Because epidemiological research into the causes of cancer takes so long and may leave crucial questions unanswered, the more direct method of laboratory experimentation is undertaken when possible - as it generally is before the introduction of a newly synthesized chemical into foods or drugs. The two types of techniques often go hand in hand. Some compounds such as soot and certain dyes are known from the findings of epidemiologists to be dangerous. Laboratory tests can identify the components of such compounds that are to blame - benzopyrene in soot, beta-naphthylamine in dyes - and such tests can also provide indisputable proof that a suspected substance or activity does cause cancer.
When materials suspected as carcinogen are being tested, scientists cannot risk a human life; laboratory experiments must be done only on animals. During the 1960s, as epidemiological evidence on the risks of smoking began to pour in, generations of mice at the Imperial Cancer Research Fund in London panted out their lives in smoke-filled enclosures and showed evidence of lung cancer. At a Veterans Administration laboratory in New Jersey in 1970, cigarette smoke was pumped directly into dog's lungs, through openings cut in their throats and generated cancerous growths. And in a bizarre experiment of the early 1980s, baboons at the Southwest Research Foundation in San Antonio, Texas, were taught to smoke cigarettes themselves in the continuing quest for hard experimental evidence.
All such trials somewhat resemble a prospective study, in which a scientist selects a group of subjects and follows their course of health or illness. But there is one critical difference. In the laboratory - but not in everyday life - the scientist can be sure that his subjects differ only in the amounts of carcinogen they are exposed to. What they eat, how they exercise or sleep, the very air they breathe - all can be rigidly controlled. Hereditary differences between animals can be reduced or eliminated; among laboratory mice for example. Scientific breeding has produced strains in which all the individuals are genetically identical. Thus, if a scientist exposes one group of mice to a suspected carcinogen, leaving a second group unexposed, and the first group develops cancer while the second does not, he has proved that the substance causes cancer - at least in mice. To move beyond this test, he might repeat it with other animal subjects - rabbits perhaps, or dogs or monkeys maybe. If a substance induces cancer in several different kinds of mammals, it probably is a carcinogen for most mammals, including human beings.
