§ 90. Still there remains a difficulty. It may be said that admitting functional change to be the initiator of variation--granting that the physiological units of an organism long subject to new conditions, will tend to become modified in such way as to cause change of structure in offspring; yet there will still be no cause of the supposed heterogeneity among the physiological units of different individuals. There seems validity in the objection, that as all the members of a species whose circumstances have been altered will be affected in the same manner, the results, when they begin to show themselves in descendants, will show themselves in the same manner: not multiform variations will arise, but deviations all in one direction. The reply is simple. The members of a species thus circumstanced will _not_ be similarly affected. In the absence of absolute uniformity among them, the functional changes caused in them will be more or less dissimilar. Just as men of slightly-unlike dispositions behave in quite opposite ways under the same circumstances; or just as men of slightly-unlike constitutions get diverse disorders from the same cause, and are diversely acted on by the same medicine; so, the insensibly-differentiated members of a species whose conditions have been changed, may at once begin to undergo various kinds of functional changes. As we have already seen, small initial contrasts may lead to large terminal contrasts. The intenser cold of the climate into which a species has migrated, may cause in one individual increased consumption of food to balance the greater loss of heat; while in another individual the requirement may be met by a thicker growth of fur. Or, when meeting with the new foods which a new region furnishes, accident may determine one member of the species to begin with one kind and another member with another kind; and hence may arise established habits in these respective members and their descendants. Now when the functional divergences thus set up in sundry families of a species have lasted long enough to affect their constitutions, and to modify somewhat the physiological units thrown off in their reproductive cells, the divergences produced by these in offspring will be of divers kinds. And the original homogeneity of constitution having been thus destroyed, variation may go on with increasing facility. There will result a heterogeneous mixture of modifications of structure caused by modifications of function; and of still more numerous correlated modifications, indirectly so caused. By natural selection of the most divergent forms, the unlikenesses of parents will be rendered more marked, and the limits of variation wider. Until at length the divergences of constitutions and modes of life, become great enough to lead to segregation of the varieties.