Progeny Means Research Articles

The recovery of recombinant inbreds outyielding the hybrid in tomato

Publications commenting on the possibility of fixing heterosis and the possibility to transgress heterosis through partition of crop yield into genetic components have prompted us to reconsider our views on the relative importance of hybrid versus inbred vigour. The three genetic components of crop yield, all assessed in the absence of competition, are: (1) yield per plant assessed by the progeny mean, (2) tolerance to stresses assessed by the progeny standardized mean, and (3) responsiveness to inputs assessed by the progeny standardized selection differential. Although fixation of heterosis using the top commercial tomato hybrids was based only on the first genetic component of crop yield and quality, inbreds outyielding the original hybrid were obtained relatively easily. The results indicated that selection for homozygote superiority on the basis of the three genetic components of crop yield and quality, would lead to the development of tomato inbreds outperforming the hybrids.

Heritability of Heat Tolerance in Winter and Spring Wheat

Cell membranes are the site for many biological activities of the plant and play a key role in heat‐induced damage to wheat (Triticum aestivum L.). This study evaluated the genetic variability of wheat using two assays of heat tolerance, and estimated their heritability by parent‐offspring regression, parent‐offspring correlation, and realized heritability using F3 plants and their F4 progeny means. One assay of heat tolerance was membrane thermal stability (MTS) which measures electrolyte leakage from leaf tissue after exposure to high temperature. Heat injury was also assessed by quantifying the reduction of triphenyl tetrazolium chloride (TTC) to formazan by mitochondrial dehydrogenase respiratory enzymes in heat‐stressed seedlings. Results from the two assays were highly associated (r = 0.62, n = 14, P < 0.05). Parent‐offspring regression and correlation heritability was intermediate to high (0.50–0.65) for TTC and relatively low (0.32–0.38) for MTS. Realized heritability, based on 15% selection intensity, was intermediate to high (0.49–0.64) for TTC and low to intermediate (0.27–0.47) for MTS. The high heritability of TTC warrants good progress from selection in early generations. The relatively lower heritability of MTS suggests the use of multiple replications during selection to limit environmental effects.

Predicting Progeny Means and Variances of Winter Wheat Crosses from Phenotypic Values of Their Parents

The efficiency of breeding programs could be increased by predicting the prospects of crosses for line development before producing and testing lines derived from them. In this study, we examined the performance of F4:n (n = 7 or 8) lines randomly derived from 30 winter wheat (Triticum aestivum L.) crosses produced by a factorial mating of five high yielding with six high baking quality cultivars. Our objectives were to (i) contrast the midparent value m̂ij with ĉij for each cross and apply tests for the presence of epistasis, (ii) compare the estimates of the variance between means of crosses (σ2c) with the average segregation variance within crosses (σ2g), (iii) determine the variation in σ2gij among crosses, (iv) evaluate the use of various parameters for predicting cij, σ2gij, and Uij, and (v) briefly describe quantitative‐genetic theory for interpretation of our experimental findings. Twenty‐two lines per cross and the respective parents were evaluated in four environments for seven agronomic and quality traits. Additionally, 44 F2:4 lines per cross were tested in hills in two environments. On the basis of first‐degree statistics, parental means were good predictors of cross means, yet significant epistatic effects were observed for most traits. In agreement with quantitative‐genetic expectations, the genetic variance between σ2c was of the same size as the variance between parental means and the average σ2g of F4:n lines within crosses for all traits but sedimentation. Here, the two parental groups differed significantly in their mean, and σ2g was significantly larger than σ2c Estimation of the segregation variance (σ2gij) for individual crosses is not recommended because they (i) have a large standard error, (ii) can be assessed reliably only in advanced selfing generations, and (iii) are expected to have a minor influence on differences in the usefulness among crosses. Among all predictors investigated, only the parental means can be recommended for predicting the usefulness of crosses.

Três ciclos de seleção entre e dentro de progênies de meios-irmãos na população de milho CMS-52

Três ciclos de seleção entre e dentro de progênies de meios-irmãos foram praticados na população de milho (Zea mays L.) de alta qualidade protéica CMS-52, nos tabuleiros costeiros dos estados de Sergipe e Bahia, no período de 1995 a 1997, visando à obtenção de uma população melhor adaptada às condições edafoclimáticas da região. As progênies foram avaliadas em látice simples 14 x 14, com recombinação das progênies superiores, dentro do mesmo ano agrícola, de modo a se obter um ciclo por ano. Os valores dos parâmetros genéticos decresceram do ciclo original para o ciclo I, mantendo-se no ciclo II com magnitudes semelhantes ao ciclo I. As altas magnitudes desses parâmetros genéticos, as altas médias de produtividades das progênies, e o ganho médio esperado com a seleção entre e dentro de progênies, por ciclo de seleção (12,3%), mostram o grande potencial da população em responder à seleção, o que permitirá a obtenção de uma população mais produtiva e melhor adaptada às condições edafoclimáticas da região. A magnitude da interação progênies x locais evidenciou a importância de se avaliarem as progênies em mais de um local, para melhorar a eficiência do processo seletivo e obter estimativas mais consistentes dos componentes da variância.

Progeny mean, heterosis and heterobeltiosis inSolanum tuberosum x tuberosum andS. tuberosum x andigena families under a short day sub-tropic environment

Solanum tuberosum x tuberosum (TxT) families were compared withS. tuberosum x andigena (TxA) families as progeny mean, heterosis and heterobeltiosis for ten important agronomic characters by evaluating 72 cross combinations (36 of TxT and 36 of TxA, from 18 common female parents), for three successive seedling and clonal generations under short day sub-tropic conditions. TxA families had more vigorous progenies, higher tuber yield, higher number of tubers, larger tubers and better general impression than TxT families. The TxA families were inferior to the TxT families for characters such as tuber colour and uniformity of tuber colour, but their progeny means were within the acceptable range. High progeny means of TxA families were associated with high heterosis and high heterobeltiosis in these families, although there were a few exceptions. High mean performance for tuber yield and tuber number were associated with high variance of these traits in TxA families. It is concluded that, unlike under long day environments, andigena adapted to short days can be used advantageously for exploiting heterosis in the potato breeding programmes of short day subtropic environments.

Genetic Variation and Selection for Shoot and Rhizome Growth Traits in a Naturalized Quackgrass Population

Quackgrass [Elytrigia repens (L.) Nevski; Agropyron repens (L.) Beanv.; Elymus repens (L.) Gould] is a perennial grass that has colonized much of the world's temperate cropland. Its persistence, perenniality, high nutritional value, and tolerance to several stress factors make it a potentially useful forage crop. This research was undertaken to assess the breeding improvement potential of quackgrass for direct use as a forage crop. Thirty‐five quackgrass parent clones were grown in a series of four greenhouse experiments (growth periods) between November 1981 and March 1984, and in a field experiment from July 1984 to June 1985. A random sample of their polycross progeny was grown in a field experiment from September 1982 to June 1984. Parent means, family means, and progeny means within families varied (P < 0.01) for all traits determined. Greenhouse traits had moderate repeatability (0.50–0.72), high clone × growth period interaction, and low correlation coefficients with field traits (r = −0.50–0.32). Thus, selection should be conducted in the field. There were large amounts of additive genetic variation for all three field traits (forage yield, plant diameter, and maturity). Realized gains from one cycle of selection were between expected gains for disomic and tetrasomic inheritance models. Late maturity was associated with low plant diameter during selection. Forage yield responded to selection, but appeared to be governed primarily by inbreeding because of small effective population sizes. Quackgrass plants that thrive in pastures and croplands still contain hidden alleles for low‐rhizome‐spreading phenotypes.

Genetics of susceptibility to facial eczema in Friesian and Jersey cattle

A total of 528 calves from 3 calf crops, sired by 34 Friesian and 28 Jersey bulls, were challenged with sporidesmin, the toxin causing facial eczema (FE), in order to determine the genetic variation in susceptibility to FE. The challenge was either by grazing toxic pasture or by oral dosing, and the response was monitored by blood sampling to measure the elevations of serum gamma‐glutamyltransferase (GGT), glutamate dehydro‐genase (GDH), and other possible indicators (enzymes or metabolites: bilirubin, aspartate transaminase, ferroxidase, bile acids, and 5'nucleo‐tidase). The last calf crop was sired, in part, by bulls phenotypically selected from the second calf crop using GGT records. Data were analysed using the repeated records option of a restricted maximum likelihood programme. A sire model was used, with sires and maternal grandsires included to determine pedigree. The uni‐variate heritability estimates for logeGGT were 0.29 ± 0.15 in Friesians and 0.77 ± 0.13 in Jerseys, and for logeGDH they were 0.20 ± 0.12 and 0.64 ± 0.13, respectively. Corresponding repeatabilities were 0.86 ± 0.01 and 0.84 ± 0.02 for logeGGT, and 0.66 ± 0.03 and 0.69 ± 0.03 for logeGDH. The genetic correlations between the two enzymes were 0.99 for Friesians and 0.95 for Jerseys. Genetic correlations between the five other enzymes or metabolites and logeGGT were generally high (0.79 to 0.99), with the exception of ferroxidase which was not correlated with logeGGT. The next best indicators of FE susceptibility after GGT and GDH were aspartate transaminase and 5'nucleotidase, which were phenotypically correlated with GGT, were heritable (in Jerseys) and repeatable, and showed elevation in response to challenge. The “High” and “Low” GGT‐selected sire groups in the last calf crop had significantly different progeny mean GGT levels in both breeds. It was concluded that selection against high GGT concentrations following FE challenge should reduce the susceptibility to FE in subsequent generations.

Estimation of Correlation Coefficients between 17 Quantitative Traits in Hazelnut

Thirty-five hazelnut (Corylus avellana L.) progenies and 41 parent genotypes were used to estimate phenotypic and genetic correlation coefficients among 17 nut, kernel, and phenological traits from 1994 to 1996. Correlation coefficients were large and positive for nut size traits (nut length, width, depth, nut weight and kernel weight). Phenological traits (days of opening of female flowers, time of catkin elongation and leafing out) were also highly correlated. But most other traits were uncorrelated with each other. Plots of progeny means and individual seedling values were inspected to verify which data points to have bigger effect on coefficients. These results and their application in the OSU hazelnut breeding program will be discussed.

Identification of superior parents and crosses in potato breeding programmes

The effectiveness of parent per se performance and their self values in the selection of superior parents and crosses in potato breeding programmes was studied by evaluating progenies of 72 crosses from 18×4 (female×male) matings, parents and their selfs for ten agronomically important characters for three successive generations. Simple correlation coefficients were computed between parent per se performance versus general combining ability effects, female per se performance versus progeny means of females, female self values versus progeny means of females, mid-parent values versus progeny means of crosses and mid-self values versus progeny means of crosses. The magnitude of the significant correlation coefficients showed that progeny means of crosses could be moderately predicted by the mid-self values for plant vigour and general impression in clonal generations. Similarly, the progeny means of females in clonal generations could be predicted by their per se performance for general impression. Female self values in the second clonal generation were moderately associated with progeny means of females in that generation for general impression and plant vigour. Parent per se performance and mid-parent values were, in general, ineffective in predicting the general combining ability and the progeny means of the crosses, respectively. The poor predictive powers of parent per se performance and mid-parent values are discussed in relation to the relative importance of specific and general combining ability effects.

Response to divergent selection for early vigour in sunflower (Helianthus annuus L.)

Early vigour (EV), biomass at the first development stages, may be a positive trait to improve sunflower yield in water-limited environments. Field experiments were carried out from 1992 to 1995 to define an index to quantify EV in sunflower plants, evaluate the response to divergent selection for this trait, study EV ×environment interaction, estimate the heritability of EV, and study the association of EV with other traits. A sunflower population with a broad genetic base and 10 sunflower lines derived from it were used. Stem volume calculated from stem diameter and height measured approximately 425 degree-days after emergence was correlated with above-ground plant biomass at the same date. In 1992, divergent selection for high and low stem volume was carried out on this population and plants from every group were recombined independently, obtaining 2 new populations. Selection response, using these and the initial population, was studied in 1993. Ten S2 lines, 5 with high-EV derived from the high-EV population and 5 with low-EV derived from the low EV population, were sown in 1995 to study the EV ×environment interaction. Differences among lines with different EV were significant in all the environments studied. In 1993, a random sample from the initial population was selfed and broad sense heritability and heritability by parent-offspring regression were estimated. Heritability based on variance component estimates of S1 progeny means was 0·61. Realised heritability (± s.e.) from a divergent selection for high and low stem volume were 0·40±0·10 and 0·52±0·15, respectively. Heritability by parent-offspring regression was 0· 40± 0·10. These values and the high variability of stem volume observed in the new populations indicate that additional gain from selection should be possible.

Assessing progenies of potato for resistance to skin spot (Polyscytalum pustulans) in the glasshouse

Crosses were made between six parents differing in susceptibility to skin spot and glasshouse-grown tubers from true seedlings of the nine resulting progenies assessed for resistance to the disease. The assessment was evaluated by a two-year comparison with the behaviour of field-grown tubers of the progenies and their parents. The correlations between progenies in the glasshouse and field were r=0.911 and 0.753 and those between the mid-parental values and the progeny means were high (r>0.864), indicating that parental differences were heritable. It is concluded that correlations were sufficient for glasshouse-grown tubers to be used for assessing large numbers of progenies in genetical studies and for eliminating highly susceptible progenies at an early stage in a breeding programme.

Host selection by winged summer females of the aphid Sitobion avenae

AbstractA total of 55 parthenogenetic lineages of the grain aphid, Sitobion avenae F. were established from single clones collected from winter wheat (agricultural host) and cocksfoot (natural host) at various locations in southern Britain. RAPD‐PCR profiles indicated that these lineages consisted of at least 15 genetically distinct clones. Twenty winged individuals (alatae) of known weight were taken from each lineage and presented with a choice of wheat and cocksfoot hosts (with a total leaf area each of 2 cm2) held in petri dishes (laboratory experiments) and flowerpots (field experiments). In both experimental designs host preferences were determined by a ranking of a proportion of counts (limited to −1 →+1) of the progeny produced on each host after 5 days. Overall, alatae tended to prefer the agricultural host. However, alatae from individual clones found exclusively on wheat generally had a higher preference for wheat than alatae from individual clones found exclusively on cocksfoot. Wheat‐derived lineages (aphid genotypes that were collected on wheat but also present on cocksfoot) showed a significantly greater preference for the agricultural host than the cocksfoot‐derived lineages. Individuals from a wheat‐derived lineage had significantly higher observed and potential progeny production on wheat than they did on cocksfoot, while individuals from a cocksfoot‐derived lineage had significantly higher potential progeny but lower mean progeny weights on wheat. In a second stage, reciprocal host transfer experiments were carried out in the laboratory, i.e. lineages collected from the agricultural host were reared for several generations on the natural host and vice versa prior to being tested for host preference. The preference of the lineages for their host of origin significantly decreased in this second trial, reversing the overall preference trends, while there was little evidence for between‐lineage variation in this change in preference. In summary these results indicate weak genotypic but strong environmental influences on alate host preference in S. avenae. This host plant conditioning effect may serve to promote host‐based genetic structuring observed in southern British populations of S. avenae.

Progeny selection for agronomic characters in early generations of a potato breeding programme

Repeatabilities of progeny means, and the univariate cross prediction method were used to study the effectiveness of progeny selection for agronomically important characters in early generations of potato (Solanum tuberosum L.) breeding. The study was based on 90 progenies (72 crosses+18 selfs) evaluated for three successive generations, i.e. seedling, first clonal and second clonal generations. Repeatabilities of progeny means were measured as correlation coefficients between generations. In the univariate cross prediction method, progeny means and within-progeny standard deviations were used to calculate the proportions of clones exceeding the target values, and correlation coefficients between generations for predicted and observed proportions of clones, were calculated. Population means varied from generation to generation. Correlation coefficients between generations for progeny means for most of the characters were significant, but moderate. These were higher than the correlation coefficients between predicted and observed proportions of clones exceeding the target values. The possibility of using progeny means as a selection parameter to reduce the number of genotypes to be examined in later stages by rejecting the poor crosses in seedling generation is discussed.

Symbiont Regulation and Reducing Ergot Alkaloid Concentration by Breeding Endophyte-Infected Tall Fescue

Ergot alkaloids in endophyte-infected (E+) tall fescue (Festuca arundinacea) are responsible for livestock toxicosis. The host plant is capable of modifying the extent to which endophyte produces ergot alkaloids, possibly via endophyte nutrition by the host. Endophytes vary in alkaloid production. Maintaining the E+ tall fescue is essential for plant survival. Therefore, crop scientists are interested in reducing alkaloid concentration of E+ tall fescue. The objectives of this study were to determine maternal and paternal effects as an estimate of plant and endophyte genotype interactions for ergot alkaloid phenotype and the heritability of the ergot alkaloid trait in divergently selected E+ tall fescue populations. Diallel crosses were conducted among four tall fescue genotypes that varied in alkaloid concentration. Five of six crosses had progeny means less than the parental means when genotypes with lowest alkaloid were the female parents. Four of six crosses had progeny means similar to the parental means of crosses when high alkaloid genotypes were the female parents. Large ranges in progeny suggest an interaction between endophyte and plant affects alkaloid phenotype. In a second experiment, a tall fescue population was screened for alkaloid concentration. Low- and high-alkaloid plants were selected and poly-crossed within each class. Mean alkaloid concentrations of the high- and low-alkaloid progeny were higher or lower than the parental population. The trait was highly heritable and alkaloid concentration was reduced by 86% in the low-alkaloid population after two generations of selection.

A comparison of performance of tetraploid progenies produced by diploid and their vegetatively doubled (tetraploid) counterpart parents

Two diploid potato parents, W5295.7 and W5337.3, and their colchicine-doubled tetraploids were used as male to cross with three cultivars: ‘Raritan’, ‘Shepody’ and ‘Superior’. Both diploids produced 2n gametes via first division restitution (FDR) during meiosis. Field experiments were carried out on seedlings of the 12 hybrid progenies in 1991 and 1992. Six of the progenies represented the tetraploid-diploid (4x-2x) hybrids and the other six tetraploid-tetraploid (4x-4x) progenies. The genetic consequence of FDR gametes was examined by comparing the mean and genotypic variance of six tuber traits of 4x-2x and 4x-4x progenies. Tuber appearance, eye depth, specific gravity and tuber size showed significant differences between the means of 4x-2x and 4x-4x progenies from either both or one of the diploid parents. The genotypic variances of these traits in 4x-2x progenies were reduced in various degrees from those observed in 4x-4x progenies. Major genes responsible for the inheritance of these traits are probably located close to the centromere. The two diploid parents may also carry alleles of different size and direction of non-additive genic effects in their 2n gametes. The 4x-2x progenies also outperformed the 4x-4x progenies with respect to the means of total and marketable yield. There were thus heterotic gene loci for high yield located close to the centromere. No marked reduction of genotypic variance, however, was observed. The results suggest that genes for yield would be scattered between the centromere and the site for maximum recombination.

Genetic divergence for cross prediction in potato

To study the effectiveness of genetic divergence for cross prediction in potato, progeny means, heterosis and specific combining ability effects were correlated with parents’ genetic distances (D values) estimated under six in vitro and four in vivo conditions, for tuber yield in 72 crosses (18 × 4) of 22 parents under autumn crop conditions for three successive generations. Genetic distances under in vitro conditions had no relationship with the progeny means for tuber yield. Whereas, those under in vivo conditions in the autumn seasons were positively associated with the progeny means. Similarly, heterosis for tuber yield had a stronger relationship with genetic distances based on an in vivo crop than those based on an in vitro crop. All correlation coefficients between genetic distances and specific combining ability effects were non-significant. The magnitudes of the significant correlation coefficients showed that genetic divergence can be used as an indirect parameter of moderate effectiveness in selecting parents to produce heterotic high yielding progenies. Such cross predictions, however, would be effective only if parents are evaluated under the conditions similar to those under which crosses are likely to be evaluated.

Screening ofPinus radiataProgenies for Resistance toSphaeropsis sapinea

SYNOPSIS The relative susceptibilities of half-sib progenies of Pinus radiata obtained from seed parents selected for apparent resistance to the pine pathogen Sphaeropsis sapinea in plantations in the Western Cape and Chile were assessed in two separate trials. In each trial, artificial inoculations with the pathogen were performed on the stems of 20 one-year-old potted seedlings. Cambial lesions which resulted on stems were measured 14 days after inoculation and the mean lesion length for each selection was evaluated as a trait indicative of the degree of field resistance of the respective parents. Substantial differences in mean lesion lengths were observed among the different progenies in both trials. There were significant differences (P<0,05) between minimum and maximum progeny means as well as among intermediate progeny means for each of the two trials. The results of this preliminary study suggest that indirect selection of P. radiata for resistance to S. sapinea from within a general breeding popu...

Postharvest Longevity of Cut-flower Gerbera. I. Response to Selection for Vase Life Components

A broad source of Gerbera × hybrida Hort. germplasm was evaluated for vase life. Senescence mode, i.e., bending or folding of stems or wilting of ligulae was also recorded for flowers evaluated. Intensive selection was practiced to improve vase life. About 10% of the plants from a sample population were selected for having flowers with high vase life. Progeny means for vase life resulting from a topcross between these plants and `Appleblossom' were used to select five plants (about 1.5% of the sample population) whose flowers had high vase life. A diallel cross using these five plants as parents resulted in a progeny population with an increase in mean vase life of 3.4 days compared to mean vase life for the initial sample population. Increases in vase life means for days to bending, folding, and wilting were 0.3, 3.5, and 1.2 days, respectively. Plants with flowers which senesced due to wilting had the longest mean vase life before and after breeding. Changes in proportion of senescence modes were observed; bending decreased, folding and wilting increased. Frequencies of bending, folding, and wilting were compared to vase life means for 10 progenies. Proportion of bending generally decreased as vase life increased.

Analysis of parental effects on mean vine maturity and chip colour of 4x.2x potato progenies

The parental effects on progenies from potato 4x.2x(FDR) crosses were analyzed to determine anether and to what extent selection criteria used for selecting tetraploid breeding lines have to be adjusted when selecting diploid breeding lines. For vine maturity and chip colour multiple regression analysis of progeny means on means of diploid and tetraploid parents revealed high coefficients of determination (R2) for both characters, indicating good predicting power of the performance of both diploid and tetraploid parents on the performance of their 4x.2x progenies. For vine maturity the multiple regression slope for the tetraploid perents was significantly larger than for the diploid parents, and progeny means were towards the tetraploid parents. This indicates that selection criteria for vine maturity may be less stringent at the diploid level. However, diploids may be earlier maturing than tetraploids and therefore selection criteria can better be similar at both ploidy levels. For chip colour, multiple regression slopes for diploid and tetraploid parents were similar and progeny means were towards the darkest coloured parent irrespective of its ploidy level. This indicates that selection of breeding lines for chip colour can be conducted similarly at both ploidy levels.

Inheritance of pseudo-self compatibility in self-incompatible garden and greenhouse chrysanthemums, Dendranthema grandiflora Tzvelv

Self incompatibility (SI) can be used to alleviate costly hand emasculation and pollination in F1 hybrid chrysanthemum seed production. SI, however, disrupts the progression of inbreedig (selfing or full-sib mating). Consequently, inbreds are selected for breakdown of the SI system or the presence of pseudo-self compatibility (PSC). PSC inbreds, recombinant inbreds, and noninbred cultivars were selfed and/or intercrossed to determine PSC expression across environments and generate 1–3 inbred generations (I1–I3). Percent PSC ranged from 0–68.8% for inbred parents, 0.2–99.7% for recombinant inbreds, and 0.6–25.7% for noninbred cultivars. There was no indication of ‘end-of-season’ PSC. The majority of parents (78%) were classified as low PSC and this trend continued in the I1 (70.1%), I2 (65.6%), and I3 (83.6%) generations; mid PSC was rarer (11.9–18.8%) and high PSC the least common (4.5–15.6%). PSC distributions were primarily continuous, rather than discrete, indicating quantitative inheritance. In several inbred families, 100% of the I1 individuals were SI; this was not correlated with parental PSC level. Inbred families derived from selfing low and mid PSC parents were the most likely to reach extinction due to inbreeding depression. High PSC was not highly heritable, since I1 progeny were predominantly SI or had low to mid PSC levels. Most 153-1 values for PSC:SI segregations (1:1, 1:3, 3:1) were not significant at the 5% level. Realized heritability (H R ) estimates for PSC ranged from a low of 0.05% to 10.19%, although increased H R values did not account for inbreeding depression or genetic mechanisms preventing selection for high PSC. The highest individual %PSC increased over the high parent with SI × low PSC or SI × mid PSC parents in all cases; progeny means did so only in SI × low PSC. Low × low crosses were split evenly between an increase and decrease in progeny or highest individual mean. Since all low × low crosses and low selfs produced 43–50% of the progeny with PSC levels higher than the parents, it appears that most low parents possess some unexpressed PSC genes. Mid PSC parents responded similarly to low PSC genotypes. An increase in PSC was found when crossing SI × PSC parents, illustrating a ‘threshold effect’. PSC × PSC crosses (high × low, low × low) produced SI, low, and mid PSC I1 individuals but did not show a heterotic effect, since the PSC parents had already surpassed the PSC threshold. Selfing high PSC parents, however, suggested different genetic control. Progeny and highest individual means behaved the same; 83% decreased and 17% remained the same for PSC levels. The drop in PSC indicated non-additive gene action. Thus, the PSC threshold with additive gene action holds when selection for higher PSC levels is being done from low and mid PSC parents, but once high PSC levels are obtained, non-additive gene action prevails.