We transform the two-matrix model, studied by P.Di Francesco and al., into a normal one-matrix model by identifying a 'formal' integral used by these authors as a proper integral. We show also, using their method, that the results obtained for the resolvent and the density are not reliable.