How far it really falls short of the truth was first pointed out by Herbert Spencer. "In all times and places, among savage and civilized," he says, "victory is followed by pillage. Whatever portable things of worth the conquerors find, they take.... The taking of women is manifestly but a part of this process of spoiling the vanquished. Women are prized as wives, as concubines, as drudges; and, the men having been killed, the women are carried off along with the other moveables." Thus "women-stealing" is an "incident of successful war." But a woman so taken has a double value. "Beyond her intrinsic value she has an extrinsic value. Like a native wife, she serves as a slave: but unlike a native wife, she serves also as a trophy." A warrior possessing such a token of prowess gains social distinction. "In a tribe not habitually at war, or not habitually successful in war, no decided effect is likely to be produced on the marriage customs." But in warlike and successful tribes an "increasing ambition to get foreign wives" will arise. Among savages, proofs of courage are often required as qualifications for marriage. Hence it is not surprising that the abduction of a foreign woman should be accepted as the best proof of all. "What more natural than that where many warriors of the tribe are distinguished by stolen wives, the stealing of a wife should become the required proof of fitness to have one? Hence would follow a peremptory law of exogamy." Spencer's interpretation, therefore, agrees with that of McLennan in finding the origin of exogamy in wife-capture and in implying that usage grows into law. But it does not, "like his, assume either that this usage originated in a primordial instinct, or that it resulted from a scarcity of women caused by infanticide.[356] Moreover, unlike Mr. McLennan's, the explanation so reached is consistent with the fact that exogamy and endogamy in many cases co-exist; and with the fact that exogamy often co-exists with polygyny;" nor does it "involve us in the difficulty raised by supposing a peremptory law of exogamy to be obeyed throughout a cluster of tribes." For if exogamy would be likely to arise in tribes usually successful in war, peaceful tribes and those usually worsted in war, though living side by side with the successful and warlike, would be naturally led to adopt the rule of endogamy. Furthermore, among tribes not differing much from one another in strength, endogamy and exogamy may coexist. "Stealing of wives will not be reprobated, because the tribes robbed are not too strong to be defied; and it will not be insisted on, because the men who have stolen wives will not be numerous enough to determine the average opinion." Spencer also maintains that the symbol of rape in the marriage ceremony does not necessarily imply the previous existence either of foreign wife-stealing or of exogamy, assigning three other reasons which singly or together may account for it.