RFLP loci associated with soybean seed protein and oil content across populations and locations

Molecular markers provide the opportunity to identify marker-quantitative trait locus (QTL) associations in different environments and populations. Two soybean [Glycine max (L.) Merr.] populations, Young x PI 416 937 and PI 97100 x Coker 237, were evaluated with restriction fragment length polymorphism (RFLP) markers to identify additional QTLs related to seed protein and oil. For the Young x PI 416937 population, 120 F₄-derived lines were secored for segregation at 155 RFLP loci. The F₄-derived lines and two parents were grown at Plains, G.a., and Windblow and Plymouth, N.C. in 1994, and evaluated for seed protein and oil. For the PI 97100 x Coker 237 population, 111 F₂-derived lines were evaluated for segregation at 153 RFLP loci. Phenotypic data for seed protein and oil were obtained in two different locations (Athens, G.a., and Blackville, S.C.) in 1994. Based on single-factor analysis of variance (ANOVA) for the Young x PI 416937 population, five of seven independent markers associated with seed protein, and all four independent markers associated with seed oil in the combined analysis over locations were detected at all three locations. For the PI 97 100 x Coker 237 population, both single-factor ANOVA and interval mapping were used to detect QTLs. Using single-factor ANOVA, three of four independent markers for seed protein and two of three independent markers for seed oil were detected at both locations. In both populations, singlefactor ANOVA, revealed the consistency of QTLs across locations, which might be due to the high heritability and the relatively few QTLs with large effects conditioning these traits. However, interval mapping of the PI 97100 x Coker 237 population indicated that QTLs identified at Athens for seed protein and oil were different from those at Blackville. This might result from the power of QTL mapping being dependent on the level of saturation of the genetic map. Increased seed protein was associated with decreased seed oil in the PI 97100 x Coker 237 population (r = –0.61). There were various common markers (P0.05) on linkage groups (LG) E, G,H,K, and UNK2 identified for both seed protein and oil. One QTL on LG E was associated with seed protein in both populations. The other QTLs for protein and oil were population specific.     

Mapping the genome of rapeseed (Brassica napus L.). I. Construction of an RFLP linkage map and localization of QTLs for seed glucosinolate content

A linkage map of the rapeseed genome comprising 204 RFLP markers, 2 RAPD markers, and 1 phenotypic marker was constructed using a F₁ derived doubled haploid population obtained from a cross between the winter rapeseed varieties Mansholt's Hamburger Raps and Samourai. The mapped markers were distributed on 19 linkage groups covering 1441 cM. About 43% of these markers proved to be of dominant nature; 36% of the mapped marker loci were duplicated, and conserved linkage arrangements indicated duplicated regions in the rapeseed genome. Deviation from Mendelian segregation ratios was observed for 27.8% of the markers. Most of these markers were clustered in 7 large blocks on 7 linkage groups, indicating an equal number of effective factors responsible for the skewed segregations. Using cDNA probes for the genes of acyl-carrier-protein (ACP) and -ketoacyl-ACP-synthase I (KASI) we were able to map three and two loci, respectively, for these genes. The linkage map was used to localize QTLs for seed glucosinolate content by interval mapping. Four QTLs could be mapped on four linkage groups, giving a minimum number of factors involved in the genetic control of this trait. The estimated effects of the mapped QTLs explain about 74% of the difference between both parental lines and about 61.7 % of the phenotypic variance observed in the doubled haploid mapping population.     

Molecular-marker-facilitated studies of morphological traits in maize. II: Determination of QTLs for grain yield and yield components

Genetic factors controlling quantitative inheritance of grain yield and its components have not previously been investigated by using replicated lines of an elite maize (Zea mays L.) population. The present study was conducted to identify quantitative trait loci (QTLs) associated with grain yield and grain-yield components by using restriction fragment length polymorphism (RFLP) markers. A population of 150 random F₂₃ lines was derived from the single cross of inbreds Mo17 and H99, which are considered to belong to the Lancaster heterotic group. Trait values were measured in a replicated trial near Ames, Iowa, in 1989. QTLs were located on a linkage map constructed with one morphological and 103 RFLP loci. QTLs were found for grain yield and all yield components. Partial dominance to overdominance was the primary mode of gene action. Only one QTL, accounting for 35% of the phenotypic variation, was identified for grain yield. Two to six QTLs were identified for the other traits. Several regions with pleiotropic or linked effects on several of the yield components were detected.     

RFLP mapping of quantitative trait loci controlling seed aliphatic-glucosinolate content in oilseed rape (Brassica napus L)

We report the RFLP mapping of quantitative trait loci (QTLs) which regulate the total seed aliphaticglucosinolate content in Brassica napus L. A population of 99 F₁-derived doubled-haploid (DH) recombinant lines from a cross between the cultivars Stellar (low-glucosinolate) and Major (high-glucosinolate) was used for singlemarker analysis and the interval mapping of QTLs associated with total seed glucosinolates. Two major loci, GSL-1 and GSL-2, with the largest influence on total seed aliphatic-glucosinolates, were mapped onto LG 20 and LG 1, respectively. Three loci with smaller effects, GSL-3, GSL-4 and GSL-5, were tentatively mapped to LG 18, LG 4 and LG 13, respectively. The QTLs acted in an additive manner and accounted for 71 % of the variation in total seed glucosinolates, with GSL-1 and GSL-2 accounting for 33% and 17%, respectively. The recombinant population had aliphatic-glucosinolate levels of between 6 and 160 moles per g^-1 dry wt of seed. Transgressive segregation for high seed glucosinolate content was apparent in 25 individuals. These phenotypes possessed Stellar alleles at GSL-3 and Major alleles at the four other GSL loci demonstrating that low-glucosinolate genotypes (i.e. Stellar) may possess alleles for high glucosinolates which are only expressed in particular genetic backgrounds. Gsl-elong and Gsl-alk, loci which regulate the ratio of individual aliphatic glucosinolates, were also mapped. Gsl-elong-1 and Gsl-elong-2, which control elongation of the -amino-acid precursors, mapped to LG 18 and LG 20 and were coincident with GSL loci which regulate total seed aliphatic glucosinolates. A third tentative QTL, which regulates side-chain elongation, was tentatively mapped to LG 12. Gsl-alk, which regulates H₃CS-removal and side-chain de-saturation, mapped to LG 20.     

Quantitative trait loci influencing protein and starch concentration in the Illinois Long Term Selection maize strains

A study was initiated to determine the number, chromosomal location, and magnitude of effect of QTL (quantitative trait loci or locus depending on context) controlling protein and starch concentration in the maize (Zea mays L.) kernel. Restriction fragment length polymorphism (RFLP) analysis was performed on 100 F₃ families derived from a cross of two strains, Illinois High Protein (IHP), X Illinois Low Protein (ILP), which had been divergently selected for protein concentration for 76 generations as part of the Illinois Long Term Selection Experiment. These families were analyzed for kernel protein and starch in replicated field trials during 1990 and 1991. A series of 90 genomic and cDNA clones distributed throughout the maize genome were chosen for their ability to detect RFLP between IHP and ILP. These clones were hybridized with DNA extracted from the 100 F₃ families, revealing 100 polymorphic loci. Single factor analysis of variance revealed significant QTL associations of many loci with both protein and starch concentration (P < 0.05 level). Twenty-two loci distributed on 10 chromosome arms were significantly associated with protein concentration, 19 loci on 9 chromosome arms were significantly associated with starch concentration. Sixteen of these loci were significant for both protein and starch concentration. Clusters of 3 or more significant loci were detected on chromosome arms 3L, 5S, and 7L for protein concentration, suggesting the presence of QTL with large effects at these locations. A QTL with large additive effects on protein and starch concentration was detected on chromosome arm 3L. RFLP alleles at this QTL were found to be linked with RFLP alleles at the Shrunken-2 (Sh2) locus, a structural gene encoding the major subunit of the starch synthetic enzyme ADP-glucose pyrophosphorylase. A multiple linear regression model consisting of 6 significant RFLP loci on different chromosomes explained over 64 % of the total variation for kernel protein concentration. Similar results were detected for starch concentration. Thus, several chromosomal regions with large effects may be responsible for a significant portion of the changes in kernel protein and starch concentration in the Illinois Long Term Selection Experiment.     

Restriction fragment length polymorphism analysis of loci associated with disease resistance genes and developmental traits in Pisum sativum L.

An F₂ population of pea (Pisum sativum L.) consisting of 174 plants was analysed by restriction fragment length polymorphism (RFLP) and random amplified polymorphic DNA (RAPD) techniques. Ascochyta pisi race C resistance, plant height, flowering earliness and number of nodes were measured in order to map the genes responsible for their variation. We have constructed a partial linkage map including 3 morphological character genes, 4 disease resistance genes, 56 RFLP loci, 4 microsatellite loci and 2 RAPD loci. Molecular markers linked to each resistance gene were found: Fusarium wilt (6 cM from Fw), powdery mildew (11 cM from er) and pea common Mosaic virus (15 cM from mo). QTLs (quantitative traits loci) for Ascochyta pisi race C resistance were mapped, with most of the variation explained by only three chromosomal regions. The QTL with the largest effect, on chromosome 4, was also mapped using a qualitative, Mendelian approach. Another QTL displayed a transgressive segregation, i.e. the parental line that was susceptible to Ascochyta blight had a resistance allele at this QTL. Analysis of correlations between developmental traits in terms of QTL effects and positions suggested a common genetic control of the number of nodes and earliness, and a loose relationship between these traits and height.     

Restriction fragment length polymorphism diversity in soybean

Summary Fifty-eight soybean accessions from the genus Glycine, subgenus Soja, were surveyed with 17 restriction fragment length polymorphism (RFLP) genetic markers to assess the level of molecular diversity and to evaluate the usefulness of previously identified RFLP markers. In general, only low levels of molecular diversity were observed: 2 of the 17 markers exhibited three alleles per locus, whereas all others had only two alleles. Thirty-five percent of the markers had rare alleles present in only 1 or 2 of the 58 accessions. Molecular diversity was least among cultivated soybeans and greatest between accessions of different soybean species such as Glycine max (L.) Merr. and G. soja Sieb. and Zucc. Principal component analysis was useful in reducing the multidimensional genotype data set and identifying genetic relationships.     

An unusual polymorphic locus useful for tagging Rps1 resistance alleles in soybean

The Phytophthora root and stem resistance locus Rps1 has been mapped to linkage group N of the USDA-ARS soybean molecular map, approximately 2 cM from locus A071-1. To determine if A071-1 polymorphisms exist that distinguish and tag different Rps1 alleles, germplasms containing the seven Rps1 alleles were screened with eight enzymes for pA071-detectable polymorphisms. Six enzymes revealed at least one polymorphic fragment. All six detected a polymorphism at A071-1 as determined by restriction fragment length polymorphism mapping, comparison to an EMBL3 clone containing locus A071-1, and Southern hybridization with probes specific for locus A071-1. Screening of the Rps1 donors and 24 Rps1-and 15 Rps1-containing U.S. soybean varieties showed that locus A071-1 exhibited three polymorphisms with each enzyme. The polymorphisms detected by one enyme did not always correlate with those detected by the other four, suggesting that multiple mutation events may be responsible for the different A071-1 polymorphisms. Although no combination of alleles distinguished Rps1-and Rps1-containing genotypes, polymorphism at A071-1 made it possible to distinguish five groups of soybean germplasms. Thus, the unusual polymorphism of locus A071-1 should useful for following Rps1 inheritance in many breeding programs.Joint contribution of North Central Region, USDA-ARS and Journal Paper no. 15383 of the Iowa Agricultural and Home Economics Experiment Station, Ames, IA 50011. Project no. 2974     

Identification of the double genome donor in spontaneous triploid tomato plants by RFLP analysis

Independent spontaneous triploid tomato plants (Lycopersicon esculentum Mill.) were collected among diploid hybrids growing in commercial greenhouses. Ploidy levels were verified by counting chromosomes, and the donor of the double genome dose was determined by restriction fragment length polymorphism (RFLP) analysis. The TG101 probe, which is tightly linked to the Tm-2 ^a locus, revealed different restriction patterns between TMV-resistant and TMV-susceptible parent lines. The parent donor which provided two genomes to the triploid was identified by comparing the relative intensity of alleles in the triploid with that in the diploid. The results indicate that both parents can serve as a double genome donor.     

RFLP mapping of major and minor genes for important agronomic characters in rice

A molecular genetic map with 233 RFLP markers which covered about 2070 cM of rice genome was constructed based on a doubled haploid (DH) population derived from anther culture of a cross between an indica variety Gui630 and a japonica variety 02428. Quantitative trait loci (QTLs) for agronomic characters such as number of panides, heading date, plant height, number of spikelets, number of grains, fertility and 1 000-grain weight were analyzed using interval mapping approach. 8 major genes and 29 minor genes were identified associating with these traits. The results also indicated that great phenotypic difference between parents was profitable in detection of major genes.     

Genetic analysis of morphological variation in Brassica oleracea using molecular markers

A cross between the open-pollinated Brassica oleracea cabbage cultivar Wisconsin Golden Acre and the hybrid broccoli cultivar Packman was used with molecular markers to investigate the genetic control of morphological variation. Twenty-two traits derived from leaf, stem, and flowering measurements were analyzed in 90 F₂ individuals that were also classified for genotype by restriction fragment length polymorphism (RFLP) markers. Seventy-two RFLP loci, which covered the mapped genome at an average of 10 map-unit intervals on all nine linkage groups, were tested individually for associations to phenotypic measurements by single factor ANOVA, and markers with significant associations (P<0.05) were used to develop multilocus models. These data were utilized to describe the location, parental contribution of alleles, magnitude of effect, and the gene action of trait loci. Single marker loci that were significantly associated (P<0.05) with trait measurements accounted for 6.7–42.7% of the phenotypic variation. Multilocus models described as much as 60.1% of the phenotypic variation for a given trait. In some cases, different related traits had common marker-locus associations with similar gene action and genotypic class ranking. The numbers, action, and linkages, of genes controlling traits estimated with marker loci in this population corresponded to estimates based on classical genetic methods from other studies using similar, or similarly-wide, crosses. There was no evidence that genome duplication accounted for a significant portion of multiple genes controlling trait loci over the entire genome, but possible duplications of trait loci were identified for two regions with linked, duplicated marker loci.     

Molecular marker analysis of Helianthus annuus L. 2. Construction of an RFLP linkage map for cultivated sunflower

A detailed linkage map of Helianthus annuus was constructed based on segregation at 234 RFLP loci, detected by 213 probes, in an F₂ population of 289 individuals (derived from a cross between the inbred lines HA89 and ZENB8). The genetic markers covered 1380 centiMorgans (cM) of the sunflower genome and were aranged in 17 linkage groups, corresponding to the haploid number of chromosomes in this species. One locus was found to be unlinked. Although the average interval size was 5.9 cM, there were a number of regions larger than 20 cM that were devoid of markers. Genotypic classes at 23 loci deviated significantly from the expected ratios (121 or 31), all showing a reduction in the ZENB8 homozygous class. The majority of these loci were found to map to four regions on linkage groups G, L and P.     

RFLP mapping of QTLs conferring cold tolerance during seed germination in an interspecific cross of tomato

Most cultivars of tomato, Lycopersicon esculentum, are sensitive to low (chilling) temperatures (0–15 °C) during seed germination; however, genetic sources of cold (chilling) tolerance have been identified within the related wild species. The purpose of this study was to identify quantitative trait loci (QTLs) that contribute to cold tolerance during germination in tomato using a backcross population of an interspecific cross between a cold-sensitive tomato line (NC84173, recurrent parent) and a L. pimpinellifolium accession (LA722) that germinates rapidly under low temperatures. A total of 119 BC1 individuals were genotyped for 151 restriction fragment length polymorphism (RFLP) markers and a genetic linkage map was constructed. The parental lines and 119 BC1S1 families (self-pollinated progeny of the BC1 individuals) were evaluated for germination at a low temperature (11±0.5 °C). Germination was scored visually as radicle protrusion at 8 h intervals for 28 consecutive days. Germination response was analyzed by the survival analysis and the times to 25, 50 and 75% germination were calculated. In addition, a germination index (GI) was calculated as the weighted mean of the time from imbibition to germination for each family/line. Two QTL mapping techniques, interval mapping (using MAPMAKER/QTL) and single-point analysis (using QGENE), were used to identify QTLs. The results of both methods were similar and two chromosomal locations (3–5 putative QTLs) with significant effects on low temperature germination were identified. The L. pimpinellifolium accession had favorable QTL alleles on chromosomes 1 and NC84173 had favorable QTL alleles on chromosome 4. The percentage of phenotypic variation explained (PVE) by individual QTLs ranged from 11.9% to 33.4%. Multilocus analysis indicated that the cumulative action of all significant QTLs accounted for 43.8% of the total phenotypic variance. Digenic epistatic interactions were evident between two of the QTL-linked markers and two unlinked markers. Transgressive phenotypes were observed in the direction of cold sensitivity. The results indicate that low temperature germination of tomato seed can be improved by marker-assisted selection.     

RFLP mapping of QTLs conferring salt tolerance during germination in an interspecific cross of tomato

 Most cultivars of tomato (Lycopersicon esculentum) are sensitive to salinity during seed germination and at later stages. Genetic resources for salt tolerance have been identified within the related wild species of tomato. The purpose of the present study was to identify quantitative trait loci (QTLs) for salt tolerance during germination in an inbred backcross (BC₁S₁) population of an interspecific cross between a salt-sensitive tomato breeding line (NC84173, maternal and recurrent parent) and a salt-tolerant Lycopersicon pimpinellifolium accession (LA722). Onehundred and nineteen BC₁ individuals were genotyped for 151 restriction fragment length polymorphism (RFLP) markers and a genetic linkage map was constructed. The parental lines and 119 BC₁S₁ families (self-pollinated progeny of 119 BC₁ individuals) were evaluated for germination at an intermediate salt-stress level (150 mM NaCl+15 mM CaCl₂, water potential approximately −850 kPa). Germination was scored visually as radicle protrusion at 8-h intervals for 28 consecutive days. Germination response was analyzed by survival analysis and the time to 25, 50, and 75% germination was determined. In addition, a germination index (GI) was calculated as the weighted mean of the time from imbibition to germination for each family/line. Interval mapping, single-marker analysis and distributional extreme analysis, were used to identify QTLs and the results of all three mapping methods were generally similar. Seven chromosomal locations with significant effects on salt tolerance were identified. The L. pimpinellifolium accession had favorable QTL alleles at six locations. The percentage of phenotypic variation explained (PVE) by individual QTLs ranged from 6.5 to 15.6%. Multilocus analysis indicated that the cumulative action of all significant QTLs accounted for 44.5% of the total phenotypic variance. A total of 12 pairwise epistatic interactions were identified, including four between QTL-linked and QTL-unlinked regions and eight between QTL-unlinked regions. Transgressive phenotypes were observed in the direction of salt sensitivity. The graphical genotyping indicated a high correspondence between the phenotypes of the extreme families and their QTL genotypes. The results indicate that tomato salt tolerance during germination can be improved by marker-assisted selection using interspecific variation. Received: 29 January 1998 / Accepted: 4 June 1998     

Molecular markers associated with seed weight in two soybean populations

Seed weight (SW) is a component of soybean, Glycine max (L.) Merr., seed yield, as well as an important trait for food-type soybeans. Two soybean populations, 120 F₄-derived lines of YoungxPI416937 (Pop1) and 111 F₂-derived lines of PI97100xCoker 237 (Pop2), were mapped with RFLP makers to identify quantitative trait loci (QTLs) conditioning SW across environments and populations. The genetic map of Pop1 consisted of 155 loci covering 973 cM, whereas Pop2 involved 153 loci and covered 1600 cM of map distance. For Pop1, the phenotypic data were collected from Plains, GA., Windblow, N.C., and Plymouth, N.C., in 1994. For Pop2, data were collected from Athens, GA., in 1994 and 1995, and Blackville, S.C., in 1995. Based on single-factor analysis of variance (ANOVA), seven and nine independent loci were associated with SW in Pop1 and Pop2, respectively. Together the loci explained 73% of the variability in SW in Pop1 and 74% in Pop2. Transgressive segregation occurred among the progeny in both populations. The marker loci associated with SW were highly consistent across environments and years. Two QTLs on linkage group (LG) F and K were located at similar genomic regions in both populations. The high consistency of QTLs across environments indicates that effective marker-assisted selection is feasible for soybean SW.     

RFLP mapping of QTLs conferring salt tolerance during the vegetative stage in tomato

 Quantitative trait loci (QTLs) contributing to salt tolerance during the vegetative stage in tomato were investigated using an interspecific backcross between a salt-sensitive Lycopersicon esculentum breeding line (NC84173, maternal and recurrent parent) and a salt-tolerant Lycopersicon pimpinellifolium accession (LA722). One hundred and nineteen BC₁ individuals were genotyped for 151 RFLP markers and a linkage map was constructed. The parental lines and 119 BC₁S₁ families (self-pollinated progeny of the BC₁ individuals) were evaluated for salt tolerance in aerated saline-solution cultures with the salt concentration gradually raised to 700 mM NaCl+70 mM CaCl₂ (equivalent to an electrical conductivity of approximately 64 dS/m and a water potential of approximately −35.2 bars). The two parental lines were distinctly different in salt tolerance: 80% of the LA722 plants versus 25% of the NC84173 plants survived for at least 2 weeks after the final salt concentration was reached. The BC₁S₁ population exhibited a continuous variation, typical of quantitative traits, with the survival rate of the BC₁S₁ families ranging from 9% to 94% with a mean of 51%. Two QTL mapping techniques, interval mapping (using MAPMAKER/QTL) and single-marker analysis (using QGENE), were used to identify QTLs. The results of both methods were similar and five QTLs were identified on chromosomes 1 (two QTLs), 3, 5 and 9. Each QTL accounted for between 5.7% and 17.7%, with the combined effects (of all five QTLs) exceeding 46%, of the total phenotypic variation. All QTLs had the positive QTL alleles from the salt-tolerant parent. Across QTLs, the effects were mainly additive in nature. Digenic epistatic interactions were evident among several QTL-linked and QTL-unlinked markers. The overall results indicate that tomato salt tolerance during the vegetative stage could be improved by marker-assisted selection using interspecific variation. Received: 4 January 1999 / Accepted: 4 January 1999     

Genetic analysis and mapping of genes controlling freezing tolerance in oilseedBrassica

Freezing tolerance is the ability of plants to survive subfreezing temperatures and is a major component of winter survival. In order to study the genetic regulation of freezing tolerance, an F2 population ofBrassica rapa and a doubled haploid population ofBrassica napus were assayedin vitro for relative freezing tolerance of acclimated and nonacclimated plants. Linkage maps developed previously were used to identify putative quantitative trait loci (QTL). Genomic regions with significant effects on freezing tolerance were not found for theB. napus population, but forB. rapa four regions were associated with acclimated freezing tolerance (FTA) and acclimation ability (FTB), and two unliked regions were associated with nonacclimated freezing tolerance (FTN). Acclimation ability was regulated by genes with very small additive effects and both positive and negative dominance effects. The allele from the winter parent at the FTN QTL had positive additive effects, but negative dominance effects. RFLP loci detected by a cold-induced and a stress-related cDNA fromArabidopsis thaliana mapped near two QTL for FTA/FTB. Further tests are needed to determine if alleles at these loci are responsible for the QTL effects we detected.     

RFLP mapping of resistance to the blackleg disease [causal agent,Leptosphaeria maculans (Desm.) Ces. et de Not.] in canola (Brassica napus L.)

We report the tagging of genes involved in blackleg resistance, present in the French cultivar Crésor of B. napus, with RFLP markers. A total of 218 cDNA probes were tested on the parental cultivars Crésor (resistant) and Westar (susceptible), and 141 polymorphic markers were used in a segregating population composed of 98 doubled-haploid lines (DH). A genetic map from this cross was constructed with 175 RFLP markers and allowed us to scan for specific chromosomal associations between response to blackleg infection and RFLP markers. Canola residues infested with virulent strains of Leptosphaeria maculans were used as inoculum and a suspension of pycnidiospores from cultures of L. maculans, including the highly virulent isolate Leroy, was sprayed to increase disease pressure. QTL mapping suggested that a single chromosomal region was responsible for resistance in each of the four environments tested. This QTL accounted for a high proportion of the variation of blackleg reaction in each of the assays. A second QTL, responsible for a small proportion of the variation of blackleg reaction, was present in one of four year-site assays. A Mendelian approach, using blackleg disease ratings for classifying DH lines as resistant or susceptible, also allowed us to map resistance in the region of the highly significant LOD scores observed in each environment by interval mapping. Results strongly support the presence of a single major gene, named LmFr ₁ controlling adult plant resistance to blackleg in spring oil-seed rape cultivar Crésor. Several RFLP markers were found associated with LmFr ₁.     

Mapping loci controlling flowering time in Brassica oleracea

The timing of the transition from vegetative to reproductive phase is a major determinant of the morphology and value of Brassica oleracea crops. Quantitative trait loci (QTLs) controlling flowering time in B. oleracea were mapped using restriction fragment length polymorphism (RFLP) loci and flowering data of F₃ families derived from a cabbage by broccoli cross. Plants were grown in the field, and a total of 15 surveys were made throughout the experiment at 5–15 day intervals, in which plants were inspected for the presence of flower buds or open flowers. The flowering traits used for data analysis were the proportion of annual plants (PF) within each F₃ family at the end of the experiment, and a flowering-time index (FT) that combined both qualitative (annual/biennial) and quantitative (days to flowering) information. Two QTLs on different linkage groups were found associated with both PF and FT and one additional QTL was found associated only with FT. When combined in a multi-locus model, all three QTLs explained 54.1% of the phenotypic variation in FT. Epistasis was found between two genomic regions associated with FT. Comparisons of map positions of QTLs in B. oleracea with those in B. napus and B. rapa provided no evidence for conservation of genomic regions associated with flowering time between these species.