during the first 3weeks of the experiment conducted in1999 showed significant differences in this parameter for seedlings resulting from the different quartiles of seeds. The average values of RGR for the first to fourth quartiles were 0.53, 0.53, 0.72 and 0.73 mm mm1 per week, respectively (P ¼ 0:008). After this period, the RGR of each quartile slowed and there was no difference in this parameter among the quartiles. Data from the trial held in 2000 showed the same trend and significant differences in the RGR of each cultivar were also found (data not presented). In general, in this study for all cultivars in each of the two trials there was a correlation between quartile (and therefore seed weight) and final plant height (see Fig. 2). Therefore, by the end of the experiment, the heaviest seeds had produced the tallest seedlings and the lightest seeds the shortest.The exception tothis generalisationwas the cv.Golden Queen in the 1999 trial where seeds of the third quartile did not grow well. Therefore, seed weight is a good predictor of seedling vigour.The seresults are similar to those of