We show here that the kinetoplast DNA (kDNA) networks from six Trypanosoma evansi strains differ from those of T. brucei by their lack of maxi-circles and absence of mini-circle sequence heterogeneity. The lack of maxi-circles is sufficient to account for the inability of T. evansi to multiply in tsetse flies, since this requires functional mitochondria containing maxi-circle gene products. Judged by restriction enzyme analysis, five of the six T. evansi strains contain mini-circles that differ less than 4% in sequence. This type A mini-circle is found in strains from East Africa, West Africa and South America. Another strain from East Africa contains a very different mini-circle (type B), which shows about the same degree of hybridization to type A mini-circles as to a mini-circle from T. brucei. We propose that the pronounced sequence heterogeneity of the mini-circles of T. brucei has arisen by recombination of strains that had diverged for long periods of time in reproductive isolation. We further propose that the homogeneous mini-circles of T. evansi (and T. equiperdum) reflect the inability of species to mate. This proposal implies that mini-circle heterogeneity indicates (infrequent) genetic exchange and that all kinetoplastid flagellates with heterogeneous mini-circles exchange DNA.