BOOK REVIEWS 405 Arthropod Fossils and Phylogeny, Gregory Edgecombe (Editor), 1998, Columbia University Press, New York, 347 p. (Hardcover $125.00) ISBN: 0-231-09654-2. Copyright ? 1999, SEPM (Society (or Sedimentary Geology) There is something about writing on arthro|X)d pliylogcny that brings out the worst in people. J.W. Hedgepeth The above gem was reinlroduccd 17 years ago in an article survey- ing the crustacean fossil record (Schr?m, 1982). Although this obser- vation is just as true today as it was then, there has been improve- ment: modern versions of these tendentious and occasionally abrasive disagreements at least are more explicit. Edgecombe's edited volume, Arthropod Fossils and Phylogeny, extends this theme, but also fol- lows in the tradition of two important and scholarly precursors. Ar- thropod Phylogeny (Gupta, 1979) and Arthropod Relationships (For- tey and Tliomas, 1997). A new treatment is deemed necessary since Arthropod Relationships makes clear that major disagreements about how to analyze arthropods and the nature of their relationships is as contentious as ever. Arthropod Fossils and Phylogeny also gives us the st?rm und drang of arlhro|>od relationships. Even if the tenor of the debate has not improved, there have been advancements in the quality and ty|H;s of evidence and approaches that have been brought to bear on the subject. Consider the recent plethora of analytical techniques for estimating phytogenies both within tlje morphological and biomolccular research programs, the inclusion of fossil and extant data, addition of lar\'al and adult char- acters, the diversity of algorithms for establishing rclaledncss, and incessant discoveries of new biomolcculcs and morphologies for inter- taxal comparisons. Of course, if that is not enough, on the other side of the problem arc arthropods themselves. The Arthropoda is the most speciose macroscopic group of organisms of all time, possessing a bewildering range of morpholog>' that encompasses the more famil- iar trilobites, lobsters, tarantulas, and beetles. This phenomenal range in lx)dy plans applies not only to the fossil record, such as the bi- 0883-1351/99/0014-405/S3.00 406 BOOK REVIEWS zarre "funnies" of the Burgess Shale (as my former mentor Jack Sep- koski once referred to them), but also many extant arthropod repre- sentatives equally stretch our sensibilities of what a conventional ar- thropod should look like. Arthro[X)ds include organisms as different as barnacles, butterflies, fluke-like pentastomids (Abele el al., 1992), and some parasitoid insects that are perennially worm-like. Many phylogeneticists would include tardigrades and possibly velvet worms as well. Perhaps the problem with arthropod phylogeny is that there is a double dose of trouble: a multitude of techniques are available to exercise on an immensely multifarious phylum. Although this context may be unduly gloomy, there are a few bright lights. I have identified several recurring and inter-related themes that the 16 authors of the volume's seven chapters mostly can agree on. While some of these themes are not expressed or supported in ev- erj' chapter, they collectively transcend the background level of ar- gumentation because they emerge at the ends of these chapters as im- portant conclusions. The first seven themes involve substantive methodological issues, and the subsequent five are results concerning arthropod phylogeny. (1) Fossils, combined with modern taxa, are essential for phyloge- netic analyses. Data sets based exclusively on extant representatives, or alternatively only on fossils, provide inaccurate results for phylog- eny reconstruction (an exception, of course, are clades that are wholly extinct, but they too should be linked at coarser levels to extant line- ages). For example, as pointed out by Walossek and M?ller in Chap- ter 5, stem-group fossils may possess sjnapomorphies no longer ex- pressed in their extant descendants and, thus, are essential for estab- lishing early developmental features that are largely inaccessible to- day. Also, recent discoveries have increased the range of known taxa, preservational types, and sampled environments and have signifi- cantly narrowed the gap between the quality of modem and fossil morphological e\'idence. Data now are available for three-dimension- ally and sofl-part-preserved embryos, larval instars, and taxa as ob- scure as pentastomids. Sampling paleoenvnronmonts such as the ben- thic flocculant zone has markedly increased the range of morpholo- gies and taxa available for examination. It is in this context that Wheeler's presentation of "total evidence" in Chapter 1 is difficult to grasp since "total" in his analysis excludes crucial fossil material. (2) Arthropod data-sets are still significantly incomplete in terms of the total range of potentially available taxa and characters. For ex- ample, important unsequenced taxa could provide crucial data to- ward resoKing or expanding relationships that are loo poorly known morphologically. Wills and colleagues in Chapter 2 indicate that mo- lecular studies have sampled too few representatives across arthro- pods to yield repeatable phylogenies. Moreover, some of this incom- pleteness is conditioned by the lack of computational power inherent in the analysis of larger data sets. Schr?m and Hof, in Chapter 6, term this general deficiency of data as the "vraagteken effect," and note that the absence of critical information results in unstable phy- logenies. (3) Arthropods have significant homoplasy, confounding phyloge- netic analyses. Consistency indices of 0.5 and lower are typical for more taxonomically inclusive studies. Bergstr?m and Hou in Chapter 4?probably the most heterodox contribution of all?cite the recur- ring problem of a mosaic pattern of character distribution as being a critical issue that needs to be addressed before further progress can be made in arthropod phylogeny. Elevated arthropod homoplasy is frustrating for those ferreting out phylogenies and undoubtedly re- flects a developmental program that allows for major subelement transpositions within modular and segmental organisms. Thus, sub- elements can be often transposed, resulting in unique combinations of appendage type or body segments, yet retaining conservative similar- ities of the larger constituent clade. Such complications undoubtedly have spurred the application of the multitude of techniques men- tioned above. (4) The issue of what are the character-based limits of modem taxa is as relevant as ever. This issue is particularly important when it is applied to fossil arthropod taxa that incorporate some but not all of the characters found in their crown-group descendants, as discussed by Schr?m and Hof for the Crustacea and by Seiden and Dunlop for opilionid Arachnida in Chapter 7. One solution is a relaxation of the definition of a crown group, which tjTjically is based on modem taxa, to include reasonably those fossil stem lineages that arc phylogeneti- cally close, the collection of which are monophyletic at a coarser level. The alternative? which has resulted in much confusion regarding ar- thropod problem?tica addressed in this volume?is to propose an in- ordinate number of unfathomable stem-groups and plesions left in phylogenctic limbo. (5) Tliere remains confusion regarding the establishment of initial ground-plan characters of a basal taxon prior to phylogenetic analysis of its constituent lineages. Constructing a ground pattern (to use Wal- ossek and Muller's term) or a hypothetical ancestor (Ramsk?ld and Chen's similar term in Chapter 3) for determining the primitive spec- trum of plesiomorphic and .synapomorphic characters is an alterna- tive to selecting an appropriate fossil or credible outgroup that ideally would reveal the same plcsiomorphies and synapomorphies of inter- est. But relevant fossil material is frequently unavailable. The major contrast to using ground patterns, appropriate fossils, or hypothetical ancestors is to arrive at hypotheses from extant taxa, such as com- puter-based analyses of unordered molecular data. Such analyses, even if combined with morphological data, according to Walossek and M?ller, neglect the contingency that evolution has on morphology. Thus, a resulting trajectory of character-state change must be evalu- ated sequentially, from the bottom-up, and not post hoc at the termi- nal twigs of a highly-branched tree. Krzeminski (1992) has intro- duced an example of this approach in an analysis of the Mesozoic phy- logeny of dipteran insects. (6) The current state of the art involves multiple approaches, con- stant experimentation and introduction of new analytical techniques. Because of the proliferation of such approaches, alternatives, for ex- ample, no longer include just parsimony versus consensus approach- es, but more appropriately which consensus technique will be used: majority mle, strict consensus, or others? In fact, journals such as Cladistics and Systematic Biology now contain a high fraction of con- tributions devoted solely to the introduction and discussion of such techniques. (7) The translation of cladograms into classifications is probably unwarranted at this time. This is because of the ongoing problems mentioned above and even uncertainties about some basic arthropod homologies. Evidence for this, as mentioned above, are major and per- sistent difficulties in obtaining reasonably stable cladograms of the same taxa by different researchers. This volume presents six other developments that encapsulate ma- jor conclusions regarding internal arthropod relationships and the phyletic integrity of the entire clade. Wlien these issues are discussed, there is agreement among the contributors of this volume on the fol- lowing issues- (1) Arthropods are almost undoubtedly monophyletic. The term Euarthropoda is typically applied to this clade and includes the Tra- cheata (Myriapoda + Hexapoda) and Schizoramia. The Schizoramia encompasses the extinct Marellomorpha, the Arachnomorpha (in- cluding trilobitomorph and chelicerate clades), and the Crustacea. Tardigrades are probably arthropods as a sister-group to the Euar- thropoda, although lobopodan onychophorans, "great appendage" an- omalocaridids, and Opabinia with a median and frontal grasping claw are best considered as more distantly related, protarthropodan line- ages (but see Ballard et al., 1992). Manton's speculative scenario of profound arthropod polyphyly (summarized in her 1977 book) is in- sightful functional morphology but it has not withstood the test of many subsequent and varied phylogenetic analyses. (2) Enigmatic Cambrian taxa of arthropodan affinities are not as weird as once thought. These taxa, such as Burgessia, Marella, Na- raoia, Sidneyia and Waptia, form basal lineages comprising either (1) an extensive pectinate series that terminates in a crown group, the Crustacea (indicated by Schr?m and Hof and suggested by Walossek and M?ller and by Bergstr?m and Hou), or (2) a larger group of four principal clades within the Arachnomorpha, with each of the three larger clades characterized by a pectinate but highly pruned topology (indicated by Wills and colleagues). Although the position within the Arthropoda of these Cambrian taxa is significantly different in these two sets of studies, the successive acquisition of apomorphies by these Cambrian taxa is a key common feature. Their possession of at least some of the characters found in subsequently-derived crustaceans or arachnomorphs make them appear less "bizarre." Interestingly, a Pa- leozoic clade of high phylogenctic distinctiveness, the Marellomorpha, consists of Burgess Shale and Hunsr?ck Slate taxa that are probably BOOK REVIEWS 407 associated with basal arachnomorphs. Also, Ramsk?ld and Chen doc- ument a diverse, Early to Middle Cambrian lolwpodan clade consist- ing of at least 15 species, including the famed Hallucigenia, wth ap- preciable higher-rank cladogenotic structure. (3) The concept of the "Uniramia" survives. I once thought that the Uniramia was a dead idea. The polyramy hypothesis advocated bj' Kukalova-Peck (1992), which advocates that each body segment of primitive arthropods had poljTamous limbs and that biramy and uni- ramy are derived conditions, has been questioned directly by Walos- sek and M?ller and implicitly by some of the other authors. These au- thors posit that ancestral arthropods bore basipods with two rami, an inner endopod and an outer, often leaf-like, exopod, both modified for diverse functions in crustaceans and largely missing in cheUcerates and tracheates. Bergstr?m and Hou extend this issue, stating that Cambrian material supports the view that ancestral arthropods had only one unsegmented exopod per body segment, and that biramy and polyramy, as illustrated in exiles of Paleozoic insects, is derived. (4) The Cnustacea is monophyletic and the best-supported mono- phyletic clades of crustaceans are the Phyllopoda, Maxillopoda, and Malacostraca. Surprisingly, there was agreement among the four studies that analyzed ostensible "crustacean" forms that the Crusta- cea is a monophyletic clade, contrary to significant recent evidence. In the detailed analysis by Schr?m and Hof limited only to crustacean forms and a myriapod/hexapod outgroup, the Phyllopoda, two major clades of Maxillopoda, and Malacostraca (minus phyllocarids) were deemed monophyletic. These results are largely consistent with Wal- ossek and M?llers views. Interestingly, fonns from the Cambrian Burgess Shale and the Orsten Fauna constitute a separate clade be- tween croNvn-group Crustacea and a noneuarthropod out-group, rem- iniscent of the Marellomorpha of Wills and colleagues. The counter- intuitive result of a sister-group relationship between isopods and amphipods also was presented. (5) A position of the Hexapoda within the Crustacea is not sup- ported. Ironically, the one major arthropod subclade not analyzed in this volume is the TVacheata, also known as the Atclocerata or Anten- nata, and consisting of the Myriapoda (millipedes, centipedes and rel- atives) and the Hexapoda (insects and a few basal groups). However, tracheates were included as data in three of the chapters, in conjunc- tion with evaluations of major relationships within the Arthropoda. Several recent molecular-based analyses (Averof and Akam, 1995; Friedrich and Tautz, 1995) and evidence from developmental biology (Panganiban et al., 1995; Popadle et al., 1996) have placed the Hex- apoda within their fellow "mandibuiates", the Crustacea?an idea that was first suggested at the turn of the century based on external morphological considerations and subsequently abandoned. \Vliee- ler's total evidence approach (modem data only) and the comprehen- sive analysis by Wills and colleagues (modem and fossil data), how- ever, indicate that the Hexapoda is joined with the Myriapoda as the monophyletic Tracheata, which supports considerable evidence from traditional studies of morphology, (furiously, the two chapters that analyzed all major arthropod groups, including tracheates, reached different conclusions. Will's cladistic analyses lodged tracheates be- tween the more derived crustaceans, arachnomorphs, and marello- morphs and the more primitive lobopods, tardigrades, and anomalo- caridids. Alternatively Wheeler's cladistic analysis and a phenetic study by Wills and colleagues placed tracheates as a terminal clade whose closest relatives are crustaceans, and more distantly, chelicer- ates, supporting the clade Mandibulata. However, as other authors in this volume have shown, the integrity of the Mandibulata evaporates when fossils arc included. Edgecombe states on the first page that this s'olume was inspired by two topical issues. "One is an explosion of research on high-level phylogeny of the Arthropoda. The other is a rekindling of interest in the role fossils play in phylogenetic analysis." The contributions in this volume certainly address these two aims. However, this is a vol- ume that may rankle participants within the various ongoing arthro- pod debates. For those of us who are removed from current skirmish- es of methodology and lack interest in the increasingly powerful ar- mamentaria of cladistic algorithms, but are interested in basic ques- tions of how did the most successful macroscopic clade of all time achieve dominance on this planet, this volume is highly recommend- ed. If one can get though the unavoidable but often dreadful arthro- pod terminology?there frequently is tenninological incongruity from chapter to chapter?this volume presents generally cutting-edge summaries of considerable fascination in aggregate. This is a fast- moving field and most of the conclusions in this volume would have been unanticipated even five years ago. Who would have thought that many Burgess Shale taxa would have been accommodated within modem crustacean or arachnomorph subgroups? Or that lobopodan protarthropods were so disparate at higher taxonomic levels during the Cambrian, and that the recent Onychophora is only a surviving remnant? Or that the position of the tracheates within the Arthropo- da remains as unresolved as ever? With a bit of patience, this book will not bring out the worst in people. REFERENCES ABELE, L.G., SPE/VRS, T., KIM, W., and APPLEGATE, M., 1992, Phylog- eny of selected maxillopodan and other crustacean taxa based on 18S ribosomal nucleotide sequences: A preliminary analysis: Acta Zool?gica, v. 73, p. 373-382. AVEROF, M., and AK/\M, M., 1995, Insect-cnistacean relationships: In- sights from comparative developmental and molecular studies: Philosophical Transactions of the Royal Society of London, v. 347B, p. 293-303. BALL^VRD, J.W.O., OLSEN, G.J., FAITH, D.P., ODGEI?S, W.A., ROWELL, D.M., and ATKINSON, P.W., 1992, E\idence from 12S ribosomal RNA sequences that onychophorans are modified arthropods: Sci- ence, V. 258, p. 1345-1348. FORTEY, R.A., and THOMAS, R.H. (editors), 1997, Arthropod Relation- ships: Chapman & Hall, London, 383 p. FRIEDRICH, M., and TAUTZ, D., 1995, Ribosomal DNA phylogeny of the major extant arthropod classes and the evolution of myria- pods: Nature, v. 376, p. 165-167. GUPTA, A.P. (editor), 1979, Arthropod Phylogeny: Van NostrandRein- hold, Ijondon, 762 p. KRZEMINSKI, W., 1992, Triassic and lower Jurassic stage of Diptera evolution: Mitteilungen der Schweizerischen Entomologischen Gesellschaft, v. 65, p. 39-59. KUKALOVA-PECK, J., 1992, The "Uniramia" do not exist: The ground plan of the Pterj'gota as revealed by Permian Diaphanopteroidea from Russia (Insecta: Paleodictyopteroidea): Canadian Journal of Zoology, V. 70, p. 236-255. MANT?N, S.M., 1977, The Arthropoda: Habits, Functional Morpholo- gy and Evolution: Oxford, London, 527 p. PANGANIB/\N, G., SEBRING, A., NAGY, L. and CARROLL, S., 1995, The development of crustacean limbs and the evolution of arthropods: Science, v. 270, p. 1363-1366. PopADic, A., Ruscii, D., PETERSON, M., ROGERS, B.T., and KAUFMAN, T.C., 1996, Origin of the arthropod mandible: Nature, v. 380, p. 395. SCHR/\M, F., 1982, The fossil record and evolution of Cmstacea: in Bliss, D.E., ed.. The Biology of Crustacea, Volume 1 (Systematics, the Fossil Record and Biogeography, L.G. Abele, ed.): Academic press. New York, p. 93-147. CONRAD C. LABANDIERA Department of Paleobiology Smithsonian Institution National Museum of Natural History Washington, DC 20560-0121 and Department of Entomology University of Maryland College Park, MD 20742