Mr. H. A. Rowland on Magnetic Distribution. 



Fig. 7. 



361 



Results from permanent magnets. 



A. Soft steel. 



B. Hard steel. 



In fig. 7 I have attempted to give the curve of distribution 

 from Table XV., and have made the curves coincide with 

 observation as nearly as possible, making a small allowance, 

 however, for the errors introduced by the shortness of the 

 bar. It is seen that the effect of hardening in a bar of these 

 dimensions is to increase the quantity of magnetism, but espe- 

 cially that near the end. Had the bar been very long, no increase 

 in the total quantity of magnetism would have taken place ; but the 

 distribution would have been changed. From this we de- 

 duce the important fact that hardening is most useful for short 

 magnets. And it would seem that almost the only use in harden- 

 ing magnets at all is to concentrate the magnetism and to reduce 

 the weight. Indeed I have made magnets from iron wire 

 w 7 hose magnetization at the central section was just as intense as 

 in a steel wire of the same size j but to all appearance it was 

 less strongly magnetized than the steel, because the magnetism 

 was more diffused ; and as the magnetism was not distributed 

 so nearly at the end as in the steel, its magnetic moment and 

 time of vibration were less. 



It is for these reasons that many makers of surveyors' com- 

 passes find it unnecessary to harden the needles, seeing these 

 are long and thin. 



We might deduce all these facts from the formulas on the 

 assumption that r is greater the harder the iron or steel. 



Having now considered briefly the distribution on electro- 



