Martin Borrill's research while affiliated with Aberystwyth University and other places

S ummary Hybridizations were carried out between samples of ten diploid Dactylis subspecies. Both pollen fertility and ‘fertility index’ of hybrids decreased from initial cross to F 2 . There was no correlation of fertility with geographical distribution of the parents, and despite the fall in hybrid fertility it was considered that gene exchange would not be completely prevented.

Chromosome counts on the progeny of crosses between diploid and tetraploid races ofDactylis show that tetraploid hybrids are produced as well as the expected triploids. The relative proportions of 4x and 3x hybrids vary greatly in different crosses, and the data suggest that parental geno-type influences the result. Overall, the frquencies of 3x and 4x hybrids are about equal, with no indication of a difference between the reciprocal 2x×4x and 4x×2x cross-combinations except perhaps in the case of diploids and autotetraploids of the same subspecies. Rare triploid hybrids are found in crosses between diploid subspecies ofDactylis. The mechanisms by which a diploid plant could donate two genomes to its offspring are discussed in relation to theDactylis situation, and the evolutionary significance of 4x hybrids formed in this way is considered.

An autopolyploid series comprising 2x, 3x and 4x was obtained by colchicining the diploid speciesA. ovatum Lag. The autotriploid, used as seed and pollen parent, was back-crossed to diploid and tetraploid. No significant overall difference was observed in fertility, but there was a tendency for particular pair crosses to the diploid as female parent to show higher levels of seed-setting. This may be associated with the relatively large proportion of haploid pollen that functions when the triploid is used as the male parent. The effectiveness, in part, of the freely liberated pollen of autotriploidAnthoxanthum in producing progeny, contrasts with the total failure of manually liberated allotriploidDactylis pollen. Triploid sterility may well relate to degree of hybridity. Genetic unbalance induced by aneuploidy was considered to be more relevant than the “genetic ratio” in seeds, as an explanation of the type of interploid progeny obtained inAnthoxanthum.

Hybridization reveals the close affinity ofAnthoxanthum odoratum and the Mediterranean annualA. ovatum. A. alpinum is genetically somewhat more isolated; occasional sterile hybrids withA. ovatum can be obtained. Morphology reveals an affinity betweenA. odoratum andA. ovatum. There are nodiagnostic characters linkingA. odoratum toA. alpinum. The perennial nature of the generality ofA. odoratum, plus phenotypic flexibility and an ability to behave as an annual in particular habitats, suggest an affinity both toA. alpinum (winter-hardy perennial) andA. ovatum (variable annual or occasional biennial). The artificial allotetraploid based on the sterile hybrid betweenA. alpinum andA. ovatum simulatesA. odoratum, more closely than do either of the diploids. Experimental observations suggest a very reasonable possibility that a similar sequence of events perhaps occurring on several occasions under natural conditions could have given rise to the tetraploid speciesA. odoratum.

The average cross fertility is low and influenced by genotype, rather than by the taxonomic position of the parents. A few genotype combinations were reasonably fertile. On the average, crosses were more successful with the tetraploid as the female parent. When the triploids, which are functionally male-sterile, were allowed to outcross to diploids and tetraploids some seed setting occurred comparable in extent to the initial crosses. Meiosis is irregular with trivalents, bivalents and univalents; the mean multivalent frequency shows a positive correlation with mean chiasma frequency. Pairing seems to be affected mainly by the genetic constitution of the parents and not by interspecific affinities. The low fertility of the triploids is not surprising in view of the apparently random assortment of the chromosomes to the gametes.

The average cross fertility is low and influenced by genotype, rather than by the taxonomic position of the parents. A few genotype combinations were reasonably fertile. On the average, crosses were more successful with the tetraploid as the female parent.

A RECENT intensive survey of Dactylis has shown that tetraploids (2n = 28) are widely distributed throughout the temperate parts of the world1. In Europe and Asia, regions where this genus is native, diploids are also found, and each of these has a comparatively limited range of distribution2.

A LIMITATION of gene exchange by incompatibility between segments of an outbreeding species-population is of primary importance in micro-evolution1, and few cases have hitherto been reported in the Gramineae. Some results of a study of the species Glyceria fluitans and Glyceria plicata may therefore be of interest.