The actual causes of variation are unknown.  Mr. Darwin favors the opinion of the late Mr. Knight, the great philosopher of horticulture, that variability tinder domestication is somehow connected with excess of food.  He regards the unknown cause as acting chiefly upon the reproductive system of the parents, which system, judging from the effect of confinement or cultivation upon its functions, he concludes to be more susceptible than any other to the action of changed conditions of life.  The tendency to vary certainly appears to be much stronger under domestication than in free Nature.  But we are not sure that the greater variableness of cultivated races is not mainly owing to the far greater opportunities for manifestation and accumulation—­a view seemingly all the more favorable to Mr. Darwin’s theory.  The actual amount of certain changes, such as size or abundance of fruit, size of udder, stands of course in obvious relation to supply of food.  Really, we no more know the reason why the progeny occasionally deviates from the parent than we do why it usually resembles it.  Though the laws and conditions governing variation are known to a certain extent, those governing inheritance are apparently inscrutable.  “Perhaps,” Darwin remarks, “the correct way of viewing the whole subject would be, to look at the inheritance of every character whatever as the rule, and non-inheritance as the anomaly.”  This, from general and obvious considerations, we have long been accustomed to do.  Now, as exceptional instances are expected to be capable of explanation, while ultimate laws are not, it is quite possible that variation may be accounted for, while the great primary law of inheritance remains a mysterious fact.

The common proposition is, that species reproduce their like; this is a sort of general inference, only a degree closer to fact than the statement that genera reproduce their like.  The true proposition, the fact incapable of further analysis, is, that individuals reproduce their like—­that characteristics are inheritable.  So varieties, or deviations, once originated, are perpetuable, like species.  Not so likely to be perpetuated, at the outset; for the new form tends to resemble a grandparent and a long line of similar ancestors, as well as to resemble its immediate progenitors.  Two forces which coincide in the ordinary case, where the offspring resembles its parent, act in different directions when it does not and it is uncertain which will prevail.  If the remoter but very potent ancestral influence predominates, the variation disappears with the life of the individual.  If that of the immediate parent—­feebler no doubt, but closer—­the variety survives in the offspring; whose progeny now has a redoubled tendency to produce its own like; whose progeny again is almost sure to produce its like, since it is much the same whether it takes after its mother or its grandmother.