As this unhinging of Claude’s faculties increased, he drifted into a sort of superstition, into a devout belief in certain processes and methods.  He banished oil from his colours, and spoke of it as of a personal enemy.  On the other hand, he held that turpentine produced a solid unpolished surface, and he had some secrets of his own which he hid from everybody; solutions of amber, liquefied copal, and other resinous compounds that made colours dry quickly, and prevented them from cracking.  But he experienced some terrible worries, as the absorbent nature of the canvas at once sucked in the little oil contained in the paint.  Then the question of brushes had always worried him greatly; he insisted on having them with special handles; and objecting to sable, he used nothing but oven-dried badger hair.  More important, however, than everything else was the question of palette-knives, which, like Courbet, he used for his backgrounds.  He had quite a collection of them, some long and flexible, others broad and squat, and one which was triangular like a glazier’s, and which had been expressly made for him.  It was the real Delacroix knife.  Besides, he never made use of the scraper or razor, which he considered beneath an artist’s dignity.  But, on the other hand, he indulged in all sorts of mysterious practices in applying his colours, concocted recipes and changed them every month, and suddenly fancied that he had bit on the right system of painting, when, after repudiating oil and its flow, he began to lay on successive touches until he arrived at the exact tone he required.  One of his fads for a long while was to paint from right to left; for, without confessing as much, he felt sure that it brought him luck.  But the terrible affair which unhinged him once more was an all-invading theory respecting the complementary colours.  Gagniere had been the first to speak to him on the subject, being himself equally inclined to technical speculation.  After which Claude, impelled by the exuberance of his passion, took to exaggerating the scientific principles whereby, from the three primitive colours, yellow, red, and blue, one derives the three secondary ones, orange, green, and violet, and, further, a whole series of complementary and similar hues, whose composites are obtained mathematically from one another.  Thus science entered into painting, there was a method for logical observation already.  One only had to take the predominating hue of a picture, and note the complementary or similar colours, to establish experimentally what variations would occur; for instance, red would turn yellowish if it were near blue, and a whole landscape would change in tint by the refractions and the very decomposition of light, according to the clouds passing over it.  Claude then accurately came to this conclusion:  That objects have no real fixed colour; that they assume various hues according to ambient circumstances; but the misfortune was that when he took to direct observation, with his brain