On the other hand, in the earlier phases of the disease, where its presence has been recognized solely by the aid of the tuberculin test, before there are any recognizable physical symptoms in any part of the animal, the milk is generally unaffected. Between these extremes, however, is found a large proportion of cases, concerning which so definite data are not available. The results of investigators on this point are conflicting and further information is much desired. Some have asserted so long as the udder itself shows no lesions that no tubercle bacilli would be present,[86] but the findings of a considerable number of investigators[87] indicate that even when the udder is apparently not diseased the milk may contain the specific organism as revealed by inoculation experiments upon animals. In some cases, however, it has been demonstrated by post-mortem examination that discoverable udder lesions existed that were not recognizable before autopsy was made. In the experimental evidence collected, a varying percentage of reacting animals were found that gave positive results; and this number was generally sufficient to indicate that the danger of using milk from reacting animals was considerable, even though apparently no disease could be found in the udder. The infectiousness of milk can also be proven by the frequent contraction of the disease in other animals, such as calves and pigs which may be fed on the skim milk. The very rapid increase of the disease among the swine of Germany and Denmark,[88] and the frequently reported cases of intestinal infection of young stock also attest the presence of the organism in milk. The tubercle bacillus is so markedly parasitic in its habits, that, under ordinary conditions, it is incapable of growing at normal air temperatures. There is, therefore, no danger of the germ developing in milk after it is drawn from the animal, unless the same is kept at practically blood heat. Even though the milk of some reacting animals may not contain the dangerous organism at the time of making the test, it is quite impossible to foretell how long it will remain free. As the disease becomes more generalized, or if tuberculous lesions should develop in the udder, the milk may pass from a healthy to an infectious state. This fact makes it advisable to exclude from milk supplies intended for human use, all milk of animals that respond to the tuberculin test; or at least to treat it in a manner so as to render it safe. Whether it is necessary to do this or not if the milk is made into butter or cheese is a somewhat different question. Exclusion or treatment is rendered more imperative in milk supplies, because the danger is greater with children with whom milk is often a prominent constituent of their diet, and also for the reason that the child is more susceptible to intestinal infection than the adult.