Although the extra-solar planets discovered so far are of the giant, gaseous, type, the increased sensitivity of future surveys will result in the discovery of lower mass planets. The detection of O2 in the atmosphere of a rocky extra-solar planet would be a potential indicator of a life. In this paper we address the specific issue of whether we would be able to detect the O2 A-band absorption feature in the atmosphere of a planet similar to the Earth, if it were in orbit around a nearby star. Our method is empirical, in that we use observations of the Earth's O2 A-band, with a simple geometric modification for a transiting extra-solar planet, allowing for limb-darkening of the host star. We simulate the spectrum of the host star with the superposed O2 A-band absorption of the transiting planet, assuming a spectral resolution of 7 km/s (typical of current echelle spectrographs), for a range of spectral signal-to-noise ratios. The main result is that we could reliably detect the O2 A-band of the transiting planet for host stars with radii 0.3 solar or less. However, using existing instrumentation and 8m telescopes, this requires target M stars with m(V) of approximately 10 or brighter for integration times of about 10 hours or less. The number of such stars over the sky is small. Larger aperture telescopes and/or improved instrumentation efficiency would enable surveys of M stars down to m(V) = 13 and greatly improve the chances of discovering life elsewhere.