The interior is very mountainous and cut into ravines so deep and precipitous that no large number of people could live there. The whole population was crowded on the strip of land which runs like a low shelf round the greater part of the island, interrupted only in three or four places, as at the Pari, by cliffs directly overhanging the sea. On this strip, less than a hundred and twenty miles long, and varying from the bare cliff, without even a beach, to one or perhaps two miles in extreme width, where the larger streams cut out a few broader valleys, Cook found in 1774 a population that he could hardly trust himself to estimate. Modern writers, without a shadow of reason, have rejected his evidence, but all other evidence confirms it. In 1767 Wallis had been astonished at the numbers of the people, and not without reason, for while he was still warping his ship into Matavai Bay he was surrounded by swarms of war-canoes. "When the great guns began to fire, there were not less than three hundred canoes about the ship, having on board at least two thousand men; many thousands were also upon the shore, and more canoes coming from every quarter."

Already in 1774, when Cook made his second voyage¹, disease and war must have begun to reduce the population from what it had been when Wallis arrived in 1767; yet Cook saw, at Pare Arue, a fleet of one hundred and sixty large double canoes, attended by one hundred and seventy smaller double canoes, preparing to set out against the neighboring island of Eimeo. This fleet, he calculated, could not contain less than seven thousand seven hundred and sixty men, allowing forty to each large canoe and eight to the small ones, and it was the

1. Second Voyage, i, 321

Indigenous Histories