ties affecting particular books or writers, or collocations of letters or words, and the like, has often shown that an exceptional smallness or largeness of evidence has been probably due to accident. On the other hand it would be unreasonable to assume that the same writer, even in the same book, always spells a word in the same way. Absolute uniformity belongs only to artificial times; and, after full allowance has been made for anomalies of evidence, the verdict of MSS is decisive against the supposition. Absolute uniformity therefore we have made no attempt to carry out, even within narrow limits; while we have assumed the existence of such a moderate or habitual uniformity in the usage of the writers as would enable us to come to a decision for the text in difficult cases. Many ancient spellings are therefore adopted in individual places on evidence which might be perilously small if they were taken alone, and if substantive readings were in question; but we have printed absolutely nothing without some good documentary authority.
402. In some departments of orthography the evidence is so unsatisfactory that the rejected spellings are but little less probable than those adopted; and thus they should in strictness be accounted alternative readings. But to have printed them in the margin along with the substantive alternatives would have crowded and confused the pages of our text beyond measure, without any corresponding gain. They are therefore reserved for the Appendix, in which a few additional remarks on some special points of orthography, especially on some forms of proper names, may fitly find a place. The alternative readings thus relegated to the Appendix under the head of orthography include not only forms of inflexion, but forms of particles, as ἄν or ἐάν, and variations in the elision or retention of the last vowel of ἀλλά and of such prepositions as end with a vowel. We have ventured to treat in the same manner variations of the indicative or subjunctive after such particles as ἵνα, ἐάν, and ὅταν, and after relatives with ἄν or ἐάν.
403. A word may be interposed here on a topic which in strictness belongs to Part III (compare § 303), but which it is more convenient to notice in connexion with orthography. Attention was called above (§ 399) to the necessity of making allowance for purely itacistic error in considering the properly orthographical testimony of MSS. But there is another more important question con-