253

question left unasked, any possible way of getting at the information; but most of what the old man knew he had forgotten, and what he dimly remembered he was incapable of explaining. I made one more futile effort, then bade him good-bye and turned away. It was late afternoon on a day of unusual calm, everything in the lonely spot was perfectly still, the sea lay below like a sheet of glass, the sun as, a globe of fire was nearing the horizon, while close at hand lay the old man gradually sinking, and carrying in his tired brain the last remnants of a once-prized knowledge. In a fortnight he was dead.
   No detailed systematic study of the tablets has as yet been possible from the point of view of the Expedition, but it seems at present probable that the system was one of memory, and that the signs were simply aids to recollection, or for keeping count like the beads of a rosary. To what extent the figures were used at will, or how far each was associated with a definite idea it is impossible to say. Possibly there was no unvarying method; certain ones may conveniently have been kept for an ever-recurrent factor, as the host in the tau, and in well-known documents, such as "he timo te ako-ako," they would doubtless be reproduced in orthodox succession. But in the tablets which we possess the same figures are continually repeated, and the fact that equivalents were always having to be found for new names, as in that of the fish-man, or ika, suggest that they may have been largely selected by the expert haphazard from a known number. As Tomenika said, "the words were new, but the letters were old," or to quote Kapiera to the same effect, they were "the same picture, but other words." It will be noted how few men are reported to have known each variety of rongo-rongo, and that while Ngaara looked at the tablets of the boys, apparently to see if they were properly cut, it was in the recitation only of the older men that accuracy was insisted on. The names which Bishop Jaussen's informant assigned to some five hundred figures may or may not be accurate, but whether the native or anyone else could have stated what the signs conveyed is another matter. It is easy to give the term for a knot in a pocket-handkerchief, but no one save the owner can say whether he wishes to remember to pay his life insurance or the date of a tea-party.
   In trying to enter into the state of society and of mind which

253