Standard Measures of Length. 



Centimeter Divisions. 



.0017 



Millimeter Divisions. 



It will be seen, from this brief account, that the construction 

 of a true centimeter at a given temperature is by no means an 

 idle or a useless problem. In fact, it has become a necessity. 



It will also be seen from what has been said, that even an 

 absolutely perfect screiv cannot be taketi as a standard. Hence, the 

 proposition of our friends abroad to depend upon a Whit- 

 worth screw, carefully kept in the custody of the Royal Micro- 

 scopical Society, is hardly feasible. A screw is not in itself, 

 a standard, and it cannot be so taken. It is simply a tool to 

 facilitate the subdivision of a given unit. Even if there were 

 no other reasons, the changes effected through temperature 

 would be fatal to its use as a standard, for there are no means 

 of measuring these changes with sufficient precision. But 

 there are other conclusive reasons, which forbid the use of a 

 screw for a standard. In a paper on the sources of errors in 

 micrometer screws, which I hope to publish presently, I expect 

 to show : 



(i) That the errors which belong to a given screw, reside, 

 for the most part, in the mounting of the screw, rather than 

 in the screw itself. Hence, if a screw is dismounted, or if any 

 change is made in the adjustment of the mounting, either 

 arbitrarily or by the action of temperature, the screw practi- 

 cally undergoes a change of pitch. 



