In this study, up to 21 morphological traits including those describing growth habit, stolons, shoots and inflorescences were measured on all of the available recreational Cynodon spp. turf varieties in Australia including 19 hybrid interspecific bermudagrass (Cynodon dactylon x C. transvaalensis) varieties and 16 bermudagrass (C. dactylon) varieties at the former Queensland Department of Agriculture, Fisheries and Forestry (DAFF) Redlands Research Facility, Queensland, Australia (27º32’S lat, 153º15’E long, 25 masl) between 2002 and 2004. During this period 5 spaced plant experiments and a single sward experiment, were established to measure a range of morphologicalagronomic characteristics.
Up to 14,248 morphological-agronomic data points were collected from the 6 experiments monitoring stolon and shoot, inflorescence and growth habit characteristics. These data were analysed by first conducting a spatial analysis on each experiment adjusting for any environmental trends and to obtain estimates of residual variances and first order auto-regression coefficients for row and column to be input into a one-stage spatial analysis for each trait. Because varieties were treated as ‘random’ these analyses produced a set of variety averages or Best Linear Unbiased Predictors (BLUPs) for each trait. The BLUPs were used to construct a two way table of varieties x traits and used for a pattern analysis i.e. clustering and ordination.
Four ‘variety groups’ were identified, following a pattern analysis of the comprehensive data set. Each of the groups included members of both tetraploid C. dactylon (2n = 36) and triploid Cynodon hybrids (2n = 27) highlighting the large morphological variation that exists within the Cynodon taxa used in this study.
The morphological-agronomic traits studied were also grouped following pattern analysis to provide 6 ‘trait groups’. Four of the 6 trait groups contained individual traits, while 2 groups contained multiple characteristics. A set of informative traits were compiled consisting of a single characteristic from each of the 6 trait groups that would provide a distinct, uniform and stable morphological measure. Trait grouping could also provide beneficial information for the selection of varieties of common knowledge (VCK) needed for comparative testing as required for trials to obtain, for example, Plant Breeders Rights in Australia or Plant Variety Protection in the USA. Comparator varieties could be chosen from within a group containing parent/source material or a variety earlier identified following observations of morphological and developmental growth. Using this method to identify VCK would remove subjective and often bias judgement.