ABSTRACT
The analysis of dust allows inference of exposures to geographical areas, environments, activities, and processes. This activity of inferential source attribution is distinguished from that of comparative source attribution, where the focus is on the degree of correspondence between two sources in relation to other possible sources. Review of source attribution efforts in the forensic and broader scientific literature shows that most efforts are limited in one or more of four principal ways, which are classified as: (a) methods based on attribution by direct comparison; (b) methods based on closed-set item classification; (c) analysis using restricted methods and characteristics, and (d) requirement of a large sample size. These limitations provide the context for the requirements of more generalized inferential source attribution. Occurring much more rarely, and almost exclusively in the forensic literature, are individual source attribution case reports that have a microscopical, multidisciplinary perspective. Collectively these are an excellent illustration of potential and their common features demonstrate that (a) a diversity of laboratory expertise and methodology is required in order for source attribution to be successful; (b) different tools need to be applied in different cases, and (c) a process must be in place that allows a facile choice among this diversity of tools, in response to the particular investigative problem and the specifics of the samples that are available. Alternative collaborative mechanisms are considered and recommendations are made for related research and programmatic application.
