This is an artificial system which was probably created with a leaning towards the Arabic months. In Edo too the familiar names of months are borrowed from the ceremonies that take place at different times[779]. Madagascar has a comparatively highly developed civilisation in which various influences cross. The Merina have the Arabic months. The history of the native calendar is said to be very complicated: Grandidier in a detailed discussion seeks to prove that the Malgassian year, which is commonly held to be a lunar year, is a solar or lunisolar one, and on the strength of certain resemblances in the names of the months derives the calendar from S. India. I give the principal data. Grandidier says that one reason for believing that the Malgassian calendar is a solar one is the fact that it is in reality agricultural. In 1638 Cauche says that the Malgassi divide their year into 4 seasons and 12 lunar months, with some intercalary days. The year is for them the time which elapses between two phases of the vegetation; for greater convenience they divide it into twelve lunar months, without caring much about the number of days composing these months, as is rightly said of the Antandroy by Vacher[780], who gives the following list, which is almost identical with that compiled by Grandidier himself in the south-east, at Iavibola, in 1866. The months have names and epithets: the latter are explained. 1, millet is cut; 2, winter begins; 3, the beans flower; 4, the tamarinds of the north are ripe; 5, the leaves fall; 6, tamarinds and beans are ripe; 7, the _Cythere_-tree flowers; 8, the bulls seek the shade of the _sakoa_; 9, the guinea-fowls sleep; 10, the rain rots the ropes (with which the calves are fastened); 11, the gourds flower; 12, the grains of the _fano_ are ripe. Rowlands[781] had already remarked that the Betsileo months depend more upon the time of the sowing and reaping of the rice and upon the flowering of certain plants than upon the phases of the moon, and that the agreement with the months of the Merina (i. e. the Arabic months) is only approximate. The same applies to the calendar of the Sakalava, the Bara, the Tanala, and the Sihanaka, which is identical with that of the Betsileo. What is here said about the calendars of the peoples of the south and the centre of the island is also true of the calendars of the northern and eastern peoples[782]. To me it seems as though we have here a series of months of the ordinary type, in which the months are named and at the same time fixed with reference to the seasons, although I do not presume to decide upon the complicated question of the Malgassian calendar. There remains one possibility, viz. that the ‘months’ are seasons with no relation to the moon, but this possibility does not seem to have been seriously considered by those who can make use of the sources, which are only to be got at with extreme difficulty.