Attachment of Thin Galcanometer Mirrors. 191 



To give the polished surface its reflecting coating a very 

 thin layer of the platinizing liquid is uniformly painted ovar 

 the surface with a clean brush, and tlie disk is placed on a 

 metal plate over a water-bath. When the coating is quite 

 drv the disk is heated in a small muffle made by bending a 

 sheet of thin sheet-iron and placing it over a large Bansen 

 flame. The heating must be continued till a fairly bricjht 

 red is attained. It is an advantage to raise the temperature 

 of the disk to a red heat as quickly as possible, for if the 

 heating is slow the platinizing compound tends to volatilize 

 before it decomposes. The surface thus obtained will be very 

 bright and will not require polishing. If the film of platinum 

 is too thin a second coating of the platinizing liquid can be 

 given. 



In this way a mirror is obtained which, however, is about 

 a milHmetre thick. To reduce it to a more suitable thickness 

 the platinum surface is temporarily protected with a coating 

 of pitch or some other varnish, and the other surface of the 

 disk is ground away on the brass-surface plate with carbo- 

 rundum till the desired thickness is obtained. It will be 

 found quite easy to prepare a mirror having a diameter of 

 one centimetre and a thickness of tw^o-tenths of a millimetre. 

 Such a mirror will w^eigh about '045 gram. 



In order to support the mirror the tag is fused, by means 

 of a small oxyhydrogen flame, to a thin rod of fused silica 

 and the magnets are cemented to this rod. 



The above method of constructing galvanometer mirrors 

 avoids the necessity of grinding and polishing the surface of 

 a very thin disk, an operation of great difficulty, and is only 

 rendered possible by using a non-corrodible reflecting surface 

 so that the light has not to traverse the quartz disk. Also 

 the excessively small coefficient of thermal expansion of quartz 

 enables us to obtain a disk of the material which is so free 

 from internal strains that w^hen we grind aw^ay one side the 

 form of the other side is not appreciably altered. Lastly, 

 the method of attachment of the mirror to the stem, since it 

 avoids all cements, entirely does aw^ay with the risk of dis- 

 tortion due to the contraction of the cement. 



Platinum surfaces prepared as above described have not 

 quite such a large reflecting power as glass backed with 

 silver. A comparison between a platinum surface which had 

 been exposed quite unprotected to the air of the laboratory 

 for a month and a silver on glass mirror made by Hilger 

 show^ed that the reflecting power of the platinum was about 

 seven-tenths of that of the silver. Anv difficultv due to this 



