Abstract A series of seven dedicated Lyman- α observations of exospheric atomic hydrogen in the martian corona were performed in March 2005 with the ultraviolet spectrometer SPICAM on board Mars Express. Two types of observations are analyzed, observations at high illumination (for a solar zenith angle SZA equal to 30°) and observations at low illumination (for a solar zenith angle equal to 90° (terminator), and near the south pole). The measured Lyman- α emission is interpreted as purely resonant scattering of solar photons. Because the Lyman- α emission is optically thick, we use a forward model to analyze this data. Below the exobase, the hydrogen density is described by a diffusive model and above the exobase, it follows Chamberlain's approach without satellite particles. For different hydrogen density profiles between 80 and 50,000 km, the volume emission rates are computed by solving the radiative transfer equation. Such an approach has been used to analyze the Mariner 6, 7 exospheric Lyman- α data during the late 1960s. A reasonable fit of the set of observations is obtained assuming an exobase temperature between 200 and 250 K and an exobase density of ∼1–4 × 10 5 cm −3 in good agreement with photochemical models. However, when considering the average exospheric temperature of 200 K measured by other methods [Leblanc, F., Chaufray, J.Y., Witasse, O., Lilensten, J., Bertaux, J.-L., 2006a. J. Geophys. Res. 111 (E9), doi:10.1029/2005JE002664. E09S11; Leblanc, F., Chaufray, J.-Y., Bertaux, J.-L., 2007. Geophys. Res. Lett. 34, doi:10.1029/2006GL028437. L02206; Bougher, S.W., Engel, S., Roble, R.G., Foster, B., 2000. J. Geophys. Res. 105, 17669–17692; Mazarico, E., Zuber, M.T., Lemoine, F.G., Smith, D.E., 2007. J. Geophys. Res. 112, doi:10.1029/2006JE002734. E05014] a supplementary hot population is needed above the exobase to reconcile Lyman- α measurements with these previous measurements, particularly for the observations at low SZA. In this case, the densities and temperatures at the exobase for the two populations are 1.0 ± 0.2 × 10 5 cm −3 and T = 200 K and 1.9 ± 0.5 × 10 4 cm −3 and T > 500 K for the cold and hot populations respectively at low SZA. At high SZA, the densities and temperatures are equal to 2 ± 0.4 × 10 5 cm −3 and T = 200 K and n = 1.2 ± 0.5 × 10 4 cm −3 and T > 500 K . These high values of the hot hydrogen component are not presently supported by the theory. Moreover, it is important to underline that the two population model remains relatively poorly constrained by the limited range of altitude covered by the present set of SPICAM measurements and cannot be unambiguously identified because of the global uncertainty of our model and of SPICAM measurements. For a one population solution, an average water escape rate of 7.5 × 10 −4 precipitable μm/yr is estimated, yielding a lifetime of 13,000 years for the average present water vapor content of 10 precipitable microns.