Top panel: comparison between surface brightness profile from (Kormendy et al. 2009; crosses) and logarithmic PN number density profile of the emission line candidates, brighter than m₅₀₀₇ = 28.3, and corrected for spatial and colour incompleteness (open circles with error bars, see Table 4 for data), as function of the distance from the M 87 centre. The black dotted line represents the contribution of Lyα emission objects to the logarithmic PN number density profile, assuming a homogeneous distribution over the surveyed area. Under the same hypothesis, filled circles show the logarithmic PN number density profile when the Lyα contribution is statistically subtracted. Red lines mark the inner, intermediate and outermost regions of the M 87 halo (see text). Bottom panel: difference between the logarithmic PN number density profiles if we had used the automatically extracted sources without the final inspection: a value of zero would mean that no variation in the number of sources is implied in the annulus at a given radius.