But this is not the case with a nearly straight-bladed broad-sword, which requires what may be termed an artificial draw, either backward or forward, in order that the cut may have its full effect. Of course the draw back is by far the most common form of the "draw;" and on reference to the accompanying sketch (Fig. 19) it will be seen that the edge, if the hand retains its position _throughout the entire sweep_, on the circumference of the circle B D, will meet the object to be cut simply as a _hit_, and not as a _cut_. This is just what we want to avoid. [Illustration: Fig. 19.] Suppose the cut is being made parallel to the plane of the paper, and that the hilt of the sword is, in the first part of the sweep, moving on the circumference of the circle from B to D. Suppose, too, that the edge first meets the obstacle to be cut at the point _n_. Then slightly before _n_ is reached the "draw" should commence, the hand coming into position at F, and the point _n_ being necessarily drawn down to _n'_ by the time the object has been severed. That is to say, the portion of the blade between _m_ and _n_ will have been made effective in the drawing cut, the point _n_ having travelled in the direction of the dotted lines till it arrives at _n'_. The point _n_ is taken at random: it might be nearer the hilt or nearer the point, according to the distance of the object aimed at. It may also be observed that the "draw" _might_ continue during the entire sweep from B to F, but a very slight consideration will show clearly the advantage of keeping the arm fully extended until the edge is quite close to the object, as, by this means, the reach is increased and the _power of the cut gains considerably_. The dynamical proof of this latter advantage would take up too much space, and I regret that it is rather outside the scope of this little work. No matter how extended the arm may be when commencing the cut--and the more extended the better in the case of a long heavy sword--the "draw" should always come in towards the end of the sweep, the first part of which is merely intended to give the required impetus to the effective portion of the cut. How is it that an apple or potato can be divided by a straight cut when placed in the folds of a silk pocket-handkerchief, which remains uninjured? Simply because there is a complete absence of "draw," and the apple or potato is broken or split in two, much as the flesh is indented by the edge of the razor whilst the skin escapes without the slightest mark.