We studied the effects of carbon and aluminum substitutions on the gaps of the two-band superconductor MgB2 by means of point-contact measurements in Mg(B1−xCx)2 and Mg1−yAlyB2 single crystals with 0 ≤ x ≤ 0.132 and 0 ≤ y ≤ 0.21. The gap amplitudes, ∆σ and ∆π, were determined by fitting the conductance curves of the point contacts with the standard Blonder-Tinkham-Klapwijk (BTK) model generalized to the two-band case. Whenever possible, their values were confirmed by the independent fit (with a single-band BTK model) of the partial contribution of the two bands to the conductance, separated by means of a suitable magnetic field B ∗ . In C-substituted crystals, the two gaps remain clearly distinct up to x ∼ 0.10, but at x =0 .132 we observed for the first time their merging into a single gap ∆ � 3 meV with a gap ratio 2∆/kBTc close to the standard BCS value. In Al-substituted crystals, we found no evidence of this gap merging. Instead, ∆π reaches the value 0.4 meV at y=0.21, where ∆σ saturates at about 4 meV. These results are compared with other recent experimental findings in polycrystals and with the predictions of the models for multiband superconductivity. Most of the present research on magnesium diboride is devoted to investigate the effects of pressure, irradiation, lattice stress, disorder and, over all, chemical substitutions on the properties of this extremely interesting two-band superconductor. The purpose of these studies is to gain an insight into the effect of changes in some physical quantities that describe MgB2, such as interband scattering, density of states, electron-phonon coupling and so on. On one hand, this might suggest possible ways to tune the superconducting and magnetic properties of MgB2 so as to make it more suitable for power applications or superconducting electronics. On the other hand, the results of these investigations might allow testing some predictions of the two-band models, that have successfully explained most of the physics of the pure compound [1, 2]. A central prediction of these models, that has not been confirmed up to now, is the merging of the two gaps into one single, BCS-like gap. This merging is expected, for example, when the interband scattering increases significantly with respect to MgB2. In principle, chemical substitutions are the simplest and most convenient way to “perturb” a system and study the effects of these perturbations. However, they usually affect both the structural and the electronic properties of the parent compound and also introduce disorder. In complex systems like MgB2, this makes it difficult to identify the main cause of a certain effect experimentally observed. Moreover, obtaining partial substitution of Mg or B atoms in MgB2 is a difficult task. Aluminum and carbon are among the few chemical species that enter the MgB2 lattice, and even in this case problems of solubility [3], phase segregation [4, 5], inhomogeneities [6] and structural transitions [7] have been reported. Most of the experimental studies on Mg(B1−xCx)2 and Mg1−yAlyB2 have been carried out in polycrystalline samples. Among them, it is worth citing the few measurements of the superconducting gaps as a function of the