their glassy surfaces. No material gives
such play of light and shade as glass does. One window is never
absolutely like another; so that while a certain uniformity in
their value may be required for breadth of effect in the drawing of
a building, there is plenty of opportunity for incidental variety
in their treatment.

A few practical hints on the rendering of windows may prove serviceable.
Always emphasize the sash. Where there is no recess, as in wooden
buildings, strengthen the inner line of sash, as in Fig. 41. In
masonry buildings the frame and sash can be given their proper
values, the area of wood being treated broadly, without regard
to the individual members. The wood may, however, be left white
if required, as would be the case in Colonial designs. In either
case the dark shadow which the sash casts on the glass should be
suggested, if the scale of the drawing be such as to permit of it.
Do not try to show too much. One is apt to make a fussy effect,
if, for instance, one insists on always shading the soffit of the
masonry opening, especially if the scale of the drawing be small.
Besides, a white soffit is not a false but merely a forced value,
as in strong sunlight the reflected light is considerable. If the
frame be left white, however, the soffit ought to be shaded, otherwise
it will be difficult to keep the values distinct. In respect of
wooden buildings there is no need to always complete the mouldings
of the architrave. Notice in Fig. 41 that, in the window without
the muntins, the mouldings have been carried round the top to give
color, but that in the other they are merely suggested at the corners
so as to avoid confusion. Care should be taken to avoid mechanical
rendering of the muntins. For the glass itself, a uniformly flat
tone is to be avoided. The tones should soften vaguely. It will be
found, too, that it is not advisable to have a strong dark effect
at the top of the window and another at the bottom; one should
predominate.

[Illustration: FIG. 41 C. D. M.]

The student after careful study of Fig. 41 should make from it
enlarged drawings, and afterwards, laying the book aside, proceed
to render them in his own way. When he has done so, let him compare
his work with the originals. This process ought to be repeated
several times, the aim being always for _similarity_, not for
_literalness_ of effect. If he can get equally good results with
another method he need not be disconcerted at the