innocent-looking hole in the courtyard of the mosque winds down into a black subterranean maze of passages, stairways, cells, and tiers of arches climbing above bottomless pits. So much earth and rubbish have sifted into this extraordinary labyrinth that its true extent can only be guessed at until it is systematically excavated. In the meantime, archæology has been very busy discussing which of the two contiguous towers that form a part of it was the tower of Anemas, and whether either of them was the tower of the emperor Isaac Angelus. The Anemas in question was a Byzantinised Arab, descendant of the Emir who surrendered Crete to Nicephorus Phocas, and he had the honour of being the first of many prisoners of state to be shut up in his tower. Whichever it may have been, however, the most unarchæological visitor is capable of enjoying a dip into that romantic darkness and the view, from the terrace, of a cypressed country beside the Golden Horn. [Illustration: The Palace of the Porphyrogenitus] On top of the hill stands the well-preserved ruin known in Turkish as Tekfour Seraï, the Palace of the Crown-Wearer. As to its real name, there has been the most fanciful variety of opinions. The palace is now generally supposed, however, to have been built in the tenth century by Constantine Porphyrogenitus. It seems to have been separate from the Palace of Blacherne, though on the analogy of the Great Palace it may have belonged to the same group. Architects as well as archæologists take a particular interest in Tekfour Seraï, because it is the only authentic piece of domestic building left of Byzantine Constantinople. The main façade is divided into three tiers of arched windows and ornamented by a mosaic of dark and light stone that recalls the brickwork of later Byzantine churches. What the general effect does not recall is the Venetian version of Byzantine civil architecture. We should not take that version too literally, of course, any more than the Venetian Gothic; but St. Mark’s is so true a transcription of a Byzantine church—without the crockets—that one has the more faith in the palaces. The difference may be chiefly one of periods. It is noticeable that the spacing of the arches of Tekfour Seraï is not like that of the _Fondaco dei Turchi_, to whose designed irregularity Ruskin drew attention. Neither has the checker-work of the façade anything in common with the plaques of porphyry and serpentine reflected in the Grand Canal. It suggests, rather, the checker-work of the Ducal Palace. The first tier of arches, too, looks like the same kind of ground arcade. Is it possible that any influence interacted between the two palaces? If so the presumption would be that it worked in Venice, under a Gothic cloak; for the Ducal Palace, or the lagoon front of it, belongs to the century after the Latin occupation of Constantinople. In the light of my question this latter detail is interesting, since the features I have noted decorate only the