PHILOSOPHICAL SOCIETY OF WASHINGTON. 131 



the sun were obtained with sextant and artificial horizon. These 

 three sets give as the correction of our "hack," or observing chron- 

 ometer, to local mean time 



h. m. s. s. 



4- I 03 26.9 ± 0.35, 



this probable error resulting from computing the eighteen observa- 

 tions siugly and treating in the usual way. The sky was nearly 

 covered with cumulus clouds, the wind fresh, raw and chilly, and 

 thermometer 39° F. Near noon the sun appeared again for a short 

 time, and nine pointings were obtained for latitude, giving the fol- 

 lowing results, each depending upon a single observation. 



65° 44/ 54" • 



50 

 38 

 54 

 44 

 52 

 53 

 60 



65 



Mean latitude, 65 44' 51 =b iZ'5 N. 



Leaving the Diomedes on the afternoon of September 10, we 

 sailed directly for Plover Bay. That night we were stopped by ice, 

 the next day delayed by calms, but on the following day, September 

 12, we reached our anchorage in Plover Bay a little before noon, just 

 in time to get a good series — 39 observations of circummeridian 

 altitudes of the sun for latitude. In the afternoon we obtained a 

 good series of time observations, but the following morning was 

 cloudy. We succeeded, however, in getting four altitudes corres- 

 ponding to those of the preceding day, thus enabling our time 

 determination to hang upon four pairs of equal altitudes, the epoch 

 being local mean midnight September 12 and 13. The times of 

 local apparent midnight from these four pairs by our " hack " were 



h. m. s. 



II 09 o. 2 



1.2 



0.3 

 0.7 



from which the probable error is found to be ± 0.15s. 



