her
side; therefore the building near it should be of the average size; and
thus the villas of the Lago di Como, being among hills from 6000 to 8000
feet high, are well proportioned, being neither colossal nor diminutive:
but a mountain 3000 feet high always looks higher than it really
is;[48] therefore the buildings near it should be smaller than the
average. And this is what is meant by the proportion of objects; namely,
rendering them of such relative size as shall produce the greatest
possible impression of those attributes which are most desirable in
both. It is not the true, but the desirable impression which is to be
conveyed; and it must not be in one, but in both: the building must not
be overwhelmed by the mass of the mountain, nor the precipice mocked by
the elevation of the cottage. (Proportion of color is a question of
quite a different nature, dependent merely on admixture and
combination).

[Footnote 48: This position, as well as the two preceding, is important,
and in need of confirmation. It has often been observed, that, when the
eye is altogether unpracticed in estimating elevation, it believes every
point to be lower than it really is; but this does not militate against
the proposition, for it is also well known, that the higher the point,
the greater the deception. But when the eye is thoroughly practiced in
mountain measurement, although the judgment, arguing from technical
knowledge, gives a true result, the impression on the feelings is always
at variance with it, except in hills of the middle height. We are
perpetually astonished, in our own country, by the sublime impression
left by such hills as Skiddaw, or Cader Idris, or Ben Venue; perpetually
vexed, in Switzerland, by finding that, setting aside circumstances of
form and color, the abstract impression of elevation is (except in some
moments of peculiar effect, worth a king's ransom) inferior to the
truth. We were standing the other day on the slope of the Brevent, above
the Prieure of Chamouni, with a companion, well practiced in climbing
Highland hills, but a stranger among the Alps. Pointing out a rock above
the Glacier des Bossons, we requested an opinion of its height. "I
should think," was the reply, "I could climb it in two steps; but I am
too well used to hills to be taken in in that way; it is at least 40
feet." The real height was 470 feet. This deception is attributable to
several causes (independently of the clearness of the medi