158 LANGUAGE OF THE NEW TESTAMENT.
is older), e<a7ra appear first in comedians, and in no
classical Greek is as common as in the N. T. St.
Matthew's a-n-o rore (also Luke xvi. 16) may be
instanced : so 0,71-0 7rpa>t, CCTTO Trepixri (which is unique,
2 Cor. viii. 10, ix. 2), eWaAcu.
When a prep, has more nouns than one depending
on it, the prep, is repeated with each of them more
frequently in the N. T. than in pure Greek. In
Luke xxiv. 27, the second O-TTO almost spoils the
sense of the first the sense is " going on through all
the prophets," and perhaps the repetition adds the
idea of drawing from each, but it would not have
occurred in a pure Greek writer. In 1 Thess. i. 5,
the second lv was of course required after dAAa, but
the third and (si vera I.) fourth have at most a rhe-
torical value. Mark xiii. 32 (true text), 1 Tim.
ii. 9, v. 19, Heb. x. 28, are said to be the only cases
where nouns separated by disjunctive conjunctions
have only one prep, between them; and every one
will see that in all these cases in the two last
especially the repetition would have been impossible,
or have altered the sense. In clauses where there
is a comparison (e.g. Actsxi. 15 CTT' avrovs . . . okrTrep
<' Tj/xas) the prep, is always repeated : always after
an adversative, except sometimes where (as 1 Pet.
i. 23) it is adjectives belonging to one subst. that are
distinguished (so ibid. ver. 1 1 after a disjunctive). In
Acts vii. 4, xx. 18 we get the prep, repeated with
the rel., though by no means the most suitable prep,
to its place in the sentence, by a curious extension
of the principle of attraction. In xiii. 2, 39 we have,
far more classically, the prep, omitted with the rel.,
being understood from the antecedent clause. In