432 BEPOET— 1892. 



in principle of a plate, A. On this moves, in the direction of the radius, 

 a roller, C. It is quite evident, if we make the plate A turn, the number 

 of turns made by the roller will be proportional to the number of turns 

 made by the plate A. 



' If we suppose the curve to be planimetrical, mounted on a rack, 

 F, working with a pinion, B, fixed on the same axis as the plate, while 



Fis. 36, 



,'B ••,flc\ ° 



=iisiiia 



from the other side we guide the roller by means of a lever, G, so that the 

 extremity, H, of this lever is on the axis of x when the roller is at the 

 centre of the plate, then we shall displace the curve while following the 

 turning of the disc and roller. 



' The plate A will turn as a function of the abscissae x x. The distance 

 from the centre of the plate to the i-oller will be at each moment pro- 

 portional to the different ordinates of the curve ; consequently, if we make 

 a wheel move endlessly with the screw D on the axis of the roller, the 

 product will be the sum of all the differentials, 



edx. 



' In practice the rack is replaced by a toothed wheel or a crank, and 

 the plate is composed of a system of two wheels turning in a contrary 

 way, compressing between them the roller. 



'At the same time, for greater convenience, the curve is fixed on a 

 cylinder similar to all those we make. The advantage over all other 

 planimeters is that, should the roller produce an error in making a great 

 number of turns, it is absolutely due to negligence, and the apparatus 

 gives the reading directly in square millimetres, the approximation being 

 at least one-millionth of the total value.' 



In order to use the instrument the paper is fixed upon the drum, a 

 knob is held in the right hand and pressed against the flange at the 

 bottom part of the cylinder ; then the crank is turned until one of the 

 ordinates, passing through any point on the diagram, comes under 

 the point with the knob. The starting-point of the diagram is then 

 marked by means of a pencil point, and after bringing the point to this 

 mark, the crank is turned until a starting-point is again reached. A 

 guiding pointer is then brought against the flange, and the reading of 

 the index gives the area of the diagram. The writer has not been able 

 to obtain one of these instruments so as to actually test them, and makes 

 the above statement solely on the high authority of the makers ; but it is 



