A scheme that has sometimes been used in the place of trap-nests, is a system of small compartments, in each of which one hen is kept. Such a scheme does not seem feasible on a large scale, but for breeders wishing to keep the records of a small number of hens, it is all right. Because of its cost, this system is wholly out of the question, except for a man following breeding as a hobby and who cannot devote himself during the day to the care of trap-nests. Having determined the best layers, it remains to breed from these and from their descendants. The tests of pullets hatched from hens are better signs of the hen's value as a breeder than is her own record. It has been surmised that a hen which lays heavily will not lay eggs containing vigorous germs. So far as the writer's experience has gone, the laying of infertile eggs is a family or individual trait not particularly related to the number of eggs laid. When we have bred from the best layers and have raised our average egg yielder to a higher level, the question arises as to whether the strain will permanently maintain the high yield or drop back to the former rate of production. Theory says that it will not drop back. As a matter of fact it will not do so, for the heavier production will be more trying on the hen's constitution, and naturally selection will gradually cause the egg record to dwindle. Hence the necessity of continued selection or the infusion of new blood from other selected strains. Whatever may be the change desired in a strain of chickens, specimens showing the trait to be selected should be used as breeders. Those characteristics readily visible to the eye have long been the subjects of the breeder's efforts. But traits not directly visible can likewise be changed by breeding. The number of eggs, size and color of eggs, rapid growth, ready fattening powers, quality of meat and general characteristics, are all matters of inheritance, and if proper means are taken to select the desirable individuals all such characteristics can be changed at the will of the breeder. It is a fact, however, often overlooked, that the more traits for which one selects, the slower will be progress. For illustration: If in breeding for egg production, one-half the good layers are discarded for lack of fancy points, the progress will be just half as rapid. A discussion of the work in breeding for egg production at the Maine Experiment Station is taken up in the next chapter. CHAPTER XV EXPERIMENT STATION WORK Our entire scheme of agricultural education and experimentation is new. The poultry work at experiment stations is very new. Ten years will about cover everything worthy of a permanent record in the poultry experiment station files. Stations Leading in Poultry Work.