One cannot pretend to deal with the whole question of the textual criticism of Strabo on the strength of part of Book IX. Some conclusions, however, may be arrived at, since the MSS. we deal with are typical and Book IX. (and VIII.) contain the most vital part of Strabo for the question.

Ib. 511 sq.

lapsuramque domum subeas, ut sanguis Aleuae,

Stella Leoprepidae cum fuit aequa uiro.

A foot is lacking at the beginning of line. Perhaps add <μ⋯κραν> to go with ναυστολ⋯, lost after the end of I. 12. Silenus expatiates on the με⋯ζονα πόνον of I. 10.

Gercke reads with the MSS. Δ and E: ‘haec adhuc Etruscis philosophisque communia sunt: in illo dissentiunt quod fulmina a Ioue (“fulmina nouem” the best MSS.) dicunt mitti et tres illi (om. Δ) manubias dant.’ Mr. Garrod (C.Q. vol. VIII. No. 4, p. 281) remarks that the soundness of ‘nouem’ is clinched by the passage he cites from Pliny, N.H. II. 138. But the suggestion he bases on this—to alter ‘illi’ to ‘Ioui’—seems unsatisfactory, as ‘mittiy’ in the first clause is left in crying need of a governing agent; ‘Ioui’ comes in too late in the sentence, and the whole is given an awkward turn that Seneca surely could and would have avoided—by writing ‘fulmina >a Ioue< nouem dicunt mitti.’ Even if ‘a Ioue’ is merely a conjecture of the Δ family, as Mr. Garrod seems to hold, it is a conjecture good enough to pass muster. Perhaps copyists then dropped ‘illi’ from a feeling that it was needless after ‘a Ioue.’ But construction practically demands it.

The subject of these lines may be found in Caes. B.C. I. 54, from which they are in part derived, though probably at second hand. The reference is to Caesar's tactics after the floods in the plain around Ilerda. He built a number of coracles after the British fashion, and had them conveyed to a point on the right bank of the Sicoris, twenty-two miles from his camp. In these boats he sent a number of men across the river, who fortified some rising ground. Then he sent a legion across and started the building of a bridge from both sides at once: hue legionem postea traiecit atque ex utraque parte pontem institutum biduo perfecit.

The fact that Greek ζ = Indo-European j in the words ζειαί, ζέω, ζ⋯λος, ζόρξ, ζυγόν, , and ζωρός may be regarded as fairly well established (Boisacq, Dictionnaire étymologique de la langue grecque, 307, 309, 196 sq., 311 sq.; cf. also Berneker, Slavisches etymologisches Wörterbuch, i. 447 sq.; most recent bibliography in Brugmann, Griechische Grammatik,4 151 sq.), despite the counter-arguments of Sommer (Griechische Lautstudien, 137–160; cf. also Pedersen, KZ. xxxvi. 103 sq., Meillet, Introduction à l'étude comparative des langues indo-européennes, 88).

77, 3 (manubia) uastat in quae includit.

E has incidit for includit–another example, eagerly welcomed by Gercke, of free interpolation in this MS. Perhaps incutitur.

IN the Byzantinische Zeitschrift xxi. 52 Paul Maas states: ‘Es ist das Verdienst von H. B. Dewing, zuerst erkannt zu haben, dasz Prokop seine Satzschliisse rhythmisch reguliert.’ That this is only partly true appears from the remark of Heisenberg in the Berliner Philologische Wochenschrift, 1901, Sp. 1481, who comments on it, and that in a case of text–criticism, and likewise from a remark of Cronert in the Rheinisches Museum, 54, 1899, 593. Dewing was the first to point out the connection between the rhythm of Procopius and Meyer's law and to collect specific statistical material, although his numbers are to a great extent of no practical value. Compare in this connection Maas B.Z. xix. 593 (Maas himself came to the conclusion later that he exaggerated in this article: тά, μή, σύν, пΕρί should, according to Maas, here not be considered as accentuated). It appears to me that the correct statement of the end–rhythm of Procopius is that given by Maas for Constantinus Manasses, B.Z. xi. 505: ‘Im Ausgang der Satzglieder musz die Zahl der zwischen den letzten beiden Hochtonen stehenden Silben eine gerade sein,’ on condition that 0 is considered as an even number.

There are few ancient texts in a condition so mutilated and unsatisfactory as that of Fronto. Not only was the Codex containing his letters used subsequently for recording the Latin translation of the Acts of the first Council of Chalcedon, and even in one case (Cod. Vat. 227) for two writings one over the other, but there are also indications that there was an earlier writing under the Fronto text, of which Studemund found traces in Cod.Ambr1. p. 107 and Hauler on p. 251 of the same. The latter detected the name Hadrian, and supposed that a work of that Emperor might have been transscribed there. However, the Ambrosian MS. has recently been thoroughly cleaned and laboriously examined by Dr. Hauler, whose authoritative edition of Fronto has long been expected. Meanwhile Naber's edition, faulty as it is, holds the field, and the references in the following notes will be to it. The numerals indicate page and line.

So in this outburst of Cornelia should line 104 be punctuated. For the poenas crudelis compare VII. 431 ‘quod semper saeuas debet tibi Parthia poenas’ and Verg. A. 6. 501 quis tam crudelis optauit sumere poenas? whence, or from ib. 585, ‘uidi et crudelis dantem Salmonea poenas’ we may suppose Lucan derived it. The feeble vulgate punctuation which puts the comma after crudelis, supposed to be vocative, well exemplifies the mischievous influence of propinquity.—I now find the correct punctuation in W. E. Weber's Corpus, but with the needless alteration of crudel e s.

Page 101, line 9: For ‘ille quidem’ read ‘ille quidam.’

〃 101, paragraph 4, line 6: For ‘uiscera sunt’ read ‘uiscera sint.’

〃 102, line 17: For ‘two emendations’ read ‘four emendations.’

〃 102, line 20: For ‘3’ read ‘4.’

〃 102, line 9 from foot: For ‘ears’ read ‘ear.’

The following paper will undertake to demonstrate an I.E. root dyu (alternating with dyu andyu)‘iungere,’ and its synonymous correlatives dyem/dyā (cf.dru drem drā ap. Brugmann, Kvg. 367), dyā-t-/dyat dyes/dyō[u]s (cf. trem/tres, ap. loc. cit.).