable fitness benefit for A. thaliana plants when compared to plants exposed to feeding damage by unparasitized P. rapae larvae. Plants that attract parasitoids and parasitoids that respond to herbivore-induced plant volatiles will both experience selective advantage, justifying the use of the term mutualism for this parasitoid-plant interaction. This type of mutualism is undoubtedly very common in nature.     

Host plant and the predator Podisus nigrispinus: when the defense compounds of the plant affect the third trophic level

Many herbivorous insects can overcome chemical plant defenses, using the plant's defensive products for their own good, as a defense against predators. Eucalyptus spp. (Myrtaceae), recently introduced in Brazil, are rich in secondary compounds; however, there are reports that these plants have been suffering from population outbreaks of defoliating Lepidoptera in Brazil. The predator Podisus nigrispinus (Dallas) (Hemiptera: Pentatomidae) has been used against herbivorous insects in eucalyptus plantations, but little is known about its establishment in the field. This study aims to investigate whether the effectiveness of this predator may be affected indirectly by compounds of eucalyptus plants, when compared to guava, Psidium guajava L., a Brazilian native species of Myrtaceae. Thus, we evaluated the performance of P. nigrispinus on larvae of Thyrinteina arnobia (Stoll) (Lepidoptera: Geometridae) reared on eucalyptus (exotic species) or guava plants (native species). Podisus nigrispinus performance (reproduction and survival) was better on larvae fed on guava than on larvae fed on eucalyptus. It is possible that the negative effect on the predator's development occurred because of the plants’ secondary compounds appropriated by caterpillars, due to the short coevolutionary history between eucalyptus and the predator. The data suggest that the chemical compounds that could help the plant's defenses against herbivores may also affect their natural enemies, especially when the interaction between plant and natural enemy involves an exotic plant recently introduced into the insect's habitat.     

Simultaneous evolution of competitiveness and defense: induced switching in Arabis drummondii

Optimality theory for plant defense against herbivores predicts an evolutionary tradeoff between the abilities to compete and defend. We tested this hypothesis by studying the effects of genetic variation in competitiveness on defense expression. Two closely related and differentially competitive congeners were compared for levels of resistance, tolerance, and secondary metabolite production. In a growth room experiment, plants of Arabis drummondii and A. holboellii were grown in the presence and absence of the common bunch grass Boutelloua gracilis, the specialist herbivore Plutella xylostella, and generalist herbivore Trichoplusia ni. Tolerance to competition, measured as growth next to the grass relative to controls in the absence of grass, was greatest for A. drummondii, the species that occurred in communities with higher densities of inter-specific neighbors. Measures of defense (resistance to herbivores, tolerance to damage, and concentrations of glucosinolates) varied inconsistently between the Arabis, species, depending on type of herbivore, competition level, and type of defense. The better competitor A. drummondii was more resistant to specialist herbivores, as in the field, and exhibited greater herbivore- and competition-induced changes in glucosinolate profiles. Further, when plants of A. drummondii were fed upon in competitive environments, the induced glucosinolate response was reduced while tolerance levels increased in an apparent switching of induced strategies. We suggest that competitiveness and defense responses are sometimes positively correlated because some defensive traits also function as competitive traits. A competitive function for defenses may also explain why defenses were affected by competition. Alternatively, since the induced response did not increase estimates of total glucosinolate content significantly, minimal defense costs might also allow the simultaneous evolution of competitiveness and defense. Finally, when faced with both herbivory and competition, some competitive species, such as A. drummondii, may switch to growth-based rather than toxin-based strategies as recent theoretical models predict.     

The defensive role of Ni hyperaccumulation by plants: a field experiment

Hyperaccumulation of Ni by plants is hypothesized to function as an elemental defense against herbivores and pathogens. Laboratory experiments have documented toxic effects to herbivores consuming high-Ni plant tissues, but this paper reports the first experiment to examine the defensive effectiveness of Ni hyperaccumulation under field conditions. The experiment was conducted at an ultramafic soil site naturally inhabited by the Ni hyperaccumulator Streptanthus polygaloides (Brassicaceae). Experimental treatments examined the response of herbivores to hyperaccumulated Ni, using exclosure and insecticide treatments to divide herbivores into groups based primarily upon herbivore size. Three soils (Ni-amended greenhouse soil, unamended greenhouse soil, ultramafic soil), three exclosure treatments (exclosure, control exclosure, no exclosure), and a systemic insecticide treatment were combined in a fractional factorial experimental design. Streptanthus polygaloides plants were grown in a greenhouse for 2 mo, transplanted into the field by inserting potted plants into holes dug on the experimental site, and periodically examined for herbivore damage during a 41-d period. Initial surveys showed greater amounts of insect damage to plants with low tissue Ni levels, confirming the defensive effect of Ni against some insect herbivores, but large herbivores (probably vertebrates) later consumed entire plants regardless of plant Ni status. We concluded that Ni was not an effective defense against these large herbivores, probably because their diets mix high-Ni S. polygaloides foliage with that of associated non-hyperaccumulating species. We suggest that such dietary dilution is one mechanism whereby some herbivores can circumvent elemental plant defenses.