The incompatibility genotypes of a further 25 plants from the South Wales population of Sinapis arvensis studied by Ford and Kay (1985) have been determined, and an additional 21 S-alleles have been found. This brings the overall number of plants analysed from the population to 35, and the total number of S-alleles found also to 35. It is estimated that there are 52 S-alleles in the population.Dominance interactions between S-alleles have been established in a total of 42 different heterozygotes (pooling data obtained during the present study with that of Ford and Kay (bc. cit.)). Co-dominance of alleles in both pollen and stigma occurred in 18 (42.8 per cent), dominance of one allele in the pollen occurred in 21(50 per cent), dominance of one allele in the stigma occurred in two (4.8 per cent), and dominance of one allele in both pollen and stigma occurred in one (2.4 per cent). There is a minimum of three levels in the pollen dominance hierarchy, and two in the stigma dominance hierarchy; dominance interactions between S-alleles in the stigma may be non-linear. There is some indication that pollen dominant alleles tend to be found in lower frequencies than pollen recessive alleles.Preliminary data show that at least some S-alleles in S. arvensis have a very wide geographical distribution.

The level of inbreeding depression has been measured at four independent stages of the life cycle (parental seed fecundity, seed viability, seedling survival and seedling growth. rate) using wild-collected parental plants from seven natural populations of Allium schoenoprasum. At each stage, progenies produced by self-fertilisation were significantly less fit than progenies produced by cross-fertilisation, and the overall inbreeding depression was severe (0.718). This suggests that a high level of cross-fertilisation usually occurs in natural populations, despite the fact that A. schoenoprasum is a clonal, self-compatible, hermaphrodite plant, with apparently ample opportunities for geitonogamous selfing.Using a white flower colour marker, the minimum outcrossing rate in a natural population has been estimated as 091. However, this is almost certainly an over-estimate due to the unequal viabilities of cross and self zygotes. An adjusted estimate of ta,,, = 080 is derived which takes account of the lower survival rate of self zygotes.

Meiotic analysis of 1017 plants of A ilium schoenoprasum, sampled as adults from 18 natural populations in Britain and continental Europe, revealed that 35 (3.4 per cent) were heterozygous for one or more major structural chromosome mutations. Nineteen different interchanges, 11 different inversions, one deletion, and one supernumerary chromosome segment were found. There was no significant difference between British and continental European populations in the frequency of chromosomal variants.There was considerable variation in meiotic behaviour between different interchanges. On average, quadrivalent formation occurred in 68 per cent meiocytes, but the range (from 13 to 95 per cent) was wide. In five interchange heterozygotes, chiasma formation in the interstitial segments occurred in a high proportion of multivalents. In three interchange heterozygotes, a significant excess of chain quadrivalents were in alternate orientation. In the inversion heterozygotes, the frequency of reverse loop chiasma formation was relatively low, occurring, on average, in 6 per cent meiocytes (with a range from 3 to 13 per cent). The pollen and ovule fertilities of inversion heterozygotes were little affected, but those of many interchange heterozygotes were significantly depressed.The distribution of structural chromosome mutations throughout the genome was not random. A disproportionately high frequency involved the acrocentric, nucleolar-organizer chromosome pair.

A breeding programme designed to establish the number of loci controlling the sporophytic self-incompatibility system in Sinapis arvensis has been completed. The programme involved analyses of diallel crosses within F1 and F2 families derived from crosses between plants from geographically remote populations, and within self families derived from the parental plants. Incompatibility reactions were determined cytologically using fluorescence microscopy. Control by a single locus has been established. However, the expression of individual S-alleles may be modified, as shown by incomplete dominance in a self family, and mutual weakening in an F1 family.