Proceedings of the Eighth International Grassland Congress
held at the University of Reading, England, 11-21 July 1960. p. 41-44
Selection for Production Characters in Ryegrass
J P Cooper
(Welsh Plant Breeding Station, Aberystwyth, Wales)
Fig. 1. Response to selection for ear emergence in Kent ryegrass.
Continued response to selection in later generations is illustrated in Figure 1 for date of ear emergence in Kent ryegrass. Selection for this character was started in Irish and Kent perennial ryegrass in 1954 and is now in the fifth generation. A similar picture is presented by both varieties; 5 generations of selection have resulted in extensive response well outside the range of the original populations, and as early as the fourth generation, the full range of Kent had been obtained from Irish and vice versa. Even 4 initial plants of the same phenotype have released more variation than is expressed phenotypically in the original variety, confirming the great heterozygosity of the initial parents. Appreciable genetic variation still exists within the selection lines and response is still continuing.
There are, however, dangers in selecting for a single character at a time. Unfavourable correlated responses in other characters may occur. In selection for date of ear emergence, for instance, male and female sterility have appeared as correlated responses in 2 of the lines, decreasing the possible selection differential (7). Similarly, selecting for rapid leaf appearance has invariably led to smaller leaves and vice versa (9); this correlation may, well prove an obstacle to selection for high total leaf area.
Genotype/environment interaction may also act as a brake on progress, as in selection for dry matter production in Italian and Irish perennial ryegrass, where a strong interaction occurs between family production and level of cutting.
The efficiency of selection for any character, however, depends also on the accuracy with which it can be measured. Such characters as date of ear emergence, habit of growth, disease resistance, and winter greenness can be recorded fairly accurately on individual seedlings or spaced plants (1). The case of production in the sward, however, is more difficult. Assessment of production on the basis of spaced plants, either by eye or by cutting, is not always correlated with yield under sward conditions, particularly where a legume is included (10, 16).
Selection techniques must therefore be developed to assess individual plants and their progenies under controlled competition and defoliation, preferably with a legume. A simulated sward technique involving close spacing of seedlings (2.5 in.) combined with different levels of defoliation has been developed at Aberystwyth for this purpose. It has been used for comparing different varieties and families, and since records are taken on individual seedlings, it can be used as a screening technique for selecting parents.