An accepted approach to elucidating patterns of historic biogeography would typically begin with either a Brooks Parsimony, or COMPONENT Analysis. This would require that the distribution information (Fig. 9) be converted into a supplementary matrix, which would then be subjected to further analysis. Though effective and widely accepted, these methods pose some conceptual disadvantages. The biogeographic analysis is performed on top of the results of a previous cladistic analysis. This imposes a number of new equally parsimonious trees on top of the supposedly equally parsimonious trees from the base analysis. In effect, one accumulates uncertainty on top of uncertainty.
There are, of course, ways to handle all these alternative trees, but there is an additional conceptual problem to this approach. This treats the biogeographic history as if it were a completely separate and subsequent set of events in the evolution of a group when in fact the spatial and distributional components are an integrated part of a taxon’s biology. One possible solution to this could be to attach a step matrix to the primary analysis that would code for all possible biogeographic movements within the recorded range of the group in question (Berge, 2000). The biogeography could then be integrated into the base analysis. This is worthy of further exploration within the Amphipoda, but must be the subject of a separate study.
A conceptually simpler approach, however, is suggested by the data at hand. The biogeographic history of the separate clades (Fig. 8, nodes A-G) can be projected onto paleomaps (see Ebach & Humphries, 2002). This is a process of straightforward inspection. But where to begin? Since our cladogram clearly points to an early development of the large fresh water African and Mediterranean cave species it leads us to look for a paleomap, in this case Triassic, that would combine both areas and show a continuous land mass with those areas present at some particular point in earth history. Actual fossil evidence for the existence of the Amphipoda does not go beyond approximately 40 million years ago (Coleman and Myers, 2000), when epigean as well as hypogean forms where trapped in amber resin. These Cenozoic fossils closely resemble the living forms of today, and therefore would seem to point to a much older amphipod origin perhaps one more in line with other groups within the Peracarida such as the isopods, tanaids, cumaceans and spelaeogriphaceans. Fossils of these latter groups date back to the Carboniferous (Schram, 1981).
Why is it that no Paleozoic or Mesozoic fossils are found among the Amphipoda? The reason is probably twofold. Either no recognizable form existed, i.e., somehow the amphipods arose from relatively recent ancestors, or the amphipods living in the Mesozoic (or earlier) occurred in habitats that were quite unsuitable for fossilization. These latter ancestors could be deep sea, interstitial, and/or cave inhabitants. We assume that the colonization of the marine interstitial by small benthic crustaceans was an ancient event. A trend can be seen in the evolution of worm-like bodies in amphipods, which invaded older areas in ancient times and are present today in areas that have been emergent at least since the late Cretaceous. This stands in contrast to forms with rounder bodies, which resemble more their benthic relatives (Vonk, 1990; Coineau, 2000).
The present day ingolfiellideans that occupy underground waters far inland on the continents are the large cave lake inhabitants of Africa below the equator. Their counterpart, in body length and micro-environmental requirements, is the one species of Metaingolfiella, known from a well in Italy. Returning to our original question of where to begin and project a common history on a paleomap, it seems justifiable to search for to a time where the African and Mediterranean freshwater forms could have had contact. The map from the Early Triassic (Scotese, 1977) reflects such a situation. The Pangaea and Gondwana geography of the Triassic can be fitted to the branches of those taxa that group together low in the clade of the ingolfiellideans (Fig. 10, Fig. 11A).
Fig. 10. Cladogram with paleogeographic characterstates projected on its branches. Tr. = Triassic; EJ = early Jurassic; LJ = Late Jurrasic; LK = Late Cretaceous.
“Above” these stem taxa on the tree, we encounter species of Ingolfiella found today in groundwater habitats along the North Mediterranean coasts. As above, if we seek a period after the Triassic, we find in the Early Jurassic a time when the central Atlantic Ocean and western Mediterranean Sea was forming from the older West Tethys Ocean. We can plot the taxa that are found nowadays in the northern Mediterranean near-coastal areas onto the northern coast of the West Tethys Ocean in the Early Jurassic geography (Fig. 11B). There are exceptions, like Ingolfiella macedonica and especially I. ischitana that occur in this same geographic area but appear higher in the cladogram (Fig. 10).
Fig. 11. Early parts of the phylogenetic tree projected on different epochs of geological history. A, three rectangles on the map of the Early Triassic represent roughly the distribution of the depicted taxa. B, The dark grey area on the map of the early Jurassic represents the area adjacent to the Recent mediterranean distribution of the earliest evolved species of the Ingolfiella clade (see Fig. 8). Notice this distribution is continuous and restrained to a limited area (maps modified after Scotese, 1997).
In a similar manner, we find that the grade on the cladogram that represents the freshwater and brackish species from the present day Atlantic Islands and Caribbean area are easily projected paleogeographically onto the northern coast of Late Jurassic Gondwana (Fig. 12A).
The extremely high sea levels of the late Cretaceous and further opening of the Atlantic Ocean (Haq et al., 1987) could have seen an evolution and dispersal of marine benthic and infaunal elements linked with the expansion of coastlines and deep waterways. From the Cretaceous onward the spread of ingolfiellids over the earth might have taken place at a faster pace (Fig. 12B) into the North Atlantic and out over the unfolding Indo-Pacific.
Fig. 12. A. The intermediate taxa of the Ingolfiella clade, the fresh and brackish water species of the Caribbean, the Mediterranean, and South America. The dark grey area represents the species distribution projected on a map of the Late Jurrassic with their nowadays occurrence. B. the most derived taxa of Ingolfiella containing marine species with a scattered distribution virtually worldwide, projected on a map of the Late Cretaceous (maps modified after Scotese, 1997).