To this we shall have occasion to revert presently, when describing the Dravidian style; but meanwhile it may be assumed that this theory represents the facts of the case more nearly than any hitherto brought forward. The Chalukyas of Kalian were situated on the border-line, halfway between the north and the south, and they, or their subjects, seemed to have practised the styles of architecture belonging to those two divisions indiscriminately--it might almost be said alternately--and we consequently find them mixed up here and at Dhumnar in a manner that is most puzzling. The last king of this race, Vicramaditya II., ascended the throne A.D. 733,[289] and died probably in or about the year A.D. 750. It was probably, therefore, before that date that these Dravidian temple-forms were introduced by the Jains at Ellora. The Kylas and other great Saiva temples were, I believe, excavated by the Cheras or Cholas, who were the Dravidian races, and, if I mistake not, superseded the Chalukyas on the death of Vicramaditya, their last king, and carried their power, as will presently be explained, up to the Nerbudda. The Jains, however, seem to have been earlier in the field, and this little shrine in the court of the Indra Subha looks very much as if it may have been the model that suggested the Kylas, the greatest of all Indian rock-cut examples of its class. CONVERTED MOSQUES. Another form in which we can study the architecture of the Jains in the north of India is the courtyards of the early mosques which the Mahomedans erected on their first entry into India. So essentially do some of these retain their former features that it might be convenient to describe them here. It is doubtful, however, in some instances whether the pillars are--some or all of them--in their original position, or to what extent they have been altered or eked out by the conquerors. Be this as it may, for our present purposes the one fact that is certain is, that none of them are now Jaina temples. All are Mahomedan mosques, and it will, therefore, be more logical, as well as more convenient, to group them with the latter rather than with the former class of buildings. Were it not for this, the Arhaí-dín-ka Jomphra, at Ajmir--so called--might be, and has been, described as a Jaina temple.[290] So might a great part of the mosque at the Kutub, Delhi. That at Canouge, however, was originally a rearrangement, and has been much altered since I knew it; that at Dhar, near Mandu, is of comparatively recent date; while the Jaina pillars, so frequently used at Ahmedabad in the 15th century, are all imported, and used in positions for which they never were intended.