Trial heterogeneity and variance models in the genetic evaluation of potato tuber yield

Genetic evaluation aims to identify genotypes with high empirical breeding values (EBVs) for selection as parents. In this study, 2157 potato genotypes were evaluated for tuber yield using 8 years of early‐stage trial data collected from a potato breeding programme. Using linear mixed models, spatial parameters to target greater control of localised spatial heterogeneity within trials were estimated and variance models to account for across‐trial genetic heterogeneity were tested. When spatial components improved model fit, correlations of errors were mostly small and negative for marketable tuber yield (MTY) and total tuber yield (TTY), suggesting the presence of interplot competition in some years. For the analysis of multi‐environment trials, a variance model with a simple correlation structure (with heterogeneous variances) was the most favourable variance structure fitted for TTY and PTY (per cent marketable yield). There was very little difference in model fit when comparing a factor analytic structure of order 2 (FA2) with either FA1 or simple correlation structures for MTY, indicating that simple variance models may be preferable for early‐stage genetic evaluation of potato yield.     

Heritability, variance components and correlation of morphological and phenological traits in hazelnut

Most traits of interest in hazelnut are quantitative, yet there is little information available on heritability, variance components, or correlations. In this study, 13 morphological and four phenological traits were observed for 3 years on 41 parental genotypes, and for 2 years on seed‐lings representing 35 progenies. Narrow‐sense heritability estimates obtained by regression of offspring means on midparent values were all high, ranging from 0.56 for amount of kernel fibre to 0.89 for nut depth. Data on the parent genotypes was used to analyse variance components and showed that genotypic variance accounted for most of the phenotypic variance for all traits, and there was very little variation among trees within genotypes. Genotype‐year interaction accounted for only a very small part of the phenotypic variance for most traits, and changes in genotype rank were few and minor. Thus, selection based on genotype means is appropriate. The largest genotype‐year interaction components were observed for time of nut maturity, time of stigma exsertion, time of catkin elongation, number of nuts per cluster and amount of kernel fibre. For these five traits, evaluation over a period of years is advised. Correlation coefficients were large and positive for four nut size traits, indicating some redundancy in these measurements. Three of the phenological traits (time of stigma exsertion, time of catkin elongation, and time of leaf budbreak) were also highly and positively correlated. A large genetic correlation (0.44 < r < 0.73) was noted between good blanching and early phenology (catkin elongation, stigma exsertion and leaf budbreak), although plots of individual seedling values showed a large degree of scattering. Genetic correlations also indicated that seedlings with small nuts also tend to have compressed nuts, short husks and catkins that elongate later, but plots of individual seedling values showed large scattering. Correlations among the six traits directly related to the objectives of the Oregon State University hazelnut breeding programme (kernel weight, per cent kernel, nut shape index, blanching ability, relative husk length and nut maturity) were very low, indicating that they can be handled as independent traits.     

Genetic variation, heritability and genetic correlations for forage quality and yield traits of Mediterranean tall fescue germplasm

Higher ruminant intake potential and in vivo digestibility as indicated by lower neutral and acid detergent fibre (NDF and ADF), respectively, and higher crude protein (CP), may be valuable targets in tall fescue breeding. Some 105 full‐sib families from a diallel cross of Mediterranean parent germplasm were grown in a heated greenhouse simulating the temperature pattern of a Mediterranean environment, to estimate genetic parameters for NDF, ADF, CP and dry matter (DM) yield in one autumn and one winter harvest. Family x harvest interaction was large for NDF and ADF, moderate for CP and absent for DM yield. Additive genetic variance was mostly larger than specific genetic variance, and tended to be small for quality traits and large for DM yield. Narrow‐sense heritability was low for CP and ADF in autumn and moderate to fairly high in the remaining cases. Predicted gains per selection cycle never exceeded 5% for quality traits while being greater than 10% for DM yield. An inverse genetic correlation between DM yield and any quality trait emerged in the presence of sizeable genetic variation.     

Observed and Predicted Genetic Gains from Single and Multi-Trait Selections Using Three Mating Designs in Pearl Millet*

The objectives of the study were to estimate the genetic components of variance, predict the response to selection and to evaluate observed response to selection using two populations namely, PSB 7 and PSB 3, of pearl millet (Pennisetum americanum [L.] Leeks). The observed response was studied using four methods of selection based on single and multitrait criteria in three types of families, namely S1, half-sib (HS) and full-sib (FS) families. One hundred SI, HS and FS families each were developed in PSB 7 and PSB 3. Plants within a population were pollinated with bulk pollen of other population to develop HS families. Crosses were made between a pair of plants using one plant from each of the two populations to develop PS families. Data were recorded on days to flower, plant height, tiller number, spike length, spike girth, seed yield per plant and spike weight. Additive genetic variance estimates were significant for all traits. The estimates of dominance and additive × location and dominance × location interaction components were not significant, except additive × location component in PSB 7 for seed yield and spike weight. The genotypic component of variance among S1 families and S1 × location interactions were significant except S1 × location interactions in PSB 7 for days to flower, spike length and spike girth. Twenty-four strains were developed by intermating superior families identified on the basis of four selection methods in each of the six types of families. The selection methods were high seed yield (YS) per se; index selection (IS) for high tiller number and long and thick spike; truncation selection (TS) for high yield among the families having days to flower and plant height lower than the population mean: and visual selection (VS) after completion of flowering for good vigour, uniformity in plant height and good inflorescence attributes. YS and IS were found to be superior to TS and VS with respect to observed genetic gains for seed yield. That advance was higher in selection among SI families as compared with FS and HS families. In general, a good agreement was noted between observed and expected responses. Expected genetic gain from FS reciprocal (R) recurrent selection was higher than the average gain from HS-R selection.     

Recurrent selection in oat for adaptation to diverse environments

In order to test if selection can improve a population's adaptation to diverse environments simultaneously, three cycles of recurrent selection based on grain yield in Iowa, Idaho, and Norway were practiced in an oat (Avena sativaL.) population developed from North American, Scandinavian, and wild species (A. sterilis L.) germplasm sources. Specific objectives were to determine if selection: increased mean yields across environments and within all environments; changed the genetic correlation of yields in different environments; and changed genetic variation for yield within the population. We evaluated 100 to 210 randomly-chosen families from each cycle of selection at three Iowa locations, in Idaho, and in Norway for two years. Grain yield within each location and mean yields across locations increased significantly over cycles of selection. Mean yields across locations expressed as a percent of the original population mean increased at a rate of 2.6% per year. Several families from the third cycle population exhibited both high mean yields across locations and consistently high yields within all locations. Average genetic correlations of yield in different environments were higher in the second cycle than in the original population. A trend of reduced genetic variation and heritability was observed in Iowa only. These results suggest that we successfully improved mean population yield both within and across locations, and yield stability across environments, and in developing families with outstanding adaptation to diverse environments. This revised version was published online in July 2006 with corrections to the Cover Date.     

Estimation of genetic parameters and prediction of breeding values in an autotetraploid blueberry breeding population with extensive pedigree data

The strategy for breeding blueberries has been based on phenotypic selection without accounting for pedigree correlation information. A central premise of classical quantitative genetics is that through knowledge of the relationships among individuals in a population, we can make inferences about breeding values (BVs), estimate genetic parameters, and support phenotypic selection. Genetic evaluations using best linear unbiased prediction (BLUP) and restricted maximum likelihood (REML) are founded on pedigree information and have been the standard method used in livestock and forest breeding. Despite theoretical and practical benefits, their application in blueberry breeding programs remains unexploited. The objective of this study was to evaluate the use of REML/BLUP in the University of Florida blueberry breeding program in order to estimate genetic parameters and predict BVs of primary selection traits. To do this, we collected phenotypic data for eight selectable traits (yield, flower bud density, fruit weight, fruit firmness, fruit diameter, fruit soluble solids, fruit pH, and fruit scar diameter) from 1996 individuals. The highest narrow-sense heritabilities were found for fruit weight, scar diameter, and yield. The majority of genetic correlations between traits were under 0.20. The largest genetic gain was achieved when selections from early stages were used as parents and the top 5% of the population was selected. Compared to traditional phenotypic selection, our results are evidencing of the importance of implementing REML/BLUP to estimate genetic parameters that help and support the breeding strategy and maximize genetic gains.     

Genetic Variability, Genotypic and Phenotypic Correlations and Path-Coefficient Analysis in Egyptian Clover (Trifolium alexandrinum L.)

Fifty-six accessions of multi-cut Egyptian clover were compared at four experiments under different environmental conditions; i.e. location and season. Considerable variation existed for all the characters studied. The genetic variance exceeded the environmental variance for all the studied traits. Heritability was high for all characters studied. The expected genetic advance from selection of the superior 10.7% of the accessions as calculated from the combined data reached to 16.10, 17.50, 5.20, 20.60, 3.52 and 21.00 for mean plant height, seasonal fresh forage yield, mean dry matter percentage, seasonal dry forage yield, mean protein percentage, and seasonal protein yield, respectively. Both phenotypic and genotypic correlations among traits showed that mean plant height was positively correlated with each of seasonal fresh forage yield, seasonal dry forage yield, mean dry matter percentage and seasonal protein yield, but negatively correlated with mean protein percentage. Negative correlation were also found between mean protein percentage with seasonal fresh and dry forage yield.
Over all experiments, pat-coefficient analysis revealed that seasonal fresh forage yield had the highest positive direct effect on seasonal protein yield (0.841), followed by mean dry matter percentage (0.461). The mean protein percentage showed the minimum direct (0.172) and indirect influence on seasonal protein yield.     

Effects of selection for maize resistance to Sesamia nonagrioides on the additive and dominant components of genetic variance

Sesamia nonagrioides causes high yield losses in maize in northwestern Spain. Therefore, a recurrent selection programme to improve resistance against this pest has been carried out in a maize synthetic population. The main purposes of this investigation were to determine which are the main components of genetic variance for stem tunnelling and yield under infestation, to determine the variance remaining after selection, and, finally, to establish the genetic relationship between traits used as selection criteria and other ear resistance and agronomic traits. A North Carolina Design I mating system was used. Individual analyses by year were carried out and additive variance was identified as the only significant component of the genetic variance for stem tunnel length while a dominant variance was significant for yield. Those variances have decreased due to selection, but not significantly. Selection could continue for more cycles as there is enough variance. Additive genetic correlation coefficients between tunnel length and yield could not be determined. Thereafter, we propose three different methodologies of the selection to determine the relationship between resistance and yield under infestation.     

Genetic gains in two selection phases of a wheat-breeding programme

Genetic gain in field crops is usually estimated as the collective contribution of several breeding programmes. Critical to an individual breeder, however, is the genetic gain realized within a single programme. Our objective was to quantify per-annum genetic gains from 1969 to 1993 within two successive phases of experimental line testing in the wheat (Triticum aestivum L.) breeding programme of the Oklahoma Agricultural Experiment Station. Data for grain yield (Phases I and II), volume weight and heading date (Phase II) were analysed according to the procedure developed by St. Martin and McBlain (1991). Genetic gains for grain yield were indistinguishable between phases, averaging 4.6 ± 1.1 and 4.2 ± 1.3% of the control mean across the 24-year period. Genetic gain for volume weight averaged only 0.2% of the check mean, reflecting lower selection pressure for that trait compared to grain yield. Inconsistent genetic gains in heading date reflect a preference for a window of desirable heading dates rather than difficulties in selecting uni-directionally for maturity. The absence of a decline in genetic gain in recent years indicates that future improvement in grain yield is likely; however, increased attention must be given to volume weight to maintain US domestic and export standards for hard winter wheat. The statistical procedure was useful in monitoring long-term trends in genetic gains and identifying gaps in programme efficiency.     

Comparing the performance of cereal varieties in organic and non-organic cropping systems in different European countries

Top ranking varieties are tested in multiple environments before and after registration in order to assess their value for cultivation and use (VCU testing). Recently, interest has increased in obtaining varieties specifically adapted to organic farming conditions. This raised the question if an independent system of trials may be required for this purpose. To help answering this question, through the exchange network of European cereal researchers SUSVAR (), a number of data sets of agronomic traits from barley, wheat and winter triticale, from trials performed in Denmark, Sweden, The Netherlands, France, Switzerland, UK and Germany, were made available and analysed using an approach based on mixed models involving parameters describing genetic correlation between the two types of experiments, i.e., organic and non-organic (high or low input). Estimated variance components and correlations were used to evaluate response to selection and index selection. The response to index selection was analysed as a function of the fraction of available trials assigned to the organic system. The genetic correlations were interpreted in terms of ranking agreement. We found high genetic correlations between both systems for most traits in all countries. Despite high genetic correlations, the chances of very good agreement in observed rankings were moderate. Combining information from both organic and non-organic systems is shown to be beneficial. Further, ignoring information from non-organic trials when making decisions regarding performance under organic conditions is a sub-optimal strategy.     

A comparison of bulk testing and direct-line testing in self-fertilizing crops

Summary Computer simulation techniques were used to compare the efficiency of bulk-yield testing with direct-line yield testing. Bulk testing was almost always more efficient than direct-line testing for a wide range of values for population parameters. The differences between the two methods were smaller when the among-bulk heritability was low, the within-bulk genetic variance was high relative to the among-bulk variance, the variance of the within-bulk genetic variance was high, and the correlation between bulk means and within-bulk genetic variances was highly negative.With the direct-line testing method, discarding lines with low means after the first year of testing was always more efficient than retaining all lines for 2 years. Selection of 25% of the lines was slightly more efficient than selection of 50%. In general, it was better to evaluate more crosses and less lines per cross with the direct-line testing procedure.     

Genetic variability of cassava progenies developed through introgression of cassava mosaic disease resistance into Ghanaian landraces

Availability of genetic diversity is important in selecting suitable cultivars for crop improvement. Progenies of crosses between 19 farmer-preferred cassava landraces and Cassava Mosaic Disease (CMD)-resistant cultivar TME 11 were evaluated and compared with their maternal parents to assess their genetic variability and genetic gains based on CMD tolerance, some growth variables and root yield. The combined analysis of variance for the traits showed significant genotypic differences amongst the progenies and the parents. Apart from a few instances, the differences in replications and years were also significant. Genetic variability components showed high heritability ranging from 92.3 to 99.6% and moderately high genetic gains of 54.1 to 99.6% for the parental cultivars. Apart from canopy spread and plant height, heritability and genetic gains of the progenies were slightly lower than the parents indicating high variability within families of the progenies. The phenotypic coefficient of variation was slightly greater than genotypic coefficient of variation indicating minimal influence of environment on the genotypes. For most of the characters evaluated, the mean performance of the progenies was higher than their corresponding parental lines. Although the genotypes differed significantly (P < 0.001) for all traits evaluated, cultivars such as Afebankye, Bosome nsia, Cedi bankye, 262 Debor, Kwadaso 25, Nkaakom 57 and Sisipe were found to constitute a pool of germplasm with adequate variability. The negative correlation between CMD incidence and other traits evaluated showed significant progress made in the breeding program.     

河南夏大豆区近30年主要大豆品种产量改良的遗传进展

提高产量是大豆育种的主要目标。研究大豆产量及其相关性状的遗传进展,对于今后制定高产育种策略有重要参考意义。本研究随机选择近30年河南主要育成品种中的18个大豆品种,进行两年产量评价试验的研究。结果表明,产量随育种年份增加总体呈递增趋势,遗传进展为17.39kghm^-2,年递增率是0.7%;有效分枝、主茎节数、百粒重、株高有弱的正向遗传进展,而单株荚数和每荚粒数表现弱的负向遗传进展,但均不显著。百粒重、主茎节数和株高与产量有显著的遗传相关与表型相关,环境相关均不显著,表明这3个性状具有较大的遗传力,随产量性状的遗传改良,这3个性状均协同提高,且不易受环境条件的影响;而有效分枝、单株荚数、每荚粒数与产量的遗传相关和表型相关均没达到显著水平,这是它们与百粒重、主茎节数、株高有极显著负遗传相关所致。     

Estimation,interrelationships and repeatability of genetic variability parameters in spring safflower using multi-environment trial data

The present investigation was carried out to estimate the interrelationships and repeatabilities of various variability parameters, i.e., genotypic (σ²g) and phenotypic (σ²p) variances, genotypic (GCV) and phenotypic (PCV) coefficients of variation, broad-sense heritability (h_b²), genetic advance (GA), and genetic gain (GG), using data derived from a number of plant characters [seed yield (SY), days to heading (DH) and maturity (DM), plant height (PH), and thousand kernel weight (TKW)] of 17 spring safflower genotypes grown in 27 environments in Iran during 2003–2005. The analysis of variance for the five characters revealed significant differences among the genotypes in most of the environments, indicating a very high variability within the genotype. High genotypic and phenotypic variances in the quantitative traits, particularly for SY, were markers of increased success. A close correlation between GCV and PCV was observed for all the traits, indicating that all of the characters were less influenced by the environment and that the variability which exists in these characters is under genetic influence; however, the relatively higher values of PCV indicate the predominant role of environment. High expected genetic advances were observed for SY, but this trait showed large fluctuations in different environments (97.4%). High and low mean values of heritability were observed for DH (74.3%) and SY (59.7%), respectively, whereas high and low mean values of genetic gain were found for SY (60.3%) and DM (4.2%), respectively. High heritability coupled with low GA as a percentage of the mean was observed for DH and MD, whereas moderate heritability with moderate GA as a percentage of the mean was exhibited for SY. Analysis of correlation among the parameters showed that they were strongly and constantly correlated with each other, but none of them were consistently well correlated with the heritability parameter (h_b²). Based on estimates of the genetic variability parameters within each trial, the ranking of genotypes in the low-performance subset for all traits differed from that in the high-performance subset. This result indicates poor repeatability for the genetic variability parameters. The estimates of the parameters in the multi-year trials were almost constant and repeatable, whereas the responses over years showed poor repeatability.     

Effect of moving mean covariance adjustments on error and genetic variance estimates and selection of superior lines in peanuts (Arachis hypogaeal L.)

Summary Effects of moving mean covariance adjustments on error control, genetic variance component estimates and selection were examined in nine trails in peanut breeding nurseries. Neighbourhood sizes of two to fourteen adjacent plots were used in two types of moving mean covariates: one based on adjacent plot yields and a second based on adjacent plot residuals after removal of replicate and treatment effects. Both covariate types reduced error variance, in one trial by up to 57%, although the covariate type giving the most error reduction varied among trials. Generally most error variance reduction occurred when six to eight adjacent plots were used.In some trials genetic variance component estimates following covariance adjustment were higher than those from unadjusted analyses and in other trials they were lower. As the neighbourhood size included in the moving mean increased from two to fourteen plots the genetic variance component estimate changed considerably in some trials. However, large rank changes did not appear to be associated with these genetic variance component changes as neighbourhood size had little influence on the selection of progenies. Also, genetic effects included in the yield covariate had little influence on selection of elite lines as essentially the same lines were chosen or rejected when residual covariates lacking these genetic effects wre used.     

Line selection for exploiting durum wheat (T. turgidum L. var. durum) local landraces in modern variety development program

Summary Pure line selection for grain yield applied within four durum wheat local landraces (Triticum turgidum L. var durum) was studied in an effort to determine: (i) the agronomic performance of selected line mixtures (bulks) and their value as potential new cultivars, (ii) the effectiveness of pure line selection for grain yield within local landraces and (iii) the correlated selection response of yield components and kernel quality traits. Four local durum wheat landraces were used. Two of them, Mytilini-1 and Mytilini-2 are still cropped in some rural areas of the island Mytilini whereas the other two, Limnos and Mavragani are maintained in the Greek Gene Bank. Following their preliminary evaluation for agronomic performance under current growing conditions, 100 individual plants from each landrace were randomly selected. Selected plants (pure lines) were field evaluated head to row during 1997–98 and 1998–99 growing seasons. Based on grain yield performance, bidirectional selection, high vs low, was practised within each landrace. Thus the five higher yielding lines and their corresponding lower ones from each landrace were identified. Selected lines, with no further progeny testing, were pooled using a balanced seed composite to form the high yielding five line bulk and its respective low counterpart. The resulting 16 bulks (eight high and eight low) along with their four source landraces and two commercial varieties, Mexicalli and Simmeto, used as checks, entered field testing for agronomic performance in two and one location during 1999–00 and 2000–01 growing seasons respectively. Data for grain yield and yield components (1000 kernel weight, number of kernels, and productive tillers) along with kernel quality data (protein content and vitreous) were recorded. Data for grain yield indicated that the landraces studied seemed to be line mixtures, as expected, since they responded to bidirectional phenotypic pure line selection. The high yielding bulks were of interesting agronomic performance, equivalent to or significantly better than that of the commercial checks and are worth further testing as potential cultivars. Selected high bulks of Mytilini 2 landrace significantly outyielded both the source landrace and the commercial checks maintaining the high protein content of the source landrace, which was significantly higher than checks. This performance provides evidence that the within landraces selection of lines combining high grain yield with high protein content seems to be feasible. In conclusion, evidence is presented that genetic variability seems to be available within the landraces studied which could be worth using in breeding programs. Furthermore the simple pure line selection breeding effort undertaken seemed to be effective in developing line bulks with promising performance and worth further testing as potential cultivars.     

Selection for high spike fertility index increases genetic progress in grain yield and stability in bread wheat

Spike fertility index (SF) has been proposed as a promising trait to be used as a selection criterion in wheat breeding programs aimed at increasing grain yield, but no actual evidence of its successful application has been reported. In this study, 146 recombinant inbred lines derived from a cross between ‘Baguette 10’ and ‘Klein Chajá’, Argentinean spring bread wheat cultivars with contrasting SF, were evaluated during three crop seasons (2013, 2014 and 2015) at Balcarce, Argentina. Grain yield, grain number/m², grain weight, and SF were measured at maturity. Changes in grain yield (i.e., responses to selection) after application of different selection strategies, including different selection criteria and selection intensities, were determined. Significant correlations were observed between grain number and grain yield, SF and grain yield, and SF and grain weight. Analysis of SF variance components showed a significant genotype?×?environment interaction, but it represented only 9% of the total variation, whereas 51% of the variation was genetic, resulting in a high narrow-sense heritability (0.84). The use of SF as a selection criterion, either solely or in combination with selection for high yield, increased yield, resulting in higher and more stable yields than if selecting for high yield alone. Our findings support the use of spike fertility index as a selection criterion for increasing genetic progress and stability of yield in bread wheat breeding programs.     

Selection for Seed Yield Production of Egyptian Clover (Trifolium alexandrinum L.) C. V. Fahl

Modified mass and family selection for seed yield production of berseem clover (Trifolium alexandrinum L.) c.v. ‘Fahl’ were applied for two generations. Two hundred plants (5 % intensity of selection) were selected for seed yield in the first season, 1985/86. In the second season, 1986/87, selection between and within half-sib families was practiced. In addition, equal parts of seeds from each of the 200 selected plants were bulked to form the C₁ modified mass selection; after establishing, the same procedure was adopted to form the C₂ generation. The cycles 1 and 2 of half-sib families and modified mass selection along with the base population family were evaluated for forage and seed yields. The realized gains from modified mass selection were 6.03 and 9.51 % for fresh forage yield, 5.57 and 10.86 % for protein yield and 13.23 and 16.19 % for seed yield in cycles 1 and 2, respectively, over the base population. The realized gain from family selection in cycle 2 as a percentage of the base population mean amounted to 11.32, 13.35, 17.47 and 3.15% for forage, protein, and seed yield and seed index, respectively. The broad sense heritability, as estimated from the variance components was 89.63, 63.03, and 76.67 % for dry forage, seed yield and seed index, respectively. Although, all these five traits (fresh, dry, protein, and seed yield and seed index) had positive correlation with each other, weak correlations were found between seed yield and forage yields. Furthermore, close associations were found among forage yield traits.     

滇中南巨桉种源/家系年度变异分析

为开展巨桉种质源引进,增加遗传多样性,对建立在滇中南地区11 个巨桉种源的173 家系开展生长性状的年度变异分析。结果表明：参试性状在所有林龄内的种源和家系间均呈极显著差异,参试性状在区组间的F 值逐年减小,区组间16 月生、28 月生和41 月生时差异极显著,53 月生时差异不显著。胸径的表型和遗传变异系数在23%~46%之间,树高的表型和遗传变异系数在19%~30%之间,单株材积的表型和遗传变异系数在60%~75%之间,参试性状遗传变异系数略低于表型变异系数。相关性分析表明,各林龄间和各参试性状间均呈极显著正相关,各林龄内性状间相关性在0.84~0.94 之间;随各林龄时间间隔增长,林龄间性状相关性有下降趋势。整个生长期内,来自昆士兰州Copperlode、昆士兰州Bambaroo 和四川乐山的种源长势良好,来自昆士兰州Ravenshoe、昆士兰州Tinaroo 和昆士兰州TullyGorge National Park的种源长势较差。     

Evaluation of floral characteristics of barley in the semi-arid climate of north Syria

Previous studies have revealed that, in areas affected by severe drought, yielding ability and stability of barley (Hordeum vulgare L.) could be improved by increasing the level of heterozygosity. One possible approach to increase the heterozygosity level is to improve the outcrossing rate. In view of that, the genetic variability and phenotypic correlation coefficients involving four floral traits related to outcrossing behaviour were investigated in 274 drought adapted barley lines from the International Center for Agricultural Research in the Dry Areas (ICARDA's) current breeding programme. A significant quantitative variation existed in anther length and width, stigma length, and anther extrusion, which indicate that selection for these traits may be effective. Anther extrusion showed weak but highly significant positive correlations with anther and stigma length. Somewhat higher positive correlations were found between anther and stigma length. The results indicate that selection for high anther extrusion and longer anthers and stigmata may increase outcrossing in cultivated barley and consequently improve yielding ability through enhanced heterozygosity.