/:'. D. Preston — Measurement of ' ih* Peruvian Arc 5 



not the importance it once had. Notwithstanding, it is be- 

 lieved that a remeasnrement would so modify it as to materially 

 change the earth's ellipticity. 



We now turn to examine the work more in detail. The 

 first hase was measured on the plain of Yaronqui, about fifteen 

 miles east of Quito. Eight days were devoted to clearing the 

 line. Irs true direction was N. 19 9 26' W. Three wooden 

 roils, each twenty feet long, with copper contact plates, pro- 

 jecting one and one half inches at each end were used in mak- 

 ing the measurement. The plates were so arranged as to 

 make contact at right angles to each other. A rope was 

 stretched for alignment, and the inclination of each rod was 

 determined by means of a level. Twenty-five days were con- 

 sumed in the work, which was exceedingly laborious because 

 the rods were laid on the ground. This course was pursued on 

 account of the violent wind. Two rods were always in posi- 

 tion. The rear one, carried forward by Indians, was brought 

 into contact with as little shock as possible ; but with a heavy 

 rod and in the hands of several untrained persons, it is difficult 

 to see how shocks were avoided. An iron toise was carried 

 along, kept in the shade, and comparisons were made always 

 daily and sometimes oftener. The temperature and humidity 

 of the air affected the wooden rods considerably. The work 

 was begun from both ends and the parties compared their rods 

 when they met in the middle. The south party, however, 

 used tressels, and it was noticed that the effect of wind on the 

 plumb-line, and the consequent error in the length of the base, 

 would be in opposite directions for the tw r o measures. In spite 

 of this fact, and with the exceedingly rough method of 

 making the contact,, the two independent measures only dif- 

 fered by about three inches for a distance of more than six 

 thousand toises, which is about TjoVoo^h part. This is a 

 degree of accuracy far beyond what we can reasonably expect 

 in work of this kind, and there must certainly have been large 

 compensating errors. 



Base measures of the present time, with all our improved 

 methods of dealing with the temperature, perfected contact 

 slides, better ways of alignment, and more skillful manipula- 

 tion of the bars by persons trained to the work, do not give 

 much better results. 



The actual measures gave somewhat less than 6273 toises for 

 the length. It was estimated that the necessary corrections 

 would increase this quantity, and in order to have their base an 

 exact number of toises, one of the end marks was moved three 

 inches and eight lines. It is hard to see what was to be gained 

 by this. Of course the round number would be broken by the 

 solution of the first triangle. Their own measures showed an 



