Discussion
A significant F-value of genotypes for seed set suggested that at least two genotypes differed significantly in the rate of seed set (Table 2). Mean comparison showed that W5 had a lower rate of seed set than the others (Table 3). Seed set in wheat x maize crosses is highly affected by 2,4-D. We observed that florets which received the hormone treatment produced seeds, but in those which did not, ovaries remained shrunken and no seed was produced (data not shown). The lower seed set in W5 crosses might have been caused by factors affecting the level of hormone reception. Thinner spikes and more closed glumes of W5 than the others appear to affect the reception of hormone in the spraying method employed. We suggest that seed set is not directly affected by genotypes of wheat parents.
In case of embryo formation irate, the highly significant F-value of genotypes suggested that the variation in response of different wheat parents resulted from different genotypes (Table 2). This was clearly due to the superiority of W1 and W3 over the other three genotypes (Table 3). Since pollen tube studies also showed a significant difference between the length of maize pollen tube growing inside the stigma of W1 and W5 (Table 4), it can be inferred that there may be some unknown crossability system different from that under Kr genes in hexaploid wheat. This system appears to be not. strong enough to completely inhibit growth of maize pollen tube (like bulbosum system) but it reduces the speed of pollen tube growth and thus affects the percentage of embryo formation. In this study W1 showed a mean performance of 12.8% for embryo formation while W5 mean was 5.3% (Table 3). This was consistent with the higher mean of pollen tube growth in W1 crosses than that of W5. Genotypic variation of hexaploid wheats for embryo production in wheat x maize crosses has already been reported by others (Laurie and Bennett 1987; Suenaga et al. 1991; Oury et al. 1993; Giura 1994; Lefebvre and Devaux 1996; Suenaga et al. 1997; Sadasivahiah et al. 1999).
In case of embryo germination rate there was no significant difference among the wheat genotypes (Table 3). The mean of W1, W3 and W5 for this trait was 63.5%,61.1% and 48.3%, respectively. The mean of W5 is appeared to be lower than the other two genotypes though the difference was not statistically significant. The morphology of W5 spikes might have affected the level of hormone reception, and this in turn might have reduced the seed set and even the development of embryos. If hormone is not trapped by wheat spikes sufficiently, the haploid embryos may not develop well, thus causing lower germination rate. The clear effect of hormone concentration on development and size of induced embryos in wheat x maize crosses has been documented by others (Sun et al. 1995; Zhang et al. 1996). Some reports (Bitsch et al. 1998; Suenaga and Nakajima 1989) referred to the effect of embryo size on its germination, ie failure of small embryos in germination. In case of W5 lower seed set and less percentage of embryo germination might also have been due to its smaller embryo size.
Plantlet formation was also independent from genotypic effect of wheat parents. It appears that once an embryo germinates some factors might affect its development to a plantlet, eg conditions of nutritive medium and incubation temperature and light.
Acknowledgments
We are thankful to Agricultural Research, Education and Extension Organization of Iran for sponsoring of the first author for this study as a part of PhD. program.