As long as the determination of time is adjusted by the phases of Nature which immediately become obvious to everyone, anybody can judge of them, and should different people judge differently there is no standard by which the dispute can be settled, because the natural phases run into one another or are at least not sharply defined. The accuracy in determination demanded by time-reckoning proper is therefore lacking. Accuracy becomes possible as a result of the observation of the risings of stars, and this observation begins even at the primitive stage, but it is not a matter that concerns everyone. It requires a refined power of observation and a clear knowledge of the stars, so that the heavens can be known. This is especially the case with the commonest observations, those of the morning rising and evening setting. The observer must be able to judge, by the position of the other stars, when the star in question may be expected to twinkle for a moment in the twilight before it vanishes. The accuracy of the time-determination from the stars depends therefore upon the keenness of the observation. In this the individual differences of men soon come into play, along with a regular science which introduces the learner to the knowledge of the stars and its uses. Thus Stanbridge reports of the natives of Victoria that all tribes have traditions about the stars, but certain families have the reputation of having the most accurate knowledge; one family of the Boorung tribe prides itself upon possessing a wider knowledge of the stars than any other[1160]. An account has been given above[1161] according to which an old chief instructed the young people of the tribe in the knowledge of the stars and the occupations which these announce. Of the Torres Straits tribes Rivers says:--When the rising of a star is expected, it is the duty of the old men to watch; they rise when the birds begin to call and watch until daybreak. As in the case of _kek_ (Achernar, the most important star), so also probably in the case of other important stars and constellations the appearance of certain other stars is a sign that the star expected will soon appear. For _kek_ the stars in question are two named _keakentonar_; when they appear on the horizon at dawn, it is known that in a few days _kek_ will shew himself, and the observation becomes especially keen. The setting of a star is observed in the same way[1162].