64 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 94 



where, at least in the last few years, the severe drought which affected 

 our Northwestern Central States was less severe, or even absent. In 

 the levels of Lake Ontario several minor features by which the yearly 

 ranges have been decidedly modified seem to be repeated each 23 years. 

 These features have been marked in figure 25 with letters. 



It is especially interesting, in view of caption 14-g, that the additional 

 cycle for Lakes Huron and Erie furnished by Mr. Horton seems, 

 when studied in connection with those commencing in 1883 ^^^ i" 

 1929, to indicate a 46-year cycle. The first and third, and so much as 

 has elapsed hitherto of the fifth 23-year cycle in the levels of these two 

 lakes since 1837 indicate a much more conspicuous and long-continued 

 low after about the fourth year than do the second and fourth cycles. 



It is not necessary to dwell upon the association which these figures 

 seem to bear to the drought in Northwest Central United States in 

 recent years. The inference, if the 46-year hypothesis is sound, is ob- 

 vious, and disquieting for the immediate future. 



20. A Fishery Test of the 23-YEAR Hypothesis 



Dr. Paul Bartsch, of the United States National Museum, suggested 

 to me that since ocean fishes live upon plankton, largely a vegetable 

 product, then if the weather is governed by 23-year cycles, the fish 

 food would probably be subject to related changes in its abundance. 

 Hence the fish population, as reflected by the annual catch, might 

 vary by 23-year cycles. On my application through the Bureau of 

 Fisheries, Dr. O. E. Sette was good enough to supply Fishery Circular 

 14, issued in 1933, and Bureau of Fisheries Document No. 1034, is- 

 sued in 1928, which give, respectively, the catches of mackerel and 

 cod taken since 1804. The catch of mackerel I read off from figure i 

 of the first cited document. The catch of cod is taken from table 2 

 of the other. 



Very great changes of scale in the mackerel catch occurred after 

 1816 and after 1885. In order to make the data fairly comparable, I 

 omitted values of the mackerel catch 1804 ^o 181 6, inclusive, and I 

 multiplied the values recorded from 1886 to 1931, inclusive, by the 

 factor 3. Five 23-year cycles remained for examination. No distortion 

 of the 23-year cycles is produced by the alteration of scale at the date 

 just noted, because it occurs at the beginning of a cycle. 



As for the cod, the catch reported was considerably smaller during 

 the first half of the nineteenth century than since. In order to make 

 my data more comparable, I omitted the years 1804 to 1811, inclusive, 

 and multiplied the values from 1812 to 1857, inclusive, by the factor 

 5/3 (again making the change of scale at the beginning of a cycle). 



