C. Barns — Simple Screw Micrometer. 269 



tion having been improvised. It furthermore contains all 

 errors of the Fraunhofer screw, and finally all errors in judg- 

 ment in bringing the centers of ellipses back to the sodium 

 line. Nevertheless, throughout the whole length examined 

 the new screw retains a satisfactory mean pitch. The oper- 

 ation of comparison can easily be completed in one after- 

 noon. 



A similar screw with the two lugs but 1/4 inch thick and 

 about 3 1/2 inches apart was now tested. The mirror mechan- 

 ism was made lighter and the adjustment screws liner (56 

 threads to the inch), the object being to quite eliminate, if 

 possible, the circular motion of the reflected ray. This is not 

 easily accomplished, owing perhaps to an insufficiently plane 

 mirror ; but on rotating the screw in successive arcs of 90° the 

 ellipses were usually available in three of the positions, though 

 they were too blurred in the fourth for use ; eventually they 

 were visible throughout. For final adjustment besides a more 

 perfect mirror (which need be only 1/2 inch in diameter) finer 

 adjustment screws than the above would be desirable. The 

 new comparison as a whole is given in fig. 3, in the same way 

 as above. The fluctuation is on the average within about '0003 

 centimeter, or half a scale part of the drum of the Fraunhofer 

 micrometer and 1/236 of the circumference of the drum of the 

 new screw. This uncertainty is wholly due to the difficulty of 

 setting the latter with this precision, as the head was impro- 

 vised in the laboratory for the purposes here in view. 



To summarize : There seems to be no doubt, therefore, that 

 a screw, satisfying the requirements of the interferometer and 

 trustworthy to about '0001 centimeter and a length of even a 

 foot or more, could be constructed by the above method. It 

 is necessary to begin with a straight rod for this purpose (or 

 preferably with a tube) of a larger diameter, say 1 or 2 centi- 

 meters. The adjustable mirror at the end should be light, 

 which was not the case in the above apparatus, ordinary thick 

 plate glass 1 1/2 inch square being used. The adjustment 

 screw with orthogonal axes must be of very fine pitch, if the 

 image is to be stationary. Such a screw of low pitch, carefully 

 cut in brass and running in sockets of indurated fiber, can be 

 made in almost any laboratory. 



The final advantage of this type of micrometer is the fact 

 that the normal to the mirror is necessarily a prolongation of 

 the axis of the screw and also coincides very nearly with the 

 incident and reflected ray. There are thus no unknown angles 

 between screw axis and normal to the mirror and exchange of 

 screws at M and J¥ is no longer necessary. When a mere 

 comparison of screws is in question, it is sufficient to reduce 

 the play of image in the telescope to a small circle, and this 

 may be done in a few minutes. 



Brown University, Providence, R. I. 



Am. Jour. Sci.— Fourth Series, Vol. XXXV, No. 207.— March, 1913. 

 19 



