Under terminal water deficit, the impact of stem carbohydrate remobilization has greater significance because post-anthesis assimilation is limited, and grain growth depends on translocation of carbohydrate reserves. The working hypothesis of this thesis is that increases in stem carbohydrates facilitate tolerance to terminal drought in wheat. The goals of this thesis are to examine this hypothesis using physiological and genetic tools; identify genes that are related to QTL for stem carbohydrate; work with wheat and barley breeders to integrate findings into the breeding program of the Department of Agricultural and Food Western Australia. The physiological data of three drought experiments (two years in a glasshouse and one year in the field) suggested the maximum level of stem water soluble carbohydrate (WSC) is not consistently related to grain weight, especially, under water deficit. The patterns of WSC accumulation after anthesis differed depending on variety and suggested that WSC degradation and translocation have different genetic determinants. Most of the carbohydrates in stem WSC in wheat are fructans. Because 1-FEH gene was an important gene in fructan degradation, the three copies of this gene (1-FEH w1, 1-FEH w2 and 1-FEH w3) were isolated from the respective genomes of bread wheat. In addition, the genes were mapped to chromosome locations and coincided with QTL for grain weight. The results of gene expression studies show that 1-FEH w3 had significantly higher levels in the stem and sheath which negatively corresponded to the level of stem WSC in two wheat varieties in both water-deficit and well-watered treatments. Strikingly, the 1-FEH w3 appeared to be activated by water deficit in Westonia but not in Kauz. The results suggest that stem WSC level is not, on its own, a reliable criterion to identify potential grain yield in wheat exposed to water deficit during grain filling. The expression of 1-FEH w3 may provide a better indicator when linked to instantaneous water use efficiency, osmotic potential and green leaf retention, and this requires validation in field grown plants. In view of the location of the contribution to grain filling of stem WSC, this is a potential candidate gene contributing to grain filling. The numerous differences of intron sequences of 1-FEH genes would provide more opportunities to find markers associated with the QTL. A new FEH gene was partially isolated from Chinese Spring and the sequence was closely related to 1-FEH genes. This gene, FEH w4, was mapped to 6AS using Chinese Spring deletion bin lines. The polymorphism of this gene was found between different bread varieties using PCRs and RFLPs, and this allowed the gene to be mapped to two populations of Hanxuan 10 × Lumai 14 and Cranbrook × Halberd. In the population of Hanxuan 10 × Lumai 14, it was close to SSR marker xgwm334 and wmc297 where the QTL of thousand grain weight and grain filling efficiency were located. This result indicated this gene might be another possible candidate gene for grain weight and grain filling in wheat.