Inbred lines and hybrids of English tick and horse beans and exotic stocks were studied with respect to pod set, floral structure and pollen production and germination to find out why most inbred lines set few or no seeds without tripping whereas most hybrids and some inbred lines set equally well with or without tripping. The most important outcome was new information on variation between genotypes in flower structure and pollen quantity and on how these two characters affect the mechanics of self-pollination. Though heterosis in terms of abundant pollen production could be a factor in the self-pollinating ability of hybrids, some of the inbred lines were capable of self pollination despite limited pollen production, because the structure of the flower permitted the pollen to reach the stigma via a short route. The floral features associated with the ability to self-pollinate in the inbred lines included a relatively short style bending at more or less a right angle, few and short stylar hairs and stigmatic papillae, and less pronounced ridges on the inside of the keel petals. The implications of these findings to the development of self-pollin ated varieties of field beans are briefly discussed. barrier to pollination, which is absent in the flowers of self-pollinating plants. The present investigation Field beans are intermediate in breeding be-was undertaken with the objective of throwing haviour between completely cross-pollinated and some light on the nature of this barrier and how completely self-pollinated crops (Fyfe & Bailey, hybrids and self-pollinating lines overcome it. Such 1951) and their seed set is markedly reduced in the knowledge is of interest both intrinsically and as an absence of pollinating bees (Riedel & Wort, 1960; aid in the development of self-pollinated varieties Scriven, Cooper & Allen, 1961;Free, 1966). Devel-of field beans, opment of completely self-pollinated varieties is being sought (Lawes, 1973) to eliminate dependence MATERIALS AND METHODS on bees for pollination, thereby avoiding fluctuations in yield caused by variations in the availaThe material studied (Table 1) comprised inbred bility and activity of bees. lines selfed for four or more generations and F t The ability of field-bean plants to set seeds with-hybrids, derived from stocks of English spring and out flower manipulation (tripping) is known as winter beans and exotic material from Sudan and autofertility and this varies with the genotype and India. The plants were grown, during the spring of level of inbreeding, hybrids being usually more 1975, singly in 18-cm pots filled with John Innes autofertile than inbred plants. Although in most No. 2 compost and placed on a capillary bench in a cases highly inbred plants set practically no seeds bee-proofed glasshouse, without flower manipulation, the existence of selfpollinating lines has been reported (Hayes & Pod set with and without tripping Hanna, 1968;Kambal, 1969; Toynbee-Clarke, Flowers were tripped by depressing the wing 1974).petals thus releasing the style and s...

Previous papers in this series described (Bond, Drayner, Fyfe & Toynbee-Clarke, 1964a) a type of male sterility in beans, segregating as a simple Mendelian recessive and showed (19646) that very good yields could be obtained from hybrids produced with its aid. With that type of male sterility, hybrid seed could only be grown on an experimental scale, because it had not been possible to produce progenies with more than 50 % male-sterile plants.The present paper describes another form of male sterility, apparently involving cytoplasmic inheritance and showing more promise for commercial exploitation.Discovery.-The male sterility was first discovered at King's College, Newcastle upon Tyne when, by the use of hilum colour markers it was shown that one plant underwent a much higher degree of natural crossing, under open pollination, than a group of fifty plants drawn from the same population (Bond, 1957). Subsequent examination of the progeny of this plant showed that the anthers were shrunken and released very little pollen.Description.-Pollen grains from male-sterile plants did not stain as deeply in acetocarmine as pollen grains from normal plants and none germinated in sugar solution. Seed setting following self-pollination of such plants was rare, but cross-pollinations produced normal amounts of pods and seeds. Glasshouse environments appeared to favour the development of pollen but even under these conditions selffertilization was difficult to obtain.No effect of the environment on the proportion of fertile plants in progenies of male-steriles could be detected. When progenies were divided into autumnsown and spring-sown portions, the date of sowing had no significant effect on the proportion of fertile plants. Similarly, plant density, position in the field and date of observation had no detectable effect on the degree of sterility of progenies.A small proportion of male-sterile plants had one or more completely fertile tillers. Such plants are described in this paper as tiller-sterile or partlyfertile, the term semi-sterile being reserved for plants with all flowers producing small quantities of viable pollen.

1. A method of producing hybrid beans with the aid of a cytoplasmic male-sterile is described.2. Most of the requirements of large scale production of hybrid seed can be met: use of genetic markers showed that the male sterility allowed 100% cross fertilization; there was no evidence of a deficiency of insect pollinators near Cambridge; and fertility restoring lines are available.3. The importance of being able to maintain a high level of sterility in the female parent line is emphasized. Methods of improving the proportion of sterile plants in the sterile phase of an inbred are discussed. The tolerable level of fertility in progenies of sterile parents depends on the level of sterility necessary in the female parent of the final cross. Five inbreds are being handled with less than 2% of fertile plants in their sterile phase.

The yields of seed from a hybrid, one of its parents and different mixtures of the hybrid and parent were compared in the case of two different hybrids. Open-pollinated controls were included in the trial, which was sown at about the normal rate for beans. With both hybrids a mixture of 11 of hybrid to 1 of parent yielded as well as the hybrid itself and mixtures of 2 of hybrid to 1 of parent yielded in both cases about 7 cwt./acro more than the mean of the open-pollinated controls. With the higher-yielding hybrid, which had less contrast in vegetative vigour between hybrid and parent, mixtures in the proportions 5:1 and 2:1 showed declining yields, but with the other hybrid they yielded as well as the pure hybrid.

When both spring and winter varieties of field beans were entered in the same trial, sown on the same date (in the spring) and harvested on the same date, the seeds of spring varieties had mean crude protein contents which were higher than those of winter varieties by 3-1 ± 0-9 % in the first trial, 2-8 + 0-7 % in the second trial, and 2-9 ± 0-3 % in the third trial.