To the Editor: Santos and Barrantes (1994) reported a 6-bp deletion be-tween nt 106and nt 111 of the mitochondrial DNA D-loop (Anderson et al. 1981), noting that it appeared to be specific for Chibcha-speaking groups (or, at least present in the ancestral population for Chibcha speakers). Merriwether (1993) described this deletion in two Aymara individuals (one from the Lluta Valley and one from Visviri) in northernmost Chile, indicating that this deletion extends considerably beyond lower Central America. The village of Visviri is located at 4,100 meters altitude in the Alti-plano, while the Lluta Valley sample is from a coastal Aymara population. We obtained D-loop sequences on> 200 Huilliche, Pehuenche, Atacameno, and Quechua Indians from Chile and Peru and did not detect the 6-bp deletion in any of these populations. The second hypervari-able region of the D-loop (where this 6-bp deletion occurs) is not sequenced as often as is the first hypervariable region in population studies, and only the lab of D. C. Wallace (Torroni et al. 1992, 1993a, 1993b, 1994a, 1994b) has consistently surveyed for the MspI-site loss at nt 104 (an MspI-site loss is caused by this deletion). Many populations surveyed by others (Ward et al. 1991, 1993; Ginther et al. 1993; Horai et al. 1993; Shields et al. 1993; Bailliet et al. 1994; Lorenz and Smith 1994; Merriwether et al. 1994) were not screened for this deletion. While Torroni et al.'s (1993a, 1994a, 1994b) putative 6-bp deletion haplotypes 51, 52, and 53 are foundon a lineage A background (char-acterized by a HaeIl 663 site gain and an MspI 104 site loss), our 6-bp deleted Aymara individuals are lineage D (AluI 5,176 site loss, HaeHI 663 site loss, and MspI 104 site loss). Kolman et al.(in press) and 0. Batista, CJ Kolman, and E. Bermingham (personal communication) report that this deletion is present at> 10% frequency in two Chibchan populations (Ngobe and Kuna) but is absent in two Choco populations (Embera and Waunaan) in Panama. The deletions observed by Kolman et al.(in press) and 0. Batista, CJ Kolman, and E. Bermingham (personal communication) all occurred on a lineage A background. Santos and Barrantes's Chibchan deletions also occur against a lineageA background (at nearly 60%). The occurrence of the 6-bp deletion on both lineage A and lineage D backgrounds could be indicative of multiple origins for the 6-bp deletion. Given the present data, one could consider the combination of a HaeIl 663 site gain (or other definitive lineage A markers) plus the 6-bp deletion (or an MspI 104 site loss) to be a Chibcha-specific combination. One should not use the 6-bp deletion alone as a marker indicating shared ancestry with the Chibchan or proto-Chibchan populations.