At the time of Dungi and his successors we have documentary evidence for a number of years with intercalation.[928] At this date Kugler stoutly denies and Weidner supports the existence of an intercalary cycle. Weidner says:--“If we denote Dungi 39 (the 39th year of his reign) by I, the following years are proved by documents to contain intercalary months:--II, V, XI, XIV, XVI, XVIII, XX, XXIII, XXVI, XXIX, XXXII, XXXV, XXXVIII. But between Dungi 43 and 49 there is at least one more leap-year to be added, most probably Dungi 46, i. e. VIII. For the period of 38 years we should then have 14 intercalary months attested. This is therefore an intercalary system that works quite well. A 19-year intercalary cycle however it cannot be, since in that case, corresponding to the former part, the years XXI, XXIV, etc. in the latter would have to be leap-years. _We have therefore to assume a 38-year intercalary cycle, which in perfection far surpasses that of 19 years._ It is the half of the well-known 76-year cycle of Callippus.” The conclusion is unwarrantable from the premises. For the intercalation which takes place just as need arises keeps the months firmly in their place in the solar year, and attains the same result as an intercalary cycle. A period of 76 Indian years will contain just as many months as a Callippean cycle. The only conclusive factor therefore is the periodicity, and this is not proved. Through an accident of tradition the leap-years are known for a period of 38 years, and it is obvious that during these 38 years an empirical intercalation, regularly carried out, kept the lunisolar year in order. The evidence that even under the Hammurabi dynasty no intercalary cycle existed is given by Kugler[929].