While logical reasoning skills are related to success in science, the relationship between thought and formal logic is not clear. Propositional Logic Test (PLT), error patterns of students' interpretations of logical proposition, and developmental patterns of the error type are described. Examined is the relationship between logical reasoning ability and school science grade in tenth-grade Western Australian students (N=226). The PLT, Test of Logical Reasoning (TOLT) and Nonsense Syllogisms Test were used for measuring the logical reasoning ability. Correlations of grade in science with the PLT (.57) and the TOLT (.63) were high and the three reasoning tests together explained approximately 45 percent of the variance in science grade. An error analysis was undertaken for each subtest of the PLT, conjunction, disjunction, biconditional and implication across three ability groups. A discussion of the use of the PLT success rate differences is included. (YP)