Results and Discussion

The total number of kernels and the number of filled kernels produced by each cross as well as their germinability are summarised in Table 2.

The various crosses between the Ae. squarrosa varieties and S. cereale cultivars produced 722 hybrid kernels (Table 2). In terms of florets pollinated this represents a kernal set of 55.28%, which agrees well with the findings of MELNYK and UNRAU (1959). These hybrid kernels were au inviable. The Ae. squarrosa kernels resulting from selfing were germinated soon after harvesting, but in spite of this a 77.38% germination was obtained. It is evident that the D-genome on its own does not suppress the crossability with rye very much, but its interaction with the R-genome results in the poor development of the endosperm and embryo which leads to kernel inviability. MELNYK and UNRAU (op. cit.) was able to grow one embryo in artificial culture into a hybrid plant.

The various tetraploid wheat x rye crosses produced 1209 hybrid kernels of which only 5, or 0.41% germinated. In comparison their synthesised hexaploid derivatives (tetraploids to which the D-genome was added by hybridisation and amphiploidy) when crossed with rye produced 1237 hybrid kernels of which 454, or 36.7% germinated (Table 2). This difference is highly significant (at 1% level). It can thus be concluded that the addition of the D-genome to the A- and B-genomes of the tetraploid wheats produces a genetic interaction or cumulative genetic effect, similar to the addition of the A-genome reported by KROWLOW (1970, 1973), which renders the endosperm/embryo developmental barrier less effective in crosses with diploid rye.

Whatever the underlying mechanism may be, it is evident that the above findings may be of practical value to the triticale breeder and supplement those of KROWLOW (op. cit.). In those cases where the 4x wheat x rye crosses fail to produce hybrids, even with the aid of the embryo culture techniques, the tetraploid wheats can be supplemented with A- or D- genomes to produce hexaploids that will give viable kernels when crossed with rye. However, the D-genome appears to be the more useful of the two genomes for this purpose, because 21.49% of the florets of the synthesised hexaploids which contain a pair of D-genomes were able to set hybrid kemels (Table 2), whereas only 3.96% of the florets of the auto-allohexaploids with an extra pair of A-genomes were able to do so (KROWLOW, op. cit.). The germinability of the hybrid kernels set by the former and latter types of hexaploids in crosses with rye was 36.7% (Table 2) and 60.99% (KROWLOW, op. cit.) respectively; consequently 7.89% of the florets of the former hexaploids (AABBDD) produced germinable hybrid kernels (Table 2) compared to only 2.41% of the florets of the AAAABB autoallohexaploids.