I have a few 429421s, and 358429 that date back to the '40s, according to the information I have gathered, the most reliable of which came from Ric Bowman. These moulds have square lube grooves, but the front drive band on every one of them is undersized. They are cut in unvented blocks. These are not difficult to find, the mould finish is not the very early style, and they have a brushed finish. I can't help but to fall back on something said earlier about Lyman/Ideal never making a Keith design to Keiths specs. I have heard that one of the big name gun writers does have an original Ideal Keith mould, and it is cut to Keiths specs. I believe this is completely possible, and that this particular mould may have been one of the original special order, or Keith prototypes. I can only hope for an article some day.
The answer to the varying versions seems to stem from Lyman/Ideals use of outside vendors to cut cherries for them. It would be interesting to see the original blueprints the machine shops had to work from, were the changes made on the design side, or the manufacturing side? I see the same issues with other designs, and they seem to share common issues, like the front driving band length. All the Keith designs seem to show variances at that point, at nose length, and in weight. Lymans catalogues in the past show weight changes for 358429 running from 170 to 168 grains over the years. Some of that may come from changing their alloy specs from Lyman #2 to linotype, but they're showing 170 gr for lino in the Third Edition of the Pistol & Revolver handbook, so that doesn't fit either.