still makes no sense, a fitting discussion follows
Bill Tindall, E.TN
>I am failing to understand why a rough surface is going to glue better than a smooth surface under any circumstances. The argument that the smooth surface will be glue starved may be true if the joint is too tight. (see below) The opposite argument is that the roughened surface will only be optimally bonded in places.
As far as the potential for PVA glue to result in a different case than hide glue, data in the literature shows that PVA glue makes stronger M&T joints, and especially so with M&T joints cut too loose.
Of course the practical side of all this is that the M&T joint is so strong that failure is unlikely even with a less than perfect fit, so long as the glue maintains its properties. I suspect the best argument for a sawed joint is speed and the rest is the rationalization.
In the class I recently took in the Mack Headley and Sons cabinet shop (Class name is "Woodworking Shops of the Shenandoah Valley") I learned an important lesson in fitting. (Every one that has not had a joint set up prematurely upon assemble raise your hand). This shop uses white glue, as does College of the Redwoods. The shop does a lot of veneering and the white glue does not stain like yellow glue (we were told). To lessen inventory they use white glue for everything. ( and I could make a compelling case for why white glue is vastly better than hide for veneering should anyone care to listen.) Anyway, everything is dry fitted before glue application. For the dry fit they wanted to be able to tap the joint together with the palm of the hand. Using this criteria for acceptable fit, I had no trouble assembling dovetail, long sliding dovetail, and M&T joints after glue application. I was particularly terrified of the 10" full sliding dovetail, but it went together effortlessly. This shop assembles hundreds of joints in a year, so this criteria has been thoroughly tested in practice. In the future, if the joint is not sawed perfect it will be pared till it meets the criteria.