The letter by Rakhmanov et al1 suggested that normal κ/λ ratios in CD21+ naive B cells and increased Igλ-chain usage by CD21J/lo B cells in common variable immunodeficiency disease (CVID) patients with expanded CD21J/lo B-cell populations (CVID group-Ia2) were evidence of normal or increased receptor editing during central selection. Although increased Igλ-chain usage could correlate with extensive secondary recombination because Igλ genes normally rearrange after Igκ genes, the best evidence of receptor editing and central tolerance defects is found in proximal B-cell populations, specifically immature B cells in marrow or new emigrant/transitional B cells in peripheral blood. These subpopulations, similar in antibody repertoire and reactivity, offer an accurate historical record of central events uninfluenced by proliferation. 3, 4 Accordingly, the analysis by Warnatz and colleagues1 of Igλ-chain usage in CD21+ naive B cells and CD21J/lo B cells from 4 CVID group-Ia patients does not inform central events but implies peripheral selection in 2 B-cell subpopulations with welldocumented histories of homeostatic expansion. 4, 5 Therefore, we