Plant–animal interactions mediate climatic effects on selection on flowering time

Abstract

Selection on flowering time in plants is often mediated by multiple agents, including climatic conditions and the intensity of mutualistic and antagonistic interactions with animals. These selective agents can have both direct and indirect effects. For example, climate might not only influence phenotypic selection on flowering time directly by affecting plant physiology, but it can also alter selection indirectly by modifying the seasonal activity and relative timing of animals interacting with plants. We used 21 yr of data to identify the drivers of selection on flowering time in the perennial herb Lathyrus vernus, and to examine if antagonistic plant–animal interactions mediate effects of climate on selection. We examined the fitness consequences of vertebrate grazing and predispersal seed predation, and how these effects varied among years and among individuals within years. Although both antagonistic plant–animal interactions had important negative effects on plant fitness, only grazing intensity was consistently related to plant phenology, being higher in early‐flowering individuals. Spring temperature influenced the intensity of both plant–animal interactions, as well as the covariance between seed predation and plant phenology. However, only differences in grazing intensity among years were associated with differences in selection on flowering time; the strength of selection for early flowering being stronger in years with lower mean intensity of grazing. Our results illustrate how climatic conditions can influence plant–animal interactions that are important selective agents for plant traits. A broader implication of our findings is that both ecological and evolutionary responses to climatic changes might be indirect, and largely mediated by species interactions.

Early flowering changes robusta coffee yield responses to climate stress and management

“…To investigate the complex interlinked relationships between flowering time, climate stress and coffee yield we used structure equation models (SEM) (Lefcheck, 2016). SEM modelling has been well utilised in phenological studies (Esch et al, 2019;Li et al, 2021;Valdés and Ehrlén, 2021). There are two key features of SEM that differentiate this approach from traditional regression modelling and make it suitable for investigating how early flowering impacts coffee yield responses to climate.…”

Spring temperature drives phenotypic selection on plasticity of flowering time

“…Our study was observational, and without experimental manipulations it is not possible to conclusively identify the causal link between April temperature and selection on plasticity. In our study system, it is possible that spring temperature influenced selection both directly, and indirectly through effects on biotic interactions, such as grazing and seed predation [ 35 ]. Spring temperature influenced the relationship between plasticity and the probability of producing any seeds, but not the relationship with the number of intact seeds.…”

“…In our analyses, we included also individuals that were grazed before flowering and estimated their expected first flowering date in order to account for the fact that some individuals (the invisible fraction [ 39 ]) were exposed to factors influencing fitness before expressing the trait of interest. It is therefore possible that grazers mediated some of the effects of spring temperature on selection by preferentially attacking plants with a certain phenology [ 35 ]. Our findings of selection against high plasticity in flowering time being strongest at low temperatures agree with the results for other species showing that more responsive genotypes have higher fitness under elevated temperatures [ 26 ].…”

“…L. vernus is self-compatible, but it lacks mechanisms for autogamy (J. Ehrlén, unpublished data). Individual flowering time influences the intensity of interactions with animals that may act as selective agents [ 35 ]. Early-flowering plants are more damaged by roe deer ( Capreolus capreolus ) than late-flowering individuals, but still produce on average more seeds per flower, possibly due to higher light and pollen availability early during spring [ 34 , 36 ].…”

“…These plants might either have flowered and were grazed before we observed the opening of the first flower, or were grazed prior to opening of the first flower. In the latter case, we calculated an ‘expected’ first flowering day to account for that grazers are potentially important agents of selection on flowering time [ 35 ].…”