ASSOCIATION BETWEEN QUANTITATIVE TRAITS AND ALLOZYME HETEROZYGOSITY IN A TETRASOMIC SPECIES: DACTYLIS GLOMERATA

Tetraploid individuals of orchardgrass (Dactylis glomerata L. subsp. hispanica Roth.) sampled from a natural population were used to evaluate the correlation between both single and multilocus heterozygosity at 7 enzyme loci, and several quantitative traits (plant size at time of collection, leaf weight, and panicle number in experimental trials). Four hundred and forty-eight plants were studied at the 7 loci and 288 of these individuals were scored for an additional eighth locus. Five genotype classes (monogenic, simplex, and duplex digenic, trigenic, and tetragenic) were distinguished according to their heterozygosity level. Multilocus heterozygosity showed a significant positive correlation with both leaf and panicle yield in experimental conditions, but not with original plant size, which was found to be markedly influenced by environmental microheterogeneity. Multilocus heterozygosity, estimated from both the number of heterozygous loci and the number of distinct alleles per locus, had a significant influence on plant performance. Individual locus effects were positive and significant at two loci (GOT1 and PX1). Panicle number increased regularly with heterozygosity level (from monogenics to tetragenics) at the GOT1 locus, as did leaf weight and panicle number at the PX1 locus. Such variation would be predicted by overdominance at these loci or at linked loci. Significant relationships between leaf yield and heterozygosity level at the GOT1 locus distinguished the homozygotes from the heterozygotes (of any class) and was thus more consistent with inbreeding effects. No significant differences were observed among the five genotype classes for any quantitative trait at the six remaining loci. At both the GOT1 and PX1 loci, heterozygosity had a significant independent effect on leaf weight and panicle number even when the correlation between these traits was removed by analysis of covariance.

Allozyme characterisation of four Mediterranean evergreen oak species

Polymorphism at 11 enzyme loci was used to compare the four evergreen oak species of the Mediterranean Basin and to establish their taxonomical relationships. As a comparison, two evergreen oak species from Tibet, located in the primary distribution area of evergreen oaks, were analysed at the same loci. Cladistic analyses led to the separation of a cluster of four species, namely the three Mediterranean Q. ilex, Q. coccifera and Q. alnifolia and the Tibetan Q. aquifolioides. The other Tibetan species, Q. semicarpifolia, and Q. suber, from the Western Mediterranean Basin, were very distinct genetically from the four other species. These results were not consistent with previous taxonomical treatment, based on morphology alone, which classified Q. ilex and Q. coccifera in "ilex group" (=subgenus Sclerophyllodrys O. Schwartz), and the other four species in "cerris group" (subgenus Cerris, according to Schwartz). Allozyme variation has thus provided useful information to clarify the taxonomy of evergreen oaks.

Flowering characteristics and fertility of interploidy progeny from normal and 2n gametes in Dactylis glomerata L

We performed an extensive experiment to compare male fertility, flowering period, seed set and germination rate between triploid and tetraploid cross-progeny of Dactylis glomerata L. originating from reciprocal diploid (2x) × tetraploid (4x) crosses. Four hundred and forty-five triploid and tetraploid individuals, constituting the cross progeny of 2x-4x reciprocal crosses between diploids from nine distinct subspecies, and tetraploids from five subspecies, were tested for male fertility and phenology under homogeneous conditions. A significant negative correlation was observed between the parental genetic distance and the frequency of sterile panicles (ST) in 3x progeny, suggesting that selection may occur against fertility in triploid progeny from interploid crosses involving closely related subspecies. Genetic distance between the parental subspecies was not significantly correlated with the frequency of fertile panicles (FT), non-spreading panicles (NT, non-dehiscent anthers with pollen inside), and ST panicles in the 4x progeny, nor with the frequency of FT and NT panicles in the 3x progeny. Among the 4x progeny individuals that produced panicles, more than 60% were fully fertile, but 24% possessed only sterile panicles. In the triploids, 63% of the reproductive plants were male-sterile, but 31% had variable proportions of fertile panicles and 6% possessed only fertile panicles. Male fertility in the progeny was significantly affected by individual ploidy level, parental diploid subspecies identity, and by the interaction between these 2 factors in the case of the fertile progeny plants. The frequency of vegetative plants was higher in 4x progeny than in 3x progeny, more particularly when the seed parent was tetraploid. However, when the seed parent was diploid, the frequencies of vegetative plants were similar in 3x and 4x progeny. Pollen viability was significantly higher in anthers from fertile panicles than in anthers from non-spreading panicles; viability of pollen from non-spreading panicles was, in turn, significantly higher than in sterile panicles. Also, seed set was higher in panicles with fertile anthers than in the two other types. However, seeds from plants with fertile, non-spreading and sterile anthers showed no differences in germination rate. Therefore, in Dactylis glomerata, tetraploid cross progeny arising from 2n gametes possess higher fertility, on average, than triploid cross progeny involving normal gametes. In addition to lower male fertility (non-spreading or atrophied anthers), the triploids also showed significantly lower pollen fertility in anthers and lower seed set in open pollinating conditions, probably because disturbance in meiosis, due to odd chromosome associations, also occurs in eggs. Moreover, triploids were observed to reproduce preferentially with individuals possessing the same ploidy levels as their seed parent. These results suggest that the production of 4x individuals, derived from 2n gametes in interploid crosses, is likely to be the main effective means of direct gene transfer in contact areas of populations or subspecies possessing distinct ploidy levels.