[1] _Hist. MSS. Com._ XIII. ii. 85. It is from a transcript of this copy
made for Dr. Gardiner that I have been permitted to take the text below.
A set of 'Instructions for the better ordering of the fleet in Sailing'
accompanies them.

[2] _British Museum, Shane MSS._ 3232, f. 81.

[3] The Sloane copy is not quite identical with that in the Portland
MSS. The variations, however, are merely verbal and in a few signals,
and are of such a nature as to be accounted for by careless
transcription.

[4] Hoste, the author of the first great treatise on Naval Tactics,
quotes Tromp's formation as a typical method of retreat; but his account
is vitiated by what seems a curious mistake. He says: 'Il rangea son
armée en demi-lune et il mit son convoi au milieu: c'est à dire que son
vaisseau faisait au vent l'angle obtus de la demi-lune, et les autres
s'étendoient de part (_sic_) et d'autre _sur les deux lignes du plus-
près_ pour former les faces de la demi-lune qui couvroient le convoi. Ce
fut en cet ordre qu'il fit vent arrière, foudroiant à droite et à gauche
tous les anglois qui s'approchent' But if with the wind aft his two
quarter lines bore from the flagship seven points from the wind, the
formation would have been concave to the enemy and the convoy could not
have been _au milieu_. (_Evolutions Navales_, pp. 90, 95, and plate 29,
p. 91.) The passage is in any case interesting, as showing that what was
then called the crescent or half-moon formation was nothing but our own
'order of retreat,' or 'order of retreat reverted,' of Rodney's time. As
defined by Sir Charles Knowles in 1780, the order of retreat reverted
was formed on two lines of bearing, _i.e._ by the seconds of the centre
ship keeping two points abaft her starboard and larboard beams