summaryIn a well‐weeded site, selection is likely to favour quick‐maturing variants of annual weeds. Variation in rate of development ‐ from sowing, with synchronous germination, to first flowering and fruiting‐was studied in experiments with Senecio vulgaris L. from the Botanic Garden of Cambridge University, England. Experiments revealed that the number of nodes on the main axis could not be used a ‘genetic marker’ for rate of development.From a large random sample of seed‐parents, progenies were raised from open‐pollinated capitula. Family lines differed significantly in rate of development, with markedly different degrees of variation within families. Twenty family lines, representing the extremes of variation, were chosen for further study. In cultivated selfed derivatives, the majority of the seed‐parents appeared to be behaving as inbred lines for rate of development. Evidence of genetic segregation was detected in only two out of 20 cases. Using the radiate locus as a marker, there was also evidence for a very small amount of outcrossing, with greater values for radiate than non‐radiate seed‐parents. In an experiment in which six inbred lines, derived from the original seed‐parents, were grown out‐of‐doors on fertile and infertile soils and on fertile soil in a glasshouse, there was considerable developmental flexibility in a number of heritable traits including rate of development.Although the methods of weed control in the Botanic Garden would seem to exert a powerful directional selective force favouring only the quickest‐maturing variants of groundsel, a certain amount of variation was nevertheless detected. Many factors are likely to be important in the maintenance of population variation in rate of development in groundsels in the Garden. Several statistically significant lines x soils and lines x places of cultivation were discovered. Thus, the lines did not all behave in the same way in the inside and outside treatments and the rank order of date of maturity of the different lines was not the same in the two soils. This suggests that, in the Garden proper, relative fitness, in the face of weeding pressures, is not fixed but may vary seasonally and in soils of different fertility. This behaviour, in a species which is able to reproduce at all seasons of the year, could be a major factor in the maintenance of population variation. Other contributory factors are of great importance, namely, spatial and seasonal variation in the intensity of weeding, developmental flexibility and the effects of predators, pests and diseases.