Benefit versus cost trade‐offs of masting across seed‐to‐seedling transition for a dominant subtropical forest species

The seed-to-seedling transition is one of the most critical phases in life cycle plants (Muscarella et al., 2013; Wang & Smith, 2002) and a major bottleneck plant demography, thereby influencing population dynamics species coexistence (Poorter, 2007). Importantly, unique mobile stage flowering for determining their spatial distribution (Swamy 2011). Elucidating ecological mechanisms that regulate important gaining better understanding both regeneration origin patterns adult trees (Fricke 2014; Muscarella Piiroinen 2017). Nonetheless, our factors influence plants, such as effect masting, still limited. Characterized by intermittent production seeds, masting common reproductive strategy many perennial (Bogdziewicz, Ascoli, 2020; Bogdziewicz, Szymkowiak, Kelly Sork, 2002; Minor Kobe, 2017; Pearse 2016; Pesendorfer Zhang 2020). Hypotheses attempt to explain selective advantages evolutionary origins are mainly related economies scale, enhanced pollination efficiency satiation seed predators during mast years (Kelly 2001; Rapp 2013). Indeed, functional benefit predator hypothesis (PSH) has been extensively supported Espelta, 2018; Linhart Xiao Zwolak, Rychlik, Wróbel 2016). However, these empirical studies have evaluated benefits at either pre-dispersal or post-dispersal stages, with few considerations potential costs To date, we lack clarity about all forest keystone (Pearse Zwolak multiphase process involves stages pre-dispersal, post-dispersal, soil bank, germination seedling establishment (Wang 2002). Multiple negative density- distance-dependent effects (e.g. from vertebrates, insects, rodents fungal pathogens) primary agents fate mortality full Lichti 2014, Tomita Yang 2015). Given diversity predators, can shape different along transition, due balancing (Figure 1). Generally, stage, insects weevils) vertebrates squirrels) main (Linhart Soler Some previous found reduces predation 2017); however, al. (2018) recently showed effectiveness on be low. In addition, knowledge far Therefore, remains unclear. At consuming an extremely high proportion total (Xiao evidence suggests seeds escaping rodent This considered when brings recruitment through scale. bank specific fungi cause lose vitality (Gallery 2007, 2010; Leishman, Masters, 2000; Masting may higher density-dependent highly aggregative bank. mass largely determines success emergence (Lebrija-Trejos Wright, 2000). Seed also trait influences (Baraloto 2005; Krishnan 2019; Lebrija-Trejos Moles Westoby, 2004). Currently, no consensus exists whether not trade-offs exist between (Bonal 2007; Koenig 2009; Ives, If, indeed, impact mass, it could counteract fitness negatively 1a). above suggest stage. seedlings survival growth, thus having profound consequence colonization populations Previous demonstrated events dominated some super-producers responsible disproportionately large As such, greater aggregation seeds. Likewise, directly affect behaviour towards dispersal, reducing dispersal (Fletcher Li Zhang, Huang, Consequently, potentially result newly germinated around adults, reduce combined reduced nutrient accumulation (Connell, 1971; Fricke Janzen, 1970; Seiwa Swamy this sense, rearrangement 1c). While create other transition. hypothesize additional caused stages. Specifically, predict (a) creates cost high-intensity predation; (b) achieving satiation; (c) decreased (d) mortality. study, tested overarching hypotheses tracing changes Castanopsis fargesii, dominant widely distributed subtropical evergreen broad-leaved forests China. We studied under natural conditions over 6-year period understand transitions. study was conducted Jinyunshan National Nature Reserve (GNNR; 29°8′18"–29°17′29″N, 118°2′14"–118°11′12″E; 8,107 ha area), Chongqing, region monsoon climate, annual mean temperature 13.6℃. Mean temperatures January 3.1℃, August 24.3℃. precipitation relative humidity 1611.8 mm 87%, respectively. Vegetation Jinyun Mountain comprised typical forest, Machilus nanmu, Polyspora speciose carlesii var. spinulosa. fargesii Yangtze River Adult C. attain heights 30 m. 0.5 g, ranging 0.1 1.8 g S1c). Seeds predominantly gravity-dispersed, but they dispersed lesser degree squirrels mice. recognized 2005). fruits ripen October November, peak rain occurs November (Yang generally begins following April May. usually starts July ends late-October S1d). September 2013, established 50 m × 100 permanent monitoring plot divided into subplots (10 10 m). 86 traps within sample late-September 2013 all, were installed centre each subplot, remaining 36 placed grid intersections S1a). closest distance two 7.1 m, which approximately equal average diameter crown trees. Next, quadrats (1 m) surface left right sides trap, respectively S1b). boundaries marked using four pvc tubes (20 cm height) nylon threads. assumed similar what actually recorded traps. measured December annually 2018. All materials collected once per week period. defined mature categorized three groups based careful examination seed: intact (seeds intact, without any injury); vertebrate-attacked (eaten (mostly Callosciurus erythraeus) before had fallen) insect-attacked (suffered invertebrate Curculio sp., obvious holes surface). precision electronic balance. (in year rain) 2014 2019. surviving viable (already elongated radicle healthy embryo cotyledon); fungi-infected (killed infection, where and/or cotyledons covered hyphae decayed); desiccated desiccation, roots withered brown) rodent-attacked (predation rodents; mostly Rattus flavipectus norvegicus). sampled every 2 weeks May Newly grown least true leaf. Seedlings quadrat censused numbered label. late-December (2015–2020), growing seasons after emergence, labelled alive. referred individuals S1e). Based information, pathway S2). selected seven microhabitat variables expected fate, including canopy openness, shrub coverage, herb litter thickness, light intensity, moisture temperature. investigated same time year. Shrub coverage estimated percentage shrubs herbs covering plot. Canopy openness fraction open sky hemisphere visible point beneath canopy. Litter thickness quadrats, ruler. Light intensity 1.3 floor Scout 6 sensor quantum bar (LightScout; Spectrum Technologies, Inc., USA). Soil HH2 meter (Delta-T Devices) layer (0–11 depth). occurring density exceeded standard deviation (LaMontagne Boutin, 2009). calculated killed according categories described (i.e. seed, rodent-attack seed) estimate seed-killing agent. rate (either seedlings) type agent, trap super top abundance, Using generalized linear mixed models (GLMMs), Gaussian family, identity linked model. trap-level included response variable, log-transformed treated fixed effect, ID random effects. significant binomial log-linked agent probability logit-linked Seedling (0/1) effect. aggregation, quadrats-level density, effects, measure partial variable odds survival, ratios coefficient (exponential coefficient). An ratio >1 indicated positive whereas <1 cumulative imputed non-mast ArcGIS 10.2 (Esri, 2010). used simple inverse weighting interpolation algorithm across entire actual numbers search radius restricted nearest (7.1 surrounding interpolated value controlled power exponent distance. more strongly weights neighbour then Moran's I evaluate (Legendre Fortin, 1989). significance level corrected multiple tests (Holm, 1979). constructed piecewise structural equation modelling (piecewise SEM) approach investigate hypothesized causal relationships among rates, failure), Piecewise SEM incorporates several nonlinear single analysis directional separation (d-separation) statistically identify missing links pathways) directed acyclic graph. initial path model built conceptual diagram S3). connect variables, being refined original dropping non-significant links, starting link, continuing stepwise until AICc <2. statistical analyses performed R version 4.0.3 (R Development Core Team, Generalized lme4 package (Bates 2014). calculation implemented ‘vegan’ (Oksanen Structural testing ‘piecewise SEM’ (Lefcheck, A 30,466 Annually, densities ranged 25.8 315.3 seeds/m2 years, occurred 2018 2a). Annual 0.43 0.65 S4) (χ2 = 23.63, p < 0.001; Figure 2b; Table S1). attacked insect increased 6.31 0.012; 3a; Insect attack removal 16.33% highest year, 1.92 times than lowest 4g). 4.32, 0.037; 3b; Vertebrate loss 8.04% 13.88% 4f). 9.25, 0.0023; 3c; Rodent 46.15% 61.12% 4h). desiccation did change significantly (Table failure 73.87, 3d; Germination 21.75% 7.4 4 k). abundance 102 513 seedlings, rates 3.83% 8.79% 4e). variation accumulated (CVtrap) 1.33 1.10, contributed 42.78% versus 36.95% years. (CVseedling quadrat) (1.11 vs. 0.70). 36.22% 25.66% Correlograms correlation class 7.2 (minimum traps) indicating aggregated only scale 5; S5). shown S6. SEMs 69% 6). transitions, likely rodent, strong indirect decreasing mass. rodents, faliure. rodents. Herb weak 13.12, 7a). positively affected 7b). 81 212 seed-to-established 1.59% 6.98% 7c). Consistent hypothesis, multilevel 8). resulted supporting hypothesis. Nevertheless, predated dramatically (15.3% 75.1% year) its roughly half resulting pattern seedlings. More importantly, smaller spatially recruited suffering (59.9% 43.3% Looking whole process, outweighed costs. Our points importance considering phase. Inferences new alone would certainly lead overly optimistic conclusion regard masting. predicts producing massive number provide enough food allows escape decrease observed consistent past Moreover, twice observations made These results recruitment, extent, relation Despite findings, First, (Figures 3 4), certain oak Marino, Second, trade-off inter-annual S4). overall disadvantage small discussed (reviewed correlated (Gao reduction 3d 4k). seen (4.5% 8.9%, During disadvantages germination. demography (Nathan Muller-Landau, 5b). became segregated Fagus crenata (Tomita associated non-random dispersal. density-dependent, microsites suffer relatively intensive increase 14 initially low indicates patches transported recent (Hirsch 2012; Huang contrast, 5a). existence led (Minor provides abundant resources affects rodents’ scatter-hoarding 2005, result, notably despite causing lost demonstrates key offspring fitness, distinct sensitivity competition, shading, drought limitation (Leishman, For end October, droughts cold winter away. understorey environmental surveyed characterized low-light intensity. early maximizing development deep root systems, able defend against unfavorable environment. establishing 7b), 2019). conspecific trees, Thus, (Murphy Terborgh 2015; Zhu fact, 2-year 7a), difference further outweigh recruitment. conclusion, effective situation germination, arrangement establishment. regarding limited reproduction, Further observation targeted warranted assess how indirectly long-term thank Xiangyang Gao, Meng Qi, Xia Zhou, Chao Yang, Xiaohan Li, Shenghe Jing Xiao, Lijuan Tian, Ting Mingming Zheng, Siwei Hu Jingmei help field laboratory. like Dr. Baraloto Stephen Yale University Glen Wheeler assistance English language grammatical editing manuscript. Chongqing Technology Innovation Application Demonstration Major Theme Special Project (cstc2018jszx-zdyfxmX0007) Y.Y. L.H. wrote manuscript input authors; involved planning supervising work; C.J., L.Z. S.Q. measurements; D.L., L.Z., K.S., B.C., E.Y. R.M. analyses. authors final peer review history article available https://publons.com/publon/10.1111/1365-2745.13722. Data Dryad Digital Repository https://doi.org/10.5061/dryad.4qrfj6q9m (Huang 2021). Please note: publisher content functionality information supplied authors. Any queries (other content) should corresponding author article.