Ovules were removed from ovaries from mid-July to early August and cultured in vitro. As a result, all ovules which were sampled and cultured by July 19 in the central female flowers and by July 22 in the lateral female flowers died before early September. However, in ovules sampled and cultured on and after July 22 and 26 in the central and lateral female flowers, respectively, nearly one ovule per ovary survived and developed on the culture medium until mid-September. The development and enlargement of the embryos within the ovules were also observed. Differences in the weight between the largest ovule per ovary and the other ovules began to become clear after July 22 in the central and lateral female flowers. The differences increased rapidly on and after July 29. The longest pollen tubes which penetrated into the ovaries reached the nucelli on July 19-22 and July 22-26 in the central and lateral female flowers, respectively. The pollen tubes which reached the zone where the nucelli were located were observed on July 15 and 19 in the central and lateral female flowers; their numbers increased rapidly during July 15-19 and July 19-22, respectively. Subsequently, their numbers gradually decreased. In the central female flowers, the ovules with a single pollen tube per ovary numbered 5, 6 and 5 on July 22, 26 and 29, respectively. A pollen tube, which penetrated into the micropyle, reached the nucellus in only one ovule in the ovary during July 22-29. The pollen tubes which had penetrated into the micropyles of the other ovules stopped growing at the same length, about half of the length of the entire micropyle, on July 29 in particular. It was suggested that when one ovule had already been fertilized, the other ovules, into which pollen tubes had penetrated, could not be fertilized owing to the cessation of the pollen tube growth.