214 Intelligence and Miscellaneous Articles. 



come to the opinion that solid steel bars do not in general furnish 

 good, permanent magnets, because, on the one hand, the core absorbs 

 much of the magnetizing force already during the magnetization, 

 and, next, because the core generally is to be regarded as an arma- 

 ture joining the two poles*. I consequently instituted some further 

 experiments, with steel tubes, which appeared, on the whole, to 

 confirm that view. 



Now, as the manufacture of such steel tubes presents no diffi- 

 culty at all, but rather they can be had in larger pieces at consider- 

 ably lower prices than solid bars (for the tubes need not be bored, 

 but can be forged out of sheet-steel, since it is indifferent whether 

 the wall is perfectly closed or not), it did not seem superfluous, in 

 regard to the possibility of turning it to practical account, to draw 

 a more exact comparison between the action of such and ordinary 

 solid magnets. 



Now t it is true that in the meantime I have learned that Nobili 

 ascertained that a small steel tube became permanently more pow- 

 erfully maguetic than a solid rod of equal external dimensions. 

 According to the brief report in Poggendorff 's Ahnalen, xxxiv. 

 p. 270 (1835) — the original memoir, unfortunately, not being ac- 

 cessible to me — the rod weighed 28*5, and the tube 16 grams ; in 

 spite of this the latter caused a deflection of a compass-needle of 

 19°, while the former produced one of only 9°'5. Still this expe- 

 riment, apart from the dimensions of the pieces having been pro- 

 portionally small and the want of a more exact determination of 

 the magnetisms, has hitherto remained isolated. Hence I should 

 nevertheless like to describe in a few words the results of my own 

 experiments. 



First I employed a rod and a tube 12 J centims. long and 13mil- 

 lims. in outside diameter. The thickness of the wall of the tube 

 amounted to 1 j millim. Both rod and tube were magnetized up to 

 saturation. The magnetisms were tested by the oscillation method, 

 the tube being loaded throughout its whole length with copper. The 

 magnetism of the rod was to that of the tube as 1 : 1*6. 



I next used a rod and a tube of 32 centims. length and 35 ruil- 

 lims, external diameter, while the thickness of the tube-wall was 

 the same as the above. I believe that here also both rod and tube 

 were magnetized to saturation, since an extraordinarily great mag- 

 netizing force was at my disposal. The magnetism of the rod 

 (tested as above) was to that of the tube as 1 : 1 # 5. 



I had, in truth, expected a greater difference in the latter case 

 than in the former, because the core was so much thicker, and this 

 circumstance would necessarily have great influence ; but I soon 

 reflected that the length of the core (the imaginary armature) must 

 also have played an essential part. Had I in the second case, with 

 the same thickness of the rod, used only a length of 12| centims., 

 I should in all probability have found a much greater difference* 



Lastly I tried, besides, how much the magnetism of the tubes 

 would be lessened if I employed a real armature, i. e. if I filled them 



* Holtz, Wied. Ann. vii. p. 71 (1879). 



