C. The date of the migrations. This is either relative or absolute: _relative_ when we ascertain whether one division of the Oceanic populations migrated before or after another; _absolute_ when we fix the chronological date of a migration. As a general rule the latter is unattainable--Iceland and a few other areas, peopled within the historical period, forming the exceptions. Respecting, then, the absolute date of the Polynesian migration, there is no reason why it should not be known _in particular islands_; for instance, in the Dangerous Archipelago, where only a small proportion of the clusters is peopled even at present, any given island may receive a population so late as this, the eleventh hour of the extension of the human species; yet it is evident that the knowledge of such a migration would throw but little light upon the broader question of the date of the Polynesian population _en masse_. Of this it may safely be said, that no important group has received its first occupants within the Polynesian _historical_ period. This, however, is but a short one. Will the longer range of the _traditionary_ period supply any such information? I think not. Nevertheless it must be added, that in Nukahiva pedigrees run up to the eighty-fifth generation, the founders of them being connected with the first occupancy of the island. Even, however, if we admit so long a genealogy as an historical fact, it only gives the date for one particular island. Proper _ethnological_ reasoning is, from its very nature, inapplicable to the investigation of a definite epoch in chronology; since it only begins where the evidence of testimony ends. Furthermore, it is only approximate, since it simply calculates, by means of an imperfect induction, the _minimum_ period required to account for differences; and the _maximum_ period that will account for resemblances; _e.g._ for the Polynesians to differ as they do from the Micronesian, a certain time must have elapsed; and for them to differ no more than they do, that time must have a limit. Applied to the _relative_ date of the Oceanic migrations, ethnological reasoning gives for even the most recent of them, a geological rather than an historical epoch; and this is as much as it is safe to say. Its other probable conclusions are more definite. 1. Occupancy had begun in Australia before migration across Torres Strait had commenced in New Guinea. 2. Occupancy had begun in New Guinea before Polynesian migration had commenced in Protonesia. The first of these facts we infer from the physical differences between the Australian and the Papuan, taken with the fact that it is scarcely likely that the Papuans of Torres Straits would have failed in extending themselves to Australia had that island been unoccupied.