Villemin's observations have been repeatedly confirmed and extended; although subjected to the severest criticism and control, their results are so constant that the law of the inoculability of tubercle is almost universally regarded as fixed. Its value as a test is evident from the statement of Cohnheim,[54] who regards as tuberculous only that which produces tuberculosis when transferred to suitable animals. The transfer may be made in various ways. Chauveau and others were successful in producing an intestinal tuberculosis by the introduction of tuberculous material into the intestinal canal of animals, especially the Herbivora. Tappeiner[55] succeeded in producing pulmonary tuberculosis, with or without general tuberculosis, in dogs, by compelling them to breathe air in which were contained minute particles of sputa from tuberculous pulmonary cavities. [Footnote 54: _Die Tuberkulose vom Standpunkte der Infections-Lehre_, 1880, 13.] [Footnote 55: _Virchow's Archiv_, 1878, lxxiv. 393.] The production of a tuberculosis of the iris, as well as of remote organs, by the inoculation of tuberculous material into the anterior chamber of the eye, was an ingenious method devised by Cohnheim and Salomonsen.[56] It permitted the direct observation of the several steps in the process of absorption of the inoculated material and development of the tubercles. [Footnote 56: Cohnheim's _Vorlesungen über Allgemeine Pathologie_, 2te Auflage, 1882, i. 707.] The objections to the various experiments above alluded to are based upon the assumption that the results of the inoculation are not tubercles, but inflammatory products resembling tubercles. It is further advocated that the inoculation of indifferent material, as bits of glass or hairs, as well as other foreign substances, will produce the so-called artificial tuberculosis, especially in rabbits and guinea-pigs. It is admitted that these animals readily become tuberculous when exposed to simple inflammatory irritants, the local action of which frequently results in the production of cheesy material. This termination is now regarded as due to faults in the method of experimentation, the animals not being thoroughly protected from the influence of the virus of tuberculosis. The objection on the ground of structure loses its force in connection with the well known differences in the structure of miliary tubercles in the human body, already mentioned. The tubercles resulting from inoculation often resemble in structure the meningeal tubercles of the brain rather than the type presented by tubercles in lymphatic glands. The development of tubercles in the iris may take place without any permanent inflammatory reaction. The association of evidences of inflammation with the development of the tubercle is therefore unnecessary.