The objective was to combine a set of crossing and evaluation techniques into a regime that permits a breeder to conduct one cycle of recurrent selection for protein or grain yield of oats (Avena sativa L.) per year. Parental strains are crossed in the greenhouse during September to December, S0 plants are propagated in the greenhouse during December to March, S0.1 lines are evaluated in a replicated multilocation experiment in the field during April to July, and data summary and selection of parental lines for a new cycle occur during August. The schedule is repeated for each cycle. Over three cycles, the gain per cycle (and per year) was 4.7% for protein yield and 5.4% for grain yield. Techniques employed to permit completion of a cycle of recurrent selection in 1 year were: (i) approach method of crossing, (ii) evaluating S0.1 lines, (iii) hill plots for field evaluation, (iv) infrared analyzer for N determinations, and (v) computer data analyses and line selection.

Cereal crops generally are considered to be a source of energy in food and feed, but they also supply 60 and 25% of the human dietary pro tein in developing and developed countries, respectively (Orr, 1978). Research on the improvement of oats (Avena sativa L.) as a protein source has been limited relative to that done on wheat (Triticum aestivum L.)> rice (Oryza sativa L.)« and maize (Zea mays L.). Oats rank sixth among the grains in world production, and its protein yield per ha exceeds that of all the legumes except soybeans (Glycine max (L.) Merrill) (Bright and Shewry, 1983). Oat groats (caryopses) are higher in protein percentage and in protein production per ha than other cereals, and their amino acid profile is superior to all cereals except rice and rye (Secale cereale L.) (Pomeranz, 1973; Youngs et al., 1983; Zarkadas, 1982). Also, unlike other cereals, as protein percentage of oats is enhanced, the amino acid profile remains unchanged (

Three oat {Avena sativa L.) populations (i.e., lines of descent), high grain yield (HG), high protein content (HP), and high protein yield per se (HGP), each developed by three cycles of Si recurrent selection, were evaluated for the effect of selection for groat-protein yield upon other agronomic traits. Selections making up the HG line of descent had high protein yield primarily due to high grain yield, and those selected for HP had high protein yield due to both high protein content and high grain yield. Selection in HGP was on the basis of protein yield per se. Selection caused increases in bundle weight, harvest index, vegetative growth rate, and seed number in all lines of descent. Heading date, plant height, and seed weight were unaffected, whereas groat percentage and test weight were decreased in HP and HGP. Heritabilities were high for heading date, plant height, test weight, and seed weight, moderate for harvest index and bundle weight, and low for groat percentage. Genetic variability generally declined from GO to G3 for all traits.Groat-protein yield and amount of protein per groat increased in all lines of descent. In HG, the increase in groat weight was due primarily to increases in the nonprotein fraction, with groat-protein content actually decreasing. In HGP, groat weight increased due to increases in protein and nonprotein fractions, and groat protein content remained constant. In HP, groat weight and amount of nonprotein per groat decreased, thus increasing groat-protein content.

Hill plots are used extensively for testing small‐grain lines in early generations when seed supplies are limited and large numbers of genotypes must be evaluated. Hill plots and multiple‐row plots estimate the value of a genotype similarly, yet little is known about the effects of seeding rate on performance of cereal genotypes in bills. We evaluated 30 oat (Avena sativa L.) genotypes for grain and protein yields, heading date, plant height, and groat protein concentratian in hill plots with seeding rates of 5, 10, 15, 20, 25, and 30 seeds per hill (equivalent to 18, 36, 54, 71, 88, and 106 kg ha‐1). In another set of treatments, barley (Hordeum vulgare L.) was sown with the five lowest oat seeding rates so that all hills were planted with 30 seeds. It was intended that the barley plants would mimic the competitive effects of oat plants at the standard seeding rate of 30 seeds per plot. The experiment was grown in a split‐plot design at two locations for 2 yr. Significant differences existed among seeding rates for all traits, but most seeding rates ranked the 30 genotypes in similar order. The use of barley to mimic the competitive effect of oat at the standard sowing rate was successful, but the extra cost of using this technique is not justified. Seeding rates for oat as low as 20 seeds per plot did not affect trait expression or genotypic rankings, and it requires one‐third fewer seeds. A reduction in seeding rate will permit small grain breeders to expand their breeding operations to include schemes such as early generation recurrent selection.