Chromium tris-2,4-pentanedionate becomes quite active as an ethylene polymerization catalyst when sublimed onto an acidic high-porosity carrier and in the presence of aluminum alkyl cocatalyst. The resultant catalyst produces a broad MW distribution of mostly linear polyethylene. Because the MW distribution reflects the active site distribution, it has been studied in this paper as the catalyst was probed by changing the organic ligand, the acidic carrier, the cocatalyst and reactor additives. The MW distribution was quite responsive to changes in the carrier and the cocatalyst, but not to changes in the organic ligand. These responses were then compared to those of a commercial relative, the Phillips chromium (VI) oxide catalyst system anchored on the same supports. Many common behaviors were observed, and also some contrasting ones. For example, the incorporation of 1-hexene, and the resulting branch placement within the MW distribution were identical between the two catalysts. On the other hand, the role of the support calcination temperature as a MW regulator was greatly reduced from the experimental catalyst, but the response to reactor H-2 was much more pronounced than that of Phillips catalysts. Thus, the experimental catalyst offers many of the same features of commercial Cr catalysts, but with some environmental and operational advantages too. (C) 2016 Elsevier B.V. All rights reserved.