ABSTRACT
The classic population concept that can be traced back to Malthus (1798) and which was developed in ecology and evolutionary biology in the early part of the 20th century (McIntosh 1985) assumes that all individuals interact equally and share the same environment. This is clearly not true for real populations at spatial scales greater than the daily movement range of individuals, for the simple reason that interactions become restricted by the physical distance. At larger spatial scales, habitat heterogeneity is typically another factor, apart from just long distances, that influences population processes and thereby population structures. Viewed from the perspective of a particular species, a landscape may be heterogeneous in many different ways. It may consist of several different types of habitat that may be used for foraging and reproduction, but typically in this case there is spatial variation in habitat quality: not all habitat types are the same. The classic metapopulation concept (Levins 1969; Hanski 1999) assumes that there is just one type of habitat, but it is fragmented into discrete patches. Incorporating spatial structure into population studies and into population models has been a major goal in ecology and evolutionary biology for the past 20
years (Tilman and Kareiva 1997; Hanski 1999), with sporadic early contributions published since the 1930s (Wright 1931; Nicholson 1933; Andrewartha and Birch 1954; Huffaker 1958).
