Ed - you're entirely correct. Erin is my only storm that I've truly been through -- the center passed directly over my house -- but I was in Orlando at the time. The most we saw was 50-55mph winds in localized areas, which pales in comparison to what you saw. I should be a bit more careful in my words, as it only takes one storm -- as we found out so many times over last season.

As for the SW Caribbean -- there is some sort of circulation, but it was not captured by the last QuikSCAT scan (sometime this morning); however, it did show some pretty strong winds on the north side -- up to 50kt (mostly rain-flagged vectors, though). There should be another scan in the next few hours that may provide more insight. The convection -- really an MCS (mesoscale convective system) -- helped provide the surface pressure falls to get the circulation going; it'll be another matter if the circulation can maintain itself. Judging from visible imagery, which in itself is somewhat obscured on the south side by high clouds & convection, I would be willing to say that it is a near-surface or surface circulation that has developed. Convection near the center has died for the most part (save to the east) for now; something's going to have to organize again to keep the circulation from opening up on the southern side, however.

Wind shear has decreased in the region, but is still a bit on the high side. Any movement northward is going to rapidly take the system into a region of higher shear; the jet also appears to be slowly approaching again from the west, not a good sign for development. 18Z NHC analysis doesn't call for a low to develop there any longer, though neither does it analyze something there now. However, while I typed up the post initially, NRL put out a 91L INVEST on the system...talk about timely. Models still try to build the ridge in the SW Caribbean, but not for another 2 days. By then, whatever is there will have either developed extratropically or simply dissipated. Nevertheless, there is a narrow window for something; I just don't see it happening. Good practice for analysis, though. Ultimately, we could well end up like last May's Hispaniola storm, as previously alluded to here on the board.