Ten chloroplast microsatellite markers were used to characterise chloroplast genetic diversity at allelic and haplotypic level in 104 accessions of Lolium perenne, other Lolium species, Festuca species and  Festulolium cultivars. Furthermore, genetic relationships between the accessions and biogeographic distribution of haplotypes were investigated using a range of Nei's population genetic diversity measures and analysis of molecular variance (AMOVA). An extremely high number (511) of haplotypes was detected in 1575 individuals. Nei's gene diversity values among L. perenne accessions ranged between 0 and 0.333. Much of the L. perenne European ecotype diversity (61%), as calculated using AMOVA, could be attributed to withinpopulation variance and this is likely caused by, and maintained by, high levels of natural and anthropogenic seed dispersal. Plastid gene pools and maternal lineages for L. perenne could be clearly identified. Evidence was found, using AMOVA, to show a likely migration route of L. perenne from Southern regions of Europe northwards.

Twelve primers to amplify microsatellite markers from the chloroplast genome of Lolium perenne were designed and optimized using de novo sequencing and in silico sequences. With one exception, each locus was polymorphic with a range from two to nine alleles in L. perenne. The newly developed primer pairs cross‐amplified in different species of Lolium and in 50 other grass species representing nine grass subfamilies.

A sufficient database of seed yield components is the prerequisite to predict which component contributes most to the seed yield complex and to dissect the biology of the complex quantitative trait seed yield. Phenotypic variation of eleven vegetative and reproductive traits was characterized for 2481 individuals from fifty Lolium perenne L. (perennial ryegrass) ecotypes and cultivars. Considerable levels of among-and within-population variation were found across several inflorescence characters. Principal component and canonical variate analysis separated ecotypes from cultivars, and cultivars generally had later dates of ear emergence, better spring and summer growth, longer rachis length and more spikelets per spike than ecotypes. Strong positive relationships were seen between inflorescence characters using both correlation and regression analyses. The strong relationship between rachis length and other inflorescence characters suggested that rachis length could be used as a performance predictor for several other characters of the spike. This phenotypic study has helped to determine basic patterns of morphological diversity and correlations between characters. It is discussed in which way further developmental genetic studies can be directed.