Background: The quantitative viral outgrowth assay (QVOA) is the gold standard for measuring the latent reservoir of HIV infection, yielding precise measurement of the reservoir of resting CD4+ T cell infection (RCI). But variability of RCI over time on ART, relevant to assess potential effects of antilatency interventions, has not been fully described. We report an analysis of 160 QVOAs. Methodology: HIV+ male participants, stably HIV RNA <50 c/ml on ART, donated resting CD4+ T cells via leukapheresis. Acute HIV infection (AHI) patients (plasma HIV RNA detected and HIV Western blot negative) were identified via the North Carolina STAT program, and initiated ART within 45 days of the estimated diagnosis (range 7-45, median 19 days); one patient initiating ART 6 months after infection (early infection) was also studied. Once HIV RNA was <50 copies/ml for 6 months, leukapheresis was performed and RCI quantitated. QVOA was performed using 30-225 million resting CD4+ T cells in 24-96 replicate cultures in limiting-dilution format. RCI was similarly measured for patients initiating ART during chronic infection (CHI). Random effects regression evaluated RCI decay and estimated sources of variability, restricting to RCI measures obtained ≥12 months after ART. Results: 17 AHI and 20 CHI were studied (ages 19-66 years), including a total of 160 independent QVOA measures (median 4 per patient at median 3 months intervals; median 1.2 years RCI follow-up). Median ART duration at first RCI was 21 months AHI, 78 months CHI. CD4 count at first RCI was 744 cells AHI, 584 CHI. RCI significantly declined over time (p<0.001) with estimated mean (95% CI) half-life = 3.6 (2.3-8.1) years. There was no evidence of more rapid decay for AHI vs. CHI (p=0.99) after ≥12 months of ART. From the model, the estimated SD for the change in log10 RCI for a patient = 0.38, implying that a >2.5 fold RCI decrease over 2 months would have likelihood 0.16 (16%) under stable ART. A >6 fold RCI decrease would have likelihood 0.023 (2.3%). Empirically evaluating the 123 pairs of consecutive RCI measures, 81 (66%) had <2 fold increase or decrease; 21 (17%) had >2.5 fold decreases and 3 (2.4%) had >6 fold decreases. Conclusions: Consistent with prior studies, RCI decayed with a half-life of 3.6 years (43 months). Based on using large numbers of cells obtained via leukapheresis, RCI was reliably estimated with longitudinal measurements generally showing <2 fold variation from the previous measure. RCI decreases >6 fold were rare on stable suppressive ART, suggesting a potential threshold to identify effects of anti-latency therapeutics on RCI.