367 suppor~ed,--which lille (borrowing a phrase from hydraulics), might be called " its mean radius." Reasons were then adduced, for believing flint slifFenitrg pieces mi~tht, with saf'ety, be so distributed, as to leave tile mean radius of the izJcluded spaces equal to seventy times the thickness of tile plate ; al:hough, in ,rll)st early slrtiettrres, it would be foulld 1o be lnueh less than what would restllt from such i, calculation.