The geographical distribution of life and the problem of regionalization: 100 years after Alfred Russel Wallace
Year: 2013
Type: article
Abstract: This editorial marks the collation of a set of papers published in the Journal of Biogeography between 1983 and 2013 into a virtual issue highlighting the enduring legacy of Alfred Russel Wallace (1823–1913) within biogeography. Widely regarded as one of the founding fathers of biogeography, Wallace died in Dorset, England, on 7 November 1913, in a house he designed and built himself (Beccaloni, 2011). By the time of his death Wallace had a global reputation that went far beyond his fame as the co-discoverer of evolution by natural selection (http://people.wku.edu/charles.smith/index1.htm). His popularity during his lifetime derived from a vast breadth of interests, including his core scientific work in evolutionary biology, biogeography and glaciology, but also the more eccentric and difficult to appreciate interests in astrobiology, spiritualism and land reform (Smith, 2011). He was a prolific writer, and several of his fundamental contributions to biogeography were also accessible to an educated public, resulting in books such as The Malay Archipelago (Wallace, 1869), The Geographical Distribution of Animals (Wallace, 1876) and Island Life (Wallace, 1880) becoming popular and broadly read. Revolutionary technical scientific papers included the 'Sarawak' paper on speciation (Wallace, 1855) and the 'Ternate' paper on natural selection (Wallace, 1858). A vast array of additional insights into Wallace's thinking is to be found scattered within his many letters, essays and book reviews spanning from 1843 until his death (see http://people.wku.edu/charles.smith/index1.htm). Some commentators feel that history has not been particularly kind to Wallace's legacy – relegating him to the role of Darwin's less inspired co-founder of the theory of natural selection (Gould, 2002) – perhaps a consequence of his turn towards teleological thinking and more metaphysical interests in his later years (Birx, 1998). Nevertheless, over the past decade or so, many scholars have attempted to resurrect Wallace's reputation as a leading Victorian scientist (Shermer, 2002; Smith, 2005; Smith & Beccaloni, 2011). The centenary of his passing has been marked by a number of commentaries (e.g. Heaney, 2013; Knapp, 2013) and by the publication of two new analyses of global zoogeographical regions that have explicitly paid tribute to Wallace's legacy as the creator of the first widely adopted map of global regions (Holt et al., 2013; Rueda et al., 2013). Rather than attempting a comprehensive overview of Wallace as a scientist and activist, here we highlight his enduring contribution to the modern science of biogeography. We do so by introducing a virtual issue of the Journal of Biogeography in which we have gathered a series of key contributions published between 1983 and 2013 on a matter to which he devoted much attention – the problem of regionalization. These papers illustrate the enduring debt we still owe to the foundational work of Wallace and his contemporaries. With the extraordinary growth of scientific data and publications in recent decades, much of this legacy of knowledge and ideas is no longer routinely attributed back to 19th century (or even earlier) foundations. However, it can often be salutary to spend time browsing the back catalogue to rediscover not only the forgotten gems and insights (e.g. Hooker, 1867; Brown et al., 2004) but also the many blind alleys. Wallace was both a highly original thinker and a great synthesizer and promoter of scientific understanding through books such as The Malay Archipelago (1869), The Geographical Distribution of Animals (1876) and Island Life (1880). To complete the set of papers highlighting his legacy we therefore include a few papers that are not primarily focused on the problem of regionalization. Reading sections of Wallace's books can lead to a rather surprising realization that many of his insights have a distinctly modern ring to them (see summary in Box 2.1 in Lomolino et al., 2010). Yet, he has been criticized retrospectively by several more recent biogeographers, notably Leon Croizat and the late 20th century vicariance biogeographers, as having been nothing more than a non-relevant old-school dispersalist (Michaux, 2011). In this context it is important to note that Wallace had no access to many of the foundational concepts (e.g. plate tectonics, Hennigian phylogenetics, phylogeography) or analytical tools (e.g. molecular systematics, stable isotopes, satellite imaging, geographical information systems, advanced statistics or mechanistic modelling) that we depend on today. It is thus intriguing to ask not only how relevant does his work remain to framing modern biogeographical scholarship, but also to what extent did his biogeographical syntheses and ideas withstand the test of time? Zoological regions are those primary divisions of the earth's surface of approximately continental extent, which are characterised by distinct assemblages of animal types. Though strictly natural… they have no absolute character as equal independent existences, since they may have been different in past ages, but are more or less conventional, being established solely for the purpose of facilitating the study of the existing geographical distribution of animals in its bearing on the theory of evolution. There is thus, in my opinion, no question of who is right and who is wrong in the naming and grouping of these regions, or of determining what are the true primary regions. All proposed regions are, from some points of view, natural, but the whole question of their grouping and nomenclature is one of convenience and of utility in relation to the object aimed at. Wallace (1894) was adamant in this paper that biogeographical regions provided a crucial framework for zoogeography and that it was unhelpful to be promoting multiple regionalization schemes for different major taxa or subordinate taxa. He did not welcome the notion of separate schemes for reptiles, butterflies, beetles, or for tribes or families being developed as competing schemes. Instead he wrote: '[l]aws of distribution can only be arrived at by comparative study of the different groups of animals, and for this study we require a common system of regions and a common nomenclature'. He then expanded on this by commenting that the purpose of the unified system of regions was to provide a single framework encapsulating the law of distribution so that, by working out the 'details in the different groups', we might identify the 'certain difficulties or anomalies' that require special explanation. He was also clear that regions could best be determined by analyses commencing with the distribution of terrestrial mammals '… supplemented in doubtful cases by that of the other vertebrates' (Wallace, 1876, p. 58). In short, he saw Schlater's scheme of six regions, and his refinement of it, as the expression of the general rules, the major natural subdivisions of the world the explanation for which '… can only be because during the more recent geological periods they have formed single more or less continuous areas, while separated either by geographical, climatal or biological barriers from the adjacent areas'. One can only imagine how this line of thinking would have been influenced had Wallace's view been informed by the modern plate tectonics model of Earth history. Wallace was aware of and discussed those parts of the world where the zoogeographical regions did not neatly fit the current distribution of continental boundaries and where transitional faunas could be discerned, but he felt it was premature to attempt definite naming of subdivisions on the basis of knowledge then available. (1) They should be founded upon, and approximate to, the great primary geographical divisions of the earth, which there is reason to believe have been permanent during considerable geological periods. (2) They should be rich and varied in all the main types of animal life. (3) They should possess great individuality; whether exhibited by the possession of numerous peculiar species, genera, or families, or by the entire absence of genera or families which are abundant and widespread in some of the adjacent regions. It is because I think that the future progress of a branch of biological study in which I take great interest will depend on our arriving at some uniformity of view as to this question of zoological regions, that I have devoted so much space to its discussion. Determining, refining and applying zoogeographical regions have been enduring themes of the biogeographical research agenda since these words were written. Within conservation practice, for example, zoogeographical regionalization schemes (generally closely similar to the Sclater–Wallace system), have been integral to the development of global and continental-scale systems for strategic conservation planning purposes in both terrestrial and marine realms (see discussions in Ladle & Whittaker, 2011; Briggs & Bowen, 2012). The regions thus hold continuing relevance to pure and applied biogeography. To highlight recent developments in the study of biogeographical regions, we turn now to our featured articles. Our first featured paper is by Charles H. Smith (1983a), who undertook an analysis based on the distribution of 115 families of mammals across the 24 subregions of Wallace's (1876) system. Smith undertook a pioneering analysis designed to assess the efficiency of this system and whether it provided a logical set of relationships among the subregions, i.e. how well they nest within the higher level regions. He constructed a similarities matrix of the subregions and applied an ordination technique, non-metric multidimensional scaling, in combination with classification techniques, to determine an optimal set of regions and subregions. This analysis produced a novel set of four regions, subdivided into a set of 10 subregions, which displayed significant differences from Wallace's original system. Smith (1983b) wrote an innovative companion paper which set out to assess the proposition that evolution might be regarded as a stochastic spatial process with inter-regional affinities explicable through chance interactions between the subregions and a deterministic distance-decay effect on the diffusion of evolutionary innovations. This approach was in some ways rather ahead of its time, as the computer power and computerized distributional datasets required to develop this line of thinking further were lacking in the early 1980s and this, combined perhaps with a somewhat impenetrable writing style, may explain the lack of attention that has been paid to Smith's efforts and especially to the 1983b paper. By way of contrast with the computational/analytical approach taken by Smith (1983a,b), our next featured article by Barry Cox (2001), published 18 years later, took the form of an expert synthesis in which the author reviews both existing zoogeographical (essentially updating Wallace's map) and phytogeographical regionalizations (updating the scheme proposed by A. L. Takhatajan, 1986), focusing especially on the comparison of the two. The six zoogeographical regions identified by Cox at the outset as the 'currently recognized set' comprise the Neotropical (South and Central America), the Nearctic (North America), the Palaearctic (Eurasia and the northern portion of Africa and Arabia), the African (Africa and Arabia south of the Sahara), the Oriental (Southeast Asia) and Australian. Cox (2001) proposed two main modifications to these regions: first, redrawing the boundary between the Palaearctic and the African region more or less along the boundaries of the African continent (contrary to Wallace); and second, withdrawing the scope of the regions back to the edges of continental plates, or to the low point of Pleistocene sea-level minima. Cox's solution to the long debate about the placing of Wallace's Line, which bisects Bali and Lombok, was to remove the oceanic islands between 'mainland' Southeast Asia and the Sahul landmass of the Pleistocene from the area encompassed by the adjacent regions, and to follow an earlier proposal to name this transitional area 'Wallacea'. Cox's major modification to the earlier phytogeographical schemes was to drop the tiny but notable Cape Region of southern Africa from the top level of the classification, giving the following set: Holarctic, African, Indo-Pacific, South American and Australian regions (note that Cox follows prior practice in using the term kingdom instead of regions for the top level for plants). The relegation of the Cape to a second tier level in the regionalization could conceivably have implications for the treatment of this area in global conservation prioritization schemes, but judging from the frequency of use of 'Cape Floristic Region' in the scientific literature, this proposal has been widely ignored. Cox's restriction of the zoogeographical, mostly mammal-based, regionalization to the continents contrasts with his inclusion of oceanic regions in the phytogeographical regionalization. Of course, as Cox recognizes, many oceanic islands display the influence of multiple colonization sources and are thus less easily assigned to a particular region than are continental areas (Whittaker & Fernández-Palacios, 2007). For recent advances in network analysis of island regions in relation to biogeographical dynamics, linking pattern to process, see Carstensen & Olesen (2009) and Carstensen et al. (2012). And, for recent advances in the use of phylogenetic data to refine the positioning of Wallace's Line, see Esselstyn et al. (2010). Cox's (2001) update has been well cited, but so too has a response by Morrone (2002), who in a very short critique proposed an alternative regionalization scheme. Morrone's analysis was rooted in a personal interpretation of cladistic and panbiogeographical studies published by other authors. He claimed that these analyses supported a fairly radical redrawing of the boundaries of the major regions, fusing zoogeographical and phytogeographical evidence together to form a single biogeographical scheme. Morrone's scheme differs from others in seeking evolutionary connections produced through the vicariant history generated by the Earth's tectonic history. This leads, for example, to his splitting Wallace's original Neotropical region into an Andean region with affinities to western Gondwanan terranes, and a Neotropical region with affinities to eastern Gondwanan terranes. The contrast between these two reviews reveals something of the gulf in philosophy, theory and method that still persists between different schools of thought within biogeography. There is clearly still a place for expert reviews and syntheses in the field of biogeographical regionalizations, as illustrated by the latest recasting of marine biogeographical regions by Briggs & Bowen (2012). However, as Cox (2010) has commented, more recent analytical, computationally intensive regionalization schemes provide a more satisfactory basis for advancing towards an agreed set of global maps as they generate a quantitative basis for drawing the boundaries and for determining at what level of the hierarchy of regions the distinctions occur: a concern that motivated Smith's (1983a,b) analyses and to which Wallace himself devoted considerable attention (Wallace, 1876). Alternative maps produced within and across such analyses also generate insight into which assumptions, steps or parameters exhibit sensitivity in terms of outcomes (below). The next papers we feature are examples of such schemes, being motivated by the development of both improved availability of systematic distributional data for many higher taxa and by ongoing developments in analytical techniques and data handling capacity. Procheş (2005) illustrates the application of cluster analysis to the regionalization of bats, a group that might be thought unlikely to follow the boundaries between regions as closely as less mobile forms of mammals. In practice, the cluster analysis identified 10 groups showing similarities with elements of both previous zoogeographical and phytogeographical mapping exercises. Procheş also usefully reviews previous regionalizations and, rather than promoting his scheme as replacing previous schemes, he follows Wallace in explaining how the differences between taxa or even ecological classifications of species may be used to 'partial out the role of dispersal abilities, body size, evolutionary age, etc., in determining global distribution patterns' (Procheş, 2005, p. 607). There have been three new attempts in the last three years to generate global regionalization schemes, two published in the Journal of Biogeography, and featured here. The first paper, by Kreft & Jetz (2010), illustrates something of a step-change in both data resolution and analytical power. Their analysis, based on a global database of species distributions of mammals, employs a combination of ordination (using non-metric multidimensional scaling) and cluster analysis, with the latter providing a means to generate a hierarchical representation of groups and their degree of relationship to one another. They also illustrate how switching the temporal/evolutionary 'depth' of the analysis from families, to genera, to species, can generate subtle but significant differences in the outcomes of analyses, as can distinguishing volant from non-volant species – very much as envisaged by previous papers reviewed herein (e.g. Wallace, 1894; Procheş, 2005). In a recent paper published in Science, Holt et al. (2013) take this approach a step further in their analysis of amphibians, birds and mammals (both separately and in combination). Again, they use cluster analyses and species-level distributional data (for 21,037 species), but complement this approach with algorithms that take into account the phylogenetic relationships between species. This, they claim, is consistent with Wallace's intent that the regions of the world express ancestral relationships among species, although in practice their means of attaining this is rather different from the approaches open to Wallace (1894), when he wrote that the '… intelligent study of the distribution of animals… derives its chief interest and importance from its relation to the theory of organic evolution… [and] must therefore include the comparative distribution of the various classes, orders, and minor groups'. The latest contribution to the problem of global regionalization is by Rueda et al. (2013). They employ an innovative approach to classification, termed affinity propagation, based on range map data for birds, mammals and amphibians, but excluding bats and migratory birds. They argue that according to Wallace's criteria for zoogeographical regions, the appropriate taxonomic resolution to use is the genus rather than species or families. they focused mostly on analyses but these with those for species and They the to Wallace's original scheme in their they found regions for mammals and birds as to the six of Wallace and the six they found for amphibians, the additional boundaries fairly closely to subregions in Wallace's scheme. The to a particular taxonomic level is within quantitative analyses of this but it is that in The Geographical Distribution of Animals Wallace that the (or possession of particular families and genera and the absence of families and genera were the key of the primary zoological regions, that he did not only using a single taxonomic level families, genera or to determine regions. These recent papers illustrate how the and resolution and taxonomic of and the and character of the analytical can generate of alternative regionalization schemes, with differences of too numerous to here. The which include (1) areas of a transitional to past of and (2) in the level at which boundaries are and may be important within a conservation biogeographical & Whittaker, 2002; Ladle & Whittaker, 2011; Briggs & Bowen, 2012). of these studies illustrates that the problem of biogeographical regionalization have a single solution and that it may be appropriate to different schemes for different or at to sensitivity analyses to determine the of the However, perhaps the feature of these schemes is the degree of between them and the various schemes, including Wallace's (1876) This is that the 19th century schemes were with much more and data and the and tools of the modern It will be to see which of these recent schemes Kreft & Holt et al., 2013; Rueda et al., 2013; Briggs & Bowen, with the of time or whether they will be as new forms of analysis to be developed et al., 2013). In our of regionalization we feel it important to note that there are other ways to the that could be regarded as biogeographical and are different from the regionalizations discussed thus One such system is the ecological of the world into or major types. for many purposes in land and such classifications may be of or may required data the regions, as is the within the scheme (see in Ladle & Whittaker, 2011). In of the and and distinctions between and regionalization schemes, we have included a paper by et al. a different to the other papers this analysis can also be regarded as following on from the work in the 19th century by Wallace and their contemporaries. We have focused on the problem of regionalization because first, as Wallace this is a fundamental step in understanding the of and in mapping at a or global and second, because the Journal of Biogeography has some of the key papers on this over the last three Of course, Wallace's influence is far more within biogeography. Lomolino et al. Box as many as biogeographical by Wallace that These (1) the classification of islands into continental continental and which a key point of many island biogeographical analyses (e.g. et al., (2) that dispersal is not only but as a means of islands et al., in while (3) the distributions of species lacking dispersal evidence of past land connections line of thinking that to for biogeography in the early 20th century as land across were but which was used more by Alfred as he developed his theory of continental and which was later to a major feature of the late 20th century vicariance biogeography & that of does not determine the degree of affinity between the or of and of the of and history in of Wallace's et al., is the closely with Wallace himself – and its a key role in his understanding – Wallace also had important working in the It was for this region that he the which that differences between closely forms on of barriers could be attributed to the role of the in the on either (Wallace, other of especially the Pleistocene took in the of the 20th some recent analyses – such as the paper featured in the virtual issue by et al. – have found for the role of as Wallace Wallace's work was by no means to the problem of understanding the of for the biological the of his contribution to this problem was and highly and is recognized as such among It was in tribute to this of his legacy that the term was to to the problem of data on distributions of species (e.g. Lomolino et al., Ladle & Whittaker, 2011). while we now have much resolution on species distributions than we did in the 19th it the that our analyses can be to, or the still and of our data on The paper featured in this virtual by et al. neatly the of such on our understanding of and of species illustrated with data for from Their paper, while only to Wallace, is included in this set because it how much work to be in the continuing to refine our knowledge of the of distribution from which we in Wallace's to much of interest about the of the natural The virtual issue is as an easily set of papers the system. It a of the papers in this article that were published in the Journal of published prior to are only by the
Source: Journal of Biogeography
DOI: 10.1111/jbi.12235
Language: English
Subfield: History and Philosophy of Science
Field: Arts and Humanities
Domain: Social Sciences
Open Access status: bronze
OpenAlex ID: W2019002046