Relationship between seed yield heterosis and molecular marker heterozygosity in soybean

 In soybean [Glycine max (L.) Merr.] heterosis has been reported for seed yield. Molecular markers may be useful to select diverse parents for the expression of heterosis and yield improvement. The objective of this study was to determine if molecular markers could be used to predict yield heterosis in soybean. From each Maturity Group (MG) II and III, 21 genotypes were selected on the basis of high yield (HY), different geographic origin (GO), and isozyme loci (ISO) and for diversity in restriction fragment length polymorphisms (RFLP), and crosses were made within MGs and selection criteria groups to obtain 6 F₁ hybrids per group. The 21 parents and the 24 F₁ hybrids of each MG were evaluated for yield in replicated tests at two locations in 2 years, and midparent heterosis (MPH) and high-parent heterosis (HPH) estimates were calculated. On the basis of hybrid performance during the first year, 12 parents (3 per selection criteria group) were chosen in each MG to conduct a second RFLP analysis using 129 probes. Genetic distances (GD_M) for pairs of the 12 genotypes were calculated with this RFLP information and correlated with MPH and HPH estimates. Significant MPH averages for seed yield were observed in the combined analysis of variance in each of the four selection criteria groups of MG II, and in the HY, ISO, and GO of MG III. Significant HPH averages were observed only in the ISO and GO groups of MG II. The greatest frequency of F₁ hybrids with significant MPH was observed in the ISO and GO groups of both MGs. For HPH, the greatest frequency was observed in the ISO group of both MGs. In both MGs, the ISO group had the largest absolute MPH value; the RFLP group had generally the smallest. The observations indicated that the expression of heterosis in seed yield might be associated with diversity in the isozyme loci present in the parents. For the genotypes included in the second RFLP analysis, correlations of GD_Ms with MPH and HPH values on an entry-mean basis were low and not significant, indicating that heterosis in yield may not be associated with genetic diversity at the molecular level as determined by RFLPs. The results suggest that in soybean, parent selection on the basis of RFLPs and isozyme loci to exploit heterosis in seed yield may not be feasible. There was no association between genetic distance estimated by the RFLP analysis and seed yield heterosis, and in spite of the observed relationship between isozyme loci and heterosis for yield, the practicality of using the isozyme markers to select parents may be limited because of the reduced number of assayable isozyme loci in soybean. Received: 8 March 1996 / Accepted: 21 February 1997     

Molecular marker heterozygosity and hybrid performance in indica and japonica rice

An essential assumption underlying markerbased prediction of hybrid performance is a strong linear correlation between molecular marker heterozygosity and hybrid performance or heterosis. This study was intended to investigate the extent of the correlations between molecular marker heterozygosity and hybrid performance in crosses involving two sets of rice materials, 9 indica and 11 japonica varieties. These materials represent a broad spectrum of the cultivated rice gene pool including landraces, primitive cultivars, historically important cultivars, modern elite cultivars and parents of superior hybrids. Varieties within each set were intermated in all possible nonreciprocal pairs resulting in 36 crosses in the indica set and 55 in the japonica set. The F₁s and their parents, 111 entries in total, were examined for performance of seven traits in a replicated field trial. The parents were surveyed for polymorphisms using 96 RFLP and ten SSR markers selected at regular intervals from a published molecular marker linkage map. Molecular marker genotypes of the F₁ hybrids were deduced from the parental genotypes. The analysis showed that, with very few exceptions, correlations in the indica dataset were higher than in that of their japonica counterparts. Among the seven traits analyzed, plant height showed the highest correlation between heterozygosity and hybrid performance and heteorsis in both indica and japonica datasets. Correlations were low to intermediate between hybrid performance and heterozygosity (both general and specific) in yield and yield component traits in both indica and japonica sets, and also low to intermediate between specific heterozygosity and heterosis in the indica set, whereas very little correlation was detected between heterosis and heterozygosity (either general or specific) in the japonica set. In comparison to the results from our previous studies, we concluded that the relationship between molecular marker heterozygosity and heterosis is variable, depending on the genetic materials used in the study, the diversity of rice germplasms and the complexity of the genetic basis of heterosis.     

Relationship of Parental Genetic Distance with Heterosis and Specific Combining Ability in Sesame (<Emphasis Type="Italic">Sesamum indicum</Emphasis> L.) Based on Phenotypic and Molecular Marker Analysis

The genetic distance analysis for selection of suitable parents has been established and effectively used in many crops; however, there is dearth of conclusive report of relationship of genetic distance analysis with heterosis in sesame. In the present study, an attempt was made to estimate the associations of genetic distances using SSR (GDSSR), seed-storage protein profiling (GDSDS) and agro-morphological traits (GDMOR) with hybrid performance. Seven parents were selected from 60 exotic and Indian genotypes based on genetic distance from clustering pattern based on SSR, seed-storage protein, morphological traits and per se performance. For combining ability analysis, 7 parents and 21 crosses generated from 7 × 7 half diallel evaluated at two environments in a replicated field trial during pre-kharif season of 2013. Compared with the average parents yield (12.57 g plant^?1), eight hybrids had a significant (P < 0.01) yield advantage across environments, with averages of 26.94 and 29.99% for better-parent heterosis (BPH) and mid-parent heterosis (MPH), respectively, across environments. Highly significant positive correlation was observed between specific combining ability (SCA) and per se performance (0.97), while positive non-significant correlation of BPH with GDSSR (0.048), and non-significant negative correlations with GDMOR (? 0.01) and GDSDS (? 0.256) were observed. The linear regressions of SCA on MPH, BPH and per se performance of F₁s were significant with R² value of 0.88, 0.84 and 0.95 respectively. The present findings revealed a weak association of GDSSR with F₁’s performance; however, SCA has appeared as an important factor in the determination of heterosis and per se performance of the hybrids. The present findings also indicated that parental divergence in the intermediate group would likely produce high heterotic crosses in sesame.     

Molecular divergence and hybrid performance in rice

This study was undertaken to determine the relationship between genetic distance of the parents based on molecular markers and F₁ performance in a set of diallel crosses involving eight commonly used parental lines in hybrid rice production. The F₁s and their parents were measured for five traits including heading date, plant height, straw weight, grain yield and biomass. The parental lines were assayed for DNA polymorphisms using two classes of markers: 140 probes for restriction fragment length polymorphisms (RFLPs) and 12 simple sequence repeats (SSRs), resulting in a total of 105 polymorphic markers well spaced along the 12 rice chromosomes. SSRs detected more polymorphism than RFLPs among the eight lines. A cluster analysis based on marker genotypes separated these eight lines into three groups which agree essentially with the available pedigree information. Correlations were mostly low between general heterozygosity based on all the markers and F₁ performance and heterosis. In contrast, very high correlations were detected between midparent heterosis and specific heterozygosity based on the markers that detected significant effects for all the five traits; these correlations may have practical utility in predicting heterosis. The analyses also suggest the existence of two likely heterotic groups in the rice germplasm represented by these eight lines.     

Prediction of hybrid grain yield performances in Indica Rice (<Emphasis Type="Italic">Oryza sativa</Emphasis> L.) with effect-increasing loci

One hundred and fifty-one rice hybrids produced in two sets of half-dialell crosses and their parents (13 cytoplasmic male sterile lines and 19 restorers) were used to predict the F₁ performances of seven yield traits through the parental genetic distances (GD) based on SSR markers. The positive loci (PL) and effect-increasing loci (IL), which were screened from SSR polymorphic loci by the F₁ traits of 32 parents, together with total loci (TL), were utilized to estimate parental GD and the models were found to predict the traits of hybrids derived from different parents, fixed parents, and different environments, respectively. The results were as follows: (1) 550 polymorphic loci were detected from 174 SSR markers: a dendrogram based on these loci could separate all the sterile and restorer lines used in the present study, which indicated that parental genetic diversity of F₁ was large; (2) the correlations between F₁ traits and parental GDs based on IL ranged from 0.61 to 0.87 with a mean of 0.76, and they were higher than those on TL or on PL; (3) predictions based on IL for F₁ traits (except grain weight per plant) derived from different environments were ideal, but worse for F₁ traits derived from different parents; and (4) IL was more effective than TL and PL in predicting traits of F₁ with fixed parents, and predictions for fixed restorer combinations were more effective than those for fixed sterile line combinations. These results should facilitate molecular prediction for hybrid yield and other traits by means of both elite sterile and restorer lines.     

A diallel analysis of heterosis in elite hybrid rice based on RFLPs and microsatellites

Hybrid rice has contributed significantly to the dramatic increase of rice production in the world. Despite this, little attention has been given to studying the genetic basis of heterosis in rice. In this paper, we report a diallel analysis of heterosis using two classes of molecular markers: restriction fragment length polymorphisms, (RFLPs) and microsatellites. Eight lines, which represent a significant portion of hybrid rice germ plasm, were crossed in all possible pairs, and the F₁s were evaluated for yield and yield component traits in a replicated field trial. The parental lines were surveyed for polymorphisms with 117 RFLP probes and ten microsatellites, resulting in a total of 76 polymorphic markers well-spaced in the rice RFLP map. The results indicated that high level heterosis is common among these crosses: more than 100% midparent and 40% better-parent heterosis were observed in many F₁s, including some crosses between maintainer lines. Heterosis was found to be much higher for yield than for yield component traits, which fits a multiplicative model almost perfectly. Between 16 and 30 marker loci (positive markers) detected highly significant effects on yield or its component traits. Heterozygosity was significantly correlated with several attributes of performance and heterosis. Correlations based on positive markers (specific heterozygosity) were large for midparent heterosis of yield and seeds/panicle and also for F₁ kernel weight. These large correlations may have practical utility for predicting heterosis.     

Relationships of differential gene expression in leaves with heterosis and heterozygosity in a rice diallel cross

Using differential display analysis, we assessed the patterns of differential gene expression in hybrids relative to their parents in a diallel cross involving 8 elite rice lines. The analysis revealed several patterns of differential expression including: (1) bands present in one parent and F₁ but absent in the other parent, (2) bands observed in both parents but not in the F₁, (3) bands occurring in only one parent but not in the F₁ or the other parent, and, (4) bands detected only in the F₁ but in neither of the parents. Relationships between differential gene expression and heterosis and marker heterozygosity were evaluated using data for RFLPs, SSRs and a number of agronomic characters. The analysis showed that there was very little correlation between patterns of differential expression and the F₁ means for all six agronomic traits. Differentially expressed fragments that occurred only in one parent but not in the other parent or in F₁ in each of the respective crosses were positively correlated with heterosis and heterozygosity. And conversely, fragments that were detected in F₁s but in neither of the respective parents were negatively correlated with heterosis and heterozygosity. The remaining patterns of differential expression were not correlated with heterosis or heterozygosity. The relationships between the patterns of differential expression and heterosis observed in this study were not consistent with expectations based on dominance or overdominance hypotheses.     

Use of DNA Markers in Prediction of Hybrid Performance and Heterosis for a Three-Line Hybrid System in Rice

Two Cytoplasmic Male Sterile lines were crossed with fourteen restorer lines of rice widely grown in the western regions of Maharashtra, India, to produce 28 F₁ hybrids which were evaluated for eight agronomically important traits, contributing to yield potential, in replicated field trials. The hybrid performance was recorded along with heterosis and heterobeltiosis. All the rice lines under investigation were subjected to marker-based variability analysis. An attempt was made to correlate genetic distance based on specific markers for each trait individually, as well as average genetic distance based on all specific markers, with hybrid performance and heterosis, by regression analysis. Specific markers could cluster the parental lines in different groups and showed significant correlation with hybrid performance. The data also supports the proposition that epistasis is the basis of heterosis. The analysis, however, revealed a lack of significant predictive values for field application.     

Comparison of phenotypic and molecular distances to predict heterosis and F1 performance in Ethiopian mustard (Brassica carinata A. Braun)

Predicting heterosis and F₁ performance from the parental generation could largely enhance the efficiency of breeding hybrid or synthetic cultivars. This study was undertaken to determine the relationship between parental distances estimated from phenotypic traits or molecular markers with heterosis, F₁ performance and general combining ability (GCA) in Ethiopian mustard (Brassica carinata). Nine inbred lines representing seven different geographic regions of Ethiopia were crossed in half-diallel. The nine parents along with their 36 F₁s were evaluated in a replicated field trail at three locations in Ethiopia. Distances among the parents were calculated from 14 phenotypic traits (Euclidean distance, ED) and 182 random amplified polymorphic DNA (RAPD) markers (Jaccard’s distances, JD), and correlated with heterosis, F₁ performance and GCA sum of parents (GCA_sum). The correlation between phenotypic and molecular distances was low (r=0.34, P≤0.05). Parents with low molecular distance also had low phenotypic distance, but parents with high molecular distance had either high, intermediate or low phenotypic distance. Phenotypic distance was highly significantly correlated with mid-parent heterosis (r=0.53), F₁ performance (r=0.61) and GCA (r=0.79) for seed yield. Phenotypic distance was also positively correlated with (1) heterosis, F₁ performance and GCA for plant height and seeds plant⁻¹, (2) heterosis for number of pods plant⁻¹, and (3) F₁ performance for 1,000 seed weight. Molecular distance was correlated with GCA_sum (r=0.36, P≤0.05) but not significantly with heterosis and F₁ performance for seed yield. For each parent a mean distance was calculated by averaging the distances to the eight other parents. Likewise, mean heterosis was estimated by averaging the heterosis obtained when each parent is crossed with the other eight. For seed yield, both mean ED and JD were significantly correlated with GCA (r=0.90, P≤0.01 for ED and r=0.68, P≤0.05 for JD) and mean heterosis (r=0.79, P≤0.05 for ED and r=0.77, P≤0.05 for JD). In conclusion, parental distances estimated from phenotypic traits better predicted heterosis, F₁ performance and GCA than distances estimated from RAPD markers. Electronic Supplementary Material Supplementary material is available for this article at and is accessible for authorized users.     

Gene expression analyses in maize inbreds and hybrids with varying levels of heterosis

Background  

Heterosis is the superior performance of F₁ hybrid progeny relative to the parental phenotypes. Maize exhibits heterosis for a wide range of traits, however the magnitude of heterosis is highly variable depending on the choice of parents and the trait(s) measured. We have used expression profiling to determine whether the level, or types, of non-additive gene expression vary in maize hybrids with different levels of genetic diversity or heterosis.     

Understanding the genetic and molecular constitutions of heterosis for developing hybrid rice

The wide adoption of hybrid rice has greatly increased rice yield in the last several decades. The utilization of heterosis facilitated by male sterility has been a common strategy for hybrid rice development. Here, we summarize our efforts in the genetic and molecular understanding of heterosis and male sterility together with the related progress from other research groups. Analyses of F₁ diallel crosses show that strong heterosis widely exists in hybrids of diverse germplasms, and inter-subspecific hybrids often display higher heterosis. Using the elite hybrid Shanyou 63 as a model, an immortalized F₂ population design is conducted for systematic characterization of the biological mechanism of heterosis, with identification of loci controlling heterosis of yield and yield component traits. Dominance, overdominance, and epistasis all play important roles in the genetic basis of heterosis; quantitative assessment of these components well addressed the three classical genetic hypotheses for heterosis. Environment-sensitive genic male sterility (EGMS) enables the development of two-line hybrids, and long noncoding RNAs often function as regulators of EGMS. Inter-subspecific hybrids show greatly reduced fertility; the identification and molecular characterization of hybrid sterility genes offer strategies for overcoming inter-subspecific hybrid sterility. These developments have significant implications for future hybrid rice improvement using genomic breeding.     

DNA divergence as a criterion of a sweet pepper (Capsicum annuum L.) selection for heterosis

The identification of perspective parental lines for the creation of high-yield hybrids is the most labor-consuming stage of selection, because it needs extensive trials of combining ability. Based on evaluation of the genetic divergence of the parental lines, the prediction accuracy of F₁ hybrids performance was investigated. The value of the divergence was calculated using biometric and molecular methods, such as inter simple sequence repeats (ISSR) and random amplified polymorphic DNA (RAPD). Based on estimates of divergence, 10 lines were chosen for cyclic cross (scheme I) and testcross (scheme II). In most crosses, the F₁ hybrids were significantly superior to the parents in the main economically valuable traits. The level of heterosis was higher among hybrids of scheme I. Analysis of the relationship between parental divergence and F₁ performance showed that the hybrid productivity of scheme I was predetermined by ISSR divergence in 86%, and productivity was caused by RAPD divergence in 69%, whereas the F₁ yield of scheme II was not related to the value of genetic distances. Since the values of DNA divergence were closely associated both with midparent level and F₁ performance, we assumed that part of the polymorphic DNA fragments of the parental lines of scheme I is related to heterotic loci (HTL), which may be considered potential key markers for the heterotic selection of the sweet pepper.     

Genetic diversity and its relationship to hybrid performance and heterosis in rice as revealed by PCR-based markers

Ten elite inbred lines (four japonica, six indica), chosen from those widely used in the hybrid rice breeding program at Human Hybrid Rice Research Center in China, were crossed to produce all possible hybrids excluding reciprocals. The 45 F₁ hybrids along with the ten parents were evaluated for eight traits of agronomic importance, including yield potential, in a replicated field trial. The ten parents were analyzed with 100 arbitrary decamer oligonucleotide primers and 22 microsatellite (simple sequence repeats, SSRs) primer sets via polymerase chain reaction (PCR). Out of the 100 random primers used, 74 were informative and amplified 202 non-redundant bands (variants) with a mean of 2.73 bands per polymorphic primer. All 22 microsatellite primer sets representing 23 loci in the rice genome showed polymorphisms among the ten parents and revealed 90 alleles with an average of 3.91 per SSR locus. Cluster analysis based on Nei's genetic distance calculated from the 291 (202 RAPDs, 89 SSRs) non-redundant variants separated the ten parental lines into two major groups that corresponds to indica and japonica subspecies, which is consistent with the pedigree information. Strong heterosis was observed in hybrids for most of the traits examined. For the 43 diallel crosses (excluding 2 crosses not heading), yield potential, its components (including panicles per plant, spikelets per panicle and 1000-grain weight) and their heterosis in F₁ hybrids showed a significant positive correlation with genetic distance. When separate analyses were performed for the three subsets, yield potential and its heterosis showed significant positive correlations with genetic distance for the 15 indica x indica crosses and the 6 japonica x japonica crosses; however, yield potential and its heterosis were not correlated with genetic distance for the 22 indica x japonica crosses. Results indicated that genetic distance measures based on RAPDs and SSRs may be useful for predicting yield potential and heterosis of intra-subspecific hybrids, but not inter-subspecies hybrids.     

Heterotic patterns in rapeseed (<Emphasis Type="Italic">Brassica napus</Emphasis> L.): I. Crosses between spring and Chinese semi-winter lines

Chinese semi-winter rapeseed is genetically diverse from Canadian and European spring rapeseed. This study was conducted to evaluate the potential of semi-winter rapeseed for spring rapeseed hybrid breeding, to assess the genetic effects involved, and to estimate the correlation of parental genetic distance (GD) with hybrid performance, heterosis, general combining ability (GCA) and specific combining ability (SCA) in crosses between spring and semi-winter rapeseed lines. Four spring male sterile lines from Germany and Canada as testers were crossed with 13 Chinese semi-winter rapeseed lines to develop 52 hybrids, which were evaluated together with their parents and commercial hybrids for seed yield and oil content in three sets of field trials with 8 environments in Canada and Europe. The Chinese parental lines were not adapted to local environmental conditions as demonstrated by poor seed yields per se. However, the hybrids between the Chinese parents and the adapted spring rapeseed lines exhibited high heterosis for seed yield. The average mid-parent heterosis was 15% and ca. 50% of the hybrids were superior to the respective hybrid control across three sets of field trials. Additive gene effects mainly contributed to hybrid performance since the mean squares of GCA were higher as compared to SCA. The correlation between parental GD and hybrid performance and heterosis was found to be low whereas the correlation between GCA_{(f + m)} and hybrid performance was high and significant in each set of field trials, with an average of r = 0.87 for seed yield and r = 0.89 for oil content, indicating that hybrid performance can be predicted by GCA_{(f + m)}. These results demonstrate that Chinese semi-winter rapeseed germplasm has a great potential to increase seed yield in spring rapeseed hybrid breeding programs in Canada and Europe.     

Relationship Between Gene Differential Expression of Leaves in Full Opening Flower Stages of Hybrids & Their Parents and Heterosis in Pest-resistant Cotton

cDNAs of leaves of 24 pest-resistant cotton (authorization) crosses and their parents in full opening flower stages were amplified and different bands were displayed by DDRT-PCR, with 45 primer combinations. The mean percentage of bands duplicated was 70.1%, which implied that there was a high proportion of pseudopositive fragments in the amplified cDNA. These pseudopositive bands can be reduced using duplicate PCR. Correlation analysis between differential gene expression and hybrid performance and heterosis showed that M4 (bands detected in one parent and F₁) was positively correlated with all yield traits, and significantly correlated with boll number; M2 (bands observed in one parent but not in F₁ and another parent) was negatively correlated with heterosis of boll number; and M3 (bands observed only in F₁) was positively correlated with heterosis of lint yield. These results showed that genes that had dominant and specific expression in top leaves of full opening flower stages were beneficial to yield formation and heterosis occurrence.     

Heterosis breeding in rice (Oryza sativa L.)

Summary Studies conducted at the International Rice Research Institute (IRRI) during 1980 and 1981 have shown up to 73% heterosis, 59% heterobeltiosis and 34% standard heterosis for yield in rice. The latter was estimated in comparison to commercial varieties: IR36 and IR42 (yield 4–5 t/ha in wet season trials and 7–8 t/ha in dry season trials). Generally speaking, absolute yield was lower and extent of standard heterosis was higher in wet season than in dry season with some exception. Yields up to 5.9 t/ha (22% standard heterosis) in the wet season and 10.4 t/ha (34% standard heterosis) in the dry season were obtained. Most of the hybrids performed better in some season while some performed better in both seasons. Hybrids showed better lodging resistance although they were 5–10 cm taller. F₁ hybrids had significant positive correlations with the parental traits viz., yield (r = 0.446), tillering (r = 0.746), height (r = 0.810) and flowering (r = 0.843). Selection of parents among elite breeding lines on the basis of their per se yield performance, diverse origin and resistance to insects and diseases should give heterotic combination. Yield advantage of hybrids was due primarily to increase in number of spikelets per unit area even though tiller number was reduced. Grain weight was either the same or slightly higher. High yielding hybrids also showed significant heterosis and heterobeltiosis for total dry matter and harvest index. For commercial utilization of heterosis in rice, effective male sterility and fertility restoration systems are available and up to 45% natural outcrossing on male sterile lines has been observed. Consequently, F₁ rice hybrid have been successfully developed and used in China. Prospects of developing hybrid rice varieties elsewhere appear bright especially in countries that have organized seed production, certification and distribution programs and where hybrid seed can be produced at a reasonable cost.     

Determination of heterotic groups for tropical Indica hybrid rice germplasm

Key message

Two heterotic groups and four heterotic patterns were identified for IRRI hybrid rice germplasm to develop hybrid rice in the tropics based on SSR molecular data and field trials.

Abstract

Information on heterotic groups and patterns is a fundamental prerequisite for hybrid crop breeding; however, no such clear information is available for tropical hybrid rice breeding after more than 30 years of hybrid rice commercialization. Based on a study of genetic diversity using molecular markers, 18 parents representing hybrid rice populations historically developed at the International Rice Research Institute (IRRI) were selected to form diallel crosses of hybrids and were evaluated in tropical environments. Yield, yield heterosis and combining ability were investigated with the main objectives of (1) evaluating the magnitude of yield heterosis among marker-based parental groups, (2) examining the consistency between marker-based group and heterotic performance of hybrids, and (3) identifying foundational hybrid parents in discrete germplasm pools to provide a reference for tropical indica hybrid rice breeding. Significant differences in yield, yield heterosis and combining ability were detected among parents and among hybrids. On average, the hybrids yielded 14.8 % higher than the parents. Results revealed that inter-group hybrids yielded higher, with higher yield heterosis than intra-group hybrids. Four heterotic patterns within two heterotic groups based on current IRRI B- and R-line germplasm were identified. Parents in two marker-based groups were identified with limited breeding value among current IRRI hybrid rice germplasm because of their lowest contribution to heterotic hybrids. Heterotic hybrids are significantly correlated with high-yielding parents. The efficiency of breeding heterotic hybrids could be enhanced using selected parents within identified marker-based heterotic groups. This information is useful for exploiting those widely distributed IRRI hybrid rice parents.     

Heterosis for biomass yield and related traits in five hybrids of Arabidopsis thaliana L. Heynh

Heterosis is of utmost economic importance in plant breeding. However, its underlying molecular causes are still unknown. Given the numerous advantages of Arabidopsis thaliana as a model species in plant genetics and genomics, we assessed the extent of heterosis in this species using five hybrids derived from five ecotypes. Parents, F(1) and F(2), generations in both reciprocal forms were grown in a greenhouse experiment with four replications. Mid-parent heterosis (MPH) and best-parent heterosis (BPH) averaged across hybrids were surprisingly high for biomass yield (MPH: 60.3%; BPH: 32.9%) and rosette diameter (MPH: 49.4%; BPH: 34.8%), but smaller for flowering date (MPH: 27.5%; BPH: 18.5%), seed yield (MPH: 18.9%; BPH: 1.7%), and yield components. Individual hybrids varied considerably in their MPH and BPH values for all traits, one cross displaying 140.1% MPH for biomass yield. MPH was not associated with parental genetic distance determined from molecular markers. Reciprocal effects were significant only in a few cases. With a proper choice of hybrids, our results encourage the use of Arabidopsis as a model species for investigating the molecular causes of heterosis.