134 Dr. S. P. Thompson on Magnetic Figures, and 



wire, and winding it upon a narrow bobbin placed upon the 

 extremity of the magnet. 



Conversely, the passage of a very feeble current through a 

 coil so placed will produce a greater change in the effective 

 intensity of the magnetic field between the core and the dia- 

 phragm than would be produced by the same current traver- 

 sing a similar coil in any other region of the field ; for here it 

 has its greatest power to shift the position of the neutral zone, 

 and to alter the distribution of magnetism in the diaphragm. 



It would therefore appear unnecessary to form an hypo- 

 thesis of molecular vibrations in the disk to account for the 

 emission of sounds by the instrument. Such vibrations do in 

 fact exist ; but their existence does not necessarily prove that 

 they play any important part in the production of the sound. 

 And it must be remembered that, so far as the disk is con- 

 cerned, they take place within the narrow range of the extreme 

 positions possible to the neutral annular zone. 



Two further experiments seem to confirm the conclusion 

 derived from the foregoing observations. If a compound dia- 

 phragm be used, consisting of concentric annuli of thin iron 

 fixed to a stretched membrane of paper, or if a small iron disk 

 thus fixed be employed, as in Bell's earliest experiment and in 

 some of the experimental telephones of M. Niaudet, a curious 

 timbre is thereby imported into the voices of speakers, though 

 their enunciation is very distinct. A similar result is found 

 to follow the employment of small thick diaphragms. In each 

 of these cases the disposition favours the lamellar distribution 

 of the magnetism. 



If, however, a compound diaphragm be employed, consisting 

 of a number of radial pieces similarly fastened to a stretched 

 membrane, tones are well rendered, but enunciation is not 

 distinct. This result is also obtained when the diaphragm of 

 iron is too large in proportion to its thickness. In these cases 

 the greater part of the magnetism is radially distributed. 



Whenever a complete theory of the telephone is framed, 

 these are points which must be taken into account. 



II. On a new Variety of Magnetic Figures. 



De Haldat showed that it was possible to produce magnetic 

 writing upon a steel plate by actually writing with the pointed 

 pole of a powerful magnet, the writing being invisible until 

 fine iron-filings were dusted over the plate. In the Physical 

 Laboratory of University College, Bristol, a small circular saw 

 has been found to afford a plate of suitable thickness and 

 quality to produce good results. The latent characters re- 

 mained for eight months after being inscribed. 



