Low and variable rates of capture are common problems when estimating abundance of freshwater turtles with capture-mark-recapture (CMR). We speculated camera traps would allow us to obtain reliable estimates of abundance by re-sighting marked Trachemys scripta elegans (Sliders) as they basked on man-made rafts during a 20-day surveillance period. We evaluated the method by releasing Sliders in a fenced enclosure to compare estimates from CMR to true abundance. We also evaluated probabilities of detection and retention of marks. Permanence of marks applied with marine epoxy satisfied assumptions for CMR. Camera traps detected 23 of 25 Sliders. Our ability to discern marks from photos was good (110 of 114 re-sightings). The proportion of marked Sliders detected per day was 0.22; detection varied with day of surveillance (1–20) and maximum air temperature the preceding day. All CMR models providing valid estimates of abundance included the true number of marked Sliders in their confidence intervals and yielded point estimates within 27% of the true value. An estimate of abundance from the top CMR model exceeded the true value by 22%, with a wide confidence interval. Model averaging improved the point estimate (17% over true) and produced a narrower confidence interval. A favorable comparison of estimated and true abundance validated camera traps as a tool for estimating abundance of adult Sliders. We believe camera traps could prove useful for detecting biases caused by primary methods of capture, refining estimates of abundance from other methods and collecting data at multiple locations consistently, simultaneously and frugally compared to manual methods alone.