Analysis of mitochondrial (mt)DNA size polymorphism in the form of variable number tandem repeats (mtVNTRs) has become an increasingly popular methodology for addressing questions in molecular ecology. When detected by PCR, mtVNTR analysis can provide a sensitive, rapid, and cost-effective measure of genetic variability that may be exploited in studies of population differentiation and biogeography. Despite the emergence of this approach, there has been little critical evaluation of its success or utility as a practical tool. In this review, we identify problematic methodological, theoretical and interpretive factors that can influence the utility of mtVNTR analysis. The reliability of the procedure is considered in terms of both detection of alleles and scoring of intra-individual allele frequencies. While many of the potential technical problems of the technique do not raise serious practical concerns, this rapid and sensitive methodology is seriously compromised by the difficulty of reliably assessing allele frequencies, of assaying only germline tissue, and in our ignorance of the mechanisms generating mtVNTR diversity. Thus, although there is a considerable potential for mtVNTR pilot studies to assess genetic diversity, the utility of the technique to resolve broader questions in molecular ecology should be treated cautiously until such a time as the system is better understood.