The effect of B chromosomes on homoeologous pairing in species hybrids

The effect of B chromosomes on chromosome pairing at meiosis was investigated in the species hybrid Lolium temulentum x L. perenne at both the diploid and tetraploid level. The presence of B chromosomes drastically reduced association of homoeologous chromosomes in both the diploids and tetraploids. This was evident from the high frequency of univalents recorded in PMC's of diploid hybrids with B's and from the predominantly bivalent association of homologous chromosomes in tetraploids of this type. In the absence of B's homoeologous pairing was extensive giving a high frequency of bivalents in the diploids and multivalents as well as bivalents and univalents in the tetraploids.     

Chromosome relationships between Lolium and Festuca (Gramineae)

With a view to eclucidating chromosome relationships between Lolium perenne (Lp), L. multiflorum (Lm) and Festuca pratensis (Fp), chromosome pairing in different diploid (2n=14), auto-allotriploid (2n=3x=21), trispecific (2n=3x=21), amphidiploid (2n=4x=28) and auto-allohexaploid (2n=6x=42) hybrids between them was analysed. At all these levels of ploidy there was very good chiasmate pairing between the chromosomes of the three species and, on the whole, there was little evidence of preferential pairing of the chromosomes of a particular species in the triploid, tetraploid and hexaploid hybrids. A critical test for this also came from the synaptic ability of the chromosomes of the single genome with those of the duplicated genome in the auto-allotriploids which formed predominantly trivalents with 2, 3 or even 4 chiasmata. Moreover, the homology between the Lp and Lm chromosomes seems strong enough to pass the discrimination limits of the B-chromosomes which do not suppress homoeologous pairing in the Lp LmLm triploid and LpLm diploid hybrids. — The triploids having two genomes of a Lolium species and one of F. pratensis had some male and female fertility which suggested genetic compatibility of the parental chromosomes resulting, presumably, in compensation at the gametic level. Also, the occurrence of comparable chiasma frequencies in the auto-allotriploids and trispecific hybrids showed that they were not markedly affected whether two doses of one genome and one of the other or all the three different genomes from the three species were present. From the trend of chromosome pairing in all these hybrids it is concluded that there is little structural differentiation between the chromosomes of the three species, no effective isolation barrier to gene-flow between them, and that they are closely related phylogenetically, having possibly evolved from a common progenitor. Taxonomic revision of the two Lolium species is suggested.     

Genome relationships in the diploid oats

Genome relationships between the three diploid oats, Avena strigosa (S.), A. longiglumis (L.) and A. prostrata (P.) were studied by chromosome pairing in diploid hybrids and in synthetic triploids and tetraploids combining these genomes. Fairly regular pairing in the diploid hybrid and typical autopolyploid behavior in the triploids and in the amphidiploid suggest small differentiation in the chromosome architecture of A. longiglumis and A. prostrata. A. strigosa diverges from the other two oats by complex chromosome rearrangements. Conspicuous preferential pairing took place in triploids with SSL, SSP and SPP genomic constitution. The low bivalent frequency in the SLL triploid suggests that preferential pairing in triploids with two S genomes is not a consequence of chromosome rearrangement but is rather of genetic origin. The presence of the three genomes in a triploid or a tetraploid caused considerable meiotic irregularities suggesting a better pairing competition of the S genome.     

Genome relationships between Hordeum vulgare L. and H. bulbosum L.

Interrelationships between H. vulgare (2x=14) and H. bulbosum (2x=14; 4x=28) were estimated on the basis of the karyotypes and the pairing behaviour of the chromosomes in diploid, triploid and tetraploid hybrids obtained with the aid of embryo culture. — A comparison of the karyotypes of the two species revealed similarities as well as differences. It was concluded that at least 4 or more of the chromosomes were similar in morphology and probably closely related. — Diploid and tetraploid hybrids are rarely obtained and their chromosome numbers tend to be unstable whereas triploid hybrids (1 vulgare + 2 bulbosum genomes) were stable and relatively easy to produce. In the diploid hybrid only 40% of the meiotic cells contained 14 chromosomes while the numbers ranged from 7 to 16 in other cells. All hybrids exhibited pairing between the chromosomes of the two species. Diploid hybrids had a mean of 5.0 and a maximum of 7 bivalents per cell in those cells having 14 chromosomes. Triploid hybrids from crosses between 2x H. vulgare and 4x H. bulbosum exhibited a mean of 1.5 and a maximum of 5 trivalents per cell. In a hexaploid sector found following colchicine treatment of a triploid the mean frequencies of chromosome associations per cell were: 5.5_I+8.0_II+0.7_III+3.7_IV+0.3_V+0.4_VI. One unstable 27 chromosome hybrid obtained from crosses between the autotetraploid forms had a mean of 1.1 and a maximum of 4 quadrivalents per cell. The chromosome associations observed in these hybrids are consistent and are taken as evidence of homoeologous pairing between the chromosomes of the two species. Interspecific hybridization between these two species also reveals that chromosome stable hybrids are only obtained when the genomes are present in a ratio of 1 vulgare2 bulbosum. Based upon the results obtained, the possibility of transferring genetic characters from H. bulbosum into cultivated barley is discussed.     

Karyotypes,nucleoli, and amphiplasty in hybrids between Hordeum vulgare L. and H. bulbosum L.

Chromosome measurements were carried out in Hordeum vulgare, H. bulbosum, and their diploid, triploid, and tetraploid hybrids. The chromosomes were classified by using relative values, and thus karyotypes were established. For comparison of these karyotypes both relative and absolute values were used. It was concluded that differential amphiplasty occurred, whereas neutral amphiplasty could not be demonstrated. In the hybrids the relative length of the parts of the chromosomes (long arm, short arm, satellite) was not changed in comparison with these lengths in the pure species. The karyotypes of both species had considerable similarities. From comparing the mean absolute genome lengths, it was, however, concluded that in the pure species, as well as in all hybrid types, the chromosomes of H. vulgare were longer than those of H. bulbosum. In the diploid and tetraploid hybrids the mean genome lengths were shorter than those in the pure species and the triploid hybrids. The differential amphiplasty was such that the secondary constriction of chromosome 6 of H. bulbosum, did not show up in the hybrids. This could be related to the suppression of nucleolar formation in the genome of H. bulbosum, because the maximum number of nucleoli in root tip cells equalled the number of satellite chromosomes. Finally it was found that the pattern of nucleolar fusion in diploid and triploid hybrids deviated from the expectation. The results were discussed in relation to chromosomal disturbances that occurred in the hybrid tissues and that resulted in elimination of chromosomes and other effects.     

TRITICUM URARTU AND GENOME EVOLUTION IN THE TETRAPLOID WHEATS

The A genome of the tetraploid wheats (AABB, 2n = 28) shows 5-6 bivalents in crosses with Triticum boeoticum (2n = 14) and various Aegilops diploids (2n = 14). The B genome has never been similarly identified with any species, and is commonly thought to have been modified at the tetraploid level. Triticum boeoticum was presumably accepted as the A-genome donor because of its morphological similarity to the wild tetraploids and because it was formerly the only known wild diploid wheat. The B donor has been thought to be Ae. speltoides or another species of the Sitopsis section of Aegilops, but these diploids show pairing affinity with A rather than B. More recently, another diploid wheat, T. urartu, was found to be sympatric with T. boeoticum throughout the natural range of the tetraploids. The synthetic boeoticum-urartu amphiploid was virtually identical morphologically with the wild tetraploid wheats, whereas various boeoticum-Sitopsis amphiploids were markedly different. But the urartu genome, like those of T. boeoticum and Sitopsis, paired with A and not with B. However, cytological evidence also shows (1) that the genomes of any plausible parental combination pair with one another, (2) that the A and B genomes of the tetraploid wheats pair with one another in the absence of the gene Ph, and (3) that homoeologous chromosomes of the tetraploids have differentiated further, presumably as a result of diploidization. Consequently, chromosome pairing at Meiosis I can be expected to give ambiguous evidence regarding the identity of the tetraploid genomes with their parental prototypes. A hypothesis regarding the expected pairing affinities between tetraploid homoeologues that have differentiated from closely related parental chromosomes is advanced to explain the anomalous pairing behavior of the A and B genomes. Triticum boeoticum and T. urartu are inferred to be the parents of the tetraploid wheats.     

Analysis of meiosis in triticale (XTriticosecale Wittmack) X rye (Secale cereale L.) F1 hybrids at three ploidy levels

Summary Triticales (XTriticosecale Wittmack) at three ploidy levels (8x, 6x, 4x, x=7) were crossed with diploid rye (Secale cereale L.) to produce a solitary hypopentaploid hybrid (2n=32), and a number of tetraploid (2n=4x=28) and triploid (2n=3x=21) hybrids. The hybrids exhibited a morphology which was intermediate between the parents. The number of bivalents ranged from 1–7 (4.65 per cell) in hypopentaploid, from 2–12 (7.13 per cell) in tetraploid and from 4–9 (6.84 per cell) in triploid hybrids. In 4x and 3x hybrids, trivalents and quadrivalents were also observed at low frequencies (range 0–1; mean 0.01–0.03 per cell). Chiasmata frequency was highest in triploid hybrids (12.44 per cell), lowest in hypopentaploid (5.37 per cell) and intermediate in tetraploids (10.54 per cell). More than 7¹¹ were found in 39.7% pollen mother cells (PMC's) in the 4x hybrids and in 5.0% PMCs in 3x hybrids. It is concluded that an increase in the relative proportion of wheat chromosomes in the hybrids had a slight suppression effect on homologous as well as homoeologous pairing of rye chromosomes. Contrary to this, the relative increase in rye complement promoted homoeologous pairing between wheat chromosomes. In triploid hybrids, the chiasmata frequency as well as the c value were the highest, suggesting that in tetraploid hybrids rye chromosomes had a reduced pairing (low frequency of ring bivalents).     

Pairing and recombination at meiosis of Brassica rapa (AA) × Brassica napus (AACC) hybrids

Interspecific crosses contribute significantly to plant evolution enabling gene exchanges between species. The efficiency of interspecific crosses depends on the similarity between the implicated genomes as high levels of genome similarity are required to ensure appropriate chromosome pairing and genetic recombination. Brassica napus (AACC) is an allopolyploid, resulting from natural hybridization between Brassica rapa (AA) and Brassica oleracea (CC), both being diploid species derived from a common ancestor. To study the relationships between genomes of these Brassica species, we have determined simultaneously the pairing and recombination pattern of A and C chromosomes during meiosis of AAC triploid hybrids, which result from the interspecific cross between natural B. napus and B. rapa. Different AAC triploid hybrids and their progenies have been analysed using cytogenetic, BAC-FISH, and molecular techniques. In 71% of the pollen mother cells, homologous A chromosomes paired regularly, and usually one chromosome of each pair was transmitted to the progeny. C chromosomes remained mainly univalent, but were involved in homoeologous pairing in 21.5% of the cells, and 13% of the transmitted C chromosomes were either recombined or broken. The rate of transmission of C chromosomes depended on the identity of the particular chromosome and on the way the hybrid was crossed, as the male or as the female parent, to B. napus or to B. rapa. Gene transfers in triploid hybrids are favoured between A genomes of B. rapa and B. napus, but also occur between A and C genomes though at lower rates.     

A CYTOGENETIC ANALYSIS OF CERTAIN POLYPLOIDS IN GRINDELIA (COMPOSITAE)

Two diploid taxa, Grindelia procera and G. camporum, and 3 tetraploid ones, G. camporum, G. hirsutula, and G. stricta, have been studied to ascertain their interrelationships. Meiosis in diploid parental strains was regular, the common chromosome configuration being 5 rod bivalents and 1 ring bivalent. The average chiasmata frequency per chromosome was 0.60. Pollen fertility was about 90% in all strains examined. Diploid interspecific hybrids had normal meiosis with an average chiasmata frequency of 0.56 per chromosome. No heterozygosity for inversions or interchanges was detected, and pollen fertility was above 85%. Meiosis in parental tetraploid strains was characterized by the presence of quadrivalents in addition to a complementary number of bivalents. The average chiasmata frequency per chromosome was 0.59 and pollen fertility was generally about 80%. Tetraploid interspecific hybrids also had quadrivalents, normal meiosis, and high pollen fertility. Close genetic relationships between the diploids and between the tetraploids are indicated, and geographical, ecological, and seasonal barriers to gene exchange exist. Attempts to obtain hybrids between diploids and tetraploids were successful in a few cases. The hybrids were tetraploid and had normal meiosis and fertility similar to parental and F₁ tetraploids. Their origin was by the union of unreduced gametes of the diploid female parent and normal pollen from the tetraploid parent. On the basis of chromosome homology, normal meiosis, plus high fertility exhibited in the diploid, tetraploid, and diploid X tetraploid interspecific hybrids, these species of Grindelia are considered to be a part of an autopolyploid complex. Gene exchange between diploids and diploids, tetraploids and tetraploids, and diploids and tetraploids is possible. Tetraploid G. camporum may have originated by hybridization between G. procera and diploid G. camporum with subsequent doubling of chromosomes and selection for the combined characteristics of the diploids.     

SYNTHETIC HYBRIDS OF NEW WORLD AND OLD WORLD AGROPYRONS. I. TETRAPLOID AGROPYRON SPICATUM × DIPLOID AGROPYRON CRISTATUM

Four interspecific hybrids of tetraploid A. spicatum × diploid A. cristatum ‘Fairway’ were obtained by controlled pollinations of emasculated and unemasculated spikes of A. spicatum. Most vegetative and spike characteristics of the hybrids were intermediate between those of the parent species. Tetraploid A. spicatum behaved cytologically as an autotetraploid, with mean chromosome associations of 0.09 I, 7.95 II, 0.03 III, and 2.98 IV being observed in 103 cells interpreted. The diploid A. cristatum was cytologically regular and formed 7 bivalents in 160 of 163 cells examined. Meiosis in the triploid hybrids was highly irregular, and these plants were completely sterile. Chromosomes of A. spicatum and A. cristatum differed sufficiently in size so that they could be distinguished in the hybrids. Seven bivalents and 7 univalents were formed in 90.5% of 262 cells interpreted at metaphase I. Mean chromosome associations of 6.87 I, 6.98 II, and 0.05 III were observed in the hybrids. Most chromosome pairing was due to autosyndesis of A. spicatum chromosomes, but A. cristatum chromosomes occasionally paired among themselves and with A. spicatum chromosomes. Tetraploid A. spicatum was considered to be an autotetraploid. On the basis of cytological evidence, A. cristatum, A. spicatum, and their interspecific hybrids were represented by genome formulae of AA, BBBB, and ABB, respectively.     

Autosyndetic pairing in Gibasis (Commelinaceae) hybrids revealed by C-banding

Two closely related species of Gibasis, G. karwinskyana and G. consobrina, and their F₁ hybrids were studied cytologically at the diploid and tetraploid level. Despite similarity in their basic karyotype, pairing was extremely limited in the diploid hybrid and almost exclusively autosyndetic in the tetraploid, except for multivalent formation due to interchange heterozygosity. The analysis was considerably facilitated by the use of C-banding techniques at meiosis, by which the chromosomes of each species could be readily identified. In the parents, quadrivalents were formed between homologous but non-identical chromosomes, which also formed autosyndetic bivalents in the hybrids. Meiotic pairing in the hybrids was unaffected by polytypy for C-bands among different populations of the parental species.     

REPRODUCTIVE ISOLATION OF TRITICUM BOEOTICUM AND TRITICUM URARTU AND THE ORIGIN OF THE TETRAPLOID WHEATS

The diploid wheats Triticum boeoticum and T. urartu are sympatric with one another throughout the geographic range of the wild tetraploids. Reciprocal crosses between ecogeographic types within each diploid species gave viable seed, but interspecific crosses consistently gave viable seed only when T. boeoticum was the female parent. Apparently urartu cytoplasm in combination with the boeoticum genome resulted in nonviable seed. The endosperm failed to develop normally despite regular endosperm fertilization. The F₁ plants obtained were completely self sterile although they showed regular intergenomic pairing (7II) at meiosis. Presumably the accumulation of cryptic differences between the two closely related genomes under reproductive isolation accounts for this sterility. The same accumulated cryptic differences could largely account for the preferential diploid pairing in the tetrapolid wheats which presumably were derived from such hybrids by chromosome doubling. The behavior of reciprocal crosses between the diploids and tetraploids suggested that T. boeoticum contributed the cytoplasm to both of the wild tetraploid species.     

Natural hybridisation in birch: triploid hybrids between Betula nana and B. pubescens

Hybridisation between diploid (2n=28) dwarf birch Betula nana L. and tetraploid (2n=56) downy birch B. pubescens Ehrh. has occurred in natural populations in Iceland. About 10% of birch plants randomly collected are triploid (2n=42) hybrids. Ribosomal gene mapping on chromosomes and genomic in situ hybridisation confirms the hybridity. However, the triploid hybrids are not morphologically distinct, i.e. they are not different from diploid and tetraploid birch plants that have intermediate morphology. The triploid hybrids have evidently played an important role in driving bi-directional gene flow between these two species. This paper reviews the extent of interspecific hybridisation in selected birch woodland populations and discusses the significance of natural hybridisation and introgression in birch.     

Cytogenetics of the hybrid Gilia millefoliata × achilleaefolia

Summary Gilia millefoliata andG. achilleaefolia, two annual diploid (n=9) species ofPolemoniaceae, crossed readily in certain combinations but not in others. The F₁ hybrids were vigorous but sterile. They gave rise, apparently by the union of unreduced gametes, to an F₂ generation of tetraploids, which were mostly fertile.Chromosome pairing in the hybrids varied markedly according to the state of nutrition of the plants. The F₁ hybrids formed fewer clear diakinesis figures, fewer bivalents, fewer chiasmata per bivalent, and more attenuated or stretched bivalents when grown in 2 pots of sand than when grown in rich soil (Table 3). A pot-bound allotetraploid individual derived from this hybrid showed the same meiotic irregularities as the starved F₁s until irrigated with a solution of mineral nutrients, after which its chromosomes paired regularly in bivalents (Table 2, Fig. 38).The capacity of the F₁ hybrids to produce polyploids also differed strikingly in the two cultures. The rate of polyploidy of the stunted sand-grown hybrids was 2381 viable tetraploid zygotes per million flowers, while the corresponding figure for the luxuriant field hybrids was only 2.7 per million flowers.For the production of polyploid progeny by diploid parents — a process which should be clearly distinguished from normal fertility — the termpolyploidy rate is proposed. It is suggested that starvation of a structural hybrid may sometimes increase its polyploidy rate by reducing chromosome pairing to the point where restitution nuclei and hence unreduced gametes can be formed.     

CHROMOSOMES AND CROSSING BEHAVIOR OF HIBISCUS CANNABINUS,H. ACETOSELLA,AND H. RADIATUS

Menzel , Margaret Y. (Florida State U., Tallahassee), and F. D. Wilson . Chromosomes and crossing behavior of Hibiscus cannabinus, H. acetosella, and H. radiatus. Amer. Jour. Bot. 48(8): 651–657. Illus. 1961.—Chromosomes of diploid H. cannabinus L. (kenaf) form 18 bivalents at metaphase I. In autotetraploid H. cannabinus (2n = 72), more than 50% of the chromosomes pair as trivalents or quadrivalents. In the tetraploid species H. radiatus Cav. and H. acetosella Welw. ex Hiern (H. eetveldeanus De Wild. & Dur.) (2n = 72), only 4% of the chromosomes pair as multivalents and the rest pair as bivalents. Vigorous, highly fertile F₁ hybrids between H. acetosella and H. radiatus are easily obtained, show complete chromosome pairing, and give rise to a freely segregating, vigorous, fertile F₂: apparently the parental species have similar genome constitutions and are closely related. Chromosome pairing in the triploid hybrids of H. radiatus and H. acetosella with H. cannabinus, in hexaploids obtained by doubling the chromosome number of H. acetosella-cannabinus F₁, and in pentaploid and tetraploid backcrosses of the hexaploids to H. cannabinus shows that the tetraploid species each contain 1 genome (A) very similar to, but not identical with, that of H. cannabinus and 1 dissimilar genome (B). Morphology, fertility, and other characteristics of the various hybrids are discussed in connection with the problem of recombining the resistance to root-knot nematodes found in the tetraploid species with the desirable fiber properties of H. cannabinus.     

Chromosome and nucleotide sequence differentiation in genomes of polyploid Triticum species

Summary The nature of genome change during polyploid evolution was studied by analysing selected species within the tribe Triticeae. The levels of genome changes examined included structural alterations (translocations, inversions), heterochromatinization, and nucleotide sequence change in the rDNA regions. These analyses provided data for evaluating models of genome evolution in polyploids in the genus Triticum, postulated on the basis of chromosome pairing at metaphase I in interspecies hybrids.The significance of structural chromosome alterations with respect to reduced MI chromosome pairing in interspecific hybrids was assayed by determining the incidence of heterozygosity for translocations and paracentric inversions in the A and B genomes of T. timopheevii ssp. araraticum (referred to as T. araraticum) represented by two lines, 1760 and 2541, and T. aestivum cv. Chinese Spring. Line 1760 differed from Chinese Spring by translocations in chromosomes 1A, 3A, 4A, 6A, 7A, 3B, 4B, 7B and possibly 2B. Line 2541 differed from Chinese Spring by translocations in chromosomes 3A, 6A, 6B and possibly 2B. Line 1760 also differed from Chinese Spring by paracentric inversions in arms 1AL and 4AL whereas line 2541 differed by inversions in 1BL and 4AL (not all chromosomes arms were assayed). The incidence of structural changes in the A and B genomes did not coincide with the more extensive differentiation of the B genomes relative to the A genomes as reflected by chromosome pairing studies.To assay changing degrees of heterochromatinization among species of the genus Triticum, all the diploid and polyploid species were C-banded. No general agreement was observed between the amount of heterochromatin and the ability of the respective chromosomes to pair with chromosomes of the ancestral species. Marked changes in the amount of heterochromatin were found to have occurred during the evolution of some of the polyploids.The analysis of the rDNA region provided evidence for rapid fixation of new repeated sequences at two levels, namely, among the 130 bp repeated sequences of the spacer and at the level of the repeated arrays of the 9 kb rDNA units. These occurred both within a given rDNA region and between rDNA regions on nonhomologous chromosomes. The levels of change in the rDNA regions provided good precedent for expecting extensive nucleotide sequence changes associated with differentiation of Triticum genomes and these processes are argued to be the principal cause of genome differentiation as revealed by chromosome pairing studies.     

Erratum to Production of diploid and triploid offspring by inbreeding of the triploid planarian <Emphasis Type="Italic">Dugesia ryukyuensis</Emphasis>

Triploidy has generally been considered to be an evolutionary dead end due to problems of chromosomal pairing and segregation during meiosis. Thus, the formation of tetraploids and diploids from triploid types is a rare phenomenon. In the present study, we demonstrated that inbreeding of the triploid planarian Dugesia ryukyuensis resulted in both diploid and triploid offspring in nature. In the triploids of D. ryukyuensis, chiasmata between homologous chromosomes were observed in both female and male germ lines. This result suggests that both diploid and triploid offspring of this species are produced bisexually by zygotic fusion between sperm and eggs. Hence, this phenomenon may be a novel mechanism in planarian for escaping the triploid state.     

The genetics of meiotic chromosome pairing in Lolium temulentum x Lolium perenne tetraploids

Summary The degree of preferential pairing of homologous chromosomes was estimated in a series of tetraploid hybrids of Lolium temulentum x Lolium perenne by means of cytological and genetic analyses. The correlations between the frequency of bivalents at first metaphase of meiosis in the hybrid tetraploids and the degree of preferential pairing calculated from the segregation pattern of isozyme alleles in a test cross was extremely high. The results showed clearly that suppression of heterogenetic pairing in these Lolium tetraploids is achieved by a genetic system involving the A chromosomes as well as the B chromosome system which has been known for some time. Certain similarities with the genetic system controlling pairing in polyploid wheats are discussed.     

Further evidence for the presence of meiotic pairing control genes in Alopecurus L. (Gramineae)

Mathematical equations applied to data on the meiotic chromosome behaviour of diploid, triploid and tetraploid Alopecurus species, their hybrids and synthesised autopolyploids confirm that chromosome pairing among homologues does not occur at random. The genotypic control of preferential bivalent formation is demonstrated and its role in natural populations discussed.     

CYTOGENETICS OF AGROPYRON FERGANENSE AND ITS HYBRIDS WITH SIX SPECIES OF AGROPYRON,ELYMUS, AND SITANION

Meiosis and mode of reproduction are described in Agropyron ferganense Drob., a perennial forage grass from Central Asia. This species is diploid (2n = 14); it exhibits normal meiosis and reproduces by cross-pollination. Hybrids were produced between A. ferganense and six species with known genome formulas: 1) North American A. spicatum (Pursh) Scribn. & Smith, an SS diploid (2n = 14), 2) Middle Eastern A. libanoticum Hack., an SS diploid (2n = 14), 3) North American A. dasystachyum (Hook.) Scribn., an SSHH tetraploid (2n = 28), 4) Eurasian A. caninum (L.) Beauv., an SSHH tetraploid (2n = 28), 5) North American Sitation hystrix (Nutt.) J. G. Smith, an SSHH tetraploid (2n = 28), and 6) South American Elymus patagonicus Speg., an SSHHHH hexaploid (2n = 42). Almost complete chromosome pairing in the A. ferganense x A. spicatum and A. libanoticum hybrids demonstrated that A. fergenanse is an SS diploid, but it is genetically isolated from the other SS diploids because of high sterility in the F₁ hybrids. S-genome diploids form a network of species that extend from the Middle East through Central Asia to western North America. Frequent occurrence of seven univalents and seven bivalents at metaphase I in the triploid hybrids of A. ferganense x A. dasystachyum, A. caninum and S. hystrix was consistent with the proposed genome formulas of SS for A. ferganense, SSHH for the three tetraploid species, and SSH for the hybrids. Chromosome pairing was highly variable in the A. ferganense x E. patagonicus hybrids; however, some cells had almost complete bivalent pairing, an expected observation in an SSHH hybrid from a cross between an SS diploid (A. ferganense) and an SSHHHH hexaploid (E. patagonicus). Various options were considered concerning the appropriate generic classification of the S-genome diploids, which are now commonly placed in Agropyron. The inclusion of these species in the genus Eiytrigia, as advocated by some Soviet taxonomists, appears to be a reasonable decision.