A few remarks may be offered upon the various methods of fastening the flag to its support. In the earliest times it seems to have been nailed to its staff, and while it was small and easily portable this entailed no inconvenience. Afterwards it became customary to form a socket of canvas, buckram, or other stout material to which the hoist was sewn, and which was slipped over the staff from the top. This method was only convenient for the smaller flags, and with larger ones it appears to have been usual to sew a band of canvas down the hoist and attach a number of ribands to this by which the flag was tied to the staff. Finally, the flag was sewn to a piece of rope which could be made fast at the ends to halliards running through a block or sheave in the cap at the top of the staff, but at first these halliards were, in the case of masthead flags, taken only down to the top, not to the deck, so that a sailor had to ascend the mast whenever the flag was taken down or put up, or when any signal had to be made aloft. Thus, during the battle with the Dutch off Lowestoft on 3rd June, 1665, the Duke of York ordered the signal to be given for the whole fleet to tack, but the sailor who had got up the mast to give the signal was so long about it that before he could let the flag fly Opdam had with his van bore up round.... This little accident lost above six hours. No doubt it was "accidents" of this nature that led, at a later date, to the introduction of longer halliards reaching down to the deck[412]. After the introduction of the driver or spanker boom about 1790 the ensign staff on the poop had to be removed in three-masted ships when the ship was under sail, and the ensign was then hoisted at the "peak" or outer end of the gaff by which this spanker sail was supported, the halliards passing from the poop through a small block made fast at the extremity of this spar. This continued to be the position for the ensign when the ship was at sea until the abolition of sails brought it again to its original position.