visible result, the last two classes of the ratio are indistinguishable, and our F_2 generation comes to consist of three classes in the ratio 9 : 3 : 4, instead of four classes in the ratio 9 : 3 : 3 : 1. +-------+-------+-------+-------+ |GCB....|GCB....|GCB....|GCB....| |GCB....|GcB....|gCB....|gcB....| |.......|.......|.......|.......| |.Agouti|.Agouti|.Agouti|.Agouti| +-------+-------+-------+-------+ |GcB....|GcB |GcB....|GcB | |GCB....|GcB |gCB....|gcB | |.......| |.......| | |.Agouti| Albino|.Agouti| Albino| +-------+-------+-------+-------+ |gCB....|gCB....|gCB####|gCB####| |GCB....|GcB....|gCB####|gcB####| |.......|.......|#######|#######| |.Agouti|.Agouti|##BLACK|##BLACK| +-------+-------+-------+-------+ |gcB....|gcB |gcB####|gcB | |GCB....|GcB |gCB####|gcB | |.......| |#######| | |.Agouti| Albino|##BLACK| Albino| +-------+-------+-------+-------+ FIG. 8. Diagram to illustrate the nature of the F_2 generation which may arise from the mating of agouti with albino in mice or rabbits. This explanation was further tested by experiments with the albinos. In an F_2 family of this nature there ought to be three kinds, viz. albinos homozygous for G (GGccBB), albinos heterozygous for G (GgccBB), and albinos without G (ggccBB). These albinos are, as it were, like photographic plates exposed but undeveloped. {53} Their potentialities may be quite different, although they all look alike, but this can only be tested by treating them with a colour developer. In the case of the mice and rabbits the potentiality for which we wish to test is the presence or absence of the factor G, and in order to develop the colour we must introduce the factor C. Our developer, therefore, must contain C but not G. In other words, it must be a homozygous black mouse or rabbit, ggCCBB. Since such an animal is pure for C it must, when mated with any of the albinos, produce only coloured offspring. And since it does not contain G the appearance of agoutis among its offspring must be attributed to the presence of G in the albino. Tested in this way the F_2 albinos were proved, as was expected, to be of three kinds: (1) those which gave only agouti, _i.e._ which were homozygous for G; (2) those which gave agoutis and blacks in approximately equal numbers, _i.e._ which were heterozygous for G; and (3) those which gave only blacks, and therefore did not contain G.