246 Barus and Iddings — Electric conductivity observed 



The porphyrite when heated decrepitates, and afterward 

 fuses in the neighborhood of 1400°, but does not become as 

 liquid as the basalt for like temperatures. It solidifies as a 

 dark brown glass with some gas bubbles. 



The rhyolite melts at about 1500°, but remains quite viscid 

 at 1700°, and solidifies as a gray glass filled with minute gas 

 bubbles. 



The amount of gas bubbles serves as an indication of the 

 relative viscosity of the different magmas at the temperatures 

 to which they were subjected. 



§ 4. Relation to temperature. — In making the present de- 

 ductions the outspoken parts of the data of Table I can alone 

 be taken, viz : in case of the magma (1) between 900° and 

 500° ; in case of (2) below 1034°, and in case of (3) between 

 1124° and 692°. Careful scrutiny of the relation of resistance 

 to temperature then shows, that the ratio, dr/dt, of correspond- 

 ing decrements of specific resistance, — dr, and of temperature, 

 dt, is proportional to the corresponding resistance at the given 

 temperature. In other words we have — r' = a + bt, whence 



ln(a + br)=zc—bt (1) 



where a, b, c, are constants. Relative to Table I these values 

 are approximately (r in megohms) 



(1) (2) (3) 



a.. —-10 —-05 +'05 



b '0134 -0125 -0091 



c 7*59 8-78 7-96 



It is probable that a merely reflects the errors of observation 

 and that its real value is zero. The constant b, which is an 

 index of the composition will be discussed in §§ 5, 6. The 

 quantity c/b, which is a temperature, has no apparent meaning. 

 These results are exceedingly interesting since a relation simi- 

 lar to equation (1) was found* for the thermal variation of the 

 viscosity of a highly viscous body like pitch. Again the acid 

 magmas at least, betray no evidences of circumflexure, or any- 

 thing that would be an electrical index of polymerism or melt- 

 ing point. An aqueous solution of zinc sulphate carried 

 under pressure through a large interval of temperature, shows 

 changes of resistancef quite akin to the present. In both 

 cases, therefore, we have similar solution phenomena, exhibit- 

 ing a thoroughly regular change of the electrical character 

 with temperature. 



Moreover the nature of the law embodied in equation (1), 

 for silicates, the enormous range of temperatures to which they 



*Proc. Am. Acad., xxvii, p. 13. 1892. 

 fThis Journal, xlii, p. 135, 1891. 



