Does introgression require interspecific hybrids to be fertile? How is this possible?

Does introgression require interspecific hybrids to be fertile? How is this possible?

We are searching data for your request:

Forums and discussions:
Manuals and reference books:
Data from registers:
Wait the end of the search in all databases.
Upon completion, a link will appear to access the found materials.

Wikipedia says that Introgression, also known as introgressive hybridization, in genetics is the movement of a gene (gene flow) from one species into the gene pool of another by the repeated backcrossing of an interspecific hybrid with one of its parent species.

Does this mean that the interspecific hybrid is fertile? But is it possible, that an interspecific hybrid is fertile?

Does this mean that the interspecific hybrid is fertile?


But is it possible, that an interspecific hybrid is fertile?


You might be confused by the idea that by definition two individuals from different species should not be able to interbreed. If so, then you should have a look at How could humans have interbred with Neanderthals if we're a different species?.

Genomic introgression through interspecific hybridization counteracts genetic bottleneck during soybean domestication

Evidence of introgression, the transfer of genetic material, between crops and their wild relatives through spontaneous hybridization and subsequent backcrossing has been documented however, the evolutionary patterns and consequences of introgression and its influence on the processes of crop domestication and varietal diversification are poorly understood.


We investigate the genomic landscape and evolution of putative crop-wild-relative introgression by analyzing the nuclear and chloroplast genomes from a panel of wild (Glycine soja) and domesticated (Glycine max) soybeans. Our data suggest that naturally occurring introgression between wild and domesticated soybeans was widespread and that introgressed variation in both wild and domesticated soybeans was selected against throughout the genomes and preferentially removed from the genomic regions underlying selective sweeps and domestication quantitative trait locus (QTL). In both taxa, putative introgression was preferentially retained in recombination-repressed pericentromeric regions that exhibit lower gene densities, reflecting potential roles of recombination in purging introgression. Despite extensive removal of introgressed variation by recurrent selection for domestication-related QTL and associated genomic regions, spontaneous interspecific hybridization during soybean domestication appear to have contributed to a rapid varietal diversification with high levels of genetic diversity and asymmetric evolution between the nuclear and chloroplast genomes.


This work reveals the evolutionary forces, patterns, and consequences of putative genomic introgression between crops and their wild relatives, and the effects of introgression on the processes of crop domestication and varietal diversification. We envision that interspecific introgression serves as an important mechanism for counteracting the reduction of genetic diversity in domesticated crops, particularly the ones under single domestication.

Factors Limiting Natural Hybridization

A variety of factors serve as reproductive barriers among plant taxa . These barriers, which can be subdivided into those acting prior to fertilization (prezygotic) or following fertilization (postzygotic), restrict natural hybridization and help maintain species boundaries.

Prezygotic Barriers.

The potential for natural hybridization is largely determined by the proximity of potential mates in both space and time. The likelihood of hybridization is therefore governed, to a large extent, by differences in the ecology (spatial isolation) and/or phenology (temporal isolation) of the individuals of interest. Even if ecological and temporal differentiation are absent, pollen transfer may be limited by differences in floral morphology (form). Differences in traits such as floral color, fragrance, and nectar chemistry can influence pollinator behavior and may discourage the transfer of pollen among different species (ethological isolation). Alternatively, the structure of the flower may preclude or limit pollination of one taxon by the pollinator(s) of others (mechanical isolation). Finally, even if pollen transfer is successful, the pollen may not germinate on a foreign stigma if it does, the pollen tubes may fail to effect fertilization due to slow growth or arrest prior to reaching the ovule (cross-incompatibility).

Postzygotic Barriers.

Assuming that fertilization occurs, the resulting hybrid progeny (offspring) may fail to survive to reproductive maturity due to developmental aberrations (hybrid inviability). If the hybrids do survive, their flowers may be unattractive to pollinators, thereby restricting further hybridization (floral isolation). Alternatively, the hybrids may be attractive to pollinators but partially or completely sterile (hybrid sterility). Finally, even if first generation hybrids are viable and fertile, later-generation hybrids may exhibit decreased levels of viability and/or fertility (hybrid breakdown).


Three types of interspecific hybridization can be defined according to the reproductive characteristics of the first generation hybrids: (i) hybridization yielding unviable or infertile offspring (ii) genome-exclusion hybridization producing F1 hybrids that exclude one parental genome during gametogenesis and, therefore avoiding genetic introgression between the parental species and (iii) hybridization generating F1 hybrids in which genetic recombination between the parental genomes proceeds normally during gametogenesis. The threat to parental species that hybridization with genome exclusion may represent is certainly not due to genetic introgression but can be caused by a demographic decline in one or both species, yet this phenomenon is poorly understood. In conservation biology, demography is considered of primary importance in determining the viability of wild populations (Lande 1988 ). In particular, the viability of small-sized populations depends on the interaction between demographic and genetic factors, both playing a role during a demographic decline. A minimum population size has to be maintained to avoid inbreeding depression in the short term and retaining sufficient genetic variation to allow adaptive changes over large timescales (Jamieson and Allendorf 2012 ).

In addition to the potential threat that interspecific hybridization between native species may represent, especially in the context of global change, interspecific hybridization may facilitate biological invasions by accelerating the population decline of native species (Hall and Ayres 2009 ). The invasion of exotic species is one of the main threats to biodiversity worldwide, in which competition, predation or even habitat modification by exotic species can lead native taxa to the brink of extinction (Mack et al. 2000 ). By modelling interspecific hybridization, it is possible to assess the future consequences on the demography of parental species or hybrid populations and to predict the conditions under which local populations or even a species can reach extinction.

Here, we present an extensive investigation of the possible outcomes on the demography of native species threatened by the introduction of an alien species with which they hybridize, and we project the impact of different changes in the biological setting to identify effective conservation actions. To this aim, we adapted and implemented a model of interspecific hybridization without genetic introgression that we have recently developed (Quilodrán et al. 2014 ). The case study of biological invasion reinforced by nonintrogressive hybridization that we investigated here is the colonization of Western Europe by the waterfrog Pelophylax ridibundus coming either from Central or Southern Europe (Holsbeek and Jooris 2010 ), apparently mediated by human activities (Luquet et al. 2011 ). In France and Switzerland, this waterfrog was introduced during the 20th century for frog legs consumption and for scientific purposes. Since then, this exotic species has been displacing local populations of native waterfrogs probably facilitated by the fact that it can hybridize with the native Pelophylax lessonae (Vorburger and Reyer 2003 ). This threat to native frogs adds to the global worldwide threat to amphibians caused by human-induced climate change (Shoo et al. 2011 ).

In nature, hybrids between P. ridibundus and P. lessonae were present in Western Europe before the recent introduction of P. ridibundus they have been named Pelophylax esculentus and have long been considered as a different species. Previous genetic studies revealed the hybrid nature of P. esculentus, which likely originated at the time of the last glaciation, about 10 000 years ago, when native populations of P. ridibundus were suggested to be present in Western Europe but went subsequently extinct (Vorburger 2001 ). The hybrid P. esculentus persisted by hybridogenesis with P. lessonae despite the extinction of P. ridibundus (Fig. 1). This hybridogenetic system is characterized by the hybrid P. esculentus, which develops germ cells that discard the P. lessonae genome before meiosis, producing only haploid gametes containing the P. ridibundus genome (Anholt et al. 2003 ). Thus, backcrosses between P. esculentus and P. lessonae generate only P. esculentus hybrids. Crosses between native P. esculentus hybrids are unsuccessful likely due to the expression of recessive deleterious mutations (Vorburger and Reyer 2003 ). This hybridogenetic system has been named the L/E system (Graf 1986 ).

The equilibrium of the L/E system was disrupted in Western Europe by the anthropogenic introduction of P. ridibundus from different places of Central and Southern Europe (Schmeller et al. 2007 ). It has been recently highlighted that populations of P. ridibundus from different geographical origins differ in the type of offspring they produce (Holsbeek and Jooris 2010 Plötner et al. 2010 ). Introduced individuals coming from Southern Europe, when mating with the native P. lessonae, produce new P. esculentus hybrids that are sterile in all types of crosses (see Fig. 1). In contrast, if the parental P. ridibundus comes from Central Europe, the resulting P. esculentus hybrids are fertile in all types of crosses. The latter situation is explained by the probable absence of recessive deleterious mutations in the genome of P. ridibundus coming from Central Europe (Holsbeek and Jooris 2010 ).

The introduction of P. ridibundus in Western Europe is likely to cause complex interactions and perturbations in the native L/E system, like ecological interspecific competition, waste of reproductive potential for native taxa when they mate with the invasive species or population replacement through hybrid production and asymmetric backcrosses with parental species. Therefore, predicting the risks to resident populations and proposing efficient protection strategies are not trivial tasks. Models that describe the native L/E system in Western Europe have already been proposed (Graf 1986 Hellriegel and Reyer 2000 Som et al. 2000 Som and Reyer 2006 ), but our model is more general and adaptable because important processes like density-dependent competition, dominance/recessive inheritance of traits and assortative mating are incorporated. It allows us to include into the system the populations of the invasive P. ridibundus coming from different origins, with their reproductive specificities, an element which was absent from previous models. Thus, using our model as a tool to simulate different scenarios of invasion and to assess their consequences in the biological system, we aimed at: i) generating a comprehensive picture of the threats that native frogs may face, ii) determining the conditions under which native waterfrogs may be at risk, and iii) identifying key factors that can be modified for deploying an efficient conservation strategy.

Interspecific Fertile Hybrids of Haplochromine Cichlidae (Teleostei) and Their Possible Importance for Speciation

INTERSPECIFIC FERTILE HYBRIDS OF HAPLOCHROMINE CICHLIDAE (TELEOSTEI) AND THEIR POSSIBLE IMPORTANCE FOR SPECIATION by M.-D. CRAPON DE CAPRONA and B. FRITZSCH (Fakultät für Biologie, Universität Bielefeld, 4800 Bielefeld-1, Federal Republic of Germany) SUMMARY Hybridization tests were carried out with the allopatric and sympatric haplochromine cichlids Haplochromis burtoni (Lake Tanganyika), H. nubilus (L. Victoria), H. elegans (L. George) and H. "black lividus" (L. Victoria). Female H. nubilus crossed with male H. burtoni produce lethal hybrids which do not survive the larval stage. The reciprocal cross, female H. burtoni x H. nubilus, produces F1, F2, F3 hybrids with a normal sex ratio, and fertile backcrosses F1 x H. burtoni and F1 x H. nubilus with a skewed sex ratio. In behavioral tests female H. nubilus spawn readily with H. burtoni males (lethal cross), whereas female H. burtoni spawn rarely with male H. nubilus (fertile cross). The inheritance of coloration, the "egg-dummies" pattern on the anal fin, and reproductive vigor in the hybrids are analysed in detail. Crosses which involve male H. elegans x female H. nubilus and male H. "black lividus" x female H. nubilus are fertile but have skewed sex ratios in favor of the females. The importance of


Netherlands Journal of Zoology (in 2003 continued as Animal Biology) &ndash Brill


Switchgrass and its relatives have the potential to be important bioenergy crops as feedstock, yet germplasm improvement remains a critical barrier to this goal. The development of improved cultivars and hybrids would benefit from the ability to combine the genetic variation of intervarietal ecotypes as well as related species. The ultimate goal would be to expand the pool of available genetic resources further through the identification and introgression of valuable biofuel traits. Here we demonstrate the recovery of an interspecific crosses between transgenic herbicide-resistant P. virgatum cv. Alamo switchgrass and wild-type P. aramrum var. amarulum (ACP) and the subsequent generation of a fertile admixture population.

Phylogenetics in Switchgrass and Related Species

There are approximately 500 species in the genus Panicum ( Huang et al., 2011 ). Genomics on switchgrass and its relatives is relatively recent ( Lu et al., 2013 ), yet there has been considerable clarification in the phylogenetic and taxanomic relationships among species from genomic studies ( Casler et al., 2007 , 2011 Huang et al., 2011 Lu et al., 2013 ). More recently, switchgrass genomic diversity, ploidy, and evolution were explored using a network-based SNP discovery protocol ( Lu et al., 2013 ). These studies yielded a linkage map, an EST database, and a set of SNP markers across 18 linkage groups and bacterial artificial chromosome libraries. These results found that the switchgrass genome is highly syntenous with the genomes of rice, sorghum (Sorghum bicolor (L.) Moench), and Brachypodium distachyon (L.) P.Beauv. ( Casler et al., 2011 ) and illustrate isolation by distance and by ploidy between switchgrass populations ( Lu et al., 2013 ).

Phylogenetic analyses ( Lu et al., 2013 ) of switchgrass indicate a tendency of south-to-north migration in North America. Ploidy levels vary within switchgrass ecotypes ( Hopkins et al., 1996 Lu et al., 1998 Narasimhamoorthy et al., 2008 ), ranging from diploid (2n = 2× = 18) to duodecaploid (2n = 12× = 108) ( Nielsen, 1944 ). All lowland ecotypes have been identified as tetraploids (2n = 4× = 36) upland types can be tetraploids or octaploids (2n = 8× = 72) ( Hultquist et al., 1996 ). Mixed ploidy levels among accessions and within cultivars have also been observed ( Narasimhamoorthy et al., 2008 ). Both geographic isolation and sexual incompatibility related to ploidy have resulted in both varietal and species-specific diversification. The switchgrass accession used in this study, Alamo, is a tetraploid member of the lowland ecotype ( Serba et al., 2013 ). Using sequence-related amplified polymorphism and EST–simple sequence repeat markers Huang et al. (2011) shows that P. amarum is a sister taxon to P. virgatum. Close genetic proximity and multiple abiotic stress resistance and yield traits result in them being excellent candidates for hybridization.

Interspecific F1 hybrids between HbR transgenic Alamo switchgrass and wild-type ACP were generated in this study. Phenotypic and molecular comparisons identified clear differences between the parents themselves and between the parents and F1 offspring, indicating the likely origin of F1 and F1BC1 individuals as interspecific hybrids. Molecular and genomic analysis of the F1 offspring indicated the presence of an ACP-specific maternal chloroplast allele and the paternal bar transgene, providing further confirmation of the F1 hybrid genome. A hybrid F1 plant was backcrossed to the wild-type Alamo and an F1BC1 population of 83 herbicide-sensitive progeny was recovered. All F1BC1 progeny were identified as backcross hybrids by genotyping, demonstrating the robustness to this approach for screening. Phenotypically each of these individuals exhibited unique characteristics and were stable to floral maturity. As predicted, the majority of the HbS plants no longer contained the bar transgene as evaluated by PCR assay and sequencing. With the application of further testing for transgenic sequence in select F1BC1 individuals, this method may allow the recovery of nontransgenic and, arguably, offspring that are not genetically modified organisms (GMOs) from wide crosses. This method provides a proof of concept for efficient selection of interspecific hybrids using a selectable transgenic as an intermediate. It is also likely that this approach will also extend to intervarietal crosses.

Accuracy of Variant Identification and Imputation

Although well-established methods for molecular marker analysis have been described previously for the development of mapping populations ( Missaoui et al., 2005 Okada et al., 2010 ) and phylogenetic comparisons ( Gunter et al., 1996 Hultquist et al., 1996 Tobias et al., 2005 , 2006 , 2008 ), here, we have applied the GBS approach for marker identification and data analysis as previously described ( Heffelfinger et al., 2014 ). The major strength of this method is that it allows a considerably higher resolution of recombination via a denser panel of markers than traditional nonsequencing-based methodologies. Although the results quantitatively assess the hybrid contribution of both parents and identify patterns of recombination in the F1BC1 individuals, questions over the rate of error remain.

Beyond the specific concerns of variant identification with GBS and low coverage sequencing in general ( Elshire et al., 2011 Heffelfinger et al., 2014 ), the switchgrass reference genome presents additional problems. The current state of the P. virgatum reference assembly is scaffolded contigs with synteny established against the related species P. hallii genome (P. virgatum version 1.1,, accessed 6 Mar. 2015). Compounding this difficulty is that switchgrass is an allotetraploid, resulting in a highly repetitive genome. The result of this is that few variants achieve a high mapping quality (>30) of those that do, the possibility of misalignment remains because of the draft nature of the assembly.

To resolve confounding issues as best as possible, a stringent set of variant filtering metrics was applied to retained markers. These included that variants were required to be homozygous within the parents, polymorphic between the parents, and heterozygous in the F1 offspring. The requirement that all three possible marker states had to be observed reduced the likelihood of PCR, sequencing, and mapping artifacts. No expectation for allele frequency or segregation was applied to the F1BC1 however, as segregation distortion would be masked.

Another concern, not related to the reference genome but caused instead by the high degree of multiplexing, was false homozygosity. False homozygosity results when only one allele of a heterozygous site is observed in the sequencing data. False homozygosity was primarily solved through the imputation and error correction method, which is relatively insensitive to a single genotype call and instead determines a regional genotype based on a “mean” value from a set of calls ( Heffelfinger et al., 2014 ). As long as homozygous miscalls did not randomly skew toward ACP or Alamo for a given region, the rate of erroneously imputed homozygous sites should be low.

A partial estimate of this error rate can be obtained from the amount of the genome across F1BC1 individuals called as ACP homozygous. Due to the nature of the cross, this genotype state is not physically possible, but may nonetheless result from false homozygous calls. Across all samples, approximately 1.92% of the genome is called as ACP homozygous. Assuming this error rate results in the same percentage of the genome being miscalled as Alamo homozygous, the total postimputation error rate caused by false homozygosity is under 4%. In reality, the percentage of the genome miscalled as Alamo homozygous or heterozygous is probably higher than 4%, however, as erroneously mapped reads may result in regions of the genome being “placed” incorrectly, even if the genotype is technically correct.

Evidence of Segregation Distortion in the F1BC1 Offspring

A primary concern in interspecific crosses is the segregation distortion caused by genetic incompatibility. If such forces were active, one would expect ACP alleles to be selectively lost in the F1BC1 because of backcrossing to the Alamo parent. Such regions with a fixed homozygous Alamo state would present severe obstacles to introgressive hybridization and trait mapping because of a local lack of ACP genetic variation and recombination. However, when the data were analyzed for such regions, there was no evidence of any segregation distortion favoring a homozygous Alamo state in the postimputation F1BC1 offspring.

Instead, the results suggested segregation distortion toward the heterozygous state in some regions of the genome. One region of heterozygosity was expected because of the selection against herbicide resistance. Selection for herbicide sensitivity would cause the T-DNA insertion site from the Alamo genome to be absent in all F1BC1 individuals. Nevertheless, other regions of fixed heterozygosity may represent segregation distortion in favor of alleles from the non-backcross ACP parent. Switchgrass exhibits both prefertilization and postfertilization self-incompatibility systems ( Martinez-Reyna and Vogel, 2002 , Martinez-Reyna and Vogel, 2008 ). Gametophytic self-incompatibility in grasses is controlled by two loci, S and Z ( Lundqvist, 1962 ). In switchgrass Martinez-Reyna and Vogel (2002) showed that in controlled octoploid × octoploid, octoploid × tetraploid, and tetraploid × octoploid crosses, postfertilization abortion occurs in many cases, 20–40 d after pollination. On the basis of this study, self-compatibility is estimated to be between 0.35 and 1.39%. None of the self-compatibility genes have been cloned ( Aguirre et al., 2012 ).

The mean fraction of the genome with the ACP allele across all offspring is slightly higher than expected (∼25%) at 31.62% (SD = ±6.35%). Furthermore, although across most of the genome, the fraction of offspring with a heterozygous call is similar to that with an Alamo homozygous call, there are several regions with varying degrees of heterozygous enrichment. This enrichment is especially pronounced on chromosome 6a and on parts of 9a and 7b. Chromosomes 4a, 5b, 8a, and 8b show more modest levels of heterozygous enrichment. One of these regions is certainly due to artificial segregation distortion caused by selection against bar in the F1BC1 offspring. Chromosome 6a is the most likely candidate, because of the absence of any homozygous Alamo calls in any of the offspring near the telomere. Self-incompatibility loci may be responsible for the other regions showing significant segregation distortion.

Recombination Observed in All F1BC1 Offspring

Across all 83 F1BC1 individuals, a total of 1341 recombination events were observed for a total genetic distance of 1615.7 cM. This estimate reflects female recombination in the F1 and is close to previous estimates of the genetic distance of Alamo switchgrass: 1733 cM (female) ( Serba et al., 2013 ) and 1645 cM (male) ( Okada et al., 2010 ). Other estimates from non-Alamo cultivars are 1376 cM ( Okada et al., 2010 ), 1508 cM ( Serba et al., 2013 ), and 2085 cM ( Liu et al., 2012 ). The estimate of 1761.4 cM is likely to be slightly high, however, because of several sources of error within this dataset ( Hackett and Broadfoot, 2003 ).

The first source of error was false homozygosity as a result of low sequencing coverage, resulting in heterozygous regions incorrectly called as ACP or Alamo homozygous. This source of error is common within GBS datasets ( Elshire et al., 2011 Heffelfinger et al., 2014 ). Even though all spurious recombination events involving an ACP homozygous event were removed, spurious Alamo homozygous recombination events were not detectable as errors and were probably retained. On the basis of the number of ACP homozygous calls, there may be as many as ∼350 regions incorrectly called as homozygous Alamo.

A second source of error is the early state of the reference genome assembly and the highly repetitive allotetraploid nature of Alamo switchgrass itself. Unlike the markers used in the other switchgrass linkage studies ( Liu et al., 2012 Okada et al., 2010 Serba et al., 2013 ), the markers used in this study were ordered by position on a draft reference genome. This makes our estimate of genetic distance sensitive to errors in the reference genome. The current switchgrass reference genome consists of contigs scaffolded by synteny with the related species P. hallii genome (P. virgatum version 1.1,, accessed 6 Mar. 2015). The early state of the reference probably resulted in misplaced contigs and unannotated paralogous artifacts. Both of these events may contribute spurious recombination events.

In spite of these potential sources of error, the observed recombination rate remains well within the expected range. Recombination primarily occurs distal rather than proximal to the centromeres, as has been reported in other species ( Akhunov et al., 2003 Tanksley et al., 1992 ). A detailed map of recombination frequency across the genome of switchgrass would be desirable on the publication and release of the reference genome.

The Question of Nontransgenic F1BC1 Offspring

The ability to use transgenics to identify hybrids rapidly would accelerate the breeding process and presents clear advantages over extensive phenotyping and genotyping-based hybrid identification. These advantages may be offset by the increased regulatory difficulties presented by the use of transgenic herbicide resistance. Although the transgenic Alamo parent and the F1 hybrid are clearly GMOs, the F1BC1 population is more difficult to classify. There are two key aspects to the question of the F1BC1: whether the plants themselves contain transgenes and whether the absence of a transgenic sequence is sufficient to call them non-GMO when a transgenic sequence was present in the earlier generation.

Selection against bar-positive HbR F1BC1 offspring should produce a nontransgenic population. Multiple assays suggest that many of the offspring are hybrids with no evidence of transgenic DNA in them. There remains the possibility that gene silencing through methylation or structural polymorphism could disrupt bar resistance while maintaining all or part of the transgene. These individuals could be easily identified by a simple PCR assay for the presence of the transgene. The absence of transforming DNA in the genome of the F1BC1 individual, however, does not address whether these plants are considered to be non-GMO from a regulatory perspective. Ultimately, it is not likely to be a scientific question but a legal one as to whether this is sufficient to consider these offspring to be non-GMO.

Future Directions

Our results are potentially broadly applicable because they demonstrate the production and recovery of interspecific hybrids using a transgenic (GMO) selectable marker. This approach can be accomplished and applied to most major crop plants and may be useful for expanding breeding opportunities, including making rapid intervarietal crosses for conferring disease and other abiotic resistance traits, interspecific crosses for the combination of broad QTL characteristics, and creating close intergeneric crosses. Such a platform could serve as a basis for combining desirable characteristics by exploiting additive genetic variation and provide a more timely approach to developing novel lines in various crop species.

In addition to the creation of novel hybrid populations, phenotyping and QTL mapping on the extant ACP–Alamo hybrid demonstrated in this study remains a priority. Genotyping-by-sequencing identifies and types markers in a single experiment, thus providing not only a detailed picture of parental contribution and local ancestry but also a framework for downstream genomics applications. A key argument for switchgrass as a bioenergy crop is its ability to thrive in marginal environments ( Casler, 2012 Casler et al., 2011 Moser and Vogel, 1995 ), the introgression of abiotic resistance into commercial varieties would be desirable. Follow-up studies will combine phenotyping of the hybrid population with the marker framework from the GBS dataset to identify and introgress novel trait loci.


Sample collection, DNA extraction, and species identification

Larvae of both Helicoverpa species were collected from 13 different Brazilian locations by active searching on host plants. The sampling included the most important soybean, cotton, and maize-producing areas in Brazil during the 2015 crop season. Detailed information about the host plant, collection date, and geographic coordinates is presented in Table 1. Upon collection, samples were preserved in pure ethanol and stored at − 80 °C until further manipulation.

Total DNA was extracted from each specimen following an adapted protocol based on the CTAB method [42]. After the DNA extraction, species identification was confirmed through a PCR-RFLP method involving the digestion of a mitochondrial fragment of the COI mitochondrial gene (

511 bp). The PCR reactions were prepared using the COI-F02/R02 set of primers, and the reaction product cut with the enzyme BstZ17I [43].

Genotyping by sequencing library preparation

A total of 172 samples of Helicoverpa species (53 H. zea and 119 H. armigera) were selected to generate two GBS libraries containing

86 insects each [44]. Before the library-preparation step, the gDNA quality and quantity were assessed in each sample by visual inspection on agarose gel 1% (p/v), followed by determination of the concentration with a Qubit® 2.0 fluorometer (Life Technologies, Carlsbad, CA, USA). We normalized DNA at 20 ng/μl and digested with a single restriction enzyme, endonuclease (MSeI). Last, we used HiSeq 2500 to sequence the pair-end libraries, which were prepared and sequenced at the Molecular & Cellular Imaging Center Genomics Facility at the Ohio State University (Wooster, OH, USA). Raw fasta files of Illumina sequences were included in the SRA-NCBI repository (PRJNA615801).

Demultiplexing, SNP genotyping, and filtering strategy

Raw-sequence reads were demultiplexed using process-radtags implemented in STACKS 2.2 [45, 46]. Reads were trimmed at 85 bp after quality checking. In the first steps of the analysis, the program rescued RAD-tags from the reads, removed reads with uncalled bases, and then discarded reads with low-quality scores (i.e., −r, −c, and -q). Several attempts were made to map the GBS reads to the reference genome (PRJNA378437) however, due to the low percentage of the alignment (< 15%), we decided to use the non-reference-based method available in STACKS. The de-novo approach to assemble loci has been extensively used in non-model system and when there is no reference genome available this strategy is also more appropriate when the percentage of alignment is low. We ran the de-novo pipeline using all default parameters, closely following the method described by Anderson et al. [23]. After running preliminary tests, we concluded that the parameter combination used by Anderson et al. [23] provided the optimal yield regarding the number of markers retained and cluster resolution. Pair-end reads were integrated into a single-end locus, organized by loci in tsv2bam and assembled into contigs using gstacks. In the last step, we generated statistical summaries and Treemix analysis using the population module, allowing a minimum of 5% individuals required within groups and 100% between groups, excluding SNPs with less than 5% frequency, using one random SNP per RAD locus. Due to the great divergence between groups and the possibility of a high degree of variation within groups caused by hybridization, filtering parameters were relaxed, allowing an overall 20% presence of SNPs (i.e., to be included, a certain SNP must be shared with 20% of all individuals independently of their location). We conducted preliminary tests to maximize data retention while minimizing the rates of missing data in both species. The impact of hybridization varies in different parts of the genome, as previous studies have shown [47]. Thus, a different set of SNPs isolated from different genome regions can potentially give different values of estimated introgression. Our approach will help identify and limit potential biases of the different imputation methods [48, 49].

Nuclear admixture, introgression, and population structure

Species were identified using the collection information, including the host plant, morphological characters, and mtDNA genotyping, followed by the analysis with SNPs. Bayesian clustering methods implemented in STRUCTURE 2.3.4 and NewHybrids 1.1 were used to identify putative hybrids and to estimate proportions of nuclear admixture and patterns of introgression [50,51,52]. For parameter settings, we set the admixture model as the ancestry model and correlated frequencies as allele-frequency models. The posterior probability (q), representing the proportion of the genotypes originating from cluster categories (K), was later used to infer the putative degree of introgression in each sample. We used individual estimates of the introgression of insects collected at different locations as a dependent variable in models to explain possible causes of the observed differences.

First, we assumed K = 2, because two gene pools could potentially contribute to the genetic makeup of each sample. However, because strong evidence supports a history of multiple invasions of H. armigera [32], we also explored levels of substructure to detect the coexistence of different gene pools that may reflect the population structure of H. armigera in Brazil. We ran the STRUCTURE analysis for a range of K values (K = 1–10), and subsequently used the Evanno method implemented in STRUCTURE HARVESTER 0.6.93 to test for the most likely number of K [53]. We used only one SNP per RAD locus (−-write-random-snp) to minimize the effect of markers on linkage disequilibrium while performing long runs of the program to ensure convergence. We set the program to discard the first 150,000 steps (burn-in) and recorded 250,000 steps in each replicate (n = 10). Replicates of each K value were aligned and averaged in CLUMPP 1.1.2 [54] and visualized in DISTRUCT 1.1 [55].

The number of clusters and the level of hybridization were also investigated using non-model-based methods such as Principal Components Analysis (PCA) with the R package adegenet and ade4, as well as pairwise FST analysis [56, 57]. Additionally, we explored the phylogenetic relationships of insects collected at different locations, taking into account possible migration events, using the program Treemix [58]. Population divergence and migration events were estimated using bootstraps to calculate parameters in different scenarios by varying the number of migration events (m = 1–6). The most likely number of migration events was determined based on log-likelihood values and plotted residuals.

Association studies with landscape and climatic variables

Given that populations of the two species are now in sympatry, interbreeding may occur at different rates, possibly related to the presences of their main agricultural hosts of soybean, cotton, and maize. To investigate the ecological context of hybridization of H. armigera and H. zea in Brazil, we considered two groups of environmental variables in our analysis: climatic and land-use variables. For the climatic variables, we used elevation and 19 locality-specific bioclimatic variables from the WorldClim database, with a resolution of 30 arc-seconds [59]. To account for the significant number of correlated inputs, a principal component analysis (PCA) was carried out to constrain the climatic variables, converting many climatic variables into a smaller set of linear, uncorrelated values. The linear models used climatic variables from the sampling location, using the first two PC coordinates, since they carry the most significant portion of the variance, while the importance of climatic variables was assessed based on their contribution to the PC axes.

Land-use (i.e. landscape) variables were obtained classifying agricultural-landscape components such as soybean, maize, and cotton. Landscapes also contained other crops such as sugarcane, tree plantations, and orange orchards, as well as non-crop elements such as native forests, pastures, water, and urban sites that were also included in the classification maps. We quantified and characterized landscape attributes, using satellite images with a maximum of 2 months of differences in the collected data. This time window was necessary due to image quality, cloudy weather, and the availability of satellite images in public databases. Two databases were used for the satellite image collections, the Instituto Nacional de Pesquisas Espaciais (INPE) and the United States Geological Survey – Earth Explorer (USGS/Earth Explorer), which provide images from CBERS 4 and LANDSAT 8, respectively.

We manipulated and classified the different attributes from satellite images using ArcGIS 10.2.2. Briefly, a buffer with a 25-km radius from the collection point was created to delimit the study area. This radius was chosen based on the relative size of regional agricultural areas, in order to prevent overlap between landscapes, and also based on the insect’s flight capacity (up to 1000 km) [60]. Different signatures based on spectral responses can be linked to landscape attributes such as maize, soybean, and cotton, and therefore a supervised classification using the maximum-likelihood classification method was selected to separate classes within 25 km. The resulting classification was carefully revised and manually curated to minimize classification errors, using information from crop calendars and by contacting growers in the respective areas. Similar to climate variables, we conducted PCA analyses of the standardized proportion of each class, using only the first two PCAs to generate the models.

We constructed linear mixed-effects models using the ‘lmer’ function in the R package ‘lme4’ to estimate the relative importance of environmental factors for H. armigera and H. zea hybridization in Brazilian croplands [61]. The average introgression rates for each population, estimated based on SNP data, were used as our response variable. We inspected the residuals of each variable for distortion in homoscedasticity, and normality by visually checking the diagnostic plot and the residual. The proportion of introgression was arcsine square root-transformed to correct for normality. First, we tested the effect of species (fixed effect = species), controlling for the effect of populations (random effect = populations) to assess the asymmetry in gene flow between the two species. Then, we used the first PCA coordinates as independent variables in a full model for the hybridization detected in H. armigera. We tested the effect of landscape and climate, using these variables as fixed factors while controlling for the effect of the population (random = populations). We checked the significance of the models by evaluating χ 2 and p-values from the likelihood-ratio test of model comparisons. The most complex models included the interaction between landscape and climate, followed by models of isolated factors, and naïve models.


Crop relatives have been used for decades for breeding, in particular to transfer genes of resistance or tolerance to pests, diseases or abiotic stress to the cultivated species [1, 2]. Introgression breeding has been extensively used in the genetic improvement of some of the most important Solanaceae crops, like potato (Solanum tuberosum L.) or tomato (Solanum lycopersicum L.). Thus, up to twelve traits have been introgressed in potato from related species like S. demissum, S. stoloniferum, S. chacoense, S. acaule, S. vernei or S. spegazzinii [3] and many more have been transferred to tomato from their wild relatives like S. peruvianum, S. cheesmanii, S. pennellii or S. chilense [4]. However, breeding programmes in the economically important common Capsicum peppers (Capsicum annuum L.) have made little use of related species for breeding as reviewed by Mongkolporn and Taylor [5]. This limitation has been mainly due to the presence of different pre-zygotic barriers which avoid fertilization (e.g. pollen-pistil incompatibilities) and/or post-zygotic barriers, which prevent the achievement of fertile hybrids, e.g. embryo/endosperm abortion, hybrid weakness or sterility [6, 7].

In this sense, C. annuum is related to about other 30 Capsicum species, of which four are also cultivated, C. baccatum L., C. chinense Jacq., C. frutescens L. and C. pubescens R. & P. [8]. By one hand, C. chinense and C. frutescens cultivars have economic importance in America, Africa, and Asia and both are phylogenetically close to C. annuum. In fact, these species make up the annuum-chinense-frutescens complex (or annuum complex), characterized by white flowers and yellow seeds [9]. On the other hand, C. pubescens and C. baccatum represent separate taxons from the annuum complex and, although they have been widely grown in the Andean region and Brazil for millennia, they are very rare outside this area nowadays [9, 10]. The former, mainly known as rocoto (with purple flowers and black rough seeds), is the least economically important [9]. Due to prezygotic barriers which prevent the growth of the pollen tube through the style, and possible postzygotic barriers, it does not cross with any of the other four species [9, 11]. The latter, commonly known as ají (with white flowers and yellow spots), has showed an extremely low/nil cross compatibility with C. annuum [6], although it has been reported as source of variation for a range of traits with potential interest for the genetic improvement of this species. These include resistances to several diseases such as: anthracnose (Colletotrichum spp.), powdery mildew (Leveillula taurica), Rhizoctonia root rot (Rizhoctonia solani), Verticillium wilt (Verticilium dahliae) and bacterial wilt (Ralstonia solanacearum syn. Pseudomonas solanacearum), viruses like PYMV and TSWV or even new flavours [12–21]. However, successful wide hybridization attempts to introgress these traits in C. annuum have been scarce [12, 22].

Postzygotic barriers have been suggested as the main cause of cross compatibility problems between both species, specifically embryo/endosperm abortion and hybrid sterility [23, 24]. In most plant species, the first barrier is caused by abnormal cell division of the zygote or slow endosperm development, which causes an incompatibility with embryo growth [25], while the second is due to a range of factors such as diverged genes, karyotypic changes, gene transposition or gene loss, sequence divergence or dosage imbalance, among others [26].

An alternative to overcome these barriers, known as genetic bridge, is based on the use of species (bridges) phylogenetically close to the two species affected by crossability barriers. The bridge species is used to obtain hybrids with one of the target species, and subsequently these hybrids are crossed to the other target species [27]. Thus, C. chinense and C. frutescens might play this role for wide hybridization between C. annuum and C. baccatum as previously suggested by Pickersgill [28].

Another strategy for wide hybridization between C. annuum and C. baccatum is the in vitro rescue of immature interspecific embryos or embryo rescue before abortion occurs [27]. This approach is technically more complex as it requires embryo excision and in vitro culture. Also, the stage at which embryo abortion occurs after hybridization may depend on the specific genotypes involved in the cross. Thus, for example within Solanaceae, while some authors could rescue interspecific embryos at the latest immature stages [29], there are also examples on which embryos had to be rescued at the earliest stages [30, 31]. However, the earlier the stage at which embryo rescue is done, the more difficult is the procedure and the lower the efficiency [32].

Furthermore, even though hybrid materials could be achieved, hybrid sterility must be also considered as an important postzygotic barrier. Full sterility or different degrees of fertility of the interspecific hybrids may vary depending on the parent genotypes. In extreme cases, when sterility is complete due to the lack of chromosome pairing during meiosis, fertility may be restored by poliploidization, enabling pairing of homologous chromosomes in the allopolyploid hybrid [33].

Unfortunately, studies on wide hybridization between C. baccatum and C. annuum and the overcoming of their compatibility barriers are very scarce [29, 34], especially regarding the range of diversity encompassed in materials used for crosses. Consequently, there is a lack of detailed practical information about the breeding process and levels of sexual compatibility among species, which might also depend on the genotypes involved or the direction of the crosses, among other factors.

Therefore, the development of approaches which allow overcoming these barriers, assessing all the difficulties which can appear during their application, will provide breeders with useful tools, practical information and a comprehensive perspective for the introgression of genes of interest from C. baccatum to C. annuum. Moreover, the use of a wide genetic diversity will contribute to offer a more complete view of these barriers.

The aim of this work was to compare comprehensively two approaches for the achievement of wide hybridization between C. annuum and C. baccatum: i) genetic bridge using C. chinense and C. frutescens as bridge species and, ii) direct hybridization between C. annuum and C. baccatum in combination with embryo rescue. In both strategies full diallel interspecific crosses were planned and a range of genetically different genotypes were used. The effects of the direction of the crosses, cross compatibility at different levels among the species involved and hybrid viability and fertility are also discussed.

Introgression: Brower’s criticisms. Part II.

1) Mallet redefined species to allow hybridization between them! p. 4: The relevance of one third of these, hybrids between H. himera and H. erato, is thrown into question by Brower on the basis that himera was only elevated to a species by Mallet himself, after designing a new species concept to allow hybridization.

An email to Brower queried this criticism: surely himera is separate from erato under most peoples’ ideas of species? His reply, on 14 Dec 2012: “No dispute that the himera x erato specimens are hybrids, or even that the two are different species under your (or my) species concept.” So Brower agrees that these are hybrids between good species, in spite of the snarky suggestion he puts forward.

As this is one of the first points made by Brower in his 2012 critique, one immediately wonders how well the rest of his criticisms will hold up. As we shall see, not very well.

2) Some butterfly collectors were dishonest! p. 4: Older specimens in the database often have poor locality information, and may have been collected by dishonest butterfly-mad people like Anton Fassl, who stole butterfly specimens from the Vienna Museum in 1906. [This criticism could be divided into two, based (i) on faulty locality data, and based (ii) on dishonesty of collectors].

Are hybrid specimens with faulty or approximate locality data thereby rendered non-hybrids? I don’t think so.

Are hybrid specimens collected, or maybe stolen by dishonest butterfly enthusiasts thereby made less hybrid? I don’t think so.

Brower produces no evidence at all that these older museum hybrids were fraudulently produced in captivity. The example of kleptomania by a butterfly collector in Vienna (as documented by Brower) is no more relevant to the finding of hybridization in the wild than is the dishonesty of bankers or politicians.

A number of key hybrids in the dataset were named as separate species. They were not even recognized as hybrids before the systematics of Heliconius was properly sorted out in the 1960s and 1970s. If these earlier entomologists didn’t know that these rare specimens were hybrids, how could they have known that it was possible to produce such specimens via hybridization in captivity?

3) The rest of the hybrids were also artificially produced! p. 4: For newer specimens (post 1960, according to Brower), “it is not unreasonable to suspect that many of these oddities could have been captive-reared for the butterfly trade, despite [Mallet et al.‘s] assurances to the contrary”.

To obtain hybrids between species in captivity, one must have a number of species flying around as adults in suitable multi-generation breeding facilities for long periods of time. Interspecific hybrids occur in small cultures only extremely rarely.

No such suitable enclosures existed, except in scientific establishments (and there only since the 1960s), prior to the 1980s. I believe the first commercial butterfly house to open world-wide was the London Butterfly House at Syon Park, London, in 1981. It was the brainchild of property magnate Clive Farrell (see Wikipedia for details). Hybrid specimens occurring after the mid-1980s have been reared in captivity. However, this potential problem was clearly outlined in the paper by Mallet et al. (2007), who selected only specimens with bona fide credentials. A number of specimens were excluded as they were clearly laboratory-reared hybrids (See supplementary information in Mallet et al. 2007).

J. Mallet and M. Linares have personally met some of the Colombian collectors of these hybrids (Mallet et al. 2007). Many of the hybrids in Ernesto Schmidt-Mumm’s collection, for example, were collected by Ernesto himself before he died, as he personally described to us. Ernesto never reared any Lepidoptera at all — they were all caught on the wing. At the time he was actively collecting, he was the only butterfly expert capable of distinguishing a hybrid, given that knowledge of the taxonomy of Heliconius was in its infancy in Colombia at the time. Keith Brown also records personally collecting a cydno-melpomene hybrid at Victoria, Caldas, Colombia, where Dr. Schmidt-Mumm had his ranch (Brown & Mielke 1972 p. 10).

A number of hybrids have in fact been collected by Heliconius biologists in the field, in spite of Brower’s contrary assertion. See comments on an ethilla x melpomene hybrid from Peru in (4) below. Mathieu Joron (pers. comm.) and Chris Jiggins (pers. comm.) have also collected hybrids in the wild recently, but only Brower felt that their existence was controversial, so no systematic investigations were felt necessary. I figure below one of these recent hybrids, a melpomene x numata hybrid.

Heliconius numata x H. melpomene hybrid collected 2007 by Owen McMillan at Nueva Jordania, San Martín, Peru. Photo courtesy of Mathieu Joron

The idea that many or most of these hybrids, collected by so many entomologists in so many different localities, not just in Colombia, but in a variety of locations from Mexico to the extreme South of Brazil, were produced fraudulently is laughable.

Even if some interspecific hybrids documented by Mallet et al. (2007) post-1980 were produced in captivity, this cannot apply to hybrids before this date.

4) Therefore, no hybrids ever occurred in the wild at all! p. 4: Conclusion to this section “In sum, so many of the Mallet et al. [(2007)] records are dubious, at least for H. cydno-H. melpomene ‘hybrids,’ that this dataset must be discounted as convincing evidence of widespread natural hybridization between those species, or more generally for ‘the species boundary as a continuum.’ ”

Even if some of the hybrids were produced in insectaries, this conclusion is absurd. As shown above, Brower’s conclusion is supported by no evidence whatsoever, but is merely a sort of paranoid conspiracy theory.

In the above-mentioned email (dated 14 Dec 2012), Brower agrees to the veracity of a fairly distant hybrid between Heliconius ethilla and H. melpomene netted by field biologists in Peru, and confirmed as an F1 hybrid using molecular genetic tools (Dasmahapatra et al. 2007). However, he doubts this hybrid is fertile. This particular hybrid is similar to another specimen from Colombia named as a new species, Heliconius hippola, by Hewitson in 1867. A number of similar hybrids between Heliconius ethilla and H. melpomene occur elsewhere, and as is usual in the melpomene-silvaniform group of Heliconius, backcross phenotypes also exist from wild collections and form a substantial fraction of existing hybrid specimens. Even if these other supposed wild caught specimens were artificially created in insectaries (for which there is no evidence), these back-cross specimens indicate fertility of F1 individuals.

Furthermore, there is evidence from insectary crosses as distant as this that backcrossing takes place and can be used to transfer colour patterns across the entire melpomene-silvaniform group. For a series of crosses involving Heliconius hecale x atthis x melpomene (with some x cydno), see Jean-Pierre Vesco’s photos in Mallet et al. (2007). See also other examples in Gilbert (2003).

For the sake of argument, let us suppose that ALL the Mallet et al. (2007) wild-caught hybrids were artificially created in captivity by fraudulent dealers. (Again, I reiterate that the idea is laughable, and only Brower would seem to argue for this). Then even so, these specimens still provide excellent evidence for the possibility of hybridization and backcrossing in the field.

B) ‘Allelic sharing’ or mixed genealogies of individually sequenced loci

5) Allele sharing gives no evidence of introgression! p. 4: “… Explained at least as well by retention of ancestral polymorphism as by recent introgressive hybridization”. “Thus, oft-cited claims of ongoing, evolutionarily significant gene flow between H. cydno and H. melpomene should be viewed with circumspection.”

All of us are well aware of the possibility that ancestral polymorphism may be confused with true genealogical reticulation due to gene flow. This is well recognized in the coalescent-based IM (isolation and migration) algorithm proposed by Jody Hey, Rasmus Nielsen and others to investigate evidence for DNA sequence flow among species. A number of studies on Heliconius have used this methodology on gene fragments revealed by PCR and Sanger sequencing (Bull et al. 2006, Kronforst et al. 2006, Mavárez et al. 2006, Kronforst 2008). These studies have concluded that some, but not all loci do indeed flow among closely related species.

The interpretation of such data is fraught. Apparent genealogical reticulation may result from ancestral polymorphisms, rather than gene flow, and deviations from the very strict assumptions used in the algorithm may indeed give spurious results.

Nonetheless, an inference of gene flow using IM does provide some evidence for gene flow. It is certainly better to use an IM algorithm which acknowledges the possibility that allele sharing may be due to ancestral polymorphism as well as gene flow, than to ignore the possibility altogether!

A typical argument against an IM-based inference of gene flow might be that the locus Mpi, to give a concrete example, is under balancing selection. Therefore related species may inherit multiple balanced alleles from a common ancestor. If so, is there evidence that this is taking place? In the loci studied by Bull et al., Kronforst et al., and Mavárez et al., and inferred to be flowing between species to date, no evidence for such balancing selection is forthcoming. Brower produces no new evidence to refute either the neutrality of variation at these loci, or the inference of gene flow.

C) Heliconius heurippa as a homoploid hybrid species, with red markings coming from H. melpomene, and yellow markings from H. cydno or a related species

6) Allele sharing gives no evidence of introgression, again! p. 5: ” ‘Support’ from the ambiguous gene genealogies mentioned earlier” is “weak.”

See above, (5). Weak evidence, maybe, but it is evidence.

7) If the allelic sharing of a new species with its parents isn’t exactly 50%, we must conclude it is not hybrid speciation! p. 5: ” ‘Classic’ homoploid hybrids are expected to exhibit mosaic genomes composed of blocks of DNA from the two parental species.” “Mallet and Jiggins et al., recognizing that H. heurippa’s widespread genomic affinity to H. cydno does not fit that model, have relaxed their concepts of HHS.”

Brower again argues here that certain authors, in the course of proposing a novel hypothesis, have tailored their concepts and definitions to fit the data, and then claimed that the hypothesis is proved. This, in Brower’s view, apparently rules out the inferences made by these authors.

Whatever the truth of this allegation about concepts, all parties must surely agree that what was proposed by Mavárez et al. was not that the genomic constitution of Heliconius heurippa consists of 50% melpomene and 50% cydno. Instead, they suggested that H. heurippa is a cydno-related form with some evidence of introgression from melpomene. The introgressed fraction of the genome includes those regions which determine red patches in the forewing band. Definitions, and the precise distinction between “hybrid speciation” and “hybrid trait speciation” are unimportant in this debate.

Keith Brown, after visiting localities near Villavicencio, Colombia, was the first modern author to recognize H. heurippa as a separate species, and to suggest hybrid origin involving cydno and melpomene. It flies alongside H. melpomene, and Brown was able thereby to refute Emsley’s earlier suggestion that heurippa merely represented an infraspecific hybrid within H. melpomene (Brown & Mielke 1972: 10).

Brower’s argument in this case, as in so many others, is simply a red herring.

8) Red patterns ancestral in H. timareta explain the red markings of H. heurippa, not hybridization! p. 5: H. heurippa is somewhat closer genomically to populations currently designated as H. timareta (many races of which have red markings) than to H. cydno (none of which have red markings) (see Nadeau et al. 2012), and therefore red markings may already be present in the cydno/timareta-like ancestor of H. heurippa.

The whole cydno/timareta group is monophyletic (Heliconius Genome Consortium 2012). However, a monophyletic group of populations, mainly West of the Andes and more Northern, is currently referred to as H. cydno. The other monophyletic group includes H. heurippa and populations Southwards, many of which do indeed have red markings, on the eastern and more Southern slopes of the Andes, which are traditionally referred to as H. timareta.

The nearest populations of H. timareta/cydno to the South of H. heurippa (which occurs near Villavicencio, Meta, Colombia) are the “H. cydno cognate” (with red basal spots on the underside, this is certainly a timareta, and is from Río Pato, Colombia) and the rayed H. t. florencia (occurring at Florencia, Caquetá, Colombia) (Giraldo et al. 2008, Fig. 6). The nearest red-banded form of H. timareta sensu lato is distributed farther South, and was named by Brower himself as a separate species H. tristero (Mocoa, Putumayo, Colombia). Since the adjacent races to the South of H. heurippa, the Río Pato taxon and H. t. florencia have yellow, not red-banded forewings, it is not clear how Brower’s argument would explain the origin of H. heurippa’s red band.

9) Even if hybridization was a likely origin of H. heurippa, there was no reason for it to happen, since H. heurippa is non-mimetic! p. 5: H. heurippa is not a mimic therefore it had no mimicry reason to acquire red patterns from H. melpomene.

This argument is spurious, another red herring. However H. heurippa evolved, it would face the same problem of being probably non-mimetic.

The empirical evidence that H. heurippa acquired its colour pattern via hybridization cannot be trumped by a theoretical argument based on mimicry that it should not occur.

H. heurippa does have a passing resemblance to the co-occuring local mimicry ring consisting of Heliconius numata messene, H. hecale ithaca, Melinaea marsaeus messenina, Melinaea isocomma isocomma, Mechanitis (mazaeus) messenoides, and a number of other species. This may have helped in the establishment of the colour pattern via a weak mimicry effect.

D) Genomic evidence that colour pattern homoplasy in Heliconius melpomene, timareta and elevatus (and also perhaps H. besckei) is explained by introgression

10) Too much data! p. 6: “It is easy to be intimidated by the overwhelming quantity of data and elaborate analyses in genomics publications.” In 2011 Brower had the same problem with overabundance of data: “Salazar et al. (2010) is an example of several disturbing emergent trends in genomic-era publications. The paper alludes to analysis of an enormous amount of data: nearly 45 kb of sequence from 30 individuals, representing nearly 3,000 individual GenBank accessions. The sheer quantity of data … makes results difficult to evaluate critically.”

However, Salazar’s (2010) phenomenally large dataset was achieved via PCR and Sanger sequencing, as used by Brower himself.

In any case, increased amounts of data must surely give more power to those who would like to reject incorrect hypotheses, rather than being a cause for concern as a “disturbing emergent trend.” Brower’s argument seems silly.

11) Homoplasy is the new parsimony! p.5-6: Some homoplasious colour patterns in the genus Heliconius have evolved via selection for mimicry, as opposed to being maintained by selection after hybrid transfer from another species. De novo truly homoplasious mimicry evolution is therefore considered by Brower to be the most parsimonious mode of origin for all similar patterns.

A basic principle of parsimony seems violated by Brower’s argument here. The principle of parsimony is that ad hoc assumptions (such as novel evolution of mimetic phenotypes) should be minimized. It doesn’t mean that parallel evolution never occurs via selection for mimicry. For example, almost certainly the same colour patterns were invented twice in H. erato and H. melpomene. The species that achieved the parallelism was probably H. melpomene, the mimic of H. erato. No hybrids between species as distant as erato and melpomene are known in Heliconius.

But when the same pattern co-occurs in related species known to hybridize and backcross in the wild and in captivity, as here, one suspects that transfer of patterns is possible and indeed likely. The principle of parsimony doesn’t prove transfer, but suggests that the hypothesis of transfer would obviate the need for parallel evolution of the same nucleotide sites.

12) All next-gen sequencing data is suspect! p. 6: “There are issues both of data quality and analytical rigour that raise concerns” in next-gen sequencing data, in general [and therefore by implication as applied to Heliconius as well].

Brower’s angst about newer genomic data in general does not seem a reasonable criticism of the question at hand. See also (10) above.

13) We must ignore any evidence from partial data! p. 6: “Unfortunately, these data matrices contain an enormous amount of missing data”.

The reason for the apparent “missing data” is that very strict filters were applied so as to reject low coverage or poor quality alignments to ensure clean data. All of the original Illumina reads are available for those who would prefer to work with original data.

But after this strict alignment procedure, surely it’s the filtered data that do support these findings that are more important? Given there are now thousands of times more data than we had before, aren’t these next-gen sequencing data thousands of times more powerful than those from PCR and Sanger sequencing?

Inferences must always be made with the data available, even if these form only a sample of the genomic information.

14) I couldn’t see any pattern in their data! p. 6: “Support for grouping of the taxa by wing pattern reveals that they are rife with ambiguity: there is not a single fixed character state difference”

Brower should look at the data more carefully. The ABBA-BABA peaks shown in Fig. 4b,c of Heliconius Genome Consortium (2012) are based entirely on fixed sites. In any case, although some sites may not follow the major phylogenetic patterns shown in Fig. 4d, Brower must be aware that conflicting signals (“rife with ambiguity”, as he calls it here) are entirely normal in phylogenetics, and do not negate results of a well supported analysis. In Fig. 4d, bootstrap support in the centre of the colour-pattern genomic region, which suggests hybrid transfer, is 100% for both postman and rayed colour patterns.

In 2011, Brower argued: “It … may not be possible to obtain a clear understanding of the evolution of mimetic phenotypes in these butterflies until we are able to examine gene genealogies for the genes that are responsible for the wing pattern elements themselves. I predict that the allele producing a red band on the forewing of H. heurippa will not be homologous (IBD) to that of sympatric H. melpomene melpomene, a pattern that would lay to rest the H. heurippa [hybrid speciation] hypothesis.”

We’ve now disproved Brower’s (2011) prediction with two different colour pattern loci in two races of H. timareta, and also in H. elevatus (Heliconius Genome Consortium 2012). There is similar evidence from the red forewing band locus in H. heurippa itself (Pardo-Diaz et al. 2012). But Brower is not satisfied, and is now apparently claiming that these regions must, in effect, be completely identical and fixed at all divergent bases, and not just genealogically identical by descent (“IBD”), to disprove his hypothesis of independent origin. Brower’s view of the purity of species is proving to be a moving target, which can never be falsified.

15) I didn’t understand what kind of phylogenetic analysis they used! p. 6: “Published trees do not make clear how many or what kind of characters support these patterns, nor what models were used to produce the trees.”

Phylogeneticists often criticize other phylogeneticists’ conclusions on the basis of methodology, especially if different kinds of analyses give different results. Here Brower criticizes our methods, and tries, but fails to show that any other method would give a different result. In fact, he agrees that “phylogenetic analyses of various walk segments do yield the published topologies [of the Heliconius Genome Consortium]”.

Neighbor-joining and maximum likelihood methods based on the nucleotide data were clearly indicated in the methods sections of the Heliconius Genome Consortium (2012) paper. The model used in maximum likelihood analysis was standard — GTR + gamma — a complex model useful due to the large amount of data available for each tree.

16) I don’t understand this other analysis either! p. 6: Use of D-statistics to infer introgression “assumes neutrality”, which is not true in regions affecting colour patterns.

D-statistics were used only in the genome-wide analysis of Fig. 3b, where it can be assumed that most of the variable sites were approximately neutral. In any case, it is hard to imagine a model of selection bias towards variable sites that are preferentially shared between the local races of timareta and melpomene in non-colour pattern regions, when this contravenes the species-wide phylogeny, unless those sites have actually been transferred via gene flow locally. Thus, these data provide genome-wide evidence of polymorphic allele sharing between local races of the two species.

Given that this is the case, it becomes an unnecessary hypothesis to argue that the fixed ABBA-BABA sites shared within the colour pattern regions (in Fig. 4b,c) were not transferred with the rest. Given that they will be shared during occasional hybridization events, an influx of adaptive variation with potential value for mimicry must have taken place.

17) Allele sharing gives no evidence of introgression, again, and yet again!! p. 6: “A similar problem [to that in (15)] occurs with use of the statistical program IM and linkage-disequilibrium tests to infer interspecific gene flow.” “When the traits of interest are under selection, as genes responsible for wing patterns manifestly are, then inferences drawn from coalescent-based methods for inferring gene flow that assume neutrality may be unreliable”

See also (5) above. Genomic evidence for allelic sharing in Heliconius Genome Consortium (2012) does not depend on coalescent-based estimates such as the use of the program IM or analysis of linkage disequilibria. The ABBA-BABA tests and D-statistics used are essentially parsimony-based tests of site patterns, and do not employ strict assumptions of neutrality, unlike coalescent-based methods such as IM.

E) “Paradigms and paradoxes:”This section of Brower’s (2012) paper depends on his strong prior bias against introgression. Brower explains that he cannot imagine how introgression might occur under his interpretations of mimicry, homoploid hybrid speciation, and natural selection for mimicry. This, in his view, militates against the empirical, genetic data which suggests it does.

18) Rare events are not possible! p. 6: “How can wing mimetic pattern alleles flow from one species to another (and apparently be the only gene regions that do so), when it has been shown that wing patterns are perhaps the key adaptations responsible for intrinsic maintenance of species boundaries by mate choice?”

First of all, colour pattern regions are not the only genomic regions to show exchange. See Fig. 3 in Heliconius Genome Consortium (2012) and point (16) above, which Brower apparently misunderstands.

Wing patterns are indeed involved in mate choice. (However, the colour pattern itself may not be the strongest barrier instead other traits such as behaviour and pheromones may be more important). In spite of rather strong barriers to hybridization, hybrids certainly do occur, both in the wild and in captivity. Although hybrids suffer many problems, including female sterility and mimetic disadvantages, male hybrids can and do backcross both in the wild and in captivity. Hybrids effectively bridge the gap between species, and the mating barriers to backcrossing are much weaker than in the original hybrid mating (Naisbit et al. 2001). Colour pattern genes can thus readily be crossed and backcrossed into rather distant species (Gilbert 2003, Mallet et al. 2007).

19) Selection for mimicry disproves speciation! p. 6: “How can wing pattern alleles spread from one species to another when such introgression does not occur across intraspecific hybrid zones in geographically differentiated species in which there are no barriers to interracial hybridization?”

Actually, abundant hybridization and introgression do occur across intraspecific hybrid zones. The whole Amazon basin, for instance is a mass of such clinal polymorphism in most Heliconius species found there (Rosser et al. 2012).

It is of course true that disfavoured colour patterns are usually selected against. Nonetheless, novel patterns do sometimes become established, with the genus Heliconius perhaps holding the record for the evolution of multiple novel warning colour patterns (Mallet 2010).

20) Non-mimetic phenotypes can never establish, and therefore non-mimetic species cannot arise via hybridization either! p. 6: “How can ‘non-mimetic’ phenotypes arise and become fixed as a result of interspecific gene flow when there is a strong selective advantage to phenotypic conformity due to Müllerian mimicry?”

We’d like to know the answer to this as well!

Heliconius timareta timareta, from near Puyo, Ecuador (photo courtesy of Bernard D’Abrera).

However, non-mimetic patterns do occasionally become established in Heliconius including in the polymorphic H. timareta timareta in Ecuador (see photos above). The argument by Brower applies to any mode of origin of the novel Heliconius heurippa colour pattern, whether via hybridization or not. This particular argument of Brower’s therefore does not single out introgression as a less likely means of heurippa‘s evolution. Another red herring.

Note that this is merely a repeat of Brower’s argument (9) above.

21) Recombination is not possible! p. 6: Non-mimetic phenotypes like Heliconius heurippa are not “explained by the introgression hypothesis: if selection for mimicry drives the process of introgression, then phenotypes resulting from introgressed alleles should be identical to those of the species from which they came.”

This would be true if mimicry was driving the fixation of introgressed alleles. However, as Brower points out, in this case, H. heurippa is probably non-mimetic, and therefore mimicry would not explain the establishment of its novel pattern. So Brower’s argument is a red herring, as the establishment of the novel colour pattern in H. heurippa has never been dependent on a hypothesis of perfect mimicry.

The possibility of introgression depends on hybridization between species, not on mimicry. The genetic evidence for introgression of colour patterns in Heliconius is independent of any mimicry hypothesis. The establishment of any introgressed colour pattern alleles may sometimes be helped by mimicry, but could also be dependent on other factors for hybrid taxa such as H. heurippa or H. timareta timareta, such as mate choice.

As already noted (see (9) and (20) above), it is possible that the establishment of the hybrid colour pattern in H. heurippa was helped along by approximate similarity of the large Melinaea, Mechanitis, and some other Heliconius with which it co-occurs.

22) Introgression is unlikely because, if it occurred, all Heliconius butterflies would share the same pattern! p. 6: “If wing patterns are promiscuously shared across species boundaries, then why has this not led to fixation of a single, shared aposematic pattern, which would represent a stable, selectively advantageous global optimum for all Heliconius butterflies?”

Once again, a very interesting question, and one which is a recurring theme in our research.

However, this argument is relevant to the evolution of novel colour patterns, whether or not hybridization was involved, and is therefore another red herring in this context. See also (9), (20), and (21) above.

23) Absence in small samples of allelic sharing indicates absence of introgression! p. 6-7: “The absence of introgressed neutral loci (e.g. microsatellites) between H. heurippa and H. melpomene does not fit the pattern of shared genetic material expected if significant hybridization had taken place between those two species or the ‘parental’ H. cydno and H. melpomene populations.”

Note, this argument appears to be the precise converse of (5), (6) and (17)! If these microsatellite loci did show allelic sharing, then presumably the “ancestral polymorphism” argument of (5), (6) and (17) would instead be deployed by Brower against the idea that allelic sharing suggests introgression. Brower’s arguments are constructed so as to be irrefutable whatever the results!

In any case, Brower does not apparently understand the use of Bayesian STRUCTURE analysis in cluster assignments, as used in Mavárez et al. (2006). Although the full microsatellite data may cluster individuals into separate taxa, this does not preclude the existence of multiple alleles at those same microsatellite loci shared between those taxa via introgression. All that is required to obtain STRUCTURE evidence for different clusters is that the allele frequencies of at least some of the loci should be different in each cluster.

24) De novo homoplasious evolution of mimetic colour patterns in different populations is more likely than introgression from another species! p.7: “Further, the biogeographical pattern, with six or seven different H. cydno cognates east of the Andes exhibiting at least three different H. melpomene-like mimetic phenotypes, implies that the extremely unusual genetic phenomena proposed to produce them must have occurred independently in multiple populations.”

Given that hybridization and introgression occur in many different locations where H. cydno or H. timareta overlap with H. melpomene, it would hardly be surprising if occasional hybrid transfers resulted in the promiscuous sharing of colour pattern loci that we observe in the genomic data.

Brower remains unconvinced by the abundant specimens showing evidence of hybridization and introgression among species of Heliconius in nature, and in captivity. Uniquely, for a systematic biologist, Brower argues that data from museum specimens collected over hundreds of years are not valid.

Brower displays a strong preference for what he was perhaps taught as an undergraduate, that species are reproductively isolated, and that species never exchange genes after separation. This belief leads him to discount all evidence ever produced to suggest hybridization and gene flow among Heliconius species. One correspondent wrote to me in an email “it’s almost entertaining to read in fact, you find yourself wondering with curiosity what he will find to discredit [in] the topic outlined in the title of each paragraph!”

Brower displays such a fanatic determination to dismiss this evidence that it leads him to make many errors of logic in interpreting the genetic and genomic data recently revealed by the Heliconius community.

It is unlikely that anyone will ever persuade Brower that he is wrong, and maybe it’s best not to try. Nonetheless, it seemed to this author important to document for a wider audience, at least informally, just how mistaken all of Brower’s arguments are.

To dismiss the simple finding of hybridization and introgression, Brower has to use so many different arguments against different aspects of the extensive data that it is impossible to answer them all in a reasonable-length article. However, the multiplicity of arguments he uses itself tells against his theme. One wonders why Brower doesn’t come to the much simpler, alternative conclusion that explains all of the data: that hybridization and introgression do indeed occur.

Brower’s articles and views on this topic are in my view becoming unreasonable.

Brower AVZ. 2011. Genetica 138: 589.
Brower AVZ. 2012. Proc R Soc B 280 online.
Brown KS, Mielke OHH. 1972. Zoologica 57: 1.
Bull V et al. 2006. BMC Biol 4: 11.
Gilbert LE. 2003. In: Boggs CL, ed. Ecology and Evolution Taking Flight: Butterflies as Model Systems. Chicago: University of Chicago Press, p. 281.
Giraldo N et al. 2008. BMC Evol Biol 8: 324.
Heliconius Genome Consortium. 2012. Nature 487: 94.
Kronforst MR. 2008. BMC Evol Biol 8: 98.
Kronforst MR et al. 2006. Evolution 60: 1254.
Mallet J. 2010. Ecol Ent 35 (Suppl. 1): 90.
Mallet J et al. 2007. BMC Evol Biol 7: 28.
Mavárez J et al. 2006. Nature 441: 868.
Nadeau NJ et al. 2012. Molec Ecol online.
Naisbit R et al. 2001. Proc R Soc B 268: 1849.
Pardo-Díaz, C. et al. 2012. PLoS Genet 8: e1002752.
Rosser N et al. 2012. Biol J Linn Soc 105: 479.
Salazar C et al. 2010. PLoS Genet 6: e1000930.

Hybridization and introgression between bread wheat and wild and weedy relatives in North America.

NATURAL HYBRIDIZATION involves successful mating in nature between individuals of two populations, or groups of populations, which are distinguishable on the basis of one or more heritable characters (Harrison, 1990 Arnold, 1997). This process thus includes cases involving crosses between individuals considered to be conspecific, but not crosses between individuals from the same gene pool that possess alternate states of a polymorphic character. The phrase "successful mating" means the production of viable [F.sub.1] progeny that possess some level of fertility (Arnold, 1997). If hybrid progenies backcross with their parents in subsequent generations, the phenomenon is termed introgression.

Introgression thus refers to "the infiltration of germplasm from one species into another through repeated backcrossing of the hybrids to the parental species" (Anderson and Hubricht, 1938 Arnold 1997). With each successive backcross the hybrid-derived plants progressively accumulate the traits of the backcrossing parent(s). The formation of fertile hybrids and backcross individuals is necessary for a successful introgression between any two taxa. Accordingly, genetic relatedness, ploidy level, and direction of hybridization between hybridizing taxa play a crucial role during the process of introgression (Rieseberg and Wendel, 1993).

A number of angiosperm taxa are believed to be derived from hybridization or introgression between closely related taxa (Clausen et al., 1945 Grant, 1981 Soltis and Soltis, 1993 Rieseberg and Carney, 1998), and even in the extant floras, the occurrence of hybridization or introgression is reported to be widespread (Knobloch, 1972 Stace, 1987 Rieseberg and Wendel, 1993 Peterson et al., 2002). For example, Knobloch (1972) identified thousands of interspecific and intergeneric natural hybridization events between species of angiosperms. Similarly, 7% of the 1264 introduced plant species to the British Isles are known to be involved in hybridization with either native species or other alien species (Stace, 1991 Abbott, 1992). A permanent transfer of genes between hybridizing taxa has been documented in 65 of 165 suggested instances of introgression (Rieseberg and Wendel, 1993). The majority of the 165 documented instances of introgression were reported for angiosperm species including dicots, monocots, and nearly all growth forms, pollination syndromes, and mating systems.

Introgression has played a role in structuring the genetic diversity of species (Wendel et al., 1989 Rieseberg, 1997), in the origin of new adaptations (Ricseberg, 1991), in the transfer of adaptations between species (Heiser, 1973 Rieseberg et al., 2003), in the formation of new ecotypes (Levin, 1967 Abbott, 1992 Rieseberg, 1997) or species (Arnold et al., 1991 Soltis and Soltis, 1999), and in the evolution of invasiveness (Anttila et al., 1998 Ellstrand and Schierenbeck, 2000).

Plant breeders artificially introgress traits from wild relatives into crop plants to develop new cultivars. Earlier, such artificial introgressions were mainly restricted to gene transfer between species that were cross compatible (Maan, 1987), as cross incompatibility problems prevented production of viable seeds among hybrids between unrelated species. However, modern biotechnology techniques, which do not depend on the sexual transfer that occurs in nature, overcome this sexual barrier by making gent transfer possible between unrelated species (Bajaj, 1990 Lorz et al., 1997). By uncoupling the process of gene exchange between species from that of sexual transfer, modern biotechniques have become the preferred method of transferring useful genes into crop cultivars from other taxa irrespective of the taxonomic relationship between donor organisms and crop plants. Although these techniques provide benefits, potential gene flow between crop plants and their wild or weedy relatives (Zohary and Feldman, 1962 Linder et al., 1998) have raised concerns regarding the negative consequences of alien genes in wild populations (Rieseberg and Wendel, 1993 Ellstrand, 2003).

In this review, we focus on three important issues concerning introgression between bread wheat and its wild relatives, especially those wild relatives with a known history of colonizing behavior in North America. First, we discuss reproductive biology and ecology of bread wheat and its wild relatives, as the potential for introgression from bread wheat to a wild relative primarily depends on the effectiveness of wheat pollen as a pollen donor (Joppa et al., 1968 Waines and Hegde, 2003). Second, since the success of wheat pollen in fertilizing a wild species depends on the degree of genetic relatedness between the two hybridizing taxa (Kimber and Feldman, 1987), we talk about the genetic mechanisms controlling intra- and intergenomic compatibility of bread wheat that influence meiosis and fertility of wheat plants and wheat hybrids. Third, we present the documented instances of introgression between bread wheat and its weedy wild relatives occurring in North America.

Reproductive Biology of Bread Wheat

The inflorescence of wheat is a determinate, composite spike with a main axis (rachis) bearing spikelets separated by short internodes (Percival, 1921 Lersten, 1987). The majority of wheat flowers in a spike are hermaphroditic, but a few are unisexual (De Vries, 1971). During flowering (blooming), flowers open (or remain closed as in the case of cleistogamous flowers) (Ueno and Itoh, 1997) and the three anthers open, releasing pollen (anthesis). Flowering continues throughout the day, with 2 to 6 d required for a spike to finish blooming (Percival, 1921). When flowers have reached blooming stage, the exact time and rate of blooming are strongly influenced by meteorological conditions such as variation of temperature, light, and humidity throughout the day (Leighty and Sando, 1924 De Vries, 1974). In addition, there are genotypic differences for blooming among wheat cultivars (Joppa et al., 1968 De Vries, 1974).

The extrusion of anthers and the duration of flower opening are significant factors affecting cross-pollination and potential gene flow, and in wheat these two factors are affected by both genetic control and environmental influence. Sage and De Isturiz (1974) indicated that anther extrusion in wheat is under the influence of a few genes, probably two, with additive effect and low heritability. Zukov (1969) found a correlation coefficient (r) of 0.93 in hexaploid wheats between the percentage of open florets and the percentage of extruded anthers. It was shown in 10 Canadian spring wheats that cultivars with high outcrossing rates tended to have a greater degree of floret opening at anthesis (Hucl, 1996). Both genotype and environment appear to influence the number of extruded anthers. In a multiyear investigation involving different genotypes and variable environments, Rajki (1962) observed the percentage of extruded anthers varied from 61 to 93% in the driest year the fewest anthers were extruded because of limited flower opening. In warm weather and at high atmospheric and soil humidity, anthers dehisce faster (De Vries, 1971). Environmental stress that results in a large proportion of extruded anthers would likely represent an ideal situation for gene flow to nearby wild relatives. Cultivars with a large number of open florets before anther dehiscence and/or with a longer duration of flower opening might be prone to greater gene flow than cultivars with a short anthesis period or with a large proportion of cleistogamous flowers (Waines and Hegde, 2003). Genotype and environmental (e.g., dry weather) interaction for flower opening likely results in a continuum of gene-flow effects among wheat cultivars.

Within wheat cultivars there is large variation for the number of pollen grains produced per anther ranging from 856 to 3867 (Cahn, 1925 Joppa et al., 1968) depending on the anther size (Bert and Anand, 1971). Although pollen production in predominantly self-pollinating wheat plants is nearly one-tenth (450 000 pollen grains) that of rye (Secale cereale L.) (4 200 000 pollen grains), an outcrossing relative (D'Souza, 1970), potential still exists in wheat for gene flow to occur, as 30 to 80% of the pollen is shed outside the flower (D'Souza, 1970 Bert and Anand, 1971). If environmental conditions are conducive and the isolation distance is not great enough between wheat and a wheat relative, gene flow could occur if pollen from one cultivar could pollinate unfertilized flowers of a receptive plant. Under optimal field conditions (20[degrees]C, 60% relative humidity), wheat pollen retains viability for approximately 30 min (D'Souza, 1970 De Vries, 1971), a duration long enough to bring about cross-fertilization with a nearby wild relative.

Wheat pollen is relatively heavy compared with other grass pollen and as a consequence wheat pollen grains typically travel short distances (Lelley, 1966). However, long-distance pollen movement was also reported by several researchers: wheat pollen grains were detected from a source population at a distance of 20 to 24 m (Suneson and Cox, 1964 De Vries, 1971), 50-60 m (D'Souza, 1970 Khan et al., 1973), or even at 1000 m (Virmani and Edwards, 1983). For example, during a 3-yr period from 1967 to 1969 at Newton, KS, collection stations were placed at 0, 3, and every 6 m afterward up to a 60-m distant from the pollen source. The average number of pollen grains collected hourly from 0700 to 1700 h on glass rods ranged from 72 to 153 grains at 0 to 3 m distance, and 33 to 43 grains at 48 m. A small number of pollen grains were also detected at 60 m from the pollen source (Khan et al., 1973). A similar study was also performed at Pioneer Hi-Bred International, Inc. (Kansas) where researchers conducted pollen trap studies using glass slides and tested the trapped pollen grains for their viability. This pollen trap study revealed that wheat pollen (viable or nonviable) could travel as far as 1000 m from a very large source population (Virmani and Edwards, 1983). In the majority of these pollen movement studies, more than 90% of pollen grains remained within 3 m of their source, and the amount of pollen collected decreased rapidly with increasing distances from the source plant (Jensen, 1968 Khan et al., 1973 Virmani and Edwards, 1983).

Furthermore, this long-distance pollen movement did not proportionately increase gene flow, as gene flow decreased rapidly after 0.5 to 1.5 m (Khan et al., 1973 Hucl and Matus-Cadiz, 2001). In a recent investigation on gene flow in wheat, Hucl and Matus-Cadiz (2001) examined gene flow rates of four wheat cultivars. Percent gene flow was determined for distances from 30 cm to 33 m from the pollen source. Gene flow rate decreased with greater distance from the pollen source, and was dependent upon wind direction and wheat genotype. Maximum gene flow rates at 30 cm (adjacent rows) were 3.8, 2.6, 0.4, and 0.2% for the four cultivars. At 27 m, gene flow was recorded in only 2 (0.095 and 0.06%) of 32 samples (2 yr x 4 cultivars x 4 directions). This emphasizes that physical movement of pollen does not necessarily result in gene flow.

Wheat Genomes and Genetics in Regards to Hybridization and Introgression

Bread wheat is an allohexaploid with three genomes, B, A, and D [genomic formula according to Waines and Barnhart (1992)]. Each genome originated from a different species: the B genome possibly from an ancestor of Aegilops speltoides Tausch the A genome from Triticum urartu Tum. ex Gand. and the D genome from Aegilops tauschii Coss. (Kihara, 1944 McFadden and Sears, 1946). The genus Aegilops contributed two-thirds, being the source of B and D genomes of bread wheat. Studies of variation in isozymes (Jaaska, 1978), nuclear DNA (Dvorak and Zhang, 1990), and organelle DNA (Mori et al., 1988 Wang et al., 1997) strongly support the idea that the B genome was derived from an S-genome species of the section Sitopsis, most likely related to allogamous Ae. speltoides.

Each one of the three genomes (B, A, and D) of bread wheat is composed of 7 chromosomes, which are labeled 1B, 1A, 1D through 7B, 7A, 7D, respectively. Thus, it is possible to place the 21 different chromosomes into seven groups of three. The chromosomes of each group are termed homeologous (= similar) and they are considered to have a common evolutionary origin (Kimber and Feldman, 1987). The term homologous refers to a pair of chromosomes within a genome that have alleles for the same genes. Although the homeologous chromosomes of the B, A, and D genomes of hexaploid wheat share several structurally identical genes, pairing between these chromosomes is prevented by a gene called Ph1 or Pairing homeologous 1 (Riley and Chapman, 1958 Sears, 1976). The Ph1 locus is mapped to the chromosome 5B linkage group (Riley and Chapman, 1958 Sears and Okamoto, 1958), and Ph1 acts as a dominant gene suppressing the pairing of homeologous chromosomes while allowing regular pairing between homologous chromosomes. Consequently, only bivalents are formed at meiosis therefore, common wheat behaves as a typical allopolyploid showing disomic inheritance. Furthermore, the Ph1 locus was also found to prevent homeologous chromosome pairing between wheat and several other related genomes in hybrids (Riley et al., 1959 Jauhar and Chibbar, 1999).

The Ph1 locus largely operates premeiotically to guarantee the association and alignment of homologous chromosomes (Feldman, 1966 Martinez et al., 2001), thus restricting synapsis and recombination to homologs rather than between homeologues. In the absence of Ph1, premeiotic alignment is presumably not restricted to homologs and can also occur between homeologues (Sears, 1977 Benavente et al., 1996), leading to the formation of one to several multivalent chromosomes (Riley, 1960 Sears, 1976) at a relatively high frequency (Riley, 1966 Benavente et al., 1998). The removal of Ph1 is necessary before homeologous chromosomes can pair but this does not guarantee synapsis and recombination. In fact, the degree of recombination achieved between homeologous chromosomes appears more related to their evolutionary divergence--for example, recombination is less likely to occur between unrelated but paired homeologous chromosomes (Feldman, 1993).

The effect of the Ph1 gene is suppressed under a variety of genetic backgrounds, as in the case of hybrids between bread wheat and some diploid Aegilops species. Some genes from Aegilops species are found to suppress the effect of the Ph1 locus and thereby allow homeologous chromosome pairing. This suppression was first observed in hybrids of bread wheat with Ae. speltoides (Feldman and Mello-Sampayo, 1967 Dover and Riley, 1972), but has since been observed in hybrids with certain genotypes of Ae. caudata L., Ae. sharonensis Eig, Ae. longissima Schweinf. & Muschl. and perhaps all diploid Aegilops and Triticum species (van Slageren, 1994). Another distant genus, Amblyopyrum muticum (Jaub. & Spach) Eig (synonym: Ae. mutica Boiss.), also effects suppression of the Phi locus (Dover and Riley, 1972).

A multi-allelic series is also known for differential suppression of the Ph1 locus in most species (Dover and Riley, 1972). But, so far, it is not clearly known if genes in polyploid Aegilops relatives are also able to suppress the Ph1 locus and thereby promote homeologous chromosome pairing and introgression. However, a few recent reports (Zemetra et al., 1998 Wang et al., 2001 Lin, 2001 Morrison et al., 2002b) indicated that, probably, Ph1 suppression also occurs at the polyploid level. Lin (2001) made artificial crosses between Ae. cylindrica (as female parent genomes CCDD) and bread wheat (genomes BBAADD) and obtained 16 to 21% hybrid seeds. Subsequently these hybrids were backcrossed (B[C.sub.1]) to Ae. cylindrica (as male parent) producing 2 to 2.5% B[C.sub.1] seeds. The author contended that formation of a few fertile [F.sub.1]s in his experiments was probably because of the pairing between shared D genomes of jointed goatgrass and wheat. If such interspecific hybrids between wheat and other polyploid Aegilops can also take place in nature, this may allow recombination of wheat transgenes into weedy Aegilops species.

In addition to the Ph1 genetic system, another genetic system called Kr is involved in controlling the ability of bread wheat to cross with rye and other genera. The ability of common wheat to cross with rye is controlled by four loci located on four chromosomes loci kr1 and kr2 (Lein, 1943) located on 5B (Riley and Chapman, 1967 Falk and Kasha, 1981) and 5A (Sitch et al., 1985) respectively locus kr3 on 5D (Krolow, 1970) and kr4 (Luo et al., 1989) on 1A (Zheng et al., 1992). These four loci do not equally contribute to the ability to cross, as their strength decreases from kr1, kr4, kr2, and kr3 in that order in other words, the alleles providing the greatest crossing ability are shown to occur on chromosomes 5B, 1A, and 5A (Luo et al., 1989 Zheng et al., 1992).

In hexaploid wheat, the ability to cross with rye is facilitated by recessive alleles or inhibited by dominant alleles at these loci (Snape et al., 1979 Fedak and Jui, 1982). This assertion was also supported by observation that noncrossable hexaploid wheat cultivars contain only dominant Kr alleles (Zheng et al., 1992). In addition to the dominant-recessive nature of the crossing alleles, wheat-rye crossing is also affected by the genotype of wheat cultivars, as there was marked variation in the amount of outcrossing between cultivars of wheat with rye--on average most of the 1400 wheat cultivars tested for crossing with rye exhibited greatly reduced seed set. Among these wheat cultivars, those producing higher seed set in crosses with rye were native to East Asia (Deng-Cai et al., 1999). Incidentally, the most popular bread wheat cultivar in wheat genetics research, Chinese Spring, was chosen for its high ability to cross with rye (Sears and Miller, 1985) and was a descendant of a land race from Sichuan Province, China (Deng-Cai et al., 1999).

Introgression in North America

In the USA, natural hybridization has been documented between Aegilops cylindrica Host. (jointed goatgrass) and bread wheat (Mayfield, 1927 Morrison et al., 2002a,b). Although another Aegilops species, Ae. triuncialis, has significant presence in northern California and also occurs in Pennsylvania, there has been no record of hybrids forming between bread wheat and Ae. triuncialis (Watanabe and Kawahara, 1999). The two Aegilops species have not been reported as occurring in Canada (Canadian Food Inspection Agency, 2003). Rye is a distant relative of wheat and feral rye populations persist in many places in the western wheat growing areas of the USA (Sun and Corke, 1992 Pester et al., 2000). There are reports of successful artificial hybrids created between wheat and rye (Florell, 1931), but natural hybrids between these two species have not been reported in the USA. Although plant breeders and agriculture extension workers observed natural wheat x rye hybrids occurring in Canada, such reports have yet to appear in the peer-reviewed publications. Therefore, in the following section we primarily discuss the biology and ecology of jointed goatgrass as related to the documented instances of introgression in North America.

Biology of Jointed Goatgrass

Jointed goatgrass generally grows 20 to 40 cm tall. As the plant matures, it can produce up to 135 tillers. The inflorescence is a narrow cylindrical spike with six to eight spikelets. Each spikelet can have two to five florets with usually one to two and sometimes three kernels per spikelet. In terms of seed production, a jointed goatgrass plant can produce up to 100 spikes, 1500 spikelets or joints and up to 3000 kernels. However, when growing in a wheat crop, even with adequate moisture, a typical jointed goatgrass plant produces approximately 130 kernels per plant (van Slageren, 1994).

Jointed goatgrass is similar to bread wheat (in particular, the winter types) in growth habit and appearance, but differs substantially in a few morphological traits (Donald and Ogg, 1991). Jointed goatgrass tends to have a more prostrate growth habit, narrower leaves, and a more compact spike than does bread wheat. Jointed goatgrass has hairs on the leaf margins, whereas bread wheat does not. The coleoptile color in jointed goatgrass is red versus green in bread wheat (Snyder et al., 2000). The joints (spikelets) differ from wheat spikelets, looking like small pieces of stick. At maturity, the spike falls intact and the spikelets then separate with a segment of the rachis still attached.

Jointed goatgrass is a weed of winter wheat, and it rarely infests spring wheat. Spring wheats grow in regions where the winters are too harsh for the winter varieties. The cold winter temperature, therefore, also prevents the successful establishment of jointed goatgrass. Winter wheats are planted in the fall in regions where the winters are relatively mild and dry. The grain begins to grow before the cold weather approaches and they become dormant during the winter and resume growth in the spring. Winter wheat is harvested in the summer. Physiologically, jointed goatgrass and winter bread wheat are analogous, having similar temperature optimums, maximum photosynthetic rates and growth rates, as well as other similar physiological characteristics. Thus both morphological and physiological similarities of jointed goatgrass to winter wheat make control very difficult once jointed goatgrass is introduced into a wheat field (Donald and Ogg, 1991).

Jointed goatgrass has definite weedy characteristics, which make it particularly troublesome. The kernels ripen before those of winter wheat and spikelets readily shatter from the plant. Studies indicate that jointed goatgrass readily germinates at soil temperatures of 10 to 35[degrees]C (Morrow et al., 1982). Further, some germination will occur at temperatures as low as 2[degrees]C and as high as 40[degrees]C. Jointed goatgrass can emerge from soil at depths as great as 16 cm (Cleary and Peeper, 1980) thus, plowing to bury the spikelets at depths great enough to prevent emergence has not been effective. In addition, jointed goatgrass seeds may remain viable in the soil for 3 to 5 yr after being shed. In terms of seed dormancy, the basal seeds of a spikelet have greater dormancy than the subterminal seeds exhibiting germination-polymorphism among seeds of a spike (Morrow et al., 1982). Seed dormancy has helped the species to persist in the soil and to rapidly colonize new areas.

Distribution in North America

Jointed goatgrass is a native of western Asia and eastern Europe and was introduced into the USA at the end of the 19th century (Carleton, 1915 Hitchcock, 1951). Currently it has colonized many states, from the east to west coasts, although most abundantly in the western and northwestern states and the plains of the Midwest (Donald and Ogg, 1991). At the time of last survey conducted in 1993 to document the distribution of jointed goatgrass, the infestation in U.S. winter wheat production areas was over 2 million hectares and was spreading unchecked at a rate of 20000 ha per year (National Jointed Goatgrass Research Program, 2003). It is most commonly found in winter wheat fields or other cereal grain fields, fencerows, roadsides, and waste areas. It also infests rangelands surrounding wheat growing areas and land in the conservation reserve program throughout the western USA (Westra and Davis, 1987). It disperses through spikelets (joints) and has been introduced to some areas by custom combines (Fenster and Wicks, 1976) and to others by planting contaminated wheat grain as seed (Swan, 1984).

Introgression (Bread Wheat x Jointed Goatgrass)

Jointed goatgrass is an allotetraploid (DDCC genome formula according to Kimber and Tsunewaki, 1988) with 28 chromosomes. Each chromosome set contains 7 chromosomes and originates from two species designated by CC and DD (Kihara, 1937 Kimber and Sears 1987). One reason jointed goatgrass and wheat are so similar is that the two species share a common ancestor, Ae. tauschii, the donor of the D genome. Hexaploid wheat also contains genomes B and A from two other species. The D genomes in jointed goatgrass and bread wheat are so similar that in the hybrid the D genomes from the two species complement each other acting as a homologous pair (Kimber and Sears, 1987 Zemetra et al., 1998) while producing a few viable hybrids (Mallory-Smith et al., 1996 Morrison et al., 2002b). In an ideal situation, the pentaploid hybrids can be expected to possess 7 pairs of D genome chromosomes, and a haploid set of A (7 chromosomes), B (7 chromosomes), and C (7 chromosomes) genomes with a hybrid genome of 2n = 5x = 35, DDCAB or BADDC (depending on whether jointed goatgrass or bread wheat is the female parent). Nevertheless, in nature there may be breakdown of strict homologous chromosome pairing, and the wheat and jointed goatgrass hybrids may exhibit an array of chromosome numbers (Zemetra et al., 1998) depending on the number of promoter or suppressor genes for homeologous chromosome pairing.

A majority of plants of both species are primarily self-pollinating, but the discovery of a few natural hybrids between bread wheat and jointed goatgrass indicates that a small amount of outcrossing does occur under natural conditions both in the USA and other world areas where both the parental species co-occur (Zohary and Feldman, 1962 Hammer and Matzk, 1993 van Slageren, 1994 Mallory-Smith et al., 1996). When formed, bread wheat x jointed goatgrass hybrids are expected to be sterile because they are pentaploid (2n = 5x = 35) and hybrids lack chromosome pairing during meiosis except for the chromosomes of the D genomes. Because of the lack of meiotic pairing and subsequent unbalanced chromosome numbers in the gametes, most (>99%) of the [F.sub.1] hybrids were completely sterile (Zemetra et al., 1998).

Nevertheless, a few fertile hybrid seeds were observed in the U.S. wheat fields (Mallory-Smith et al., 1996 Snyder et al., 2000 Morrison et al., 2002b) and glasshouse experiments (Zemetra et al., 1998 Wang et al., 2001). Furthermore, even seeds were obtained on subsequent backcross generations (Zemetra et al., 1998 Wang et al., 2002). For example, a backcrossing program was initiated in the glasshouse between wheat x jointed goatgrass [F.sub.1] hybrids (Zemetra et al., 1998) and either jointed goatgrass or wheat was used to determine the potential for seed set and the restoration of self-fertility. Seed set occurred on the wheat x jointed goatgrass hybrids with either wheat or jointed goatgrass as the pollen parent and the frequency of seed-set (2.0 and 2.2%) was similar between the two pollen sources. Furthermore, these glasshouse seed-set data were similar to the overall frequency of seed set (2.2%) on natural hybrids observed in the field, indicating that the seed set observed in the glasshouse was not an artifact of controlled emasculation-pollination. Backcross individuals with either species as the recurrent parent exhibited a similar level of seed set (4.6- 5.1%). The level of seed set on B[C.sub.2] plants differed between the two recurrent parents, with jointed goatgrass B[C.sub.2] plants having three times the level of partial female fertility (37.4%) compared with wheat B[C.sub.2] plants (13.7%). Although the relative female fertility level increased from 2% in the [F.sub.1] hybrids to 37% in the backcross generations in glasshouse conditions, some of the techniques used in the glasshouse to achieve a higher hybrid seed production and to rescue B[C.sub.1] seeds are too artificial to relate the results to the same set of plants growing under natural conditions. For example, in the glasshouse a partially emerged head of wheat, jointed goatgrass, or backcross plant was emasculated and pollinated only with one of the selected parents however, in nature, pollen grains from different sources such as self- and cross-pollen compete to fertilize the ovules. In the glasshouse, a 5 % seed set was accomplished on B[C.sub.1] plants after B[C.sub.1] seeds were rescued through an embryo rescue technique.

In a similar glasshouse study (Wang et al., 2001), the fertility of wheat x jointed goatgrass hybrids and backcross progenies were tested for their seed set as a measure of gene flow from wheat to jointed goatgrass. Although this study did not provide the number of hybrid seeds generated via artificial crossing in the glasshouse, the authors report 100% male sterility and 0.87% female fertility for [F.sub.1] hybrid plants. When jointed goatgrass was used as the male parent to generate backcross generations, the seed-set percentages on [F.sub.1], B[C.sub.1], and B[C.sub.2] plants were 0.9, 4.4, and 18, respectively. Furthermore, there was a dramatic increase in fertility restoration when the B[C.sub.2][S.sub.1] plants (56% seed set) and the subsequent generation (B[C.sub.2][S.sub.2] 79% seed set) plants were selfed. Such a high proportion of seed set in the backcross generations indicates that a wheat transgene with either a higher fitness or a neutral effect could spread in a jointed goatgrass population if it escaped. For example, a soft red winter wheat was mutagenized and plants resistant to an imidazolinone-class herbicide were selected from the mutagenized population (White and Morrison, 1998). The herbicide kills both bread wheat and jointed goatgrass deficient for the herbicide-resistant mutant gene. In a 1995-1996 herbicide efficacy study, jointed goatgrass pollen fertilized some imidazolinone-resistant wheat plants near Pullman, WA (Ball et al., 1999). Two herbicide-resistant hybrids were discovered in the next generation. These hybrids produced seven viable seeds in the first backcross generation between the hybrid and wheat plants. The hybrid plants were, on average, 15 cm taller and more robust than the surrounding herbicide-resistant wheat cultivar (Seefeldt et al., 1998). The subsequent glasshouse experiments showed that a few viable seeds were produced via selfing or open pollination (Snyder et al., 2000).

One interesting aspect regarding wheat and jointed goatgrass hybrid derivatives is the genes located on the D genome of wheat have a greater chance of retention in the [F.sub.1] and backcross generation plants, compared with genes of the A or B genomes. In both [F.sub.1] and backcross hybrids, normal bivalents are formed between the D genome chromosomes of wheat and jointed goatgrass, while the majority of chromosomes of the A and B genomes of wheat and the C genome of jointed goatgrass form univalents at meiosis-I (Zemetra et al., 1998 Wang et al., 2001, 2002). The results of this study suggest that normal bivalents could also be expected between A and B genomes of wheat and wild relatives possessing these genomes.

This indicates that, at least in theory, some genes could be transferred between wheat and jointed goatgrass, especially if those genes are located on the D genome of bread wheat. However, for genes located on the A or B genomes of bread wheat, the probability of their incorporation in the hybrids and hybrid-derived generations is substantially lower and very stochastic (Lin, 2001 Wang et al., 2002). Lin (2001) tested the transferability of genes present on different genomes (B and D) of wheat into jointed goatgrass (C and D genomes). Accordingly, the author transformed line 31 of 'Bobwhite' wheat with the glufosinate (herbicide)-resistant Bar gene in the B genome, and line 71 individuals of Bobwhite wheat with the glufosinate-resistant Bar gene in the D genome. Under glasshouse condition, these transformed wheat lines were used as male parents to pollinate jointed goatgrass as female parents. Three ways of pollination were performed between the parents: natural pollination on unemasculated jointed goatgrass females, natural pollination on emasculated jointed goatgrass females, and artificial pollination on emasculated jointed goatgrass females. Subsequently, the same three-way pollination method was also repeated, using fertile [F.sub.1] plants as female parents and jointed goatgrass as pollen donor, to generate backcross progenies. The unemasculated and naturally pollinated spikes of jointed goatgrass did not produce any seeds. The emasculated and naturally pollinated jointed goatgrass spikes set 16 and 17% seeds with male parental lines 31 and 71, respectively and emasculated and artificially pollinated jointed goatgrass spikes set 21 and 18% seeds with male parental lines 31 and 71, respectively. But among B[C.sub.1] crosses, only the emasculated and artificially pollinated [F.sub.1] plants produced some seeds. The cross (jointed goatgrass x line 31) x jointed goatgrass set 2% seeds, and the cross (jointed goatgrass x line 71) x jointed goatgrass set 2.5% seeds out of 50 pollinated spikes under each of the two crossing schemes. When the B[C.sub.1] plants were tested for the presence of the glufosinate-resistant Bar genes, two-thirds of B[C.sub.1] plants possessed the resistant gene from the cross [(jointed goatgrass x line 71) x jointed goatgrass] where line 71 wheat carried the resistant Bar gene on the D genome, whereas none of the B[C.sub.1] plants possessed the resistant gene from the cross [(jointed goatgrass x line 31) x jointed goatgrass] where line 31 wheat carried the resistant Bar gene on the B genome. These results suggest that the transfer of transgenic genes located on the A or B genomes of wheat to other related plants is less likely via sexual hybridization.

Introgression between Bread Wheat and Rye

Natural introgression between bread wheat and rye has rarely been reported in North America (Leighty, 1915) and appears infrequent in other areas of the world (Rimpau, 1891 Dorofeeva, 1966). Nevertheless, several successful artificial crosses were performed between these two genera (Falk and Kasha, 1981). For example, Falk and Kasha (1981) and Tanner and Falk (1981) reported a high correlation between the ability of specific wheat cultivars to cross with S. cereale L.

Genetics of Wheat x Rye Hybrids

The ability of wheat and rye to cross was reported to be controlled by one (Taylor and Quisenberry, 1935) or by two recessive genes, kr1 and kr2 (Lein, 1943) in wheat. In contrast, dominant genes in S. cereale control crossing with wheat (Tanner and Falk, 1981). In addition, suppression of Ph1 locus was found to induce wheat and rye homeologous chromosomes to pair and recombine (Naranjo and Fernandez-Rueda, 1996).

Wheat--rye hybrids are easier to obtain than rye--wheat hybrids (Backhouse, 1916 Leighty and Sando, 1928). Most Triticum--Secale hybrids are completely male sterile and highly female fertile. The chromosomes of Triticum species do not pair with chromosomes of Secale species, and meiosis in the [F.sub.1] was highly irregular (Maan, 1987). Occasionally, fertilization involving unreduced functional gametes results in selfed seed that produces fertile amphiploid progeny (Yakubtsiner, 1952 Sadykov, 1952). Although genes from Ph and Kr loci are directly involved in the regulation of meiosis in wheat--wheat and wheat--alien hybrids they may also indirectly affect seed set in the hybrids. Therefore, it is possible that interactions between the pairing and crossing genes in wheat and rye may reduce or enhance seed set depending on the direction of hybridization.

Fertility of Interspecific and Intergeneric Wheat Hybrids

Triticum and its related genera consist of diploid and polyploid species with various genome combinations. Therefore, the extent of hybrid fertility between species of these genera is a function of ploidy and genomic compatibility between hybridizing taxa. Although many species of Triticum and related genera are cross-compatible, and hybrids between species belonging to the same as well as different genera can be produced (Knobloch, 1968, p 47-52 Maan, 1987 van Slageren, 1994), most interspecific and intergeneric hybrids involving species of Triticum are highly male sterile but partially female fertile (Maan, 1987). This female-only fertile condition may encourage introgression. Therefore, the [F.sub.1] hybrids sometimes produce backcross progeny from pollination with parental or other related species (Maan, 1987). Polyploid species are more fertile acting as females than as males in crosses between diploid and polyploid species (Maan, 1987). Also, there is substantial intra- and interspecific variability in cross-compatibility among the species of Triticum and related genera. Certain species of Triticum are cross-compatible only as a male or as a female with other species of Triticum and related genera, regardless of the ploidy differences between the parental species (Kihara, 1937 Gill and Waines, 1978).

Implications of Introgression in Consideration of Current and Possible Future Development of Transgenic Wheat

Bread wheat is a major food crop of mankind in terms of the area under cultivation and megagrams of grain harvested (Braun et al., 1998). The crop is cultivated in all the continents and, in several places, bread wheat and wild relatives grow side by side (van Slageren, 1994). The fact that hybrids between wheat and jointed goatgrass are partially fertile raises the question of whether a wheat gene could be transferred to jointed goatgrass if jointed goatgrass were the recurrent parent in the backcross. This is important because one possible control strategy for jointed goatgrass is the use of genetically engineered herbicide resistant wheat. On the basis of our analysis (see Appendix 1), we estimated that the probability of recovering B[C.sub.2] seed with jointed goatgrass as the recurrent parent would be approximately one plant out of 1.54 million plants if a transgene is present on the D genome of wheat. In reality, however, the hybridization rate in B[C.sub.2] might be very much lower than what was predicted by our analysis. For example, the haploid genomes A, B, and C in the [F.sub.1], each of which contains seven chromosomes, rarely segregate intact to one of the meiotic poles. Such irregular meiosis tends to produce higher numbers of sterile pollen and eggs than was predicted by our analysis (Appendix 1). Moreover, in the field, the frequency of successful backcross seed production could be much lower since our assumptions of 1% seed set of [F.sub.1] hybrids and 5% seed set in B[C.sub.1] plants are from experimental data obtained under optimum conditions in the glasshouse (Wang et al., 2001). Furthermore, the extent of backcrossing with jointed goatgrass is very limited in the field because of competition between wheat and jointed goatgrass pollen (Snyder et al., 2000) and also because of a limited number of jointed goatgrass plants in the field, as agronomic practices to control weeds eliminate many of the jointed goatgrass plants. Although partial introgression may occur in approximately one in 1.54 million plants from our probability analysis, there is no empirical evidence linking the invasiveness of species such as jointed goatgrass to their introgression with bread wheat. The recent experimental evidence only suggests that introgression is possible between jointed goatgrass and bread wheat.

One concern associated with genetically modified wheat is that when/if a transgene is introduced into a population of wild relatives, there may be an undesired ecological effect associated with the presence of the transgene, such as increased fitness or competitiveness. The perceived concern is the potential for a greater ecological risk if the transgene involved has a potential to enhance the fitness of hybrids in a weedy wild population, thereby increasing the weed potential of the receiving species. Escaped transgenes for traits like seed color and flavor, per se, may not bring any fitness advantage to the wild populations, while a drought tolerance, or an insect or herbicide resistant transgene has, under certain conditions.(e.g., insect or herbicide induced selective pressure) the potential to enhance the fitness of a wild population if it escaped. But bread wheat and its wild relatives have to overcome several internal and external constraints to produce a viable hybrid or introgressed wild population in nature. In the first place, a predominantly self-pollinating bread wheat and its wild relatives coupled with their asynchronous flowering phenologies effectively prevent any large-scale pollen flow between wheat and wild populations. Besides, cross incompatibility between species, and hybrid sterility in the [F.sub.1] and subsequent backcross generations, further reduce viable hybrids between wheat and wild relatives.

The discussion, so far, points to the process of introgression between bread wheat and wild relatives, in the present scenario, as a two-step process. The first, and also the most important step, is the formation of viable hybrids with a fitness equal or higher than the parents. The available data on hybrid frequency do not suggest the formation of wheat and wild wheat hybrids in any significant frequency either in the USA or in other wheat growing areas of the world when bread wheat and wild relatives occur in sympatry. Even the reported incidence of hybrids are mostly based on morphological traits therefore, the hybrid frequency data may be an under or overestimation of the actual figure. Many natural hurdles preclude hybrid formation between wheat and wild relatives, namely, asynchronous flowering, gametic or zygotic incompatibility, reduced hybrid fitness, or hybrid sterility. For example, many wild wheats have distinctly different sets of genomes that do not pair with wheat chromosomes during gamete formation in [F.sub.1] hybrids or may cause developmental instability thus reducing the ability of the hybrids to survive in the parental habitats.

Technological innovations bring their own set of benefits and risks to the environment, and no technology is 100% safe. The same is true for transgenic crop plants that contain novel traits incorporated by the tools of biotechnology and for crop cultivars produced by traditional plant breeding methods (Hegde and Ellstrand, 2002). Therefore, when considering a new product, it is important to consider the risk and benefits of the current cropping system compared with the risks and benefits of the new system. For most of the fitness enhancing traits such as disease, insect, and herbicide resistance, the potential harm to the environment is not very much different. So, some of the modern wheat cultivars with fitness enhancing traits can be used as a model system to evaluate the risk of escape of an introduced gene to wild populations. This kind of investigation will benefit both transgenic risk analysis research and the environment.

The following illustrates a possible route and frequency of introgression of a transgene located on the D genome of bread wheat into jointed goatgrass (JGG) (see text for details).

1. Direction of gene flow: From bread wheat (pollen donor) to JGG (female parent).

2. [F.sub.1] hybrid fertility: No [F.sub.1] hybrids were obtained through natural pollination (Zemetra et al., 1998 Wang et al., 2001 Lin 2001). However, in artificial pollination experiments the average female fertility values were 0.9% (Wang et al., 2001) and 2% (Zemetra et al., 1998). We considered a value of 1% [F.sub.1] female fertility for the purpose of our calculations.

3. Restoration of fertility in backcross generation 1 (B[C.sub.2]) by crossing [F.sub.1] individuals with JGG varied from 4.4% (Wang et al., 2001) to 4.6% (Zemetra et al., 1998). For our calculations, we have considered a value of 4% female fertility for B[C.sub.1] individuals.

4. All [F.sub.1] gametes containing only one genome are considered sterile.

JGG [female parent: 2n = 4x = 28 (7D+7D, 7C+7C)] x

bread wheat [male parent: 2n = 6x = 42

Genome combination in [F.sub.1] seeds:

7A, 7B, 7C, 7D+7D (35 chromosomes)

[F.sub.1] plants (DD C B A) x JGG

Potentially, all the B[C.sub.1] individuals can undergo backcrossing with JGG and produce some seeds (Appendix Table 1). However, for the illustration purpose, the B[C.sub.1] plant with the genome combination DDCC is the best candidate to generate fertile B[C.sub.2] individuals in the subsequent round of backcrossing with JGG. The frequency of B[C.sub.1] individuals with DDCC genome combination is (1/7 x 1/1100)

1/7700. If these individuals have an average 4% female fertility (1/7700 x 4/100 = 4/770 000) with equal survival of B[C.sub.1] gametes, the results will be as follows:

B[C.sub.1] female plant (DDCC) x JGG (DDCC)

Thus in an ideal situation where the transgene is located on the shared D genome of wheat, at the end of the second backcross generation, 4 out of 770 000 plants can be partially fertile B[C.sub.2] individuals.

If the transgenic trait is placed only on one of the two D genomes of wheat (as is the practice in transgenic corn), then such a chromosomal arrangement creates a hemizygous condition for the transgenic trait. Then there is a one-half probability of the transgene being present among [F.sub.1] individuals, 1/4 among B[C.sub.1] individuals, or 1/8 among B[C.sub.2] individuals. Therefore, the overall probability of getting a transgenic B[C.sub.2] individual from the cross [B[C.sub.1] (DDCC) x JGG] is (4/770 000 x 1/8) nearly one in 1.54 million.

Partial funding for this project was provided by Monsanto Company, California Agricultural Experiment Station, and J.G.W.

Abbott, R.J. 1992. Plant invasions, interspecific hybridization and the evolution of new plant taxa. Trends Ecol. Evol. 7:401-405.

Anderson, E., and L. Hubricht. 1938. Hybridization in Tradescantia. III. The evidence for introgressive hybridization. Am. J. Bot. 25: 396-492.

Anttila, C.K., C.C. Daehler, N.E. Rank, and D.R. Strong. 1998. Greater male fitness of a rare invader (Spartina alterniflora, Poaceae) threatens a common native (Spartina foliosa) with hybridization. Am. J. Bot. 85:1597-1601.

Arnold, M.L. 1997. Natural hybridization and evolution. Oxford University Press, New York.

Arnold, M.L., C.M. Buckner, and I. Robinson. 1991. Pollen-mediated introgression and hybrid speciation in Louisiana irises. Proc. Natl. Acad. Sci. USA 88:1398-1402.

Backhouse, W.O. 1916. Note on the inheritance of "crossability". J. Genet. 6:91-94.

Bajaj, Y.P.S. (edit.). 1990. Biotechnology in agriculture and forestry 13: Wheat. Springer-Verlag, New York.

Ball, D.A., F.L. Young, and A.C. Ogg. 1999. Selective control of jointed goatgrass with imazamox in herbicide-resistant wheat. Weed Technol. 13:77-82.

Benavente, E., B. Fernandez-Calvin, and J. Orellana. 1996. Relationship between the levels of wheat-rye metaphase-I chromosomal pairing and recombination revealed by GISH. Chromosoma 101: 365-373.

Benavente, E., J. Orellana, and B. Fernandez-Calvin. 1998. Comparative analysis of the meiotic effects of wheat ph1b and ph2b mutations in wheat x rye hybrids. Theor. Appl. Genet. 96:1200-1204.

Beri, S.M., and S.C. Anand. 1971. Factors affecting pollen shedding capacity in wheat. Euphytica 20:327-332.

Braun, H.J., T.S. Payne, A.I. Morgounov, M. van Ginkel, and S. Rajaram. 1998. The challenge: One billion tons of wheat by 2020. p. 33-40. In A.E. Slinkard (ed.) Proc. 9th Int. Wheat Genet. Symp., Vol. 1, Saskatoon, SK, Canada. Univ. Ext. Press, Saskatoon, SK, Canada.

Cahn, E. 1925. A study of fertility in some common varieties of wheat with respect to anther length and amount of pollen in parents and offspring. J. Am. Soc. Agron. 17:591-595.

Canadian Food Inspection Agency. 2003. Website: http://www., verified 23 March 2004.

Carleton, M.A. 1915. Hard wheats winning their way. p. 319-420. USDA Year-book 1914. U.S. Gov. Print Office, Washington, DC.

Clausen, J., D.D. Keck, and W.M. Hiesey. 1945. Experimental studies on the nature of species. II. Plant evolution through amphiploidy and autoploidy, with examples from the Madiinae. p. 1-163. Carnegie Institution of the Washington Publ. 564, Pasadena, CA.

Cleary, C.L., and T.F. Peeper. 1980. Growth and control of jointed goatgrass. Proc. So. Weed Sci. Soc. 33:23.

Deng-Cai, L., C. Yen, Y. Juan-Liang, Z. You-Liang, and L. Xiu-Jin. 1999. The chromosome location of high crossability genes in tetraploid wheat Triticum turgidum L. cv. Ailanmai native to Sichuan, China. Euphytica 108:79-82.

De Vries, A.Ph. 1971. Flowering biology of wheat, particularly in view of hybrid seed production--a review. Euphytica 20:152-170.

De Vries, A.Ph. 1974. Some aspects of cross-pollination in wheat (Triticum aestivum L.). 4. Seed set on male-sterile plants as influenced by distance from the pollen source, pollinator: Male-sterile ratio and width of the male-sterile strip. Euphytica 23:601-622.

Donald, W.W., and A.G. Ogg. 1991. Biology and control of jointed goatgrass (Aegilops cylindrica), a review. Weed Technol. 53:3-17.

Dorofeeva, V.F. 1966. Spontaneous (wheat-rye) amphidiploids. (In Russian.) Vaskhnil 2: 21-23.

Dover, G.A., and R. Riley. 1972. Prevention of pairing of homoeologous meiotic chromosomes of wheat by an activity of supernumerary chromosomes of Aegilops. Nature 240:159-161.

D'Souza, V.L. 1970. Investigations concerning the suitability of wheat as pollen-donor for cross-pollination by wind as compared to rye, Triticale, and Secalotricum. (In German.) Zeitschrift fur Pflanzenzuchtung 63:246-269.

Dvorak, J., and H.B. Zhang. 1990. Variation in repeated nucleotide sequences sheds light on the phylogeny of the wheat B and G genomes. Proc. Natl. Acad. Sci. USA 87:9640-9644.

Ellstrand, N.C. 2003. Dangerous liaisons: When crops mate with their wild relatives. Johns Hopkins University Press, Baltimore, MD.

Ellstrand, N.C., and K.A. Schierenbeck. 2000. Hybridization as a stimulus for the evolution of invasiveness in plants? Proc. Natl. Acad. Sci. USA 97:7043-7050.

Falk, D.E., and K.J. Kasha. 1981. Comparison of the crossability of rye (Secale cereale) and Hordeum bulbosum onto wheat (Triticum aestivum). Can. J. Genet. Cytol. 23:81-88.

Fedak, G., and P.Y. Jui. 1982. Chromosomes of Chinese Spring wheat carrying genes for crossability with Betzes barley. Can. J. Genet. Cytol. 24:227-233.

Feldman, M. 1966. The effect of chromosomes 5B, 5D, and 5A on chromosomal pairing in Triticum aestivum. Proc. Natl. Acad. Sci. USA 55:1447-1453.

Feldman, M. 1993. Cytogenetic activity and mode of action of the pairing homoeologous (Ph1) gene of wheat. Crop Sci. 33:894-897.

Feldman, M., and T. Mello-Sampayo. 1967. Supression of homoeologous pairing in hybrid of polyploid wheat x Triticum speltoides. Can. J. Genet. Cytol. 9:307-313.

Fenster, C.R., and G.A. Wicks. 1976. Jointed goatgrass, p. 2. Nebr. Univ. Coop. Ext. Serv. G 75-210.

Florell, V.H. 1931. A genetic study of wheat x rye hybrids and back crosses. J. Agric. Res. 42:315-339.

Gill, B.S., and J.G. Waines. 1978. Paternal regulation of seed development in wheat hybrids. Theor. Appl. Genet. 51:265-270.

Grant, V. 1981. Plant speciation. Columbia Univ. Press, New York.

Hammer, K., and F. Matzk. 1993. Variation in breeding systems in Triticeae. p. 51-58. In A.B. Damania (ed.) Biodiversity and wheat improvement. John Wiley and Sons, New York.

Harrison, R.G. 1990. Hybrid zones: Windows on evolutionary process. Oxford Surv. Evol. Biol. 7:69-128.

Hegde, S.G., and N.C. Ellstrand. 2002. Transgenic crop plants and the environment: Benefits and risks, p. 41-62. In S. Van Gundy (ed.) Benefits and risks of food biotechnology. California Council on Science and Technology, Sacramento, CA.

Heiser, C.B. 1973. Introgression re-examined. Bot. Rev. 39:347-366.

Hitchcock, A.S. 1951. Manual of the grasses of the United States (2nd ed., rev. A. Chase). USDA Misc. Publ. 200.

Hucl, P. 1996. Out crossing rates for 10 Canadian spring wheat cultivats. Can. J. Plant Sci. 76:423-427.

Hucl, P., and M. Matus-Cadiz. 2001. Isolation distances for minimizing out-crossing in spring wheat. Crop Sci. 41:1348-1351.

Jaaska, V. 1978. NADP-dependent aromatic alcohol dehydrogenases in polyploid wheats and their diploid relatives--origin and phylogeny of polyploid whetas. Theor. Appl. Genet. 53:209-217.

Jauhar, P.P., and R.N. Chibbar. 1999. Chromosome-mediated and direct gene transfers in wheat. Genome 42:570-583.

Jensen, N.F. 1968. Results of a survey on isolation requirements for wheat. Ann. Wheat Newsl. 15:26-28.

Joppa, L.R., F.H. MacNeal, and M.A. Berg. 1968. Pollen production and pollen shedding of hard red spring (Triticum aestivum L. em. Thell.) and durum (T. durum Desf.) wheats. Crop Sci. 8:487-490.

Khan, M.N., E.G. Heyne, and A.L. Arp. 1973. Pollen distribution and the seedset on Triticum aestivum L. Crop Sci. 13:223-226.

Kihara, H. 1937. Genomanalyse bet Triticum and Aegilops. VII. Kurze Uberscicht uber die Ergebnisse der Jahre 1934-1936. (In Japanese.) Mem. Coll. Agric. Kyoto Imp. Univ. 41:1-61.

Kihara, H. 1944. Discovery of the DD-analyser, one of the ancestors of vulgare wheats. (In Japanese.) Agric. Hort. 19:13-14.

Kimber, G., and E.R. Sears. 1987. Evolution in the genus Triticum and the origin of cultivated wheat, p. 154-164. In E.G. Heyne (ed.) Wheat and wheat improvement. 2nd ed. Agron. Monogr. 13. ASA, CSSA, and SSSA. Madison, WI.

Kimber, G., and K. Tsunewaki. 1988. Genome symbols and plasma types in the wheat group, p. 1209-1211. In T.E. Miller and R.M.D. Koebner (ed.) Proc. 7th Int. Wheat Genet. Syrup., Vol. 2, Cambridge, England. Inst. Plant Sci. Res., Cambridge, UK.

Kimber, G., and M. Feldman. 1987. Wild wheat: An introduction, p. 99-103. Special Report 353. College of Agric., Univ. of Missouri, Columbia.

Knobloch, I.W. 1968. A checklist of crosses in the Graminae. Department of Botany and Plant Pathology, Michigan State University, East Lansing, MI. p

Knobloch, I.W. 1972. Intergeneric hybridization in flowering plants. Taxon 21:97-103.

Krolow, K.D. 1970. Untersuchungen uber die Kreuzbarkeit zwischen Weizen und Roggen. (In German.) Z. Pflanzenzucht 64:44-72.

Leighty, C.E. 1915. Natural wheat-rye hybrid. Am. Soc. Agron. 7: 209-216.

Leighty, C.E., and W.J. Sando. 1924. The blooming of wheat flowers. J. Agric. Res. 27:231-244.

Leighty, C.E., and W.J. Sando. 1928. Natural and artificial hybrids of a Chinese wheat and rye. J. Hered. 19:23-27.

Lein, A. 1943. The genetical basis of the crossability between wheat and rye. Z. Indukt. Abstam.-u-VererbLehre. 81:28-59.

Lelley, J. 1966. Observation on the biology of fertilization with regard to seed production in hybrid wheat. (In German.) Dr Zuchter 36:314-317.

Lersten, N.R. 1987. Morphology and anatomy of the wheat plant, p. 33. In E.G. Heyne (ed.) Wheat and wheat improvement. Agron. Monogr. 13. ASA, CSSA, and SSSA, Madison, WI.

Levin, D.A. 1967. Hybridization between annual species of Phlox: Population structure. Am. J. Bot. 54:1122-1130.

Lin, Y. 2001. Risk assessment of Bar gene transfer from B and D genomes of transformed wheat (Triticum aestivum) lines to jointed goatgrass (Aegilops cylindrica). J. Anhui Agril. Univ. 28:115-118.

Linder, C.R., I. Taha, G.J. Seiler, A.A. Snow, and L.H. Rieseberg. 1998. Long-term introgression of crop genes into wild sunflower populations. Theor. Appl. Genet. 96:339-347.

Lorz, H., D. Becker, and S. Lutticke. 1997. Molecular wheat breeding by direct gene transfer. p. 285-289. In H.J. Braun et al. (ed.) Wheat: Prospects for global improvement. Kluwer Academic Publishers, Norwell, MA.

Luo, M.C., C. Yen, and J.L. Yang. 1989. The crossability of landraces of common wheat in Sichuan with Aegilops tauschii and rye. J. Sichuan Agric. Univ. 7:77-81.

Maan, S.S. 1987. Interspecific and intergeneric hybridization in wheat. p. 453-461. In E.G. Heyne (ed.) Wheat and wheat improvement. 2nd ed. Agron. Monogr. 13. ASA, CSSA, and SSSA, Madison, WI.

Mallory-Smith, C.A., J. Hansen, and R.S. Zemetra. 1996. Gene transfer between wheat and Aegilops cylindrica, p. 441-445. In Proc. 2nd Int. Weed Control Cong., Copenhagen, Denmark. Published by Dep. Weed Control and Pesticide Ecol. Copenhagen, Denmark.

Martinez, M., N. Cunado, N. Carcelen, and C. Romero. 2001. The Ph1 and Ph2 loci play different roles in the synaptic behaviour of hexaploid wheat Triticum aestivum. Theor. Appl. Genet. 103:398-405.

Mayfield, L. 1927. Goat grass--A weed pest of central Kansas wheat fields. Kans. Agric. Student 7:40-41.

McFadden, E.S., and E.R. Sears. 1946. The origin of Triticum spelta and its free-threshing hexaploid relatives. J. Hered. 37:81-49, 107-116.

Mori, N., T. Terachi, and K. Tsunewaki. 1988. Organellar genome differentiation in wild tetraploid wheats, Triticum dicoccoides and T. araraticum, p. 109-114. In T.E. Miller and R.M.D. Koebner (ed.) Proc. 7th Int. Wheat Genet. Symp., Cambridge, England. Inst. Plant Sci. Res., Cambridge, UK.

Morrison, L.A., L.C. Cremieux, and C.A. Mallory-Smith. 2002a. Infestation of jointed goatgrass and its hybrids with wheat in Oregon wheat fields. Weed Sci. 50:134-144.

Morrison, L.A., O. Riera-Lizarazu, L. Cremieux, and C.A. Mallory-Smith. 2002b. Jointed goatgrass (Aegilops cylindrica Host) x wheat (Triticum aestivum L.) hybrids: Hybridization dynamics in Oregon wheat fields. Crop Sci. 42:1863-1872.

Morrow, L.A., F.L. Young, and D.G. Flora. 1982. Seed germination and seedling emergence of jointed goatgrass (Aegilops cylindrica). Weed Sci. 30:395-398.

Naranjo, T., and P. Fernandez-Rueda. 1996. Pairing and recombination between individual chromosomes of wheat and rye in hybrids carrying the Ph1b mutation. Theor. Appl. Genet. 93:242-248.

National Jointed Goatgrass Research Program. 2003. Website: http://, verified 18 February 2004.

Percival, J. 1921. The wheat plant. Duckworth, London.

Pester, T.A., P. Westra, R.L. Anderson, D.J. Lyon, S.D. Miller, P.W. Stahlman, F.E. Northam, and G.A. Wicks. 2000. Secale cereale interference and economic thresholds in winter Triticum aestivum. Weed Sci. 48:720-727.

Peterson, C.D., D.A. Pearman, and T.D. Dines. 2002. New Atlas of the British flora. Oxford University Press, London.

Rajki, E. 1962. Effect on pollination of the amount of pollen on the surface of the stigma. (In Hungarian.) Novenytermeles 11:35-44.

Rieseberg, L.H. 1991. Homoploid reticulate evolution in Helianthus: Evidence from ribosomal genes. Am. J. Bot. 78:1218-1237.

Rieseberg, L.H. 1997. Hybrid origins of plant species. Annu. Rev. Ecol. Syst. 28:359-389.

Rieseberg, L.H., and J.F. Wendel. 1993. Introgression and its consequences in plants, p. 70-109. In R.G. Harrison (ed.) Hybrid zones and the evolutionary process. Oxford University Press, Oxford. UK.

Rieseberg, L.H., O. Raymond, D.M. Rosenthal, Z. Lai, K. Livingstone, T. Nakazato, J.L. Durphy, A.E. Schwarzbach, L.A. Donovan, and C. Lexer. 2003. Major ecological transitions in wild sunflowers facilitated by hybridization. Science 301:1211-1216.

Rieseberg, L.H., and S.E. Carney. 1998. Plant hybridization. New Phytol. 140:599-624.

Riley, R. 1960. The diploidization of polyploid wheat. Heredity 15: 407-429.

Riley, R. 1966. The genetic relationship of meiotic behavior in wheat and its relatives, p. 395-408. In J. MacKay (ed.) Proc. 2nd Int. Wheat Genet. Syrup., Lund, Sweden. Berlingska Boktryckeriet, Lund, Sweden.

Riley, R., and V. Chapman. 1958. Genetic control of the cytogenetically diploid behavior of hexaploid wheat. Nature 182:713-715.

Riley, R., and V. Chapman. 1967. The inheritance in wheat of crossability with rye. Genet. Res. (Cambridge) 9:259-267.

Riley, R., V. Chapman, and G. Kimber. 1959. Genetic control of chromosome pairing in intergenic hybrids with wheat. Nature 183:1244-1246.

Rimpau, W. 1891. Kreuzungs produkte landwirtschaftlicher kulturpflanzen. (In German.) Landwirtschaft. Jahrb. 20:335-371.

Sadykov, A.M. 1952. Origin of rye in wheat plant. (In Russian.) Ph.D. Thesis submitted to Institute of Genetics, Moscow, Russia.

Sage, G.C., and M.J. De Isturiz. 1974. The inheritance of anther extrusion in two spring wheat varieties. Theor. Appl. Genet. 45:126.

Sears, E.R. 1976. Genetic control of chromosome pairing in wheat. Annu. Rev. Genet. 10:31-51.

Sears, E.R. 1977. An induced mutation with homoeologous pairing in common wheat. Can. J. Genet. Cytol. 19:585-593.

Sears, E.R., and M. Okamoto. 1958. p. 258-259. In Genetics Society of America & Genetics Society of Canada (ed.) Intergenomic chromosomal relationships in hexaploid wheat. Proc. X Intl. Congr. Genet. Vol II, Montreal, Canada. Aug. 20-27. Univ. of Toronto Press, Ontario, Canada.

Sears, E.R., and T.E. Miller. 1985. The history of Chinese Spring wheat. Cereal Res. Commun. 13:261-263.

Seefeldt, S.S., R. Zemetra, F.L. Young, and S.S. Jones. 1998. Production of herbicide-resistant jointed goatgrass (Aegilops cylindrica) x wheat (Triticum aestivum) hybrids in the field by natural hybridization. Weed Sci. 46:632-634.

Sitch, L.A., J.W. Snape, and S.F. Firman. 1985. Intrachromosomal mapping of crossability genes in wheat (Triticum aestivum). Theor. Appl. Genet. 70:309-314.

Shape, J.W., V. Chapman, J. Moss, C.E. Blanchard, and T.E. Miller. 1979. The crossability of wheat varieties with Hordeum bulbosum. Heredity 42:291-298.

Snyder, J.R., C.A. Mallory-Smith, S. Balter, J.L. Hansen, and R.S. Zemetra. 2000. Seed production on Triticum aestivum by Aegilops cylindrica hybrids in the field. Weed Sci. 48:588-593.

Soltis, D.E., and P.S. Soltis. 1993. Molecular data and the dynamic nature of polyploidy. Crit. Rev. Plant Sci. 12:243-273.

Soltis, D.E., and P.S. Soltis. 1999. Polyploidy: Recurrent formation and genome evolution. Trends. Ecol. Evol. 14:348-352.

Stace, C.A. 1987. Hybridization and the plant species, p. 115-127. In K.M. Urbanska (ed.) Differentiation patterns in higher plants. Academic Press, New York.

Stace, C.A. 1991. New flora of the British Isles. Cambridge Univ. Press, Cambridge, UK.

Sun, M., and H. Corke. 1992. Population genetics of colonizing success of weedy rye in Northern California. Theor. Appl. Genet. 82:321-329.

Suneson, C.A., and E.L. Cox. 1964. Promiscuity ha barley and wheat. Crop Sci. 4:233-234.

Swan, D.G. 1984. Jointed goatgrass, p. 4. Pacific Northwest Ext. Publ. PNW0256.

Tanner, D.G., and D.E. Falk. 1981. The interaction of genetically controlled crossability in wheat and rye. Can. J. Genet. Cytol. 23: 27-32

Taylor, J.W., and K.S. Quisenberry. 1935. Inheritance of rye crossability in wheat hybrids. J. Am. Soc. Agric. 27:149-153.

Ueno, K., and H. Itoh. 1997. Cleistogamu in wheat: Genetic control and the effect of environmental conditions. Cereal Res. Comm. 25:185-189.

van Slageren, M.W. 1994. Wild Wheats: A monograph of Aegilops L. and Amblyopyrum (Jaub. & Spach) Eig (Poaceae). Wageningen Agric. Univ. Papers 94-7, Wageningen, the Netherlands.

Virmani, S.S., and I.B. Edwards. 1983. Current status and future prospects for breeding hybrid rice and wheat. Adv. Agron. 36: 145-214.

Waines, J.G., and D. Barnhart. 1992. Biosystematic research in Aegilops and Triticum. Hereditas 116:207-212.

Waines, J.G., and S.G. Hegde. 2003. Intraspecific gene flow in bread wheat (Triticum aestivum L.) as affected by reproductive biology and pollination ecology of wheat flowers. Crop Sci. 43:451-463.

Wang, G.Z., N.T. Miyashita, and K. Tsunewaki. 1997. Plasmon analyses of Triticum (wheat) and Aegilops: PCR-single-stranded conformational polymorphism (PCR-SSCP) analyses of organcllar DNAs. Proc. Natl. Acad. Sci. USA 94:14570-14577.

Wang, Z., R.S. Zemetra, J. Hansen, A. Hang, C.A. Mallory-Smith, and C. Burton. 2002. Determination of the paternity of wheat (Triticum aestivum L.) x jointed goatgrass (Aegilops cylindrica Host) B[C.sub.1] plants by using genomic in situ hybridization (GISH) technique. Crop Sci. 42:939-944.

Wang, Z., R.S. Zemetra, J. Hansen, and C.A. Mallory-Smith. 2001. The fertility of wheat x jointed goatgrass hybrid and its backcross progenies. Weed Sci. 49:340-345.

Watanabe, N., and T. Kawahara. 1999. Aegilops species collected in California and Oregon. USA. Wheat Inf. Serv. 89:33-36.

Wendel, J.F., P.D. Olson, J. Mc, and D. Stewart. 1989. Genetic diversity, introgression, and independent domestication of old world cultivated cottons. Am. J. Bot. 76:1795-1806.

Westra, P., and T. Davis. 1987. Jointed goatgrass in winter wheat, p. 104. Colorado State Univ. Cooper. Ext. Bull. 3.

White, R.H., and R.A. Morrison. 1998. IMI wheat: New weed management technology. Proc. West. Soc. Weed Sci. 51:106-107.

Yakubtsiner, M.M. 1952. Presence of rye grains in ears of wheat. (In Russian.) Agrobiology 1: 24-38.

Zemetra, R.S., J. Hansen, and C.A. Mallory-Smith. 1998. Potential for gene transfer between wheat (Triticum aestivum) and jointed goatgrass (Aegilops cylindrica). Weed Sci. 46:313-317.

Zheng, Y.L., M.C. Luo, C. Yen, and J.L. Yang. 1992. Chromosome location of a new crossability gene in common wheat. Wheat Inf. Ser. 75:36-40.

Zohary, D., and M. Feldman. 1962. Hybridization between amphidiploids and the evolution of polyploids in the wheat (Aegilops-Triticure) group. Evolution 16:44-61.

God Provides for His Creation

What does all this mean? God is the provider, as memorably expressed in Philippians 4:19 : “ My God will supply every need of yours according to his riches in glory in Christ Jesus. ” God provides for his creatures, too (see Psalm 145:15 Luke 12:24 ).

God ’s provision makes sense because God is love. Recent scientific findings are giving us a glimpse into how creatures might have flourished even after the cataclysmic destruction of the worldwide Flood. It is amazing to see how animals in all their wonderful variety have persisted since creation , reflecting God ’s nature as Jehovah Jireh (“the Lord will provide”).

If he did that for creatures that are not made in his image, how much more must he have provided for us! Not only has he designed his world to provide humankind with food, water, shelter, and resources, he has provided for our salvation.

Death and suffering were not always a part of our world. They entered because Adam declared independence from God . His sin brought great strife and loss to all God ’s creatures. But God never stopped providing, never stopped equipping us to survive in this cursed world ( Genesis 3).

Our Creator knew we could never be in right relationship with him on our own terms. But his love knows no bounds, so he miraculously became one of us and dwelt among us ( John 1:14 ). As the God -Man, Jesus provided a way to God that we could never have blazed ourselves. He became one of us and took the penalty of mankind’s sins, so that we wouldn’t have to suffer it ourselves. What he asks in return is our surrender and love.

As you marvel at the way species change to survive in a dangerous world, and how current evidence supports creation research over the last 100 years, never forget the main truth. It is our Creator who gave us the ability to learn about him and his ways, and who loved us enough to provide the essential tools to live an abundant life with him, both now and forever.

Watch the video: ΔΗΜΗΤΡΗΣ ΚΟΝΤΟΛΑΖΟΣ = ΜΑ ΠΩΣ ΕΙΝΑΙ ΔΥΝΑΤΟΝ (August 2022).