This paper discusses the problems faced when making interpretations of human stable isotope values due to the various explanatory alternatives that arise when reading archaeological data. These interpretative issues are analysed and discussed using the isotopic results for approximately 100 human individuals from archaeological contexts spanning from 5000 BC to 1540 AD in central Chile, supported by data for more than 50 plant and animal samples to establish a local C₃ and C₄ baseline. A number of assumptions are frequently used to establish the bridge between isotopic results in human tissues and their corresponding diets. The problem is that different assumptions lead to different dietary reconstructions. Past feeding experiments on herbivores, pigs, rats and mice give different results, so we need to be cautious when applying these models to human isotope data. One specific problem concerns estimates of % C₄ from collagen and apatite data, a very important issue when looking for evidence of maize in archaeological contexts, which was one of the major objectives we had in the isotopic analyses of archaeological specimens in central Chile. We conclude that the opportunity for estimating the actual percentage of C₄ foods in human diets is limited, since a specific apatite fractionation value for humans cannot be experimentally determined, while maize consumption is underrepresented in bone collagen. This may be addressed in our study by sampling more specimens of wild gramineae to establish baseline plant values, more humans that could have had a low maize intake, and more Archaic period individuals when there was certainly no maize in the region. Copyright © 2009 John Wiley & Sons, Ltd.