138 PROCEEDINGS OF THE AMEllICAN ACADEMY. 



with the ring, may be irradiated. Thus the gas will transpire past not 

 only the inner edge of the ring, but also the outer edge, and yet the 

 whole bulb is not heated. There are then really two concentric annuli, 

 and no space by which difference of pressure can be effaced except 

 through the transpiration spaces themselves. 



Before packing the box with wool, the lamp was placed at its proper 

 distance, about 80 cm., and in such a position that the light from it fell 

 perpendicularly upon the vane. Then a bright metal diaphragm was 

 placed on the glass opening in the door of the box just large enough so 

 that the base of the cone of light from the lamp would cover the desired 

 area. The only part of the bulb illuminated, other than the mere ring 

 contiguous to the outer edge of the mica ring, was, then, that imme- 

 diately within the box and opposite the hole in the diaphragm, so that 

 the heated portion of the bulb was at least 16 cm. away from the vane. 

 The spot of glass irradiated was only about 4 cm. in diameter. To 

 minimize the heating of this by absorption, the light was made to pass 

 through about 2.5 cm. of glass and about 7 cm. of water. The water 

 was kept running from one bottle to another through a water window 

 placed between the lamp and the apparatus. The temperature of the 

 water suffered no perceptible change in the course of an experiment 

 which lasted some three hours. 



Perhaps the greatest difficulty experienced in the mere handling of 

 this apparatus is that caused by the great inertia of the suspended 

 system. It is easily seen that the least jar given to the apparatus in 

 turning the control magnet may give to the system a momentum which 

 is easily many times greater than the force which it is desired to meas- 

 ure. To make conditions as favorable as possible, the platform on which 

 the magnet was placed was finally fastened so that it and the magnet 

 were entirely free from the glass vessel and from the containing box ; 

 but the more or less jerky motion of the armature as it moves under 

 the influence of the magnet still gives the same trouble, though in a 

 much less degree. This difficulty is greatly increased at low pressure ; 

 for when the density is small, the viscosity of the air is so small that 

 the vane, though large, is a very inefficient damper, and so when the 

 suspended system suffers a slight disturbance it moves freely until it 

 strikes one side of the containing vessel ; then it rebounds, strikes the 

 opposite side, and so on. It is true that the suspended system might 

 have been made lighter than it is, but, at best, it must be rather heavy, 

 for the form which it must assume in order that it may be put 

 together is such as to preclude its being made as light as it should be 

 for convenience. 



It is clear also that the equilibrating torsion given to the fibre will 



