4. One of the most striking of them regards what is called ν ἐφελκυστικόν, the “ν attached,” which has been held to be an arbitrary and secondary adjunct. This letter, however, which is “of more frequent occurrence at the end of words, is itself of such a weak and fleeting consistency, that it often becomes inaudible, and is omitted in writing” (Donaldson, Greek Grammar, p. 53, 2nd edit.). Hence, though, through the difficulty of pronunciation, it became usual to neglect it before a consonant, it always comprised _a real portion of the word to which it was annexed_, and the great Attic poets are full of verses which cannot be scanned in its absence(324): on the other hand, the cases are just as frequent where its insertion before a consonant would be fatal to the metre. In these instances the laws of prosody infallibly point out the true reading, and lead us up to a general rule, that the weak or moveable ν is more often dropped before a consonant than otherwise. This conclusion is confirmed by the evidence of surviving classical manuscripts, although but few of them are older than the tenth century, and would naturally be conformed, in such minute points, to the fashion of that period. Codices of the Greek Testament, and of the Septuagint, however, which date from the fourth century downwards, present to us this remarkable phenomenon, that they exhibit the final ν before a consonant full as often as they reject it, and, speaking generally, the most ancient (e.g. Evan. אABCD)(325) are the most constant in retaining it, though it is met with frequently in many cursive copies, and occasionally in almost all(326). Hence arises a difficulty, on the part of modern editors, in dealing with this troublesome letter. Lachmann professes to follow the balance of evidence (such evidence as he received) in each separate case, and, while he usually inserted, sometimes omitted _nu_ where he had no cause for such inconsistency except the purely accidental variation of his manuscripts; Tischendorf admits it almost always (N. T., Proleg. p. liii, 7th edition), Tregelles (I think), as also Westcott and Hort, invariably. Whether it be employed or not, the practice should at any rate be uniform, and it is hard to assign any reason for using it which would not apply to classical writers, whose manuscripts would no doubt contain it as often as those of the N. T., were they as remote in date(327). The same facts are true, and the same remarks equally apply to the representing or withdrawing of the weak ς in οὕτως before a consonant. Each of the aforenamed editors, however, for the sake of euphony, prefers οὕτω before σ at the beginning of the next word, except that Tregelles ventures on οὕτως σε δεῖ in Acts xxiii. 11. Cod. א has οὕτω about fourteen times in the N. T.