3.3) What plant traits are measured in vertebrate and invertebrate focussed studies?
Across all studies, 209 unique plant traits were measured. Of these 209, 115 were unique to vertebrate focussed studies, 50 unique to invertebrate focussed studies. Of these traits, 44 were measured commonly across vertebrate and invertebrate studies (Appendix C). Leaf traits were the most frequently measured (115), but stem (34), whole plant (24), root (16), flower/fruit (13) and seed traits (7) were also measured.
Key areas of difference were between the proportion of biochemical and morphological traits examined, with the former a greater focus in invertebrate, and the later, in vertebrate studies (Figure 5). This difference in traits studied between vertebrate and invertebrate studies may be due to the scale at which these two herbivore guilds operate, explained largely by differences in body size (Kotanen & Rosenthal, 2000a). Due to their greater capacity for defoliation at larger scales, vertebrates can influence plant morphology readily (e.g., plant height or biomass) and thus plants are potentially more likely to change their morphology in response to herbivory. In contrast, invertebrate herbivory usually occurs over relatively longer time periods (e.g., a caterpillar feeds on one leaf longer than a cow), at smaller scales and targeted to specific tissues (Hulme, 1996; Oduor et al., 2010). This may allow the plant more time to distribute re-mobilise nutrients and induce a biochemical defence against herbivores, for example using VOCs (Ameye et al., 2018) or other secondary metabolites. Nevertheless, there are some instances where insects can have huge impacts on plant morphological traits, such as biomass. One example are migratory insects (e.g., grasshoppers) that can reach high densities and in regard to extent of consumption, act like vertebrate herbivores (Tscharntke & Greiler, 1995). Many vertebrate studies were also conducted over several years (Appendix A) and are therefore more likely to capture changes to plant morphology than invertebrate focussed studies, which were mostly conducted within a month or less.
Physiological traits, such as photosynthetic capacity and water use efficiency were only examined in relation to vertebrate herbivory (Figure 5). This is perhaps surprising as invertebrate impact on plant physiology has been relatively well-studied for crops (Nabity et al., 2008; Peterson et al., 1998; Thomson et al., 2003; Velikova et al., 2010), and findings have shown invertebrate herbivory to be both negatively and positively associated with photosynthetic rate. Similarly, across all studies only one phenological trait, flowering onset, was examined and only in four vertebrate focussed studies. This may reflect constraints within research budgets and priorities as phenological research generally requires experiments to run over several years, for which funding is generally limited (Hughes et al., 2017; Lindenmayer et al., 2012). The lack of focus on plant phenological responses to herbivory in grasslands is concerning as phenological patterns can influence regeneration capacity (Rawal et al., 2015), community composition (Lavorel & Garnier, 2002), and adaptation across trophic levels (Bagella et al., 2013; Wray & Elle, 2015). For example, research in the alpine rocky ecosystems in Colorado found herbivory by mule deer delayed flowering phenology in a perennial herb which consequentially reduced invertebrate seed predation and overall increased plant fitness (Freeman et al., 2003). Understanding how plant phenology and herbivory interact with climate is also an important area of future research, particularly under uncertain climatic conditions (Hamann et al., 2021; Lemoine et al., 2017). With advances in remote sensing technologies, studies on plant phenological studies are now easier and cheaper to perform and we see these technologies already starting to be used to answer other phenological questions (Dronova & Taddeo, 2022).
At the individual trait level and ignoring additional responses from multiple species or sites examined within studies, the five most common traits assessed in vertebrate focussed studies were specific leaf area (SLA), plant height, leaf nitrogen, LDMC and leaf area (descending order of use; Figure 4; Table 2). For invertebrate focused studies the most common traits were VOCs, SLA, aboveground plant biomass, LDMC and leaf nitrogen (Figure 4, Table 2). Specific leaf area, LDMC and leaf nitrogen were common focal traits across both vertebrate and invertebrate studies. These traits are often referred to as ‘soft’ traits as they are relatively easy to measure and have been found to correlate with traits which are harder to measure such as relative growth rate (Hodgson et al., 1999; Perez-Harguindeguy et al., 2016). These ‘soft’ traits also represent important components of the leaf economic spectrum (Hodgson et al., 1999; Wright et al., 2004) and inform us about the plant’s individual response to abiotic and biotic factors, and in the context of herbivory can inform us about the species ability to tolerate or avoid herbivory.
The study of VOCs was a key point of difference between vertebrate and invertebrate focussed studies, with eight studies examining VOCs in relation to invertebrate herbivory and only one in relation to vertebrate herbivory (Zhang et al., 2014). Analysis of VOCs is usually done via dynamic headspace sampling (Chen et al., 2003) under laboratory-controlled conditions, which may limit its capacity for use on vertebrate herbivores. Zhang et al. 2014 however used this system to first identify and isolate the VOCs released from grass speciesArtesmia fridgida and then apply these VOCs to control plants during a selection experiment with domestic sheep. Some studies have also successfully employed the head-space sampling system in a field setting to examine VOC production from grass leaves with high and low levels of invertebrate herbivory (see Kigathi et.al. 2009). It may be possible to use a similar strategy to examine VOC production in response to vertebrate herbivory. In response to invertebrate herbivory, VOC production can act as a signal to attract vertebrate and invertebrate predators (Kessler & Baldwin, 2001; Mäntylä et al., 2008) and to communicate the potential for herbivore attack to neighbouring plants (Baldwin et al., 2002). In other studies, not reviewed here, vertebrates have been shown to use plant VOCs to find food and increase feeding efficacy (Bedoya-Pérez et al., 2014; Stutz et al., 2016). Overall, the influence of VOCs were mostly studied in relation to invertebrates and expanding this research to further examine their response to, and effect on vertebrate herbivory would be an interesting avenue for future exploration.
The huge diversity and high proportion (61%) of traits examined from a single study was in part due to traits being generally characteristic to particular plant families or functional groups, such as rhizome length (Amiaud et al., 2007), woody density of shrubs and woody forbs within grasslands (Whitworth-Hulse et al., 2016), thorns (Woodward & Coppock, 1995) or latex (Rasmann et al., 2009). Some studies researched unique aspects of plant-herbivore interactions. For example, Ribeiro et al. (2017) examined the influence of metals and micronutrients on invertebrate herbivore selection and this study accounted for 31 (15 %) of the single study traits. They found most metals examined did not affect invertebrate selection, although aluminium, iron, magnesium, manganese, and total leaf metals were found to have a negative effect on herbivore selection. Other studies examined specific morphological attributes, such as leaf symmetry or grass blade width, or specific anatomical aspects of a plant, such as spine angle. For example, Santos et al. (2013) found herbivory by gall midge was negatively correlated with leaf symmetry in the plant Bauhinia brevipes . They also found less symmetric leaves had lower leaf nitrogen, which is thought to be favourable for gall development. These results reveal interesting relationships between morphological and biochemical traits and their interaction with herbivores.
Using a common list of traits across research groups exploring plant-trait herbivore interactions would help to standardise trait measurements and improve comparability across studies and this has been flagged in previous reviews on the topic (Diaz et al., 2007) but at the same time, researchers should remain open to new emerging plant traits as technologies advance and our understandings improve. Shortlists of traits that should be favoured or disfavoured by herbivory have been identified (Coley et al., 1985; Weiher et al., 1999). Although these are mostly focussed on vertebrate herbivory, used alongside results from this review, a list of potential focal plant traits which are relevant to both vertebrate and invertebrate herbivores can be determined.