CHANGES OF CLIMATE 467 



mate, known as the Bruckner cycle, after Professor Bruckner, of Berne, who 

 has made a careful investigation of the whole subject of climatic changes 

 and finds evidence of a thirty-five-year periodicity in temperature and 

 rainfall. Bruckner began with the long-period oscillations in the level 

 of the Caspian Sea. He then investigated the levels of the rivers flow- 

 ing into the Caspian, and next the dates of the opening and closing of 

 the rivers of the Bussian empire, and finally extended his study over a 

 considerable part of the world, including data concerning mean tem- 

 peratures, rainfall, grape harvest, severe winters, and the like. The 

 dates of opening and closing of Russian rivers go back in one case to 

 1559 ; the dates of vintage to the end of the fourteenth century, and the 

 records of severe winters to about 1000 a.d. In a cycle whose average 

 length is thirty-five years there comes a series of years which are some- 

 what cooler and also more rainy, and then a series of years which are 

 somewhat warmer and drier. Bruckner has found that the price of 

 grain averages 13 per cent, higher in the wetter lustrum than in the 

 drier. This thirty-five-year period is not to be thought of as being a 

 perfectly systematic recurrence, in exactly that term of years. The 

 interval in some cases is twenty years ; in others it is fifty. The average 

 interval between two cool and moist, or warm and dry periods, is about 

 thirty-five years. Moreover, not only the intervals, but the intensities 

 of the individual periods vary. The mean amplitude of the tempera- 

 ture fluctuation, based on large numbers of data, is a little less than 

 2°, which makes it greater than that obtained by Koppen for the sun- 

 spot period, and it is natural to expect it at a maximum in continental 

 climates. The fluctuations in rainfall, also, are more marked in in- 

 teriors than on coasts. The general mean amplitude is 12 per cent., 

 or, excluding exceptional districts, 24 per cent. In western Siberia 

 more than twice as much rain may fall in wet as in dry periods. 

 Regions whose normal rainfall is small are thus most affected. In 

 years of minimum precipitation they may become uninhabitable, and 

 the population may be forced to move away, perhaps never returning, 

 and allowing towns and irrigating works to fall to decay. Slight 

 fluctuations in rainfall are most critical in regions having a normal 

 precipitation barely sufficient for agriculture. The extent of land 

 cultivated, and the returns of agriculture here fluctuate directly with 

 the temporary increase or decrease of rainfall. A supplementary study 

 of the newer rainfall observations for Russia and for the United States, 

 as well as for certain stations in central Europe and eastern Siberia, 

 has given Bruckner satisfactory confirmation of his earlier conclusions 

 in the fact that he finds a decrease of rainfall over these districts as a 

 whole, beginning about the middle of the decade 1880-90. The time 

 of the ' boom' in western Kansas and Nebraska, and in eastern Colorado, 



