Annual ryegrass (Lolium multiflorum Lam.) is a cool season annual bunchgrass possessing high palatability, excellent forage quality, good seedling vigor, persistence under close grazing and impressive dry matter yield. It has a high growth rate during the reproductive phase in spring but makes relatively little growth during winter. Availability of standing forage for grazing in the southeastern USA is limited during the cooler months and improved winter growth would be a great benefit to cattle producers. The objective of this study was to evaluate the progress from selection for increased winter dry matter production in annual ryegrass resulting from two cycles of phenotypic recurrent selection. Intermating 50 plants from the six top performers in Alabama Annual Ryegrass Trials formed the base population for the selection experiment. Bulk-harvested seed from the first synthetic generation was polycrossed again to create a random mating population (C0) for the selection experiment. A 1200-plant nursery was established in late October of each year at the Plant Breeding Unit of the E.V. Smith Research Center, Tallassee, AL with plants spaced 90 cm in all directions. The nursery was subdivided into 48 blocks of 25 plants each (5 x 5 arrangement). Plants were harvested and evaluated for dry matter determination 750 growing-degree-days (GDD) after transplanting. The best entries selected from each block were then polycrossed in isolation and bulk-harvested seed used for the next cycle. Seed for populations C0, C1, and C2 was increased in isolation at the same location during the 2007/8 cropping season. In autumn 2008, seeded plots were established at five locations in Alabama (Belle Mina, Winfield, Tallassee, Headland, Fairhope) to assess changes in forage yield in response to selection. Populations C0, C1, and C2 as well as check cultivars Gulf, Marshall, and Shiwasuaoba were established in a RCB design (r = 4) at each location. Plots were fertilized with 50 lbs N acre-1 at seeding and after each harvest. The first harvest occurred 1000 growing-degree-days (GDD) post seeding, allowing 250 GDD for establishment and 750 GDD for dry matter accumulation. Thereafter, the target GDD was 750. A significant linear improvement over 2 cycles was obtained at three of the five locations, ranging from 39 lbs acre-1 cycle-1 at Belle Mina, AL to 426 lbs acre-1 cycle-1 at the selection location in Tallassee, AL, and 354 lbs acre-1 cycle-1 at Headland, AL. No change in first-cut yield was observed at Fairhope, AL. The explanation for this non-response may be that average temperature at this location close to the Gulf of Mexico never dropped below 40 F and was in the 50-60 F range for most of the dry matter accumulation period prior to harvest 1 in early January 2009. The negative response (-113 lbs acre-1 cycle-1) at Winfield correspondingly may be related to the extreme cool temperatures at that location indicated by the fact that it took > 5 months to accumulate 1000 GDD compared to 3 months at Tallassee and 2.5 months at Fairhope. Shiwasuaoba out-yielded C2 at all locations, ranging from 222 lbs acre-1 at Belle Mina to 1763 lbs acre-1 at Headland. Population C2 yielded at least 940 lbs acre-1 more than Marshall at Tallassee and Headland, whereas the differences at other locations were 162 lbs acre-1 or less in favor of Marshall. The trends for Gulf vs. C2 were similar to Marshall. We can thus conclude that selection in annual ryegrass for improved winter productivity was successful.