Bottled and packaged waters are an increasingly significant component of the human diet. These products are regulated at the regional, national, and international levels, and determining the authenticity of marketing and labeling claims represents a challenge to regulatory agencies. Here, we present a dataset of stable isotope ratios for bottled waters sampled worldwide, and consider potential applications of such data for regulatory, forensic and geochemical standardization applications. The hydrogen and oxygen isotope ratios of 234 samples of bottled water range from −147‰ to +15‰ and from −19.1‰ to +3.0‰, respectively. These values fall within and span most of the normal range for meteoric waters, indicating that these commercially available products represent a source of waters for use as laboratory working standards in applications requiring standardization over a large range of isotope ratios. The measured values of bottled water samples cluster along the global meteoric water line, suggesting that bottled water isotope ratios preserve information about the water sources from which they were derived. Using the dataset, we demonstrate how bottled water isotope ratios provide evidence for substantial evaporative enrichment of water sources prior to bottling and for the marketing of waters derived from mountain and lowland sources under the same name. Comparison of bottled water isotope ratios with natural environmental water isotope ratios demonstrates that on average the isotopic composition of bottled water tends to be similar to the composition of naturally available local water sources, suggesting that in many cases bottled water need not be considered as an isotopically distinct component of the human diet. Our findings suggest that stable isotope ratios of bottled water have the power to distinguish ultimate (e.g., recharge) and proximal (e.g., reservoir) sources of bottled water and constitute a potential tool for use in the regulatory monitoring of water products. Copyright © 2005 John Wiley & Sons, Ltd.