In either case when the sound of the pistol is heard or the first motion noticed, the timekeeper has to arrive at a determination to start the watch from the sensations of his eye or ear; he has then by will to convey that determination to-the muscles of his finger, and the watch is then started by him. Recent experiments in physiology have shown not only that the process of thinking and volition takes an appreciable time, but that there is also an appreciable difference between the rapidity of thought and volition in different individuals. Practice and experience, as in everything else, tend to quicken the rapidity of the action of the will or thought, which travels quicker over beaten tracks. We of course wish to steer clear of physiological discussion as far as possible, and to confine ourselves to practical experience of timekeeping; but it is necessary to point out that the individual qualities of timekeepers may make the results they arrive at untrustworthy. As a general proof of the truth of what we say, we may call attention to the fact that novices who try to time races invariably make the times 'fast,' i.e. they are very slow in starving the watch, although they are not so slow in stopping it at the moment the tape is reached, as by watching the runner up to the tape they know almost the exact moment when the pressure of the finger to stop the watch will be required.

Before quitting the subject of the liability of the timekeeper to error, we would suggest that there should be a definite rule laid down by the A.A.A. for the guidance of starters as to what point should be taken for the start - the flash of the pistol, the report, or the first motion of the runner. Most timekeepers profess to start from the motion of the runner's body, but we are inclined to think with some of them this is a mere profession. Obviously they must only look at one runner, and if he were left upon the mark they would be 'out' altogether, a fact which we never yet heard an official timekeeper admit. In any case the starting from the motion of the body is fallacious, as one runner usually begins to move perceptibly before the others.

Besides these complications, there is the further set of difficulties created by the watch itself, though with the splendid pieces of mechanism which are now made expressly for this purpose the uncertainty is minimised. The dial of the stopwatch is usually marked into fifths, and with some of the older watches there was this fact to be considered: the hand 'jumped' from fifth to fifth round the dial, there being of course a fifth of a second between each jump. Obviously, therefore, there might be almost a whole fifth of difference between two occasions when the same time was registered on the dial. For instance, upon one occasion when 10 1/5 was registered on the dial, the hand might, just at the instant before the watch stopped, have made its jump to the fifth, while upon the other occasion the hand might have stopped at an infinitesimal period before it took its jump to the two-fifths mark upon the dial. It is clear, therefore, that, as a good sprinter travels a shade over two yards in a fifth of a second at the end of a hundred yards race, two men might each have been fairly timed at 10 1/5 seconds in a race, although the one was two yards in front of the other.

As for a very long time there were nothing but watches which jumped fifths even at first-class meetings, it is obvious that as regards the times made at these meetings, they are quite untrustworthy as to one-fifth of a second's variation. At the present day there are some watches (although there are not many belonging to private individuals, their cost being large) which, at any rate as far as the human eye can distinguish, travel evenly over the dial. As a matter of fact, the wheel which regulates the motion has such small cogs that the jumps are divided into twenty or thirty per second. The hand, therefore, can really be stopped at any place on the dial between the fifths, and the dial then has to be surveyed through a magnifying-glass, and a conclusion arrived at how far it has travelled between one fifth and another.

All this process, however, though exceedingly wonderful and the result of admirable workmanship, is also liable to error. The slightest warp of the straightness of the hand between the fifths may lead to doubt, and even supposing the mechanism perfect the human eye can and does make mistakes in deciphering the result which the machine has registered. The watches of which we speak are of course so made that the 'seconds' hand travels round the whole of the dial; but the watch is made to be carried in the pocket, so that there is very little space between the fifths of seconds on the dial, as each fifth occupies 1/300 of the circumference of the dial. It is thus obvious that the human eye can easily be deceived in reading the dial, and it must be recollected that as soon as the next race comes on the watch has to be got ready again for this, so that it becomes impossible to verify afterwards the result registered on the dial. On account of the difficulty of reading the results on the watch many timekeepers carry a magnifying glass with them, and although this makes the reading easier and perhaps more accurate, the whole process becomes still more complex.

On the whole, therefore, it appears that there is every possibility of a mistake being made to the extent of a fifth of a second, or even rather more, in the timing of races; and some years ago when timekeepers had less experience and timing instruments were inferior, there was even a greater possibility of error than there is now. We cannot help thinking how strange it is that in these days of science no more satisfactory method of timing races has been invented than that of a watch started by a man who observes the start from some distance off. In these days of electric science it seems to the unscientific mind of the present writer that it ought not to be difficult to time by an apparatus, which could be fixed on every recognised running-ground, set in motion by the firing of the pistol, and stopped by the breaking of the thread at the winning-post. In such an apparatus the difficulty of deciphering the result marked on a small dial need not occur, as the hand or hammer registering the result might work upon a dial of any size.