Viscosity and Rigidity of Iron and of Steel. 185 



out. The following example may be given from very many 

 results of my own. In the case of Nos. 3, 4, the scale readings 

 for currents alternately off and on, were 



TZ= 



0-7° 



1-5° 



2-3° 



2-2° 



1-5° 



o-r 



(off) 



130 



110 



90 



80 



50 



60 



(on) 



150 



200 



185 



180 



128 



100 



(off) 



80 



85 



70 



70 



38 



58 



(on) 



160 



188 



180 



178 



130 



100 



(off) 



80 



82 



70 



72 



40 



60 



This is very perceptible in the first part of the measurements 

 where r or co increase from zero to the maximum ; it generally 

 becomes smaller in amount when co decreases from the maximum 

 to zero. The method of varying co from to the high value 

 and then from the high value to zero leads to a cycle of results. 

 In the case of ISTos. 7, 8, for instance 



Gl>X10 3 = 



o-o 



0-7 



1-4 



2-2 



2-9 



3-2 1 



go in- 



cpxl0 6 = 



5 



40 



57 



63 



66 



65 



' creasing 



o?X10 3 



3-2 



2-9 



2-2 



1-4 



0-7 



o-o 



go de- 



cpXIO 6 



67 



64 



61 



51 



28 



3 



" creasing 



Observation, which might perhaps be grouped with Mr. 

 Ewing's* phenomena. But in some of the many other like 

 experiments which I made there is an element of vagueness. 



Table 5 shows finally that the influence of magnetism on the 

 rigidity of soft iron is greatest when co is zero. It is obvious 

 since <p increases at a retarded rate while z or co increase uni- 

 formly, that f± must continually decrease. The large values of 

 fi differ considerably being greater as a rule for smaller diam- 

 eters. The minimum values of p. obtained are of about the 

 same magnitude, / a=0 , 0021. If the conditions of indefinite per- 

 manent set are such that the obliquity of the external fiber, 

 attain a fixed value, then these conditions imply that the 

 rigidity of steel is changed by magnetization by a fixed mini- 

 mum amount. 



Steel. — In the following table 6, 1 give the values of .<p which 

 hold for steel for the large rate of twist r=6°. The wires are 

 carefully annealed as described elsewhere. Diameter 2p= 

 0'083 cm . A soft iron wire for which maximum co is only 

 0*0021 is also tested for comparison. 



The curious result of this table is this, that during the first 

 phasef of annealing, the effect of magnetization on the rigidity 

 of steel is almost nil. This effect becomes of marked impor- 



*EwiDg: Trans. Roy. Soc., ii, 1885, p. 523-4. 

 f Thi3 Journal, xxxi, p. 443, 1886. 



