346 Messrs. J. C. McConnel and D. A. Kidd. [June 21, 



in this case, owing to the greater length of the arms, be about twice 

 as great as in the extension experiments, and yet it was found to be 

 scarcely perceptible. For practical convenience in setting up the 

 apparatus the wire was found immensely superior. The trouble of 

 fixing in position a delicate arrangement of brittle glass fibres, in an 

 awkward place like the back of the ice chamber behind the bar of ice, 

 can hardly be realised by any one who has not tried it. 



In the first few experiments the loops m and g were not used, and 

 t)ie indicator and connecting fibres were simply hooked over the 

 needles a and b. And in Experiments 2, 3, 4, and 6, no efficient 

 precautions were taken to prevent slipping along the needle. It is to 

 be remarked, however, that any such slipping would produce an 

 apparent contraction, and, owing to the sudden alteration of the rate 

 of extension, any slipping of importance could hardly escape detection.* 

 Such cases are either omitted or specially mentioned. The lever h 

 and counterpoise were found rather troublesome, and will probably be 

 dispensed with next year, by putting the connecting fibre on the other 

 side of the needle. 



Our second apparatus, which we shall call the rough apparatus, was 

 of much simpler construction. Instead of the collars we used two 

 iron plates, each about 12 cm. square with a hole 2*5 cm. square in the 

 centre. The bar to be tested was passed through the hole and frozen 

 on to a block of ice on the other side of the plate. The upper plate 

 was suspended by cords attached to holes at the corners, and from the 

 lower plate was suspended by similar cords a bucket, in which various 

 weights could be placed. In Experiments 3 and 4, the four cords 

 were simply knotted together, and hung over an iron hook fastened 

 to a single cord. But it was difficult in this way to ensure that the 

 line of action of the tension should be the central line of the bar of 

 ice, and we thought it likely that the bending in Experiment 4 

 was due to this cause, so we adopted the contrivance shown in 

 fig. 4. 



A is the upper iron plate, F the bar of ice attached to the block of 

 ice E.f B is a wooden plate with holes at the corners and a hole at the 

 centre, in exactly the same relative positions as the holes in the corners 

 and the centre of the square hole in A. CCCC are four cords of equal 

 length, and D the main cord by which the whole is upheld. When the 

 arrangement is in equilibrium, the cords C will be vertical as well as 

 the cord D, so the line of action of the tension, which is the central 

 line of the cord D, will pass through the centre of the square hole 

 in A, even though the two plates be not quite horizontal. The same 

 remarks apply to a similar arrangement for the lower iron plate. If 



* In almost every experiment far more readings were taken than are recorded 

 below. 



f This block was thicker than in the figure. 



