Mars Express observations give ozone abundances that are smaller than those from the ground-based infrared heterodyne and HST observations at low and middle latitudes. Both ground-based and Mars Express observations of the O 2 dayglow at 1.27 ฮผm, which originates from photolysis of ozone, are in mutual agreement after correction for the local time variability. Therefore a problem appears: whether the MEX ozone data are compatible with (1) the observed O 2 dayglow intensities and (2) the photochemical model by Krasnopolsky [Krasnopolsky, V.A., 2006. Icarus 185, 153-170] within uncertainties of its reaction rate coefficients. That model involves heterogeneous loss of H 2 O 2 on water ice and agrees with the observations of the O 2 dayglow, H 2 O 2 , and the ground-based and HST ozone. The answers are 'yes' to both questions. A version of the model is given that fits the MEX ozone as well as the observed O 2 dayglow and H 2 O 2 . Laboratory studies of two reaction rate coefficients could indicate a preferable version of the model and a preferable set of the ozone data (MEX versus the ground-based and HST). The predicted seasonal behavior of H 2 O 2 is different from that in the model by Lefevre et al.