most varied and intricate kind,
and that at present the nature of the facts themselves is in many cases
very far from being accurately ascertained. We are constantly reminded
in reading 'Supernatural Religion,' able and vigorous as it is, how
much of its force depends rather upon our ignorance than our knowledge.
It supplies us with many opportunities of seeing how easily the whole
course and tenour of an argument may be changed by the introduction of
a new element. For instance, I imagine that if the author had given a
little deeper study to the seemingly minute and secondary subject of
text-criticism, it would have aroused in him very considerable
misgivings as to the results at which he seemed to have arrived. There
is a solidarity in all the different departments of human knowledge and
research, especially among those that are allied in subject. These are
continually sending out offshoots and projections into the neighbouring
regions, and the conclusions of one science very often have to depend
upon those of another. The course of enquiry that has been taken in
'Supernatural Religion' is peculiarly unfortunate. It starts from the
wrong end. It begins with propositions into which _a priori_
considerations largely enter, and, from the standpoint given by these,
it proceeds to dictate terms in a field that can only be trodden by
patient and unprejudiced study. A far more hopeful and scientific
process would have been to begin upon ground where dogmatic questions
do not enter, or enter only in a remote degree, and where there is a
sufficient number of solid ascertainable facts to go upon, and then to
work the way steadily and cautiously upwards to higher generalisations.

It will have been seen in the course of the present enquiry how
many side questions need to be determined. It would be well if
monographs were written upon all the quotations from the Old
Testament in the Christian literature of the first two centuries,
modelled upon Credner's investigations into the quotations in
Justin. Before this is done there should be a new and revised
edition of Holmes' and Parsons' Septuagint [Endnote 359:1].
Everything short of this would be inadequate, because we need to
know not only the best text, but every text that has definite
historical attestation. In this way it would be possible to arrive
at a tolerably exact, instead of a merely approximate, deduction
as to the habit of quotation generally, which would supply a
fi