6fj0 WINDSOFTHEGLOBE. 



taking into account the velocity of the winds, it will be found that in the last case 

 the mean direction generally approaches nearer to the actual direction of the 

 prevailing wind. 



Besides it is seen that generally the ratio of resultant is greater in the second 

 case than in the first. 



Thus it follows that, when we have the mean direction of the wind at a given 

 place, calculated from the number of observations only, we may infer that, if the 

 velocity was known, it would modify the result in so far as to make the mean 

 direction nearer to that of the prevailing wind, and the ratio of resultant greater. 



Unfortunately most of the above deductions apply only to the United States, as it 

 was the only country for which Prof. Coffin made his calculations from the original 

 journals. As to printed meteorological journals, they were very scarce until within 

 a late period, and many of these were not to be had in the United States. There- 

 fore published means and abstracts had to be relied upon, and these gave only the 

 number of observations for each wind. It might be thought that the results of the 

 self-registering anemometers now in use in so many meteorological stations would 

 give abundant material for the answer to this question, but, owing to the recent 

 introduction of these instruments in some cases, and to discontinuity of record in 

 others, comparatively few tabulated records of velocity of winds have been printed. 



Yet it seems that the angle between the mean direction calculated, taking into 

 account the number of observations only, and that in which the velocity is con- 

 sidered, seldom exceeds 15°. In case of a very small ratio of resultant it can be 

 much greater, but this small ratio itself shows that the mean direction is not 

 much to be relied upon. 



All this leads to the conclusion that it is possible to calculate the mean direction 

 of the wind from the number of observations only witliout incurring a large error. 

 The map, Plate 13, shows the resultant direction for the number of observations 

 only, as also for velocity, in the United States. I must also explain in what sense 

 I use the words " polar" and " equatorial" winds. Polar designates a wind blowing 

 from a higher latitude towards a lower ; and equatorial, a wind blowing from a 

 lower towards a higher latitude. I use these terms in the way which is most 

 generally admitted, to avoid confusion. This agrees also with the manner in 

 which winds are generally designated, so far as we call north wind one that blows 

 from the north towards the south, and not rice versa. 



It will be remembered that in the " Winds of the Northern Hemisphere" Prof. 

 Coffin used the words " polar" and " equatorial" in the opposite sense. 



Another question, to my mind, more difficult to answer, is as to the value of the 

 observations on the motion of clouds. They may serve two ends: 1, to ascertain 

 the motion of an upper current of the air ; 2, to observe the lower current, free 

 from the irregularities often found immediately above the surface of the earth. 



Naturally enough, in this case all depends on the height of the clouds observed. 

 Very seldom, if ever, in discussing observations from a journal, can even the 

 approximate height of the clouds observed be ascertained. This alone detracts 

 very much from the value of such observations. Besides this, the cases must be 

 taken into account wlien there were no clouds, or, the clouds being very high, no 



