IS 14.] which serves as a Magnetometer. 435 



I suspend by an oiled silk thread a moderately even glass rod 24 

 inches long. The point of suspension is eight inches from the left 

 end of the rod, so that the other arm of the lever is 16 inches in 

 length, and at the extremity of it is fixed a silver pointer to show 

 the motion of the arm. To the shorter end of the lever is fixed a 

 small leaden weight, secured from oxidation by being varnished, and 

 from the extremity of the longer arm of the lever a small glass 

 scale is suspended by silken threads. 



The we'ght is so regulated that the pointer stands always at 0. 

 (PI. XXV. fig. /•) Hy putting one grain of apothecaries' weight into 

 the glass scale, the pointer sinks a little; and still farther on adding 

 two, three, ibur, &e. grains. The positions of the pointer with each 

 weight in the scale are marked on a circle against which the pointer 

 plays. As the distance between each of these positions is 1^. inch, 

 1 divide it into 100 parts, and thus I obtain a means of determiring 

 the change of weight amounting to -j-J— part of a grain. To prevent 

 the lever from being agitated backwards and forwards, it is made to 

 play between two perpendicular glass rods placed at a. Suppose, 

 now, that I have melted or heated a small piece of matter by means 

 of a fire urged by oxygen gas, and that I wish to ascertain what 

 change has taken place in its weight ; my instrument puts it in my 

 power to do so. In order to determine the magnetic energy of a 

 body, I take a piece of it weighing one grain = 100 of my divi- 

 sions, and put it into the small glass scale. Then, by means of a 

 screw, I apply a magnet capable of supporting one pound of iron, 

 always in a direction exactly under the scale in which the substance 

 to be examined is deposited. When the magnet comes within the 

 sphere of attraction of the body, the pointer usually starts a little 

 up, as, for example, to U5. I now cautiously push the magnet 

 forward with the screw. At last the scale descends towards the 

 magnet. 1 then observe the pointer again. Supposing it to stand 

 at 155, I consider 55 as indicating the magnetic energy of the body; 

 so that in the present case this energy is a little more than equal to 

 half the weight of the body. 1 repeat the experiment six or ten 

 times, and never find more than -j-L- part of difference. 



In making this communication, 1 do not pretend to point out any 

 thing new in the use of the lever, but only to make known a con- 

 venient application of what is already well understood. It will be 

 understood, 1 suppose, that the glass rod employed must not be 

 thicker than a common thermometer tube, and that the t'lass scale 

 itself must be small and light. 



2l2 



