As an illustration of the second mode of vengeance, we may perhaps cite the Serbians of recent times who adopted a restricted form of vendetta and who often allowed murder to remain unpunished.[14] The restricted system seems to have existed amongst the Araucanians[15] of South America, and amongst the Jivaros Indians,[16] but only when the identity of the murderer could be established. In this latter case we find the alternative operation of a more civilised with a more barbarous form of vengeance. But we must not assume that these forms coexist as alternatives everywhere. A French authority holds[17] that the essential motive of collective punishment was the production or identification of the murderer. ‘So long,’ he says,[17] ‘as the murderer is unknown, so long is the responsibility collective and diffused.’ We cannot accept this statement as an explanation of the origin of unrestricted vendetta. We admit that collective penalties of a minor kind would form a strong inducement for the discovery of the criminal. It was for this reason perhaps that an Anglo-Saxon law[18] levied a fine on the whole ‘hundred’ if the murderer was not produced. But it is one thing to bring pressure to bear on a district, whether by a fine, as in this case, or by an oath, as in the instance mentioned in Deuteronomy[19]; it is quite another thing to destroy a whole town or village if the murderer was unknown. We shall see[20] that the Homeric Achaeans often waited long for vengeance, and often allowed the homicide to go unpunished, rather than visit with unjust punishment the innocent relatives of the slayer. In this system there is no trace of collectivity. The relatives have not even to pay a sum of money. The flight of the slayer is not indeed accepted as a substitute for the normal penalty, which is death, but it postpones indefinitely, if not for ever, a vengeance which the slayer alone can suffer.