APPENDIX 535 up to date. But the list on which he worked seems to be later than the list used by Philotheus, for it includes the theme of Langobardia. (3) Incomplete enumerations of the themes, in the reign of Romanus I., are given by some Arabic writers, esjiecially by Ibn Khordadbeh (see M. Rambaud, L'empire grec, p. 18^). (4) The Treatise on the Themes. We must criticise Constantine for including Sicily and Cyprus, which did not belong to the Empire, and at the same time omitting Dalmatia, where there was the semblance of a province. Constantine raises the Optiraaton to the dignit}- of a theme, but apologizes for doing so ; it is only a quasi-theme. In this he was justified ; for, though the Optimaton was not governed by a strategos but by a domesticus, and was not in a line with the other themes, it was a geographical province. But the most serious matter that calls for criticism is Constantine's incon- sistency in stating definitely that Charsianon and Cappadocia are themes, and j-et not enumerating thern in his list. He discusses them under the heading of the Armeniac theme, but they should have headings of their own. This unac- countable procedure has led to the supposition that these two themes were temporarily merged in the Armeniac, out of which they had originalh* been evolved. (5) Anum.ber of notices in the treatise de Adrainistratione supjdy material for reconstructing a list of the themes c. a.d. 950-2. (6) To these sources must be added, the seals of the various military and civil oflBcers of the themes. M. Gustave Schlumberger's imjjortant work, Sigillographie byzantine (1884), illustrates the lists. Sardinia i)assed away from the empire in the 9th century, but it seems to have never formed a regular theme. Ve have however traces of its East-Roman governors in the 9th cent. A seal of Theodotus, who was " hypatos and diuc of Sardinia, has been preserved: and also seals of archons of Cagliari, with the curious style APXONTI MEPEIAS KAAAPE02. [Rambaud, L'empire grec au di.xieme siecle. p. 175 sqq. ; Bmy, Later Roman Empire, vol. ii. p. 33i) sqq. ; Diehl, L'origine du regime des themes dans l'empire byzantin (in Etudes d'histoire dix moven age, dedit5es k Gabriel Monod, 1896) ; Schlumberger, Sigillographie byzantine, passim (1884).] 4. CONSTANTINE PORPHYROGENNETOS ON THE ADMINISTRA- TION OF THE EMPIRE— (P. 62-91) The treatise of Constantine Porphyrogennetos on the Administration of the Roman Emjiire is one of the most interesting books of the Middle Ages, and one of the most jjrecious for the earl}- mediaeval history of south-eastern Europe. The author >vrote it as a handbook for the guidance of his son Romanus. Internal evidence allows us to infer the exact date of its com- jjosition. Cha])s. 1-29 were composed between a.u. 948 and 950; chap. 45 was composed in 952. The work was probabl}- published in 953. In his preface i Constantine promises his son instruction on four subjects. He wille.plain (1) which of the neighbouring nations maj' be a source of danger to the Empire, and what nations may be plajed off against those formidable neighbours ; (2) how the unreasonable demands of neighbouring peojjles may be eluded. (3) He will give a geographical and ethnographical description of the various nations and an account of their relations wth the Empire ; and (4) enumerate recent changes ami innovations in the condition and administra- tion of the Empire. This programme is followed. A summary of the contents of the book will probably interest readers of Gibbon, and it may be divided under these four heads. 1 P. 6C, ed. Bonn.
