perhaps what we should expect: in longer quotations it would be
better worth the writer's while to refer to his cumbrous
manuscript. These purely mechanical conditions are too much lost
sight of. We must remember that the ancient writer had not a small
compact reference Bible at his side, but, when he wished to verify
a reference, would have to take an unwieldy roll out of its case,
and then would not find it divided into chapter and verse like our
modern books but would have only the columns, and those perhaps
not numbered, to guide him. We must remember too that the memory
was much more practised and relied upon in ancient times,
especially among the Jews.

The composition of two or more passages is frequent, and the
fusion remarkably complete. Of all the cases in which two passages
are compounded, always from different chapters and most commonly
from different books, there is not, I believe, one in which there
is any mark of division or an indication of any kind that a
different source is being quoted from. The same would hold good
(with only a slight and apparent exception) of the longer strings
of quotations in cc. viii, xxix, and (from [Greek: aegapaesan] to
[Greek: en auto]) in c. xv. But here the question is complicated by
the possibility, and in the first place at least perhaps
probability, that the writer is quoting from some apocryphal work
no longer extant. It may be interesting to give one or two short
examples of the completeness with which the process of welding has
been carried out. Thus in c. xvii, the following reply is put into
the mouth of Moses when he receives his commission at the burning
bush, [Greek: tis eimi ego hoti me pempeis; ego de eimi
ischnophonos kai braduglossos.] The text of Exod. iii. 11 is
[Greek: tis eimi ego, oti poreusomai;] the rest of the quotation