268 CARNEGIE INSTITUTION OF WASHINGTON. 



instruments with frequent changes of observers might indicate some improve- 

 ment in methods of observations or adjustment. 

 The following conclusions are drawn: 



1. The dip-of-horizon tables in conmion use which ignore air-water tempera- 

 ture-differences are sufficiently accurate for the navigator. 



2. Extraordinary values may possibly occur in certain regions, where one 

 should be ready to detect them either by observing stars in different azimuths 

 or by special instruments or special attachments to the sextant. Even if 

 the direction of the abnormal value is uncertain, the knowledge of its existence 

 is a factor of safety. 



3. If aerial navigation across the oceans is eventually realized, and astro- 

 nomic methods are used, then simple means of measm-ing the dip-of-horizon 

 might be desirable if not absolutely essential. 



