**6. Traits Related to Direct and Indirect Selection for Heat-Tolerant Genotypes in Tomato Plants**

As discussed earlier, many physiological and biochemical traits are related to heat tolerance in tomato plants. However, most of the previous studies were based on a few genotypes and the results were genotype-dependent; therefore, these traits cannot be general predictors in screening or selecting heat-tolerant genotypes in the tomato. There have been a limited number of correlation studies between fruit set and/or yield as well as other traits in HS that use a large number of tomato genotypes having various degrees of heat tolerance. A correlation study is particularly important for indirect selection to identify heat-tolerant genotypes. In this section, we examine promising target traits for direct and indirect selection for heat-tolerant genotypes, based on correlation studies with a large number of genotypes. To our knowledge, no such correlation study was conducted in tomato plants among fruit traits and soluble sugars, glycine betaine, secondary metabolites, and SOD activity. The proline content did not show any significant correlation with traits related to heat tolerance in HS (Table 2). Therefore, this will not be discussed here.

#### *6.1. Traits for Direct Selection for Heat Tolerance*

Reproductive rather than vegetative growth is more vulnerable to HS in many crops including the tomato [69]. Vegetative traits were also not significantly correlated with fruit yield and fruit set in HT among 38 [56] and 13 tomato genotypes [37], respectively, despite fruit set and yield being considered as the main targeted traits for heat-tolerance screening in tomato plants. Therefore, most studies on heat tolerance in tomato plants have been focused on reproductive growth. HS significantly decreased the fruit set and the number of fruits per truss [31,37,56,99–101], which were significantly correlated with reduced fruit yield [56]. However, the effect of HS on flower number per truss is somewhat controversial; there were reports of a reduction [37,102,103], no change [20,28,29,104], and even an increase [56] in flower number under HS. The obvious decrease in fruit set or

number but not in flower number under HT among tomato genotypes indicates that floral development may not be the key physiological trait that confers heat tolerance on tomato plants.

Fruit set and yield have been the main traits used for screening heat tolerance among tomato genotypes [63], although fruit drops after fruit set have also increased under HT [37,102]. Various tomato genotypes have been screened, based on fruit set and yield [24,63,68]. However, the fruit set is calculated from the ratio of the number of fruits divided by that of flowers and can be significantly affected by a reduction [37,102,103] or increase in the number of flowers [56]. Therefore, the fruit set in HS is inevitably affected by the changes in the number of flowers in HS, which can also affect the correlation between fruit set and other traits, resulting in a lower correlation with fruit yield than fruit number [56]. Therefore, breeders working with tomato genotypes who have increased the flower number per truss under HS should consider fruit number per truss instead of fruit set for the trait to achieve direct selection for heat tolerance.

Different fruit traits must be considered, depending on cultivars with different fruit sizes, in the screening of heat-tolerant genotypes since some traits associated with heat tolerance differ according to tomato fruit size [56]. For example, both the fruit number per truss and fruit set were significantly and positively correlated with fruit yield in cherry tomato genotypes (>50 g) but only the fruit number per truss was significantly correlated with fruit yield in large fruit types (<100 g) [56]. This is because an increase or decrease in flower number per truss does not significantly affect the fruit set in cherry tomato genotypes, which have much more flowers per truss than large fruit genotypes; however, the increase or decrease by one or two flowers in large fruit genotypes significantly affects the fruit set. In addition, in cherry tomato types, heat-tolerant genotypes can be pre-selected, simply by looking at the previous fruit yield data collected in optimum temperature conditions, because a significantly positive correlation was observed between fruit yields in CK and MCHS conditions that was not observed among large fruit genotypes [56].
