178 MINNESOTA BOTANICAL STUDIES. 



The electro-magnetic balancing mechanism consists of a 

 weight holder and an electro- magnet, together with the contact 

 point on the scale beam, mercury cup, wiring, batteries and the 

 necessary fittings. 



The weight holder is a coiled brass tube that holds about one 

 hundred and twenty-five weights. At the lower end of this 

 brass coil is a lever that can turn back and forth on pivot. 

 One end of this lever is connected by a link to the armature of 

 the magnet, and the other end, which is held in place by a 

 spring, when the circuit is open, has a weight pocket that takes 

 one weight from the weight tube each time the circuit is closed 

 and carries it laterally about five-sixteenths of an inch and lets 

 it drop, through a hole in the brass plate, on to the scale pan. 

 As soon as the circuit is opened again by the readjustment of 

 the scale beam the lever returns to its position and receives 

 another weight from the tube, and is again ready to drop it 

 into the scale pan as soon as the necessary increase in weight 

 to close the circuit at the other end of the beam has taken 

 place. 



The weight holder has a calibre one-sixteenth of an inch 

 larger than the diameter of the weights used. It is screwed to 

 the frame of the electro-magnet and extends upward and to 

 the outside of the case for the reception of the new weights at 

 its exterior end. It is made air and water tight from the ex- 

 terior by means of a rubber stopper that fits into the case. 

 The weight holder can therefore be taken out and replaced by 

 one of greater or less calibre, depending upon the size of the 

 weights used, thus if two tubes, five-sixteenths and one-fourth 

 inch respectively, are used, two sets of weights can be put in, 

 viz. : one-fourth inch weighing about one gram, and three-six- 

 teenth inch weighing about one half a gram. Larger or 

 smaller sizes could be used, but for growth or transpiration the 

 above sizes are sufficiently delicate to give a good curve. 



The weights used are steel balls, the same make and size as 

 are used in bicycle bearings. These are perfectly accurate, 

 not varying in diameter more than one two thousandths of an 

 inch, and in weight on an average not more than one thous- 

 andth of a gram. Sets of these balls can of course be weighed 

 and verified by the experimenter himself. 



The electro-magnet has a single coil, and one end of the core 

 is joined with the frame of the magnet which forms the return 

 magnetic circuit. The other end of the core is contracted in the 

 form of a paraboloid. 



