Progeny Means Research Articles

Comparison of resistance toFusarium spp. of glasshouse-and field-grown tuber progenies of potato

Field-grown tubers of 22 progenies ofSolanum tuberosum L. generated in a crossing programme involving seven parents differing in resistance toFusarium coeruleum Lib. ex Sacc. andF. sulphureum Schlect. (=F. sambucinum Fuckel, teleomorphGibberella pulicaris (Fr.) Sacc.) were wound-inoculated with a cornmeal + sand culture of each pathogen. Parental genotypes were also included. The mean lesion size of each progeny was compared in 2 years of tests, as well as with published data on glasshouse-grown tubers. ForF. coeruleum there was a high correlation between years as well as with the glasshouse data, but no such correlations were apparent withG. pulicaris. Furthermore, parental and GCA values, as well as progeny means and mid-parent scores, also correlated highly forF. coeruleum but not forG. pulicaris. Glasshouse-grown tubers of 11 wildSolanum spp. were also inoculated with both pathogens. Some resistance to one or other, or both, was apparent, particularly inS. chacoense.

Estimation of variance components and heritability of bunch yield and yield components in the oil palm (Elaeis guineensis Jacq.)

AbstractIn an experiment involving 24 single‐cross progenies of oil palm conducted at the Nigerian Institute for Oil Palm Research, 9‐year data of fresh‐fruit yield and its two components, annual bunch number and single‐bunch weight, were recorded. Means of progenies of each of the three traits obtained from 1–9‐year records were used in analyses of variance according to a repeated‐measures design. The results were then used to estimate variance components due to progenies and progeny × year interactions. These estimates were then used to estimate heritability for yield and yield components based on progeny means, which were assumed to be calculated from 1–9 years of harvests. The results revealed that heritability estimates of fresh‐fruit yield and its two yield components were not reliable based on the first‐year harvest data. The single‐bunch weight had a slower rate of improvement than the other two traits when more years were used for the estimation. Heritability can be expected to be over 80% for fresh‐fruit yield and bunch number based on 1‐, 3‐, and 4‐year means, respectively. Six‐year data are required for single‐bunch weight to reach the same heritability measure.

Genetic studies of 11 quantitative characters in apricot

A comprehensive programme of apricot genetic improvement was started in 1969 at the INRA Fruit Research Station in Bordeaux. A half-diallel of the first generation, consisting of 15 f 1 progenies and five S 1 progenies, issued from six parents. Progeny means over 3 years were used for a genetic study. General combining ability (GCA) was highly correlated with the parental phenotype ( r = 0.8–1 in ten out of the 11 characters studied) as was progeny mean with midparent value ( r = 0.7–0.9 in nine out of the 11 characters). Heritability, estimated by regressing the progeny mean on midparental phenotype, was high except for fruit overcolour ( h 2, $ ̃ 0.3 ); h 2 ranged from 0.5 to 0.8 for seven characters and was greater than 0.9 for flowering date, maturity date and juiciness. The mean value of S 1 progenies, as a function of the parental phenotype, did not differ from that of F 1 progenies, but one progeny showed a poor yield which was ascribed to an inbreeding effect. The high correlations of GCA with parental phenotype, and of progenies with midparent, and some high or very high heritabilities indicated suitability of choosing parents based on their phenotype, and also the high potential for genetic improvement in the apricot.

Comparison of ancestral and current-generation inbreeding in an experimental strawberry breeding population

Progenies from first-generation self, half-sib, full-sib, and cross fertilizations were generated to evaluate the magnitude of inbreeding depression for vegetative and production traits in strawberry. Tests were conducted to determine the linearity of trait mean depression with inbreeding rate (ΔF) over this range of inbreeding values, as an indication of the presence of non-additive epistasis. A control population, for which a similar range of coancestry had accumulated over several cycles of breeding and selection, was also generated to compare the consequences of ancestral and current-generation inbreeding. Trait means for crosses among current-generation half-sibs, full-sibs, and selfs were 2-17%, 3-12%, and 14-45% lower than for unrelated crosses among the same set of parents, respectively. Linear regression of progeny means on current generation ΔF was significantly negative for all traits and explained 17-44% of the variance among progeny means. Mean depression was largely linear over the range of inbreeding rates tested in this population, indicating the absence of epistasis for the traits evaluated. Conversely, (F) regressions of progeny means on pedigree inbreeding coefficients, where coancestry had accumulated over several cycles of breeding and selection, were uniformly non-significant and explained 0-10% of the variance among cross means. Further, multiple regression of progeny means for current-generation relatives on pedigree F failed to improve fit significantly over regression on current-generation ΔF alone for all traits. Together, these results suggest that pedigree inbreeding coefficients are poor predictors of changes in homozygosity when populations are developed through multiple cycles of breeding and selection. They also imply that inbreeding depression will be of minor importance for strawberry breeding populations managed with adequate population sizes and strong directional selection.

Inheritance of Latent Period Length in Maize Infected withExserohilum turcicum

Extended latent period length is an important component of partial resistance of maize to northern leaf blight (NLB) caused by Exserohilum turcicum. Latent period length is easily measured on seedling plants under both field and greenhouse conditions and has been shown to be a relatively stable trait over a range of temperature and light conditions. The inheritance of latent period length was studied in F 2 and backcross generations of crosses of the experimental inbred line 69-1 (highly resistant, long latent period) and Mo17 (intermediately resistant, intermediate latent period) with the highly susceptible inbred line A632. Studies were conducted under both winter greenhouse conditions and in the field. Differences in mean latent periods between the two parental inbred lines varied from 14.1 days (A632 × 69-1; 1993 field experiment) to 1.8 days (A632 × Mo17; 1993 greenhouse experiment). Analysis of generation means of log-transformed latent periods revealed that over 92% of the variation among generation means could be explained by additive gene action and that dominance and epistatic effects were negligible. Heritability and gene number estimates were in agreement with previously published estimates for partial resistance to NLB measured as reduced disease severity on adult plants. Selection on some sort of progeny mean basis would probably be the most efficient and rapid means of selection for increased latent period and could potentially be more effective than selection for reduced NLB severity after anthesis

Comparative performance of FDR and SDR progenies from reciprocal 4x-2x crosses in potato

Numerically unreduced (2n) gametes from first division restitution (FDR) are considered to be superior to 2n-gametes from second division restitution (SDR) because they transfer a larger proportion of the total parental heterozygosity and epistasis intact to the tetraploid progeny. This supposed superiority was investigated by comparing 12 sets of reciprocal 4x-2x crosses. Each diploid parent used in a reciprocal set produced 2n-pollen by FDR and 2n-eggs by SDR. Six agronomic characters were investigated. FDR progenies (from 4x.2x) were found to have higher mean yields due to more and bigger tubers. With respect to underwater weight, the overall progeny mean of FDR progenies was significantly higher than that of SDR progenies (from 2x.4x). However, the absolute difference found between both overall progeny means was too small to be of practical significance. No differences between FDR and SDR progeny means were found for vine maturity and chip colour. In addition to the progeny mean, within-progeny variation is important in potato breeding. For vine maturity a higher within-progeny variation was detected in SDR progenies, whereas within-progeny variations for yield, underwater weight and chip colour were not different in FDR and SDR progenies. With regard to vine maturity, we conclude that SDR 2n-gametes are superior to FDR 2n-gametes because, with the same progeny means of FDR and SDR progenies, the within-progeny variation was higher in SDR progenies. Therefore the assumed superiority of FDR 2n-gametes was confirmed for yield but was not observed for vine maturity, underwater weight and chip colour.

Genetic Analysis of Japanese Persimmon Fruit Weight

The expected proportion of individuals in progeny having genotypic values for fruit weight over a given selection criterion to the total individuals derived from a cross was estimated by multiple-regression analysis in which inbreeding coefficient (F) and midparental (MP) value were independent variables and progeny mean was the dependent variable in Japanese persimmon (Diospyros kaki Thunb.). A total of 117 seedlings from 39 crosses was used. Genetic differences of progenies among crosses could be explained solely by F and MP, the effect of the former being greater than the latter. The expected proportion of progenies with large fruit decreased as MP decreased and severely decreased as F increased. Based on the parental mean of 35 fruit on a single tree for 3 years, the proportion of individuals in progeny with fruit weight >200 g was estimated as 34%, 21%, and 12% for 0, 0.125, and 0.25 F values, respectively, in individual from a cross with MP = 200 g.

Genetic Variances of Agronomic Traits in Tetraploid Crested Wheatgrass under Competitive Conditions

Limited research has been conducted on characterizing the genetic variability of important agronomic traits of forage grasses grown in seeded swards. This study was conducted to estimate the degree of genetic control, i.e., the beritability (h2), of several agronomic, morphological and physiological traits of crested wheatgrass [Agropyron desertorum (Fischer ex Link) Schultes] under sward conditions. Thirty randomly selected half‐sib families from a genetically broad‐based crested whcatgrass population were evaluated at one location for 2 yr in single‐row plots that were spaced 38 cm apart and seeded at a rate of 90 seed m−1. Heritabilities were determined for the following traits based on the progeny means of the 30 families: fresh forage yield, dry matter yield, tiller height, tiller number, flag leaf width, spikelets per spike, proline content, first‐cut vigor, regrowth vigor, and flag leaf pubescence. Heritability estimates of fresh forage yield were similar, i.e, 80 and 82%, when based on 30 and 300 half‐sib families, respectively, indicating that a sample size of 30 families provided a reliable estimate of genetic variance in this crested wheatgrass population. With the exceptions of tiller number and proline content (h2 = 0%), all traits demonstrated moderate to high heritabilities, ranging from 53 to 86%. Confidence interval widths ranged from 14 to 116% as large as the heritability point estimate. Signilicant family × year interactions were detected for some traits that possessed moderate to high heritabilities. The interaction variance components, however, were significantly less than their corresponding genetic variance for all traits except first‐cut vigor. The results indicate that for most traits, evaluation for multiple years would not be needed.

Ergovaline and peramine production in endophyte-infected tall fescue: Independent regulation and effects of plant and endophyte genotype

Peramine and ergovaline have ecological and economical significance because they possess insect and/or mammalian toxicity properties. The relationship between these endophytically derived alkaloids in tall fescue (Festuca arundinacea Schreb.) is unknown. Seasonal and plant tissue effects on the concentration of peramine and ergovaline was investigated in field and greenhouse experiments. The relationship between the alkaloids and the regulatory effects of endophyte and plant on their content were investigated among progeny of reciprocal crosses between high- and low-ergovaline and peramine plant-endophyte combinations. Variation in peramine concentration ranged from 750 to 1742 µg/kg in greenhouse-grown plants, and there was no seasonal trend in peramine content. There was a correlation (r=0.69) between peramine and ergovaline content among leaf tissues of field-grown plants, but there was no correlation between the alkaloids in the culm (r=0.20) or panicle (r=0.17) tissues. Mean leaf ergovaline concentration of progeny derived from the low-ergovaline parent (163 µg/kg) was less than the midparent value (228 µg/kg), but mean of progeny from the high-ergovaline parent was not different from the mid-parent value. Ranges within each progeny set were approximately double their mean. Mean leaf peramine concentrations of the progeny sets were similar to the mid-parent value (3354 µg/kg) but ranges were from 1716 to 8753 µg/kg. There was no correlation between leaf peramine and ergovaline (r=0.01). These data suggest that endophytically produced alkaloids are independently regulated and are controlled by both plant and endophyte genotype.

Assessing progenies of potato for resistance to early blight

Progenies from crosses between cultivars varying widely in resistance to early blight (Alternaria solani Sor.), were assessed for resistance as true seedlings in a glasshouse in Scotland. The resistance of a representative sample of surviving genotypes from each progeny was compared with samples of the same progenies not previously exposed to the fungus, both in the glasshouse in Scotland and in the field in Israel. The exposed population was more resistant. Resistance was identified more effectively in adult plants from tubers in the glasshouse than in true seedlings and agreement between glasshouse and field assessment was better when progenies were compared rather than individual genotypes. The mid parent and progeny mean scores of the unexposed population were correlated at both sites, thus confirming that the resistance is heritable. Selecting resistant individuals at the seedling stage is suggested as a useful tool for resistance breeding, having first chosen the best parents for crossing.

Inheritance of 'Hullability' in Sunflowers (Helianthus annuus L.)

Improvement of the feed value of sunflower seedmeal by hulling achenes before crushing is of economic importance and varieties with good ‘hullability’ (ease of hulling) would be economically significant. Knowledge of the genetics of this characteristic is therefore necessary. Studies of heritability were made on a factorial cross of six female and six restorer lines. Hullability was measured as the ratio of hull removed by a laboratory huller against the total quantity of hull in the achene. The genotypes studied showed a wide range of hullabilities. There were considerable environmental effects but only small interactions; correlations between results in different locations were significant. Additive genetic effects predominated, although there were some interactions between parental lines. Narrow-sense heritabilities were 0.73–0.85 for genotype means, 0.52 when calculated from individual plants, and 0.36–0.76 from parent-progeny regressions. Early selection based on progeny means or a combination of individual and progeny values would appear to be the best method of breeding for improved hullability of sunflower achenes.

Glasshouse‐based selection to identify agronomically adapted, uniform true potato seed (TPS) progenies

SummaryTwenty‐five true potato seed progenies and five clonally maintained cultivars were grown in a glasshouse. Tubers of each clone of each progeny were grown in a field experiment. At harvest, one seed tuber was taken from each plant and field grown for a second year. After each experiment, the tubers from every plant were weighed and counted and their colour, shape, form (regularity) and skin finish assessed. Progeny means and variances for each character were calculated. There were significant differences between means for each character in all experiments, except mean yield in the glasshouse, variances for each character in both field experiments and tuber number variances and colour variances in the glasshouse. In general, Pearson's product‐moment correlations between glasshouse and field assessments were moderate or poor (r ≤ 0.6) and inconsistent between experiments. Spearman's rank correlations between glasshouse and field assessments for progeny means were similar to the linear correlations, while those for variances were somewhat better. Rank correlation gave more consistent results between experiments. Ranking glasshouse means will allow some worthwhile progeny selection for tuber colour, shape and skin finish. It can be a useful screen for progenies with the highest levels of uniformity for all characters.

Ergovaline Variability in Acremonium‐Infected Tall Fescue Due to Environment and Plant Genotype

Competitiveness and persistence of tall fescue (Festuca arundinacea Schreb.) is superior when infected with the endophyte, Acremonium coenophialum Morgan‐Jones & Gams. Acremonium‐derived ergopeptine alkaloids impair animal productivity, and if reduced or eliminated would improve animal performance. This study was conducted to determine: (i) environmental influence on ergovaline concentration in A. coenophialum‐infected tall fescue genotypes, and (ii) the extent that plant genotype influences ergovaline production by the endophyte. A field experiment evaluated five endophyte‐infected tall fescue genotypes grown at three locations (environments) during 2 yr. Plant genotypes, harvested every 28 d, differed in ergovaline concentration (P < 0.05) during spring months. Rank correlations of genotypes were high among harvest dates, locations, and years. A greenhouse experiment was used to evaluate ergovaline concentration among progeny from a reciprocal cross between a high‐ and a low‐ergovaline tall fescue genotype. Parents differed in ergovaline concentration by 630 μg kg−1 but produced similar F1 progeny means that were not different from the midparent (415 μg kg−1). Both F1 populations had similar ranges for ergovaline (122‐938 μg kg−1). Although subject to environmental influence, ergovaline production by tall fescue—Acremonium associations was partially dependent upon plant genotype.

Heritabilities of Morphological and Agronomic Traits in Western Wheatgrass

Limited research has been directed towards characterizing the phenotypic and genotypic variability of different traits in North American plant species. This study was conducted to estimate the degree of genetic control, i.e., the heritability (h2), of several agronomic and morphological traits of ND-WWG931 western wheatgrass [Pascopyrum smithii, (Rydb.) Love] and to provide insight into appropriate sample sizes needed to estimate genetic parameters. Thirty randomly selected half-sib families of ND-WWG931 western wheatgrass were evaluated over 2 years and 2 locations in seeded single-row plots. Heritabilities were determined for the following traits based on the progeny means of the 30 families: dry matter yield, tiller height, spikelets per spike, vigor, spike density, spike pubescence, and spikelet color. Spike density, dry matter yield, and vigor had relatively high heritabilities (h2 = 79, 72, and 67%, respectively) and were estimated with the greatest precision (90% confidence interval width range: 33 to 64% as large as the point estimate). Spike pubescence, spikelets per spike, tiller height, and spikelet color demonstrated moderate to low heritabilities (h2 = 55, 49, 33, and 0% respectively) and were estimated with the least precision as demonstrated by relatively wide confidence limits. The genetic variance components for spike density, forage yield, vigor, and spike pubescence exceeded twice their standard errors indicating that selection for these traits should be effective in ND-WWG931. Heritability estimates of fresh forage yield were essentially the same, i.e., 61.9 and 61.5%, when based on either 30 or 270 half-sib families, respectively, indicating that a sample size of 30 families was adequate to provide reliable estimates of genetic variance in ND-WWG931. These data provide general insight into the population genetics of a North American plant species and demonstrate an approach to determine the genetic variability within plant materials that are being used for rangeland revegetation.

Divergent selection for carbon isotope discrimination in crested wheatgrass

Because plant growth on semiarid rangelands is frequently water-limited, breeding for enhanced water-use efficiency (WUE, kg dry matter gained per kg water transpired) should improve forage production on these areas. In crested wheatgrass [Agropyron desertorum (Fischer ex Link) Schultes], variation for carbon isotope discrimination (Δ) has been negatively associated with WUE, suggesting that selection for lowered Δ would increase WUE. To determine the potential of altering Δ through breeding, we selected nine clones from a crested wheatgrass breeding population based on their Δ values, equally subdivided them into three groups (low, medium, and high Δ), and made a series of diallel crosses within each group. The parental clones and single-cross progenies were established in the field as replicated spaced plants on 1-m centers in spring 1989. Forage dry matter yield and Δ were determined in 1990 and 1991; leaf gas exchange traits were determined for low and high Δ classes in 1990, and for low, medium, and high Δ classes in 1991. The previous ranking of Δ classes was confirmed in the present studies. Combined across years, values for Δ, leaf CO2 exchange rate (CER), and stomatal conductance (gs) were significantly (P < 0.05) lower in the low than high Δ class. When data were combined across clones and progenies, Δ was negatively correlated with leaf intrinsic water-use efficiency (WUEi, expressed as the CER/gs ratio) in 1990 (r = −0.87**, df = 14) and in 1991 (r = −0.83**, df = 23). Forage yield and Δ were not correlated, suggesting that the two traits may be under separate genetic control. Progenies from crosses among the low-Δ clones had significantly lower Δ values than progenies from either the medium-Δ or high-Δ clones during each year and when the data were combined across years. Divergent selection for low Δ also was reflected by improved WUEi in the subsequent generation. Moreover, narrow-sense heritability values for Δ were in excess of 0.75 and correlations between progeny means for Δ and means of the corresponding midparents were significant in each of the three analyses (r = 0.87** in 1990, 0.91** in 1991, and 0.92** in the combined analysis). These results complement earlier findings and confirm that genetic improvement for WUE can be effectively achieved in crested wheatgrass through indirect selection for Δ. Key words: Agropyron desertorum, water-use efficiency, parent-progeny correlations, drought, carbon isotope ratio, grass breeding, forage yield

AVAILABILITY AND HERITABILITY OF SALT TOLERANCE IN RABBITEYE BLUEBERRY

To assess the variability and heritability of salt tolerance in current rabbiteye blueberry (Vaccinium ashei) germplasm, about 30 plants from each of 15 crosses involving 8 parents were grown in sand culture. All plants were fertilized five days per week with a solution containing essential micro-and macronutrients but having an excess (250 ppm) Na level. After nine weeks, each plant was rated for vigor and chlorosis, dry weights of roots and shoots were recorded, and Na content of the roots and leaves were determined. The regression of progeny means on mid-parent means for leaf Na content produced a regression coefficient of 0.30, an estimate of heritability. There appeared to be little variability for Na content in the leaves of this germplasm so the moderate level of heritability would not be expected to produce much change in leaf Na by selection. Furthermore, there was no correlation between leaf Na content and either plant weight or rating which would facilitate selection of plants with low salt content in the leaves. variability existed for plant response to excess Na (plant rating) but heritability of the plant rating was estimated to be zero.

INHERITANCE OF BACTERIAL CANKER RESISTANCE IN THE TOMATO LINE HAWAII 7998

Clavibacter michiganensis subsp. michiganensis (Smith) Davis et al. incites bacterial canker of tomato and is a significant disease in many tomato growing areas of the world. In previous studies, the accession Hawaii 7998 displayed a high degree of resistance to bacterial canker in North Carolina. Hawaii 7998 and the bacterial-canker-susceptible breeding line NC EBR-3 were crossed to develop reciprocal F1, F2 and BC populations for genetic analysis of bacterial canker resistance. Plants were evaluated for resistance in field and greenhouse studies. The field study indicated no maternal effect. Generation means and variance analysis indicated that foliar resistance was moderate and quantitative. Narrow sense heritability (h2) was estimated as 0.24. By regressing F3 progeny means onto F2 plant values, h2 for foliar resistance was calculated as 0.35. When data for both F2 and F3 generations were transformed into standard units, h2 was 0.44. In greenhouse studies generation means and variance indicated wilt resistance to be quantitative in nature, but h2 was estimated as low (0.09). The F1 generation mean was similar to the resistant parent.

Variance Components and Heritability of Nut and Kernel Defects in Hazelnut

AbstractNut samples of hazelnut (Corylus avellana L.) cultivars, selections, and seedlings were cracked and scored for frequency of good kernels and 7 types of defects. Blanks, poorly filled nuts, and moldy kernels were the most common defects. Variance components were calculated using defect counts from trees of 17 genotypes which were sampled in each of 4 years. The largest components were for among genotypes, genotype × year interaction, and between samples within trees. Year effects were large only for poorly filled nuts. The genotype × year interaction term was large for black tips, shrivels, and blanks. Variance components indicate limited genetic variability for brown stain, but moderate to large amounts for all other defects. Narrow sense heritability estimates from regression of progeny means on mid‐parent values exceeded 50 % for frequency of blanks (51 %), moldy kernels (61 %), kernels with black tips (60 %), and doubles (84 %), but was low for frequency of shrivelled kernels (22 %), poorly filled nuts (25 %), and brown stain (15 %). The low heritability of poorly filled nuts appears to indicate that the frequency of poor fill in the first crop or two produced by a young seedling does not reliably predict the genotype's susceptibility to this disorder rather than a lack of additive genetic variance. Heritability was intermediate for frequency of good kernels (42 %). Phenotypic correlation coefficients indicate that nut weight, kernel percentage, fiber rating, and pellicle removal rating are, for applied breeding purposes, inherited independently of the frequency of the 7 types of defects.

Combining Ability and Heterosis in U.S. Southern Long‐Grain Rice

Combining ability and heterosis estimates are excellent statistics to help discern breeding program goals and direction. Currently, combining ability and heterosis estimates for U.S. southern long‐grain rice (Oryza sativa L.) are lacking, and, thus, our objectives were to estimate these parameters for rice yield, plant height, panicle number, panicle length, panicle weight, and primary branch number. In 1989, two 4 × 4 crossing factorials were completed, and the 32 F1 hybrids and their 16 parents were evaluated in 1990 at two Arkansas locations: the Rice Research and Extension Center near Stuttgart and the Cotton Branch Experiment Station near Marianna. The 16 parents constituted a representative sample of elite U.S. southern long‐grain rice germplasm available to rice breeders. Hill plots consisting of single plants were planted 61 cm apart in a square grid pattern. General combining ability (GCA) was more important than specific combining ability (SCA), accounting for 70% of the variation for yield, 89% for plant height, 84% for panicle number, 60% for panicle length, 69% for panicle weight, and 85% for primary branches. Specific combining ability was significant for all traits except panicle number, indicating that nonadditive gene action was important as well. When crossing combinations are being decided, the sole use of midparent means would not adequately predict progeny mean performance. High parent beterosis for yield ranged from ‐ 16 to 68%. Heterosis for yield was due primarily to beterosis for panicle weight rather than panicle number. U.S. southern long‐grain rice yields could be improved through the use of hybrid rice varieties.

Identification of superior parents in a potato breeding programme

An incomplete diallel cross was used to study components of genetic variation in potatoes for a range of characters after early and late harvest. The progenies were also used to evaluate five predictors of progeny performance, namely the mean seedling performance, the mid-parent value and the means of the selfed progenies, of the diploid progenies and of the test-cross progenies. For almost all characters, the general combining ability effects were predominant, although the specific combining ability effects present were greater at late than at early harvests. The seedling performance for tuber yield, number of tubers and average tuber weight did not show any relevant relationship to the field performance. The midparent value provided, in general, satisfactory predictions of the mean progeny performance obtained in the diallel, except for ware tuber yield. The selfed and the diploid progenies did not improve the prediction of progeny means compared to the mid-parent value. The predictions based on the test-crosses surpassed those of the mid-parent value, particularly for tuber yield at ware potato harvest. Methods to identify superior parents are discussed.