Williams, Watkin.
The validity of many of the widely accepted publications dealing with heterosis in the tomato in relation to its utilisation in breeding may be questioned because of the nature of the material chosen for the crosses. In several instances species crosses involving L. pimpinellifolium have been used, while in others, lines derived from pimpinellifolium x esculentum hybrids have been included as parental material. Objection to the use of species crosses can be raised from two standpoints. Firstly, from the practical standpoint, crosses involving small fruited, wild species can under no circumstance yield hybrids of commercial importance because of the partial phenotypic dominance of small fruit. The total yield of such crosses is of no consequence if the fruit is too small for economical harvesting and distribution. Secondly, from the genetical standpoint, the behaviour of hybrids involving species is unreliable experience on which to predict the behaviour of intra-specific crosses. One can expect various kinds of genetical interaction in crosses between diversified species, which never, or at most quite rarely, exist within a species.
With these thoughts in mind an investigation was made of the expression of quantitative characters in three hybrids of the cultivated tomato. Two of the three hybrids chosen, Ware Gross (Potentate x E.S.1.) and Hertford Cross (Potentate x Leafmould Resister), have been found to be highly acceptable varieties under glass in England, while the third hybrid, No. 5942 (Vetomold x Harbinger), is one of a new series of hybrids undergoing testing at the present time. All three are therefore legitimate material on which to draw conclusions regarding the expression of hybrid vigour in relation to tomato breeding.
The following five character expressions were studied in all three hybrids: flower number, date of first flower, average fruit weight, fruit number and yield per plant. In hybrid No. 5942 the period from first flower to first ripe fruit was also recorded. The experiments were in randomised blocks with thirty replicates.
None of the hybrids exceeded the parental limits of expression in respect of any one of the characters with the exception of yield per plant. Yield per plant in all three hybrids was greater than in the heaviest yielding parent. Yield per plant, unlike the other expressions studied here, is a compound expression. It involves fruit number and average fruit weight and the apparent heterosis for yield is simply due to a favourable combination of these component characters from the parents. In two of the hybrids fruit number equalled the rud-parent and was combined with an average fruit weight in excess of the mid-parent. In one of the parents of both these hybrids high fruit number was combined with low fruit weight and vice versa in the other. Under these circumstances heterosis for fruit yield is inevitable. In the third hybrid phenotypic dominance of high fruit number from one parent was combined with an increased fruit weight derived from the opposite parent again inevitably leading to heterosis for yield per plant.
Schematically the system prevailing in these hybrids can be represented as follows:-
________________________________________________________________
Unit Character (A) Unit Character (B) Compound
Line (Fruit number) (Fruit weight ) Character
(Yield)
A x B
________________________________________________________________
Parent 1 High Expression Low Expression c (c > d)
________________________________________________________________
Parent 2 Low Expression High Expression d (d < c)
________________________________________________________________
Hybrid Mid-parent to Mid-parent to c + x
dominance of high dominance of high (= Heterosis)
expression expression
________________________________________________________________
The important characteristics of the relationship between compound characters
and their simple unit components in these hybrids are: (1) the levels of
expression of the component characters from opposite parents alternate, (2)
there is no heterosis in the unit characters, and (3) the heterosis expressed
by the compound character is the rest of multiplicative action on the
phenotypic level. It is suggested that multiplicative gene interaction which
has been frequently invoked to explain such results is therefore an illusion.Heterosis for yield in F1 tomato hybrids appears from these data to depend not on the genetic system but on the way the parental characters are introduced into the cross.