Obsidian hydration dating (OHD) originally promised to be a rapid, inexpensive, simple and reliable method for dating obsidian artefacts. The perceived elegance of OHD resulted in rapid acceptance and widespread application despite questions concerning its theoretical rigor. With increased usage it has become evident that, while economical, simple and fast, OHD is unreliable. Here results of a secondary ion mass spectrometry (SIMS) study of obsidian and synthetic glass artefacts are presented that explain why OHD has not lived up to expectations. The results suggest that the standard OHD equations are inappropriate and that traditional optical measurements are inherently flawed. Although significant refinement will be needed, the results suggest that both chronological and palaeoclimatic data might be obtained from glass hydration rims through an improved analytical methodology and more rigorous treatment of the data.