Figure 23. — Upper: Maryland marsh in March 1938, at the peak of the nmskrat- 

 population cycle as indicated by density of houses and badly eaten condition of the 

 marsh. Lower : Same marsh at low of the muskrat cycle, April 1944. 



so plentiful that they deplete their 

 food supply. During these years of 

 abundance their numbers build 

 gradually to a peak, but not simul- 

 taneously over the country, and 

 then abruptly decline (fig. 23). 

 Such major fluctuations cannot be 

 attributed to trapping variation. 

 Canadian fur records accumulated 

 over a period of many years show 

 that fluctuations in muskrat popu- 



lations tend to move in more or less 

 definite 10-year cycles. More recent 

 but incomplete records from various 

 parts of the ITnited States confimi 

 the fact that distinct periods of 

 abundance and scarcity exist, but 

 indicate that the cycle in some in- 

 stances is longer than 10 years. 

 These cycles are not fully under- 

 stood at present, but undoubtedly 

 the two major factors responsible 



17 



