HALL. — THERMO-ELECTRIC HETEROGENEITY IN ALLOYS. 549 



means of a formula furnished us by Professor Crafts, which formula 

 has not yet been published. 



Observations. 



Observations were sometimes made with junctions 1 (see Figure 1) in 

 the naphthalin pot and junction 2 in the steam pot, and sometimes with 

 the opposite arrangement. The first was called " the parallel " arrange- 

 ment ; the second, the " crossed " arrangement. A complete run for 

 any one day included a set of observations made on each couple, with 

 the parallel arrangement, and another similar set made with the crossed 

 arrangement, the temperatures being kept, as nearly as might be, 

 constant during the whole run. 



If on one day the observations were taken first with the " parallel " 

 arrangement and then with the " crossed " arrangement, on the next day 

 of work, as a rule, the observations were taken first with the " crossed " 

 arrangement and then with the " parallel " arrangement. 



The thermo-electric current from each of the four couples was observed 

 by means of an armored Siemens-Halske galvanometer, the sensitiveness 

 of which was at first taken before beginning the observations with the 

 four couples, between the two sets of observations, and after the second 

 set. This sensitiveness was found to vary but little from hour to hour, 

 or even from day to day ; and after a time the between-sets test was 

 often omitted. All the resistance coils used during this work were 

 carefully compared with international-standard-ohm coils, and were found 

 correct to within one tenth of one per cent. 



The usual order of observations was as follows, (3-2), for exam- 

 ple, meaning three galvanometer deflections in one direction and two 

 in the other direction: A (3-2), B (3-2), C (3-2), D (5-4), C (3-2), 

 B (3-2), A (3-2), a reversal of the couples from "parallel" to 

 " crossed " or vice versa, then a wait of one hour for equilibrium of 

 temperature to re-establish itself, then, as before, A (3-2), B (3-2), 

 C (3-2), D (5-4), C (3-2), B (3-2), A (3-2). 



Observations of the pressure gauges, when artificial pressure condi- 

 tions were used, and of the barometer, accompanied these galvanometer 

 observations and gave data for calculating the temperatures of the 

 junctions. 



The following summary of the results of the observations made July 

 21, 1905, is a fair sample of the summaries usually obtained. The num- 

 bers given under " Sensitiveness " ace galvanometer deflections observed 

 on reversal of a current of known constant strength used in getting 



