220 Journal of Philology. no more introduce a finite concessive sentence than ai or nep separately. It is to be remarked that koX in this combination signifies " even" not " and/' and the practice of separating the two particles, as in Kal paa nep KexoXauevos, and the like, shows that the concession is not in the particles themselves, but in the par- ticiple. In point of fact, although concessive sentences are of perpetual occurrence in Greek, as in other languages, there are only three examples, so far as I know, where the existing text exhibits Kai irep with a finite verb. And as the corruption in each case is obvious and the remedy easy, really good scholars in this country will be surprised to hear that the most recent Greek grammarians in Germany, and certain Englishmen who pin their faith on the Germans, persist in teaching that nal irep may, though rarely, be used with the finite verb! The three passages to which I refer are, (a) Pindar Nem. TV. 36: epna <ai irep exei j3a0eta irovrias akpa peacrov ; (6) Plato Syitip. p. 219 C : Kai irep eKeivo ye aprjv n ehai, a uvbpes diKaarai' StKaoral yap eore rfjs 2a>Kparovs vireprjcpapias. (c) Theophrastus Charact. C. II. I Kai irep c'tis Kal aXkos exeis ""P * Ta ^ T V peKaivav ttjv rpia. To begin with the last, as the most recent and least important of the three passages, we have obviously a corruption for Kai rot tbnp ns Kal 5os <x* k.t.X., the ir(p having left its usual place in this construction. In the passage of Plato also (which is omitted in the Vatican MS.) the position of ye shows that Kai irep here has taken the place of Kai rot, which is constantly followed by ye, as in Eurip. Orest. 77 I Kai rot areva ye rfjs YXwatuvrjorpas uopov. Plato Phcedo, p. 68: /caf roi <papev ye abvvarov elvai. Besides, the passage is corrective rather than merely concessive, and therefore Kai rot is the better combination of particles. And so entirely is irep in the concessive construction limited to the participle, that even when it is added to Kai rot, which generally takes the finite verb, the participial construction follows ; as in Herod, vm. 53 : Kai roi irep ajroKprjpvov e6vrot rov apiov. Such being the case, it may seem surprising that, like all the other editors of Pindar, I allowed Kai irep to stand in the passage quoted above, without any remark, and without suggesting the correction, which I be- lieve to be as certain as it is necessary : epira Ketirep exei, as in the passage from Soph. Ajax 563 which I have quoted in my note : clokvov epna, Kel ravvv rqKoiirbs oi^vct. The fact is that when I was writing my notes on Pindar some 14 years ago, I was consulted
