@Visisthebest: Yep. I understand and understood. My remark was to confirm that I did not have to do your extra step. For me, the variation in this behavior for the various responders is very strange. Historically, the performance has been more like what Surrogate and Croc observe. I'm surprised it mostly works for you. The concern is how reliable is this approach? It strikes me as being less than robust. At this point, I'd say, "User beware!"