In the course of a study of the transmission of tomato canker (Phytomonas michiganensis (E.F.S.) Bergey et. al) Prof. P. A. Ark of the Division of Plant Pathology, University of California at Berkeley made reciprocal grafts between Lycopersicon pimpinellifollium (Jusl.) Mill. and Cyphomandra betscea Sendt. In one case, in which L. pimpinellifolium was used as stock, an adventitious shoot of the stock produced fruit about twice as large as ungrafted sister plants. This increased fruit size was retained through several seed generations.
Three possibilities suggest themselves as explanations for the origin of the large fruited line: (1) the Cyphomandra scion caused a mutation in the tomato stock which was expressed somatically and was transmissable; (2) such a mutation occured in the tomato stock independantly of the graft; or (3) the original seed of the tomato culture contained both large and small fruited genotypes.
Two kinds of experiments were started to determine which possibly is most likely. First, an attempt was made to induce an increase in fruit size by grafting between tomato and Cyphomandra. A large number of reciprocal grafts using both pimpinellifolium and an esculentum variety with Cyphomandra were made. To data there is no evidence that Cyphomandra has any influence on fruit size of the grafted tomatoes or on their progeny.
Secondly, an investigation of the genetic differences between the large and small pimpinellifolium lines was begun. Reciprocal crosses were made and the F1 which had the large fruited line as female parent was backcrossed to both parents and was also grown to give F2 seed. In 1950 a field planting was made which included both parents, both F1's, F2 and the two backcrosses.
A preliminary study of the 1950 field data shows that the parental lines differ by many genes, each with a small effect, as there is no evidence of segregation into discontinuous classes even in the backcross generations. The mean fruit weight of F1 is somewhat less than the midparental value and the other generations have means below that expected on the basis of strictly additive gene action. In other words, potence is in the direction of small fruit size which indicates that at least many of the genes for large fruit size are recessive. It is therefore extremely unlikely that such a large number of recessive mutations could have arisen in the adventitious shoot either as a result of the graft or spontaneously. It is even more unlikely that the original shoot from the stock could have been homozygous for such mutated loci.
Since the history of the original pimpinellifolium seed lot is obscure, it could and very probably did contain some seed of a large fruited pimpinellifolium line from Peru. Thus, all the evidence is consistant with the assumption that the seedling used in the original graft was genotypically large fruited.