741

Use of half-tetrad analysis to discriminate betweeIl two types of 2n egg formation in a potato haploid

JOANNA E. WERNER Department of Biological Sciences, University of California, Santa Barbara, CA 93106, U.S.A.

AND

DAVID S. DOUCHES AND ROSANNA FREYRE Department of Crop and Soil Sciences, Michigan State University, East Lansing, MI48824, U.S.A.

Corresponding Editor: G. Fedak

Received January 13, 1992

Accepted May 7, 1992

WERNER, J. E., DOUCHES, D. S., and FREYRE, R. 1992. Use of half-tetrad analysis to discriminate between two types of 2n egg formation in a potato haploid. Genome, 35: 741-745.

r

The ratio of the first division restitution (FDR) to second division restitution (SDR) 2n eggs was estimated in 4182t, a haploid (2n = 2x = 24) of Solanum tuberosum L. that produces 2n eggs by the two modes. The segregation of three genes previously mapped relative to their centromeres, Pgm-2 (2.0 cM), Mdh-1 (33.5 cM), and 6-Pgdh-3 (30.1 cM) was analyzed in the tetraploid offspring of a 2x x 4x cross. Based on the segregation of the Pgm-2 locus, 39.70/0 of the progeny originated from FDR 2n eggs and 60.30/0 from SDR. Segregation patterns of the two distal loci within the FDR-derived 4x subpopulation indicated that the gene-centromere recombination rate during megasporogenesis was significantly reduced for Mdh-1 when compared with a previous estimate during microsporogenesis. In the SDR­derived 4x subpopulation, the gene-centromere recombination rates for Mdh-1 and 6-Pgdh-3 were not significantly different from previous estimates. Tetraploid progeny generated from one 2x x 4x cross where the 2x parent produces 2n gametes by two modes can be used to make an unbiased comparison of the potential breeding value of FDR and SDR gametes.

Key words: potato, megasporogenesis, first division restitution, second division restitution, isozyme.

WERNER, J. E., DOUCHES, D. S., et FREYRE, R. 1992. Use of half-tetrad analysis to discriminate between two types of 2n egg formation in a potato haploid. Genome, 35 : 741-745.

Chez l'haplolde 4182t (2n = 2x = 24) de Solanum tuberosum L., nlquelle possede deux modes de production d'oospheres 2n, Ie rapport entre la premiere division de restitution (PDR) et la seconde division de restitution (SDR) des oospheres 2n a ete evalue. Chez les descendants tetraploldes issus d'un croisement 2x x 4x, la segregation de trois genes anterieurement cartographies en fonction de leur position par rapport au centromere, soit les Pgm-2 (2,0 cM), Mdh-1 (33,5 cM) et 6-Pgdh-3 (30,1 cM), a ete analysee. Sur la base de la segregation du locus Pgm-2, 39,7070 des des­cendants sont issus d'oospheres 2n derivees de la PDR et 60,3070 de la SDR. Les patrons de segregation des deux locus distaux al'interieur de la sous-population 4xderivee de la PDR indiquent que Ie taux de recombinaison gene-centromere au cours de la gynosporogenese a ete significativement reduit pour Ie locus Mdh-1 par comparaison aux estimes ante­rieurs etablis lors de l'androsporogenese. Chez les sous-populations 4x derivees de la SDR, les taux de recombinaison pour les locus Mdh-1 et 6-Pgdh-3 n'ont pas differe significativement des estinles anterieurs. Les descendants tetra­ploldes resultant d'un croisement 2x x 4x, dont les parents 2x produisent des gametes 2n de fa~on bimodale, peuvent etre utilises pour etablir une comparaison non biaisee de la valeur potentielle d'amelioration des gametes PDR et SDR.

Mots cles : pomme de terre, gynosporogenese, premiere division de restitution, seconde division de restitution, isozyme. [Traduit par la redaction]

Introduction since FDR and SDR differ in their genetic consequences. Occurrence of 2n gametes (gametes with the sporophytic FDR 2n gametes transmit greater heterozygosity from the

chromosome number) has been well documented in potato, 2x parents to the 4x progeny than SDR 2n gametes; higher Solanum tuberosum (2n == 4x = 48) (den Nijs and levels of heterozygosity have been considered to contribute Peloquin 1977; Watanabe and Peloquin 1989; Werner and to the higher yields of 4x progeny generated from 4x x 2x Peloquin 1991). The identification of a high frequency of (FDR) in comparison to 4x from 4x x 2x (SDR) crosses both 2n gamete producing plants and 2n gametes per se (Mok and Peloquin 1975; Mediburu and Peloquin 1977; among many wild species and haploids extracted from the Peloquin et ale 1989). Various methods have been employed cultivated tetraploids indicates the importance of sexual to determine the mode of 2n gamete production. Cytologi­polyploidization in the origin and evolution of the polyploid cal and genetic analyses have revealed that 2n pollen in series in potato. These findings led to the development of potato is predominantly formed owing to the parallel modern breeding methods of the cultivated tetraploid spindles mechanisnl which leads to FDR-type gametes (Mok potato; production of 4x progeny via 4x-2x (2x x 4x, et al. 1976), while SDR is the principal mode in the forma­4x x 2x) and 2x-2x crosses as a means to effectively exploit tion of 2n eggs (Stelly and Peloquin 1986; Werner and the abundant genetic diversity of the diploid Solanum species Peloquin 1991). The last authors, using microscopic observa­(Peloquin et al. 1989). tions, reported the occurrence of 2x potato clones (haploids

There are two main modes of 2n gamete formation in of S. tuberosum and wild species) that produced 2n eggs by potato: first division restitution (FDR) and second division more than one mechanism representing the two genetic restitution (SDR). Determination of the mode is important, modes. Printed in Canada / Imprime au Canada

742 GENOME, VOL. 35, 1992

Diploid clones that form 2n gametes provide a means to perform half-tetrad analysis (HTA) in 4x progeny from 2x-4x crosses as a genetic test, either for gene mapping in relation to the centromere when the mode is known, or to determine the mode(s) of 2n gamete formation when gene­centromere relationships are estimated. The basis for HTA is that two strands of a bivalent from the diploid parent are recovered together in tetraploid progeny. HTA has been applied to a nunlber of organisms that form 2n gametes, e.g., maize (Rhoades and Dempsey 1966; Nel 1975), rain­bow trout (Thorgaard et al. 1983), paradise fish (Gervai and Csanyi 1984). In potato, gene-centromere linkages have been estimated for a yellow tuber flesh marker (Y) (Mok et al. 1976; Stelly and Peloquin 1986; Douches and Quiros 1987), blue pigmentation locus (P) (Mendiburu and Peloquin 1979), and 11 isozyme loci (Douches and Quiros 1987, 1988). Because of their codominant expression, isozyme nlarkers offer unbiased estimations compared with most morpho­logical markers. In addition, the linkage data indicate that these markers are distributed over at least 8 of 24 chromo­some arms in the potato genome with a wide range of distances from the centromeres (0.9-33.5 cM)

A marker tightly linked to the centromere is of special value for HTA when a 2x clone producing 2n gametes het­erozygous for a marker is crossed to a monoallelic 4x clone (Aa x aaaa). In this case, no recombination is expected between the centronlere and the locus. If 2n gametes are formed by FDR, all progeny would be simplex for the marker (Aaaa); on the other hand, for 2n gametes formed by SDR, half of the progeny would be duplex (AAaa) and the other half nulliplex (aaaa). If a 2x clone produces 2n gametes by both modes, the segregation of a tightly linked marker in a 2x-4x cross would provide an estinlate of the ratio of FDR and SDR 2n eggs within a 2x parent. The situa­tion is more conlplex for a gene distal to the centromere (single exchange tetrad is expected), since the differences between the two modes of 2n egg formation are less pronounced (25070 of nulliplex with FDR and 0070 with SDR) (Mendiburu and Peloquin 1979).

The objective of this study was to apply HTA to deter­mine the frequency of FDR and SDR 2n eggs in a haploid of Solanum tuberosum that had been cytologically deter­mined to produce 2n eggs by both modes. Cytological classification alone can often lead to a biased estimation, while HTA takes into account only those 2n gametes which contribute to the offspring.

Materials and methods

A haploid 4182t (2n = 2x = 24) extracted from the advanced breeding line W231 of Solanum tuberosum has been cytologically determined to produce 2n eggs by three mechanisms: delayed first meiotic division and omission of the second meiotic division (Warner and Peloquin 1991) and a synaptic variant (Werner and Peloquin 1987). This clone was cytologically reexamined to deter­mine the frequencies of the different mechanisms by the whole ovule clearing technique as described by Stelly et al. (1984).

The haploid has been determined to be heterozygous for three unlinked isozyme loci: Pgm-2, 6-Pgdh-3, and Mdh-l, previously mapped 2.0, 30.1, and 33.5 cM from their centromeres, respec­tively (Douches and Quiros 1987). The tetraploid used as a pollen parent in 2x x 4x crosses, NDD277-2, is homozygous (nulliplex) for all three markers. The 2x x 4x crosses were performed in the greenhouse during the winter months with supplemental lighting (16-h daylength). The average greenhouse temperature was 25°C

TABLE 1. Segregation of 4x progeny from the 4182t x NDD277-2 (2x x 4x, pgm_2223 xpgm_22222222

) cross for the Pgm-2 locus

Locus No. of 070 of genotype Mode 4x progeny 4x progeny

pgm_22222223 (simplex) FDR 116 39.7 pgm_22222323 (duplex) SDR 98 33.6 pgm-22222222 (nulliplex) SDR 78 26.7

Total 292 100

The leaf tissue from 4-week-old greenhouse seedlings (2n = 4x) was sampled for electrophoretic analyses according to Quiros (1981) and Douches and Ludlam (1991). x2 tests and 95070 confidence intervals were calculated to determine the gene-centromere recom­bination rates for both FDR and SDR 2n eggs. The segregation of the Pgm-2 locus (2.0 cM from the centromere) was used to discriminate between FDR- and SDR-derived progeny. All sinlplex progeny (pgm_2 2222223

) were designated FDR-derived, while duplex and nulliplex progeny (pgm_22222323 and Pgm_22222222

,

respectively) were SDR-derived. The segregation of the distal loci (Mdh-l and 6-Pgdh-3) was examined in each of the subpopula­tions defined by the segregation of the Pgm-2 locus to estimate recombination rates in the 2x parent.

Results Cytological analysis indicated that in the 4182t haploid,

there were 56.2 and 56.4070 2n megaspores and 2n mega­gametophytes, respectively. The rate of aborted ovules was 29.2070. Among a total of 421 ovules observed, 108 repre­sented stages of meiosis leading to 2n megaspores that allow classification of a developing megaspore into either FDR or SDR mode: 15070 of the cells analyzed at meiosis displayed delayed first meiotic division (both FDR and SDR), 36.4070 synaptic variant (FDR), and 48.6070 omission of the second meiotic division (SDR).

Two hundred and ninety-two seedlings were electropho­retically examined. Figure 1 presents the banding patterns for the PGM system of segregating 4x progeny from the 2x x 4x cross. The segregation ratio of the Pgm-210cus in the 4x offspring indicated that 116 plants were simplex (pgm-22222223 ), 98 duplex (pgm_22222323 ), and 78 nulliplex (pgm_22222222 ). There was no significant difference between the frequency of duplex and nulliplex progeny (x 2 = 2.27, p = 0.52). The results indicate that 39.7070 of the progeny originated from FDR 2n eggs and 60.3070' from SDR (Table 1).

The segregation data for the FDR-derived subpopulation are summarized in Table 2. The observed recombination rates for the two distal loci (21.6070 for Mdh-l and 25.9070 for 6-Pgdh-3) were lower than expected (33.5 and 30.1070, respectively); however, the gene-centromere recombination rate for the 6-Pgdh-3 locus was not significantly different from the expected rate. The expected values represent the distance of the locus in relation to the centromere based on the frequency of recombinants as previously determined by Douches and Quiros (1987). The segregation of the SDR­derived subpopulation is summarized in Table 3. Accord­ing to paired t-tests, the observed gene-centromere recom­bination rate for the Mdh-Ilocus (28.7070) was not signifi­cantly lower (p > 0.05) than expected (33.5070), whereas the observed recombination rate for the 6-Pgdh-310cus (35.0070) was not greater than expected (30.1070).

743 WERNER ET AL.

.J

FIG. 1. Segregating 4x progeny from the cross 4182t x NDD277-2 (2x x 4x) for the Pgm-2locus (two lower bands). The nulliplex, duplex, and simplex genotypes are designated by N, D, and S, respectively.

TABLE 2. Gene-centromere recombination rates for distal isozyme loci in FDR­derived 4x progeny of the 2x-4x mating

Gene-centromere Genotypic classification recombination rate (0/0)

Progeny Locus size Nulliplex Duplex Simplex Observed Expecteda

Mdh-I 116 12 13 91 21.6±3.8 33.5 ±4.4 6-Pgdh-3 116 15 101 b 25.9±4.1 30.1 ±4.7

QFrom Douches and Quiros (1987). bSum of both duplex and simplex genotypes.

TABLE 3. Gene-centromere recombination rates for distal isozyme loci in SDR­derived 4x progeny of the 2x-4x mating

Gene-centromere Genotypic classification recombination rate (%)

Progeny Locus size Nulliplex Duplex Simplex Observed Expecteda

Mdh-I 176 34 41 101 28.7±3.4 33.5 ± 6.0 6-Pgdh-3 173 26 147 b

35~0±3.6 30.1 ±6.3

QFrom Douches and Quiros (1987). bSum of both duplex and simplex genotypes.

Discussion The mode of 2n egg formation is of practical significance.

For 2x clones producing gametes with the unreduced chro­mosome number exclusively by one mechanism, cytologi­cal analysis is a quick and sufficient method to determine the mode of 2n gamete formation. However, when a 2x parent produces 2n eggs by both modes, microscopic examination may be inconclusive. Cytological estimation does not take into consideration the number of ovules that abort owing to improper chromosome distribution or nuclear restitution events necessary for the formation of FDR 2n eggs. The percentage of aborted ovules in the reported haploid accounts for the discrepancy in the num­ber of FDR and SDR 2n eggs between cytological and genetic analyses. Both delayed meiotic division and the synaptic variant have been associated with abortion of ovules (Werner and Peloquin 1991). On the other hand, clones producing 2n eggs owing to omission of the second meiotic division, a mechanism that has balanced divisions during mega­sporogenesis, usually have a low number of aborted ovules (Werner and Peloquin 1991). HTA together with cytologi­

cal analysis allows for unbiased estimation of viable FDR and SDR 2n gametes that contribute to the 4x progeny in interploidy crosses.

The separation into SDR and FDR subpopulations was based on one proximal locus (Pgm-2; 2 cM). Misclassifica­tion is occasionally expected owing to recombination. With a population size of 292, assuming a 2010 recombination rate, approximately two FDR-derived and four SDR-derived progeny would be misclassified. Douches and Quiros (1987) used two other proximal loci (00t-1 and Sdh-l) to discriminate between FDR and SDR 2n eggs in 2x X 4x crosses. Examination of these additional loci would reduce the misclassification, but these loci were not heterozygous in the haploid parent. Since the number of misclassifications would have been small relative to the total population size, the observed ratio between FDR- and SDR-derived prog­eny was not distorted and should not have significantly altered the observed gene-centromere recombination rate estimates for Mdh-1 and 6-Pgdh-3.

FDR 2n gametes transmit more heterozygosity from the 2x parent to the 4x progeny than SDR gametes (Peloquin

I

744 GENOME, VOL. 35, 1992

et ale 1990). The level of heterozygosity may be important in tetrasomic polyploids; maximum heterozygosity leads to heterosis for polygenic traits for which inter- and intra-locus interactions are important, e.g., total tuber yield in potato. Based on the results reported herein, the cited superiority of FDR 2n gametes per se does not correlate to the superiority of 4x FDR-derived progeny. There is no evidence for ganletophytic selection or developing seed competition between embryos that would favor FDR over SDR 2n gametes. Survival of the particular mode of 2n gametes seems to be unaffected by the mechanism of 2n gamete formation.

Lower gene-centromere recombination rates than previ­ously estimated were observed for the distal loci in the FDR subpopulation (Table 2), whereas recombination rates in the SDR subpopulation were not conclusive (Table 3). These results may be attributed to the synaptic variant that is expressed at megasporogenesis. Several synaptic variants have been identified in potato that affect microsporogenesis (Johnston et al. 1981; Okwaugwu and Peloquin 1981; Iwanaga 1984; Ramanna 1983). In general, cytological observations of these mutants indicate a lack of pairing during diplotene and diakinesis stages, followed by random distribution of the chromosomes during the first division. For example, the sy3sy3 genotype leads ultimately to pollen sterility unless acconlpanied by the parallel spindles mech­anism (Okwaugwu and Peloquin 1981). Douches and Quiros (1988) estimated 89070 reduction in recombination in a synaptic mutant of the potato (sy3sy3, psps) that produced FDR 2n pollen. The reduction in reconlbination for the distal loci reported herein was considerably less pronounced (35.5070 for Mdh-l, 13.6070 for 6-Pgdh-3). This may in part be due to the facts that two mechanisms lead to FDR 2n eggs in 4182t haploid and it is unknown whether the delayed meiotic division mechanism affects recombination. To esti­mate the effect that the synaptic variant has on recombina­tion during megasporogenesis, one needs to measure the recombination rate in diploid clones that produce 2n eggs solely by this mechanism. Jongedijk et al. (1991) determined that sex differences in recombination did not occur in desynaptic (ds-lds-l) diploid potato clones. Under these assumptions that the same level of recombination occurs in the synaptic variant affecting megasporogenesis as in the synaptic variant expressed in microsporogenesis (sy3sy3), and the delayed meiotic nlechanism does not reduce recom­bination, one could estimate that about 40070 of the viable FDR 2n eggs were derived from the synaptic variant. Cyto­logical analysis, however, indicated 70.8070 of FDR 2n eggs to be produced by the synaptic variant and 29.2070 owing to delayed meiotic division. Therefore, considering low viability of the 2n eggs derived from delayed nleiotic division (Werner and Peloquin 1991), it is unlikely that the synaptic variant reduced recombination in megasporogenesis to the same extent as those variants observed in microsporogenesis.

The predominant mode of 2n egg formation in potato is SDR (omission of the second meiotic division). However, Werner and Peloquin (1991) reported that of the 2n egg producing clones, 42070 of the haploids and 47070 of the wild species formed 2n eggs by two modes, mainly owing to synaptic variant and omission of second division. The mix­ture of FDR and SDR 2n eggs in one clone was also indicated by Stelly and Peloquin (1986) based on genetic analysis, and by Conicella et al. (1991) from cytological observation.

Furthermore, based on the number of mechanisms described for 2n gametes in potato, 2n eggs are formed by more divergent mechanisms than 2n pollen. Genes controlling 2n egg formation are less exposed to selection pressure than genes controlling 2n pollen production owing to both the course of gametogenesis and the biology of fertilization. Occurrence of FDR 2n eggs among SDR 2n eggs might be also an evolutionary means to introduce more heterozygos­ity from the parent to the progeny in a manner similar to occasional outcrossing that was found in otherwise self­pollinated plants.

From the practical point of view, occurrence of 2n potato clones producing 2n gametes by two modes creates the ideal conditions for unbiased comparison of the value of both FDR and SDR 2n gametes. In the experinlents conducted to date, owing to unavailability of desired clones, the breeding value of FDR and SDR 2n ganletes has been par­tially confounded with allelic diversity of the 2x parents since different 2x clones producing either FDR or SDR 2n pollen were used in 4x-2x crosses (Mok and Peloquin 1975). When clones that form 2n gametes by two genetically different modes are utilized, the 4x progeny obtained from 2x(FDR) x 4x and 2x(SDR) x 4x crosses would be gener­ated from the same genotypes. Moreover, the mode of 2n gamete formation in the 2x parent that led to 4x progeny can be determined by HTA. Therefore, the differences detected between 4x progeny from such a cross could be attributed to the differences in transmission of heterozygos­ity through 2n gametes.

Acknowledgements The haploid was kindly supplied by Dr. S. J. Peloquin,

University of Wisconsin, Madison. This research was sup­ported in part by the Michigan Agricultural Experiment Station and the Michigan Potato Industry Commission.

Conicella, C., Barone, A., DelGiudice, A., et ale 1991. Cytologi­cal evidence of SDR-FDR mixture in the formation of 2n eggs in a potato diploid clone. Theor. Appl. Genet. 81: 59-63.

den Nijs, T.P.M., and Peloquin, S.J. 1977. 2n gametes in potato species and their function in sexual polyploidization. Euphytica, 26: 585-600.

Douches, D.S., and Ludlam, K. 1991. Electrophoretic characteriza­tion of North American potato cultivars. Am. Potato. J. 68: 767-780.

Douches, D.S., and Quiros, C.F. 1987. Use of 4x-2x crosses to determine gene-centromere map distance of isozyme loci in Solanum species. Genome, 29: 519-527.

Douches, D.S., and Quiros, C.F. 1988. Genetic recombination in a diploid synaptic mutant and a Solanum tuberosum x S. chacoense diploid hybrid. Heredity, 60: 183-191.

Gervai, J., and Csanyi, V. 1984. Artificial gynogenesis and map­ping of gene-centromere distance in the paradise fish, Macropodus opercularis. Theor. Appl. Genet. 68: 481-485.

Iwanaga, M. 1984. Discovery of a synaptic mutant in potato haploids and its usefulness for potato breeding. Theor. Appl. Genet. 68: 87-93.

Johnston, S.A., Hanneman, R.E., and Ruhde, R.W. 1981. Iden­tification of an asynaptic variant in Solanum commersonii. Am. Potato J. 58: 506.

Jongedijk, E., Hutten, R.C.B., Van Der Wolk, J.M.A.S., et ale 1991. Synaptic mutants in potato, Solanum tuberosum L. III. Effect of the Ds-I/ ds-Ilocus (desynapsis) on genetic recom­bination in male and female meiosis. Genome, 34: 121-130.

745 WERNER ET AL.

Mendiburu, A.D., and Peloquin, S.J. 1977. The significance of 2n gametes in potato breeding. Theor. Appl. Genet. 48: 137-143.

Mendiburu, A.D., and Peloquin, S.J. 1979. Gene-centromere map­ping by 4x-2x matings in potatoes. Theor. Appl. Genet. 54: 177-180.

Mok, D.W.S., and Peloquin, S.J. 1975. Breeding value of 2n pollen (diplandroids) in tetraploid x diploid crosses in potatoes. Theor. Appl. Genet. 46: 307-314.

Mok, D.W.S., Peloquin, S.J., and Mendiburu, A.D. 1976. Genetic evidence for the mode of 2n pollen formation and S-locus map­ping in potatoes. Potato Res. 19: 157-167.

Nel, P.M. 1975. Crossing over and diploid egg formation in the elongate nlutant of maize. Genetics, 79: 435-450.

Okwaugwu, C.O., and Peloquin, S.J. 1981. A nlethod of transfer­ring the intact parental genotype to the offspring via meiotic mutants. Am. Potato J. 58: 512-513.

Peloquin, S.J., Yerk, G.L., Werner, J.E., and Darmo, E. 1989. Potato breeding with haploids and 2n gametes. Genome, 31: 1000-1004.

Peloquin, S.J., Yerk, G.L., and Werner, J.E. 1990. Ploidy manip­ulation in the potato. In Chromosomes: eukaryotic, prokaryotic, and viral. Vol. II. Edited by K.W. Adolph. CRC Press, Boca Raton, Fla. pp. 167-178.

Quiros, C.F. 1981. Starch electrophoresis technique used with alfalfa and other Medicago species. Can. J. Plant Sci. 61: 745-749.

Ramanna, M.S. 1983. First division restitution gametes through fertile desynaptic mutants of the potato. Euphytica, 32: 337-350.

Rhoades, M.M., and Dempsey, E. 1966. Induction of chromosome doubling at meiosis by the elongate gene in maize. Genetics, 54: 505-522.

Stelly, D.M., and Peloquin, S.J. 1986. Diploid female gametophyte formation in 24-chromosome potatoes: genetic evidence for the prevalence of second meiotic division restitution mode. Can. J. Genet. Cytol. 28: 101-108.

Stelly, D.M., Peloquin, S.J., Palmer, R.G., and Crane, C.F. 1984. Mayer's hemalum-methyl salicylate: a stain clearing technique for observation within whole ovules. Stain Technol. 58: 155-161.

Thorgaard, G.R., Allendorf, F.W., and Knudsen, K.L. 1983. Gene-centromere mapping in rainbow trout: high interference over long map distances. Genetics, 103: 771-783.

Watanabe, K., and Peloquin, S.J. 1989. Occurrence of 2n pollen and ps gene frequencies in cultivated groups and related species in tuber-bearing Solanums. Theor. Appl. Genet. 78: 329-336.

Werner, J.E., Peloquin, S.J. 1987. Frequency and mechanisms of 2n egg formation in haploid Tuberosum - wild species F1

hybrids. Am. Potato J. 64: 641-654. Werner, J.E., and Peloquin, S.J. 1991. Occurrence and mecha­

nisms of 2n egg formation in 2x potatoes. Genome, 34: 975-982.