To study the effect of image resolution, eye tracking, and image registration on central macular thickness reproducibility (rCMT) among spectral-domain and time-domain optical coherence tomography (SD-OCT and TD-OCT) instruments. Seventy-six eyes were imaged (44 normal, 32 maculopathy) either twice using four SD-OCT and one TD-OCT devices or three times using Spectralis SD-OCT (with and without eye tracking) (Heidelberg Engineering, Inc., Heidelberg, Germany). Cirrus images (Carl Zeiss Meditec, Dublin, CA) were further analyzed with three-point image registration. All instruments had superior rCMT in normal versus pathologic eyes (P < .001). No difference in rCMT was noted among instruments in normal eyes (P = .92), but TD-OCT was superior to SD-OCT (P = .017) in pathologic eyes. Cirrus image registration improved rCMT for normal eyes (P = .04), with borderline improvement in pathologic eyes (P = .06). Spectralis eye tracking improved rCMT in normal (P = .01) and pathologic (P = .004) eyes. Higher image resolution with SD-OCT may not improve rCMT, but image registration and eye tracking options may improve rCMT. Copyright 2012, SLACK Incorporated.