I. PROCLIVI SCRIPTIONI PRAESTAT ARDUA: the more difficult the reading the more likely it is to be genuine. It would seem more probable that the copyist tried to explain an obscure passage, or to relieve a hard construction, than to make that perplexed which before was easy: thus in John vii. 39, Lachmann’s addition of δεδομένον to οὔπω ἦν πνεῦμα ἅγιον is very improbable, though countenanced by Cod. B and (of course) by several of the chief versions. We have here Bengel’s prime canon, and although Wetstein questioned it (N. T., vol. i. Proleg. p. 157), he was himself ultimately obliged to lay down something nearly to the same effect(246). Yet this excellent rule may easily be applied on a wrong occasion, and is only true _ceteris paribus_, where manuscripts or versions lend strong support to the harder form. “To force readings into the text merely because they are difficult, is to adulterate the divine text with human alloy; it is to obtrude upon the reader of Scripture the solecisms of faltering copyists, in the place of the word of God” (Bp. Chr. Wordsworth, N. T., vol. i. Preface, p. xii)(247). See Chap. XII on Matt. xxi. 28-31. Compare also above, Vol. I. i. § 11. II. That reading out of several is preferable, from which all the rest may have been derived, although it could not be derived from any of them. Tischendorf (N. T., Proleg. p. xlii. 7th edition) might well say that this would be “omnium regularum principium,” if its application were less precarious. Of his own two examples the former is too weakly vouched for to be listened to, save by way of illustration. In Matt. xxiv. 38 he(248) and Alford would simply read ἐν ταῖς ἡμέραις τοῦ κατακλυσμοῦ on the very feeble evidence of Cod. L, one uncial Evst. (13), _a_ _e_ _ff_, the Sahidic version, and Origen (in two places); because the copyists, knowing that the eating and drinking and marrying took place not in the days of the flood, but before them (καὶ οὐκ ἔγνωσαν ἕως ἦλθεν ὁ κατακλυσμός ver. 39), would strive to evade the difficulty, such as it was, by adopting one of the several forms found in our copies: ἡμέραις πρὸ τοῦ κατακλ., or ἡμέραις ταῖς πρὸ τοῦ κατακλ., or ἡμέραις ἐκείναις πρὸ τοῦ κατακλ., or ἡμέραις ἐκείναις ταῖς πρὸ τοῦ κατακλ., or even ἡμέραις τοῦ νῶε. In his second example Tischendorf is more fortunate, unless indeed we choose to refer it rather to Bengel’s canon. James iii. 12 certainly ought to run μὴ δύναται, ἀδελφοί μου, συκῆ ἐλαίας ποιῆσαι, ἢ ἄμπελος σῦκα; οὔτε (_vel_ οὐδὲ) ἁλυκὸν γλυκὺ ποιῆσαι ὕδωρ, as in Codd. אABC, in not less than six good cursives, the Vulgate and other versions. To soften the ruggedness of this construction, some copies prefixed οὅτως to οὔτε or οὔδε, while others inserted the whole clause οὕτως οὐδεμία πηγὴ ἁλυκὸν καί before γλυκὺ ποιῆσαι ὕδωρ. Other fair instances may be seen in Chap. XII, notes on Luke x. 41, 42; Col. ii. 2(249). In the Septuagint also the reading of א συνεισελθόντας 1 Macc. xii. 48 appears to be the origin both of