The roof of the temple has so entirely disappeared that Baron Hügel doubted if it ever possessed one.[310] General Cunningham, on the other hand, has no doubts on the subject, and restores it in stone on his plate No. 14. The absence, however, of any fragments on the floor of the temple that could have belonged to the roof, militates seriously against this view; and, looking at the tenuity of the walls and the large voids they include, I doubt extremely if they ever could have supported a stone roof of the usual design. When, too, the plan is carefully examined, it will be seen that none of the masses are square; and it is very difficult to see how the roof of the porch could, if in stone, be fitted to that over the cella. Taking all these things into consideration, my impression is, that its roof--it certainly had one--was in wood; and knowing how extensively the Buddhists used wooden roofs for their chaitya halls, I see no improbability of this being the case here at the time this temple was erected. [Illustration: 161. View of Temple at Marttand. (From a Photograph.)] The courtyard that surrounds and encloses this temple is, in its state of ruin, a more remarkable object than the temple itself. Its internal dimensions are 220 ft. by 142 ft.,[311] which are respectable, though not excessive; they are not much more than those of the temple of Neminatha at Girnar (Woodcut No. 126), which are 165 ft. and 105 ft., though that is by no means a large Jaina temple. On each face is a central cell, larger and higher than the colonnade in which it is placed (Woodcut No. 162), but even then only 30 ft. in height to the summit of the roof, supposing it to be completed, and the pillars on each side of it are only 9 ft. high, which are not dimensions to go wild about, though their strongly-impressed Grecian aspect is certainly curious and interesting. [Illustration: 162. Central Cell of Court at Marttand. (From a drawing by General A. Cunningham.) No scale.]