The growth of populations and the spread of urbanization, resulting in new agricultural structures, have entailed a concentration of livestock production and recourse to new techniques. Of some importance among these techniques is the enteric or parenteral administration of substances in very low doses. These substances include anabolic agents, some of which, like many natural feeds, exhibit hormonal activity. They may be divided into two classes: --those of the DES type, synthetic compounds non-existent in the natural state, --natural agents, which are normally distributed throughout the animal and human organism, and hence in food of animal origin---milk, meat, eggs. The compounds belonging to the second class may also be synthesized and the main toxicological consideration is that they then have to meet clear-cut standards of identity and purity. A compound belonging to the first class, diethylstilbestrol (DES), administered to rats in doses as small as 60 mug/kg of feed or even smaller, causes in general lower growth rates as well as alterations in the genital system and reproductive functions. In long-term experiments (12 months) using rats and mice and applying so-called toxicity "de relais" tests, developed and described by the authors, it also appeared that meat from calves in which DES pellets were implanted under normal rearing conditions, inhibits growth and reproduction in mice and rats fed a diet containing 20% of this meat. Studies in which the livers from treated calves constituted 6% of the diet of these two rodent species also led to the conclusion that fertility was impaired in the second reproduction test. The authors also recall cases of vaginal cancer observed in young girls whose mothers had been treated with DES during pregnancy. Compounds belonging to the second class (estradiol-progesterone and estradiol-testosterone) gave no evidence of harmful effects upon rats when mixed with their rations during short and medium-term trials. Similar results were obtained from studies with rats and mice fed a diet containing 20% veal or 6% liver of calves in which pellets of these natural hormones had been implanted. In these studies, which lasted 23--24 months with rats and 347 days with mice, the toxicity "de relais" test was applied under similar conditions to those undertaken using meat and livers of implanted calves. The report also refers to trenbolone and to "zéralénone" but no toxicity "de relais" tests were carried out with the meat and livers of animals treated with these substances. The authors reached the following conclusions: DES should not be used as an anabolic agent in livestock production. Natural hormones prepared through synthesis or otherwise could be used as anabolic agents in livestock production since no risk to the consumer has been witnessed when they were administered at doses similar to those used in the author's experiments. However, it is recommended that they should be applied under veterinary supervision...