A utotetraploid alfalfa (2n = 4x = 32) is the third largest crop by acreage in the United States with an estimated 28 million kg of seed produced per year (NASS, 2009). Although alfalfa tends to outcross (i.e., xenogamous), it is a self-compatible species meaning that self-pollination or "selfing" is possible. Selfing in alfalfa is more likely to occur in the absence of pollen from another genotype, suggesting that, when present, outcross pollen may outcompete self pollen (Viands et al., 1988). Selfing has been shown to occur in insect-pollinated alfalfa seed production fields. Reported field-wide selfing rate estimates range from 9 to 53% (Burkart, 1937, 15% selfing;Johansen, 1963, multiple studies 9 to 26% selfing; Bradner and Frakes, 1964, 34% selfing; Pedersen, 1968, 53% selfing; Knapp and Teuber, 1993, 24% selfing; and Brown and Bingham, 1994, 28% selfing). Most of the pre-1990s studies used white flowered "sentinel" plants in fields of purple flowered alfalfa plants to estimate selfing rates. The pollinators used in these studies varied widely or were not mentioned and ABSTRACT Alfalfa (Medicago sativa L.) self-pollination (i.e., selfing) causes inbreeding depression. Determining factors influencing alfalfa seed production selfing rates could inform potential mitigation strategies to reduce selfing. We measured in situ selfing rates from seed sampled from random plants in a commercial alfalfa seed production field pollinated by leafcutter bees (Megachile rotundata F.). Alfalfa selfing rates were estimated by genotyping ~24 progeny from each of 38 maternal plants. Maternal plant distance to pollinator domicile, pod position on racemes, raceme position on stems, and seeds per pod were noted during seed and tissue collection. Selfing rates averaged 11.8% with individual selfing rates ranging from 0% to 52.2%. Seed from pods collected from upper parts of racemes had lower selfing rates (9.1%) compared to pods from lower parts of racemes (15.1%). When "low" self-compatible (<15% selfing rate in 3+ seeded pods) and "high" self-compatible (≥15% selfing rate in 3+ seeded pods) plants were examined separately, however, this pattern remained significant only for low self-compatible plants (upper raceme selfing rates 3.1% vs. lower raceme 8.3%). Low self-compatible plants had higher selfing rates in 1-2 seeded pods (12.9%) compared to 3+ seeded pods (3.8%) while high self-compatible plants showed no differences in selfing rates based on seed number per pod. Genetic differences for self-pollen's ability to outcompete outcross pollen when growing down the style best explained observed differences between low and high self-compatible plants. Best management practices and selection could help reduce but not eliminate selfing in alfalfa seed production fields.
