then sent a pedlar, Allan's cousin, back to William Stewart, got
3_l._, added, in the evening of the 16th, more money of his own, and
sent it to Allan. There was a slight discrepancy between the story of
the maid, Kate MacInnes, and that of Donald Stewart, as to what
exactly passed between them, concerning Allan, on the night of the
murder, and whether Allan did or did not give her a definite message
to Donald. The prosecution insisted on this discrepancy, which really,
as James's advocate told the jury, rather went to prove their want of
collusion in the manufacture of testimony. Had their memories been
absolutely coincident, we might suspect collusion--that they had been
'coached' in their parts. But a discrepancy of absolutely no
importance rather suggests independent and honest testimony. If this
be so, Allan and James had arranged no trysting-place on May 11, as
they must have done if Allan was to murder Glenure, and James was to
send him money for his escape.

But there was a discrepancy of evidence as to the hour when the pedlar
sent by James to Fort William on May 15 arrived there. Was he
despatched after the hour when Donald Stewart swore that he gave
Allan's message to James of the Glens, or earlier, with no knowledge
on James's part of the message carried by Donald? We really cannot
expect certainty of memory, after five months, as to hours of the
clock. Also James did not prove that he sent a message to Allan at
Coalisnacoan, bidding him draw on William Stewart for money; yet on
Friday, May 15, James did, by the pedlar, bid William Stewart give
Allan credit, and on Saturday, May 16, Allan did make a pen from a
bird's feather, and ink with powder and water, and write a letter for
money, on the strength of James's credit, to William Stewart. This is
certainly a difficulty for James, since he suggested John Breck
MacColl, a tenant of Appin's at Coalisnacoan, for the intermediary