A study of polyembryonic tomatoes was conducted with two major objectives. The first was to determine means of obtaining a higher frequency of polyembryonic seeds. The second was to study various growth traits of twin plants, especially to observe differences between twins which might be attributable to a nutritional competition within polyembryonic seeds. The variety Beefsteak, a seed lot of which was previously observed to produce a high incidence of twins, was found in this study to transmit this trait to its progeny. F1 weed of reciprocal crosses with low incidence lines showed that inheritance of a high incidence of twinning was neither dominant nor maternal since in no case was there a high incidence of twinning in any of the progenies. Self-pollinated seed taken from F1 plants of these crosses indicated a more complex inheritance than would be afforded by a single gene. Reciprocal crosses between the Beefsteak line and Premier both produced a higher incidence than crosses between the same high twinning line and Red Jacket. It was found that detection of polyembryonic seeds was better accomplished where seeds had been separated by weight. Heavy seeds were found to contain a higher proportion of twins. Detection of polyembryonic seeds was further improved by soaking heavy seeds in Chlorox, a 5.25 percent solution of sodium hypochlorite, for five minutes. Chlorox clarifies the seed coat tissue and enables the observer to see the outline of the embryo more clearly in single embryo seeds than in polyembryonic seeds. Chlorox treated, heavy seeds in which the embryo could not be clearly discerned through the seed coat produced twins at a rate of 1 in 17 seeds; whereas, non-treated and randomly selected seeds of the same seed lot produced twins at the rate of 1 in 125 seeds. A distinction was made between the each twin pair according to which seedling completely emerged first from the seed coat. It was found that the twin that first completely emerged was later in flowering and slower in growth. Differences between the twins were not significant for height to inflorescence, number of flowers in first and second inflorescences, or number of nodes below first inflorescence.

Covers not scanned.