After ripping the stocking #2 and reknitting the clocks and the heel I decided to knit a test foot. This time using big light yarn so I can see what's going on. This is the result:

The heel looks wrong because the yarn is thick, but it fits perfectly. When it's not worn it also buckles exactly like the original. I slipped the first stitch of every heel row and picked up one stitch per two rows (one st per sl1.)

The foot (after the heel) can be divided into three parts: gusset, foot and toe. In the original the gusset decreases take 21% (of the foot length) being themselves 30% long. The foot decreases and increases take 49% and the toe takes 30% of the foot length. In my test the gusset decreases are too long.

Another feature that differs from modern socks, is that the top of the foot has less stitches than the bottom. It means the decreases at the toe are tricky, but with a little bit of math they shouldn't be too hard to figure out. Sigh. I hate math. I tried to get the foot proportions right by just knitting, but it didn't work. So, math it is. I should have calculated the percentages in the first place, but... Have I told you I hate math?

One thing I got right was the balance between top and bottom and the angle of foot increases. The toe needs more tweaking, especially the beginning and the balance between decreases and grafting.

I can't decide should I knit another foot to test the calculations or should I trust my equasions and continue with stocking #2?