Improvement is after all the keynote to the breeder's operations. He is aiming at the production of a strain which shall combine the greatest number of desirable properties with the least number of undesirable ones. This good quality he must take from one strain, that from another, and that again from a third, while at the same time avoiding all the poor qualities that these different strains possess. It is evident that the Mendelian conception of characters based upon definite factors which are transmitted on a definite scheme must prove of the greatest service to him. For once these factors have been determined, their distribution is brought under control, and they can be associated together or dissociated at the breeder's will. The chief labour involved is that necessary for the determination of the factors upon which the various characters depend. For it often happens that what appears to be a simple character turns out when analysed to depend upon the simultaneous presence of several distinct factors. Thus the Malay fowl breeds true to the walnut comb, as does also the Leghorn to the single comb, and when pure strains are crossed all the offspring have walnut combs. At first sight it would be not unnatural to regard the difference as dependent upon the presence or absence of a single factor. Yet, as we have already seen, two other types of comb, the pea and the rose, make their appearance in the F_2 generation. Analysis shows that the difference between the walnut {157} and the single is a difference of two factors, and it is not until this has been determined that we can proceed with certainty to transfer the walnut character to a single-combed breed. Moreover, in his process of analysis the breeder must be prepared to encounter the various phenomena that we have described under the headings of interaction of factors, coupling and repulsion, and the recognition of these phenomena will naturally influence his procedure. Or again, his experiments may show him that one of the characters he wants, like the blue of the Andalusian fowl, is dependent upon the heterozygous nature of the individual which exhibits it, and if such is the case he will be wise to refrain from any futile attempt at fixing it. If it is essential it must be built up again in each generation, and he will recognise that the most economical way of doing this is to cross the two pure strains so that all the offspring may possess the desired character. The labour of analysis is often an intricate and tedious business. But once done it is done once for all. As soon as the various factors are determined, upon which the various characters of the individual depend, as soon as the material to be made use of has been properly analysed, the production and fixation of the required combinations becomes a matter of simple detail.