388 Allen and Crenshaw — Stokes Method for the 



Further, each sulphide in the mixture must reduce only its 

 proportional fraction of the solution ; in a 50 per cent mixture 

 the pyrite must reduce one half of the ferric solution and the 

 two sulphides must reduce the two portions at the same rate, — 

 otherwise the one which worked faster would encroach on the 

 province of the other and the results woujd not be linearly 

 related. In other words, the results are the same as if one 

 should first separate the two minerals, divide the solution 

 between them in proportion to their percentages, and allow 

 each constituent to reduce its own portion separately. 



The rate at which the solution is reduced by the two sul- 

 phides must then be the same for each mineral, but the rate at 

 which each mineral is decomposed is quite different; that is 

 proportional to the quantity of iron (x — a) which each yields to 

 the solution. This quantity, in terms of the standard perman- 

 ganate solution, is 4*20 for pyrite (Elba No. 1) and 10'57 for 

 marcasite. Marcasite is therefore decomposed 2'5 times as 

 rapidly as pyrite, while the solution is reduced at the same 

 rate by each. 



(b) Electrolytic effect in the Stokes reaction. — From the 

 findings of Gottschalk and Buehler,* and of K. C. Wells,f one 

 would expect an electrolytic effect here. If there is one, it 

 must be very small. The results in fig. 3 do indeed appear to 

 form a slightly convex curve which means of course that the 

 marcasite seems to have a little more than its proportionate 

 effect. Possibly this may be attributed to electrolytic influence. 



(c) Stokes'' curve. — If the relation between x and y is linear, 

 x — by + c. By substituting in this equation the values x and 

 a? 100 for x, when y = and y = 100 respectively, i. e., when we 

 have pure marcasite and pure pyrite, we obtain 



Now substituting this value for x in equation (2) (page 374) 

 the latter becomes 



(4) = (p-frz)a-(p+l$-66)x, 



fc + lMqp^ft) 



or if we replace x , a? 100 and a by their experimental values, viz., 



^oo ah hh ^ 0/./T1 i 10*57 jo — 129*8 -,,. 



4:7-28, 41-17 and 36'71, we have y= Ag1 / , , A1Q . This 



* -0611 p + 1-018 



curve, which is evidently a hyperbola, is plotted in fig. 5. The 



plotted curve is a true hyperbola calculated from the values 



which fall on the line drawn between the values of {x— a) for 



pyrite and marcasite respectively ; the points plotted are the 



values of p calculated directly from experimental values of x. 



* Gottschalk and Buehler, Econ. Geol., vii, 28, 1912. 

 fK. C. Wells, ibid., vii, 571, 1913. 



