Although most laboratory studies have found respiration to be of limited value in polygraph tests, Timm (1979) found thoracic (upper chest) respiratory tracings to be slightly superior in detecting deception compared to skin resistance responses. In Timm's study, a different objective procedure was used to quantify responses than had been used in other studies. A similar analysis procedure was used in the current study. The study randomly selected charts from 25 'truthful' and 25 'untruthful' subjects in a verified case file. The curvillinear lengths of the respiratory tracings measured on the charts were determined using a Mnemonics electronic digitizer (model #1224). This equipment permitted the determination of those lengths by tracing over the respiratory pattern located within specified points with a magnifying cursor. It was hypothesized that the smallest response would occur more frequently on the control questions for 'truthful' subjects, while the smallest response for 'untruthful' subjects would occur on the relevant questions. The data were also analyzed to determine whether thoracic or abdominal respiration was superior for detection of deception. Thirteen percent of the 46 classifications based on thoracic respiration indicated the subjects were innocent when they were actually guilty, and 15 percent indicated the subjects were guilty when in fact they were innocent. Although the levels of detection efficiency were less than perfect, the study demonstrates the potential utility of developing empirically based objective criteria for analyzing polygraph charts. Tabular data and 18 references are provided.