Now, although Fornander's chronology has no direct bearing upon the date of the Melanesian arrival if, as I have shown, the Polynesians had no settlement in the group, the method of calculating dates should be the same for both races. Our only guide for events that happened in Polynesia before Tasman's voyage, 1642, is in the natives' genealogies, calculating by generations. They contain two obvious tendencies to error. It was very rare for a man of consequence to carry the same name throughout his career. Adoption, any notable exploit, or succession to a title were constant excuses for such changes, and it is quite possible that in the older genealogies the same hero is recorded twice under different names. Moreover, it is by no means certain that the names were not those of the reigning chiefs, and seeing that the succession often went to the next brother when the son was not of an age to wield the power, it is highly doubtful whether every name represented a generation. I know one genealogy where, in the portion relating to historical times, one of the recorded names was younger brother to the chief who precedes him.[15] This may account for the great diversity of readings found in the same genealogy, one version being shorter than another. On the other hand, there is the tendency to omit the names of remote personages whose short reign or insignificant character have failed to stamp themselves on the memory of posterity. There is thus a double tendency to error--on the one side to multiplication of generations, and on the other to curtailment by omissions. But even supposing that Fornander's genealogies are correct, it is difficult to see how he could arrive at an approximate date without showing more discrimination in fixing the length of a generation. All his dates are calculated upon a generation of _thirty years_, because that is the average length generally assigned in Europe. But Polynesia is not Europe, and generations in Polynesia, where men marry much earlier, are less than thirty years, as he might have discovered by taking the average in historical times. This I have done both in Tonga and Fiji, with the result that the generations in both races average from twenty-five to twenty-seven years. The Tui Tonga family is a very fair guide, because the office went invariably from father to son, and the holder was so sacred that he was never cut off by a violent death. The generations of this family since 1643 average twenty-seven years, while those of the temporal sovereign, the Tui Kanakubola who were often the victims of rebellion, average only twenty years apiece. The history of Hawaii was so bloodstained, that it is unlikely that Hawaiian generations averaged more than twenty-five. Five years in a generation makes a vast difference, for the date given by Fornander for the Polynesians' arrival in the Pacific is set forward from the fifth to the seventh century, and for their arrival in Hawaii from the eleventh to the thirteenth.