26 



EIGHTH REPORT — 1838. 



four inches in diameter. The water was measured in a glass tube, in 

 which one-thousandth of an inch of rain fallen was measurable. When 

 the snow gauges became useless, they were used as rain gauges, by at- 

 taching a funnel to them, or were finally replaced by rain gauges simi- 

 lar to those described. The quantity of water collected was measured 

 after each rain, and the direction of the wind during the rain was fre- 

 quently noted. To illustrate the effects which are attributed to cur- 

 rents of air deflected by the tower, Professor Bache has taken from the 

 journal of the latter months of observation the records of the quanti- 

 ties of rain collected by four similar gauges, placed at the four angles 

 of the tower, under different circumstances as to the direction of the 

 wind. These are selected so as to present, as far as possible, a case of 

 rain with each principal direction of the wind. 



On this table the author remarks, — 1. Tliat it illustrates the very 

 great differences between the quantities of rain collected at the different 

 angles of the tower. In one extreme case the quantity collected at the 

 S.E. angle was 2^ times that at ^he N.W. angle. 2. That, in general, 

 the gauges to leeward received more rain than those to windward. 

 Thus, with a north wind, the gauges at the S.E. and S.W. angles re- 

 ceived more rain than those at the N.E. and N.W. angles. With a 

 N.E. wind the gauge at the S.W. corner of the tower received the most 

 rain. In the case given in the table, the ratio of the quantities is nearly 

 2*1 to 1. With an easterly wind the N.E. and S.E. gauges received 

 less than the N.W. and S.W. With a south-easterly wind the S.E. 

 gauge received the least, and the N.W. the greatest quantity of rain, 

 and so on, nearly in the order stated in the general remark. 3. As the 

 more considerable rains accompany certain winds, it is not to be ex- 

 pected that averages of any number of observations exposed to such 

 errors Avill lead to an accurate result of the quantity of rain falling at 

 a certain height above the surface. In fact, tiie averages from a period 

 of nine months do not agree nearly so well as those from the selected 

 specimens in the table. These give ratios of 1, 1*19, r24', and 1*20, 



