Hi Rob, Bernd, when I bought the DRO for my mill, I thought I was doing it primarily to be able to accurately machine with it, drilling holes where they are supposed to be, and putting slots right on target and the like. My first experience was with bolt circles, and I thought, "what a great extra", and went on. It was only when I stripped a fifty one tooth gear in a valve reseating machine, and realised I needed to be able to do differential indexing to get fifty one, and of course don't have a differential indexing setup, that it occured to me I could do a fifty one hole bolt circle. Rob, when you are using your indirect indexing, meaning not using the indexing plate, but using the worm, the index plates, and a hole circle, in the case of a forty to one dividing head, you divide the number of holes by forty, and you then select an index plate with holes sufficient to account for any "carry over". Since the smallest plate with fifty one as a factor is a fifty one hole plate, one must have it, or a multiple of it to do the indexing indirectly. With the fifty one hole plate, one completes the circle by moving the pin forty holes in the fifty one hole circle, not counting the hole the pin is in as one of those holes. Most indexing books use examples that are almost useless, such as twelve divisions, or sixteen, where there is no challenge in finding an answer. Think of it this way: forty fifty oneths, turned fifty one times will bring the indexing pin around fifty one times, and will equate to the full circle, and account for forty teeth being used to get all the way around. By the way, I did make a direct indexing plate before the light bulb lit, but didn't use it as my mill power down feed chose to keep feeding long after the lever popped out, and gave me one hole three times as deep as the other fifty, and I didn't trust it's accuracy. I see this as opening the way to transposition gears for doing metric gears on my American lathe, and getting the gears as accurate as the ones you can buy at highly inflated prices. It struck me that it is even more accurate than "differential indexing", which actually only provides an approximation for most counts between fifty and a thousand. One can put an hundred line plate on an indexing head, with a cursor line, and calculate the divisions on the plate to make any count, and with a hundred division plate, and a forty tooth worm, this is accurate to one part in four thousand, and with a simple vernier on the cursor plate, bring it to one in forty thousand, however that is extremely tedious and very easy to lose track of exactly where you are at, and would best be left to a stepper motor setup, than done by hand, although many people have done good work with such a setup. mad jack