It would be necessary, in order to prevent the effects of intercrossing, that only a single variety should be turned loose in its new home.

Nevertheless, as our varieties certainly do occasionally revert in some of their characters to ancestral forms, it seems to me not improbable that if we could succeed in naturalising, or were to cultivate, during many generations, the several races, for instance, of the cabbage, in very poor soil--in which case, however, some effect would have to be attributed to the DEFINITE action of the poor soil --that they would, to a large extent, or even wholly, revert to the wild aboriginal stock.

Whether or not the experiment would succeed is not of great importance for our line of argument; for by the experiment itself the conditions of life are changed.