The plan of a northern temple is always a square internally, and generally the same form is retained in the exterior; but very rarely, if ever, without some addition. In some instances it is only a thin parallel projection, as at A in the diagram (No. [229]). Sometimes it has two such slices added, as at B; but in the oldest examples these are only half the thickness shown here. From this they proceeded to three projections, as at C, the oldest examples being the thinnest. In more modern times the thickness of the projections became equal to their distance from each other, as at D; so that the temple became in plan practically a square, the sides of which were parallel to the diagonal of the original square or to the line E F G. Even, however, when this was the case, the cell always retained its original form and direction, and the entrance and windows kept their position on what had thus practically become the angles of the building. This is the case with the temple at Benares, shown in Woodcut No. 228, and generally also with the Jaina temples, and especially the case with the temple on the Takht-i-Suleiman at Kashmir. Although the depth and width of these offsets vary considerably even in the same design, the original square is never lost sight of; the four central angles, as at F, being always larger and more strongly accentuated than the others, and their line is always carried through to the summit of the pyramid.
It will be observed that by this process we have arrived at the same form or plan for a solid building that was attained by the arrangement of pillars described above, page 216. In fact, the two forms were elaborated simultaneously, and were afterwards constantly used together. My impression is, that the pillared arrangement is the oldest, and led to the deepening of the additions to the solid square till the two became identical in plan. Whether this were so or not, it is one of the most distinguishing features of northern Hindu architecture.
In the very centre of India, near a place marked Adjmîrghur on the map, is a sacred tank, from which it is said that the Soane flows to the north, the Mahanuddi to Cuttack in the Bay of Bengal, and the Nerbudda to the Indian Ocean. All these rivers certainly have their sources in the hill. The spot has always been held sacred, and is surrounded by temples—as far as can be gathered from the imperfect accounts available—of great age. On the south and east of this hill extends the great and fertile table-land of Chutteesghur. This is now, and has always been, so far as our knowledge extends, one of the principal seats of the native tribes. My conviction is, that if that country and the surrounding valleys could be examined, much older forms of these temples might be discovered—some perhaps so old as to betray the secret of their origin; but, till this is done, the Bengali devala must be relegated—like the Irish round towers[413]—to the category of unexplained architectural puzzles.
CHAPTER II.
ORISSA.
CONTENTS.
History—Temples at Bhuvaneswar, Kanaruc, Puri, Jajepur, and Cuttack.
The two provinces of India, where the Indo-Aryan style can be studied with the greatest advantage, are Dharwar on the west, and Orissa on the east coast. The former has the advantage of being mixed up with the Dravidian style, so as to admit of synonyms and contrasts that are singularly interesting, both from an ethnological and historical point of view. In Orissa, on the contrary, the style is perfectly pure, being unmixed with any other, and thus forms one of the most compact and homogeneous architectural groups in India, and as such of more than usual interest, and it is consequently in this province that the style can be studied to the greatest advantage.
One of the most marked and striking peculiarities of Orissan architecture is the marked and almost absolute contrast it presents to the style of the Dravidian at the southern end of the peninsula. The curved outline of the towers or vimanas has already been remarked upon, but, besides this, no Orissan towers present the smallest trace of any storeyed or even step-like arrangement, which is so universal further south, and the crowning member is never a dome, nor a reminiscence of one. Even more remarkable than this, is the fact that the Orissan style is almost absolutely astylar. In some of the most modern examples, as for instance in the porches added to the temples at Bhuvaneswar and Puri in the 12th and 14th centuries, we do find pillars, but it is probably correct to state that, among the 500 or 600[414] original shrines at Bhuvaneswar, not one pillar is to be found. This is the more remarkable, because, within sight of that capital, the caves in the Udayagiri (ante, p. 140) are adorned with pillars to such an extent as to show that their forms must have been usual and well known in the province before any of the temples were constructed. When we recollect that no great temple in the south was considered complete without its hall of 1000 columns, and many besides this had hundreds dispersed about the place, and used for every conceivable purpose, the contrast is more striking, and shows what a complete barrier the Chalukyas, whoever they were, interposed between the two races on this side of India, though not on the other. As a rule, every Orissan temple consists of two apartments, similar in plan, as shown in the diagram ([Woodcut No. 124]). The inner one is generally a cube, surmounted by a tower, here called Bara Deul, or Dewul, corresponding with the vimana of the south, and in it the image or images of the gods are enshrined; in front of this is a porch, called Jagamohan, equally a cube or approaching it, and surmounted by a pyramidal roof of varying pitch. The peculiarities are illustrated in the diagram ([Woodcut No. 124]) just referred to, which purports to be an elevation of the celebrated Black Pagoda at Kanaruc. It is only, however, an eye-sketch, and cannot be depended upon for minute detail and correctness, but it is sufficient to explain the meaning of the text. Sometimes one or two more porches were added in front of this one, and called Nât and Bhog mandirs (mantapas), but these, in almost every instance, are afterthoughts, and not parts of the original design. Be this as it may, in every instance in Orissa the tower with its porch forms the temple. If enclosed in a wall, they are always to be seen outside. There are gateways, it is true, but they are always subordinate, and there are none of those accretions of enclosures and gopuras that form so marked a characteristic of the southern style. There generally are other shrines within the enclosures of the great temples, but they are always kept subordinate, and the temple itself towers over everything to even a greater extent than that at Tanjore ([Woodcut No. 191]), giving a unity and purpose to the whole design, so frequently wanting in the south.
Other contrasts will come out as we proceed, but, in the meanwhile, few examples bring out more clearly the vast importance of ethnography as applied to architecture. That two people, inhabiting practically the same country, and worshipping the same gods under the guidance of the same Brahmanical priesthood, should have adopted and adhered to two such dissimilar styles for their sacred buildings, shows as clearly as anything can well do how much race has to do with these matters, and how little we can understand the causes of such contrasts, unless we take affinities or differences of race into consideration.
History.