590 



SCIENCE. 



[N. S. Vol. XVIII. No. 462. 



the cable were sent on January 1, 1903. The 

 cable was opened for public use on Janu- 

 ary 5, 1903. The first observations in the 

 telegraphic determinations of longitude 

 were secured on April 20. 



The cable company cooperated with the 

 government by aiding the observers in vari- 

 ous ways, and by granting the free use of 

 the cable during a short period each night 

 for their work. 



Between April 20 and June 13 eleven 

 determinations of the difference of longi- 

 tude were made, five with the observers in 

 one position and six after they had ex- 

 changed places for the purpose of elimina- 

 ting the effects of their relative personal 

 equation. The total range of the eleven 

 results was only 0.17 s. No result differed 

 from the mean by as much as one tenth of 

 a second. 



This degree of accuracy corresponds to 

 that usually attained in determinations of 

 the difference of longitude of two places 

 connected by a telegraph line overland. 

 Until within a few years such a degree of 

 accuracy had not been possible for deter- 

 minations made through long cables. The 

 increase in accuracy has apparently been 

 due to improvements made in the record- 

 ing apparatus used by the cable companies. 



The difference of longitude of the transit 

 piers in San Francisco and Honolulu, re- 

 spectively, was found to be, from the field 

 computation, 2 h. 21 m. 38.92 s. 



A revision of the computation still re- 

 mains to be made, but the changes in this 

 result will be very small. The chances are 

 even for and against the difference stated, 

 being within 0.03 s. of the truth. 



The longitude of San Francisco from 

 Greenwich is fixed by four transatlantic de- 

 terminations by the telegraphic method, 

 1866, 1870, 1872 and 1892, and by a com- 

 plicated telegraphic longitude net stretched 

 across the United States. 



The longitude of the Transit of Venus 

 pier at Honolulu, as fixed by the new de- 

 termination from San Francisco, is 10 h. 

 31 m. 27.24 s. west of Greenwich. Taking 

 into account all the uncertainties in the 

 chain of observations between Greenwich 

 and Honolulu, the chances are even that 

 the value stated is correct within 0.06 s., 

 and it is almost certain that it is not in 

 error by as much as 0.2 s. In the latitude 

 of Honolulu 0.06 s. of longitude corre- 

 sponds to 85 feet, or 1 s. to 1,418 feet. 



The observers, Messrs. Smith and Morse, 

 are still engaged in determining the differ- 

 ences of longitude over the three other 

 spans of the cable between Honolulu and 

 Manila. A good determination of the dif- 

 ference Manila-Guam has already been se- 

 cured. When their work is complete the 

 longitude of Manila will have been deter- 

 mined telegraphically in both directions 

 around the world from Greenwich, and 

 the longitude girdle of the earth will be 

 complete. 



The following resume of various deter- 

 minations of the longitude of Honolulu is 

 condensed from an account furnished by 

 Mr. W. D. Alexander, formerly Surveyor- 

 General of the Hawaiian Islands, and now 

 an assistant in the Coast and Geodetic Sur- 

 vey. It is especially interesting as show- 

 ing the comparison of various determina- 

 tions of a large difference of longitude by 

 a variety of methods and extending over 

 three and a half centuries. 



In the following tabular statement the 

 new determination of the longitude of the 

 Transit of Venus station, 10 h. 31 m. 27.2 s., 

 is taken as being correct in deriving the 

 errors of the various determinations. In 

 eases in which the older determinations 

 referred to some other point than the 

 Transit of Venus station, they have been 

 reduced to that station by using the rela- 

 tions in position which are now known. 



