A careful determination of g by means of the ordinary type of Atwood's machine does not, as a rule, lead the average student in a Physical Laboratory to a better result than 930 or 940 cm/sec2. From the point of view of successful teaching, it is somewhat unfortunate that, after bestowing reasonable care and attention to his work, a student should be unable to obtain a result approximating satisfactorily to what he knows to be the correct figure. Not unnaturally he takes it for granted that the actual numerical result obtained from his experimental labours is quite immaterial as long as the processes involved are clearly comprehended, and to him Experimental Physics is anything but an exact science. On the other hand, to set before the ordinary student a complicated apparatus specially designed for reaching an accuracy of 0·1 per cent, would be proceeding to the other extreme, and one could hardly expect much benefit to be derived from its use.