But Colonel Djevad is not more in error than the two greatest French authorities on Ottoman architecture. Saladin, in his summary of Ottoman history, instructs his readers as follows: ‘Alaeddin III, conquered by the Mongols, abandoned the sovereignty to Osman.... When the Osmanlis penetrated into Anatolia ... in proportion to the extension westward of the Ottoman Empire, we shall see the influence of Byzantine architecture increase.... Little by little, as the Turks approached Constantinople, this impregnation of the influence of Byzantium had an increasingly greater importance in the development of Ottoman art.’[727] This misconception of the origin of the Osmanlis leads him to state: ‘It is then indispensable to study the Seljuk monuments of Konia, which have _necessarily_ served as models to the first Ottoman monuments.’[728] From his premisses, Saladin has argued rightly. But his historical facts are wrong. Even if they were not, his conclusion could still be proved wrong. The refutation of his statement exists in the two earliest Ottoman buildings, the school and the kitchen for the poor at Nicaea, the date of whose construction Seadeddin places in 1331.[729] Both of these are typically Byzantine. In Brusa there is no Ottoman building of the Seljuk type which can be proved to have been constructed prior to Mohammed I (1413-21).[730] Parvillée, to whom the whole world owes a debt of gratitude for his able reconstruction of the precious historic monuments of Brusa, starts his scholarly work on Ottoman architecture in the fifteenth century with these words: ‘Towards the end of the thirteenth century the Seljuk Empire disappeared. On its ruins arose that of Osman.’ He not only follows Hammer: he uses his very words![731] From the historical point of view, I maintain that the Byzantine influence was an indissoluble factor in the evolution of Ottoman architecture from the very beginning. In this I am supported, from the expert architect’s point of view, by the two German authorities on this subject.[732] The Seljuk, Arab, and Persian influences entered in at a considerably later period. There exists in tradition and in law an intimate connexion between the House of Osman and the Grand Tchelebi of Konia. This has been pointed to as a confirmation of the hypothesis that the Ottoman sovereigns derived their authority originally from the Seljuks of Rum. I do not deny the force of tradition. In the absence of early records, the beginning of this connexion must remain a moot question. But the evidence from outside sources makes reasonable my doubt as to the existence of this connexion before the reign of Mohammed I or Murad I. There are two other arguments which might be adduced in this appendix, the questions of Osman’s title as an independent ruler, and of the chieftainship as an elective office among the Turkish tribes. But both of these have already been discussed in the text and the foot-notes of the chapter on Orkhan. APPENDIX B