The Vrn-H2 locus is a major determinant of flowering time in a facultative × winter growth habit barley (Hordeum vulgare L.) mapping population

With the aim of dissecting the genetic determinants of flowering time, vernalization response, and photoperiod sensitivity, we mapped the candidate genes for Vrn-H2 and Vrn-H1 in a facultative × winter barley mapping population and determined their relationships with flowering time and vernalization via QTL analysis. The Vrn-H2 candidate ZCCT-H genes were completely missing from the facultative parent and present in the winter barley parent. This gene was the major determinant of flowering time under long photoperiods in controlled environment experiments, irrespective of vernalization, and under spring-sown field experiments. It was the sole determinant of vernalization response, but the effect of the deletion was modulated by photoperiods when the vernalization requirement was fulfilled. There was no effect under short photoperiods. The Vrn-H1 candidate gene (HvBM5A) was mapped based on a microsatellite polymorphism we identified in the promoter of this gene. Otherwise, the HvBM5A alleles for the two parents were identical. Therefore, the significant flowering time QTL effect associated with this locus suggests tight linkage rather than pleiotropy. This QTL effect was smaller in magnitude than those associated with the Vrn-H2 locus and was significant in two-way interactions with Vrn-H2. The Vrn-H1 locus had no effect on vernalization response. Our results support the Vrn-H2/Vrn-H1 repressor/structural gene model for vernalization response in barley and suggest that photoperiod may also affect the Vrn genes or tightly linked loci.     

High-density AFLP map of nonbrittle rachis 1 (<Emphasis Type="Italic">btr1</Emphasis>) and 2 (<Emphasis Type="Italic">btr2</Emphasis>) genes in barley (<Emphasis Type="Italic">Hordeum vulgare</Emphasis> L.)

Wild relatives of barley disperse their seeds at maturity by means of their brittle rachis. In cultivated barley, brittleness of the rachis was lost during domestication. Nonbrittle rachis of occidental barley lines is controlled by a single gene (btr1) on chromosome 3H. However, nonbrittle rachis of oriental barley lines is controlled by a major gene (btr2) on chromosome 3H and two quantitative trait loci on chromosomes 5HL and 7H. This result suggests multiple mutations of the genes involved in the formation of brittle rachis in oriental lines. The btr1 and btr2 loci did not recombine in the mapping population analyzed. This result agrees with the theory of tight linkage between the two loci. A high-density amplified fragment-length polymorphism (AFLP) map of the btr1/btr2 region was constructed, providing an average density of 0.08 cM/locus. A phylogenetic tree based on the AFLPs showed clear separation of occidental and oriental barley lines. Thus, barley consists of at least two lineages as far as revealed by molecular markers linked to nonbrittle rachis genes.Electronic Supplementary Material Supplementary material is available for this article at An erratum to this article can be found at     

Varietal and chromosome 2H locus-specific frost tolerance in reproductive tissues of barley (<Emphasis Type="Italic">Hordeum vulgare</Emphasis> L.) detected using a frost simulation chamber

Exposure of flowering cereal crops to frost can cause sterility and grain damage, resulting in significant losses. However, efforts to breed for improved low temperature tolerance in reproductive tissues (LTR tolerance) has been hampered by the variable nature of natural frost events and the confounding effects of heading time on frost-induced damage in these tissues. Here, we establish conditions for detection of LTR tolerance in barley under reproducible simulated frost conditions in a custom-built frost chamber. An ice nucleator spray was used to minimize potential effects arising from variation in naturally occurring extrinsic nucleation factors. Barley genotypes differing in their field tolerance could be distinguished. Additionally, an LTR tolerance quantitative trait locus (QTL) on the long arm of barley chromosome 2H could be detected in segregating families. In a recombinant family, the QTL was shown to be separable from the effects of the nearby flowering time locus Flt-2L. At a minimum temperature of −3.5°C for 2 h, detection of the LTR tolerance locus was dependent on the presence of the nucleator spray, suggesting that the tolerance relates to freezing rather than chilling, and that it is not the result of plant-encoded variation in ice-nucleating properties of the tiller surface. Electronic supplementary material  The online version of this article (doi:) contains supplementary material, which is available to authorized users.     

Molecular mapping of the intermedium spike-c ( int-c) and non-brittle rachis 1 ( btr1) loci in barley ( Hordeum vulgare L.)

Lateral spikelet fertility and a non-brittle rachis are key characters in studying the evolution of barley. The fertility of lateral spikelets is controlled predominantly by the alleles at the vrs1 locus on chromosome 2HL and is modified by the alleles at the int-c locus on chromosome 4HS. The non-brittle rachis is controlled by alleles at two tightly linked loci, btr1 and btr2, on chromosome 3HS. This paper presents the integration of the int-c and btr1 loci in molecular linkage maps of barley. The int-c locus was mapped to the end of chromosome 4HS, 8.2 cM distal from the MWG2033 locus. The analysis was followed by a composite interval mapping of quantitative trait loci, which verified the position of the int-c locus. Linkage analysis using recombinant inbred lines showed that the btr1 locus is flanked between two AFLP loci, e14m27.4.1 and e15m19.7, with map distances of 3.1 cM and 4.2 cM, respectively. The molecular markers will expedite further high-density mapping of the int-c and btr1 loci.     

Characterization and molecular mapping of genes determining semidwarfism in barley

The semidwarf trait is desired in cereal breeding programs for increased lodging resistance. We characterized 27 brachytic (brh) semidwarf mutants in barley (Hordeum vulgare L.) and located the genes on barley chromosome linkage maps. All brachytic genes were transferred into the two-rowed cultivar Bowman by backcrossing four to seven times and selecting for semidwarf plants. The brachytic lines were evaluated for 10 phenotypic traits: plant height, awn, peduncle, and rachis internode length, leaf length and width, lodging, grain yield, number of kernels per spike, and kernel weight. We intercrossed the lines to determine which mutants were at independent loci and which were alleles at the same locus. F2 populations from 18 brh semidwarfs were constructed for genetic mapping using simple sequence repeat (SSR) markers. The brachytic semidwarf near-isogenic lines were significantly shorter than their normal counterparts and most had lower yields (16/27); shorter awns (26/27), peduncles (26/27), and rachis internodes (24/27); and reduced kernel weight (22/27). Twelve of the lines had shorter penultimate leaves and 15 had reduced lodging. Four lines had increased kernels per spike, while one had fewer kernels per spike. Allelism tests and mapping comparisons indicated that the 27 semidwarfs comprise 18 independent genetic loci. SSR mapping placed these loci in five of the seven barley chromosomes. Knowledge of the effects and locations of these brachytic semidwarf genes will help barley breeders select appropriate lines for barley improvement.     

Morphological and physiological analysis of cleistogamy in barley (Hordeum vulgare)

We previously reported that cleistogamy/chasmogamy (CL/CH) of barley ( Hordeum vulgare L.) is controlled by either two tightly linked genes or one gene with multiple alleles. To clarify the morphological and physiological mechanisms of barley CL, we analysed the lodicule size and auxin response of two cultivars whose CL/CH was controlled by two different genes, cly1 and Cly2 . In both cases, lodicules of the CL parent were smaller than those of the CH parent. Analyses of lodicule size and flowering phenotype of f ₁ plants and segregation analyses of the mapping population indicated that lodicule size co-segregated with the flowering phenotype. Indole-3-acetic acid (IAA) and other synthetic auxins, such as 2,4-dichlorophenoxyacetic acid, induced abnormal flowering in CH ears, in which florets remained open for a few days instead of the normal hour or so, but not in CL ears. This auxin effect also co-segregated with the flowering phenotype. Analyses of auxin-related compounds in the floret organs revealed that the anther contained high levels of IAA, whereas indole-3-carboxylic acid, a putative decarboxylated metabolite of IAA, accumulated only in lodicules of CH plants just at flowering. These results indicate that lodicule size and auxin response are pleiotropic effects of the CL gene, which may play a role in auxin response or metabolism.     

QTL mapping of chromosomal regions conferring reproductive frost tolerance in barley (<Emphasis Type="Italic">Hordeum vulgare</Emphasis> L.)

Spring radiation frost is a major abiotic stress in southern Australia, reducing yield potential and grain quality of barley by damaging sensitive reproductive organs in the latter stages of development. Field-based screening methods were developed, and genetic variation for reproductive frost tolerance was identified. Mapping populations that were segregating for reproductive frost tolerance were screened and significant QTL identified. QTL on chromosome 2HL were identified for frost-induced floret sterility in two different populations at the same genomic location. This QTL was not associated with previously reported developmental or stress-response loci. QTL on chromosome 5HL were identified for frost-induced floret sterility and frost-induced grain damage in all three of the populations studied. The locations of QTL were coincident with previously reported vegetative frost tolerance loci close to the vrn-H1 locus. This locus on chromosome 5HL has now been associated with response to cold stress at both vegetative and reproductive developmental stages in barley. This study will allow reproductive frost tolerance to be seriously pursued as a breeding objective by facilitating a change from difficult phenotypic selection to high-throughput genotypic selection.     

Flt-2L, a locus in barley controlling flowering time, spike density, and plant height

Flowering time represents an important adaptive trait for temperate cereal crops and may also impact on frost damage in cereal reproductive tissues by enabling escape or by influencing accumulation of genuine tolerance. The Flowering time-2L (Flt-2L) quantitative trait locus (QTL) on the distal end of barley chromosome arm 2HL overlaps with QTL for rachis internode length and reproductive frost damage. Flt-2L was also found to be associated with plant height. By combining marker analysis with phenotyping in progeny families of selected Amagi Nijo × WI2585 F₆ recombinants, we were able to map quantitative flowering time, rachis internode length, and plant height effects on 2HL as discrete Mendelian traits. The three developmental characters showed codominant modes of expression and perfectly cosegregated with one another in a 1.3-cM marker interval, indicating control by the same gene or closely linked genes. Twelve genes were identified in the related intervals in the rice and Brachypodium distachyon genomes. The HvAP2 gene cosegregated with Flt-2L and represents a plausible candidate for Flt-2L, since it is highly similar to the wheat domestication gene Q which has similar developmental effects. These data will contribute to isolation of the Flt-2L gene(s) and help establish the basis of the frost damage QTL. Electronic supplementary material  The online version of this article (doi:) contains supplementary material, which is available to authorized users.     

High resolution mapping of Dense spike-ar (<Emphasis Type="Italic">dsp.ar</Emphasis>) to the genetic centromere of barley chromosome 7H

Spike density in barley is under the control of several major genes, as documented previously by genetic analysis of a number of morphological mutants. One such class of mutants affects the rachis internode length leading to dense or compact spikes and the underlying genes were designated dense spike (dsp). We previously delimited two introgressed genomic segments on chromosome 3H (21 SNP loci, 35.5 cM) and 7H (17 SNP loci, 20.34 cM) in BW265, a BC₇F₃ nearly isogenic line (NIL) of cv. Bowman as potentially containing the dense spike mutant locus dsp.ar, by genotyping 1,536 single nucleotide polymorphism (SNP) markers in both BW265 and its recurrent parent. Here, the gene was allocated by high-resolution bi-parental mapping to a 0.37 cM interval between markers SC57808 (Hv_SPL14)–CAPSK06413 residing on the short and long arm at the genetic centromere of chromosome 7H, respectively. This region putatively contains more than 800 genes as deduced by comparison with the collinear regions of barley, rice, sorghum and Brachypodium, Classical map-based isolation of the gene dsp.ar thus will be complicated due to the infavorable relationship of genetic to physical distances at the target locus.     

Molecular and comparative mapping of genes governing spike compactness from wild emmer wheat

The development and morphology of the wheat spike is important because the spike is where reproduction occurs and it holds the grains until harvest. Therefore, genes that influence spike morphology are of interest from both theoretical and practical stand points. When substituted for the native chromosome 2A in the tetraploid Langdon (LDN) durum wheat background, the Triticum turgidum ssp. dicoccoides chromosome 2A from accession IsraelA confers a short, compact spike with fewer spikelets per spike compared to LDN. Molecular mapping and quantitative trait loci (QTL) analysis of these traits in a homozygous recombinant population derived from LDN × the chromosome 2A substitution line (LDNIsA-2A) indicated that the number of spikelets per spike and spike length were controlled by linked, but different, loci on the long arm of 2A. A QTL explaining most of the variation for spike compactness coincided with the QTL for spike length. Comparative mapping indicated that the QTL for number of spikelets per spike overlapped with a previously mapped QTL for Fusarium head blight susceptibility. The genes governing spike length and compactness were not orthologous to either sog or C, genes known to confer compact spikes in diploid and hexaploid wheat, respectively. Mapping and sequence analysis indicated that the gene governing spike length and compactness derived from wild emmer could be an ortholog of the barley Cly1/Zeo gene, which research indicates is an AP2-like gene pleiotropically affecting cleistogamy, flowering time, and rachis internode length. This work provides researchers with knowledge of new genetic loci and associated markers that may be useful for manipulating spike morphology in durum wheat.     

Structure–function analysis of the barley genome: the gene-rich region of chromosome 2HL

A major gene-rich region on the end of the long arm of Triticeae group 2 chromosomes exhibits high recombination frequencies, making it an attractive region for positional cloning. Traits known to be controlled by this region include chasmogamy/cleistogamy, frost tolerance at flowering, grain yield, head architecture, and resistance to Fusarium head blight and rusts. To assist these cloning efforts, we constructed detailed genetic maps of barley chromosome 2H, including 61 polymerase chain reaction markers. Colinearity with rice occurred in eight distinct blocks, including five blocks in the terminal gene-rich region. Alignment of rice sequences from the junctions of colinear chromosome segments provided no evidence for the involvement of long (>2.5 kb) inverted repeats in generating inversions. However, reuse of some junction sequences in two or three separate evolutionary breakage/fusion events was implicated, suggesting the presence of fragile sites. Sequencing across 91 gene fragments totaling 107 kb from four barley genotypes revealed the highest single nucleotide substitution and insertion–deletion polymorphism levels in the terminal regions of the chromosome arms. The maps will assist in the isolation of genes from the chromosome 2L gene-rich region in barley and wheat by providing markers and accelerating the identification of the corresponding points in the rice genome sequence. Electronic supplementary material  The online version of this article (doi:) contains supplementary material, which is available to authorized users.     

Identification and mapping of cleistogamy genes in barley

Cleistogamy is a closed type of flowering with ensured self-pollination and an important trait to study evolutionary development in flower organs, reproduction systems, gene flow, and disease control. Still, very limited information is available about the genetic control and regulatory mechanism of this trait in barley. In this work, from the eight crosses between cleistogamous and chasmogamous accessions, five crosses generated chasmogamous F₁ plants and their F₂ plants segregated as 3 chasmogamous:1 cleistogamous, whereas three crosses generated cleistogamous F₁ plants, and their F₂ plants segregated as 1 chasmogamous:3 cleistogamous. Although a single gene was responsible for the control of cleistogamy in these two groups of crosses, the direction of dominance was opposite, suggesting two genes, cly1 and Cly2, for the genetic control of cleistogamy in barley. Epistatic type of gene interaction between the two loci was detected. In the analysis of 99 recombinant inbred lines of Azumamugi × Kanto Nakate Gold and doubled haploid lines of Harrington × Mikamo Golden, where in both crosses F₁ was chasmogamous, the cly1 locus has been mapped on chromosome 2HL. Using the analysis of the F₂ population of Misato Golden and Satsuki Nijo where F₁ was cleistogamous, the Cly2 locus was mapped in the same region of chromosome 2HL. Because the cly1 and Cly2 loci were mapped in the same region in these three different mapping populations, it was concluded that the expression of cleistogamy is under the control of two tightly linked genes or different alleles of the same gene.     

Characteristics and inheritance of radiation-induced mutations in barley II. Two short-awned mutations

Two mutations were obtained in a two-rowed, spring-type barley variety, Asahi No. 5. KM 114 was obtained in the X₂ population of X-irradiated dormant seeds (15 Kr) and KM 218 was found in X₂ of gamma-irradiated seeds (15 Kr). KM 114 had very short awns, short plant height, almost normal spike size, and narrow and thinner seeds with pointed tip. The mutant gene is Mendelian recessive. KM 218 plants were normal or slightly taller, had somewhat larger spike, with somewhat longer rachis internodes, medium-long awn, and slightly longer, narrower and thinner kernel than those of the parental variety. At maturity, grains are exposed in many seeds (semi-naked kernels). This mutant also has a conspicuous diagnostic trait; the abnormal development of rachilla into extra florets, some of which set seeds. This mutant gene is also Mendelian recessive. Genes for these two mutants are not allelic and are inherited independently as indicated in F₂ and F₃ analyses. Genes of these two mutations will serve as good marker genes in genetic and linkage studies in barley, because their character expression is consistent and the recessive homozygotes are viable and fertile.     

Effects of loci on chromosomes 2 (2H) and 7 (5H) on developmental patterns in barley (Hordeum vulgare L.) under different photoperiod regimes

 Heading-date in cereals is the final result of a number of interacting characters that include vernalization requirement, photoperiod sensitivity, and earliness per se. Progress in developing adapted varieties may be achieved by determining the chromosomal locations of genes controlling these characters. Nineteen doubled-haploid (DH) lines from the Dicktoo×Morex mapping population were phenotyped in controlled- environment photoperiod experiments to determine the role of two previously detected QTLs on the developmental patterns of barley. The QTLs are hypothesised to represent the effects of the Ppd and Sh2 loci on chromosomes 2 (2H) and 7 (5H), respectively. Alleles at the Ppd locus were found to be vary in response to photoperiod duration. Vernalization had some effect on alleles at both loci. The presence of early and late- flowering transgressive segregants in this mapping population can be explained by interactions between the Ppd and Sh2 loci. The Ppd and Sh2 loci are hypothesised to be homoeologous with the Ppd and Vrn1 loci of wheat. Received: 1 August 1996 / Accepted: 15 November 1996     

Geographic distribution of alleles at the Ga2 locus for segregation distortion in barley

Summary A distorted segregation of esterase alleles at the complex loci, Est1, Est2 and Est4, was found in an F₂ population. This distortion is typical for cross combinations between the Ga2Ga2 and ga2ga2 genotypes responsible for segregation distortion, since the Ga2 locus is linked with the complex loci encoding the esterase isozymes. The segregation of esterase isozyme patterns in F₂ populations between 473 varieties of barley and a tester of ga2ga2 genotype was examined, and the genotypes inducing segregation distortion were detected. Varieties with a ga2ga2 genotype are widely distributed throughout the world, whereas Ga2Ga2 varieties are found only in eastern and southern regions of Asia, from Japan to North India, with a low frequency. In varieties collected from these regions, some associations were detected between alleles at the Ga2 locus and esterase isozyme patterns. Additionally, most of the Ga2 barley varieties are naked and possess a BtBtbt2bt2 genotype for a non-brittle rachis.     

Molecular marker assisted genetic analysis of head shattering in six-rowed barley

Head shattering in barley (Hordeum vulgare L.) has two forms; brittle rachis and weak rachis. Brittle rachis is not observed in cultivated barley since all cultivars carry non-brittle alleles at one of the two complementary brittle rachis loci (Btr1;Btr2). Weak rachis causes head shattering in barley cultivars and may be confused with brittle rachis. Brittle rachis has been mapped to the chromosome 3 (3H) short arm while map position(s) of the weak rachis is unknown. Two major and a putative minor QTL for head shattering were mapped using the Steptoe × Morex doubled haploid line population. The largest QTL, designated Hst-3, located on the chromosome 3 (3H) centromeric region, is associated with a major yield QTL. The Steptoe Hst-3 region, when transferred into Morex, resulted in a substantial decrease in head shattering. High-resolution mapping of Hst-3 was achieved using isogenic lines. Brittle rachis was mapped with molecular markers and shown to be located in a different position from that of Hst-3. The second major QTL, designated Hst-2 S, is located on chromosome 2 S. This locus is associated with an environmentally sensitive yield QTL.     

Elucidation of the origin of ‘agriocrithon’ based on domestication genes questions the hypothesis that Tibet is one of the centers of barley domestication

Wild barley forms a two‐rowed spike with a brittle rachis whereas domesticated barley has two‐ or six‐rowed spikes with a tough rachis. Like domesticated barley, ‘agriocrithon’ forms a six‐rowed spike; however, the spike is brittle as in wild barley, which makes the origin of agriocrithon obscure. Haplotype analysis of the Six‐rowed spike 1 (vrs1) and Non‐brittle rachis 1 (btr1) and 2 (btr2) genes was conducted to infer the origin of agriocrithon barley. Some agriocrithon barley accessions (eu‐agriocrithon) carried Btr1 and Btr2 haplotypes that are not found in any cultivars, implying that they are directly derived from wild barley through a mutation at the vrs1 locus. Other agriocrithon barley accessions (pseudo‐agriocrithon) carried Btr1 or Btr2 from cultivated barley, thus implying that they originated from hybridization between six‐rowed landraces carrying btr1Btr2 and Btr1btr2 genotypes followed by recombination to produce Btr1Btr2. All materials we collected from Tibet belong to pseudo‐agriocrithon and thus do not support the Tibetan Plateau as being a center of barley domestication. Tracing the evolutionary history of these allelic variants revealed that eu‐agriocrithon represents six‐rowed barley lineages that were selected by early farmers, once in south‐eastern Turkmenistan (vrs1.a1) and again in the eastern part of Uzbekistan (vrs1.a4).     

Microsatellite mapping of genes that determine supernumerary spikelets in wheat (T. aestivum) and rye (S. cereale)

The wheat and rye spike normally bears one spikelet per rachis node, and the appearance of supernumerary spikelets is rare. The loci responsible for the ‘multirow spike’ or MRS trait in wheat, and the ‘monstrosum spike’ trait in rye were mapped by genotyping F₂ populations with microsatellite markers. Both MRS and the ‘monstrosum’ trait are under the control of a recessive allele at a single locus. The Mrs1 locus is located on chromosome 2DS, co-segregating with the microsatellite locus Xwmc453. The placement of flanking microsatellite loci into chromosome deletion bin 2DS-5 (FL 0.47–1.0) delimited the physical location of Mrs1 to the distal half of chromosome arm 2DS, within the gene rich region 2S0.8. The Mo1 locus maps about 10 cM from the centromere on chromosome arm 2RS. The similar effect on phenotype of mo1 and mrs1, together with their presence in regions of conserved synteny, suggest that they may well be members of an orthologous set of Triticeae genes governing spike branching. The practical importance of the MRS spike is that it produces more spikelets per spike, and thereby enhances the sink capacity of wheat, which is believed to limit the yield potential of the crop.     

Large deletions within the first intron in <Emphasis Type="Italic">VRN-1</Emphasis> are associated with spring growth habit in barley and wheat

The broad adaptability of wheat and barley is in part attributable to their flexible growth habit, in that spring forms have recurrently evolved from the ancestral winter growth habit. In diploid wheat and barley growth habit is determined by allelic variation at the VRN-1 and/or VRN-2 loci, whereas in the polyploid wheat species it is determined primarily by allelic variation at VRN-1. Dominant Vrn-A1 alleles for spring growth habit are frequently associated with mutations in the promoter region in diploid wheat and in the A genome of common wheat. However, several dominant Vrn-A1, Vrn-B1, Vrn-D1 (common wheat) and Vrn-H1 (barley) alleles show no polymorphisms in the promoter region relative to their respective recessive alleles. In this study, we sequenced the complete VRN-1 gene from these accessions and found that all of them have large deletions within the first intron, which overlap in a 4-kb region. Furthermore, a 2.8-kb segment within the 4-kb region showed high sequence conservation among the different recessive alleles. PCR markers for these deletions showed that similar deletions were present in all the accessions with known Vrn-B1 and Vrn-D1 alleles, and in 51 hexaploid spring wheat accessions previously shown to have no polymorphisms in the VRN-A1 promoter region. Twenty-four tetraploid wheat accessions had a similar deletion in VRN-A1 intron 1. We hypothesize that the 2.8-kb conserved region includes regulatory elements important for the vernalization requirement. Epistatic interactions between VRN-H2 and the VRN-H1 allele with the intron 1 deletion suggest that the deleted region may include a recognition site for the flowering repression mediated by the product of the VRN-H2 gene of barley.     

Genetic Diversity Analysis of Tibetan Wild Barley Using SSR Markers

One hundred and six accessions of wild barley collected from Tibet, China, including 50 entries of the two-rowed wild barley Hordeum vulgare ssp. spontaneum (HS), 29 entries of the six-rowed wild barley Hordeum vulgare ssp. agriocrithon (HA), and 27 entries of the six-rowed wild barley Hordeum vulgare ssp. agriocrithon var. lagunculiforme (HL), were analyzed using 30 SSR markers selected from the seven barley linkage groups for studying genetic diversity and evolutionary relationship of the three subspecies of Tibetan wild barley to cultivated barley in China. Over the 30 genetic loci that were studied, 229 alleles were identified among the 106 accessions, of which 70 were common alleles. H. vulgare ssp. spontaneum possesses about thrice more private alleles (2.83 alleles/locus) than HS (0.93 alleles/locus), whereas almost no private alleles were detected in HL. The genetic diversity among-subspecies is much higher than that within-subspecies. Generally, the genetic diversity among the three subspecies is of the order HS > HL > HA. Phylogenetic analysis of the 106 accessions showed that all the accessions of HS and HA was clustered in their own groups, whereas the 27 accessions of HL were separated into two groups (14 entries with group HS and the rest with group HA). This indicated that HL was an intermediate form between HS and HA. Based on this study and previous works, we suggested that Chinese cultivated barley might evolve from HS via HL to HA.