Bee species visiting Medicago sativa differ in pollen deposition curves with consequences for gene flow

The key determinant for forming pollen density and using crop barriers to mitigate wind-blown pollen dispersal in alfalfa (Medicago sativa L.)

The potential for commercial cultivation of genetically engineered (GE) alfalfa has raised ecological concerns due to the possibility of introgression of GE alleles into conventional populations. The main objectives of this study were to determine the key affecting factors (i.e. size of pollen source, number of pollinating bees) on forming alfalfa pollen cloud density and test the mitigating effect using maize barrier rows on alfalfa pollen dispersal. The results showed that the mean pollen densities of alfalfa pollen source (Ø = 10 or 20 m) were statistically similar when treated with the same number of worker bumblebees and increased accordingly with increasing the worker bees (887 and 853 pollens m-3 h-1 for Ø = 10 and 20 m with 100-150 worker bees, respectively; 1040 and 1070 pollens m-3 h-1 for the two plots with 200-300 worker bees, respectively), demonstrating that the number of worker bees but not the size of the pollen source was the key determinant for forming alfalfa pollen density. A maize barrier row established at 0.5 m from the alfalfa edge consistently decreased downwind pollen densities (percent pollen density of pollen source) to 0.2-4.4 % at 1-9 m compared to 3.4-25.4 % and 7.5-37.8 % at the same distance range for the upwind and downwind sites without maize barrier rows, respectively. Based on the pods formed on the emasculated alfalfa flowers (due to pollen dispersal) located at various distances from the pollen source and subsequent prediction model, the pollen density threshold value for fertilizing alfalfa recipient under the wind-blown condition was determined of 65.8 pollens m-3 h-1 at 14.7 m from the pollen source edge. The results would help in understanding the pollination biology (minimum pollen density for fertilizing alfalfa recipient ovule) and the process of pollen-mediated gene flow and helpful in developing management strategies to reduce the pollen density and thus mitigate the gene flow in alfalfa.

Pollen carryover, pollinator movement, and spatial context impact the delivery of pollination services in apple orchards

Assessing the relative contributions of different pollinator taxa to pollination services is a central task in both basic eco-evolutionary research and applied conservation and agriculture. To that end, many studies have quantified single-visit pollen deposition and visitation frequency, which together determine a pollinator species' rate of conspecific pollen delivery. However, for plant species that require or benefit from outcrossing, pollination service quality further depends upon the ratio of outcross to self-pollen deposited, which is determined by two additional pollinator traits: pollen carryover and movement patterns among genetically compatible plant individuals. Here, we compare the pollination capacities of managed honey bees, native bumble bees, and native mining bees in apple-a varietally self-incompatible commercial crop-when pollen carryover and pollinator movement patterns are considered. We constructed simulation models of outcross pollen deposition parameterized using empirically measured single-visit pollen deposition, visitation frequency, and probabilities of intertree movement exhibited by each pollinator type, as well as pollen carryover patterns simulated based on parameters reported in the literature. In these models, we also explicitly specified the spatial relationships among cross-compatible trees based on field-realistic orchard layout schemes. We found that estimated pollination service delivery was considerably reduced for all pollinator types when pollen carryover and pollinator movement patterns were considered, as compared to when only single-visit pollen deposition and visitation frequency were considered. We also found that the performance of different pollinator types varied greatly across simulated orchard layout schemes and pollen carryover scenarios, including one instance where bumble and mining bees reversed their relative rankings. In all simulations, native bumble and mining bees outperformed managed honey bees in terms of both outcross pollen delivery per unit time and per flower visited, with disparities being greatest under scenarios of low pollen carryover. We demonstrate the degree to which pollination studies may reach inaccurate conclusions regarding pollination service delivery when pollen carryover and pollinator movement patterns are ignored. Our finding of the strong context dependence of pollination efficiency, even within a single plant-pollinator taxon pair, cautions that future studies in both basic and applied pollination biology should explicitly consider the ecological context in which pollination interactions take place.

Differential ability of three bee species to move genes via pollen

Since the release of genetically engineered (GE) crops, there has been increased concern about the introduction of GE genes into non-GE fields of a crop and their spread to feral or wild cross-compatible relatives. More recently, attention has been given to the differential impact of distinct pollinators on gene flow, with the goal of developing isolation distances associated with specific managed pollinators. To examine the differential impact of bee species on gene movement, we quantified the relationship between the probability of getting a GE seed in a pod, and the order in which a flower was visited, or the cumulative distance traveled by a bee in a foraging bout. We refer to these relationships as 'seed curves' and compare these seeds curves among three bee species. The experiments used Medicago sativa L. plants carrying three copies of the glyphosate resistance (GR) allele as pollen donors (M. sativa is a tetraploid), such that each pollen grain carried the GR allele, and conventional plants as pollen recipients. Different foraging metrics, including the number of GR seeds produced over a foraging bout, were also quantified and contrasted among bee species. The lowest number of GR seeds set per foraging bout, and the GR seeds set at the shortest distances, were produced following leafcutting bee visits. In contrast, GR seeds were found at the longest distances following bumble bee visits. Values for honey bees were intermediate. The ranking of bee species based on seed curves correlated well with field-based gene flow estimates. Thus, differential seed curves of bee species, which describe patterns of seed production within foraging bouts, translated into distinct abilities of bee species to move genes at a landscape level. Bee behavior at a local scale (foraging bout) helps predict gene flow and the spread of GE genes at the landscape scale.