One of the most important aspects in biological sciences is the investigation of long-term changes in nature. It is commonly accepted that the universe, the planetary system, and all the organisms living on earth, are the result of a process of continuous modification and development. The architecture of recent organisms provides a key for the understanding and reconstruction of their history through the billions of years, of the process called evolution. Evolution is the gradual change of organisms through generations and time. We cannot directely observe evolution in nature. So we have to reconstruct how organisms could have changed and how their variety emerged under anatomical and structural-functional constraints. For any evo-lutionary transformation it is considered that the process has to be continuous. Organisms are not engines, which can be stopped for rebuilding. Organismic evolution can be likened to a gradual change in the running engine.
Evolutionary research should not only consider obvious similiarities. The architectural elements by which an organism is designed are much more important. In particular, this means the mechanical properties of tissues, their arrangement and their functional connections. (In German this complex of structure and function is called Konstruktion or Bauplan; the latter term is sometimes used as a German loan-word in English scientific papers and text books).

This approach provides new visions of biological research. Common knowledge of biological investigations can be integrated, but in addition, this approach provides answers to the question: is a gradual change of one bauplan into another possible? For example: can a dinosaur evolve into a bird? This question can be answered with a yes. We can show that the dinosaur bauplan supports the possibility of this evolutionary transformation. Another question: Can a dinosaur evolve into a mammal? This has to be answered with a no. We can show that the bauplan does not support this kind of transformation.
This way of reasoning, its philosophical background, and the results on the reconstruction of evolutionary pathways, were developed by a group of scientists in the Senckenberg-Museum, Frankfurt, in cooperation with several colleagues. The new approach has been called "The Frankfurt Evolutionary Theory" and "Engineering Morphology" (in German: Konstruktions-Morphologie).

The results are summarized in a graphic design. It is published as a poster, and presented in the museum as a wall-hanging in room 206 and in a modified form in room 104 of the Senckenberg museum. The top part of the poster shows the formation of the earth and the development of early organisms. The most ancestral animals are located in the center. From here the evolutionary pathways originate and branch off. Not all branchings of the animal kingdom can be presented in the design, we decided to include those leading to well-known animals or to some of specific zoological interest. On these branches you find the most important steps in the evolutionary transformation of organisms represented by technical drawings. At the end of each pathway a most recent or a known fossil representative is shown in a naturalistic form. These stand in contrast to the model-like, hypothetical and technical drawings. All extant animals are equally remote of their point of origin. Each animal by itself has attained its own stage of sophistication in the evolutionary process. We no longer share the traditional anthropocentric view of the world, in which man has taken the position of the top of a tree growing from so-called lower to so-called higher animals.

There seems to be a great resurgence of interest in metazoan phylogeny at the present time, and one trigger for this has been the Burgess Shale and Chengjiang discoveries. Where do these strange creatures fit in? We have also the present debate about whether the Cambrian explosion of life is real or an artefact, and the apparent conflict between the molecular and palaeontological data. So, there is a very real interest at the present time in how the various phyla diverged, as well as when. The new Senckenberg poster is a very original contribution to the present discussion, but before making comments thereupon, let me consider something of the history of ideas about metazoan phylogeny. When I was an undergraduate, over forty years ago, we readily learned that the invertebrate phyla were very much separate from one another, each with a common plan of organisation, but that the relationships between them remained much more obscure. Standard textbooks, such as G.S. Carter's General Zoology of the Invertebrates lucidly expounded the model most in favour at that time, and an elegant and straightforward concept it was. From a protozoan ancestor came the sponges, a side branch which led only to more sponges. Diploblastic animals represented a higher grade of organisation and probably gave rise to triploblasts, represented in their simplest form by acoelomate platyhelminthes. From these came coelomate triploblasts, which could be divided into two main groupings. In the `Echinoderm Superphylum' (properly deuterostomes) which includes echinoderms, hemichordates and chordates, the larva is a pluteus, cleavage is radial, development equipotential, and the coelom enterocoelic, i.e. arising from pouches in the enteron. In the `Annelid superphylum' which includes most other groups, shared characters include a trochophore larva, spiral cleavage, mosiac development, and a schizocoelic coelom, arising from splits in the mesoderm. In concluding his book Carter comments "We can say nothing of the differences between the adults of the ancestral groups of the two superphyla...all these early metazoans were simple, soft-bodied, triploblastic, bilaterally symmetrical, unsegmented, bottom-living animals with slight cephalisation, a ventral mouth, an anus, and probably tubular excretory organs. Beyond these statements it is not possible to go". That was written a long time ago. Where do we stand now? The problem was tackled afresh by Pat Willmer (1990) in her Invertebrate Relationships. Her elegant analysis leaves us with a picture fairly close to that of the traditional model, though refined and extended. Willmer believes that convergence is so common that the use of cladograms may distort or conceal original relationships, and that molecular biology may hold the ultimate key. Needless to say the cladists do not like this at all. More recent views, such as that espoused by Simon Conway Morris, again retain many of the older concepts. Thus sponges are an early side branch (now known from the Ediacara fauna), and the diploblasts, a separate two-layered clade. While the deuterostomes (the old Echinoderm Superphylum) retain their integrity, what Carter regarded as the `Annelid Superphylum' now separates the Ecdysozoans (arthropods, priapulids, nematodes, lobopodians, and anomalocarids) from the Iophotrochozoans (nemerteans, platyhelminthes, molluscs, haleriids, annelids, and brachiopods). The platyhelminthes no longer seems to have a key position. One can readily draw cladograms for the animals within these groupings, but the relationships of these major packages to one another still remain obscure. We seem, therefore, to have advanced, and no doubt increasing molecular data may help us in understanding more about invertebrate relationships, though the problem of convergence returns to haunt us. Metazoans have skeletons, whether rigid or hydrostatic, and as is clear from the arguments developed by Thomas and Reif in their `Skeleton Space' papers, there are only a limited number of functional skeletal types. And Simon Conway Morris, in a recent paper, comments "The question we need to ask is whether a structure (molecular or organismal) is similar because it shares a common ancestry, or because there is no (or very few) alternative ... how do we balance the process of change against the emergence of form?" Can we escape from this conundrum? Now I am sorry about this long preamble, but it is all relevant to this present review. For over the last years a rather different approach has been developing in Frankfurt-am-Main, that of Evolution and Engineering Morphology. The basic concept here is that the great variety of organisms that arose and changed during evolution could only do so under specific structural and functional constraints, best understood in terms of engineering analogues. This approach, though in some ways paralleling more traditional concepts, offers unique perspectives. While some of these will certainly prove controversial, they are surely worthy of attention. The poster illustrating these is large, 118cm long, 84cm wide, and attractively presented, with the complete organisms themselves represented in pale yellow, but where sections through the animals are shown, they are picked out in pale blue with the internal cavities in black. At the top are serial snapshots of Planet Earth, from its origin in colliding meteorites and asteroids to the young blue planet. Corresponding illustrations are shown of steps leading to early organisms, biofilms, protocells, archaea and bacteria, protoeukaryotes (motiloids as they are called here), and to representatives of the four eukaryote kingdoms. So far, there is nothing very controversial. The rest of the display is given to the evolution of multicellular animals. So what were the most primitive of all animals? According to the Frankfurt school, these were gallertoids, animals with a gelatinous body of many cells and cell-complexes, stablised and fused with connective tissue. Only with such an organisation could muscle contractions on one side be transmitted to the other, where muscles would extend automatically. Some living animals, Trichoplax, and possibly flagellates, also the fossil Vendozoa, may be close to the ancestral gallertoid. All the evolutionary pathways in the animal kingdom are regarded as modifications of the ancestral gallertoid, and each successive development acts as a springboard to the next. So, a schematic gallertoid is represented in the centre of the poster, and subsequent developments are presented as radially diverging from it. While the arrangement reflects in a broad and general sense the major groupings to which we have already referred (diploblasts on the right, deuterostomes on the left), it is what lies in between that fits less well with traditional categories. But let us remember that it is engineering principles that are being considered here, not direct phylogeny. Let us then consider some of these evolutionary pathways. Gallertoids of compact body shape, and with cilia, could become sessile or mobile, leading to the swimming ctenophores on the one hand, and to sponges, stromatoporoids and corals on the other. Elongated gallertoids, on the other hand, replaced ciliary propulsion with lateral bending. Internally widened cavities would result in a system of flexible membranes with a fluid filling. From the engineering point of view such a coelomate organisation is highly efficient and could lead on to new developments, only possible on the basis of what already exists. A segmented coelomate hydroskeleton such as this would lead to further functional types, and the bulk of invertebrates (diploblasts and deuterostomes excepted) are seen here as derived from such an ancestor. Arthropods, for example, arose from such an ancestral type when they developed appendages, platyhelminthes by flattening the body, reduction of the gut and loss of coelom, molluscs by development of a creeping foot and radula, compression of the coelom, and ultimately loss of metameric segmentation. Deuterostomes, according to this model, evolved along a different line, from a segmented coelomate organism, but with a rostral filter-feeding mechanism, subsequent development of a notochord, and with subsequent key innovations leading to the echinoderms and chordates known from the Ordovician onwards. Before commenting further, let us hear what the authors themselves have to say. "The evolution of animals is presented here is a new way, which is not like that of the standard textbooks. The engineering-like explanation of evolution and the reconstruction of evolutionary pathways contrasts sharply with morphometric and genetic phylogenetic trees. However, explanations are provided, which are missing in most traditional presentations.
The results which are the basis of this poster have been developed since 1970 by the group "Kritische Evolutionstheorie". Continuous discussions with philosophers and testing improved the coherence of the methodology". Now not everybody is going to like this approach. If you are looking for cladograms you will not find them here. The Ecdysozoans are split across several lines of descent, and diploblasty and triploblasty are not directly referred to. But that is not the point. This is something different. It is the product of a very able group, working in some degree of isolation, and focusing on issues which have not otherwise received much attention. This novel way of looking at the relationships of metazoans, whether or not we agree with the actual pathways showing how animals diverged, merits careful consideration. Animals, living and fossil, are not just sources of data for plotting on cladograms, at least I do not regard them so. Undoubtedly the Senckenberg Poster will fuel an interesting discussion. But let it be the kind of debate which engenders more light than heat.

Euan Clarkson
University of Edinburgh