The timing of grape picking is a critical point for to production of high quality wines. Berry growth and major solutes are routinely measured on a mix of randomly sampled berries, in order to characterize the advancement of ripening at the vineyard scale, and optimize the date of harvest with respect to the specific conditions of the millesime. Most if not all knowledge on grape berry metabolism emerged from this procedure. Since berries do not ripen simultaneously, averaging induces a statistical bias leading to underestimation of water, sugars and organic acids fluxes in individual berries. Moreover, detachment of fruits not only blocks sugar import whose speed, even underestimated, largely exceeds the measured variable, but also reverses the direction of vacuolar transport. The influence of temperature on the nature of the respiratory substrate and on global carbon balance needs to be revisited because of these two possible artefacts. This preliminary study aims at evaluating the heterogeneity of developing berries and to decipher the relations between growth, sugar storage, and malate breakdown, at the cluster and berry levels, rather than on the average population. The scattering of sugars and malic acid concentrations shows that heterogeneity among berries is maximal at veraison. Berry volume shows considerable variationwithin and between clusters and doubles during the first period of sugar accumulation, while malate disappears. Sugar accumulation then continues at a constant volume (rate?), and its concentration tends to homogenize during this process, while malate content is maintained at a very low value. Finally, phloem unloading completely stops and sugars concentrate as the result of shriveling, malate being totally exhausted at this stage. Preliminary results on respiration show that stopping phloem unloading would increase the O2 demand by 25%.