CHAP. IV. MENDUT. 429 ten, the keystones are adorned with a mask, as in this last example. About two and a half miles from Boro-Budur is a small temple of a different class known as Chandi Mendut. It stands on a platform n ft. high, measuring 71 ft. wide and 83 ft. deep. The temple itself is cruciform, measuring 29 ft. 6 in. wide and 41 ft. 7 in. deep, the porch projecting more than is usual. This temple preserves its stone roof, the cell is 23 ft. square inside and at a height of 13 ft. from the pavement horizontal courses of stone, thirty-seven in number and 28 ft. in height projecting one in front of the other, from an inverted pyramid of steps which is terminated by a hollow cone. Externally the roof still rises to a height of about 50 ft. above the platform, and con- sisted probably of three storeys with a series of twenty-four miniature pagodas round the lower storey, sixteen around the second storey, eight round the third half sunk in an octagonal wall, being crowned with a larger dagaba. The walls are deco- rated with bas-relief figures of Hindu deities, groups of three or five in the larger central panels and single figures in the side panels all under canopies of slight projection. The sides of the platform are carved with figures and ornament in a series of panels. Inside the cell are three colossal figures about n ft. high each. The central one is Buddha, curly-headed of course, and clad in a diaphanous robe. The two other colossi, having only two arms each, are almost certainly intended for Bodhisat- twas. These three may have been placed in the cells at a later date. On one of the faces, externally, is Lakshmi, eight-armed, seated on a lotus, with attendants. On another face is a figure, four-armed seated cross-legged on a lotus, the stem of which is supported by two figures with seven-headed snake-hoods. It is in fact a slightly altered repetition of a group inserted among the older sculptures on the fagade of the cave at Karl. That insertion I have always believed to be of the 6th or yth century ; this group is certainly slightly more modern. The curious part of the matter is, that the Mendut example is so very much more refined and perfect than that at Karle. The one seems the feeble effort of an expiring art ; the Javan example is as refined and elegant as anything in the best ages of Indian sculpture. The same remarks apply to the sacred tree under which the figure is seated. Like all the similar conventional trees at Boro-Budur, they are complicated and refined beyond any examples known in India. The great interest, however, of this little temple arises from the fact that it almost certainly succeeded immediately to Boro- Budur. If it is correct to assume A.D. 650-750 as the period during which that temple was erected, this one must have been
