(3) While lying on the back, raising the legs from the floor to a vertical position and lowering them again, repeating to the point of physical exhaustion. (4) Raising a 5-lb. dumb-bell (with the triceps) in each hand from the shoulder up to the highest point above the head, repeating to the point of physical exhaustion. (5) Holding the arms from the sides horizontally for as long a time as possible. (6) Raising a dumb-bell (with the biceps) in one hand from a position in which the arm hangs down, up to the shoulder and lowering it again, repeating the motion to the point of physical exhaustion. This test was taken with four successive dumb-bells of decreasing weight, viz., 50, 25, 10 and 5 lbs. respectively. (7) Running on the gymnasium track at a speed to suit the subject, to as great a distance as possible. The mental test consisted of adding specified columns of figures as rapidly as possible, the object being to find out whether the rapidity of performing such work tended to improve during the experiment. PERCENTAGE OF IMPROVEMENT IN ENDURANCE (EXACT OR UNDERSTATED) OF EIGHT MEN. AVERAGE. B Lq. Lw. M P R T W Jan.-Mar. 33+ 36 50 -- 26 18+ 66+ 33 Jan.-June 84+ 84+ 181 29 56+ 89+ 80+ 107+ The figures of this table show an undoubted increase in endurance, both for the first half and more especially for the whole period of the experiment. Three methods of estimating the increase of endurance between January and June were used. These may be put together in the following table: PERCENTAGE OF INCREASE OF ENDURANCE, JANUARY TO JUNE, BY THREE METHODS. AVERAGE SIX TESTS. B E Lq. Lw. M P R T W 85 13 194 95 212 56+ 73 66 109 OMITTING DOUBTFUL CASES "+" 84+ ... 84+ 181 29+ 56+ 89+ 80+ 107+ "PURE" ENDURANCE OF BICEPS. ... ... 62 ... 50 ... 170 200 100+ The first line of this table tells us the average of the recorded improvement in endurance shown for each man. The average of these averages is 101 per cent. for the entire club, and is probably within the truth; for most of the individual figures which go to make up this result are understatements, not overstatements. The second line shows the average improvement in tests in which there is no doubt that the figure is at least not too high, though it may be too low. The average of these is 89 per cent., and is therefore certainly too low an estimate of the average improvement for the eight men who improved at all. The third line shows the increase of _pure_ endurance (that is, endurance considered apart from strength) for the five men for whom the figures were available. The average of these is 116 per cent. We are quite safe in saying, therefore, that the average improvement of the eight men who improved was 90 per cent.