216 BEI.L SYSTEM TECIlNICylE JOVRNAI. 



two eiirs have about the same ainpHtudes, differing by not more than 

 25 TU. Data obtained with 22 observers are summarized. The 

 evidence indicates that these beats are not due to cross conduction 

 but are of central origin and the result of the sense of binaural local- 

 ization of sound by phase. If the beats are slow (less than 1 per 

 sec.) they are generally recognized as an alternate right and left 

 localization, though some observers may report one or more intensity 

 maxima during the beat cycle. Such maxima are explained as the 

 result of one's interpreting the sound as louder when localization 

 is more definite. Fast beats (more than 1 per sec.) are generally 

 recognized as an intensity fluctuation. They are explained by assum- 

 ing that the sound appears louder when the phase relations are such 

 that it is normally best localized in the position toward which the 

 attention is directed. This explanation is supported by observations 

 made with a constant source rotating around the head of a listener. 



Effect of Tension Upon Magnetization and Magnetic Hysteresis in 

 Permalloy. O. E. Buckley and L. W. McKeehan.^ Wires of 

 five nickel-iron alloys containing 45, G5, 78.5, 81 and 84 per cent. 

 Ni, 60 cm. long and 0.1 cm. in diameter, were studied by a ballistic 

 method, for tensions up to 10,000 lb. per in.^ and fields up to satura- 

 tion (10 to 20 gauss). Permalloy with 81 per cent. Ni is nearly in- 

 different to tension in its magnetic behavior; permalloy with less 

 nickel is more easily magnetized and has less hysteresis when under 

 tension, while 84 per cent, permalloy is more difficultly magnetized 

 and has greater hysteresis when under tension. The saturation 

 values are independent of the tension. In 78.5 per cent, permalloy, 

 under a tension of 3,560 lb. per in.^, saturation is reached at only 2 

 gauss (and is practically complete at 0.2 gauss) and the hysteresis 

 loss is only 80 ergs per cm.^ per cycle, so small that it may be regarded 

 as due to slight inhomogeneity rather than to any essential features 

 of the magnetization process. Relation to crystal oruntation. X-ray 

 examination proves that this abnormally low loss is not due to any 

 peculiar orientation of the crystal axes as the crystals are found to be 

 oriented at random. Magnetostriction behavior can be deduced 

 from these results. Above 81 per cent. Ni, permalloy contracts like 

 Ni while below 81 per cent, Ni, permalloy expands like Fe. 



Demagnetizing factor for a wire with a length 600 times the diam" 

 eter, was determined experimentally and found to vary from a maxi' 

 mum of 1.6 XUr^ to a low value, the changes being like these pre- 

 viously described by Benedicks for iron. 



^ Physical Review, Vol. 26, No. 2, Aug., 1925. 



