290 PROCEEDINGS OF THE AMERICAN ACADEMY 



that, while the different results obtained in comparing two standards 

 by one observer and with a given instrument usually indicate marvel- 

 lous precision in the comparisons, a different observer with a different 

 instrument will probably get results equally accordant inter se. but 

 which nevertheless do not agree with those obtained by the first ob- 

 server. Until all the sources of error involved in comparisons are 

 investigated and eliminated, it will be useless to expect an agreement 

 between different observers. Among the points which demand inves- 

 tigation, the following require special attention. 



(a.) The magnifying power of the microscopes employed, which is 

 the best adapted to secure the greatest absolute accuracy in measure- 

 ments. 



In all the earlier comparisons, microscopes of very low power 

 were employed, varying from forty to sixty diameters. The Inter- 

 national Commission, relying largely upon the recent investigations 

 of Forster, have decided upon the low power of from forty to fifty 

 diameters. M. Tresca, of the French section, on the contrary, is a 

 firm believer in high powers; he prefers a power of about 400. The 

 writer has had considerable experience on this subject, and always with 

 results favorable to high powers. With a proper illumination, and 

 with lines having smooth edges, a magnifying power of 900 can be 

 used with great ease, even in the comparison of two meters upon a 

 longitudinal comparator. 



New microscopes have been recently attached to the microscopes of 

 the meridian-circle of Harvard College Observatory. In order to be 

 able to read the divisions of the circle, it was necessary to have one 

 eye-piece with the same power as that furnished by the maker. A 

 second eye-piece, giving nearly double the magnifying power, was 

 attached to a swinging arm in such a manner that either eye-piece can 

 be used at will. A sufficiently extended series of observations has 

 now been made to justify the conclusion that the high-power eye- 

 piece gives the most accurate results. Again, in the investigation of 

 the errors of one of the circles of the instrument, a metal plate, having 

 a graduated arc of 15°, is attached to the opposite circle under a one- 

 inch objective, to which is attached the interior illuminator fur viewing 

 opaque objects, invented by Mr. R. B. Tolles, of Boston, in 1807, and 

 also invented independently by M. Tresca. in 1871. The lines under 

 this objective are sharply defined. The value of one division of the 

 micrometer is only 0".12 against 1".0 for the regular microscopes. It 

 is the experience of the writer, that it is quite as easy to make every 

 reading fall within one division in one case as in the other. "With 



