Modern house mice taxonomy can be regarded to have begun with a heroic review by Ernst & Henrietta Schwartz (1943), who studied around 5000 specimens in museums from around the world. They reduced a confusion of over a hundred different taxa to a single species, Mus musculus with four wild subspecies and a number of commensal forms arising from them. This has proved over-simplistic. Five species are now recognized in Europe and the Near East alone, of which the most widespread is Mus musculus itself, divided into several subspecies. The most important and widespread of the latter is M. m. domesticus, the west European and North African form. Many of the taxa can interbreed, with various amounts of introgression but without complete fusion. The hybrid zone between the light-bellied musculus (which is a nomen nudum, but is regarded as the typical form because it is found around Linnaeus’ home in southern Sweden) and the more southern and western dark-bellied domesticus has been studied in depth. As domesticus retains its integrity in this situation, it may be proper to regard it as a full species.
Further developments have emerged, as new techniques became available (Berry & Scriven 2005). In 1969, Alfred Gropp found that mice living in an isolated Alpine valley (originally described as Mus poschiavinus, largely on the grounds of their darker colour than normal) had 26 chromosomes rather than the typical diploid complement of 40, the difference in number being the result of seven Robertsonian fusions. Many chromosomal races have since been found largely, but not entirely, in Italy and Switzerland (Piálek et al. 2005). Some of these races are limited in their range, but the discovery of six races on Madeira, an island 57-km long by 22-km wide, 741 km2 in extent which cannot have been occupied by mice for very long, was extremely surprising.
Around the time that Gropp reported the discovery of Robertsonian chromosomes, Robert Selander applied electrophoretic techniques to proteins from population samples of mice from barns in Texas. He found that house mice were among the more variable mammals and that his molecular data threw light on their social structure. This pioneering work has now expanded into an ever-increasing repertoire of molecular tools.
These karyotypic and molecular results have been complemented by a growing knowledge of mouse ecology. Early studies pointed to house mice having a home range of only a few tens of square metres, but mark-release-recapture and gene flow studies have shown that individuals may reproduce after moving a considerable distance (several hundred metres in a short time). However, churning in particular populations is dwarfed by the readiness by which new populations are established by small groups carried as human commensals (Pocock et al. 2005). Phylogeographical studies of house mice are now appearing at an accelerating rate (e.g. Rajabi-Maham et al. 2008; Searle et al. 2009); we are now beginning to gain an overall picture of the genetic patchwork that is Mus domesticus (plus its introgressing congeners) today.
The mice of Madeira are showing the potential value of such studies. The original report of chromosomal races speculated that the differentiation had arisen from drift in isolates in the steep valleys radiating from the old volcanoes, which rise to a height of 1862 m (Britton-Davidian et al. 1990, elaborated in Britton-Davidian et al. 2005) – a suggestion reminiscent of that made by Henry Crampton to account for differentiation in Partula populations, which he studied for almost 50 years on the Society Island of Moorea, a hypothesis comprehensively demolished by Johnson et al. (1993). More detailed studies of the Madeiran mice, extending to allozyme and mtDNA analyses, indicated no lack of diversity, implying that the colonization was by a large number of founders or that there had been more than one colonizing event (Gündüz et al. 2001; Britton-Davidian et al. 2007). The authors favoured the latter explanation, not least because the island mice allozyme comparisons showed them to be more closely related to animals from their likely source on mainland Portugal, while their mtDNA haplotypes suggested immigrants from northern Europe.
This conclusion is strengthened by a paper in this issue (Förster et al. 2009). Previous comparisons largely lacked information about variation in the genomes of Portuguese mainland mice; they were based on only nine mice from central Portugal. This new study fills this gap with 76 animals caught at 14 sites in Portugal, mainly from ports which had historical links with Madeira (Fig. 1), allowing more precise comparisons with samples from Madeira and its neighbouring island of Porto Santo. The authors found four mitochondrial (D-loop) lineages in the mainland animals, one of which dominates. The diversity of the samples in Portugal was similar to that found elsewhere (in Italy, Greece and Turkey). They concluded that there were probably multiple colonizations into Portugal, with a most likely common ancestor from the eastern Mediterranean area. Estimates of expansion time concur with zooarchaeological findings that mice reached the Iberian Peninsula at least 2500 years ago.
The island mice are very different. Only one animal had a haplotype of the dominant Portuguese clade; 99% of the mice had sequences common in Denmark, Sweden and Germany, more than half having haplotypes identical with those found in northern Europe. Estimates of the time of expansion of the Madeiran population are imprecise, but suggest a date more than 900 years ago, much earlier than 1419 ad, the traditional date of discovery of Madeira by Portuguese voyagers. However, the previous finding that the nuclear genome of the island mice is much more like the Portuguese variation than north European complicates this picture (Britton-Davidian et al. 2007). Förster et al. suggest that it might be explained by the asymmetric spread of nuclear genes following introduction of mice from Portugal into an already established ‘north European’ population, along the lines of the situation uncovered by Jones et al. (1995) in an introduction experiment on the Isle of May (Scotland).
The Förster et al. paper shows how molecular analyses can throw light on unexpected ecological processes as well as historical events. The Madeiran mice have yielded some surprising results and there is plenty more to discover about the evolutionary events on the island.