and S. 0 . FEJER. 1985. Recurrent selection in doubled-haploid populations of barley (Hordeurn vulgtrre L.). Can. J. Genet. Cytol. 27: 172-177.Cycle zero (C,,) of recurrent selection in barley (Hordeurn \ulgtrrc~ L.) was initiated by diallel mating of seven highly selected parents. A total of 398 doubled-haploid (CoDH) lines were derived from 2 1 crosses and were evaluated along with their parents in C,, experiment. Seven doubled-haploid lines (DH) wcre selected from the cycle zero (Co) experiment and intercrossed to form cycle I (CI). From the 2 1 crosses of the diallel, 260 doubled-haploid lines ( C I DH) wcre derived and were evaluated along with the Co and C I parents. ' The frequency distribution of the standardized means of the DH lines from C0 and C , indicated a slight response to selection for seed yield. Genetic analysis of the CIDH population showed high additive genetic variance for yield per hill, plant height, and yield per spike, and a high proportion of additive x additive epistasis for spikes per hill. days to heading, and 100-seed weight. Seven doubled-haploid lines were selected from different high-yielding crosses represented by CIDH lines. High selection pressure was applied for yield per hill, yield per spike, and spikes per hill. Further response to selection is expected in later cycles. ' The seven selected doubled-haploid lines will be used as the parents of the next recurrent selection cycle. Kev word.^: recurrent selection, doubled haploids, additive, epistasis, heritability. Hordeurn. PATEL, J. D., E. REINBERGS et S. 0 . FEJER. 1985. Recurrent selection in doubled-haploid populations of barley (Hordeurn vulgure L.). Can. J . Genet. Cytol. 27: 172-177. Le cycle zero (Co) d'une selection recurrente chez I'orge (Hordeum \~ulgarc) L.) a Ct C aniorcc par croisements diall6liques de sept parents hautement sklectionn@s. Un total de 398 lignees haplo'ides doublees (CODH) furent dCrivCes de 2 1 croisements; ces lignces, ainsi que les parents. furent evaluees au cours de I'expcrience C0. Sept ligndes haplo'ides doublCes (DH) furent s6lectionn@es de I'exp6rience du cycle (Co) et croisees entre elles pour former Ic cycle 1 (CI ). Dcs 21 croisements diall@liques, 260 lignees haploi'des doublkes (CIDH) furent d6rivi.e~ et Cvaluees en meme temps que les parents de C,, ct de C I . La frequence de distribution des moyennes standardisees des lignces DH de Co et C I n'indique qu'une faible reponsc quant ii la selection pour le rendement en grains. L'analyse genetique de la population de CIDH montre une variance gknktiquc ajoutke @levkc pour le rendement par butte, pour la hauteur des plantes et le rendement par cpi, ainsi qu'une proportion klevke d'kpistase additive x additive pour les epis par butte, les jours jusqu'ri la montaison et le poids de 100 grains. Sept lignkes haploi'des doublies ont kt6 selectionnees de diffcrents croisements ii rendement klev6 reprksentCs par des lignces CIDH. Des criteres de selection serres ont kt6 utilises pour le rendement par butte, Ie rendement par cpi et Ie rendcment...

From a series of 4 × 4 diallel crosses in the red raspberry (Rubus idaeus L.), it was determined that inheritance was predominantly additive for fruit yield and weight, fall-fruiting habit, day of flowering and total soluble solid content. However, it was mainly nonadditive for numbers of cane branches and laterals. Inbreeding depression was found for the yield of seedlings from sib-crosses within individual diallels. Reciprocal differences were noted for some characters.

Heritability and genetic variance components for 20 fruit and plant characters were investigated in 64 progenies produced among 31 North American and 1 German strawberry clone under the crossing scheme similar to North Carolina Design II (Comstock and Robinson, 1952). The data were analyzed based on the assumption of random and/or mixed models of the parents. The estimation of the genetic variance components indicated that for more than half the 20 characters nonadditive variance (dominance + epistasis) constituted approximately 50% or more of the total genetic variance and that in most cases epistasis played an important role in the nonadditive variance. Heritability estimates were low (less than 18%) for total berry yield and some fruit quality characters such as firmness, easy capping, pH value, soluble solids, and external and internal appearances but were high (more than 37%) for a number of yield component characters such as average berry weight, berries/flower stalk, yield/flower stalk and flower stalk number, suggesting that substantial improvement in total berry yield could be achieved through selection on these yield components or a linear function of them.

The performance in fruit yield characteristics and fruit appearance of seven Premier S5 inbred selections was studied in their crosses with four tester varieties, Redcoat, Senga Sengana, Sparkle and Jerseybelle. Some of the progenies produced by crossing Premier S5, inbreds with testers were superior in a number of yield characters to those produced when Premier was crossed to the same testers. This suggests that inbreeding, although expensive and time consuming, may be a useful step in one type of approach to breeding for higher yield in the cultivated strawberry.General combining abilities of the inbred selections and of the testers were both significant for all but one character, while specific combining ability of inbreds × testers was significant for six of the eight characters. The estimates of general combining ability effects of the inbreds and the testers were found to be of limited value in predicting the most desirable single crosses. The significance of this is discussed together with a proposed procedure for exploiting inbreeding in strawberry breeding.

Soil treatments with nitrogen, lime and fumigation of mixed grain seeded in a 1:1 ratio of barley and oats were applied for 2 yr. They induced relatively few changes in agronomic characters even though a sloping terrain was employed the first year compared with a flat area the second year. Fumigation reduced spot blotch development the first year but not the second year. Kernel yield and protein content were relatively higher in mixtures when compared with average values for pure stands but they were not as high as the highest component.