the picture in this case would be divided into two distinct parts, one
half cold and the other warm, it was necessary to carry some of the
mellow colours of the great group into the cold part of the picture, and
a part of the cold into the great group; accordingly Titian gave Ariadne
a red scarf, and to one of the Bacchantes a little blue drapery." As
there is no picture more splendid, it is well to weigh and consider
again and again remarks upon the cause of the brilliancy, given by such
an authority as Sir Joshua Reynolds. With regard to his rule, even among
artists, "adhuc sub judice lis est." He combats the common notion of
relief, as belonging only to the infancy of the art, and shows the
advance made by Coreggio and Rembrandt; though the first manner of
Coreggio, as well as of Leonardo da Vinci and Georgione, was dry and
hard. "But these three were among the first who began to correct
themselves in dryness of style, by no longer considering relief as a
principal object. As these two qualities, relief and fulness of effect,
can hardly exist together, it is not very difficult to determine to
which we ought to give the preference." "Those painters who have best
understood the art of producing a good effect, have adopted one
principle that seems perfectly conformable to reason--that a part may be
sacrificed for the good of the whole. Thus, whether the masses consist
of light or shadow, it is necessary that they should be compact, and of
a pleasing shape; to this end some parts may be made darker and some
lighter, and reflections stronger than nature would warrant." He
instances a "Moonlight" by Rubens, now, we believe, in the possession of
Mr Rogers, in which Rubens had given more light and more glowing colours
than we recognize in nature,--"it might easily be mistaken, if he had
not likewise added stars, for a fainter setting sun." We stop not to
enquire if that harmony so praised, might not have been preserved had
the resemblance to nature been closer. Brilliancy is produced. The fact
is, the _practice_ of art is a system of compensation. We cannot exactly