438 Intelligence and Miscellaneous Articles. 



For splitting the films I know of no better plan than the old 

 device for splitting a sheet of paper in two, viz. to glue the mica 

 between two pieces of fine cambric and (when they are dry) tear them 

 asunder ; a film remains attached to each, and can be separated by 

 immersing the whole in hot water. 



It is easy in this way to get films so thin that they only give two 

 " Talbot's bands " between the Fraunhofer lines D and F, when 

 the light reflected from them at an angle of 45° is examined in the 

 spectroscope ; and I have obtained many (but not very large) films 

 thinner than this, showing brilliant colours of Newton's first order. 

 Such films are, of course, very fragile; but they can with care be 

 floated upon a glass plate immersed in the water, and then lifted 

 out and dried. H. G. Mad an. 



Eton College, May 15, 1883. 



TO CUT A MILLIMETRE-SCREW. BY CHARLES K. WE AD. 

 To the Editors of the Philosophical Magazine and Journal. 



GrENTLEMEN, 



The simple method that Mr. Bosanquet has used in his " Ar- 

 rangement for dividing Inch- and Metre-Scales," described in the 

 March number of this Magazine, seems to me of so much value 

 that I send you a brief description (communicated to ( Silliman's 

 Journal, March 1882) of another application of the same prmciple, 

 i. e. of the use of a wheel of 127 teeth or divisions. 



In this connexion two points may be worthy of record here : — 

 First, that Eogers's measurements, reported at the Montreal Meeting 

 of the American Association, and confirmed in a general way by 

 E. S. Pierce, make the metre = 39-37015 inches; this is longer 



than what I have called the "mechanical" metre ( =- — — — inches j 



V 2x127 ) 



by only 1 part in 560,000. Secondly, the absolute error of some 

 fine screws is many hundred times this theoretical error: the 

 screw of my dividing-engine is too short by 1 part in 700 ; of a 

 screw from Perreaux at Johns Hopkins University, Eowland states 

 that the " thread is \ millim., and the head has 250 divisions ; 501 

 divisions give 1 millim. almost exactly." The screw of the engine 

 on which Rutherford's well-known gratings were cut furnishes 

 17,296 lines to the inch instead of 17,280 ( = 48 x 360) ; so the error 

 here is about 1 part in 1100. In all these cases, and in some others 

 that I have noticed, the screw 7 is too fine. From the details given 

 about the Rutherford screw in the article " Spectrum " in ' The 

 American Cyclopedia,' it would seem that the shortening here was 

 due to the hardening of the taps and dies used in making the screw. 

 In the mechanical journals methods have been described for giving 

 to a lathe feed-screw a slow longitudinal motion to correct for any 

 error of pitch ; but such methods could hardly prove useful on phy- 

 sical apparatus. I am, yours respectfully, 

 Ann Arbor, Mich., C. K. Wead. 



April 14, 1883. 



The increasing attention paid to the metric system renders it 



