The findings confirm that Y STR haplotypes are highly polymorphic and have a high power of discrimination in forensically relevant U.S. populations. The three Texas subpopulations provide data for the upper bound of the effect of population substructure. For haplotypes that comprised at least 10-16 markers, analyses showed the effects of population substructure were small, such that there was little or no need to correct for population substructure when estimating the conditional probability of a Y STR haplotype using the counting method. The counting method with a correction for sampling error was apparently sufficiently conservative; however, estimates of the conditional probability of partial profiles, depending on the number of markers and haplotype sharing, may require a correction for population substructure. The authors recommend a pragmatic maximum Fst value approach or a specific marker Fst value approach for conditional probability estimates for partial profiles where substructure correction is warranted. Both approaches are valid and conservative. The data also indicate that frequency estimates of autosomal and Y STR profiles can be combined by multiplying under the assumption of independence. DNA was obtained from unrelated male donors from paternity-testing cases submitted to the DNA Identification Lab at the University of North Texas Health Science Center, Ft. Worth, TX. Population affinity was assigned by self-reports. The samples consisted of 950 African-Americans, 957 Caucasians, and 1,005 Hispanics. The descriptions of materials and methods address sample collection procedures, sample preparation, Y STR typing, autonomic STR typing, and statistical analyses. 7 tables and 27 references