Viticultural practices such as trunk girdling and shoot topping along with defoliation, shading and node number per vine treatments were used to alter the carbohydrate physiology of mature Chardonnay grapevines growing in the cool climate of Canterbury, New Zealand. The timing of vine defoliation in the season previous to fruiting decreased concentrations of over-wintering carbohydrate reserves (mostly starch) in both the trunks and roots of grapevines. Roots were particularly sensitive, with defoliation as early as 4 weeks after bloom in the previous season reducing starch concentrations to 1.5%Dwt at bud burst compared with 17%Dwt in non-defoliated vines. In contrast, partial vine defoliation as early as bloom in the previous season reduced root starch concentrations to 4-7%Dwt at bud burst compared with 15%Dwt in non-defoliated vines. Vine shading and trunk girdling treatments at bloom in the previous season, resulted in small reductions in root starch concentrations (16%Dwt) compared with non-shaded and non-girdled vines (19%Dwt), but shoot topping did not. Study across three growing seasons established that higher concentrations of over-wintering trunk and root carbohydrate reserves were associated with warmer and sunnier weather in the previous growing season. Individual shoot leaf removal at either the beginning or towards the end of the inflorescence initiation period, reduced shoot starch concentrations to 3-6%Dwt compared with 11 %Dwt for no leaf removal, such reductions persisted through to the following season. Shoot topping at the start of the initiation period had no effect on shoot carbohydrate accumulation, but trunk girdling temporarily increased shoot starch concentrations during the first 31 days after treatment. Reductions in over-wintering trunk and root carbohydrate reserves were associated with a reduction in inflorescences per shoot and flowers per inflorescence in the following season, the reduction as much as 50% compared with non carbohydrate stressed vines. While there were strong linear or curvilinear relationships between the concentration of starch in trunks and roots at bud burst and inflorescences per shoot and flowers per inflorescence, in case the of inflorescences per shoot, there was not an immediate cause and effect because inflorescences were initiated in the previous season. Individual shoot leaf removal during the inflorescence initiation period illustrated that leaf removal directly inhibited the initiation of inflorescences in latent buds. Shoot carbohydrate measurements showed a strong curvilinear relationship to the number of inflorescences per shoot, with a threshold starch concentration of 10-12%Dwt during the inflorescence initiation period required for a maximum number of inflorescences per shoot. Furthermore, examination of individual node positions emphasised the importance of the subtending leaf on the initiation of inflorescences within the latent bud. The number of inflorescences per shoot post bud burst was reduced on vines that were both carbohydrate reserve stressed (by previous season's defoliation) and had a high node (108) number retained per vine after winter pruning compared with little or no reduction in inflorescences per shoot on carbohydrate reserve stressed vines that had a low (20) node number per vine. The reduction in inflorescences per shoot on high node vines was associated with reduced carbohydrate reserves and reduced shoot vigour (thinner and lighter shoots). Flowers per inflorescence were reduced by as much 50% in response to lower overwintering carbohydrate reserves. Fewer flowers per inflorescence were attributed to a reduction in primary branching of the inflorescence and also a reduction in flowers per branch. Strong linear relationships between the concentrations of starch in trunks and roots and flowers per inflorescence indicate that the determination of flowers per inflorescence, unlike inflorescences per shoot, may be dependent on the level of overwintering carbohydrate reserves. This is most likely due to changes in branching of the inflorescence and individual flower formation occurring during the bud burst period. Per cent fruitset was not affected by reductions in carbohydrate reserves, so fewer inflorescences per shoot and flowers per inflorescence resulted in reduced vine yield. The findings of this thesis indicate that changes in the level of carbohydrate production and partitioning in response to a range of viticultural management practices and seasonal weather contribute to seasonal variation in grapevine flowering and yields in New Zealand's cool climate environment. The relationships between carbohydrate reserves and flowering illustrate the potential to use this information to predict grapevine flowering and forecast yields. The practical implications of this research illustrate that the viticulturist must manage grapevines not only for the current crop, but also for subsequent crops by maintaining sufficient carbohydrate reserves for balanced growth flowering and fruiting from season to season.