where the country is more or less of a plain. There is evidence too that many of the cities grew by accretion beyond the limits intended by the first builders, and this fact will explain much of the irregularity observable in plan. But throughout, the assemblage of buildings round a court is a prominent feature, and the court would seem to have been the unit of growth (see Pl. XXVI, 2; p. 338). It is a noticeable fact that none of the sites in the east and centre of the Maya area exhibit any defensive qualities. The courts are open at the corners, and the sites selected for building are evidently not chosen with any strategic insight. In the west, however, the case is somewhat different, and we find settlements, such as Iximché and Utatlan, built as it were on peninsulas almost surrounded by inaccessible barrancas and connected with the "mainland" only by a narrow neck which could be easily defended by a mere handful of men. This fact goes far towards indicating that the ruins in the centre were centres of religious and ceremonial life rather than cities in the modern sense of the word.
The question as to how far Maya buildings were definitely oriented is rather complex. On the whole far fewer indications of the practice occur in the Maya region than in the Mexican. To speak generally the sites at Seibal and in southern Yucatan, as far as surveyed, are more carefully oriented than in northern Yucatan and elsewhere. The buildings at Copan are not oriented, but here and at such sites as Palenque, the surrounding hills and forests obscure the true horizon, and the arrangement of courts may yet be found to square with the apparent point of rising of certain heavenly bodies. An accurate survey of Tikal would throw a good deal of light upon the problem, since the situation of the ruins in comparatively level country, combined with the great height of the pyramids, would have enabled the inhabitants to obtain a far truer horizon than at most sites.
