Aerological data from the Air Mass Transformation Experiment (AMTEX) conducted during the period 14 February–1 March 1975 have been analyzed to study the mixed layer and capping inversion layer associated with mesoscale cellular convection (MCC) created in a cold continental air mass heated from below by a warm ocean surface. Six horizontal analyses of the depth of the mixed layer were superimposed on their respective satellite images which contained MCC events in the AMTEX region. In all six cases, open cells tended to lie in the troughs and closed cells in the ridges of the convective depth contour pattern. Vertical cross sections revealed that the actual transition from a coexisting region of open cells to an adjacent region of closed cells may actually occur at a point of inflection where the mixed layer depth contour pattern changes from a plano-concave shape to a plano-convex shape. Mean convective depths for open and closed cells were 1529 and 2066 m, respectively, based on 15 soundings in regions of open cells and 16 soundings in regions of closed cells. Also, vertical profiles of potential temperature in regions of open and closed cells were obtained, with the convectively mixed layer characterized by a superadiabatic lapse rate near the surface, adiabatic through the subcloud layer and moist adiabatic through the cloudy regions. Mean vertical potential temperature gradients and geometric thicknesses of the inversion layer for open and closed cells were, respectively, 21.9 K km−1 (437 m) and 25.5 K km−1 (480 m). The time cross sections of potential temperature indicated two types of undulations in the depth of the mixed layer due to 1) the passage of regions of coexisting open and closed cells of synoptic-scale proportions, and 2) the elevations and depressions of the inversion base associated with the passage of individual MCC. Three cases using the potential temperature field and the relative humidity field were used to construct observational “thermodynamic” models of an open cell. Features were (i) the inversion base and top were higher in the wall (cloudy) region than in the central (clear) region, (ii) the potential temperature pattern contained in the inversion layer followed the shape of the inversion base, (iii) in each case, the potential temperature pattern was reversed above the inversion layer, and (iv) the relative humidity pattern was in phase with the inversion base at all levels. Quantitative estimates were made for all terms in Arakawa’s formulation for the time rate of change in the depth of the planetary boundary layer (PBL). Local changes observed in the depth of the PBL were ±2.8,μb s−1. Values for the advective term were in the range of ±5 μb s−1. Large-scale vertical motion was always downward leading to a decreasing thickness in the PBL depth with time. Average values for open cells and closed cells were, respectively, 3.6 and 2.5 μb s−1. The entrainment term was always positive and tended to counteract the effects of the large-scale vertical motion. Mean values for the open and closed cell regions were 1.7 and 1.1 μb s−1, respectively. The penetrative convection term was assumed negligible based on photographic evidence. The net effect of the entrainment and vertical motion terms is to support greater thickness of the PBL in regions of closed cells compared to regions of open cells, which was consistent with the observations. Implications of a complete budget analysis, including radiation and latent heat released, are also discussed.