Monolith reactors used in catalytic after-treatment systems (e.g. TWC, DOC, LNT and SCR) are often studied in laboratory experiments using monolith samples of the same cell density, washcoat thickness, and catalyst loading but with a smaller length compared to the full-scale reactor. The flow rate in the laboratory scale operation is selected so that the same space velocity is maintained as in the full-scale system. However, matching space velocity between the laboratory-scale and full-scale reactor leads to similar performance only under ideal conditions such as isothermal or adiabatic operation with negligible heat and mass dispersion (i.e. plug flow conditions) and small adiabatic temperature rise. In this work, it is shown that when thermal effects are significant (or adiabatic temperature rise is large) and heat and mass transfer and dispersion effects are not negligible, similarity may not exist and the ignition/extinction behavior of the two systems can be qualitatively different. Further, when laboratory-scale data is used for kinetic parameter estimation, failure to consider these effects in the reactor model can lead to false kinetic parameters and inaccurate prediction of reactor performances. Another major reason for possible discrepancy between laboratory-scale and full-scale reactors, external heat exchange with furnace (surroundings), is also investigated. We present analysis, simulations, and analytical results quantifying the impact of these effects on the scale-up of monolith reactors. (C) 2015 Elsevier B.V. All rights reserved.