When Pellegrini partially abandoned caricature and took up portraiture he attempted to become a master of painting too soon, and, inspired by Whistler's facility, imagined that it would be easy to overcome very quickly the difficulties of a lifetime. Occasionally, of course, he succeeded legitimately, as in the case of "Gillie"[3] Farquhar; but, generally speaking, if Pellegrini had a sitter who was an admirable subject for caricature, he was unconsciously liable to put what he saw into his portrait. His successes were great; he was undoubtedly--when he had a "sympathetic" subject, a genius in caricature. That pleasure, or sympathy, is one of the main elements in the success of a caricaturist. Just as a subject may offer great temperamental difficulties, so it frequently happens that--for some inexplicable reason--he will at once afford an opening which a practised caricaturist will know immediately how to turn to account. It is this element of chance which lends a charming uncertainty to the caricaturist's art; and it is this element also which explains in many cases the strange success or failure of an impression, the apparent fluctuations of an artist's talent in preserving a likeness or translating a personality into terms of comedy. Thus it often happened that I was fortunate in my own choice of a man, and thus, on the other hand, that when I was sent off in a hurry to seize the peculiarities of a man, I found he required a great deal of study, and so was obliged to leave out the caricature and put as many characteristics in as I could. The "characteristic portrait," although without the same qualities as the caricature, is sometimes more successful with one type of man. Nature is followed more accurately, the humour is there, if there is humour in the subject, and the work is naturally more artistic in touch and finish, and probably a better drawing in consequence. The caricature done from memory is wider in scope; one is not distracted from the general impression by the various little fascinations of form one finds in closer study. In fact, I consider that in order that the cartoon should have a perfect result, it must be drawn firstly from memory. Of course, little details and characteristics can be memorised by a thumb-nail sketch, or notes upon one's shirt cuff, and for this reason I usually watch my subject all the time. I make notes, keeping him under observation and making the note at the same time. The sketch made in these circumstances is frequently useless in consequence; but it seems to impress upon my brain the special trait I have noticed.