June 13, 1901] 



NATURE 



169 



of the cast steel is high. When the density of the flow is found, 

 having a previous knowledge of the magnetic quality of the 

 steel we can get the permeability for each section. Then the 

 product of current, into number of turns of wire, necessary to 

 force the magetism through all the diflferent sections into which 

 the magnetic circuit has been divided can be found, and, adding 

 all these together, we can obtain the current that must be sent 

 round the magnet a given number of times to produce the 

 desired effect. The following table gives the flow, density of 

 flow, &c., at the different sections (Fig. 4), the " ampere 

 turns," i.e. current, into number of turns, given in column 5, 

 being for the particular section, together with the similar one, 

 on the other side of the ring. 



For Magnet, Fig. 4. 



Column 2 shows what a large amount of magnetism leaks out 

 from the one side of the ring and passes over to the other, not 

 passing through the air gap at all. Column 3 shows how the 



Fig. 4. — (Scale, quarter full size). 



area of the iron has to be increased, owing to this leakage, for 

 sections further away from the air gap. The current for this 

 magnet is to be 5 "4 amperes, and this flowing 1,500 times 

 round should allow an ample margin to produce the field 

 required. 



It is interesting to compare the leading particulars of these 

 two magnets. 



NO. 1650, VOL. 64] 



Area of air gap, sq. cm. 



E.-cciting current, amperes 



Turns of wire 



Product of current, into number of 

 turns, ampere turns 



Length of wire, yards 



Weight of wire, lbs 



Weight of iron in magnet, lbs 



Total weight, iron and copper, lbs. ... 



Cost of iron and copper — 



Iron @ 2 •2a'. per lb. ; copper @ 

 IS. per lb 



Cost of working magnet continuously, 

 per year, with current at dd. per 

 Board of Trade unit 



^11901 /2 II O 



■/130 o o ^25 O O 



Since the magnet was intended to work in connection with a 

 submarine cable, it is quite safe to say it would require current 

 throughout the year, and then, as the above table shows, there 

 would be a saving of more than 100/. a year by using the magnet. 

 Fig. 4, if the current could be got at dd. per Board of Trade 

 unit — it would, however, very likely cost much more under these 

 conditions — to say nothing of the saving in first cost. 



Another example of the importance of having sufficient iron 

 in the magnetic circuit is afforded by the alterations that were 

 made to an electro-magnet belonging to the Central Technical 

 College. The cross section of its yoke was less than half that 

 of the cores, due to the fact that the magnet was made before 

 the theory of the magnetic circuit was understood. Recently a 

 new yoke of proper cross section was made for the magnet, and 

 it was found by experiment before and after the alteration that 

 under precisely the same conditions the strength of the field 

 had been just doubled simply by adding a few pounds of iron in 

 the right place. , 



But at the same time it should be remembered that the mag- 

 netic properties of different specimens of iron vary enormously, 

 and it is important that any iron which is to be used for a magnet 

 should first be tested. Fig. 5 shows a very convenient 

 apparatus for doing this. It consists of a massive iron frame 

 into which you can slide a bar of the sample to be tested, the 

 bar passing through a thin brass tube on which a known number 

 of turns of wire are evenly wound. Also a few turns of fine 

 wire are wound on the middle of the tube, shown at r. Fig. 5, 



Fig. 5. 



and connected to a ballistic galvanometer. A current is sent 

 through the large coil, causing magnetism to flow through the 

 bar, returning by the massive frame the cross section of which 

 is so large that practically the whole of the magnetomotive 

 force due to the current is used to send the magnetism through 

 the bar ; therefore, dividing the total ampere turns by the length 

 of the bar we obtain the ampere turns necessary to send the 

 magnetism through one centimetre of that specimen of iron. 

 Several different strengths of current are sent through the mag- 

 netising coil, and in each case the flow of magnetism produced 

 is measured by suddenly switching olf the current, consequently 



