OCTOBEE 25, 1907] 



SCIENCE 



555 



lator, both of which are represented in the 

 accompanying figures. Their construction 

 will be clearly understood by referring to the 

 figures. 



fiQ.^- 



Figs. 1 and 2. Side view of a median vertical 

 secticm of mercury and glycerin electric thermo- 

 reg-ulators respectively. One half natural size. 

 General outline, wall of glass tube; a, mercury 

 or water; 5, cork; c and d-e, copper or platinum 

 wire about 0.5 mm. in diameter; /, solid glass 

 rod; m, mercury; g, glycerin, air, toluole, or 

 chloroform. 



Fig. 3. An enlarged view of the lever d-e, with 

 its connections. 



In constructing the lever the ring at d must 

 be made large enough, so that the end e, 

 where the circuit is broken, will fall with 

 its own weight. This ring should consist of 

 several coils so as to make the ring end of 

 the lever heavier than the opposite end and 

 give space enough for good electrical contact. 

 Five coils, two on either side of the middle, 

 is sufficient. I have also found it advisable 

 to make a single turn loop in the lever near 

 the middle, to connect with the ring on the 

 wire c. This insures bearing always on the 

 same point and prevents possible change in 

 the adjustment of the regulator by the sliding 

 of the ring on the lever. 



Copper wire has proved very satisfactory in 

 the construction of the lever and its connec- 

 tions. I have, however, had occasion to use 

 it only about one week. It may be that con- 

 tinuous usage for a long period will cause 

 sufficient oxidation at the point e where the 

 circuit is broken, to prevent electrical con- 

 tact. If this proves to be true, platinum will 

 have to be substituted for copper. 



If the cork 6' is cut in half crosswise and 

 a little vaselin put between the two parts as 

 represented in the figures, the glass rod will 

 slip through the cork very easily and there 

 will be no difficulty with leakage. 



The glycerin regulator will be more compact 

 and easier to manipulate if the second bend 

 in the glass tube is made in a plane at right 

 angles to that in which the first bend lies, and 

 the third in a plane at right angles to the 

 second, so that the two portions i and h will 

 lie near each other and a cross-section of the 

 four straight portions will form a rectangle. 



If glycerin and mercury are used in the 

 regulator, represented in Fig. 2, it can be 

 easily filled by first nearly filling the entire 

 tube with glycerin, then pouring the desired 

 quantity of mercury into the lever end, and 

 then drawing the glycerin above the mercury 

 off and washing it out thoroughly with water 

 and alcohol. The space above the mercury 

 must be clean and dry or the cork will be 

 likely to stick. 



Both regulators are adjusted in the same 

 way. They are immersed in the water in the 

 bath, then the glass rod f is partially or en- 

 tirely withdrawn. This completes the circuit 

 and the temperature begins to rise. As soon 

 as it has reached the desired point, the glass 

 rod is pushed in very slowly until the circuit 

 is broken. The temperature now falls until 

 the circuit is again complete. The adjustment 

 is thus seen to be very simple. It must, how- 

 ever, be remembered that owing to currents 

 -in the water due to unequal heating, the tem- 

 perature of the water in the region of the 

 thermometer compared with that in the region 

 about the regulator is likely to change and so 

 it is to be expected that the regulator will not 

 maintain the temperature recorded when the 



