The comparison of two multiple regression models is based on the assumption that geochemical composition of the drainage basin alluvial sediments is derived primarily from the underlying bedrock lithology. The parent material is integrated with both stream sediment and overbank sediment geochemistry via the two essentially different approaches as regards the drainage basin geomorphological data: (1) as the relative area of influence representing a portion of the catchment basin occupied by a specific rock type and (2) as the relative “line” of influence representing a narrow zone of the underlying bedrock traversed by the perennial streams which form the active stream network. The model comparison is established on the goodness-of-fit test for both experimental designs and for the same set of data. Both experiments converge on the linear approach as the more appropriate model in evaluating the lithologic influence on the analysed sample media in small mountainous watersheds.