The analysis of time series of wood cell anatomical features (such as the earlywood vessels of ring-porous trees) is a successful approach to understand the effect of environmental factors on tree growth and thus constitutes a valuable source of information about past environmental conditions. However, despite the rising interest in analyzing wood anatomical time series, little or no attention has been paid to establish an adequate sample of cells in order to minimize the risk of missing a valuable environmental signal. In order to contribute to such methodological bases, this paper is aimed at (1) identifying a representative sample of earlywood vessels within a tree, which encode the same climatic information, and (2) assessing if it is preferable to obtain the sample of vessels along one or two radii. Four individuals of sessile oak (Quercus petraea (Mattuschka) Liebl.) and sweet chestnut (Castanea sativa Mill.) were harvested and all their earlywood vessel lumina were measured along two 40-mm wide radial strips. Measured vessels were selected stepwise while increasing the tangential width of the wood section from 1 to 40 mm, analyzing at each step (1) the common signal of chronologies and (2) the correlation to the main climatic variables controlling growth. Additionally, both radii in each tree were analyzed together and separately. The results showed that a total tangential width of 10 mm was enough to stabilize the climatic signal with improvement when distributed along two different radii, but a slightly larger tangential width was required to reach an optimal common signal. We suggest that, at least for the case of these two species growing at this specific climatic context, two 5-mm increment cores ensure a representative vessel selection.