The detection of 56Co emission from SN 1991T has been previously reported at a level near the COMPTEL sensitivity threshold. The spectral analysis method, fitting the count spectrum to a background model plus a 56Co emission template, is subject to possible systematic effects which had not been thoroughly studied at that time. To better evaluate the significance of that ∼3.3σ detection, the same method has been applied to a grid of points with 5° spacing, out to 35° from the pointing direction, in each of 103 observing periods from phases 1 through 3. A dozen instances were found with a 56Co signal as significant as that for either of the two observations of SN 1991T alone (∼2σ). Nothing was found as significant as the combined observations of SN 1991T. The strongest instrumental background artifact in the vicinity of the two principal 56Co lines, attributed to 27Mg, falls between the 56Co lines. It fills in the valley between those lines, and so will obscure real 56Co emission rather than producing false...