scribam, iam enim charta ipsa ne nos prodat pertimesco: itaque posthac, si erunt mihi plura ad te scribenda, aλyopiais obscurabo. Another suggestion is made by Cicero Att. II 19. 5. (posthac ad te aut, si perfidelem habebo cui dem, scribam plane omnia, aut, si obscure scribam, tu tamen intelleges; in eis epistolis me Laelium, te Furium faciam; cetera erunt év airypois,) and modified in the next letter. (II 20. 5 quod scripseram me1 Furio scripturum, nihil necesse est tuum nomen mutare. Me faciam Laelium et te Atticum, neque utar meo chirographo neque signo, si modo erunt eiusmodi litterae, quas in alienum incidere nolim.)

The two periods of Cicero's life when such precautions would have been most needful, for Atticus' sake if not for his own, were the months of his exile (696-7) and those which elapsed between the June day 706 when he finally followed Pompey over seas and his pardon by Cæsar more than a year after the battle of Pharsalia. These letters comprise the third and eleventh books ad Atticum.

Turning now to the internal evidence of the text, we are struck by the entire absence in these two books of those Greek epithets and quotations which occur so frequently in most of the other letters to Atticus. We know from Att. X 8. 1. that Cicero's ever prudent friend felt so keenly the danger which attended their correspondence in 705 as to have doubted the desirability of writing at all, and we know that Cicero disregarded the delicate intimation even while admitting its wisdom. Still harder would it have been for him in 696 or 707 to deny himself such an outlet for his conflicting emotions as was afforded by these most free and intimate epistolary outpourings; and that he compromised the matter by employing a cipher seems at least a plausible theory.

We need not suppose that he is referring to the comparative laboriousness of following a code, when he makes use-as he does so often at these times-of such phrases as plura scribere non possum and non queo plura scribere, and still less that he was too much disturbed in mind during those trying times to have the patience requisite for employing a difficult cipher, for he undoubtedly had at hand an expert amanuensis, who possessed the key to all his "enigmas," and who was able to use the most complex. But might we not satisfactorily explain the absence of Greek words by supposing that he adopted some such simple expedient as that of moving the letters of the Latin alphabet a certain number of

1 Reading of M., variously altered by editors who have thought emendation necessary.

places forward or back? Greek words could not have been left intact on such a page, for their significance would have been suggestive to the inquisitive reader, while, if they too had been transposed, according to the code, but in their own alphabet, the small, isolated groups of foreign characters must have given the clue to the cipher. Nor could the letters of the Greek words have been transposed and then written in Latin characters, because of the different order in which the letters occur in the two alphabets as well as the presence in the Greek of the doubleconsonant symbols. The only feasible way would have been to keep to the vernacular, as Cicero has done in the third and eleventh books.

LOUISE DODGE.

MAMATPAI.

In the current volume of this Journal XXII, p. 195 foll., two American scholars have tried with greater or less probability to carry back to their Indian original form a number of Indian glosses in the Lexicon of Hesychios. With respect to one of them μαμάτραι οἱ στρατηγοί παρ ̓ Ἰνδοῖς, I would propose another explanation which gives both a simpler account of the word itself and one more in accordance with phonetics. In my opinion, μaμárpai represents Skrt. mahāmātrāḥ, a well-known term to denote a minister of high rank. Mahāmātrāḥ samṛddhe cāmātye hastipakā dhipe. 'Mahāmātra signifies as well a high minister as an elephant-driver.' So the Medinikośa. In literature the word is very common in both acceptations. In the Amorakośa commentary found in the edition of Vamanacharya Jhalakikar, Bombay, 1890, the right etymology is given (p. 181) mahati mātrā yeşām te mahāmātrāḥ. Its translation by σrparηyoì may have been made by Megasthenes or some other source of information about Indian matters in the time of the Diadochoi; and as στpaτnyòs is employed in a wider sense than to denote mere military power, it may have been considered an adequate term for rendering tolerably well the name by which the Indian high officials were designated. Cp. σrparnyòs as equivalent of the στρατηγὸς Roman praetor. That mahāmātra in Greek transcription must become paμárpa (as to the accent, cp. mátrā), is almost evident.

GRONINGEN.

J. S. SPEYER.

REVIEWS AND BOOK NOTICES.

Assyrian and Babylonian Letters belonging to the Kouyunjik Collections of the British Museum, by ROBERT FRANCIS HARPER, Ph. D., Professor of the Semitic Languages and Literatures in the University of Chicago. Part V. The University of Chicago Press, Luzac and Co., London, 1900. The value of the letters and dispatches to students of Assyriology is not easily overestimated. They frequently supplement the historical inscriptions with valuable details, and, in some instances, are the only source of information in regard to important events; they cast much light upon the administrative methods of the Assyrian government, and upon the practical workings of the state religion; and, although with few exceptions of an official character, they furnish valuable imformation concerning AssyroBabylonian life and customs. From the standpoint of philology they constitute a rich mine, yielding a wealth of material to be found in no other class of cuneiform texts. At first, owing to the superior attractions of the historical, religious, grammatical, and lexicographical texts, the letter tablets were little studied, and it is only within the last fifteen years that they have their due share of attention.

Father Strassmaier in his Alphabetisches Verzeichniss (1886), S. A. Smith in his Assyrian Letters (1888) and in his Keilschrifttexte Asurbanipals, and Dr. Hugo Winckler in his Sammlung von Keilschrifttexten (Part II, 1894), have published a considerable number of these texts, and Prof. Friedrich Delitzsch, in a series of articles in Beiträge zur Assyriologie (Vols. I & II, 1889-91) laid the foundation for their scientific study. But to Prof. Harper belongs the credit of conceiving and carrying into execution the plan of publishing a complete corpus of Assyrian and Babylonian letters, thus making the whole mass of these interesting_texts available for study. The first volume of Prof. Harper's Letters appeared in 1892, and five volumes have now been published containing, in all, 538 texts edited with great care and skill, and printed in a manner that leaves nothing to be desired. The fact that at least three additional volumes will be required to complete the series is evidence both of the magnitude of the undertaking and of the wealth of material available. The fifth volume, which has recently appeared, measures fully up to the standard of excellence set by its predecessors.

Among the writers of the 103 letters contained in it may be

mentioned Tem-Ašur, who is probably to be identified with the eponym of the year 717 B. C.; Tab-çil-Eshara, governor of the city of Asshur, who filled the office of eponym in 714; Ashurrêçu'a who, under Sargon, held a military command on the northern frontier of Assyria, and is mentioned in the correspondence of Sennacherib; Arad-Nabû, a priestly official contemporary with Esarhaddon; and Bel-ibni, governor of the Gulf District in 650, who played an important part in the Elamite wars of Ashurbanipal. It should be noted, by the way, that Nos. 460 (K. 1250) and 462 (K. 1374), although the writer's name is broken away in both instances, were certainly written by Belibnî. The subject matter, the general style, and several marked peculiarities of expression leave no room for doubt as to their authorship. No. 469 (48-11-4, 282), although badly mutilated, is especially interesting. It contains an appeal to the King from the people of Erech who state that a dispute about some houses, gardens, and other property had been decided in their favor by "thy father Ashurbanipal" (obv. 12-13; rev. 1). The King addressed must, therefore, have been either Ashur-etil-ilâni or Sin-shar-ishkun (the Saracus of Abydenus), and the letter affords new evidence of the fact that the rule of Assyria was maintained in Babylonia for some time after the death of Ashurbanipal.

Very few textual errors have escaped the editor's watchful care. In No. 521, rev. l. 21, ar (ar-ra-ti) should be read instead of bi, and, in No. 469, rev. 1. 2, the context shows that the first character must be di (di-i-nu) not ki. Both errors are trivial and the present writer has discovered no others. In the preface, Prof. Harper states that Part VI will probably be ready within the present year, and it is to be hoped that this expectation may be realized. The appearance of a new volume of the Letters is ever a welcome event.

CHRISTOPHER JOHNSTON.

Textes et Monuments Figurés Relatifs aux Mystères de Mithra, publiés avec une Introduction Critique par Franz Cumont, Professeur à l'Université de Gand. Bruxelles, H. Lamertin. Two Volumes, 4°: Volume II, Textes et Monuments, 1896, pp. viii, 554; Volume I, Introduction, 1899, pp. xxviii, 377.

When a certain scholar of international reputation, during a recent Winckelmannsfest at the German Archaeological Institute at Rome, pictured as the ideal of scholarship that in which there should be a union of the untiring industry and patience of the Teuton and the brilliant intuition of the Latin, he gave utterance to a sentiment which is common among scholars of the Latin nations, who, while they admire and imitate German scholarship,

find in it a certain heaviness and a tendency to rest content with the bare collection and presentation of material. After an examination of M. Cumont's two beautiful volumes on Texts and Monuments Relative to the Mysteries of Mithras, we cannot but think that their author, if indeed he has not fully realized this ideal, has at least more nearly approached it than any scholar who has yet written on a like subject. His geographical position typifies his scholarship: living on ground common to Teuton and Latin, and doctus sermones utriusque linguae, he exhibits in the highest degree the characteristics of both German and French scholarship.

An examination of M. Cumont's work properly begins with the second volume, which was first issued, and contains the texts and monuments which constitute the sources of our knowledge of the cult of Mithras. The contents are presented under three heads-Textes Littéraires, Textes Épigraphiques, and Monuments Figurés. Under Textes Littéraires, the author gives oriental, Greek, and Latin literary sources, arranged according to the alphabetical order of their authors' names. Greek and Latin sources are grouped together under one head. Under oriental sources, only Armenian texts are given, and those in translation. To have transcribed all the texts which form his oriental sources would have necessitated the presentation, not only of a great part of the Avesta, but of the Pahlavi writings, and would have drawn the author into a task which he prefers to leave to those whose knowledge of the oriental languages will permit them to do justice to it. He therefore limits himself to the transcription of a translation of the Armenian texts, and refers the reader by foot-notes to the Avesta and other oriental sources. Concluding the literary texts is a collection of Textes Douteux, passages which seem to contain allusions to Mithras, but are not beyond doubt. Following is an appendix containing Noms Théophores to the number of one hundred and six, classified according to territory.

Under Textes Épigraphiques are arranged in two divisions oriental, and Greek and Latin inscriptions. Three inscriptions in Persian constitute the oriental epigraphic sources, while there are five hundred and forty-seven in Greek and Latin. These are classified according to provincial distribution in Asia, Europe, and Africa. The number of Greek inscriptions is exceedingly small. Thirty-six Inscriptions Douteuses, five Inscriptions Fausses, with a concordance for use as a guide to C. I. L., C. I. G., etc., follow.

The third part of Volume II-Monuments Figurés—is a catalogue of all the known Mithraic monuments. Temples, grottoes, coins, amulets, paintings, statues, reliefs, altars, and all other objects having to do with the worship, are classified in the same manner as the inscriptions, are minutely described, and abundantly illustrated by, four hundred and ninety-three cuts, and nine plates in heliotype. This is the most valuable and important part