This article is a commentary on Liu et al. (2021).
In their recent review in this journal, Liu et al. (2021) argue that sexual differences in the allocation to reproduction by male and female plants may cause sexual differences in their responses to biotic (e.g. herbivory) and abiotic stress (e.g. moisture stress). Their general argument is that the nett reproductive costs are higher for females because they not only flower but must also produce fruits/cones/seeds (their Figure 3). They suggest (their Figure 2) that females can ameliorate their higher costs of reproduction such as by maximizing resource acquisition and resource gain. Many other previous reviews have also argued for higher female costs (Barrett and Hough 2013; Barrett et al. 2010; Juvany and Munné-Bosch 2015; Obeso 2002). My aims in this commentary are (i) to note that theory predicts equal sexual allocation to reproduction, (ii) to argue that some of examples cited by Liu et al. (2021) do not suggest the theory is incorrect and finally, (iii) to provide a Cape (South Africa) perspective to show that differences in allocation cannot be invoked ad hoc to explain vegetative or other differences between the sexes. Vegetative differences between the sexes must ultimately be related to differences in sexual function, as is the case in animals, and not to differences in nett allocation to reproduction.
There are many reasons why the sexes should allocate equally. Firstly, at the most basic level evolution is about fitness, thus the life of all males and females is ultimately, equally and totally allocated to reproduction. Secondly, in biparental organisms (with limited in-breeding) because the genetic benefits of sex are an almost equal contribution to the genome of the off-spring, theory predicts that the sex ratio should be equal (Hamilton 1967). Also, because the genetic benefits of sexual reproduction are equal, the ‘costs’ or allocation must be equal (or else one sex is cheating). Say males allocated less to reproduction, they would have more resources to allocate to vegetative growth and thus grow larger or live longer, none of which would benefit them unless it ultimately contributed to their fitness via reproduction, in which case this increased vegetative growth is a reproductive allocation. Liu et al. (2021) acknowledge the expectation of sexual equality in allocation but then proceed to use nett allocation differences to explain the interesting differences they noted between the sexes.
In animals there is no clear difference between vegetative and sexual traits. For example, the brain is also a sexual organ needed for mating success and therefore it is meaningless to compare male and female allocation to reproduction, because it must include the lifetime of the entire individual. Male allocation is clearly not limited to the sex organs. It is an illusion that clear differences between vegetative and sexual allocation exist in plants. For example, the architecture of plants must simultaneously serve sexual purposes for pollen dispersal, pollen capture, seed dispersal as well as for vegetative purposes such as capturing light. There are correlations (Corners Rules) between the sizes of leaves, branches and inflorescences and plant architecture (such as degree of branching) which indicate that selection on any one of these traits will affect the others (e.g. Midgley et al. 2019). This makes it difficult to determine whether allocation is to reproduction or not. The size of a branch and its associated hydraulics (such as xylem size) is determined by both its associated flowers and leaves. Also, unlike the case in many animals, the sexes in many plant species cannot avoid competing and thus they need to be competitively equal to prevent distorting sex ratios, especially to male dominated populations. In animals, stark differences between the sexes (e.g. size, colour and ornaments) are explained by factors such as sexual selection, rather than unequal allocation to reproduction. The same explanations should apply to plants.
The argument that the costs of reproduction may differ between the sexes in plants continues because it is difficult to measure, even on subsequent growth, and to quantify across their entire lifetime. For instance, for direct costs there is the issue of which currency to use (which limiting nutrients) and the fact that the sexes allocate differently in time with males typically making an earlier commitment (Obeso 2002). For indirect evidence of the costs of reproductive allocation, such as the impact on size and sex ratios, there too are problems such as the role of natal sex ratios, clonality, age and sex determination effects (e.g. certation).
Liu et al. (2021) cite two references in support of the hypothesis of higher direct costs of reproduction of females and both are problematic. Firstly, Leigh and Nicotra (2003) argued that female allocation was about 10 times that of males in their study taxon. However, they found no differences in size or sex ratios or long-term water use patterns in their study taxa. It is highly unlikely that such large direct allocation differences will have no indirect impacts such as on growth, mortality or have no eco-physiological consequences. Leigh and Nicotra (2003) used biomass as their currency of allocation (not nutrient contents as advised by Obeso 2002) and they measured biomass when fruits were ripe, i.e. when pollen (which has low biomass but high nutrients) had already been shed; these factors may inflate the female contribution. In the second cited paper, Bañuelos et al. (2004) investigated differences in herbivory between the sexes. However, although they found secondary chemistry differences, they found no sexual differences in degree of herbivory and thus this paper does not support the higher costs hypothesis. Liu et al. (2021) state that ‘as discussed in the study of Juvany and Munné-Bosch (2015), higher reproductive investment in females may cause a decrease in the energy invested into growth and defense, thus reducing stress tolerance’. However, two of the papers cited by Juvany and Munné-Bosch (2015) to indicate higher costs of female reproduction, Leigh and Nicotra (2003), as well as Pickering and Arthur (2003) also used the largely discredited biomass allocation methodology rather than nutrient contents. Like Leigh and Nicotra (2003), Pickering and Arthur (2003) also found that the total biomass of male and female plants was similar, suggesting no nett differences in the cost of reproduction. The point here is that the evidence that the direct costs of reproduction differ between the sexes is weak and paradoxically some of the cited papers suggest that there are no indirect costs.
Liu et al. (2021) cite Yule and Burns (2019) who argued that male Aristotelia serrata plants have higher levels of sap herbivory than females and this is because males have more resources than females. If males have extra resources, why then are males not better defended? If females had been more heavily infested by larvae would the ad hoc explanation be that they had fewer resources to allocate to defence? A recent large meta-analysis found no consistent differences in herbivory, chemical defences and the sexes in dioecious plants (Sargent and McKeough 2022). These authors found that as the sample of studies increased the sexual differences dwindled. This is in line with the initial studies focussing on outliers. The same might apply to sex ratios, with initially only unequal sex ratios being seen as interesting. In the Cape (South Africa) dioecious plants are extremely abundant and can dominate entire mountainsides. Despite this abundance there have only been three published measurements of adult sex ratios and one of adult size ratios (Midgley and Cramer 2022) and here ratios were equal. I suggest that the lack of studies is because populations appear to be 1:1 in size and number and are thus considered relatively uninteresting.
Based on Olano et al. (2017), Liu et al. (2021) argue that differences between males and females in water and carbon requirements lead to females being less water stress tolerant than males. Olano et al. (2017) showed anatomical differences between males and females in Juniperus, which they interpreted as being due to higher reproductive allocation by females. However, they provided no field data to verify any nett differences in moisture stress. This is relevant because as noted below, males and females may have different phenology. For example, in the dioecious Cape Restionaceae, water use efficiency (WUE, as measured using δ13C) can differ between the sexes with females being less WUE (Araya et al. 2010). This was interpreted ad hoc as the physiological consequence of higher female allocation to reproduction (Araya et al. 2010). van Blerk et al. (2022) using actual water use measurements as well as fine scale δ13C measurements between the sexes of a Cape Restionaceae species, showed that these differences were not due to inherent differences in WUE, but were due to the different phenology of male and female culm growth (males grow earlier and allocate more fully to reproduction earlier, when soils are drier). They concluded that earlier growth is more costly to males, but because culms of both sexes reached the same size, there was no nett difference between the sexes. The same phenological difference applies to extremely vegetatively different sexes in Leucadendron rubrum, where female leaves can be up to 10 times larger than males. Harris and Pannell (2010) suggested that females had greater hydraulic efficiency than males and that this was due to greater reproductive costs in cone-bearing females. However, WUE and nett hydraulic differences do not differ between the sexes (Midgley 2010; Roddy et al. 2019). In part this may be because males commit to reproduction and growth early when conditions are dry and hot (November–March) whereas females grow and produce seeds in the cooler, moister period (January–May) (de Kock et al. 1994). Thus, males and females may be different (e.g. vegetatively, phenologically) but that there is no nett difference in resource use and there is no evidence that one sex allocates more to reproduction than the other.
A better direct test for the differences in allocation is whether females in dioecious species produce double the recruits as do co-occurring hermaphrodites. Say males allocate y units to reproduction and females x units. For a dioecious female to have the same fitness with its allocation to reproduction (x + y), compared with two hermaphrodites (2x), would imply that male and female allocation is the same (x + y = 2x). Bruijning et al. (2017) showed no differences in reproductive output and population growth rate (fitness) of dioecious versus hermaphroditic taxa in the large, detailed and long-term Barro Colorado Forest data set. They assumed (and largely found) 1:1 sex ratio and that dioecious females produced double the fruit/seed load of hermaphrodites. This indicates no direct (e.g. fecundity) or indirect (e.g. mortality) costs of female reproduction. It is to be expected that dioecious and hermaphrodite forest species that coexist, must have the same overall fitness and therefore that dioecious females have the same fitness as two hermaphrodites. In the fire-prone Cape fynbos vegetation, most Proteaceae species die in fires and recruit immediately after fires as seedlings and this includes hermaphrodite (e.g. Protea) and dioecious taxa (e.g. Leucadendron, Aulax). They must have the same fitness or else the situation would be unstable (Midgley et al. 2019). In conclusion, we should be sceptical of studies which show, or invoke ad hoc, differences in nett allocation to reproduction between the sexes in plants. It is not expected, impossible to measure and meaningless.
Acknowledgements
I thank Steve Johnson (UKZN), Mike Cramer (UCT) and Joe White (RBGK) for comments.
Conflict of interest statement. The authors declare that they have no conflict of interest.
