Downloaded from geology.gsapubs.org on 18 January 2009 Geology Testing the marine and continental fossil records M. J. Benton and M. J. Simms Geology 1995;23;601-604 doi:10.1130/0091-7613(1995)023<0601:TTMACF>2.3.CO;2 Email alerting services click www.gsapubs.org/cgi/alerts to recieve free email alerts when new articles cite this article Subscribe click www.gsapubs.org/subscriptions/index.ac.dtl to subscribe to Geology Permission request click http://www.geosociety.org/pubs/copyrt.htm#gsa to contact GSA Copyright not claimed on content prepared wholly by U.S. government employees within scope of their employment. Individual scientists are hereby granted permission, without fees or further requests to GSA, to use a single figure, a single table, and/or a brief paragraph of text in subsequent works and to make unlimited copies of items in GSA's journals for noncommercial use in classrooms to further education and science. This file may not be posted to any Web site, but authors may post the abstracts only of their articles on their own or their organization's Web site providing the posting includes a reference to the article's full citation. GSA provides this and other forums for the presentation of diverse opinions and positions by scientists worldwide, regardless of their race, citizenship, gender, religion, or political viewpoint. Opinions presented in this publication do not reflect official positions of the Society. Notes © 1995 Geological Society of America Testing the marine and continental fossil records M. J. Benton Department of Geology, University of Bristol, Bristol BS8 1RJ, United Kingdom M. J. Simms Department of Geography and Geology, Cheltenham and Gloucester College of Higher Education, Francis Close Hall, Cheltenham GL50 4AZ, United Kingdom ABSTRACT phylogenies of echinoderms was tested for The fossil record of continental vertebrates is as good as that of echinoderms at the fossil-record quality (Table 1). Data for 72 family level, as shown by tests of the match of cladistic and stratigraphic data and of listed cladograms of vertebrates have been relative completeness. If echinoderms and vertebrates are typical of their environments, the published (Benton and Storrs, 1994, 1995). continental fossil record is not worse than the marine, despite the fact that, at a local level, For purposes of comparison, the nine clad- fossils are usually more abundant in marine sequences than in continental successions. The ograms of marine taxa (Actinopterygii, explanation of this paradox may be that vertebrates have attracted more intensive study Gnathostomata, Sarcopterygii, Sauroptery- than echinoderms, and thus the level of knowledge of their fossil record is some decades gia, Teleostei) were excluded, leaving 63 ahead of that of echinoderms. This finding validates the use of different kinds of fossil data trees of continental vertebrates. in broad-scale phylogenetic studies. If skeletonized marine shelf invertebrates have a markedly better fossil record than INTRODUCTION branching as indicated by phylogenetic data, continental animals, then the echinoderm It has often been asserted that the fossil and (2) the relative completeness of fossil test cases should show more examples of sta- record of marine shelf benthic organisms is records based on independent evidence for tistically significant rank order correlation better than that of continental organisms the size of gaps. than the vertebrates. The results of SRC (Benton, 1985; Flessa, 1990; Jablonski, The test of stratigraphic (age) and phylo- tests (Fig. 2, A and B) apparently show the 1991; Raup, 1979; Valentine, 1969). This as- genetic (clade) evidence about the origins of opposite. Fewer echinoderm cladograms sumption has been made by scaling up field groups (Norell, 1992, 1993; Norell and No- (38%) showed significant (P , 0.05) match- observations on Phanerozoic rocks. Typi- vacek, 1992a, 1992b) consists of comparing ing of clade order and age order than did the cally, limestones and clastic rocks laid down the rank order of nodes on a published clad- continental vertebrate cladograms (63%). on the shallow continental shelf yield abun- ogram (Fig. 1, A and B) with the rank order The same is true for highly significant cor- dant fossils of skeletonized invertebrates, of group appearances as documented in the relations (P , 0.01), found in 26% of cases such as brachiopods, molluscs, corals, ar- paleontological literature (Fig. 1C). The for echinoderms but in 41% of cases for con- thropods, bryozoans, and echinoderms match of clade and age data is assessed by tinental vertebrates. (Fu¨rsich, 1990; Kidwell, 1986). Continental the Spearman rank correlation (SRC) test. The low pass rate for echinoderm clad- sedimentary sequences generally yield less This approach depends upon the obser- ograms was surprising, especially in compar- abundant faunas of freshwater fishes and vation that there are three essentially inde- ison with cladograms of continental verte- molluscs, terrestrial insects, and vertebrates pendent methods of disentangling the se- brates. This result is almost certainly an (Behrensmeyer and Hill, 1980; Retallack, quence of events in the history of life: (1) the artifact of the small size of many of the echi- 1984). This differentiation may largely be an order of fossils in the rocks (stratigraphic noderm cladograms, 13 (21%) of which in- effect of the nature of the sediments: sedi- data); (2) cladograms, based generally on clude only 4 taxa. Such a small sample size mentation in river systems and lakes is assessments of the sequence of acquisition cannot yield critical values for the SRC co- highly episodic compared to the more con- of morphological characters; and (3) molec- efficient (Sprent, 1989). When the small tinuous deposition on marine shelves and ular phylogenies, founded on sequencing of (n 5 4) echinoderm cladograms are ex- particularly in abyssal areas of oceans nucleic acids or proteins, or on DNA-DNA cluded, the SRC scores match more closely (Sadler, 1981). hybridization. If it is accepted that these those discovered for continental vertebrates We test here the idea that groups of or- three approaches, stratigraphic, cladistic, (Fig. 2A), but still show a poorer pass rate ganisms known from a rich supply of fossils and molecular, are essentially independent, than do the cladograms of continental in the field necessarily have a fuller and bet- then mutual cross testing should be possible. vertebrates. ter-documented picture of large-scale phy- The comparison tests do not assume that The SRC test of matching between clade logeny than groups represented by sparse any one technique is better than another: and age order considers only one aspect of fossil materials. The test groups were echi- they merely compare the matching and as- the quality of the fossil record. Another cru- noderms (marine invertebrates) and tetra- sume that, if a large enough sample is used, cial feature is the relative completeness of pods (continental vertebrates). These the results will have statistical validity. particular examples, and this may be as- groups were selected for comparison be- Equivalent tests are available (Huelsenbeck, sessed by comparing the proportion of cause there is a sufficient number of clad- 1994) to test the quality of cladograms known fossil records to gaps, the relative ograms available for each and because both against stratigraphic data. completeness index (RCI) of Benton and groups consist of multielement taxa, the Comparative studies (Benton, 1994, 1995; Storrs (1994) (Fig. 1). skeletons of which may be preserved com- Benton and Storrs, 1994, 1995; Gauthier et If marine invertebrates have a better fos- pletely or may break up before burial. al., 1988; Norell, 1992, 1993; Norell and No- sil record than continental vertebrates, the vacek, 1992a, 1992b) have shown a good RCI values for the former should be mark- TESTING THE QUALITY OF THE match between cladistic branching order edly higher than those for the latter. This FOSSIL RECORD and the order of first fossil representatives does not appear to be the case. The RCI The quality of the long-term global-scale for 55%–73% of cladograms of vertebrates. values for echinoderms and vertebrates fossil record may be tested by comparison of A first attempt is made here to extend the range from 25% to 100%, but continental (1) the order of origin of groups from the tests to cladograms of nonvertebrates. vertebrates have more complete fossil stratigraphic record with the order of A sample of 63 cladograms and molecular records than do echinoderms (Fig. 2, C Geology; July 1995; v. 23; no. 7; p. 601–604; 3 figures; 1 table. 601 Figure 1. Methods for assessing quality of fossil record by comparing branching or- der in cladograms (A–C) with stratigraphic data and by comparing relative amount of gap and known record (C). Cladistic rank is determined by counting sequence of pri- mary nodes in cladogram (A). In many cases, published cladograms do not con- form to simple pectinate pattern in which all terminal taxa are simple side branches of single main stem. Frequently, there are more complex topologies in which some branches subdivide further (A), or some nodes may be partially unresolved, and give rise to more than one branch. In these cases, cladogram is reduced to pectinate form (B), and groups of taxa that meet main axis at same point are combined and treated as single unit. Stratigraphic se- quence of clade appearance is assessed from earliest known fossil representative of sister groups, and clade rank and strat- igraphic rank may then be compared (C). Minimum implied gap (MIG, diagonal rule) is difference between age of first repre- sentative of lineage and that of its sister, because oldest known fossils of sister groups are rarely of same age. MIG is min- imum estimate of stratigraphic gap, as true age of lineage divergence may lie well before oldest known fossil. Relative com- pleteness of fossil record may be as- sessed by comparing proportion of known range (standard range length, SRL) to ghost range, in form of relative complete- ness index (RCI), defined as: ((MIG) RCI 5 1 2 3 100%.