Sowing winter oil-seed rape cv. Victor late in autumn (late September or October) in seven seasons from 1970 to 1977 gave enormously variable seed yields, from 120 to 450 g/m 2 . All crops made little growth before winter, and yield was related to the size of the crop at flowering, a function of the length of time for radiation interception and growth between the 'beginning of spring', when mean temperatures rose consistently above 5 °C, and full flower in late May. A late spring in 1970 gave the poorest growth and lowest yield, whereas in 1977 an early spring coincided with late flowering to give exceptional growth, and yields higher than from any early sowing.Crops sown in early autumn (before mid-September) produced more consistent seed yields, 280-360 g/m 2 , except in the dry year of 1976. All grew well in autumn, overwintered with a large leaf area, and once temperatures rose in spring, rapidly reached peak area and full flower in early May. They were all large at flowering, and yield was apparently limited more by post-flowering events.With all sowings numbers of pods and seeds were largely determined during a 3-week phase in late May and early June, extending from full flower until most pod hulls had finished growing. Late sowings produced 3000-6000 pods/m 2 , and the number of seeds retained per pod varied widely, from 7 on a poorly grown crop to 22 on a well grown crop, thus expressing the yield potential determined by crop size at flowering. Early sowings, however, produced apparently excessive numbers of pods (6000-12 000/m 2 ) and few seeds per pod (6-10), so that yield varied little, regardless of crop size. Early in the phase, when the number of seeds was determined, the mass of yellow flowers at the top of the crop reflected or absorbed up to 60 % of incoming radiation, and then the large number of pods increasingly shaded each other and competed for assimilate, resulting in heavy seed losses. A high-yielding crop type may therefore need to incorporate the restricted pod production and good seed retention of some well-grown latesown crops with the reliability and desirable agronomic features of early-sown crops.Final seed weight varied more between seasons (3-7-5-3 nag) than between sowings. Seed growth mainly took place after the number of seeds had been determined, the duration depending on temperature, but rate of growth apparently more on assimilate supply, a function of environmental factors and the number of competing seeds.

The aim of the experiments reported here was to investigate the possibility of minimizing yield losses due to late sowing by increasing leaf area and radiation interception during the pre-flowering period. The use of large seed did increase leaf area, radiation interception and crop weight at flowering. Normally this would be expected to lead to increased yield of late sowings, but hot, dry conditions severely affected all crops during seed growth. Applying fertilizer nitrogen in the seed bed boosted pre-flowering growth of early but not late sowings, probably because low temperatures prevented a response in the latter. Increasing plant population density did improve leaf cover and radiation interception until just before flowering, but leaf canopies in late-sown, low density crops were then able to expand more and function for longer owing to less shading by flowers and pods. Seed retention was improved and, even with as few as 8 plants/m 2 in 1973-4, a worthwhile yield was still obtained. INTRODUCTIONu 8 e l a r g e s e e d ' ^ t o a p p l y n i t r o 8 e n i n t h e s e e d b e d and thus to promote leaf expansion; and (3) to The preceding paper (Mendham, Shipway & increase the seed rate and hence number of plants, Scott, 1981) examined the response of oil-seed rape to give quicker ground cover, (cv. Victor) to delayed sowing in autumn, in seven This paper examines crop responses to these seasons from 1970 to 1977. Using standard cultural treatments in experiments conducted in three of the treatments it was shown that the wide variation in seasons covered by the preceding paper. Each yield of late sowings (after mid-September) was experiment included both early and late sowings, largely due to differences in the amount of radiation The results reported in the previous paper would intercepted and growth made by crops in the pre-not lead us to expect that early sowings would flowering period. Techniques for improving the respond greatly to such treatments. AH early sowyield of late-sown crops should therefore be ings intercept most incoming radiation during the directed mainly at increasing the leaf area, and spring period and are large by flowering time. The ability to intercept radiation, over the period up to main restriction on yield seems to occur in the postflowering, particularly in the critical spring period, flowering period when heavy pod and seed losses At this time leaf area indices are often very small, occur, apparently because the profuse flower and and yet the crop is progressing rapidly towards pod canopy creates deep shade and strong comflowering. An increase in leaf area index could be petition for assimilate. A study of the way lateachieved by increasing the speed of establishment, sown crops respond to agronomic treatments should the number of plants/m 2 , their rate of leaf pro-add to our understanding of the physiology of the duction or the size of individual leaves. Possible crop as well as indicating whether such treatments approaches to achieve these objectives are: (1) to are likely to...