134 PROCEEDINGS OP THE AMERICAN ACADEMY. 



the magnet which controls the armature, A, Figure 2, and M, Figure 4. 

 A cathetometer was used to determine when the swinging disk in the 

 viscosity apparatus was midway between the other two. As the mica 

 vane could not be well viewed with the cathetometer, its proper ad- 

 justment was judged by its symmetry of position with respect to the 

 surrounding ring of mica. 



Method of Experiment. 



Viscosity Apparatus. — This instrument is held in position at the 

 top by a snugly fitting collar fixed in the top of the enclosing box. 

 It is fixed at the bottom, as shown in Figure 5, so that by giving the 

 arm, L, a slight, slow angular movement, the disk can be set rotating 

 without giving it a serious pendulous motion. This is especially true 

 at the higher pressures, but more care in starting is necessary when 

 the gas becomes rarer. About three quarters of an hour's training is 

 always given to the fibre before observations begin. After allowing 

 everything to become steady, the position of rest for every swing is 

 observed, and when the damping is rapid about seven or eight swings 

 is all one can get for any one start given to the disk. Sufiicient ac- 

 curacy is obtained by using the results got from starting the disk four 

 times. This applies only to the work at the higher pressures ; as will 

 be seen, the plan is changed when the density diminishes, for then many 

 more arcs can be obtained before the amplitude of swing becomes too 

 small. 



In getting the mean decrement, the logarithms of the fifth, sixth, 

 seventh, and eighth arcs are taken respectively from the first, second, 

 third, and fourth, and the result in each case divided by four. At the 

 lower pressures, where many more arcs can be obtained in a series, and 

 where as a consequence the error in observing any one is increased, 

 the eleventh is taken from the first, the twelfth fi-om the second, and 

 so on, and the results divided each by ten. The decrement used is 

 then got by taking the mean of some sixteen such decrements. 



Transpiration Appat-atus. The source of light used for the purpose 

 of illuminating the blackened face of the mica vane and annulus was 

 a twenty-five candle power incandescent lamp, suited to a voltage of 

 forty. The current used was that from a storage battery. 



The zero position of this instrument is that position of the armature, 

 M, Figure 4, which allows the vane to hang in the plane of the annulus, 

 when there is no irradiation of any part of the apparatus. To read 

 this position, there is a pointer attached to the magnet which controls 

 the armature. The pointer coincides with a radius of a graduated 

 circle placed on the upper magnet platform, concentric with the tube 



