Legs – Part 2


There are several ways to construct an octagon, and I’m not going to say this way is best.  If you divide the full width of the octagon into seven parts, then come in 2/7ths from each edge, those points are pretty close to where you need to be.  When I did this in CAD using an overall width of 2–5/8″, the points were only off by about 1/64″.  I doubt I could get any closer with the mathematically-correct and socially-acceptable four-circles method, considering the challenge of balancing compass points on the corners of a piece of wood, the relative sharpness of graphite, the vagaries of transferring marks from endgrain down the sides of the legs, etc.

I made a cradle to hold the legs and started planing the chamfers.  I began with a few passes of the fore plane.  Then I used the block plane with chamfering attachment.  This was not strictly necessary, but it allowed me to check that I was staying close to 45°.  Then it was back to the fore plane to get closer to the lines.  I then went more carefully with the jointer plane, and finally a little work with the smooth plane to finish.


In the book, Schwarz says if you don’t have a lathe you can pare the tenons with a chisel.  So that’s what I set out to do.  First I cut the shoulder with a tenon saw.  Then I sawed some of the waste.  Then I chopped some of the waste with a chisel and mallet.  Then I did some paring.  I started with a shallow depth, well above the final shoulder, and it seemed to be going alright.  It wasn’t exactly pretty, but it fit the test hole.

But after working on the rest of the tenon for a while… well, I just wasn’t happy.  My chisel wanted to follow the grain, which sometimes dived within the desired diameter.  I found it challenging to keep the evolving pseudo-cylinder aligned with the leg; it was starting to look like I was veering off course.  It sure wasn’t very tidy.  In a table where the legs would be permanently set in place and the tenons forever hidden, it wouldn’t matter so much, especially if they were wedged to make up for inconsistencies.  But because of the portable nature of this desk, the tenons are at least occasionally a show surface, and they need to be close to a friction fit.