GEOGRAPHICAL DISTRIBUTION OF RAIN AND WIND. 



183 



of Prof. W. von Bezold. The expcrimeutal rain field really cousists of tlie city of 

 Berlin iiiul the country arouud especially to the westward, embracing a region of 

 about 15 kilometers s<iuare, within which are located Berlin, Spandau, and Potsdam. 

 The forests on the westward, the intermediate gardens and fields, the valley of the 

 river Spree, and the citv of Berlin, otter a great variety of surfaces but without any 

 mountains or hills. The average height of the ground al)Ove sea level is scarcely 50 

 meters, the average distance of the gauges from each otlier, namely, the mean of all 

 possible combinations is about 4.5 kilometers, the maximum distance being 11 and 

 the minimum 0.5. 



Within this area Hellmann has twenty-one stations, some of which rejiresent sev- 

 eral gauges. His work began in 1884, and he adopted as the standard height of the 

 mouth of the gauge 1.07 meters above the ground. I have selected for study the eleven 

 stations for which complete records for 1886-'87 are given in Hellmann's Reports. The 

 accompanying Table VII shows the rainfall for each station for each year and the de- 

 partures of each station from the annual mean of the eleven. From these departures 

 we get the probable error of any one annual rainfall as plus or minus 6 per cent of 

 its own value. That is to say, assuming that the same quantity of rain and snow 

 fell uniformly over the whole of this small region and that the gauges, if unaftectedby 

 any error, should therefore agree among themselves perfectly, then their failure to 

 do so is such that it is an even chance that any given rainfall is discordant from the 

 average by plus or minus 6 per cent. At first Hellmann suggested that the records 

 of stations 1, 2, 3, and 4, which Avere in the open land east of the forests showed that 

 less precipitation fell there than over the forests, attbrdiug an argument for the idea 

 that the forest attracted an extra amount of rain ; but of the other stations there 

 were also some that were jirotected by the forests, and next year all of these reported 

 large rainfalls. Now all of th>sse gauges were at a standard height in open regions 

 sucli that only the variations in wind proper or in the currents induced by neigh- 

 boring obstacles could conceivably aflect the catch of the gauge; moreover the diff"er- 

 ences between the stations were greatest in the winter and least in the summer 

 months and all the study of the configuration of the ground around the stations tends 

 to show that the diff"erences in the catches of the gauges were due to the irregular- 

 ities of horizontal distribution of the strength of the wind as influenced by the sur- 

 roundings. In other words instead of studying the geographical or horizontal distri- 

 bution of the total annual rainfall it was safe to assume that that had been uniform for 

 each year over this small area, and that we are studying simply the horizontal distri- 

 bution of a deficiency in catch or a rain-gauge error due to very local winds at the 

 mouths of the gauges. 



Tap.lk VII. 



This conclusion is confirmed by examining the records in the summer months 

 separately from those iu the winter. Local showers are frefiueut during the summer 



