The usefulness of radiocarbon dates in archaeology greatly depends on both the stratigraphic relationship of the sample submitted and on the origin and homogeneity of the measured carbon. For very small samples, stratigraphic relationships can raise additional problems of movement. In chemically well-characterized materials, the best example being collagen, the carbon source can be reasonably well purified. Many samples, however, survive as a complex mixture of high molecular weight polyphenolic materials, with properties between charcoals, humic acids, and lignins. Charred bone, eg, which rarely contains useful quantities of amino acids, and charred seeds, as well as ‘charcoal,’ frequently come into this category. For such samples, the likelihood of contamination by percolating soil humics is high. It is often possible to extract chemically different fractions and to compare the dates obtained. A less exact comparison can also be made for different samples from the same context. The results suggest that ‘humic’ acid dates can be reliable in a surprisingly frequent number of situations, and that where direct comparison is possible, the reliability can be individually assessed.