These experiments were designed to evaluate DQ alpha typing for analysis of evidentiary specimens. Bloodstains were exposed to a variety of conditions and environmental insults. These conditions included exposure to many different types of substrates, various microorganisms that could be encountered in evidentiary stains, sunlight, and a variety of chemical contaminants. Varying amounts of genomic deoxyribonucleic acid (DNA) were amplified to test the sensitivity of DQ alpha typing. The sensitivity of the PCR technique raises the concern that DNA from sources other than the evidentiary material could be detected. A series of experiments was done to evaluate the question of DNA contamination. Purified DNA samples with different DQ alpha types were mixed in different ratios to determine the ratio at which it could not be determined whether an allele was from the sample or the contaminant. Samples were exposed to a variety of situations that could lead to contamination such as extensive handling and exposure to coughing or sweaty clothing, to other wet bloodstains, and to saliva. The DQ alpha types were determined from 469 individuals from three sample populations (Caucasian, black, and Hispanic), and the genotype frequencies were compared with frequencies previously reported by others. DNA samples from old cases (which had previously been analyzed by restriction fragment length polymorphism (RFLP) typing of variable number of tandem repeat sequences) were typed. All samples that were excluded by DQ alpha typing were also excluded by RFLP analysis, and all samples that were included by RFLP analysis were included by DQ alpha typing. Finally, the problem of allele dropout, or the failure to detect particular alleles, was noted and alleviated by performing the typing under appropriate conditions. The results of these validation experiments indicate that typing of the DQ alpha gene by PCR and detection of specific alleles can be accomplished, when the typing is done using proper protocols, without producing false positive or false negative results. 18 references and 7 tables (Author abstract)