'The Geo. H. Thomas version specifies ball ended screws working in conical recesses. This arrangement tends to push the gib upwards against the face of the slide locking it firmly in place.' I don't think this is correct. As far as I can see, the gib strip could pivot on the screw point or ball end, but is constrained to follow the path of the screw's movement. Please shoot me down in flames If I've misunderstood, because I've thought about this subject a bit (and posted previously) and am surprised that a definitive solution doesn't seem to have been made widely known. OK tapered gibs are the answer…
GHT was a fan of dowelled gib strips, which seem to me to be a good idea – or at least some form of gib movement control is. The screw thrust is parallel to the dowel's axis, so any pushing of the gib strip in any other direction (eg 'upwards'
would be prevented – but I think that the strip would, in fact, be pushed downwards by pointed screws in conical pockets. (We may in fact agree, but are using 'upwards' and 'downwards' reversed…)
If a ball is interposed twixt screw (flat-ended) and gib strip (not dimpled), the gib should be pushed normal to its surface, so this is a problem for dowelled gibs*, unless the dowel is aligned to be normal to the gib strip's surface too. My ancient Superior surface grinder has its gib-strip adjusting screws arranged so that they bear at right-angles to the gib strips. It's about the only nice bit of design in the whole machine (but that's another story…).
* assuming the dowel fits closely into a hole – a 'vertical' slot would be OK – just constraining 'horizontal' displacements.