A simple continuous wave near-infrared algorithm for estimating local hemoglobin oxygen saturation in tissue (%StO2) is described using single depth attenuation measurements at 680, 720, 760, and 800 nm. Second derivative spectroscopy was used to reduce light scattering effects, chromophores with constant absorption, baseline/instrumentation drift, and movement artifacts. Unlike previous second derivative methods which focused primarily on measuring deoxyhemoglobin concentration; a wide 40 nm wavelength gap used for calculating second derivative attenuation significantly improved sensitivity to oxyhemoglobin absorption. Scaled second derivative attenuation at 720 nm was correlated to in vitro hemoglobin oxygen saturation to generate a %StO2 calibration curve. The calibration curve was insensitive to total hemoglobin, optical path length, and optical scattering. Measurement error due to normal levels of carboxyhemoglobin, methemoglobin, and water absorption were less than 10 %StO2 units. Severe methemoglobinemia or edema combined with low blood volume could cause StO2 errors to exceed 10 StO2 units. Both a broadband and commercial four-wavelength spectrometer (InSpectraTM) measured %StO2. The InSpectra tissue spectrometer readily detected limb ischemia on 26 human volunteers for hand, forearm, and leg muscles. A strong linear correlation, r2>0.93, between StO2 and microvascular %SO2 was observed for isolated animal hind limb, kidney, and heart.