But there is also direct evidence that the intercalation took place empirically, i. e. as need arose. Ungnad has shewn this from a comparison of the known leap-years. Best known of all is the letter of Hammurabi to Siniddinam:--“Since the year has a deficiency, let the previous month be entered as Elul II. And instead of bringing the taxes on the 25th Tishritu to Babylon, let them be brought to Babylon on the 25th Elul II”[930]. For the empirical correcting of the position of months the stars are used among the primitive peoples, and so also in Babylonia. A tablet in the British Museum[931] gives the following injunction:--“The constellation _dilgan_ rises heliacally in the month _nisan_. As often as this constellation remains invisible, its month shall be forgotten”. The same injunction is given in regard to other constellations from which months are named. The expression that the month Nisan is to be ‘forgotten’ reminds one of the description of the intercalary month as the ‘lost’ or ‘forgotten’ month among certain tribes of N. American Indians, and of the expression of the Masai. The forgotten month is not the intercalary month in our sense, i. e. not the second of two months that have arisen by doubling; it is the first. This month must be passed over, not counted, forgotten, its name must be transferred to the following month, so that the year may run properly. The establishing of the months by means of phases of the stars is so abundantly demonstrated for primitive peoples in the preceding pages that no words need be wasted in describing the method of its carrying out. It is a method that works perfectly well but is entirely empirical, and where recourse is had to this method we know that the regulation by a definite intercalary cycle does not exist. With a more extended development of the method a still better result can be obtained, and this is the direction that the Babylonians have taken. The regulation runs:--“If on the first day of the month _nisannu_ the constellation of the Pleiades and the moon are together, the year shall be an ordinary one. If on the third day of the month _nisannu_ the constellation of the Pleiades and the moon stand together, the year shall be a full one (i. e. a leap-year)”[932]. The meaning and effect of this rule are explained by Schiaparelli. But this too is an empirical rule, aimed at an empirical, not a cyclical, intercalation. Where an intercalary cycle exists, no such rule is needed.