HYDROLOGY OF NEW YORK 



91 



altitude is by no means universal. The writer can not say posi- 

 tively, because he has not examined the vast number of records 

 with reference to this point, but he thinks it quite possible that 

 the reverse is more nearly true. That is, owing to distance from 

 the ocean, prevailing direction of wind, and other causes, it is 

 probable that for the entire country precipitation decreases with 

 higher altitude rather than increases. 



The decision of this question will depend to some extent upon 

 steepness of ascent. Thus on Mount Washington, which is 

 projected into the air far above the surrounding mountains, the 

 rainfall is about S3 inches. In other cases, where the ascent is 

 gradual, no increase is apparent. The same is also frequently 

 true of sharp ascents. On Longs Peak, in Colorado (elevation 

 14,271 feet.) the rainfall in 1899 was 16.7 inches. 



Moreover, the writer has mostly avoided comparatively small 

 differences in rainfall — those not exceeding 2 to 2.5 inches. In 

 such cases the difference is too small to be any certain guide. 

 Specially is this true in the case of the northern plateau, where 

 there is still a great lack of stations. The differences between 

 high altitudes and low should be as much as 5 or 6 inches. Again, 

 whether the excess rainfall occurs in the winter or summer 

 months must be taken into account. If it occurs in the summer, 

 even 3 inches of rainfall may not make more than 0.1 or 0.2 inch 

 in the stream. Rainfall and runoff observations are not yet, nor 

 are they likely to ever be, definite enough to take into account 

 an annual difference of much less than about 1 to 1.5 inches. 

 Again, the writer has ceased to be excessively particular about 

 the total of the annual rainfall. Assuming some considerable 

 length of record, small errors have relatively slight effect. This 

 matter is referred to here because nearly all rainfall records — 

 at any rate in the United States — have more or less error in them, 

 and while it is desirable to have records as reliable as possible, 

 a few errors do not affect a record very seriously. It is neverthe- 

 less very desirable to know the history of the record in order to 

 insure the degree of confidence to be placed in it. 



