Evidence for Coevolution between Mammalian Herbivores and Plant Secondary Compounds in Boreal Forests


Troy Gerhardt                                                                    
April 24, 1994

Colorado State University
Fort Collins, Colorado 80523

Introduction

      Browsing mammals are generally highly selective feeders, consuming 
only a small portion of the plant species and plant parts available to 
them.  The basis for this selectivity remains relatively unknown, 
despite attempts to explain diet selection in terms of nutrients or 
energy (Bryant and Kuropat 1980, Palo and Robbins 1991).  Recently, the 
importance of plant secondary compounds as determinants of browsing 
patterns has been recognized (Palo and Robbins 1991).  Such chemically 
mediated interactions between mammalian herbivores and woody plants may 
lead to evolutionary responses in both the browsers and the food plants.  
A major evolutionary response of plants to herbivory has been a 
production and diversification of secondary compounds that deter 
feeding.

     Browser-plant interactions in boreal regions are of interest 
because of the combination of often intense herbivory by several 
mammalian species and because of the relatively low temperatures, light 
levels, and nutrient levels under which boreal plants grow.  Also, in 
these systems, the effects of secondary compounds in a variety of plant 
species have been proposed to expand beyond bite selection by individual 
browsers to influence community and ecosystem properties (Bryant 1987, 
Bryant et al. 1991a, Tahvanainen et al. 1991, Bryant et al. 1992, Molvar 
et al. 1993).  Bryant and Kuropat (1980) reviewed the plant 
characteristics that affected the winter foraging patterns of boreal 
vertebrates and suggested that because boreal forests are taxonomically 
monotonous, secondary compound diversity would also be relatively low.  
Still, they concluded that vertebrate browsing patterns were strongly 
influenced by secondary compounds, and more recent studies have 
succeeded in determining the distribution and effects of specific 
compounds on the feeding patterns of hares, moose and voles browsing on 
boreal plants.    

     This review will consider the potential for coevolutionary 
interactions between plant secondary compounds and browsing mammals in a 
relatively simple system.   I will first briefly present several 
theoretical ideas on plant-animal coevolution and then present the 
evidence on interactions between mammalian browsers and woody plants in 
boreal regions that may bear on the coevolutionary questions.    

Theoretical coevolution

Definitions of coevolution

     Coevolution between plant defenses and the herbivores that feed on 
them has been discussed at multiple scales.  Trends of variable chemical 
deterrence are noticeable within the plant kingdom as a whole, with the 
complexity of secondary compounds generally increasing from ferns 
through herbaceous angiosperms (Harborne 1988).  Within plant families, 
trends of increasingly complex and effective chemical deterrents are 
evident (Berenbaum 1983).  Additionally, several considerations of 
coevolution discuss the relationship between the evolution of new 
secondary compounds and adaptive radiation in both plant and herbivore 
(Ehrlich and Raven 1964, Rhoades and Cates 1976).

       Reciprocal adaptations between plant defense and herbivore 
browsing can occur at more restricted scales.  In a simplified case, 
genes that result in plant secondary compounds that deter herbivory will 
spread through the plant population if selective forces from herbivory 
are strong.  Herbivores will be prevented from feeding on the plants or 
will suffer ill effects when they do feed on the defended plants until 
they evolve means to handle these chemicals.  If selective forces favor 
this adaptation, the ability to feed on plants containing the deterrent 
compound will spread through the herbivore population, and the plant is 
again unprotected from these herbivores.

Criteria for documenting coevolution

     The likelihood and documentation of such evolutionary interactions 
depend on the existence of several conditions. 

     1) Plants show intraspecific variation in the amount of herbivore 
damage (Marquis 1992).

     2) Variability in damage correlates with the presence or 
concentration of a secondary compound or compounds (Marquis 1992).

     3) Damage correlates with fitness (Marquis 1992).

     4) Plants show additive genetic variance for the production of the 
compound, and thus for damage (Marquis 1992). 

     5) Herbivores have intraspecific variation in their ability to 
handle secondary compounds.

     6) The ability to handle the compound shows additive genetic 
variance.

     7) The ability to handle the compound correlates with fitness.

     Demonstration of these and the micro-evolutionary processes leading 
to reciprocal adaptations has not been done in a natural system.  Most 
studies document, or attempt to document, only a few of these 
conditions.  For most studies, correlations of damage with plant fitness 
are lacking (Endler 1986), and the importance of herbivory as a 
selective force is questioned repeatedly (Marquis 1992).  Thus to 
evaluate coevolution in natural systems requires the use of the 
available evidence that applies to the listed requirements as well as 
indirect evidence.

Browsers and food plants in the boreal forest

     The vegetation of resource-limited areas such as the boreal region 
is characterized and dominated by slow growing, woody plants that are 
relatively poor food for mammals (Bryant et al. 1991).  The diversity of 
plant species is low: species such as the deciduous willows (Salix 
spp.), birch (Betula spp.), alder (Alnus spp.), and poplar (Populus 
spp.) and the evergreen Scots pine (Pinus sylvestris) often form 
monospecific stands.  

     Three main groups of mammalian herbivores in these forests are 
rodents (voles and lemmings), hares (Lepus spp.) and cervids (moose, 
Alces alces; Tahvanainen et al. 1991).  All feed on woody plants but 
because of body size and dental differences, select different portions 
of plants.  Browsing by all three groups is particularly heavy in the 
winter months and at the peaks of hare or rodent population cycles.

Variation in secondary compounds and browsing resistance 

Variation in secondary compounds: potential for defense

     Substantial variation in secondary compounds exists among 
congeneric plant species and within distinct populations of a single 
species (Bryant et al. 1989, Rousi et al. 1991).   Experiments on birch 
and willow found that genetic differences do exist, a necessary 
condition for coevolution in these systems.  

     Families of birch seedlings (B. pendula) from Finland and Siberia 
and crosses between these families differed in the amounts of 
papyriferic acid-containing resin on their bark (Rousi et al. 1993).  
All trees were grown from seed in a greenhouse to control for phenotypic 
responses to variable environmental conditions.  Individuals of B. 
pendula, B. platyphylla, and B.pendula x B. platyphylla hybrids raised 
in a similar environment showed significant differences in the amount of 
resin produced (Rousi et al. 1991).  B. platyphylla (originating from 
eastern Asia, and often considered a subspecies of B. pendula) showed 
particularly large amounts of resin and thus secondary compounds.    
Frequency distributions of number of resin droplets within families and 
crosses were consistently unimodal, suggesting additive genetic 
variation for defense compounds (Rousi et al. 1991).  Experiments on 
other boreal plants such as willow (Bryant et al. 1989, Nichols-Orians 
et al. 1993, Danell pers comm) and Scots pine (Rousi 1989) have found 
similar variation in defense compounds between families and genotypes.     

Variation in browsing resistance and correlations with secondary 
compound production

     The potential for coevolution involving secondary compounds depends 
on a correlation between the production or concentrations of secondary 
compounds and the level of browsing damage.  For the studies discussed 
above, this relationship was strong.  Among B. pendula families and B. 
pendula x B. platyphylla hybrids, resin content  was inversely related 
to mountain hare (Lepus timidus) preference (Rousi et al. 1991).  Resin 
levels of  B. platyphylla were approximately twice as high as those in 
any other family or cross, and browsing damage was about one third the 
average for other families (Rousi et al. 1991). In a separate study, 
variation in both vole (Microtus agrestis) and hare (L. timidus) 
browsing on different genetic lines of birch (B. pendula) was correlated 
with levels of secondary compounds (Rousi et al. 1993).  

     Other studies have shown similarly strong correlations between 
levels of mammalian browsing and levels of specific secondary compounds 
in boreal plants.   Levels of papyriferic acid in birch (Reichardt et 
al. 1984), salicaldehyde, 6-hydroxycyclohexenone, and 1,2- 
cyclohexadione in balsam poplar (Populus balsamifera; Reichardt et al. 
1990, Clausen et al. 1992), pinosylvin and pinosylvin methyl ether in 
green alder (Alnus crispa; Clausen et al. 1986) and the phenolic 
glycosides salicortin, picein, salidroside and acetyl salicortins in 
willow (Salix spp.; Tahvanainen et al. 1985) have all been shown to 
regulate the preference of mammalian browsers for plants.  

     In most of these studies, the genetic variation among plants was 
either not known, not measured, or not reported.  However, any 
demonstration that secondary compounds correlate with herbivore 
preference, and therefore relative broswing damage, is relevant to 
coevolutionary questions.  The distribution of secondary compounds 
within a plant (Palo 1984) and between adult and juvenile forms of the 
same plant (Bryant et al. 1991a,b) is not uniform and browsing is 
directed towards those parts that are least defended (see reviews in 
Bryant et al. 1991 a,b, Tahvanainen et al. 1991).  In the past decade, 
most studies on variability in defense and resistance have focused on 
phenotypic variation due to environmental conditions rather than genetic 
differences (Bryant et al. 1983, 1987, Waring et al. 1985, Reichardt et 
al. 1991).  Thus, although few studies document both genetic variation 
in chemistry and resistance, there is strong evidence that browsers 
discriminate based on levels and types of secondary compounds as well as 
evidence that secondary chemistry is at least partially genetically 
determined.  

 Estimating plant fitness

Determining the correlation between chemistry, damage and plant fitness

     As mentioned earlier, the relationship between the amount of 
herbivore damage and plant fitness is poorly documented and debate over 
the relationship is inconclusive (Fritz and Simms 1992).  Of importance 
for evaluating coevolution is determining the cost of producing the 
observed levels of secondary compounds in comparison to the benefits 
gained from being defended against herbivores (Zangerl and Bazzaz 1992, 
Simms 1992).  In many cases, the ability to determine this relationship 
is hindered by uncertainty about the relevant currency to measure (Simms 
1992).  

     Although I could find no direct estimates of this relationship for 
boreal systems, several lines of indirect evidence may be used to 
evaluate this chemistry-damage-fitness relationship.  In terms of 
fitness, the value of different plant parts varies.  Similarly, the 
value of a plant part likely changes depending on the stage of plant 
development.  Studies on boreal plants have been exceptionally detailed 
in measuring the variation in secondary chemistry between plant parts 
and between different developmental stages (Palo 1984, Bryant et al. 
1991a,b).  The available evidence suggests that juvenile plants are 
generally more defended and less palatable to browsers than adult plants 
(Bryant et al. 1991a,b, Tahvanainen et al. 1991).  Intra-plant variation 
also exists within a single season: the upper branches of adult trees 
are beyond the reach of even moose and generally have significantly 
lower levels of secondary compounds. When offered to herbivores along 
with lower branches in the juvenile stage or branches from juvenile 
plants, these less defended branches are preferred (Bryant et al. 
1991b).  Additionally, boreal plants tend to have higher concentrations 
of secondary compounds in winter, the time of the most intense browsing 
(Palo 1984).  These observations all suggest that plant production of 
secondary compounds is related to the levels of browsing on those plant 
parts and to the fitness value of different plant parts. 

Biogeographic evidence

     Biogeographic studies provide additional indirect evidence on the 
fitness value of deterrence provided by secondary compounds. Because 
past selection is difficult to verify, geographic patterns of  secondary 
compound variation are difficult to interpret.  However, the available 
evidence at these large scales does correlate with the estimated 
importance of browsing as a selective force.  

     Bryant et al. (1989) reported biogeographic differences in 
secondary compounds for birch and willow.  They compared the levels of 
various secondary compounds in several species and populations of birch 
and willow and correlated chemical differences to browsing by two 
species of hares.  Both levels of secondary compounds and  resistance to 
browsing hares varied across geographic regions.  For Finnish mountain 
hares (Lepus timidus) birch grown in a common garden decreased in 
palatability from Iceland (Betula pubescens), to Finland (B. pendula, B. 
pubescens, B. nana) to Siberia (B. middendorffii).  Concentrations of 
secondary compounds (papyriferic acid and 3-0-malonylbetulafolientriol 
oxide I) also varied and were inversely correlated with hare preference.  
For Alaskan snowshoe hares (L. americanus), preference was again 
inversely correlated with concentrations of secondary compounds and 
decreased from Finnish birch (B. pendula and B. pubescens) to Alaskan 
birch (B. nana and B. resinifera).  With willow twigs, both snowshoe and 
mountain hares preferred willows from Finland (Salix caprea and S. 
phylicifolia), which showed lower levels of several phenolic glycosides 
(salicortin, picein, salidroside, and acetyl salicortins; Tahvanainen et 
al. 1985), than those from Alaska (S. alaxensis and S. arbusculoides).  

     Both the concentrations of defensive compounds and the palatability 
of different plants were related to the level of browsing history found 
in the habitat from which the plants were collected (Bryant et al. 
1989).  Iceland had no browsing mammals before Norse colonization and 
plants there still experience relatively low levels of browsing.  In 
contrast, Alaskan and Siberian plants experience much higher levels of 
herbivory, due in part to the especially intense browsing during the 
peak of the 10-year hare population cycles.  Additionally, both Alaska 
and Siberia supported a large and diverse browsing fauna throughout the 
Pleistocene (Bryant et al. 1989).  Although this study did not directly 
measure fitness, it does imply that differences in plant defense are 
correlated with levels of browsing, and thus that in areas that 
experience heavy browsing, fitness is improved by producing chemicals 
that deter browsers.

The herbivores: hares, moose, and voles

Genetic variation and additive variance

     In contrast to the plants, the herbivores' side of any potential 
coevolutionary interaction in boreal systems is relatively unexplored.  
Because many estimates of browsing use wild animals, measurements of 
intraspecific variation in the ability to deal with secondary compounds 
are lacking.  The design of these studies also precludes determining 
whether any variation between individual browsers is genetically 
determined.

     Studies of congeneric hares (Lepus spp.) have shown differences in 
willingness to ingest secondary compounds and differences in the effects 
of specific compounds.  Alaskan snowshoe hares (L. americanus) and 
Finnish mountain hares (L. timidus) showed similar preferences for 
willow and birch twigs that differed in their levels of secondary 
compounds (Bryant et al. 1989).  However, unlike snowshoe hares, 
mountain hares refused to eat the most heavily defended plants at all.  
Birch and willow from geographic regions that were most intensely 
browsed had higher levels of secondary compounds and were avoided by 
mountain hares even when no other forage was offered.  This suggests 
that snowshoe hares, which inhabit regions where plants are relatively 
heavily defended, can eat greater masses of secondary compounds than 
Finnish mountain hares.  Bryant et al. (1989) propose that this 
difference is a result of better detoxification mechanisms in snowshoe 
hares, and if so, chemical coevolution has occurred. 

     Similar differences were found between mountain hares and European 
hares (L. europaeus) feeding on diets containing extracts of birch (B. 
pendula) secondary compounds.  European hares typically feed on grasses 
instead of woody plants; mountain hares feed heavily on birch.  Birch 
secondary compounds are known to cause sodium imbalances and reduced 
digestibility of dry matter and protein in hares (Iason and Palo 1991).  
In both hare species, the birch extract decreased digestibility by 
approximately similar amounts.  However, sodium balance in mountain 
hares was unaffected by increasing levels of extract (up to 80% that 
found in live birch twigs) while that of European hares decreased 
severely with increasing levels of extract suggesting a lower tolerance 
for birch secondary compounds (Iason and Palo 1991).  Together these 
studies provide evidence that at least between species, the ability to 
deal with specific secondary compounds varies.  The differences in 
abilities to ingest or detoxify high levels of secondary compounds was 
correlated with the importance of foods containing these compounds in 
the hares' natural diet.

Correlation between diet and fitness

     The evidence for the effects of secondary compounds on fitness has 
not been measured directly for boreal mammals.  However, the evidence 
presented above suggests that they may be important.  Both the fine-
scale selectivity of browsers among similar plants and plant parts that 
vary in their levels of secondary compounds (Bryant et al. 1991a,b, 
Bryant et al. 1992) and the potentially severe effects of these 
chemicals on the herbivores' nutritional status (Iason and Palo 1991) 
suggest that the ability to avoid or detoxify secondary compounds is a 
result of selection.

Conclusion

     Most studies on boreal browsing systems involve either moose, 
hares, or voles.  Studies on the chemical aspects of other boreal 
mammal- (reindeer/caribou, white-tailed and Sitka deer, roe deer, 
lemmings) plant interactions are scarce.  Moose, hares, and vole 
preferences are generally well correlated (Bryant and Kuropat 1980, 
Tahvanainen et al. 1991, Bryant et al. 1991a), suggesting that they may 
exert similar selective pressures on plant chemistry.  However, 
differences in the digestive systems of each probably result in 
different fitness consequences for feeding on or avoiding different 
secondary compounds.  Thus, the reciprocal selective pressure of the 
plants on the herbivores may be unequal.  The effects of specific 
secondary compounds on moose are unavailable and no comparisons have 
been made between either populations or subspecies of these browsers.  
Although the direction of selection on plant chemistry may be similar 
among herbivores, the intensity of this selection probably varies.  
Differences in body size and especially population sizes between these 
herbivores would tend to influence their selective intensity since the 
relative importance of food quality versus food quantity may depend on 
both.   The importance of the 10-year hare population cycles and 
browsing by large mammals during the Pleistocene as strong selective 
pressure was emphasized by Bryant et al. (1989).

The conditions necessary for coevolution have been studied to varying 
degrees in the boreal system.  Variation in plant secondary chemistry is 
relatively well explored compared to variation in browsers abilities to 
handle or avoid them.  Perhaps the single best study supporting boreal 
coevolution is the biogeographic study of  Bryant et al. (1989) that 
found differences in secondary compound levels correlated to browsing 
among geographically distinct plant populations that differed in the 
current and historical levels of browsing they experience.   The 
available evidence suggests that the observed interactions between 
plants and mammalian herbivores could be the result of chemical 
coevolution.

                               REFERENCES

Berenbaum, M.  1983.  Coumarins and caterpillars: a case for 
coevolution.  Evolution 37:163- 179.

Bryant, J.P.  1987.  Feltleaf willow-snowshoe hare interactions: plant 
carbon/nutrient balance and floodplain succession.  Ecology 68:1319-
1327.

Bryant, J.P. and P.J. Kuropat.  1980.  Selection of winter forage by 
subarctic browsing vertebrates: The role of plant chemsitry.  Ann. Rev. 
Ecol. Syst. 11:261-285.

Bryant, J.P., F.S. Chapin, III, and D.R. Klein.  1983.  Carbon/nutrient 
balance of boreal plants in relation to vertebrate herbivory.  Oikos 
40:357-368.

Bryant, J.P., F.S. Chapin, III, P.B. Reichardt, and T.P. Clausen.  1987.  
Response of winter chemical defense in Alaska paper birch and green 
alder to manipulation of plant carbon/nutrient balance.  Oecologia 
(Berlin) 72:510-514.

Bryant, J.P., J. Tahvanainen, M. Sulkinoja, R. Julkunen-Tiitto, P.B. 
Reichardt, and T. Green.  1989.  Biogeographic evidence for the 
evolution of chemical defense by boreal birch and willow against 
mammalian browsing.  Am. Nat. 134:20-34.

Bryant, J.P., P.J. Kuropat, P.B. Reichardt, and T.P. Clausen.  1991a.  
Controls over the allocation of resources by woody plants to chemical 
antiherbivore defense.  Pages 83-102 in Palo, R.T. and C.T. Robbins, 
eds. Plant Defenses Against Mammalian Herbivory.  CRC Press. Ann Arbor, 
MI.

Bryant, J.P., K. Danell, F. Provenza, P.B. Reichardt, T.P. Clausen, and 
R.A. Werner.  1991b.  Effects of mammal browsing on the chemistry of 
deciduous woody plants.  Pages 135-154 in Tallamy, D.W. and M.J. Raupp, 
eds. Phytochemical Induction by Herbivores.  John Wiley & Sons, Inc. New 
York, NY. 

Bryant, J.P., P.B. Reichardt, and T.P. Clausen.  1992.  Chemically 
mediated interactions between woody plants and browsing mammals.  J. 
Range. Manage. 45:18-24.

Clausen, T.P., P.B. Reichardt, and J.P. Bryant.  1986.  Pinosylvin and 
pinosylvin methyl ether as feeding deterents in green alder.  J. Chem. 
Ecol. 12:2117-2131.

Clausen, T.P., P.B. Reichardt, J.P. Bryant, and A.R.E. Sinclair.  1992.  
Chemical defense of Populus balsamifera: a clarification.  J. Chem. 
Ecol. 18:1505-1510.

Danell, K.  Department of Wildlife Ecology, The Swedish University of 
Agricultural Sciences, Umea, Sweden.

Ehrlich, P.R. and P.H. Raven.  1964.  Butterflies and plants: a study in 
coevolution.  Evolution 18:586-608.

Endler, J.A.  1986. Natural selection in the wild.  Princeton University 
Press. Princeton, NJ.

Fritz, R.S. and E.L. Simms, eds.  1992.  Plant Resistance to Herbivores 
and Pathogens: Ecology, Evolution, and Genetics.  University of Chicago 
Press, Chicago, IL. 590 pages.

Harborne, J.B.  1988.  Introduction to Ecological Biochemsitry, 3rd 
edition.  Academic Press. San Diego, CA.  356 pages.

Iason, G.R. and R.T. Palo.  1991.  Effects of birch phenolics on a 
grazing and browsing mammal: a comparison of hares.  J. Chem. Ecol.  
17:1733-1742.

Marquis, R.J.  1992.  The selective impact of herbivores.  Pages 301-325 
in Fritz, R.S. and E.L. Simms, eds.  Plant Resistance to Herbivores and 
Pathogens: Ecology, Evolution, and Genetics.  University of Chicago 
Press, Chicago, IL.

Molvar, E.M., R.T. Bowyer, and V. Van Ballenberghe.  1993.  Moose 
herbivory, browse quality, and nutrient cycling in an Alaskan treeline 
community.  Ecology 94:472-479.

Nichols-Orians, C.M., R.S. Fritz, and T.P. Clausen.  1993.  The genetic 
basis for variation in the concentration of phenolic glycosides in Salix 
sericea: Clonal variation and sex-based differences.  Biochemical 
Systematics and Ecology 21:535-542.

Palo, R.T.  1984.  Distribution of birch (Betula spp.), willow (Salix 
spp.), and poplar (Populus spp.) secondary metabolites and their 
potential role as chemical defense against herbivores.  J. Chem Ecol. 
10:499-520.

Palo, R.T. and C.T. Robbins, eds.  1991. Plant Defenses Against 
Mammalian Herbivory.  CRC Press. Ann Arbor, MI. 192 pages.   Reichardt, 
P.B., J.P. Bryant, T.P. Clausen, and G.D. Wieland.  1984.  Defense of 
winter-dormant Alaska paper birch against snowshoe hares.  Oecologia 
(Berlin) 65:58-69.

Reichardt, P.B., J.P. Bryant, B.R. Mattes, T.P. Clausen, F.S. Chapin, 
III, and M. Meyer.  1990.  Winter chemical defense of Alaskan balsam 
poplar against snowshoe hares.  J. Chem. Ecol. 16:1941-1959.

Reichardt, P.B., F.S. Chapin, III, J.P. Bryant, B.R. Mattes, and T.P. 
Clausen.  1991.  Carbon/nutrient balance as a predictor of plant defense 
in Alaskan balsam poplar: Potential importance of metabolite turnover.  
Oecologia (Berlin) 88:401-406.

Rhoades, D.F. and R.G. Cates.  1976.  A general theory of plant 
antiherbivore chemistry.  Rec. Adv. Phytochemistry 10:168-213.

Rousi, M.  1989.  Susceptibility of winter-dormant Pinus sylvestris 
families to vole damage.  Scandinavian Journal of Forest Rearch 4:149-
169.

Rousi, M., J. Tahvanainen, and I. Uotila.  1991. A mechanism of 
resistance to hare browsing in winter-dormant European white birch 
(Betula pendula).  Am. Nat. 137:64-82.

Rousi, M., J. Tahvanainen, H. Henttonen, and I. Uotila.  1993.  Effects 
of shading and fertilization on resistance of winter-dormant birch 
(Betula pendula) to voles and hares. Ecology 74:30-38.

Simms, E.L.  1992. Costs of plant resistance to herbivory.  Pages 392-
425 in Fritz, R.S. and E.L. Simms, eds. Plant Resistance to Herbivores 
and Pathogens: Ecology, Evolution, and Genetics.  University of Chicago 
Press, Chicago.

Tahvanainen, J., E. Helle, R. Julkunen-Tiitto, and A. Lavola.  1985.  
Phenolic compounds of willow bark as deterrents against feeding by 
mountain hare.  Oecologia (Berlin) 61:403-410.

Tahvanainen, J., P. Niemela, and H. Henttonen.  1991.  Chemical aspects 
of herbivory in boreal forest-feeding by small rodents, hares, and 
cervids.  Pages 115-131 in Palo, R.T. and C.T. Robbins, eds.  Plant 
Defenses Against Mammalian Herbivory.  CRC Press. Ann Arbor, MI. 

Waring, R.H., A.J.S. McDonald, S. Larsson, T. Ericsson, A. Wiren, E. 
Arwidsson, A. Ericsson, and T. Lohammer.  1985.  Differences in chemical 
composition of plants grown at constant relative growth rates with 
stable mineral nutrition.  Oecologia (Berlin) 66:157-160.

Zangerl, A.R. and F.A. Bazzaz.  1992.  Theory and pattern in plant 
defense allocation.  Pages 363-391 in Fritz, R.S. and E.L. Simms, eds. 
Plant Resistance to Herbivores and Pathogens: Ecology, Evolution, and 
Genetics.  University of Chicago Press, Chicago