An increasing number of studies dealing with environmental and dietary reconstruction involve measurement of intra‐tooth variation of isotope ratios. The sampling procedure typically consists of collecting a sequence of horizontal bands perpendicular to the growth axis of the tooth. The objective is to obtain a temporal sequence of the changes recorded in dental tissues during tooth development. This paper examines some issues associated with this sampling strategy and consequences for interpretation of the data. Time resolution is influenced by the pattern and duration of enamel mineralization, in a way that might, however, depend on the species and the tooth analysed. Although it may be impossible to sample discrete amounts of time, a chronological order seems to be respected, which should be well enough when the objective is to detect changes in isotope ratios with time. Absolute data must be interpreted with caution. Such issues do not affect studies of inter‐individual variability as long as the sampling procedure applied to the compared specimens is consistent. A new sampling strategy is tested on modern goat teeth, involving drilling in an oblique direction. The results obtained from obliquely drilled samples are very similar to those obtained from the horizontal sampling procedure. More work is still needed to determine the value of alternative sampling strategies. Copyright © 2003 John Wiley & Sons, Ltd.