at once occur to the teacher that there are other and greater difficulties to counterbalance this advantage. The boy's notions regarding Socrates are very vague, and no amount of talk by the teacher would avail to make plain to him just why the Athenians should think him obnoxious and desire to put him to death. A concrete portrayal of the living Socrates, as he went about interrogating every man he met as to the grounds of his beliefs, would prepare the student much more effectively to grasp the real meaning of that fateful trial.

In casting about for such a means, one might be tempted to try the Memorabilia; but Xenophon was apparently too conscious of the need of an apology for the life of Socrates. Hence he gives us a rather distorted picture, calculated to lead us to think of the master as a preacher. Certain of the lesser Platonic dialogues might also receive some consideration, but they are not so well adapted for the purpose as is the Euthyphro. The discussions of temperance, courage, and friendship, in Charmides, Laches, and Lysis, do not touch the quick, as does the discourse on piety. For here we have a most vital question; and the logical subtleties, which render the dialogue difficult to the student, are well calculated to impress him with the baffling sense of confusion and distrust with which the colloquies of Socrates filled the unschooled minds of his Athenian auditors. In view of these considerations it would seem that it is a just matter for regret that the Euthyphro is so little read in American schools.

Two years ago, when seeking the source of the runes, I discovered the origin of the Latin letters G and Z. In studying the various Latin texts said to contain the letter 2, I succeeded in showing that the reputed to stand in the line of the Salian Hymn beginning cozeulodorieso was only a medieval spelling for c, and that the coceulod orieso of the Basel manuscript was the correct reading. I had accomplished my purpose and, because of lack of time, resisted the temptation to study further this most interesting bit of early Latin. When, however, Professor Kittredge of Harvard University expressed an interest in my interpretation of these two words but wished to know how they could be explained as a portion of a hymn to Janus, I felt impelled to extend my study to the context. This paper may, therefore, be regarded as a humble offering made to Latin scholars by an Anglicist at the suggestion of a fellow Anglicist.

In attempting to decipher an archaic Latin inscription our chief difficulty consists in the meagreness of our knowledge of preclassical Latin. In dealing with a piece of classical Latin that has come down to us in medieval manuscripts, we have to reckon with the possibility that the medieval scribe has made errors in the transmission of the text. But in confronting a piece of literature like the Salian Hymn, we have to contend with both these disadvantages. Indeed, the fact that the original text was archaic Latin, and thus largely unintelligible to the medieval copyist, much increased the danger of his altering it in copying, whether unintention

1 Read at the Special Meeting held at Philadelphia, December, 1900. For the literature of the subject see Maurenbrecher, Carminum Saliarium Reliquiae (1894); Birt, Rheinisches Museum, Vol. LII (1897), Ergänz. p. 193; Spengel's edition of Varro (1885), p. 127, etc.

=

ally or with the aim of putting it into what he supposed would be a more correct form. (The same phenomena appear in the Old-English manuscript runic alphabets, which were copied by scribes who were of an antiquarian turn of mind. but ignorant of the real character of what they were copying.) Moreover, we have also to deal with abbreviations and their resolution. And these again may be ancient or medieval. We have early evidence that there were abbreviations in the ancient received text of the Salian Hymn (Festus 244 (205), cf. below), and it is clear that others, for example uo uero, were introduced later, when the d of the ancient uerod was no longer recognized as a part of the word. The various manuscripts differ from one another in the abbreviations they present and in the resolutions they offer of abbreviations in the texts from which they were copied. Hence, we have to consider the preclassical forms of Latin, early Latin abbreviations, the misinterpretation and alteration of these by medieval copyists, medieval abbreviations, and the shifting value of some of these, for example, -3=-us, later -m, or -que. Nevertheless, the vicissitudes that these three lines have suffered (except at the hands of commentators) are not many, and hence the number of corrections that have to be made is really small.

The text, as might be expected, is best preserved in the three best manuscripts, namely F, V, and (cf. Spengel, p. 127 f.) and may be given as :

coceulodorieso omia uo adpatula coemisse.

coceulod orieso. As shown in the paper alluded to above (The Origin of the Latin Letters G and Z, Transactions of the American Philological Association, 30, p. 39, etc., to which I refer the reader for a full presentation of the matter), the z for in F (and in the inferior a and M) is due to the fact that in the Middle Ages c before e was pronounced ts, for which sound-group and c were equivalent spellings, as The r for c in the inferior manu

they still are in German.

scripts G, H, and b, is due to the fact that in the eleventh century, and for some time after, the letters and c were very similar in form (Wattenbach, Anleitung, p. 46).

omia, omnia, needs no explanation.

o. As stated above, the d of uerod was not understood and was supposed to be the remnant of an original ad, which was therefore written and attached to the following patula; hence the reading no adaptula, in which no is one of the usual abbreviations for vero (Wattenbach, p. 82). In the same way, the s of meliosum below prevented the medieval scribe from recognizing in this spelling the familiar meliorum; and so, not understanding the -um, he separated it and made eum of it. Compare also the attempt of the writer of manuscript M to make good Latin of the error arcs (for arx) by writing ea res (Spengel, p. xiii). uero in the form ů (Wattenbach, p. 82) was also taken for ǹ non, and so we find non written out in b. Such a ů or ǹ was also taken for enim, and so we find · n ⋅ in G, and the full enim in the related H.

•n• =

=

coemisse appears to be a misinterpretation (on the basis of esse, Wattenbach, p. 82) of cocmie or coemise, which was meant for coemis es with final s written over the preceding vowel (Wattenbach, p. 60). This could easily happen, especially in the mechanical copying of a text that was not understood (Wattenbach, p. 84-85); but an early copyist may have regarded the -o of orieso as marking a first singular of some verb or other, and coemisse as an infinitive depending on it. A spelling like coemise is reflected in the ceruse in M, which is evidently a copy of an older ceruse or ceruse for ceruses; cf. below.

ian is for ianus (or, if the abbreviation is old, for ianos) and may have been written an' in the original text (Wattenbach, p. 70), or without sign of abbreviation. The abbreviation tan or an' was also misread and appears as tam in M and B, and this amcusianes was then copied lamcusianes in H.

cusianes for cusiatios doubtless arose by false interpretation of cusiari or cusiario, for there was a striking similarity

between 7 and n, while it is often almost impossible to distinguish between final -o and final -é.

The spellings duonus, cerus, dunus, and ianus are due to the early substitution of the classical -us for the archaic -os. That this was done early in the transmission of the text is shown by the fact that it evidently antedates the splitting up of meliosum into melios eum, which appears in all the manuscripts. The us is frequently abbreviated. Thus, while we find dunus in V (misdivided dinnıs in p), we find dun in H, dun; in F, dung in a, and in M dumque, a mistaken resolution of dung on the model of at3 = atque (Wattenbach, p. 71).

dunus for duonus (or dunos for duonos) is an example of a very common error. As I shall show in a forthcoming treatise on miswriting, the recurrence of a letter may lead to its omission in one of the two places or to the omission of an adjoining letter, as well as to other changes that cannot be considered here. I need but refer to a few of the many instances of this process that were so sadly misinterpreted by Birt in his Sprach man avrum oder aurum?, Rheinisches Museum, 52, Ergänz.; for example, a(u)rum 56, Claudius 89, fa(u)nus 90, nunq(u)am C.I.L. ix. 1524; na(u)tae 91, fa(u)nam 90, ca(u)sa 87, qu(i)bus, s(ci)urus, turp(i)us 180; pingu(i)um 180, ca(u)l(i)culi 93. (This is also the true explanation of the frequent confusion of au with a and u in Gothic (Braune, § 105, A 2) and of ei with e and i (Braune, § 7, A 2-4), as I shall sometime show in detail.)

The ueniet preserved in V appears elsewhere as ue uet. This corruption is easily explained. There is no mistake more common than the confusion of u and n. Thus ueniet became ueuret, and this became ueuet, exactly as duonus became dunus, as shown above. The anusue that thus arose was changed in M to anusque.

The po of Vand p is incorrectly expanded in F to pom and joined to melios (thus producing pommelios), in a and M to pos, and in G and H to post, the proper abbreviation of which was (Wattenbach, p. 76). That the pō of V and

is the original, we learn from Festus, who tells us, page