• Asteraceae;
  • RGR;
  • toxicity


  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

1. Most theories of plant strategies assume the presence of certain ‘trade-offs’. One such evolutionary trade-off assumes a decrease in growth rate with increasing investment in chemical defences in species adapted to different levels of habitat fertility.

2. To test this hypothesis, we grew 31 herbaceous species of Asteraceae under controlled conditions of temperature (25 °C), humidity (80%), light (500 μmol m–2 s–1) and photoperiod (16 h day–1) in a modified Hoagland hydroponic solution. The plants grew from seed for 35 days post-germination and were harvested at 14, 21, 28 and 35 days. Relative growth rate (RGR) was calculated as well as a general measure of potential phytochemical toxicity (LC50) using an alcohol extraction of secondary compounds followed by Brine Shrimp bioassay and an assay of total phenolics.

3. The interspecific correlation between RGR and the potential phytochemical toxicity was weak and non-significant (rS = 0·12, P = 0·53). The correlation between RGR and total phenolics was weak, positive but significant (rS = 0·40, P = 0·03).

4. These results suggest that such an evolutionary trade-off does not exist in this group of Asteraceae.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Studies measuring the potential relative growth rates (RGR) of different species have uncovered an apparent paradox. When growing plants of species typical of resource-poor and resource-rich environments under the same controlled conditions, free of competition and with free access to mineral nutrients, those species typical of resource-poor habitats seem physiologically incapable of attaining a high RGR (e.g. Grime & Hunt 1975; Poorter & Remkes 1990). Although it is easy to understand the ecological and evolutionary advantages of a high potential RGR when growing in a resource-rich environment, it is not obvious what advantage could be obtained by a physiological incapacity to grow rapidly even when resources are abundant. The explanation suggested by Lambers & Poorter (1992) is that selection in nutrient-poor habitats favours plant attributes that increase the retention of resources already obtained and that these attributes impose physiological or morphological constraints on a high potential RGR.

In this context, many authors (Bloom, Chapin & Mooney 1985; Gulmon & Mooney 1986; Bazzaz et al. 1987; Coley 1988; Herms & Mattson 1992) have suggested that plant species adapted to resource-poor habitats should have evolved increased defences against herbivores and that the costs associated with producing such increased defences necessarily reduce their potential RGR, i.e. their relative growth rate when grown in resource-rich environments. The logic of this evolutionary hypothesis is that resource-poor habitats favour plant species with long-lived leaves because leaf replacement, and therefore the replacement of the lost mineral nutrients, is much more costly in these habitats. Therefore, species adapted to such habitats should be better defended against herbivores in order to avoid leaf losses (Fritz & Simms 1992; Simms 1992) and the re-allocation of scarce resources to these defensive attributes would reduce the potential RGR. The converse of this logical argument is that any selection for an increase in potential RGR results in a re-allocation of resources to growth and away from defence. Defence through the production of secondary compounds, and its concomitant reduction in RGR, has frequently been invoked.

Hypotheses dealing with RGR and the production of secondary compounds involve two related, but potentially different, questions. One question asks how a given plant should modify its allocation to secondary compounds vs growth as resource levels vary. This is not the question that we ask. The second question involves how species that are adapted to habitats with different resource levels should have evolved different levels of allocation to secondary compounds vs growth. Implicit in this question is the assumption that the trade-off is the result of physiological constraints between growth and defence, not constraints that are imposed from the environment. In order to measure constraints on growth that are physiological rather than the result of suboptimal resource levels, it is necessary to measure potential RGR; that is, the RGR when the plant is experiencing a constant high resource level. This second question derives from the evolutionary hypothesis, described above, that species adapted to low-resource environments have low potential RGR (i.e. a low RGR even when such plants are grown in a high-resource environment) in part because they produce higher levels of chemical defences. This second question is the topic of the present contribution.

Several studies have shown that herbivores tend to avoid species with a low potential RGR in favour of species with higher potential RGR (Coley 1982, 1983, 1988; Sheldon 1987). These studies are typically performed by measuring the amount of tissues eaten of different plant species when presented to the herbivore. There are a number of complications that make it difficult to interpret these results in the context of the stated growth vs chemical defence hypothesis. First, herbivore preference is affected both by the defensive capacity of the plant tissue and by the nutritional advantage of eating it. Because species with high RGR typically also have higher tissue nitrogen concentrations, it is not clear to what extent the negative correlation is owing to increased defensive ability vs decreased nutritional status. Second, defence against herbivores can be obtained through structural adaptations (increased leaf thickness, thick lignified cell walls, the production of fibres, spines, etc.) or alternatively through the production of toxic secondary compounds not involved in such structural adaptations. Because intact tissues are presented to the herbivore, it is not known to what extent defence is obtained by structural vs chemical means. For instance, specific leaf mass is often negatively correlated with potential RGR (Poorter & Remkes 1990) and specific leaf mass is a component in herbivore deterrence (Lucas & Pereira 1990). It is much less clear if potential RGR is negatively correlated with chemical defence. Third, many studies are based on plants growing in the field in habitats differing in resource supply rates (McKey et al. 1978; Coley 1983; 1986, 1987; Coley et al. 1985; Mauffette & Oechel 1989). Because low-resource environments would reduce observed growth rates directly, then any negative correlations between growth and defence might simply be owing to common responses of both to resource limitation, without any physiological or morphological trade-off being present.

The purpose of the present study is to test for the hypothesized negative correlation between RGR and the potential for chemical defence of plants in an interspecific context. Several conditions are required to test this evolutionary hypothesis properly. First, it is necessary to compare species grown in identical controlled conditions in order to exclude the possibility that any observed correlations are owing to common responses to changing environments. We grew 60 individuals each of 31 species in hydroponic culture under controlled conditions of nutrient supply, irradiance level, temperature and relative humidity. Second, plants can produce a large number of different chemical compounds. Not all such compounds are related to defence and the defensive potential (i.e. toxicity) of different secondary compounds varies enormously. It is therefore necessary to have a measure that integrates a large proportion of such defensive compounds rather than one that detects only one class of compounds while excluding the effects of structural defences or nutritional differences. The defensive secondary compounds of the herbaceous Asteraceae used in this experiment are relatively well-known and carbon-based (Marby & Bohnmann 1977). The compounds include sesquiterpene lactones and tri-terpenes, derivatives of caffeic acid, acetylenic compounds (except for those in the tribes Senecioneae and Cichorieae) and essential oils (except in the latex-bearing Chichorieae). They contain no true tannins and few alkaloids (none in the species used in the present study). We measured the potential toxicity of constitutive plant compounds by performing a general extraction of these compounds coupled with a general larval Brine Shrimp bioassay of plant toxicity (Meyer et al. 1982; Alkofahi et al. 1989; Arnason, Marles & Aucoin 1991) as well as the more traditional total phenolic concentration. This bioassay is widely used as an initial screening technique in pharmacology and is known to be sensitive to a large number of secondary compounds, including those found in the Asteraceae (Arnason et al. 1991; He et al. 1997). Third, it is necessary to use a test organism that is sensitive to any secondary compound produced by the plant that has a toxic potential; this is because the objective is not to quantify the actual defensive ability of a plant against a particular herbivore (that may have evolved defences against some potentially toxic compounds) but rather to measure the potential toxicity of any secondary compound. The use of the larvae of Brine Shrimp (a marine crustacean) is a standard test organism in this regard. Fourth, it is necessary to measure the whole-plant growth potential of each species rather than the growth of only one plant part. Hydroponic culture allowed us to measure whole-plant relative growth rate (RGR). Finally, it is necessary to compare species within a monophyletic group because secondary compounds have a phylogenetic component. We therefore used species from the monophyletic Asteraceae (Cronquist 1977; Palmer et al. 1988).

Materials and methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References


We worked with 31 different species from seven tribes (Table 1) of the Asteraceae. Taxonomy follows Cronquist (Gleason & Cronquist 1991). This study is restricted to herbaceous species that inhabit open sunny habitats but with differing levels of soil fertility, such as agricultural fields, meadows, waste places, roadsides and river banks. Of the 31 species, there are two biennial, 10 annual and 19 perennial growth forms (Marie-Victorin 1964).

Table 1.  . Average relative growth rates (RGR), measurable toxicity of secondary compounds in a Brine Shrimp bioassay (1/LC50; μg ml–1) and total plant phenolic content (%GAE, dry mass) of 31 herbaceous species of Asteraceae; tribes and subtribes are also shown. Plants were grown in hydroponic culture under controlled conditions of temperature (25 °C), relative humidity (80%), light intensity (500 μmol m–2s–1 of PAR) and photoperiod (16 h) and harvested during four periods post-germination. The average position along a gradient of soil nitrogen content is shown in parentheses beside the species name of 10 species. A beside the species name indicates an annual life history Thumbnail image of

Seeds were collected from wild populations across south-western Quebec, Canada. The seeds were stored in paper bags in a refrigerator at 4 °C prior to germination. We estimated the germination rates and percentages for each species prior to the experiment so as to better synchronize germination within a 1 week period.


All experiments were conducted in a Conviron (PGW36) growth chamber. Seeds were germinated on wet filter paper. Within 2–3 days of germination, seedlings were transplanted individually into separate small blocks of rock wool (2 cm × 2 cm × 4 cm) that served as a support medium. Rock wool is a mineral fibre, sterile and inert without phytotoxic substances, and commonly used in hydroponic culture. To minimize algal growth and reduce evaporation, aluminium foil was placed around each seedling on the upper surface of the rock wool. Plants were supplied with a photosynthetic photon flux density (PPFD) of 500 μmol m–2s–1 (provided by a combination of fluorescent tubes and incandescent bulbs) for 16 h each day. This provided a daily integrated photon flux of 28·8 mol m–2. The temperature was maintained at 25 °C day and 20 °C night and the relative humidity was 80%.


We used an aerated standing nutrient system consisting of 15 polyethylene containers (36cm×36cm ×30cm). Each container was divided in 144 compartments (2·5 cm × 2·5 cm × 21·5 cm) using polyethylene sheets perforated with small holes. There was therefore c. 10 cm of undivided space at the bottom of each container, thus allowing free circulation of the hydroponic solution between compartments. The four corner compartments of each container were used to introduce aeration tubes and to monitor the temperature, pH and nitrate concentration. Therefore each container held 140 plants. Each compartment contained a block of wool rock (2 cm × 2 cm × 4 cm) which functioned as a support medium and roots hung freely in the solution. Aquarium pumps were used to aerate and circulate the solution inside of each container. Each container was filled with 30 litres of modified Hoagland solution (Hoagland & Arnon 1950). This solution consisted of 3 mM Ca(NO3)2·4H2O, 2 mM NH4H2PO4, 5 mM KH2PO4, 2 mM MgSO4·7H2O, 9·07 μM MnSO4, 0·765 μM ZnSO4·7H2O, 46·4 μM H3BO3, 0·09 μM Na2MoO4.H2O, 0·01 μM CuSO4 plus 36 μM FeSO4·7H2O as iron-EDTA. The solution was topped up daily with the same solution as required to compensate for water loss as a result of evaporation and transpiration. The nutrient solution in each container was completely renewed every week; the pH of a freshly prepared solution was 6·1. The pH and the nitrate concentration was monitored daily with a NO3 selective electrode (model 800522 Orion Research Inc., Boston, MA, USA) for re-adjustment.


The experiment was in the form of randomized blocks. Each container formed one block. The 140 individuals were randomly assigned to positions within each container. One plant per species per container was randomly chosen for each harvest period (therefore 15 plants per species per harvest). For each species, 15 randomly chosen plants were harvested at each of 14, 21, 28 and 35 days after transplanting into the hydroponic system. Of these 15 plants, a sufficient number were randomly chosen for the bioassay, which required 1 g fresh mass. The remaining plants, varying from five to 13 per harvest date, were used to estimate RGR. At each harvest, plants were separated into leaves, stem, roots and reproductive tissues (if present). Leaf blades and flowers were placed in a plant press and roots and stems were placed in paper bags. These were allowed to dry at 80 °C in a forced air-drying oven to a constant dry mass for a minimum period of 48 h. Dry masses of all plant parts were measured to the nearest 0·1 mg.

The relative growth rate (RGR, g g–1day–1) of each species was estimated as the slope of the linear regression of the natural logarithm of seedling dry mass on time. Units are grams of new biomass produced per gram of pre-existing biomass per day (g g–1day–1). Thus, RGR was a mean taken over the 14–35 day growth period.


Many of the secondary compounds produced by the Asteraceae are toxic or show other significant physiological activity (Heywood, Harbone & Turner 1977). The biological hypothesis of interest to us did not involve the actual chemical defence against a particular herbivore but rather the potential chemical defence. Because a bioassay based on a herbivore that had already evolved defences against some compounds or species would not faithfully represent the potential toxicity of each species, it was necessary to use a ‘naive’ indicator species that was both sensitive to any potentially toxic compound and that would not bias the comparisons owing to previous co-evolutionary relationships. We therefore used a bioassay using Brine Shrimp larvae (nauplii of Artemia salina Leach) to measure the chemical toxicity of the extracts of each species (described in detail in Arnason et al. 1991). Brine shrimp are not natural pests of plants, so they provided a convenient invertebrate assay (Alkofahi et al. 1989). The Brine Shrimp test is routinely employed to monitor the extraction and isolation of the bioactive plant compounds (Meyer et al. 1982; Massele & Nshimo 1995; Nick, Rali & Sticher 1995; Capson, Coley & Kursar 1996; Desmarchelier et al. 1996; Demirezer & Kuruuzum 1997; He et al. 1997) including those found in the Asteraceae.

Fresh tissues (bulked by species for each harvest period) were placed in 95% ethanol for a minimum 24 h period after weighing. Ethanol was used because its intermediate polarity allows most biologically active secondary compounds of this family to be extracted, including phenolics, alkaloids, acetylenes, terpenes and other less common secondary compounds (Liskens & Jackson 1992). Leaves, stems and roots were analysed separately using what sufficient fresh biomass was available. Sometimes more than one individual was necessary for 1 g fresh mass. The tissue samples were homogenized to increase the efficiency of the extraction process. These extracts were filtered through Whatman no. 1 filter paper using a Buchner funnel and aspiration. The residue was obtained after evaporation of the ethanol in vacuum and then brought back into solution in 50% ethanol to achieve a ratio of 1 ml solvent per 1 g (fresh mass) of tissue. The final extract solutions were stored in a freezer at – 4 °C prior to bioassay to avoid lost of solvent volume.

Four-day-old Brine Shrimp nauplii were used in the bioassay. Approximately 40 nauplii in 4 ml of brine solution were placed in a test tube. A further 100 μl addition was added based on serial logarithmic dilutions (0:100, 1:99, 10:90, 100:0 μl:μl of extract:ethanol). The concentration required for 50% mortality in the Brine Shrimp assay after 24 h (LC50; μg ml–1) was calculated using probit regression (SAS, Inc. 1990). These values were then transformed to their inverse (1/LC50). Thus, larger values indicate a greater toxicity and therefore a lower concentration of tissue extract is needed to produce 50% mortality within 24 h.


Dried material was bulked per species per harvest per treatment and ground in a Brinkman mill to pass a 500 μm mesh and dried again at 80 °C for a minimum 24 h prior to use. The method used in this project of quantitative analysis for total phenolics is a modification of the Price and Butler method (Price & Butler 1977; Price, Van Scoyoc & Butler 1978). The method exploits an oxidation-reduction reaction in which the phenolate ion is oxidized. The ferric ions are reduced to the ferrous state and detected by the formation of the Prussian Blue complex (Fe4[Fe(CN)6)]3) with a potassium ferricyanide-containing reagent.

Extracts were prepared by maceration of 0·5 g of ground dried tissue in 10 ml of methanol-HCl (8% concentrated HCl in methanol) in test tubes at room temperature for 1 h. The tissue material and the extractant were initially mixed in a vortex for 2 min. This procedure improved the results of extraction. After 1 h of maceration, the samples were centrifuged at 712·5 ×g for 2 min and the supernatant analysed by the Price and Butler method. The timing of the analyses was carefully controlled in order to avoid possible errors owing to changes in absorbance with time. Because values were calibrated using gallic acid, units are percentage phenolic content (g g–1) in gallic acid equivalents (% GAE). When plant material was sufficient, three replicates of each sample were analysed. Some dilutions were carried out if necessary. Leaves, stems and roots (and flowers as present) were analysed separately. Total plant concentrations were calculated by multiplying the dry mass proportion of each tissue type by its phenolic content.


To quantify the type of soil that our species typically occupy we used data from a large vegetation survey conducted in 1995 by the second author (B.S.). The survey consisted of 224 quadrats (50 cm × 50 cm each) obtained in 23 different sites, dominated by wild herbaceous vegetation, within a 30 km radius of the University of Sherbrooke (Quebec, Canada). All species within each quadrat were identified and their abundance within the quadrat was visually estimated as percentage cover. A soil core was taken from the centre of each quadrat and total Kjeldahl nitrogen (% dry mass), available (Bray) phosphorus and exchangeable potassium (Ammonium acetate method) were measured after passing the soil through a 2 mm sieve. Soil methods are given in Bigham (1996). Only 10 of the 31 species in the present study were represented in this vegetation survey. To quantify the average soil nutrient concentrations of these 10 species, we calculated a weighted average as:

inline image

where mij is the average soil concentration of nutrient i in quadrat j, cj is the percentage cover of the species in quadrat j and the summation is over all quadrats.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

The nutrient solution was monitored daily for changes in pH and nitrate concentrations. The average pH was 5·5 with a standard deviation of 0·2. The average nitrate concentration was 7·9 mM with a standard deviation of 1·1. Samples of the hydroponic solution were taken from each container weekly and measured for toxicity using the Brine Shrimp bioassay. The values of the measurable toxicity for the hydroponic samples were never different from the controls. This means that there were no detectable secondary compounds or other toxic substances diluted in the hydroponic solution. Typical soil nutrient levels for the 10 species whose data were available ranged from 0·015 to 0·611% total nitrogen, 5–42 mg kg–1 available phosphorus and 13–68 mg kg–1 exchangeable potassium.

Table 1 gives the mean values of each measured variable for each species. RGR varied 2·1-fold between the slowest (Bidens cernua, RGR = 0·108 g g–1 day–1) and the fastest growing species (Artemisia vulgaris, RGR = 0·226 g g–1 day–1). Total phenolic concentrations varied 3·5-fold between 0·55 and 1·90% GAE (g g–1) for Senecio vulgaris and Lactuca muralis, respectively. Measurable potential toxicity (1/LC50; μg ml–1) in the Brine Shrimp test varied 133-fold between 0·01 (i.e. an LC50 above the detectable limit of 100 μg ml–1) to 0·333 (i.e. an LC50 of 3 μg ml–1) for Bidens frondosa and Achillia millefolium, respectively.

Figure 1 illustrates the relationship between the average total phenolic content and the average RGR. There was a positive non-parametric correlation between the average total phenolic content per species and its average RGR (rs = 0·40, P = 0·03) as well as with leaf phenolic content (rs = 0·47, P = 0·007) but not with root phenolic content (rs = 0·20, P = 0·28). The Spearman correlations between average RGR and total phenolic content per species for plants aged 14, 21, 28 and 35 days were 0·36, 0·38, 0·40, 0·20, respectively. Note that this result is contrary to the hypothesized direction of the correlation. Therefore, species with more phenolic compounds in their tissues (especially their leaves) tended to have higher RGR values as well, but there was no relationship between RGR and phenolic content after controlling for differences in root:shoot partitioning based on a multiple regression. Total phenolic concentration differed significantly between leaves and roots based on a paired t-test. The average total leaf phenolic concentration was 0·93% while the average for roots was 0·53%. There were not enough tissues to allow a separation of measured toxicity into root and leaf tissues.


Figure 1. . Relationship between means of RGR (g g–1day–1) and means of total soluble phenolics [% phenolic GAE (g g–1)] for 31 species of Asteraceae. Plants were grown under controlled conditions of temperature (25 °C), RH (80%), light intensity (500 μmol m–2s–1 of PAR) and photoperiod (16 h day–1) in a full-strength Hoagland hydroponic solution and harvested at days 14, 21, 28 and 35 post-germination.

Download figure to PowerPoint

There was no significant relationship between mean RGR values of these species and the mean measurable toxicity of their tissues (rs = 0·12, P = 0·53). The Spearman correlations between average RGR and measurable toxicity per species for plants aged 14, 21, 28 and 35 days were –0·06, 0·02, – 0·19 and –0·03, respectively. An ANCOVA relating measurable toxicity to plant age, controlling for species-level differences, detected no significant trend in toxicity levels over time. The mean tissue nitrogen content was negatively correlated with the mean tissue phenolic content (rs = – 0·42, P = 0·02; Fig. 2) but this trend was diluted when looking only at leaf tissues (rs = – 0·31, P = 0·09) or only at root tissues (rs = 0·18, P = 0·32). Finally, the mean measurable toxicity values per species were never significantly related to either total phenolic or nitrogen concentration, measured on a whole-plant basis or separated into leaf and root tissues.


Figure 2. . Relationship between means of total soluble phenolics [% phenolic GAE (g g–1)] and means of leaf Nitrogen content (%) for 31 species of Asteraceae. Plants were grown under controlled conditions of temperature (25 °C), RH (80%), light intensity (500 μmol m–2s–1 of PAR) and photoperiod (16 h day–1) in full-strength Hoagland hydroponic solution and harvested at days 14, 21, 28 and 35 post-germination.

Download figure to PowerPoint

It is conceivable that the species with different types of morphological defences or life histories might also differ in their degree of chemical defence. To test for relationships between morphological defences and toxicity, we nested each species within one of three types of physical defence (hairs, spines or both) and conducted a nested ANOVA. There were marginally significant differences between these three groups in terms of the total phenolic concentrations of their tissues (P = 0·05) but no differences were detected in terms of measurable toxicity. In both cases, there were significant differences (P < 0·01) between species within each physical protection type. To test for relationships between life history (annuals vs biennials or perennials) and toxicity we conducted a t-test (allowing for unequal variances). The annuals had tissues that were half as toxic on average as the other species (P = 0·006). The average toxicity of the annual species was 0·029 ± 0·005 (i.e. an LC50 of 34·5 μg ml–1) while the average toxicity of the other species was 0·057 ± 0·011 (i.e. an LC50 of 17·54 μg ml–1).

The Spearman correlation coefficients between RGR of the 10 species for which data were available and their typical values of soil nutrients were 0·84 (P = 0·005), 0·55 (P = 0·10) and 0·64 (P = 0·05), respectively, for N, P and K. The Spearman correlation coefficients between measurable toxicity of these same 10 species and their typical values of soil nutrients were 0·17, 0·11 and –0·05, respectively, for N, P and K; the probability levels are above 0·5 for all three values.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

This study combines a wide set of species grown in systematic and standardized conditions and non-limiting resources availabilities. We tested if there was any correlation between relative growth rate and potential constitutive chemical defence in 31 species of Asteraceae under controlled and enriched environmental conditions. The results show that when these species were grown under productive conditions, there is no negative interspecific correlation either between their growth potential (RGR) and the potential toxicity of their secondary compounds or between RGR and total phenolic concentration. Of our two measures, we consider the bioassay to be a better measure of potential chemical defence but have included total phenolic concentration for comparison with other studies. This is because many phenolic compounds either have no obvious defensive role or a defensive role that is related more to structural than to chemical defence (for instance, lignin). As well, the commonly used measures of total phenolics are not truly quantitative measures but only rough indices, and many nonphenolic compounds contribute to chemical defence. Although the Artemia bioassay is probably not very sensitive to compounds, such as many tannins, that serve as feeding deterrents rather than as toxins, tannins are absent from the herbaceous Asteraceae (Cronquist 1988; Waterman & Mole 1994). As with every bioasay, the results are somewhat specific to the test organism chosen. Future work, involving other naive test organisms, are needed to determine better the generality of our results.

In order to place this in an ecological context, we correlated the measurable toxicity measures and the typical soil conditions in which the species occur, based on the 10 species for which we had data. The correlations were weak and clearly non-significant. That the distributional data are ecologically realistic is shown by the strong positive correlations of these soil nutrients and the RGR values; it is known that the potential RGR of plant species are positively correlated with their typical distributions along gradients of soil fertility (Grime & Hunt 1975). There was therefore no evidence that species typical of more fertile habitats were any less well defended by their secondary compounds than those typical of more unproductive habitats. A good example of this can be found in the two species from the genus Bidens; both occur in high-light environments on nutrient-rich soils with free access to water yet one species (Bidens cernua) had highly toxic tissues while its sister species (Bidens frondosa) had no detectable toxicity in its tissues, according to the Brine Shrimp test.

Our results appear to contradict a generally held assumption in plant ecology. For instance, a screening of 43 species and 67 plant traits by Grime et al. (1997) included a measure of leaf palatability to two common generalist herbivores. In that study the correlation between RGR in productive conditions and leaf palatability was – 0·65 and the correlation between leaf palatability and yield under nitrogen stress was + 0·69. However, it seems likely that the palatability index was determined mostly by leaf morphological attributes rather than by secondary defence metabolites because the highest correlation with leaf palatability was with leaf tensile strength (– 0·72). McCanny et al. (1990) also measured a Food Quality Index for 30 herbaceous species to a generalist herbivore as well as RGR under productive conditions. Their index was obtained by extracting the plant metabolites and then mixing them in a standard diet, thus removing any effect of leaf structure or morphology. In that study, there were no significant correlations between their food quality index and either RGR or habitat fertility, in accordance with our results.

Our results do not mean that there is no trade-off between potential growth rate and a more general defensive capacity because it is possible for structural defences to impose a constraint on growth independent of secondary compounds. Similarly, our results do not mean that a negative correlation between growth and chemical defence cannot occur as the external resource supply varies. This could occur either because decreasing resource levels both decrease growth and increase the production of toxic compounds or because there is a true physiological trade-off between the two that only occurs when resources are already limiting. Rather, our results imply that there is no detectable trade-off among herbaceous species in the ability to defend plant tissues chemically vs the potential for rapid growth. It seems unlikely that the observed inability of plants typical of low-resource environments to grow rapidly even when resources are not limiting is owing to the cost associated with chemical defence.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

We thank D. Wong, N. Faucher, P. Lavoie, N. Pitre M. Rommer and C. Cooney for assistance. This research was funded by the Natural Sciences and Engineering Research Council of Canada.

  1. To whom reprint requests should be addressed.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  • 1
    Alkofahi, A., Rupprecht, J.K., Anderson, J.E., McLaughlin, J.L., Mikolajczack, K.L. & Scott, B.A. (1989) Search for new pesticides from higher plants. American Chemical Society Symposium Series 387, 2543.
  • 2
    Arnason, J.T., Marles, R.J. & Aucoin, R.R. (1991) Isolation and biological activity derived phototoxins. Photobiological Techniques (eds D. P. Valenzeno, R. H. Pottier, P. Mathis & R. H. Douglas), pp. 187–196. Plenum Press, New York.
  • 3
    Bazzaz, F.A., Chiariello, N.R., Coley, P.D. & Pitelka, L.F. (1987) Allocating resources to reproduction and defence. Bioscience 37, 5867.
  • 4
    Bigham, J.M. (1996) Methods of Soil Analysis. Part 3. Chemical Methods. Soil Science Society of America, Madison.
  • 5
    Bloom, A.J., Chapin III , F.S. & Mooney, H.A. (1985) Resource limitation in plants—an economic analogy. Annual Review of Ecology and Systematics 16, 363392.
  • 6
    Capson, T.L., Coley, P.D. & Kursar, T.A. (1996) A new paradigm for drug discover in tropical rainforests. Nature Biotechnology14, 1200.
  • 7
    Coley, P.D. (1982) Rates of herbivory on different tropical trees. Ecology of a Tropical Forest: Seasonal Rhythms and Long-Term Changes (eds E. G. Leigh, A. S. Leigh & D. M. Windsor), pp. 123–132. Smithsonian Institution, Washington, DC.
  • 8
    Coley, P.D. (1983) Herbivory and defensive characteristics of tree species in a lowland tropical forest. Ecological Monography 53, 209233.
  • 9
    Coley, P.D. (1986) Costs and benefits of defense by tannins in a neotropical tree. Oecologia 70, 238241.
  • 10
    Coley, P.D. (1987) Interspecific variation in plant anti-herbivore properties: the role of habitat quality and rate of disturbance. New Phytologist 106 (suppl.), 251263.
  • 11
    Coley, P.D. (1988) Effects of plant growth rate and leaf lifetime on the amount and type of anti-herbivore defence. Oecologia 74, 531536.
  • 12
    Coley, P.D., Bryant, J.P. & Chapin III F.S. (1985) Resource availability and plant antiherbivore defense. Science 230, 895899.
  • 13
    Cronquist, A. (1977) The compositae revisited. Brittonia 29, 137153.
  • 14
    Cronquist, A. (1988) The Evolution and Classification of Flowering Plants, 2nd edn. The New York Botanical Garden, New York.
  • 15
    Demirezer, O.L. & Kuruuzum, A. (1997) Rapid and simple biological activity screening of Rumex species—evaluation of bioguided fractions of r-scutatus and pure compounds. A Journal of Biosciences 52, 665669.
  • 16
    Desmarchelier, C., Mongelli, E., Coussio, J. & Ciccia, G. (1996) Studies on the cytotoxicity, antimicrobial and DNA-binding activities of plants used by the eseejas. Journal of Ethnopharmacology 50, 9196.
  • 17
    Fritz, R.S. & Simms, E.L. (1992) Plant Resistance to Herbivores and Pathogens: Ecology, Evolution, and Genetics. University of Chicago Press, Chicago.
  • 18
    Gleason, H.A. & Cronquist, A. (1991) Manual of Vascular Plants of Northeastern United States and Adjacent Canada. Van Norstand, Princeton, NJ.
  • 19
    Grime, J.P. & Hunt, R. (1975) Relative growth rate: its range and adaptive significance in a local flora. Journal of Ecology 63, 393422.
  • 20
    Grime, J.P., Thompson, K., Hunt, R., Hodgson, J.G., Cornelissen, J.H.C., Rorison, I.H., Hendry, G.A.F., Ashenden, T.W., Askew, A.P., Band, S.R., Booth, R.E., Bossard, C.C., Campbell, B.D., Cooper, J.E.L., Davison, A.W., Gupta, P.L., Hall, W., Hand, D.W., Hannah, M.A., Hillier, S.H., Hodkinson, D.J., Jalili, A., Liu, Z., Mackey, J.M.L., Matthews, N., Mowforth, M.A., Neal, A.M., Reader, R.J., Reiling, K., Ross-Fraser, W., Spencer, R.E., Sutton, F., Tasker, D.E., Thorpe, P.C. & Whitehouse, J. (1997) Integrated screening validates primary axes of specialisation in plants. Oikos 79, 259281.
  • 21
    Gulmon, S.L. & Mooney, H.A. (1986) Costs of defence and their effects on plant productivity. On the Economy of Plant Form and Function (ed. T. J. Givnish), pp. 681–698. Cambridge University Press, Cambridge.
  • 22
    He, K., Zeng, L., Ye, Q., Shi, G.E., Oberlies, N.H., Zhao, G.X., Njoku, J. & McLaughlin, J.L. (1997) Comparative SAR evaluations of annonaceous acetogenins for pesticidal activity. Pestic Science 49, 372378.
  • 23
    Herms, D.A. & Mattson, W.J. (1992) The dilemma of plants: to grow or defend. Quarterly Review of Biology 67, 283335.
  • 24
    Heywood, V.H., Harbone, J.B. & Turner, B.L. (1977) The Biology and Chemistry of the Compositae, Vol 1. Academic Press, New York.
  • 25
    Hoagland, D.R. & Arnon, D.I. (1950) The water culture method of growing plants without soil. California Agricultural Experiment Station Circular 347.
  • 26
    Lambers, H. & Poorter, H. (1992) Inherent variation in growth rate between higher plants: a search for physiological causes and ecological consequences. Advances in Ecological Research 21, 187261.
  • 27
    Liskens, H.F. & Jackson, J.F. (1992) Plant Toxin Analysis. Springer-Verlag, New York.
  • 28
    Lucas, P.W. & Pereira, B. (1990) Estimation of the fracture toughness of leaves. Functional Ecology 4, 819822.
  • 29
    Marby, T.J. & Bohlmann, F. (1977) Summary of the chemistry of the compositae. Biology and Chemistry of the Compositae, vol. 1 (eds V. H. Heywood, J. B. Harbone & B. L. Turner), pp. 1097–1104. Academic Press, New York.Marie-Victorin, F. (1964) Flore Laurentienne, 2nd edn. Les Presses de l’Université de Montreal, Montreal.
  • 30
    Massele, A.Y. & Nshimo, C.M. (1995) Brine shrimp bioassay for biological activity of medicinal plants used in traditional medicines in Tanzania. East African Medical Journal 72, 661663.
  • 31
    Mauffette, Y. & Oechel, W.C. (1989) Seasonal variation in leaf chemistry of the coast liveoak Quercus argrifolia and implications for the California oak moth Phryganidia californica. Oecologia 79, 439445.
  • 32
    McCanny, J.S., Keddy, P.A., Arnason, T.J., Gaudet, C.L., Moore, D.R.J. & Shipley, B. (1990) Fertility and the food quality of wetland plants: a test of the resource availability hypotheses. Oikos 59, 373381.
  • 33
    McKey, D., Waterman, P.G., Mbi, C.N., Gartlan, J.S. & Struhsaker, T.T. (1978) Phenolic content of vegetation in two African rain forests: ecological implications. Science 202, 6163.
  • 34
    Meyer, B.N., Ferrigni, N.R., Putnam, J.E., Jacobsen, L.B., Nichols, D.E. & McLaughlin, J.L. (1982) Brine shrimp: a convenient general bioassay for active plant constituents. Journal of Medice Plant Research 45, 3134.
  • 35
    Nick, A., Rali, T. & Sticher, O. (1995) Biological screening of traditional medicinal plants from Papua New Guinea. Journal of Ethnopharmacology 49, 147156.
  • 36
    Palmer, J.D., Jansen, R.K., Michaels, H.J., Chase, M.W. & Manhart, J.R. (1988) Chloroplast DNA variation and plant phylogeny. Annals of Missouri Botany Garden 75, 11801206.
  • 37
    Poorter, H. & Remkes, C. (1990) Leaf area ratio and net assimilation rate of 24 wild species differing in relative growth rate. Oecologia 83, 553559.
  • 38
    Price, M.L. & Butler, L.G. (1977) Rapid visual estimation and spectrophotometric determination of tannin content of sorghum grain. Journal of Agriculture Food Chemistry 25, 12681273.
  • 39
    Price, M.L., Van Scoyoc, S. & Butler, L.G. (1978) A critical evaluation of the vanillin reaction as an assay for tannin in sorghum grain. Journal of Agriculture Food Chemistry 26, 12141218.
  • 40
    Sheldon, S.P. (1987) The effects of herbivorous snails on submerged macrophyte communities in Minnesota lakes. Ecology 68, 19201931.
  • 41
    Simms, E.L. (1992) Costs of plant resistance to herbivory Plant Resistance to Herbivores and Pathogens: Ecology, Evolution, and Genetics (eds R. S. Fritz & E. L. Simms), pp. 392–425. University of Chicago Press., Chicago.
  • 42
    Waterman, P.G. & Mole, S. (1994) Analysis of Phenolic Plant Metabolites. Blackwell, Oxford.