In some still rarer cases fertility continues during several generations. Then, however, a curious phenomenon is exhibited, called by M. Naudin, who discovered it, _Disordered variation._ With the _Linaria communis_ and the _Linaria purpurea_ he produced a hybrid, the descendants of which he was able to follow through seven generations, in each of which several individuals reverted to the characters either of the original male or female. The others neither resembled the primitive types nor the hybrid resulting from their crossing, nor the plants of which they were the immediate offspring, nor was there any resemblance between the plants themselves. Thus the crossing does not produce a race, even in cases where it allows a certain amount of fertility; it only produces _varieties_ incapable of transmitting their individual characters. In order to establish a series of generations presenting a certain amount of uniformity, the hybrid must lose some of its mixed characters, and resume the normal livery of the species, as M. Naudin says; in other words, it must return to one of the parent types. IV. The same facts which we have just noticed among plants, occur also among animals. We must observe in the first place, that the only two species, the crossing of which displays anything approaching to regular fertility, the horse and the ass, merely produce a hybrid almost entirely devoid of fertility. It is more than 2000 years since Herodotus regarded the fertility of mules as a prodigy, and almost 1800 years since Pliny expressed the same opinion. And yet in some works we read that the fertility of the mule is displayed in the present day; that it often propagates in hot countries, especially in Algeria. The true value of these singular assertions will be recognised if we recall the effect which was produced in 1828 upon the whole Mussulman population of Algeria by the announcement that a mule had conceived near Biskra. The astonishment was general; the Arabs gave themselves up to long fasts to conciliate the wrath of heaven, thinking the end of the world had come. Fortunately the mule miscarried; but long afterwards the Arabs still spoke with terror of this event. If this fact were occasionally repeated in Algeria it would never have produced such an impression upon a people so curious about everything connected with the horse. The impression itself proves that the facts are in our days similar to what they were in the time of Herodotus.