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.     

Heritability of fruit shape in pears

Heritabilities for each of three different height and width ratios describing fruit shape were estimated from measurements on a population of seedlings comprising 17 families of European and Asian pear parentage in the Hort Research pear cultivar breeding programme. In families with European cultivars as parents, the pyriform curvature in the upper, or stem end, half of the fruit was also quantified and heritabilities estimated. Heritabilities were computed using parent-offspring regression and from consideration of the variance components. Heritabilities estimated from parent-offspring regression were generally high and higher than those estimated from variance component analysis, although within the limits of error. Heritability estimates for the height: maximum width ratio were > 0.5 by both methods of estimation indicating that fairly rapid genetic gain could be expected. This revised version was published online in July 2006 with corrections to the Cover Date.     

Inheritance of taste characters and fruit size and number in a diploid Actinidia chinensis (kiwifruit) population

We investigated the variability and inheritance of taste components and fruit size and number in a seedling population of diploid Actinidia chinensisPlanch. We found significant variation in all characters measured. Narrow-sense heritability estimates were high for vitamin C content, titratable acidity (TA), soluble solids content (SSC), fruit weight and dry matter content (DM), but low for glucose content, quinic acid content, and fruit number per vine. The genetic correlations between sugars (except myo-inositol), acids (except quinic acid), TA, vitamin C content, SSC and DM tended to be positive and moderate to high. However, correlations between these characters and fruit weight tended to be negative and moderate to high. The standardised coefficients () of multiple regression showed that SSC was highly dependent on the levels of fructose content, sucrose content and TA. Estimates of relative selection efficiencies (E) suggested that selection using SSC could be an efficient and convenient alternative to direct selection for improvement of sugar levels in kiwifruit. However, selection strategies based on SSC need to take account of its moderate negative correlation with fruit weight. Male and female parents useful for improving acid and sugar composition and fruit size were identified. One family was outstanding, as it combined large fruit size with high levels of SSC and DM.     

Effectiveness of selection for plant characters of barley in simulated segregating rows

Summary Single plants of 12 barley (Hordeum vulgare L. and H. distichum L.) cultivars were randomized to form simulated segregating (mixed) rows. There were 16 mixed rows 15 cm apart, each row containing 24 plants 5 cm apart. In addition, every third row was a control row of one of the cultivars, Bonanza. Each plant was harvested separately.The cultivars differed significantly for all seven characters measured. These seven characters fell into 3 performance groups. Group I characters, seeds/head, 1000 kernel weight, days to ripe and height, showed a higher variance in the mixed rows than in the control rows, had high average heritabilities, and were consistent in performance from one row to another when measured by correlation amongst the cultivars. In Group II, yield/plant and seeds/plant, variance, heritabilities and performance consistency were all low or non-significant. Group III, heads/plant, fell between Groups I and II. These results suggested that single plants with the desired genotype could be successfully selected for characters in Group I only.The frequency with which the most desirable genotype would actually have been chosen with 12.5% selection intensity generally confirmed the above conclusions but indicated that the degree of differences among genotypes was also important in selecting the most desirable genotypes.Contribution No. NRG 77–9 from the Beaverlodge Research Station.     

Heritability estimates of spaced-plant traits in three perennial ryegrass (Lolium perenne L.) cultivars

Summary Seed yield in perennial ryegrass is low and unpredictable. Spaced-plant traits suitable for indirect selection for total seed yield in drilled plots would be very useful. The objectives of this investigation were to evaluate genetic variation for seed yield components and other traits among clones from three perennial ryegrass cultivars differing in seed yield and their open-pollinated progenies. Per cultivar, a random set of 50 genotypes was cloned and on each genotype seed was generated by open pollination. Clonal ramets of the parents were observed for 17 traits in 1986 at two locations. In 1987 and 1988, parents and progenies were observed as single plants in a randomized complete block design with two replications. There was little cultivar-environment interaction for most traits. The parents differed significantly for almost all traits. Half-sib (HS) families differed for only three to five traits. Broad-sense heritabilities (h² _b), based on variance components of the parents, were moderate to high; earliness had the highest h_b ². Narrow-sense heritabilities (h² _n), based on variance components among HS-families, were low to moderate and mostly not significant; for most traits h² _n estimates varied between years and cultivars. Flag leaf width and date of first anthesis showed the highest h² _n. Narrow-sense heritability estimates from parent-offspring regressions (h² _nPO) ranged from non-significant to high, depending on year and cultivar; they were generally higher than the corresponding h² _n estimates. Generally, h² _nPO was highest for earliness, flag leaf width, ear length and the number of spikelets per ear. Breeding methods that capitalize on additive genetic variance, such as mass selection, should result in improvement for these traits.     

Heritability of oil content and the onset of flowering in a population of Dimorphotheca pluvialis

To develop efficient breeding strategies for the improvement of oil content and onset of flowering in Dimorphotheca pluvialis, narrow-sense heritabilities of these characters were estimated. Forty plant progenies were tested in 2 years and heritabilities were estimated from parent-offspring regression as well as from half-sib family variance components analyses. For oil content, the heritabilities estimated by means of parent-offspring regression were 0.36 and 0.33 for 1993 and 1994, respectively. The heritabilities obtained from the variance components analyses were 0.41 and 0.23 for the two years. Heritability estimates for onset of flowering showed more variation. From parent-offspring regression, the heritabilities were 0.72 and 0.37, and from variance component analysis the estimates were 0.53 and 0.19 for 1993 and 1994, respectively. The obtained heritability values and realized responses suggest that particularly in early selection generations, even with mass selection, considerable progress can be expected. Since correlation studies revealed neither phenotypic nor genotypic correlation between oil content and onset of flowering, selection for both characters can be undertaken independently.     

Genetic variance of six agronomic characters in three lentil (Lens culinaris Medic) crosses

Summary Three lentil (Lens culinaris Medic) crosses (cross 1 = Chilean × PI 297784; cross 2 = Tekoa × PI 212611; cross 3 = Precoz × PI 212611) were advanced in a hierarchal design from the F₂ generation to the F₅ generation and tested in replicated field experiments. The purpose was to estimate the genotypic variances and covariances for several characters and to partition these into components due to additive, dominance, and additive × additive variances.Additive genetic variance was the major component of variance in cross 2 for all characters, except plant height and seed weight. Unexpectedly, estimates of dominance variance appeared to be high in crosses 1 and 3. Estimates of the additive × additive component seemed very small in all three populations. The dominance variance component estimates were consistently high for plant height in the three crosses and for seed weight in crosses 2 and 3. Where parental means were close, additive variance estimates were often low and in many cases negative. The unexpected high ratio of dominance variance to additive variance indicated that dominance variance was important in the crosses we studied.Cooperative investigations of Agricultural Research Service, US Department of Agriculture and the College of Agriculture, Washington State University, Pullman, Washington 99164, Work sponsored by Fellowship from Univ. of Jordan, Amman, Jordan. Scientific paper no. 5825.Formerly: Department of Agronomy and Soils, Washington State University, Pullman, WA., USA. Presently: Department of Plant Production and Plant Protection, University of Jordan, Amman, Jordan.Statistical Services, Washington State University, Pullman, WA 99164, USA.US Department of Agriculture, Agricultural Research Service, Washington State University, Pullman WA 99164, USA.     

The stability of the number of tillers of barley varieties and its relation with consistency of performance under semi-arid conditions

Summary In 1981/82 60 barleys varieties were grown at 18 environments (three locations, three seed rates, two soil fertility regimes) and in 1982/83 22 varieties were grown again at 21 environments (seven locations, three seed rates). Grain yield was recorded only in the second year. The estimates of variances s_x ² (untransformed data) and s² _logx (transformed data) for number of tillers and grain yield varied significantly among varieties. When untransformed data were used, high tillering varieties had a higher variance for number of tillers per unit area than low tillering varieties. When the variation was measured by CV of untransformed data or by s² _logx on transformed data a reverse relation occurred, i.e. the correlation coefficient between number of tillers and the variation of number of tillers was negative.High grain yielding varieties had a high mean number of tillers and a low variance for number of tillers per unit area when transformed data were used to compute variances. All interactions between variety, seed rate and location for grain yield were significant. The regression analysis of variety grain yield (Y) on environment mean grain yield (X) gave regression coefficients, b, ranging from 0.51 to 1.69. There was no significant correlation between regression coefficient and grain yield. The highest yielding varieties had b values around 1.0. High grain yielding varieties had low variance of yield over environments and low values for deviations from the regression when transformed data were used. However, the correlations between mean yield and variance for yield or mean yield and deviations from regression were positive when untransformed data were used.From this study it is concluded that high mean number of tillers per unit area and low variance for number of tillers could be used in selecting varieties with consistently high yield at varying environments. Techniques are proposed for application early in the breeding programme, i.e. in segregating population or in nurseries.     

Reciprocal recurrent selection applied to Coffea canephora Pierre

Summary Genetic parameters were estimated in trials planted in connection with a reciprocal recurrent selection programme applied to Coffea canephora in Côte d'Ivoire. Narrow sense heritability values, estimated from a breakdown of the analysis of variance, were high for architectural characters (0.22 to 0.78) and medium for tree vigour (0.13 to 0.40) and for bean weight (0.15 to 0.28). Calculated heritability values for the first two harvests were high (h²>0.7), but much lower for the following harvests (less than 0.2). Heritability estimates by parent-offspring regressions gave variable results, though they were often similar to those obtained by variance analysis, especially for bean weight. Phenotypic and genetic correlations between tree vigour traits and productivity were high. Bean weight was not correlated with vigour or production. Canopy diameter of 4 year-old trees was closely correlated with cumulated productivity from 2 to 5 years in the absence of development competition between trees, but the correlation was low if trees were in competition with each other. The consequences of the results for the choice of characters to be selected among intergroup hybrids or the parents of both populations are discussed.     

Analysis of quantitative variability for seed yield and related characters in Lotus corniculatus L. cv. Leo

Summary Quantitative variability for seed yield and six other characters was analysed in Lotus corniculatus L. cv. Leo. The material consisted of 144 polycross progenies and 100 parents.Wide variability existed for all characters. The characters with the greatest variability were seed yield, forage grading and plant height. The polycross progeny test was employed to study the general combining ability of the parents. Highly significant differences existed for all seven characters under study.Parent-offspring genotypic and phenotypic correlations were high and significant for all characters except genotypic correlations for seed yield and seeds per pod. High h² values (broad sense) were obtained for seed size and days to flowering. Traits with moderate to high h² were seed yield (71% in parents, 64% in progenies), plant height, forage grading, and seeds per pod. The character pods per inflorescence had the lowest h².Positive estimates of % MathType!MTEF!2!1!+-% feaafiart1ev1aaatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn% hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr% 4rNCHbGeaGak0dh9WrFfpC0xh9vqqj-hEeeu0xXdbba9frFj0-OqFf% ea0dXdd9vqaq-JfrVkFHe9pgea0dXdar-Jb9hs0dXdbPYxe9vr0-vr% 0-vqpWqaaeaabiGaaiaacaqabeaadaqaaqaaaOqaaiqabo8agaqcaK% aaavaabeqaceaaaeaacaqGYaaabaGaaeiraaaaaaa!3A89!\[{\text{\hat \sigma }}\begin{array}{*{20}c} {\text{2}} \\ {\text{D}} \\ \end{array} \] were obtained only for seed size. The ratio of dominance variance to additive variance indicated partial dominance for this character. Except for seed yield, in all other cases these estimates had very high sampling errors. In all cases except pods per inflorescence and seeds per pod high positive estimates of % MathType!MTEF!2!1!+-% feaafiart1ev1aaatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn% hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr% 4rNCHbGeaGak0dh9WrFfpC0xh9vqqj-hEeeu0xXdbba9frFj0-OqFf% ea0dXdd9vqaq-JfrVkFHe9pgea0dXdar-Jb9hs0dXdbPYxe9vr0-vr% 0-vqpWqaaeaabiGaaiaacaqabeaadaqaaqaaaOqaaiqbeo8aZzaaja% qbaeqabiqaaaqcaauaaiaaikdaaKaaGeaacaqGbbaaaaaa!3B30!\[\hat \sigma \begin{array}{*{20}c} 2 \\ {\text{A}} \\ \end{array} \] were obtained.The data indicated that it may be possible to simultaneously improve seed yield and maintain forage yield. Seed yield had positive and significant associations with seed size, seeds per pod and pods per inflorescence. The associations of days to flowering with forage grading (negative) and with pods per inflorescence (positive) were also significant.     

Agro-morphological characterisation of common vetch and its close relatives

The Vicia sativa aggregate, which includes the common vetch, is a complex of six closely related tax a which combines cultivated, weedy and wild forms. In this study 454 accessions with an emphasis on Mediterranean and Central Asian accessions have been compared for 22 key agro-morphological and phonological characters. The agronomic potential of the different members of the aggregate is assessed and useful variation existing in the collection is determined. Geographic patterns of the agronomic variation are studied to determine sources of useful variation. In both sub sp. cordata and sub sp. sativa accessions were found which had higher yields than the checks, consisting of three promising accessions. Sub sp. macrocarpa needs a longer growing season than the other tax a to reach its full potential. The other three tax a in the aggregate all had quite low yields. Important variation in shattering of pod sand in tolerance to moderate frost and resistance to the Sitona weevil was found in most subspecies of the aggregate. Variation in phenological characters appeared to be related to the latitude from which the accession was collected. This revised version was published online in July 2006 with corrections to the Cover Date.     

Genetic analysis of heading date and other agronomic characters in barley (Hordeum vulgare L.)

Genetic analyses of heading date, tiller number, plant height, grain yield, kernel weight, and plump and thin kernels were made in three six-rowed barley crosses (Hordeum vulgare L.) involving four cultivars. Six populations, P₁ , P₂ , F₁ , F₂ , BC₁ , and BC₂ , from each cross were grown and evaluated at Fargo and Prosper, North Dakota, 1982. Parental means within crosses generally were different except for tiller number. Comparison of generation means suggested that late heading was dominant to early, high kernel weight was dominant to low, and kernel plumpness was influenced by additive gene action. The relationship between yield and heading date was not consistent among crosses and positive r values were quite low. It should be possible to select early maturing, high yielding segregates with plump kernels. Heterosis over the mid-parent was quite similar among crosses for heading date, but there was no heterosis over the high parent. Inbreeding depression was fairly constant for heading date, but was less consistent for yield. The lack of uniformity for estimates of inbreeding depression can be related to environmental variation and to its influence on type of gene action. The ratio of additive to dominance variance was inconsistent among crosses for heading date and yield. These data suggest selection for these characters should be delayed past the F 2 generation. Broad sense heritabilities for heading date ranged from 42 to 86%. Values obtained for grain yield were more consistent among broad sense than narrow sense estimates. Genetic advance estimates were low due to lack of additive variance. This revised version was published online in July 2006 with corrections to the Cover Date.     

Estimates of heritability for,and relationships between,root and shoot characters of white clover. I. Replicated clonal material

Summary A total of 749 genotypes from a number of white clover (trifolium repens L.) cultivars and populations, many collected from dryland areas, were cloned and grown in field titles. Several morphological characters, including leaf size, number and diameter of large nodal roots (taproots), and proportion of root weight that was classified as taproot were measured. There was large variation between lines and genotypes for all characters measured, and differences between genotypes within lines are also reported. Broad sense heritability estimates were higher (>0.5) for leaflet width, petiole length and stolon diameter, than for all root character heritability estimates which were between 0.2 and 0.4.     

Parent selection for yield and bunch index in the oil palm in West New Britain

Summary Selection criteria of dura female parents and pisifera male parents were investigated by employing a step-wise multiple regression procedure. Independent parental variables entered in the regression equation were (1) yield (Y) and (2) bunch index (BI), i.e. the proportion of total dry matter used for fruit bunches, of the female parents. As secondary characters of both parents were included: (3) leaf-Mg level (LMG), (4) leaf area ratio (LAR), i.e. the ratio of new leaf area produced to new dry matter used for vegetative growth, and (5) sex ratio (SR), i.e. the ratio of female to total inflorescences. The joint contribution of LMG in both parents accounted for 80% of the variance in Y of the offspring. The amount of the total variance in BI of the offspring was substantially explained, in decreasing order, by LMG of the pisifera parents, and Y and LAR of the dura parents. The first two variables explained 70% of the variation in offspring BI. SR and BI appeared to be not significant for explaining Y as well as BI of the offspring. A practical implication is that for pisifera testing programs, pisifera can be efficiently screened on the exclusive basis of LMG. Dura tester parents should be selected for high, and little variable, values of LMG, Y and LAR.     

A genetic analysis of the number of cells,length of cell,and gibberellic acid sensitivity in sugar beet and their relation to bolting mechanism

Summary Three series of full-sib progenies and parental inbred lines or populations were used to study the inheritance of the number of cells and length of cell along the hypocotyl length, and the response to gibberellic acid in young plants of sugar beet. Variance components were determined by using a factorial cross design. There were no significant estimates of non-additive genetic variation for either cellular characters. Both additive and non-additive genetic variance were responsible for the variability of stem length after GA3 treatment. It is concluded that additive type genes predominantly control the number of cells in all series. The length of cell had significant additive variances in most cases, but it was indeed markedly influenced by environmental factors. Number of cells (cell division rate) and GA3 sensitivity in young plants even before vernalization were related to bolting tendency. Bolting-susceptible genotypes generally expressed higher stem length in the response to added GA3 as compared to the bolting-resistant genotypes. However, some genotypes bolted easily in the field but reacted weakly to the GA3 treatment and vise versa. Genotypes that were susceptible to bolting and/or sensitive to GA3 had a specific range (intermediate) of cell number. There were, however, some genotypes containing an intermediate number of cells which demonstrated low bolting. These responded faintly to gibberellin treatment. The results suggest that several physiological requirements have to be fulfilled before bolting can occur, and that other plant characters interfere with the bolting phenomenon.     

Usefulness of the grid system in phenotypic selection for smooth bromegrass fiber concentration

Summary Grid selection was investigated as a means of reducing environmental variation prior to selection for low neutral detergent fiber (NDF) concentration in smooth bromegrass (Bromus inermis Leyss.). Eleven populations of 350 plants each were arranged into 10 grids of 35 plants each. Adjustment of NDF concentration by block means and standard deviations reduced the phenotypic variation to 50 to 94% of the original variance. Adjustment also reduced or eliminated spatial patterns of phenotypic variation in five of eight populations with significant patterns to original NDF values. Selection for low NDF tended to identify plants from blocks with a low mean NDF in some populations. When selection was based on t-scores, the number of plants representing each block was a random variable in each population. Use of grid selection with a t-score adjustment should increase the efficiency of phenotypic recurrent selection for low NDF concentration.     

Comparison of diploid and tetraploid potato families derived from Solanum phureja x dihaploid S. tuberosum hybrids and their vegetatively doubled counterparts

Summary To determine whether in potatoes the tetraploid level is preferable to the diploid level, especially regarding tuber yield, four diploid (2n=2×=24) Solanum phureja x dihaploid S. tuberosum hybrid parents and their vegetatively doubled, tetraploid (2n=4×=48) counterparts were intermated, which resulted in two F₁ hybrid families at both levels of ploidy. The parents and clones of the F₁ families and their offspring were used in crosses in such a way that in addition Sib₁, Sib₂, F₁×Sib₁, BC₁ and Sib₁×Sib₁ families were produced. Of the first clonal generation of the 12 2 x families and their 12 counterpart 4 x families two tubers per clone were planted in three replications in a field experiment at Sturgeon Bay in 1969; of the parents six tubers were planted in each replication. Data were recorded on 16 characters, including plant height at four different times.The ANOVA's showed significant clone effects within each family for all characters. Computed from all family means as well as from the family means per ploidy level, differences due to family were also significant for all characters except one.As at the 2 x level and at the 4 x level of ploidy the mean phenotypic correlations between characters were of similar magnitude, it is concluded that they are independent of ploidy level.With the exception of eye depth, the mean coefficient of variation was greater at the 2 x than at the 4 x level of ploidy, indicating the greatest response to selection for those characters at the 2 x level. From the differences in family means between the 4 x and the 2 x level of ploidy it was apparent that the 4 x families generally had significantly taller plants, later maturity, fewer tubers, higher mean tuber weight, more tuber yield and more dry matter yield than their counterpart 2 x families. In contrast, the 4 x parents had on the average shorter plants, lower mean tuber weight, much lower tuber yield and lower dry matter yield than their 2 x counterparts.The phenotypic correlation and Spearman's rank correlation between the family means of the 2 x and the 4 x level of ploidy were positive for almost all characters and significant for nearly half of them.From the results it is concluded that 1. in potatoes the 4 x level of ploidy is preferable to the 2 x level, and 2. the performance of 4 x families is predictable from the performance of their counterpart 2 x families.Based on results mentioned in the literature and on the present results, a continued use of S. tuberosum dihaploids in potato breeding needs to be dissuaded.     

Intraspecific hybridization in Pinus caribaea var. hondurensis II. Genetic parameters

Genetic parameters of growth and stemstraightness to seven years of age wereestimated from two field populations ofpure Pinus caribaea var. hondurensis provenances and theirintraspecific hybrids, (Mountain Pine Ridge(MPR) × coastal provenances). There were noconsistent differences in the geneticparameter estimates obtained for theparental provenances compared to those fromthe MPR × coastal inter-provenance hybridcrosses. Narrow-sense heritabilities ofgrowth traits (0.06–0.33) andstraightness (0.21–0.50) were within therange of estimates that have beenpreviously reported for this species andother tropical pines. In general, additivevariance was found to be more importantthan dominance variance for growth andstraightness traits. There was no strongevidence to affirm that the superiority ofthe MPR × coastal hybrid could be caused bydominance variance. The low magnitude ofthe type-B genetic correlations within theMPR × coastal hybrid crosses suggested thatthe genotype-by-environment interactionmight be of practical importance inbreeding. For future hybridization of theMPR and coastal provenances, a recurrentselection scheme based on the utilizationof additive genetic variance of both parentpopulations is recommended.     

Estimation of genetic parameters for in vitro traits in wheat immature embryo cultures involving high X low regeneration capacity genotypes

Summary Estimates of gene actions were obtained for five in vitro traits of immature wheat (Triticum aestivum L.) embryo cultures from a cross of two wheat cultivars and the resulting reciprocal, F₁, F₂ and backcross populations. The contribution of additive gene effects to in vitro traits was not as important as the dominance gene effects. Epistatic gene effects were relatively more important than either additive or dominance gene effects. Of the individual types of digenic epistatic effects, the dominance x dominance estimates were relatively larger in magnitude for all in vitro culture traits measured. The maternal effect played a minor role in the inheritance of the in vitro studied traits since the difference among the reciprocal values was not significant. It is shown from the generation mean method that epistasis played a major role in the inheritance of most of the traits under study. The negative values of additive and dominance genetic variance were estimates of zero. Heritability estimates, in broad sense, were relatively high for the in vitro studied traits. In some cases, heritability estimates in broad and narrow senses are almost equal since the estimation of dominance genetic variance led to negative values. According to the results of the gene effects, dominance and epistasis were important for the shoot formation trait. Selection would be effective among the isolated genotypes on individual basis.     

Variation in durum wheat populations from three geographical origins. I. Material and spike characteristics

Summary Sixteen durum wheat landraces. i.e. 6 from Algeria, 5 from Ethiopia and 5 from Italy were analysed in order to ascertain the amount of variation for all 11 spike characters. For each population the differences among spike progenies were significant for all the observed characters. Principal component analysis showed that 4 factors accounted for 80% of the total variation. Component 1 accounted for 32% of the total variation and was associated with spike width, spike density, number of kernels per spikelet and kernel weight. Component 2 accounted for 24% of the total variation and was associated with number of fertile spikelets, total number of spikelets and spike length. In a scatter diagram of all 1127 spike progenies for the first two principal components, the material from Ethiopia and Italy clustered at two opposite sides, while the material from Algeria had the largest swarm and overlapped the other two groups of material. Differences among populations from the same origin were found. The ecological and anthropological causes that may have played a role in the creation of the observed variation in the examined populations, are discussed.