338 



POPULATIONS 



the temperature rises above 26.6°, and is 

 accompanied by a saturation deficiency of 

 over 8 mm." (p. 149). The climatic controls 

 of the flea populations in question may also 

 involve other stages of the Ufe history. 



Reference should be made to the study 

 of Holdaway (1932) dealing with Trifoo/mm 

 confusum. This aflFords another example of 

 the eflFect of humidity on population size. 



PRECIPITATION: EXCESS AND DEFICIENT 



The relation of rainfall, as well as other 

 climatic factors, to the development or sup- 

 pression of locust outbreaks has been the 

 subject of considerable attention in various 

 parts of the world. The topic is too com- 

 plex, and the literature too extensive, to re- 

 view in detail, but a particular instance 

 can be sketched to illustrate how the 

 amount of precipitation functions as a con- 

 trolling, density-independent factor. These 

 observations are taken from a paper by 

 Key (1942) dealing with the Australian 

 plague locust, Chortoicetes terminifera. 



A most severe outbreak of this species 

 took place in Eastern Australia in 1937- 

 1938, the second season of a cycle which 

 had its start in the spring of 1936-1937. An 

 area of 123,000 square miles was infested 

 in New South Wales, 28,000 square miles 

 in Queensland, and 1700 square miles in 

 Victoria. In the following year (1938- 

 1939), the incidence of the locust plague 

 was relatively unimportant. Ecologically, 

 the infested areas were diverse. As the 

 swarms migrate they tend to follow a hu- 

 midity gradient, and about 65 per cent of 

 the swarms fly toward and arrive in moister 

 regions than those they started from. The 

 rate of spread of the infestation is highest 

 at the start of the outbreak, at which time 

 it may exceed 100 miles a day. 



The life cycle of the locusts proceeds 

 without essential interruption only so long 

 as the efi^ective rainfall and temperature 

 remain within certain limits. These, then, 

 are two of the critical meteorological con- 

 ditions the eflFect of which on the popula- 

 tion is essentially density-independent. Key 

 illustrates this point in part as follows: "In 

 New South Wales, hatching of the over- 

 wintered eggs normally takes place when 

 the temperature rises above the threshold 

 for development, for the moisture condi- 

 tions at this time are usually favourable. 

 Interruption of the life-cycle by drought 

 quite frequently occurs during one or more 



of the summer months, in which case the 

 precise incidence of the dry period deter- 

 mines its eflFect on the population. If it oc- 

 curs during the hopper stage, heavy mor- 

 tality results, whereas if it commences dur- 

 ing the egg stage, the population is little 

 aflFected unless the dry period exceeds 

 three months" (p. 85). 



As is true of many ecological phenomena, 

 a final, particular event frequently de- 

 pends upon many preceding successful in- 

 teractions. This is true for locust outbreaks: 

 an outbreak does not develop unless a 

 number of conditions have been satisfied. 

 Thus, although rainfall is definitely impor- 

 tant, it is only one factor in a complex sys- 

 tem. Andrewartha (1940, p. 76) suggests 

 this point nicely in the following quota- 

 tion: 



"There is considerable evidence that favour- 

 able weather for several successive years is 

 necessary before a major outbreak can occur. 

 Swarms are likely to develop in the outbreak 

 area when rain is adequate during the warm 

 months. Two or more favourable seasons in this 

 way may be required to produce large or dense 

 swarms. Similar conditions are necessary for 

 swarms to develop in the intermediate breed- 

 ing areas. For the outbreak to continue its 

 development in the agricultural districts a dry 

 autumn is required. The whole sequence is 

 necessary for a major plague. The cycle may be 

 broken at any point; when this occurs the 

 incipient outbreak will be destroyed." 



Errington (1939) has studied the eflFect 

 of drought on muskrat populations in 

 Iowa. The summer of 1934 was a season 

 of severe drought, followed by dry years 

 in 1936 and 1937. Muskrat populations 

 during these three seasons are compared 

 with those of other, more normal, years. 

 Even when their habitats begin to dry out, 

 most of the muskrats remain in their orig- 

 inal home ranges. Some do move to new 

 areas and, by and large, are killed off at a 

 higher rate than those that stay behind. 

 As the drought becomes more severe a 

 number of events mav take place. There 

 is an increase in "intraspecific strife," which 

 usually assumes the form of accentuated 

 fighting between old and young of both 

 sexes. Fighting between males is prevalent 

 in the spring, primarily as a manifestation 

 of sexual unrest and aggression. Fighting 

 may be heightened after dry summers and 

 leads to incapacitating wounds and actual 

 death. This is brought about especially 



