248											22
                                                                      
                                                                      	is free and unconnected with other parts of the archi-
                                                                      tecture, it is well to give it some degree of curvature;
                                                                      allowing in its angle and in the depth of the abacus for
                                                                      any dimutive of strength caused by the curvature.  Tak-
                                                                      ing therefore the typical square capital 1 and curving
                                                                      its slo[p]pe in the directions of the dotter lines, we have
                                                                      2nd and 3rd and these will, entirely differ in expressions
                                                                      according to the placing of the point of greatest curva-
                                                                      ture:  Thus taking the single line 2 which is a line of
                                                                      delicate curvature departing at first as little as pos-
                                                                      sible from the right line;  and increasing its rate of
                                                                      curvature to its other extremity and placing this line
                                                                      between the points a b in four different directions;  we
                                                                      have the capitals 2a and b, 3[s]a and b.  Now the choice
                                                                      among these forms is almost purely aesthetic and de-
                                                                      pends upon an imagined relation between the lines a b and
                                                                      the uprights of the shaft:  Now the natural lines which
                                                                      are most frequently seen in this relation are the boughs
                                                                      and stems of trees, and it is s[a]o general fact that
                                                                      in trees whose boughs are concave to their stems[,] l
                                                                      the greatest curvature is next the stem but in trees
                                                                      whose boughs are convex ao stems, t2 the greatest
                                                                      curvature is furthest from the stem.  Hence these two
                                                                      forms are commonly b[l]est where the shaft is slender and
                                                                      capital wide;  but when

[Version 0.05: May 2008]