ing to gain the wind of the enemy, for the headmost
and sternmost ships to signify when they can weather them by Signal
17, p. 66; or if to windward of the enemy and on the contrary tack,
for the sternmost ship to signify when she is far enough astern of
their rear to be able to lead down out of their line of fire.'

'No. 234. When coming up astern and to windward of the enemy to engage
by inverting the line'--that is, for the ship leading the van to
engage the sternmost of the enemy, the next ship to pass on under
cover of her fire and engage the second from the enemy's rear, and so
on.

FOOTNOTES:

[1] The first attempt to provide a convenient Signal Book separate
from the Instructions was made privately by one Jonathan Greenwood about
1715. He produced a small 12mo. volume dedicated to Admiral Edward
Russell, Earl of Orford, and the other lords of the admiralty who were
then serving with him. It consists of a whole series of well-engraved
plates of ships flying the various signals contained in the Sailing and
Fighting Instructions, each properly coloured with its signification
added beneath. The author says he designed the work as a pocket
companion to the Printed Instructions and for the use of inferior
officers who had not access to them. Copies are in the British Museum
and the R.U.S.I. Library.

[2] _Catalogue_, Nos. 252/27 and 252/26.

[3] A still earlier Signal Book attributed to Lord Howe is in the
United Service Institution, but it is no more than a condensed and
amended form of the established one. Its nature and intention are
explained by No. 10 of the 'explanatory observations' which he attached
to it. It is as follows; 'All the signals contained in the general
printed Signal Book which are likely to be needful on the present
occasion being provided for in this Signal Book, the signals as
appointed in the general Signal Book will only be made either in
conformity to the practice of some senior officer present, or when in
company for the time being with other ships not of the fleet under the
admiral's command, and unprovided with these particular signals.' It was
therefore probably issued experimentally, but what the 'present
occasion' was is not indicated. It contains none of the additional
signals of 1782-3.

[4] Knowles was of course too old in 1830 for his memory to be trusted
as to details. A note in his handwriting upon a copy of his code in
possession of the present baronet gives its story si