567 



Any error in laying down curves from these data can scarcely have 

 exceeded two or three hundredths of an inch. 



3. Observations less accurate. The heights of the stations of 

 some observers are not known so nearly. Other corrections have 

 been applied only in a few cases, the observations sometimes recorded 

 only to the nearest tenth, as at a few lighthouses, not being deemed 

 sufficiently reliable. 



Returns in which the barometrical observations are evidently erro- 

 neous (from comparison with other neighbouring and contempora- 

 neous observations) have been rejected. 



On the whole it may be safely assumed that the observations from 

 which the curves are laid down are less than a tenth in error. 



4. Lighthouses. The heights of the lantern above the sea- level, 

 and of the tower, being known, the heights of the barometers have 

 been ascertained, and corrections for the heights have been applied. 



Comparisons of Wind Scales. 



Sea. Wind. Land. 



to 3 Light to 1 



3 to 5 Moderate 1 to 2 



5 to 7 Fresh 2 to 3 



7 to 8 Strong 3 to 4 



8 to 10 Heavy 4 to 5 

 10 to 12 Violent 5 to 6 



Wind. No. Velocity. 



Pressure Ibs. (avoirdupois). (land scale). Miles (hourly). 



i- i 10 



5 2 32 



10 3 45 



21 4 65 



26 5 72 



32 6 80 



These comparisons of scales have been used in the wind charts, 

 and have been found convenient as well as sufficiently accurate. 



