Species identification of blow flies is the initial step used in forensic investigation. Identification techniques usually are based on morphology and molecular analysis, the former requiring an experienced taxonomist while the latter laboratory expertise and considerable expense. Landmark-based geometric morphometric analysis is being widely used as an alternative technique in species identification as it is simple to perform at low-cost and requires only basic training. However, it is recommended using the same individual in digitizing the landmark coordinates to reduce measurement error, but the results from using many individuals to digitize the landmark coordinates have not been addressed in detail. Here, intra-individual and inter-individual variations in the digitization wings of Chrysomya megacephala and Chrysomya rufifacies, the two forensically important blow fly species, were assessed. Five individuals digitized 19 landmarks on the right wing of each species using the same photograph, performed in duplicate. Analysis of the results using Procrustes ANOVA demonstrated low (8.9-19.0%) measurement error of intra-individual variations but higher (23.1%) for that of inter-individual variations. Based on discriminant function analysis, C. megacephala wing shape was clearly differentiated from that of C. rufifacies with high reliability. A cross-validation test indicated high (93.3-100%) accuracy of recognition of C. megacephala intra-individual variations and 100% accuracy for inter-individual variations, while that of recognition of C. rufifacies was 100% for both intra- and inter-individual variations. Thus, although measurement error was obtained from both intra- and inter-individual variations, correct identification of C. megacephala and C. rufifacies was achieved.
