Paleoenvironmental records provide insights into fundamental questions about structural changes in past human populations. Three problems with proxies lower the resolution of such records: 1) spatio-temporal discontinuity, 2) reflecting multiple factors at once, and 3) decomposition beyond recognition. Analyses of specific lipid molecules that "don't do anything" in soils and sediments are overcoming these obstacles. These are products of decomposition that resist further chemical change. Stable hydrogen isotopic compositions (δD) of molecules with long carbon chains in fatty acids and alkanes are providing records of paleoprecipitation. The records' accuracy may be undermined by changes in cover of their plant sources. Correcting for changes in C3 and C4 plant cover did not alter δD trends from fatty acids in Ethiopian soil sections and reflected shifts in precipitation amounts. In contrast, δD values of specific alkanes in soils along an elevation transect of Ethiopian agricultural stations did not reflect precipitation trends unless differences in C3 and C4 plant cover were accounted for. Pentacyclic triterpene methyl ethers (PTMEs) include potential biomarkers for specific plant types. The PTME miliacin is specific to broomcorn millet in the French Lake Bourget area. In addition to providing precipitation inferences that are not affected by changes in plant sources, its contents in sediments provide insights into millet agriculture by Bronze Age peoples where artifacts such as seeds have decomposed. Another PTME, crusgallin, is found only in grasses and its δD values are showing greater sensitivity to changes in precipitation than fatty acids from soils near ancient Ethiopian civilizations.