Peritubular capillaries (PCs) with a circumferentially multilayered basement membrane have been suggested as an ultrastructural indicator of chronic renal allograft rejection (CR). The authors validated this lesion as a marker for CR, by analyzing its quantitative features, specificity, and sensitivity in 169 renal biopsy specimens. The mean number of circumferential layers (PCcirc) and the incidences of the grades (mild: 2 to 4, moderate: 5 to 6, severe: 7 or more layers) were investigated in biopsy specimens involving CR (CR(Bx), n = 46), acute rejection (n = 11), normal kidneys (n = 20), psoriatics treated with cyclosporine (n = 13), renal transplants with chronic cyclosporine toxicity (n = 12), native kidney diseases (NKD, n = 56), and transplant nephrectomies attributable to CR (Cr(nephr), n = 11). CR was diagnosed with regard to the clinical features and the presence of intimal fibrosis in 41 biopsy specimens or transplant glomerulopathy in 35 biopsy specimens (cg; identified only by electron microscopy in 10 cases). NKD included chronic glomerulonephritis, chronic tubulointerstitial nephritis, benign nephrosclerosis, thrombotic microangiopathy, diabetic nephropathy, and renal disease in elderly patients (median age, 72 years). All PCs around glomeruli were sampled (median, 14 profiles per case). PCs with a moderate/ severe lesion appeared as serrated profiles with a thick, ribbon-like basement membrane layer in semithin plastic sections. The numbers of circumferentially multilayered PCs were significantly characteristic of CR (PCcirc in CR(Bx): 2.87+/-1.83 SD; range, 0 to 7.36; P < .001 v other groups). A severe lesion occurred exclusively in CR (in 12% of the PCs in CR(Bx), and in 38% in CRnephr). A moderate lesion was observed in 0.6% of the PCs in NKD, 16% in CR(Bx), and 21% in CRnephr. Three or more PCs with a moderate lesion were encountered only in CR. A mild lesion was not suggestive of CR at all. In CR(Bx), 27 cases showed a severe lesion or 3 or more PCs with a moderate lesion (cpc; sensitivity: 59%). Four of the 27 cases lacked cg. The cumulative incidence of cpc and cg was 85%. In transplants with cyclosporine toxicity, the presence of cpc verified the coexistence of CR in 7 specimens. In conclusion, cpc is a specific marker of CR. The incidence of cpc increases as CR progresses. The lesion may be caused by a low-grade rejection injury to the PCs. Careful analysis of semithin sections promotes the better sampling of cpc. An ultrastructural demonstration of cpc and cg defines CR more precisely than does light microscopic evaluation per se.