Masting, also called mast seeding or masting behavior, is a widespread reproductive strategy exhibited by some long-lived plant species (Kelly and Sork, 2002; Hacket-Pain and Bogdziewicz, 2021; Koenig, 2021). Masting is defined as the massive, intermittent and synchronous production of seed crops at irregular supra-annual intervals (Herrera et al., 1998; Kelly and Sork, 2002; Koenig and Knops, 2005; Crone et al., 2011), which has been extensively documented in a number of perennial plant species around the world (Kelly, 1994; Kelly and Sork, 2002; Crone and Rapp, 2014; Pearse et al., 2016; LaMontagne et al., 2020). This massive and synchronous variation in seed crops, which is mainly triggered by genetic regulation and/or environmental cues (Burns, 2012; Fleurot et al., 2023), can alter plant-animal interactions and increase the probability of seed survival as well as seedling establishment (Wang and Yi, 2022; Xiao et al., 2022). Ultimately, mast seeding contributes a great importance to population dynamics of perennial plant species (Espelta et al., 2008; Pearse et al., 2021; Zhang et al., 2022).

Due to the crucial role of masting in the regulation of population dynamics of both plant and seed predators (Bogdziewicz et al., 2016; Zeng et al., 2021; Liu et al., 2022; Yu et al., 2023), understanding the ultimate causes of such a widespread reproductive strategy is of great importance to explain the evolution and advantages of mast seeding of plants (Visser et al., 2011; Huang et al., 2021; Seget et al., 2022). It has been suggested that masting can evolve only when costs are minimized followed by the maximized returns (Seget et al., 2022). i.e., greater reproductive efficiency at high reproductive effort (Norton and Kelly, 1988). Three hypotheses based on economies of scale (i.e., increased pollination efficiency and decreased seed predation in mast years) have been proposed on the evolutionary advantages of masting (Kelly et al., 2001; Kelly and Sork, 2002; Linhart et al., 2014; Bogdziewicz et al., 2016; Seget et al., 2022). I: the pollination efficiency hypothesis posits that mast seeding can be selected by increased pollination rates in high flowering years (Kelly et al., 2001; Bogdziewicz et al., 2020). II: the predator satiation hypothesis suggests that satiating predators benefits plants by escaping seed predation through starvation of seed predators between mast years and satiation of predators during mast events (Bogdziewicz et al., 2016, 2018; Greenberg and Zarnoch, 2018; Zhang et al., 2022; Zwolak et al., 2022), and ultimately increases seed dispersal by animals and the chance of seed survival as well as seedling recruitment (Kelly and Sork, 2002; Koenig et al., 2003; Fletcher et al., 2010; Xiao et al., 2013; Zhang et al., 2022). III: the animal dispersal hypothesis, despite less widely accepted than the predator satiation hypothesis, postulates that mast seeding may enhance seed dispersal, increase dispersal distances and finally improve the chances that seeds escape post-dispersal predation (Vander Wall, 2001, 2002; Jansen et al., 2004; Burns, 2012; Hanya and Bernard, 2013; Mendoza et al., 2015), consistent with the predator satiation hypothesis.

Given the importance of economies of scale (EOS) in the evolution of mast seeding, any selective pressures that provide EOS may have the potential to select for mast seeding. Previous work has shown that wind-pollinated species provide economies of scale as opposed to animal-pollinated ones (Kelly and Sork, 2002), indicating that mast seeding intensity has increased selection pressure mainly from pollination vectors. Plant height, one of the most important life-history traits, has been documented to be selected by pollination vectors (Bickel and Freeman, 1993; Donnelly et al., 1998; Zu and Schiestl, 2017; Hernández-Villa et al., 2020). Evidence has been provided for the important role of taller strategy in promoting effective pollen dispersal and seed dispersal either for wind- or animal-pollinated plants (Culley et al., 2002; Friedman and Barrett, 2009; Thomson et al., 2011). Despite of various benefits of taller strategy, the cost of the increased difficulties in transporting water and nutrients between roots and crowns make taller plants more susceptible to various abiotic stresses such as irregular droughts than smaller ones (Kenzo et al., 2006; Cao et al., 2012). In addition, decreased stomatal conductance and assimilation rate of leaves with increasing tree height not only prevent carbon uptake but deplete carbon reserves, leading to drought-induced C-starvation in taller trees (Grote et al., 2016). Moreover, taller plants are usually long-lived and need to put more effort into growth rather than reproduction, which is likely to come at a price of seed production failure. Therefore, comparing with shorter plant species, such a trade-off between benefits and costs is likely to result in selection of mast seeding in taller plants, favoring those with drastic variations in annual seed reproduction. A renowned example has been the giant trees from Dipterocarpoidae exhibiting extremely longer intervals of highly synchronized mast seeding (Wang et al., 2024). Therefore, taller plants are expected to act as a hidden player to make large reproductive efforts more efficient, providing EOS and selection pressure for mast seeding.

Although an increasing number of studies has made great contributions to reveal the underlying mechanisms of masting behavior (Kelly and Sork, 2002; Crone et al., 2009; Turnbull et al., 2011; Kelly et al., 2013; Miyazaki et al., 2014; Pearse et al., 2014, 2016; Fleurot et al., 2023), the potential role of adult plant height in selecting for mast seeding has been largely overlooked. Here, we first collected data of adult plant height and mast seeding intensity of 158 species of spermatophytes with different life forms and pollination vectors. Because comparative studies have identified a moderate phylogenetic signal of masting, i.e., similarity in masting intensity among congeneric plants (Koenig et al., 2016; Fernández-Martínez et al., 2019; Pearse et al., 2020; Dale et al., 2021), we then fitted multivariate phylogenetic generalized linear mixed model (PGLMM) to test whether there is a close relationship between adult plant height and mast seeding intensity at population level rather than individual level, which has been a conventional evaluation method of masting. We specifically predict that mast seeding intensity will co-evolve with adult plant height and that taller plant species will exhibit higher mast seeding intensity. In addition, anemophily adopting taller strategy will show higher mast seeding intensity than entomophily.

In total, 158 seed plant species were analyzed in our study, belonging to 95 genera and 48 families. Among the plant species, 80 of them are pollinated by wind, while 78 are entomophilous species. Additionally, 123 species belong to angiosperm and 35 are gymnosperm.

Our results provide evidence that mast seeding intensity is closely related to adult plant height than ever thought, supporting our first prediction stating that masting evolves predominantly in taller species. Our analyses also support our second hypothesis stating that mast seeding is stronger in anemophily than in entomophily that exhibits smaller plant height. Although mast seeding is related to phylogeny, life form, and type of Spermatophytes, they contribute little to variation of seed production of perennial plants.

We found that the phylogenetic signal of the mast seeding intensity is distributed unimodally like a Brownian model, implying that the common ancestry can potentially account for the high frequency of mast seeding. The phylogenetic signal observed in mast seeding intensity may suggest moderate phylogenetic conservatism for masting behavior, as observed by previous studies (Tanentzap and Monks, 2018; Pearse et al., 2020). The strong phylogenetic signal observed in adult plant height and the evolutionary correlation between mast seeding intensity and adult plant height convey the main findings in our study that mast seeding is more likely to evolve in tall plants, like gymnosperms, trees, and wind-pollinated species in our study.

Regardless of plant phylogeny, we found a strong positive association between mast seeding intensity and adult plant height, indicating that taller plants tend to have highly intermittent synchronous production of large seed crops. Economies of scale have been demonstrated to provide the ultimate factor driving the evolution of masting (Norton and Kelly, 1988; Koenig, 2021). How plant height selects for evolution of masting is of great importance to our deep understanding of masting from ecological and evolutionary perspectives. As EOS is the main selective force for masting (Kelly and Sork, 2002), the close correlation between adult plant height and mast seeding intensity implies that increased plant height may also provide EOS through either enhancing pollination efficiency and/or increasing satiation of seed predators. Strong correlations between plant height and flower number (Gomez, 2003; Trunschke et al., 2017), as well as increased attraction of pollinators, pollination success, and modularity of frugivorous network with plant height have been well observed (Klinkhamer et al., 1989; Weber and Kolb, 2013; Dupont et al., 2014; Paixão et al., 2023), which are important for enhancing pollination efficiency. Evidence has shown that increased plant height significantly promotes pollen dispersal and then strongly increases pollination success (Friedman and Barrett, 2009, 2011; Harder and Johnson, 2009; Sletvold et al., 2010; Aljiboury and Friedman, 2022), which also provides EOS. Additionally, plant height is positively correlated with seed dispersal distances (Burd and Allen, 1988; Thomson et al., 2011), which benefits from density-dependent seed survival via increased seed dispersal distances (Jansen et al., 2014), providing an alternative EOS. Moreover, plant height not only reduces herbivory but also has significant suppressive effects on pre-dispersal seed predation (Cariveau et al., 2004; Zhang et al., 2023), reflecting contributions of economies of scale in satiating predators. In addition, taller strategy of plant species incurs a number of costs, including disadvantaged water transport in vascular system, increased breakage risk, reduced investment in leaf biomass, and decreased leaf nitrogen content (Falster and Westoby, 2003). Consequently, the cost incurred by taller strategy of plant species is expected to serve as a potential selective pressure for mast seeding. Plant height is also strongly correlated with lifespan (Moles et al., 2009); therefore taller strategy of long-lived plants will benefit from mast seeding rather than regular seed production because they need to continuously allocate resources for the rest of their lives. Therefore, predominant appearance of masting behavior in taller plant species in our study would explain the higher masting intensity in gymnosperms, trees, and wind-pollinated species that widely adopt a taller strategy.

Our study clearly showed that mast seeding intensity of the anemophilous species is much higher than that of entomophilous ones, consistent with the previous studies predicting that wind-pollinated species had higher interannual variability of seed crops than animal-pollinated ones (Herrera et al., 1998; Kelly and Sork, 2002). These observations are possibly due to the fact that animal-pollinated species are usually at high risk to satiate the pollinators (Sork et al., 1993; Kelly et al., 2001; Xiao, 2022). The strong selective effect of the pollination vector (e.g., anemophily and entomophily) on mast seeding intensity indicates that, in the plant taxa, both pollination vector and plant height play important roles in the evolution of the mast seeding of plants. In our study, phylogeny, life form and type of Spermatophytes had much less power to explain mast seeding intensity than adult plant height and pollination vector, though they were also closely related to mast seeding intensity. This pattern can be partially explained by the overshadowing effects of adult plant height and pollination vectors. Our results also provide evidence that wind pollination exerts selection pressure on plant height of both gymnosperms and angiosperms, which can increase pollination efficiency and may also promote seed dispersal and survival rates of plants by overcoming pollen limitation or seed predation. Therefore, it seems that plant height provides direct selection pressure on mast seeding intensity of anemophilous and entomophilous plants, although the overall contributions of pollination vectors to the interannual variation in seed production are higher than plant height.

We admit that temperature, precipitation, soil fertility, seed predators, and plant genetic regulation may influence mast seeding. Moreover, plant height is a highly plastic trait within species as well as among species. However, our study using mean values, which may avoid the influence of these variants at individual level, clearly demonstrates that adult plant height is positively correlated with mast seeding intensity of 158 plant species in 95 genera and 48 families. For the first time, we demonstrate that adult plant height may act as a potential selective pressure for mast seeding, especially in wind-pollinated species. Pollination vectors appear to be critical in selecting plant height across Spermatophytes and have an indirect effect on mast seeding intensity of perennial plants. Given the selection force of pollination vectors on the evolution of plant height, the delicate correlations between plant height and mast seeding intensity provide a novel avenue for better understanding of the evolution of masting across plant species, which has important implications for the population dynamics, community structure as well as spatial distribution of plants and the associated fauna. This study will broaden our understanding of the evolution of masting by illustrating its relationship between functional traits of perennial plants, which highlights an ambitious direction for further research into other related traits (e.g., leaf area, photosynthetic capacity and transport of water and minerals in xylem) underlying variations in seed crops across plant species.

The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found below: https://doi.org/10.6084/m9.figshare.22353517.

HQ: Data curation, Formal analysis, Funding acquisition, Writing – original draft. XY: Data curation, Formal analysis, Writing – original draft, Writing – review & editing.