Defensive traits may evolve differently between sexes in dioecious plant species. Our current understanding of this process hinges on a partial view of the evolution of resistance traits that may result in male-biased herbivory in dioecious populations. Here, we present a critical summary of the current state of the knowledge of herbivory in dioecious species and propose alternative evolutionary scenarios that have been neglected. These scenarios consider the potential evolutionary and functional determinants of sexual dimorphism in patterns of resource allocation to reproduction, growth, and defence. We review the evidence upon which two previous reviews of sex-biased herbivory have concluded that male-biased herbivory is a rule for dioecious species, and we caution readers about a series of shortcomings of many of these studies. Lastly, we propose a minimal standard protocol that should be followed in any studies that intend to elucidate the (co)evolution of interactions between dioecious plants and their herbivores.
Sexual systems in angiosperms range from hermaphroditism (monomorphic populations of plants with bisexual flowers) to dioecy (dimorphic populations of male and female individuals) and include almost all imaginable combinations and gradations (Table
Terminology for flowers and sexual systems.
Term | Description |
---|---|
Flowers | |
Pistillate | Unisexual flower with functional pistils only (female flower; may have vestigial, sterile stamens (staminodia)) |
Staminate | Unisexual flower with functional stamens only (male flower; may have vestigial, sterile pistils (pistilodia)) |
Bisexual, perfect | Bisexual flower with both functional pistils and stamens |
Sexual system | |
Monomorphic | One kind of plant (floral morph) in the population |
Hermaphrodite | Most commonly applied to plants with bisexual flowers, but all monomorphic populations consist of hermaphrodite individuals |
Monoecious | Pistillate and staminate flowers on same plant |
Gynomonecious | Both bisexual and pistillate flowers on same plant |
Andromonoecious | Both bisexual and staminate flowers on same plant |
Trimonecious | Bisexual, pistillate, and staminate flowers on same plant |
Dimorphic | Two kinds of plants (floral morphs) in the population |
Dioecious | One morph male (with staminate flowers only); the other female (with pistillate flowers only) |
Gynodioecious | One morph female, the other hermaphrodite (with either bisexual flowers or both pistillate and staminate flowers) |
Androdioecious | One morph male, the other hermaphrodite (as above) |
Trimorphic | Three floral morphs in the population |
Trioecious | Males, females, and hermaphrodites |
Modified from Dellaporta and Calderon-Urrea 1993.
Sex-biased herbivory may be one of the selective pressures conducive to the evolution of dioecy, and it can also be a consequence of sex-specific selection on patterns of resource allocation in dioecious species. Considering only the gynodioecy pathway of the evolution of dioecy, we can think of three possible scenarios regarding the role of herbivory in each of the two steps involved in this pathway (Figure
Possible scenarios of the inception of morph- or sex-biased herbivory in the evolution of dioecy via the gynodioecy pathway. Symbols represent hermaphrodite, male, or female morphs in a population (rectangle). Arrows represent evolutionary pathways between populations with different sexual systems. The first step in the pathway is the transition from a hermaphroditic to a gynodioecious population by the successful invasion of females (left-most set of arrows). The second step in the pathway is the transition from a gynodioecious to a dioecious population following the successful invasion of males and the disappearance of hermaphrodites (right-most arrows). Letters indicate different evolutionary paths.
The second step in the gynodioecy pathway to dioecy is the successful establishment of male individuals (female-sterile mutants) in a gynodioecious population followed by the loss of the hermaphroditic morph, thus resulting in a dioecious population. Upon the evolution of two unisexual morphs, defensive traits may evolve differently in each sex and eventually become sex linked [
This paper focuses on the origin of sex-biased herbivory in dioecious species. Therefore, we will not delve into morph-biased herbivory in gynodioecious species, which would be the topic of a different essay. However, we do recognize that sex-biased herbivory—indicative of sexual dimorphism in resistance against herbivores—is likely related to morph-biased herbivory in the ancestral gynodioecious population from which it evolved (Figure
The above view for the origin of greater resistance against herbivores in females is based directly on the principle of allocation: resources freed from the male function are used for the female function, growth, and defence. In contrast to this view, the finding of male-biased herbivory in dioecious populations has been explained on the basis of sex-specific selection of resistance traits, where the main difference between sexes that drives the sex-specific selection is the cost of reproduction. In this alternative view, female individuals of dioecious species are expected to have lower herbivory levels than males because the higher cost of reproduction of females confers a selective advantage to females with traits that reduce herbivore attack [
The first review of the empirical studies on the topic of sex-biased herbivory concluded that “males are more likely than females to be preferentially used by herbivores” and suggested that male-biased herbivory was widespread among dioecious species [
In addition, Ågren et al. cautioned against publication bias, whereby studies that found differences between genders could be more likely to be published than those that did not; and taxonomic bias, an overabundance of studies from certain genera or families. In this instance, the taxonomic bias is correlated with an ecological bias for studies of temperate species, despite dioecy being more prevalent in tropical ecosystems [
More recently, sex-biased herbivory in dioecious species was tested by means of a meta-analysis of 33 studies encompassing 30 species, 19 of which were previously included in Ågren et al.’s 1999 review [
Here we propose alternative evolutionary scenarios that could result in female-biased herbivory or lack of intersexual differences in herbivory levels. We invite the reader to reconsider the evidence for male-biased herbivory in dioecious plants and recommend a standard protocol for evolutionary-ecological studies of sex-biased herbivory in dioecious species that addresses the shortcomings listed below. We contend that taking for granted the generality of male-biased herbivory in dioecious species is hampering our progress in this field.
While male-biased herbivory has been explained as a consequence of sex-specific selection of resistance determined by the cost of reproduction of each sex, few of the reviewed studies (Table
List of dioecious species of angiosperms studied for dimorphic herbivore damage, and information on assessment of reproductive allocation, growth rate and resistance.
Species | Sex with greatest | Reference | Review | |||
---|---|---|---|---|---|---|
Damage | Reproductive allocation | Growth rate | Resistance | |||
Alismatales | ||||||
Araceae | ||||||
|
M | nm | nm | nd (N, C:N, leaf total phenolics) |
[ |
3 |
F | F (total dry mass) |
[ |
||||
| ||||||
Arecales | ||||||
Arecaceae | ||||||
|
nd | nm | nm | nm |
[ |
1 |
F |
[ |
|||||
M (leaf production) |
[ |
|||||
| ||||||
Asterales | ||||||
Asteraceae | ||||||
|
nd | nm | nm | nm |
[ |
|
nm | nd (leaf production) |
nm |
[ |
3 | ||
|
nd | nm | nm |
[ |
3 | |
nd (shoot length) |
[ |
|||||
|
M, F, nd, depends on herbivore | nd (flowers/shoot) | M | F (resin) |
[ |
1 |
| ||||||
Brassiclaes | ||||||
Capparaceae | ||||||
|
nd | nm | nm | nm |
[ |
3 |
Caryophyllaceae | ||||||
|
M | M (during flowering) | F | nm |
[ |
1 |
F | nm | nd (length of infected shoots) | nm |
[ |
3 | |
Chenopodiaceae | ||||||
|
F | nm | nm | nm |
[ |
3 |
M | nm | nm; nd (height, width, fresh weight in spring), F (FW in winter) | nm |
[ |
4 | |
(F) |
[ |
|||||
|
F | nm | nm | nm |
[ |
1 |
Nyctaginaceae | ||||||
|
M | M | nd (stem production) | nm |
[ |
3 |
Polygonaceae | ||||||
|
M | ? | ? | ? | T. Elmqvist unpublished data | 1 |
|
M, F, nd | nd | nm | nm |
[ |
1 |
F (ramet production) |
[ |
|||||
| ||||||
Fagales | ||||||
Myricaceae | ||||||
|
M, nd | nm | nm | nm | L. Ericson unpublished data | 1 |
nd | nm | nm | F (1-digestibility), nd (phenolics, p-glycosides, tannins) |
[ |
||
| ||||||
Laurales | ||||||
Lauraceae | ||||||
|
nd | nm | nd, M, depending on year | nm |
[ |
1 |
M (flowers/shoot), F (N and biomass) | M (plant volume) | F (phenolics on leaves, but nd on stems) |
[ |
|||
| ||||||
Malpighiales | ||||||
Salicaceae |
||||||
|
M | nm | nm | M (phenolics), nd (p-glycosides, tannins, digestibility) |
[ |
1 |
|
M, | nm | nm | F (1-digestibility), nd (phenolics, p-glycosides, tannins) |
[ |
1 |
nd | nd |
[ |
||||
nd | nm | nm | nm |
[ |
4 | |
|
M, nd, varies by year | nm | nm | nm |
[ |
1 |
nd |
[ |
|||||
nd (phenolic glycosides) |
[ |
|||||
|
M | nm | nm | nm |
[ |
3 |
|
nd | nm | nm | nm |
[ |
4 |
|
nd | nm | nm | nm |
[ |
3 |
|
M (4 of 5 spp. of sawflies) | M (shoot length) | F (phenols, marginally significant) |
[ |
1 | |
nd (miners, gallers) | nm | nm | nm |
[ |
1 | |
|
M (at high plant density) | nm | nm | nm |
[ |
1 |
M | nm | nd (new shoots) | nm |
[ |
1 | |
M (in high productivity habitat; decreases at higher herbivore pressure) | nm | nd (biomass) | nm |
[ |
1 | |
|
M | nm | nm | F (phenolics) |
[ |
1 |
|
nd | nm | nm | nm |
[ |
4 |
|
M marginal | nm | nm | nm |
[ |
3 |
nd | nd |
[ |
||||
|
nd | nm | nd (regrowth after pruning) | nm |
[ |
3 |
nd | nm | nm | nm |
[ |
4 | |
|
nd | nm | nm | nm |
[ |
4 |
| ||||||
Pandanales | ||||||
Pandanaceae | ||||||
|
M | nm | nm | nm |
[ |
2 |
|
M | nm | nm | nm |
[ |
2, 3 |
| ||||||
Rosales | ||||||
Eleagnaceae | ||||||
|
M ? | ? | ? | ? | L. Ericson unpublished data | 1 |
M |
[ |
|||||
Rhamnaceae | ||||||
|
nd | M (anthraquinones) |
[ |
3 | ||
F | nd if age < 10 y |
[ |
||||
Rosaceae | ||||||
|
M, nd | nm | nm | nd |
[ |
1 |
F but varies with fruit set |
[ |
|||||
Urticaceae | ||||||
|
M ? | ? | ? | ? | T. Elmqvist unpublished data | 1 |
| ||||||
Sapindales | ||||||
Sapindaceae | ||||||
|
M | nm | M (growth rings) | nd (astringency, total phenols, nitrogen, toughness), F (index of defence) |
[ |
1 |
variable: F near streams; M away from streams |
[ |
|||||
nd |
[ |
|||||
F |
[ |
|||||
|
nd | nm | nm | nd (N) |
[ |
1 |
Simaroubaceae | ||||||
|
M | nm | nm | two flavonoid compounds on female flowers not present in male flowers |
[ |
3 |
F: female, M: male, nd: no statistically significant intersexual differences, nm: not measured, CT: condensed tannins, TNC: total non-structural carbohydrates, N: nitrogen content (herbivores usually attracted to greater concentrations).
1: Ågren et al. 1999 [
2: Ågren et al. 1999 [
3: Cornelissen and Stiling 2005 [
4: Not mentioned in any of 1–3 above.
More recently, McCall [
In essence, the presumed chain of evolutionary events that lead to male-biased herbivory in dioecious plants stems from the reallocation of those resources freed upon the loss of a sexual function in unisexual mutants towards defence. In most studies, defence has been equated to resistance. However, defence may also occur through tolerance [
There is one other possibility that has not been emphasized enough in the proposed models of the evolution of defence in dioecious species: while one possible consequence of a greater allocation of resources to reproduction in females is reduced allocation to growth, it is also possible that the main reduction in allocation is to defence. In this case, there would be no detectable detriment to growth. Consequently, female plants would suffer more damage (if they are less resistant; Figure
Evolutionary changes in the rate of resource acquisition in female individuals may occur through increased photosynthetic rates, canopy area, rates of mineral nutrient uptake, as well as greater branching of roots, and enhancement of mycorrhizal associations [
Alternatively, a stage in which female individuals have heavier damage levels because of resource limitation for resistance may be a transient evolutionary stage prior to the invasion of mutants whose greater defence levels are attained at the expense of growth. In this case, we should observe female-biased herbivory in younger dioecious lineages and male-biased herbivory in those lineages in which there has been enough time for selection to reshape the patterns of resource allocation to reproduction, defence, and growth. We should be able to test this by means of relative dating of lineages with male- or female-biased herbivory.
Similarly, as long as there has not been selection on prereproductive growth rates following the evolution of unisexuality, we should not see differences in growth rates or other physiological vegetative traits between males and females before their first reproductive event. It is difficult to test this prediction without reliable morphological or genetic markers that allow juveniles to be sexed so that their performance can be compared on the basis of sex. Some sex-linked markers may be, effectively, sex-related traits expressed before the onset of reproduction. Whether the presence of these markers implies the existence of sex chromosomes is still an area in need of further investigation [
In short, without fitness gain curves for each sex, it is difficult to predict accurately which sex should evolve greater resistance against herbivores and whether we should expect or not male-biased herbivory in dioecious species [
It has not escaped our attention that the evolution of defence in gynodioecious species can be approached from a similar perspective to the one presented above for dioecious species [
The collection of studies cited in the reviews of herbivory in dioecious species [
Cornelissen and Stiling’s meta-analysis of sex-biased herbivory includes 30 species, 28 of which are angiosperms. Focusing only on angiosperms, 13 of the 28 species were not considered previously in Ågren et al.’s review (Table
In addition to this taxonomic bias, a critical reexamination of that list of species casts serious doubt on the conclusion that male-biased herbivory is a rule in dioecious species: only 13 of those species were reported invariably to have male-biased herbivory. This list includes three
Perhaps the most serious problem with several studies of herbivory and dioecy has been the failure to make the connection between sex-biased herbivore damage and intersexual differences in growth rate, precisely because the latter is the purported cause of the former. Of the 30 species of angiosperms in Table
Only 12 of the 30 species listed were assessed for intersexual differences in reproductive allocation in terms of reproductive effort (the proportion of biomass or other currency devoted to reproductive structures relative to the total biomass or expenditure in the selected currency of an individual). Reproductive effort was greater in females of 10 species and in males for the other two species. In some species, reproductive effort was measured during flowering, but allocation to fruit production was not considered (e.g.,
The only species that have been assessed for foliar damage, growth rate, and reproductive allocation in the same study are
In some species, reproductive allocation, growth rate, and/or resistance were reported after the initial publication of sex-biased herbivory. However, even with these studies, the number of species for which we have a more complete picture of the causal links amongst these attributes remains low: nine more species (
In summary, the majority of studies on the topic of sex-biased herbivory have neglected the purported causal connections between bias in reproductive allocation, differential growth rate, resistance, and herbivore damage. Also, some authors seemed to confuse theoretical expectations with empirical evidence of greater female reproductive allocation: while Lloyd and Webb [
Using the search terms herbiv* and dioec* for entries between January 1998 and May 2012 on the Web of Science, we found nine studies encompassing 14 species that were not included in either of the previous reviews of the topic. Of these, only the study on the three species of
Studies of defence on dioecious species published after 2004, or published earlier but not mentioned in Ågren et al or Cornelissen and Stiling’s reviews.
Species | Sex with greatest | Herbivores | Reference | |||
---|---|---|---|---|---|---|
Damage | Reproductive allocation | Growth rate | Resistance | |||
Arecales | ||||||
Arecaceae | ||||||
|
M | F | F | F | Chrysomelid beetles | [ |
|
M | F | M | F | Chrysomelid beetles | [ |
|
M | F | M | F | Chrysomelid beetles | [ |
| ||||||
Aquifoliales | ||||||
Anacardiaceae | ||||||
|
nd; marginally F after flowering | nd | nd | nm | lepidopteran larvae and leaf spot (fungal pathogens) | [ |
| ||||||
Sapindales | ||||||
Anacardiaceae | ||||||
|
F | nd (wood/reproductive shoot) | nm | nd (wood density, branch breakability) | Elephants | [ |
|
F | nm | nm | M (N, TNC) | Cerambycid beetle | [ |
| ||||||
Malpighiales | ||||||
Salicales | ||||||
|
nd | nm | nm | M (mortality of herbivore) | Leaf galler | [ |
|
nm | F | nd | nd (phenolics, CT) | Reindeer | [ |
|
nd | nm | nd | nm | Muskox | [ |
|
nd | F | nm | nm | Insects | [ |
| ||||||
Laurales | ||||||
Lauraceae | ||||||
|
nd | nm | nm | nm | Unspecified | [ |
|
nd | nm | nm | nm | Unspecified | [ |
|
nd | nm | nm | nm | Unspecified | [ |
|
nd | nm | nm | nm | Unspecified | [ |
| ||||||
Unplaced (Euasterids I) | ||||||
Hydrophyllaceae | ||||||
|
nd | nm | nm | nm | Larvae of lepidoptera (2 spp.) and coleoptera (1 sp.) | [ |
F: female, M: male, nd: no statistically significant intersexual differences, nm: not measured, CT: condensed tannins, TNC: total non-structural carbohydrates, N: nitrogen content (herbivores usually attracted to greater concentrations).
New studies must clearly allude to the theoretical framework from which the prediction of sex-biased herbivory levels (resistance) stems—resource allocation theory, in particular, sex allocation. The claim that male-biased herbivory is expected because it has been reported as a pattern, whether implicit or explicit, lacks heuristic value because it does not address the causes of such pattern. Moreover, a plethora of factors may modify the expected pattern, as shown above.
Clearly, we need to increase the taxonomic breadth of the studies of herbivory in dioecious species. There are several ways to achieve greater taxonomic representation. We could direct our attention to those families with the greatest number of dioecious species or those with the greatest proportion of dioecious species. The first alternative will miss families with low species richness that may have a high proportion of dioecious species. The second method will miss families with high species richness but low proportion of dioecious species. One possible compromise is to focus our studies on the families with the greatest number of dioecious species among those with a large proportion of dioecious species, for instance, 50% or more (Table
Total number of species, number of dioecious species, proportion of dioecious species, and estimated 2% of dioecious species in the top 30 most species-rich families with a proportion of dioecious species greater than 0.5 (from unpublished data from S. Renner, University of Munich).
Family | Total species | Dioecious species | Proportion of dioecious species | 2% of dioecious species |
---|---|---|---|---|
Arecaceae | 815 | 778 | 0.955 | 16 |
Pandanaceae | 777 | 777 | 1.000 | 16 |
Lauraceae | 1123 | 776 | 0.691 | 16 |
Menispermaceae | 577 | 577 | 1.000 | 12 |
Ebenaceae | 487 | 487 | 1.000 | 10 |
Anacardiaceae | 594 | 439 | 0.739 | 9 |
Salicaceae | 436 | 435 | 0.998 | 9 |
Myristicaceae | 367 | 365 | 0.995 | 7 |
Clusiaceae | 590 | 365 | 0.619 | 7 |
Restionaceae | 387 | 364 | 0.941 | 7 |
Aquifoliaceae | 400 | 300 | 0.750 | 6 |
Smilacaceae | 215 | 205 | 0.953 | 4 |
Cucurbitaceae | 390 | 197 | 0.505 | 4 |
Flacourtiaceae | 209 | 192 | 0.919 | 4 |
Burseraceae | 234 | 175 | 0.748 | 4 |
Cecropiaceae | 184 | 174 | 0.946 | 3 |
Thymelaeaceae | 236 | 119 | 0.504 | 2 |
Vitaceae | 155 | 118 | 0.761 | 2 |
Loranthaceae | 147 | 114 | 0.776 | 2 |
Meliaceae | 181 | 105 | 0.580 | 2 |
Theaceae | 155 | 94 | 0.606 | 2 |
Proteaceae | 84 | 84 | 1.000 | 2 |
Hydrocharitaceae | 123 | 75 | 0.610 | 2 |
Monimiaceae | 108 | 74 | 0.685 | 1 |
Rhamnaceae | 140 | 71 | 0.507 | 1 |
Nepenthaceae | 70 | 70 | 1.000 | 1 |
Siparunaceae | 93 | 68 | 0.731 | 1 |
Myricaceae | 52 | 51 | 0.981 | 1 |
Chloranthaceae | 57 | 51 | 0.895 | 1 |
Casuarinaceae | 96 | 51 | 0.531 | 1 |
| ||||
Total | 9482 | 7751 | 155 |
In addition to the taxonomic bias, there is a preponderance of studies of woody plants. While this is understandable because most dioecious species are woody, we should strive for representation of the herbaceous component. With increased research on herbaceous dioecious species, we can address the influence of life history traits on the evolution of dioecy and defence.
Lastly, we propose that all studies aimed at assessing whether herbivory levels differ between sexes and whether these differences are a consequence of differential growth rates (in turn resulting from differential allocation to reproduction) should conduct, at least, the following measurements and observations: (1) levels of herbivory, measured as precisely as possible (preferably for more than one growing season in perennials); (2) species of herbivores responsible for most of the damage; (3) growth rates, measured either as RGR for whole individuals or from increments in branch length or leaf production; (4) reproductive allocation, measured both as the number of reproductive structures (flowers and pollen production for males, flowers, and fruits for females), and also as reproductive effort (the proportion of individual or shoot biomass allocated to reproduction, and when possible N and P allocation to reproductive structures); and finally (5) the most important resistance characters that could be influencing the levels of herbivory and measure them quantitatively. In addition, these studies could add an experimental component in which plants are damaged at least at the highest rate seen in the surveys of natural damage, so as to measure tolerance to herbivory as well as resistance [
The study of the evolution of sex-biased herbivory is hampered by the notion that male-biased herbivory in dioecious species is a rule. We have shown that the evidence used to support this conclusion has important shortcomings. We have presented other possible evolutionary outcomes with regards to sex-biased herbivory in the transition from hermaphroditic populations to dioecious ones. We have also discussed how these different outcomes can be predicted under different theoretical assumptions. Therefore, future studies of herbivory in dioecious species should be based on a clear theoretical framework. In particular, we urge that all new studies of herbivory in dioecious species include assessments of reproductive allocation, growth rates, and resistance traits deemed to differ between sexes and, therefore, determine sex-biased herbivory. In addition, tolerance should also be considered as a potentially important defence mode that can vary between sexes. In this manner, we should be able to explain better the results of any given study. The advancement of our knowledge about sex-related defence in plants should help us gain a better understanding of the evolution of sex-related traits in general.
The authors are grateful to César Domínguez, Mauricio Quesada, Nicholas Buckley, and Susana Magallón for their insight on the topic and the fruitful discussions. Thanks are due to Susan Renner for allowing them to use her data on the taxonomic distribution of dioecy across the angiosperms, to Caroline Tucker for sharing her expertise on Proteaceae, and to two anonymous reviewers for helpful suggestions to improve this paper. They are grateful to Alberto Civetta for inviting them to submit this paper.