76 Mr. E. H. Barton. Electrical Interference Phenomena 



conditions adopted in this one or were in other ways subsidiary to it. 

 The construction of the abnormal part was such as to reflect a large 

 fraction of the wave- energy incident npon it. Its position was so 

 chosen as to avoid the interferences of higher order already noticed. 

 Thus the true effect sought was large, the disturbing ones small. The 

 arrangement of the apparatus is shown in fig. 1. The details are as 

 follows. 



The abnormal part consisted of tinfoil sheets each 32 cm. deep and 

 50 cm. long. These were hung opposite each other upon the two 

 wires of the secondary. Lest a shift of the middle point of the sheets 

 along the wires might have a slight disturbing effect on the electro- 

 meter throws, the abnormal part was always lengthened or shortened 

 by the same amount at each end, thus leaving the middle point M 

 undisturbed. The length SAM was throughout the experiment 

 101 m., MD being 63 m.* so as to comply with condition (B). The 

 tinfoil sheets were kept properly spaced by very accurately cut wood 

 separators, one for each half metre's length of the sheets. The same 

 separators, but at longer intervals, were used throughout the line. 

 The vertical edges of consecutive sheets were made to overlap about 

 2 cm. and allowed to hang in simple contact. 



Lest this contact should be insufficient and so involve an error, the 

 sheets were built up in this way to a length of 2*25 m. (a quarter of a 

 wave-length), and the ratio determined of the wave-energy trans- 

 mitted by it to that incident upon it. Then an abnormal part, con- 

 sisting of zinc plates soldered together was substituted, and the 

 fraction of wave-energy transmitted again determined. In each case, 

 to insure accuracy, twenty-one electrometer readings were taken 

 alternately with and without the abnormal part. The two ratios 

 thus determined were : f 



(1) For the tinfoil sheets 0'135 0'007 



(2) For the zinc plates soldered 0'135 0'002 



Hence this apparently imperfect contact seemed sufficient for the 

 case, and was adopted in the main experiments now under considera- 

 tion. A spark gap of 2 mm. was used throughout both in the experi- 

 ments just described and in what now follows. 



The method of taking the observations so as to eliminate the 

 possible errors due to the irregularities in the sparking of the primary 

 was as follows. Several electrometer throws were taken, at first 



* When making the corresponding experiment, described in the former pa,per 

 p. 96, SAM was taken equal to MD. I had not, tip to then, seen that it was 

 desirable to avoid that arrangement. 



f The probable errors were calculated by the formula) given in Kohlrausch's 

 ' Leitfaden der Praktischen Phjsik,' Leipzig, 1892, pp. 13, and proved in Dr. B. 

 Weinstein'e ' Handbuch der Phvsikalischen Maassbestimmungen,' vol. i. 



