during the winter, especially if low leniperalure ob- 

 tains, and may intUicnce southward distribution of 

 northern animals in the summer. The length of the 

 season between spring and autumn killing frosts, or 

 between dates when its limits of temjierature tolerance 

 are reached in the spring and autumn, may determine 

 whether or not a species can complete its life cycle at 

 a given latitude. .A corollary consideration is whether 

 the accumulation of heat is sufficient to furnish cold- 

 blooded animals and plants sufficient energy for 

 growth and reproduction. 



Biotic barriers consist of changes in vegetation, 

 food, competitors, and jiredators. The adaptations and 

 behavior patterns of many animals fit them to niches 

 in specific types of vegetation ; should the previously 

 amenable vegetation change, the animals may have 

 great difficulty in ada])ting to it. Tree squirrels, for 

 instance, are replaced by ground squirrels in prairies 

 and deserts. Most animals have a great adaptibility 

 to food, so food is not so limiting a factor. Some 

 insects, however, such as certain aphids, are narrowly 

 limited to particular species of plants as a source of 

 food. Where their food plant is not present, they can- 

 not exist. 



Competition between species is also a potent force 

 in controlling distribution. The boundary between the 

 ranges of the house wren and Bewick's wren in eastern 

 Xorth America is not sharply defined, varying as a 

 function of competition between the two species. 

 Either species can live in the range of the other, but 

 in the Xorth, the house wren usually wins in com- 

 petition for territory and nest-sites. In the South, the 

 Bewick's wren is more successful (Kendeigh 1934). 

 Predators, such as the great horned owl and the 

 swifter hawks, tend to urge the bobwhite to confine 

 itself to a forest-edge habitat, where it is less vulner- 

 able to attack. Trypanosome parasites carried by the 

 tsetse fly are effective barriers against successful in- 

 troduction of domestic ungulates in certain parts of 

 Africa, and the rabbits introduced into Australia have 

 limited the range and greatly reduced the abundance 

 of several species of native marsupial. 



Dispersal nf youns; 



The dispersal of a species is primarily accom- 

 plished in the immature stages. This is obviously true 

 of eggs and spores, but banding and marking studies 

 have shown that among the higher animals — birds and 

 mammals — it is also the young which disperse the 

 species. Once a bird has reached sexual maturity and 

 nested, it has strong tendencies to return to the same 

 area in following years. The distribution of young 

 birds is not random, however, as they tend to return 

 to the general vicinity of their birthplaces rather than 



uniformly over the range of the species. Thus only 

 0..S per cent of ^^7 adult house wrens recovered a year 

 after banding (Kendeigh 19411)) nested farther than 

 3.3 km (2 miles) from the .site where they had nested 

 the year of banding, but 15 per cent of the 182 birds 

 hande<l as nestlings were recovered at greater dis- 

 tances, the longest of which were ^2 km (20 miles), 

 .^6 km (3.=; miles). SO km (.=^0 miles) and 1120 km 

 (700 miles). Dispersal distances for young of other 

 species are proportionally comparable (Haartman 

 1949). 



Of small mammals, it is characteristic that once an 

 individual has selected a homesite, it rarely leaves it 

 for another (Burt 1940, Blair 19.S3). It was observed 

 that in the months following tJie time at which they 

 had been cajjtured, marked, and released, 95 per cent 

 of 133 adult woodland white-footed mice resumed 

 habitation witiiin 183 m (200 yd) of the site of cap- 

 ture. 



Rate of dispersal 



If dispersal from birthplace were typically lim- 

 ited to one direction, then a simple mean of the dis- 

 tances to which the young disperse before they breed 

 would give the dispersal rate per generation. It is the 

 case, however, that dispersal proceeds peripherally in 

 all available directions and may extend to surprising 

 distances (Bateman 1950). 



The area exposed to invasion and the average time 

 required to saturate that area increase proportionally 

 as the square of the linear distance (d-) from the 

 center, since the total area within which the individual 

 could settle is nd'-. Therefore, the equation 



for computing mean dispersal distance seems correct 

 (Haldane 1948), although other equations have been 

 suggested (Haartman 1949, Burla et al. 1950, Dice 

 and Howard 1951 ). Consider the data on house wren 

 nestling recovery distances, presented above. Exclud- 

 ing the truly extraordinary distance of 1120 km, and 

 observing that only about 93 per cent of young female 

 wrens nest when they are one year of age, we compute 

 by this equation an annual dispersal rate of approxi- 

 mately 8 km (5 miles) for this species. The mean dis- 

 persal distance of one group of 154 young woodland 

 white-footed mice (Burt 1940), according to the 

 above equation, is about 176 m (192 yd). However, 

 mice born in the spring mature sexually very quickly 

 and breed in late summer or autumn, so the annual 

 dispersal rate must be somewhat greater than this 

 figure indicates. 



Dispersal, migration, and ecesis 



149 



