SUCTION AL RESPONSE AND ASCENT OK SAl" 36/ 



abscissa the time. The slope of the curve thus gives the 

 rate of movement. As long as the suction is uniform, the 

 slope remains constant. If any exciting agency increases 

 the rate of suction, there is an immediate flexure in the curve, 

 which thus becomes steeper. A depressing agent lessens 

 the slope of the curve. And when suction is abolished, the 

 record becomes horizontal. 



By this arrangement, then, we are enabled, simply and 

 accurately, to obtain a direct and continuous record ; and as 

 the necessity for taking readings is obviated, a large number 

 of experiments can be performed very quickly, with little 

 trouble. The flexure in the curve affords immediate visible 

 indication of the effect of any particular agency. The value 

 of each division of the ordinate is found once for all by 

 determining the volume of unit-length of the potometer-tube. 

 A previous determination at the beginning of the experiment 

 of the rate of movement of the drum, gives us the time- 

 value of each division of the abscissa. Knowing these, we 

 can determine the absolute rate of suction at any period of 

 the curve required. Responsive variations of suction are 

 more easily detected when the normal curve is almost 

 equally inclined to the ordinate and abscissa — that is to say, 

 when it makes an angle of about 45° with either. This is 

 most easily accomplished if we keep the potometer-tube 

 always the same, and merely adjust the speed of the drum. 



The Balanced Shoshungraph.— According to the simple 

 method of making records which has just been described, we 

 observe the responsive effect by means of flexures produced 

 in the curve under the action of various agencies. If the 

 effect of an agency be slight, the change in the slope of the 

 curve will be proportionately small and liable to escape 

 detection. In order to bring the sensitiveness of the instru- 

 ment to its highest, I have devised the Method of Balance, 

 by which the slightest responsive variation is made to exhibit 

 itself in a marked manner. For the purpose of many delicate 

 investigations, it is not so necessary to know the normal rate 

 of suction, as the variations positive and negative in that rate. 



