April 10, 1908] 



SCIENCE 



571 



weak that the sugar added can not be de- 

 tected by its taste, and the arrangement is 

 the simplest way to avoid an otherwise un- 

 certain and troublesome correction. 



The Resistance Temperature Coefficient 

 and the Coefficient of Expansion of 

 Carion: G. W. Stewart, University of 

 North Dakota. 



With the exception of the diamond and 

 charcoal, carbon in its various forms eon- 

 ducts electricity metallically. Its resist- 

 ance temperature coefficient is, however, 

 negative. This paper accounts for the ap- 

 parent negative coefficient by assuming the 

 changes in resistance to be due to the ex- 

 pansion of the particles of carbon. 



Assuming this explanation to be the cor- 

 rect one, experiments were performed to 

 obtain the coefficient of expansion of car- 

 bon. The carbon used was in the form of 

 films made of commercial lampblack and a 

 lacquer called "zapon L." The apparent 

 resistance temperature coefficient of the 

 films and the effect of the expansion of the 

 hard rubber base upon which the films 

 were placed were measured, and the actual 

 coefficient of expansion of the carbon par- 

 ticles was computed. The result obtained 

 shows the coefficient to be about 0.0003, 

 which is from ten to thirty times that of 

 the pure metals. 



The Temperature Coefficient of the Moving 

 Coil Galvanometer: Anthony Zeleny 

 and 0. HovDA, University of Minnesota. 

 The temperature coefficient of a moving 

 coil galvanometer having a cast-iron mag- 

 net, was determined in order that the 

 change in the sensibility could be calculated 

 when the instrument is used at different 

 temperatures. 



Since the temperature coefficient depends 

 upon the magnet, the suspensions and the 

 coil, and in different relations for different 

 kinds of measurements, the coefficients of 



these different parts were determined sepa- 

 rately. These are combined in their proper 

 relation to determine the temperature co- 

 efficients for current, potential and ballistic 

 measurements. 



The Effect of Tension on Thermal and 



Electrical Conductivity: N. F. Smith, 



Olivet College. 



Two metal bars called A and B of the 

 same material, each %6 of an inch in diam- 

 eter and about one meter long, are mounted 

 horizontally about 10 cm. apai't. One end 

 of each bar is held in a clutch made from 

 a heavy block of copper which is heated 

 and maintained at a constant temperature. 

 By means of thermo-electric couples sliding 

 on the bars a point is determined on B 

 which has the same temperature as a given 

 point on A when the steady state is reached. 

 B is then subjected to a stretching force 

 while the condition of A remains un- 

 changed. When the steady state is again 

 reached the couple on B is shifted till it is 

 again at a point where the temperature is 

 the same as at the given point on A. The 

 stretching force is increased, step by step, 

 up to the maximum which the bar wiU 

 withstand. It is assumed that the thermal 

 conductivity is proportional to the square 

 of the length to the position of the thermo- 

 electric couple. At each step the electrical 

 resistances of the two bars are compared by 

 a modification of Kelvin's double bridge 

 method. 



Observations have been made on bars of 

 several different metals and each shows an 

 increase in the thermal conductivity with 

 the stretching force. The maximum in- 

 crease is about 1.7 per cent. At the same 

 time the electrical conductivity diminishes, 

 the variation being about the same as that 

 found by other experimenters. The length 

 of time that the stress is applied has a 

 marked effect upon the thermal conduct- 

 ivity. 



