some changos which he made in the types may be mentioned here. They are thus described by Wroth (xliii.): “He restored the cross (now the patriarchal cross)' on somo specimens, and on the folles an inscription-in this case OEOFILE AVROVSTE SV NICAS-takes the place of the familiar mark of value M. He also introduces on coins the legend Kύριε βοήθει τω σω δούλη 80 familiar on Byzantine seals and other monuments. On some of his coins Theophilus describes himself and his son Constantine as the doctor of Christ: Justinian II., on his solidi, had called himself Servus Christi,"

#, not the cross potent to which appeared on the older coinage.

APPENDIX VII

THE FALL OF THEODORA (chronology)

MICHAEL III. came to the throne January 21, 842, and died Septembor 23, 867, so that his whole reign lasted twenty-five years, eight months. For the last year and four months, Basil was his colleaguo (from May 26, 866), so that the rest of his reign, including both the period of his minority and his sole reign after Thcodora's fall, lasted twenty-four years, four months. Now, according to the contemporary chronicler Goorge the Monk (801), he reigned fourtoon years with Theodora, ten years and three months by himself. There is an error of a month, but here we are helped by the Anonymi Chron. Synt., ed. Bauer, p. 68 (cp. also an addition to the Chronography of Nicephorus, cd. do Boor, p. 101), where the joint reign is given as fourteen years, one month, twenty-two days. These figures are probably correct,' and so we can fix the meeting of the Senate which signalised the formal deposition of Thcodora to March 15, 836. In any case, these data seem to be independent, and they show that the deposition fell, not in 857 as Schlosser and Finlay supposed, but early in 856. This is the conclusion rightly supported by Hirsch (61). It bears out the narrative of the chroniclers (Simconand Gen.) who connect Theodora's fall from power immciliatoly with the murder of Theoktistos, who wils still alive at the time of Michael's marriage, to which we cannot assign illl carlier dute than 855. The two events must thus have been in chronological proximity.

But a serions ditlicnlty has arisen through the connexion of the deposition of Ignating from the Patriarchate and the expulsion of Theodora from the Palace. This connexion rests on good authority, the Libellus of Ignatius (composed by Thcoynostos) addressed to

I The other figures given by this unva a' here is omitteil.

The error source here are incorrect: Michael is may have arisen in the additions to the said to have reigned aloue eleven years, Chron. of Nicephorus from a repetition onu month, nine days. Thus the total of uriva a' in the preceding notice. The reign would be twenty-five years, three list stops with Basil I., Ro that the commonths, instead of twenty-five years, piler must have writtei. soon after A.D. cight monthis. Ju tle Col. Mauritesis 886.

Pope Nicolas (Mansi, xvi. 296): “When the sovran, persuaded by Bardas, wished to ostracize his mother and sisters from the Palace, he ordered me to tonsure them, but I would not oboy, because they wero unwilling ; for this reason too I was driven from the Church.” In accordance with this statement of the Patriarch is his biographer's intimation that there was not a long interval (nac Tù pekpóv) botween the two events (Vita Ignatii, 225).

According to the older view which was still held by Hirsch, Ignatius was deposed in November 857, so that if these statements iro trio, the tonsuring of the Imperial ladies cannot be placed before 857. Hirsch therefore (luc. cit.) rejects them as inaccurato. But it is quite impossible to set thom ilsido.

We know now that the deposition of Ignatius falls in November 858 (not 857), and this seems to make the difficulty still greater. The Patriarch could never speak is he does of a refusal to comply with the Emperor's wishes carly in 856 as the cause of his deposition near the close of 858.

The key to the solution of the difficulty is simple enongh. Both the chronological statement of Gcorgo the Monk (who wils writing some ten years later) and the evidence of the Patriarch are perfectly correct. The fall of Thcoclora from power is a distinct event, chronologically divided by an interval, from her expulsion from the Palace. The end of the joint reign fell in the beginning (perhaps March) of 856, ind was marked by the meeting of the Senate recorded in ('ont, Gcorg. 823. But Theoclora continued to live in the Palace and wils expelled at a much liter period. This seems to be the obvious inference from the data.

It is true that any one reading the chronicles of Genesios and Simeon would infer that the expulsion of Thcodora from the Palace ensued almost immediately upon the fall of Theoktistos. Gen. 90 kai μετα βραχ τα κατά την δέσποιναν έκταράττεται διο του παλατίου Ernstpuniçerill Kth. But the chronology of these writers is extremely vagne; they furnish very few absolute dates, and they had no precise information as to the intervals between events. Such phritses ils metà Bpuxó and perù puikpów generally conceal their ignorance. Morcover, if we look more closely at the statements of Simeon (Cont. Geory. 823), we find that they assume im interval (which may be either short or long) between the murder of Theoktistos and the expulsion of Thcodora. (1) Michael tried to pacify his mother, who was irreconcilable; then (2) he endeavoured to distress her: he expelled three of his sisters to Karianos, and the youngest, Pulcheria, to the monastery of Gastria ; afterwards he tonsured them all and confined them in Gastria. (3) lle was recognized by the Senate ils sole ruler, and created Bardils Domestic of the Schools. (1) He sent Theodorat also to Gastrin. Although this incroiini is confused and cannot be right in detail, yet it assumes a distinct interval during which Theodora lived in the Palace after her fall from power. And we may accept the statement, which was not likoly to be inventod, that the removal of her daughters to Karianos precedod hor own expulsion. Against this we need not press the actual words of Thoognostos (quoted above), which are accurato onough for his purpose if wo suppose that all the ladies were tonsured at the same time.

As this last event was connected with the deposition of Ignatius, it can hardly have been prior to 858. It is, however, worth noticing that the anthor of the l'ita Ignatii (258) assigns fifteon years and oight months to the joint reign of Michiel and Theodora. The period is ono yenur, suven months, too long. But it is a possible hypothesis that he reckoned not to her fall from power but to her expulsion. In that case the date of her expulsion would be about Angust or September 857. This would mean that Ignatins remained Patriarch for some fourteen months after his refusal to obey the Emperor's command. And it may be thought that this is quite possible, since that refusal wis certainly only one of the offences which Ignatins committed in the eyes of Michael and Bitrdits, iind wo might suppose that it simply began a breach between the Patriarch and the Court. But this is not probable, and does not do justice to the drift of the passinge in the Libellus.

If we look more closely at the chronological text in the l'ila Ignutii, wo observe that there is an error. Nine years :ure assigned to Michael alono, which, with the fifteen years, eight months, of the joint reign, makes twenty-four years, eight months, just a year too little. My conjecture is that the author intended to count the joint reign its extending to the expulsion of the Empress from the Palace, but that he miscalculated by a year. He ought to have written sixteen years, eight months. This would bring is to Angust or September 858 for the expulsion-a date which precerles the fall of Ignatius loy just about the interval we might expect.

APPENDIX VIII

THE WARFARE WITH THE SARACENS IN A.D. 830-832

The events and chronology of theso years have boon carofully studied by Vasil'ov, from the Grook and Arabic writers; but ho was not acquainted with the original Syriac Chronicle of Michael Syrus, knowing it only through the Armenian abbreviation and the compilation of Bar-Hebraeus, nor does he seem to have realised its importance for the reign of Theophilus, and especially for tho last

years of Mamun, Michael's source was the lost Chronicle of Dionysios of Tell-Mahre, the Monophysito Patriarch of Antioch (4.1), 818-815), who was not only a contemporary but was a friend of Mamun and was with him at times during theso years. He visited the Caliph in his camp at Kasin in the antumn of A.D. 831 (Michael Syr. 74), and accompanied him in the following February to Egypt (ib. 76). The cvidence of Michael is thorefore of the highest importance,

It appears that in the spring of A.1), 830, Theophilus--with Theophobos and his new l'ornimenian allies--crossed the mountains and captured and burned the town of Zapetra, perhaps massicring many of the inhabitants.' Mamun lost no time in retaliating. In the same year, marching lwy Mosul, Zongma, Membij, and

I This capture of Zaprtra, not men. Mamun in Cilicia ; further sliceesses in tioned by the lireek writers, is recorileil Romanin. . This brings 11s to the begin. by Michael Syr. 7 1, and must be accepted. ning of Ani. Sel. 1114 = October 832. There is, however, some chronological It is clear that the enpture of the four confusion in this chapter of Michnel. forts is here datel to the summer of 11:11 Immediately after his notice of the anil Manuel's flight to the same year arression of Theophilus he recorils : (1) =October 829 to October 830. It would without data, the capiture of Zapetri; follow that the capture of Zapetra full in (2) " in the following year" the revolt of 1140, 1.0, before October 829, i.e. before Manuel, aunt Man's capture, in or after the accession of Thcopiilus, Michael June, of four forts; (B) in May 1142 = 831, las iutroulucul a superlluous year. The the siege of Lulon; (4) in 1143 =()«tober true dates are : 1141 = 830, capture of 831 to October 832, Man's sleparture %:lpetrn, and Mam's capture of the for Damarilla, on bearing that Exypt foris ; 1142 (after October 1, 830), Mny, Binal revolloid: the capture of Lulon; siege of Lulon, etc. (Michel llistes liye "sit this period'" the riurn of Manuel to Seleucial your'n, which began on October 1). The pobiilus; the combassy of Theophilus;