Takhtajan, A., Crovello, T.J., Cronquist, A. (1986). Floristic Regions of the World. University of California Press, Berkeley, CA.
Von Marilaun, A.K. (1863). Das Pflanzenleben der Donaulander. Wagner, Innsbruck.
Wagner, A. (1844–1846). Die geographische Verbreitung der Saugethiere. Abhandlungen der Königlich Bayerischen Akademie der Wissenschaften. Mathematisch-Physikalische Klasse, 4(1), 1–146; (2), 37–108; (3), 1–114.
Wallace, A.R. (1876). The Geographical Distribution of Animals: With a Study of the Relations of Living and Extinct Faunas as Elucidating the Past Changes of the Earth’s Surface. Macmillan, London.
Zimmermann E.A.W. (1777). Specimen zoologiae geographicae, Quadrupedum domicilia et migrationes sistens. Theodorum Haak, Leiden.
Zimmermann E.A.W. (1778–1783). Geographische geschichte des menschen, und der allgemein verbreiteten vierfüssigen thiere, 3 Volumes. Weygandschen Buchhandlung, Leipzig.
1 1. For a detailed account of the history of 18th and 19th century biogeographies, see Ebach (2015).
2 2. I avoid using the phrase “Father of”. Coining a term does not justify ownership of the whole field. Herman Jordan and Clinton Hart Merriam used the term only once. Jordan casually refers to the term as though it was already in use, and Merriam uses it to describe his “Bio-geographic map”.
3 3. Later, Gareth Nelson was to note that “the concepts of station and habitation are important in Candolle’s view, for they define two different sciences, which persist into the modern era […]. No matter, the terms as used by Candolle, have modern counterparts: ecological and historical biogeography. Ecological biogeography is the study of stations; historical biogeography, the study of habitations” (Nelson 1978, p. 280, footnote 31, 281).
4 4. The Tableau lacks the actual lines, but instead has a table on either side of the cross-section depicting the temperatures at elevation. Essentially, Humboldt has created a sophisticated isothermal line.
2
Analytical Approaches in Biogeography: Advances and Challenges
Isabel SANMARTÍN
Real Jardín Botánico, CSIC, Madrid, Spain
2.1. Introduction
The last decades have seen an explosion of analytical approaches in biogeography. Amid the plethora of new methods competing for attention, a researcher is expected to get lost. From parsimony-based cladistic and event-based biogeography, we have moved into the expanding world of parametric model-based methods. This chapter mainly focuses on the latter, which are less than a decade old, but I also review previous approaches, as they provide a background on the shifting focus from phylogenetic relationships and Earth history to the integration of other disciplines (ecology, paleontology and population genetics), to understand historical processes that shaped Earth’s biodiversity.
2.2. From narrative dispersal accounts to event-based methods (EBM)
As seen in Chapter 1, the introduction of the idea of evolution (Darwin 1859) provided early biogeographers with an explanation of why geographic regions sharing similar environments harbor different biotas. Since continents did not move their positions over geological time, it must be organisms that moved over Earth’s surface to achieve their present distribution. These early biogeographic reconstructions were mostly narrative dispersal accounts. A paradigm changed in the mid-twentieth century with the introduction of the concept of “plate tectonics” (the 1960s). If a global process is responsible for the current distribution of biodiversity on Earth, we should see its effects in the form of congruence in biogeographic relationships across species. This general process was termed vicariance. The first analytical school, cladistic biogeography, aimed to find general patterns of relationships among areas of endemism, indicating shared biogeographic history (“area cladograms”, Nelson and Platnick 1981). Dispersal and extinction were considered processes that depend on biological characteristics intrinsic to the species and, therefore, cannot generate shared distribution patterns (Humphries and Parenti 1999). Cladistic biogeographic methods are allegedly process-free: inference of the area cladogram is done with no consideration to the biogeographical events that may have generated the pattern. If any, these are inferred a posteriori through comparison of the area cladogram with the individual species patterns (Brooks 2005). Uncoupling the inference of biogeographic patterns from the underlying evolutionary processes made it difficult to compare alternative biogeographic scenarios (Sanmartín 2012).
The next biogeographic school was “event-based biogeography” (EBM, Ronquist 1997, 2003). Biogeographic processes or events are tied to weights or “costs”, and the analysis consists of finding the pattern of area relationships with the minimum cost in terms of these processes. Four biogeographic events are considered in EBMs (Figure 2.1): vicariance, duplication, dispersal and extinction. The last two are tied to a speciation event and have also been termed “partial dispersal”, or “sorting, extirpation, and range contraction” for partial extinction. Within dispersal, we may distinguish “jump dispersal”, where a lineage migrates from one area to another (A to B) followed by speciation, and “range expansion”, where a lineage expands its range, leading to a temporally widespread distribution (A to AB); the latter is termed “geodispersal” when it affects multiple lineages (Lieberman 2003). Two biogeographic events are not considered in EBMs because they leave no observable traces in the phylogeny, that is, no descendants survive in the ancestral range (Sanmartín 2012): “full dispersal”, colonization of an area that is not followed by speciation, and “full extinction”, when the lineage entirely disappears from its ancestral range, that is, lineage extinction (Figure 2.1).
Figure 2.1. Biogeographic processes. Four types of biogeographic processes are considered in event-based biogeography: vicariance (allopatric speciation in response to a general dispersal barrier affecting multiple lineages); duplication (speciation within the area, i.e. sympatry, or allopatric speciation in response to a temporary dispersal barrier); extinction (the disappearance of the lineage from part of its ancestral range); dispersal (colonization of a new area by crossing a pre-existent barrier). Two processes: full dispersal and full extinction (right) cannot be modeled by EBMs because they leave no observable traces in the phylogeny
2.2.1. Parsimony-based tree fitting
The two most popular EBMs are parsimony-based tree fitting and dispersalvicariance analysis (Ronquist and Sanmartín 2011). Parsimony-based tree fitting, implemented in the software TreeFitter, was born from methods used in host–parasite tree reconciliation, with which it shares many similarities (Ronquist 2003). In tree fitting, a taxon cladogram is fitted onto the area cladogram by searching for the sequence of events that explain the tip distributions according to the area cladogram and with the minimum cost. The area cladogram may be a hypothesis of relationships based on geological history, in which case we measure how much the observed distributions depart from geologically predicted vicariance (Sanmartín and Ronquist 2004). Alternatively, it may be an unknown parameter, in which case, tree fitting consists of finding the area relationships and the sequence of biogeographic events that explain the tip distributions in the phylogeny with the minimum cost. Searching for the optimal (minimum-cost) area-cladogram implies enumerating and fitting