A number of crosses were made between diploid and tetraploid Lolium perenne plants in order to determine the degree of cross compatibility between the two different ploidy levels. The range of hybridization undertaken involved diploid Γ diploid, and tetraploid Γ tetraploid control crossings, diploid Γ tetraploid crosses, tetraploid Γ diploid crosses and mixed pollinations. The seed set, the germination capacity of resultant hybrid seeds, and the chromosome constitution of derived seedlings was determined. In addition attempts were made to culture 12 day-old hybrid embryos from the diploid Γ tetraploid and reciprocal crosses.
The crossing results obtained indicated no barrier to fertilization between diploids and tetraploids, but the irregular and poor development of the seed accompanied by much reduced germination, indicated unsatisfactory endosperm establishment of large numbers of triploids, as confirmed by the result of culturing embryos dissected from 12-day old "seeds".
In preliminary studies undertaken to determine the extent to which diploid and tetraploid ryegrass varieties intercross when grown in Close proximity for seed production, GRIFFITHS and PEGLER (1966) observed a very low incidence of triploids amongst the progenies of diploid plants exposed to the pollen of tetraploid plants and also amongst the progenies of tetraploid material exposed to the pollen of diploid plants.
in these and subsequent studies it was noted that only a very small proportion of the so-called fully formed seeds derived from diploid Γ tetraploid crosses did actually germinate. When, in 1964, propagules of clones of S.24 perennial ryegrass, well separated from one another, were placed in drills of the tetraploid perennial ryegrass variety Reveille, the proportion of florets which formed caryopses, and of caryopses which germinated was as follows (Table 1).
The data (Table 1) clearly indicate that a considerable propor~.ion of the S.24 florets had been fertilized by 2n pollen and had formed caryopses, but most of these had proved to be defective. As noted in the previous studies, the frequency of triploid seedlings was low, thus suggesting incompatibility between the n and 2n gametes for the production of viabte zygotes.
Of the total progenies classed as tetraploids in the early seedling stage, 80 % proved to be ryegrass Γ tall fescue F1 hybrids. These arose as a result of free crossing with