The Kayan use a somewhat different method. The weather-prophet lets in a beam of light through a hole in the roof of his chamber in the long-house, and measures the distance of the patch of light from the point vertically below the hole. Thus they obtain a measurement similar to that given by the shadow on a sun-dial[1065]. Still more elaborate is the method used by some of the Klementan by which time is determined from the position of a star. A tall bamboo vessel is filled with water and then inclined until it points directly towards a certain star. It is set upright again, and the level of the water left in the vessel is measured. In order to determine the seed-time the vessel is provided with an empirically given mark at a certain height, and when the level of the water coincides with the mark after the inclining of the vessel towards the star, it is the time for sowing[1066]. The writers omit to say that the observation must take place at a certain time of day, e. g. morning or evening twilight. Then it becomes possible to determine the season by the height of the star above the horizon. All this is neither primitive nor native. In Bali and Java the Brahmin and Islamite priests observed the sun-dial, and from there the practice came to Borneo. Where the idea of using a vessel of water for measurement originated I am unable to determine, but it is much too refined to be a primitive invention. The only genuinely primitive method is the observation of the annual course of the sun and the solstices by the aid of certain landmarks on the horizon. This method is found in all parts of the world, but only among certain peoples. It has never attained real importance for the regulation of the calendar: the development of the calendar to greater accuracy proceeds by the indirect way of the lunisolar time-reckoning. By way of appendix a few notices of the aids used in calculating may be collected. They are almost always quite simple--knots in a string, the tally, or the joints of the body. The use of the tally in counting the years has already been dealt with above[1067]; this use is certainly later, each stick attaining so to speak an individual life. It is otherwise with the counting of the days, where the question usually is to determine the number of days which will elapse before an assembly or some other undertaking previously agreed upon, so that all may arrive together. The same reckoning may also occasionally serve a second purpose.