destroyed. Almost as bad is the effect when a rainbow cuts off a corner of a picture, for this suggests at first sight an accidental interference with the work.b Of all artists Rubens seemed to know best how to use a rainbow. He adopts three methods. The first and best is to put the bow entirely in the sky; the second to throw it right into the background where part of it is dissolved in the view d; and the third to indicate the bow in one part of the picture, and overshadow it with a strong sunlight thrown in from another part. Any of these forms seems to answer well, but they practically exhaust the possibilities in general design. A section of a rainbow may be shown with one end of it on the ground, because this is observable in nature £; but to cut off the top of the arch as if there were no room for it on the canvas is obviously bad, for the two segments left appear quite unnatural.8 The small rainbows sometimes seen at waterfalls are occasionally introduced into paintings, but rarely with success because they tend to interfere with the general view of the scene. Such views are necessarily near ground, and so a bow must seriously injure the picture unless it be placed at the side, as in Innes's fine work of Niagara Falls (the example of 1884). a In the Rivers of England series. b The Rainbow of Millet, and a similar work of Thoma. c Harvest Landscape, Munich Gallery. d Harvest Landscape, Wallace Collection, London. e Landscape with a Rainbow, Hermitage, Petrograd. f Rubens's Shipwreck of Æneas, Berlin Gallery. 8 A. P. Van de Venne's Soul Fishery, Amsterdam. The use of a rainbow as a track in classical pictures is sometimes effective, though the landscape is largely sacrificed owing to the compulsory great width and bright appearance of the bow, which must indeed practically absorb the attention of the observer. The best known picture of this kind is Schwind's Rainbow, which shows the beautiful form of Iris wrapped in the sheen of the bow, and descending with great speed, the idea being apparently taken from Virgil. To use the top of the rainbow for a walking track is bad, as the mind instinctively repels the invention as opposed to reason.b But if fleeting natural phenomena become disturbing to the observer of a picture, how much more objectionable are the quickly disappearing effects of artificial devices, as the lights from explosions. In a battle scene covering a wide area of ground, a small cloud of smoke here and there is not out of place, because under natural conditions such a cloud lasts for an appreciable time; but no good artist will indicate in his work a flash from a gun, for this would immediately become stagy and unreal to the observer. Nor can fireworks of the ordinary kind be properly represented in a picture. The beauty of these fireworks lies in the appearance out of nothing, as it were, of brilliant showers of coloured lights, and their rapid disappearance, to be replaced by others of different form and character, the movement and changes constituting important elements in their appreciation. But the painter can only indicate a Æneid V., where Juno sends Iris to the Trojan fleet. b Thoma's Progress of the gods to Walhalla. them by fixed points of light which necessarily appear abnormal. Stationary points of light can have no resemblance whatever to fireworks, and if the title of the picture forces the imagination to see in them expiring sparks from a rocket, the impression can only last a moment, and will be succeeded by a revolt in the mind against so glaring an impossibility as a number of permanent sparks. The only painted firework display that does not appear abnormal is a fountain of fire and sparks which may be presumed to last for some time, and therefore would not quickly tire the mind." a See examples by La Touche, notably La Fête de Nuit, Salon, 1914. CHAPTER XIII STILL-LIFE Its comparative difficulty-Its varieties-Its limitations. RIGHT through the degrees of the art of the painter till we reach still-life, the difficulty in producing the art is in proportion to the general beauty therein, but in the case of still-life the object is much less readily gained than in simple landscape which is on a higher level in painting. The causes of this apparent anomaly appear to be as follows:-Firstly in miniature painting one does not expect such close resemblance to nature as in still-life which usually represents things in their natural size: secondly, in still-life the relative position of the parts can never be such as to appear novel, whereas in landscape their position is always more or less unexpected: thirdly, in still-life the colours are practically fixed, for the painter cannot depart from the limited variety of tints commonly connected with the objects indicated, while in landscape the colouring may vary almost indefinitely from sun effects without appearing to depart from nature. The beauty in still-life paintings may arise from several causes, namely, the pleasure experienced |