"I have no hesitation in saying that, taking full account of its imperfection, tuberculin is the most valuable means of diagnosis in tuberculosis that we possess.... I have most implicit faith in it, when it is used on animals standing in their own premises and undisturbed. It is not reliable when used on animals in a market or slaughter-house. A considerable number of errors at first were found when I examined animals in slaughter-houses after they had been conveyed there by rail, etc. Since that, using it on animals in their own premises, I have found that it is practically infallible. I have notes of one particular case, where twenty-five animals in one dairy were tested, and afterwards all were killed. There was only one animal which did not react, and it was the only animal not found to be tuberculous when killed." Two instances of the validity of this test will suffice. In 1899, it was applied to 270 cows on some farms in Lancashire. Of these cows, 180 reacted to the test, 85 did not react, and 5 were doubtful. Tuberculous disease was actually found, when they were killed, in 175 out of the 180 = 97.2 per cent. (_Lancet_, 5th August 1899.) In 1901, Arloing and Courmont published a critical account of the whole subject, and gave the following facts. In 80 calves, which on examination after death were found not tuberculous, the test was negative: in 70 older cattle, which were tuberculous, the test was positive in every case but one, though the dilution of the serum was 1 in 10.[20] It would be easy to add instances of the value of this test, for it is practised far and wide over the world. [20] For references to this paper, and to evidence put forward against the validity of the test, and for criticism of such evidence, see Gould's _Year-Book of Medicine and Surgery_, 1902 (Philadelphia, W. B. Saunders & Company). 7. More recently, the discovery of the "opsonic index," and its use by Sir Almroth Wright and others, has given a great advance to the observation and treatment of cases of tuberculosis. The administration of the "new tuberculin" is now timed and measured with an accuracy which was absolutely impossible a few years ago. It is a far cry, from the present method of counting how many tubercle-bacilli are taken up by a single blood-cell, back to Villemin's rabbits. Every inch of the way, from 1881 onward, the pathological study of every form of tuberculosis, medical or surgical, human or bovine, has been dependent on bacteriology; that is to say, on experiments on animals. IV DIPHTHERIA