Late Pliocene faunal turnover in the Turkana Basin, Kenya and Ethiopia

Science

Volumn 278, Issue 5343, 1997, Pages 1589-1594

  Behrensmeyer, Anna K ^a   Todd, Nancy E ^a   Potts, Richard ^a   McBrinn, Geraldine E ^a  

^a DEPARTMENT OF PALEOBIOLOGY   (United States)

Author keywords

[No Author keywords available]

Indexed keywords

LAKE WATER;

FAUNAL TURNOVER; FOSSIL MAMMAL; MAMMAL; PLIOCENE; SPECIES TURNOVER;

ANALYTIC METHOD; AQUATIC FAUNA; ARTICLE; CLIMATE; ETHIOPIA; FOSSIL; PRIORITY JOURNAL; SAMPLING; TURNOVER TIME;

ETHIOPIA, TURKANA BASIN; KENYA, TURKANA BASIN;

EID: 0030659511     PISSN: 00368075     EISSN: None     Source Type: Journal    
DOI: 10.1126/science.278.5343.1589     Document Type: Article

References (60)

1. E. S. Vrba, in Ancestors: The Hard Evidence, E. Delson, Ed. (Liss, New York, 1985), pp. 63-71.
2. _, in (43), pp. 405-426.
3. _, J. Mammal. 73, 1 (1992).
4. _, in (44), pp. 385-424.
5. P. B. deMenocal, Science 270, 53 (1995); P. B. deMenocal and J. Bloemendal, in (44), pp. 262-288.
6. P. B. deMenocal, Science 270, 53 (1995); P. B. deMenocal and J. Bloemendal, in (44), pp. 262-288.
7. M. L. Prentice and G. H. Denton, in (43), pp. 383-403.
8. S. Stanley, Paleobiology, 18, 237 (1992).
9. E. S. Vrba, in (44), pp. 24-48.
10. T. D. White, in (44), pp. 369-384.
11. L. Bishop, thesis, Yale University, New Haven, CT (1994).
12. R. Bernor and M. Armour-Chelu, in African Biogeography, Climate Change and Early Hominid Evolution, T. Bromage and F. Schrenk, Eds. (Oxford Univ. Press, Oxford, UK, in press).
13. E. Delson, Courier Forschungsinst. Senkenb. 69, 199 (1984).
14. A. Turner, Acta Zool. Cracov. 38, 45 (1995).
15. _ and B. Wood, J. Hum. Evol. 24, 147 (1995).
16. C. S. Feibel, F. H. Brown, I. McDougall, Am. J. Phys. Anthropol. 78, 595 (1989).
17. F. H. Brown and C. S. Feibel, in (43), pp. 325-341; F. H. Brown, in (44), pp. 319-330.
18. F. H. Brown and C. S. Feibel, in (25), pp. 1-30.
19. We included available published information on identifications of mammals from Pliocene-Pleistocene localities of the northern Turkana Basin but excluded information on biostratigraphic ranges from other regions (we recognize that in some cases these ranges provide different FADs or LADs for Africa as a whole). Some taxonomic groups, such as the Elephantidae [N. Todd, thesis, George Washington University, Washington, DC (1997)], the Suidae (10), and the Equidae (11), are being revised, and ongoing work was incorporated where there is reference to northem Turkana Basin localities and specimens. The Evolution of Terrestrial Ecosystems (ETE) Program database is explained by J. Damuth [ETE Database Manual (ETE Consortium, Washington, DC, 1997)], and data relevant to this report are available at eteweb.lscf.ucsb.edu and www.sciencemag.org/ feature/data/973775.shl
20. The end points of each species range were established as follows: (i) a table of taxa was constructed with the use of all northern Turkana Basin locality records in the ETE database, (ii) bounding dates for the age of a locality were transferred to the table as minimum and maximum dates for each taxonomic record, and (iii) all the records for each taxon from multiple localities were sorted, and the oldest and youngest dates were used to establish the biostratigraphic range; thus, range lines in Fig. 3 are plotted between the maximum older date for the stratigraphic member or submember of first occurrence and the minimum younger date for the member or submember of last occurrence. These dates represent the oldest date at which a taxon could have appeared and the youngest date at which it could have disappeared based on currently available information. Nearly all species have continuous records through the members in each region between their FAD and LAD; that is, there are few breaks in the biostratigraphic records based on this data set. Species first recorded in the upper, middle, or lower parts of the same member are assigned the same FAD age based on the date for the lower boundary or, for species last recorded in a member, a LAD age based on the date for the upper boundary (unless there are intermediate dated horizons within the member). We did not interpolate dates for localities on the basis of stratigraphic distance below or above dated tuffs because of the variability of sediment accumulation rates in fluvial and lacustrine deposits [P. M. Sadler, J. Geol. 89, 569 (1981)] and because the deposits in the Turkana Basin have major disconformities representing significant periods of time that may not be apparent in any particular stratigraphic section (15).
21. Geological and paleomagnetic analysis shows that 120 m of lacustrine deposits below the KBS Tuff (dated at 1.88 ± 0.02 Ma) and above the unconformity are normally magnetized and thus younger than the base of the Olduvai Subchron (1.91 Ma). Feibel et al. (15) also correlate these lacustrine deposits to the Shungura Formation in the Omo region, where lacustrine sedimentation begins around 2.0 Ma. On the basis of both lines of evidence, they place the base of the lacustrine deposits at East Turkana at 2.0 Ma, which then requires a gap on the order of 450,000 years between the upper and lower Burgi members [this interpretation was reconfirmed by C. S. Feibel (personal communication)].
22. M. Beden, Les Faunes Plio-Pleistocenes de la Vallée de l'Omo (Éthiopie), vol. 2, Les Elephantides (Cahiers de Palaeontologie, Éditions du CNRS, Paris, 1987).
23. Y. Coppens and F. C. Howell, Les Faunes Plio-Pleistocènes de la Basse Vallée de l'Omo (Éthiopie), vol. 1, Perissodactyles, Artiodactyles (Bovidae) (Cahiers de Palaeontologie, Éditions du CNRS, Paris, 1985).
24. Y. Coppens and F. C. Howell, Les Faunes Plio-Pleistocènes de la Vallée de l'Omo (Éthiopie), vol. 3, Cercopithecidae de la Formation de Shungura (Cahiers de Palaeontologie, Éditions du CNRS, Paris, 1987).
25. J. M. Harris, Koobi Fora Research Project, vol. II, The Fossil Ungulates: Proboscidea, Perissodactyla, and Suidae (Oxford Univ. Press, New York, 1983).
26. J. M. Harris, Koobi Fora Research Project, vol. III, The Fossil Ungulates: Geology, Fossil Artiodactyls and Paleoenvironments (Oxford Univ. Press, New York, 1991).
27. J. M. Harris, F. H. Brown, M. G. Leakey, Contrib. Sci. 399, 1 (1988).
28. H. B. Wesselman, Contrib. Vertebr. Evol. 7, 1 (1984).
29. The maximizing approach (Table 2) gives reasonable benefit of doubt to all published taxonomic designations and thereby maximizes the sample of ranges contributing to the analysis of faunal turnover. The maximizing approach (i) treats each named taxon with a distinct range as a separate entity but combines "Genus species" with Genus cf. species (same names) if the range of one completely includes the range of the other; (ii) includes taxa designated by tribe or genus as species "A," "B," and so forth (for example, Kobus sp. D, West Turkana); (iii) excludes records referred to only as "indet. sp." under a family or tribe name (for example, "indet. sp. Cercopithecidae"); (iv) retains taxa designated only as Genus "sp." if this taxon provides unique range information, even when there are other named species of the same genus; and (v) includes taxa from any of the three regions. The minimizing approach represents our conservative assessment of the actual numbers of distinct taxa. The minimizing approach (i) combines named taxa that are likely to represent the same species (for example, one range for Connochaetes gentryi and Connochaetes cf. gentryi); (ii) eliminates all taxa referred to only as "sp. nov. A"; (iii) retains taxa identified only to genus (for example, Oryx sp.) if it is the only occurrence of that genus or if it provides unique range information; (iv) combines records for a genus under "Genus sp." if this genus provides a unique record of the range of that genus (for example, if the genus occurs in more than one region, but none of the species do); and (v) includes only taxa that occur in more than one of the three fossiliferous regions (to reduce the effects of regional sedimentation and taphonomy on FADs and LADs). (Similar conventions for the use of taxonomic information have been described by B. Van Valkenburgh and C. Janis [in Species Diversity in Ecological Communities, R. E, Ricklefs and D. Schluter, Eds. (Univ. of Chicago Press, Chicago, IL, 1993), pp. 330-340], A. Turner and B. Wood (74), and J. Alroy (37).} In spite of the differences between the maximizing and minimizing approaches, both approaches show essentially the same pattern between 3.3 and 1.5 Ma; correlation between the turnover percentages was significant at the 99% confidence interval [Pearson's correlation coefficient r = 0.958 with 9 degrees of freedom (df)].
30. The range charts do not include taxa that were represented at only one locality in the Turkana Basin (amounting to 80 out of the original 246 taxa in the raw data) because a sample size of one occurrence is too small to contribute reliable information on the beginning or end of a species' range. To test whether this omission biased the analysis against FADs caused by short-lived migrations or evolutionary events connected with a turnover pulse between 2.8 and 2.5 Ma, we analyzed turnover in the subset of 80 single-locality taxa. Although there is a peak of 70% in FADs between 2.4 and 2.2 Ma (primarily the result of rare taxa occurring in Omo Members E through G), there is an even distribution of 15 to 20% FADs from 2.4 Ma back to 3 Ma. This even distribution argues against a bias affecting turnover between 2.8 and 2.5 Ma.
31. P. W. Signor and J. H. Lipps, Geol. Soc. Am. Spec. Pap. 190, 291 (1982); C. R. Marshall, Paleobiology 23, 165 (1997).
32. P. W. Signor and J. H. Lipps, Geol. Soc. Am. Spec. Pap. 190, 291 (1982); C. R. Marshall, Paleobiology 23, 165 (1997).
33. J. Alroy, Palaeogeogr. Palaeoclimatol. Palaeoecol. 127, 285 (1996); personal communication.
34. A Turkana Basin locality typically consists of a geographically delimited concentration of fossils from one or more source horizons. Number of localities rather than number of specimens (NISP) is used as a measure of overall fossil abundance because NISP is subject to taphonomic biases that may have serious effects on relative abundance, such as inflation caused by counting multiple bones from the same skeleton [A. K. Behrensmeyer, in Taphonomy: Releasing the Data Locked in the Fossil Record, P. Allison and D. E. G. Briggs, Eds. (Plenum, New York, 1991), pp. 291-335].
35. E. S. Vrba, in Fossils in the Making, A. K. Behrensmeyer and A. P. Hill, Eds. (Univ. of Chicago Press, Chicago, IL, 1980), pp. 247-271; (1); P. Shipman and J. M. Harris, in (43), pp. 343-381 ; L. M. Spencer, J. Hum. Evol. 32, 201 (1997).
36. E. S. Vrba, in Fossils in the Making, A. K. Behrensmeyer and A. P. Hill, Eds. (Univ. of Chicago Press, Chicago, IL, 1980), pp. 247-271; (1); P. Shipman and J. M. Harris, in (43), pp. 343-381 ; L. M. Spencer, J. Hum. Evol. 32, 201 (1997).
37. E. S. Vrba, in Fossils in the Making, A. K. Behrensmeyer and A. P. Hill, Eds. (Univ. of Chicago Press, Chicago, IL, 1980), pp. 247-271; (1); P. Shipman and J. M. Harris, in (43), pp. 343-381 ; L. M. Spencer, J. Hum. Evol. 32, 201 (1997).
38. E. Delson, in Cambridge Encyclopedia of Human Evolution, J. S. Jones, R. D. Martin, D. Pilbeam, S. Bunney, Eds. (Cambridge Univ. Press, Cambridge, UK, 1992), pp. 220-221; in Colobine Monkeys: Their Ecology, Behavior and Evolution, A. G. Davies and J. F. Oates, Eds. (Cambridge Univ. Press, Cambridge, UK, 1985), pp. 11-43.
39. E. Delson, in Cambridge Encyclopedia of Human Evolution, J. S. Jones, R. D. Martin, D. Pilbeam, S. Bunney, Eds. (Cambridge Univ. Press, Cambridge, UK, 1992), pp. 220-221; in Colobine Monkeys: Their Ecology, Behavior and Evolution, A. G. Davies and J. F. Oates, Eds. (Cambridge Univ. Press, Cambridge, UK, 1985), pp. 11-43.
40. M. G. Leakey, in Theropithecus: The Rise and Fall of a Primate Genus, N. Jablonski, Ed. (Cambridge Univ. Press, Cambridge, UK, 1993), pp. 85-123.
41. Taxonomic designations follow B. A. Wood [Koobi Fora Research Project, vol. IV, Hominid Cranial Remains (Clarendon, Oxford, UK, 1991); G. Suwa, T. D. White, and F. C. Howell [Am. J. Phys. Anthropol. 101, 247 (1996)]; and B. Wood et al. (45). Although A. (P.) aethiopicus and H. rudolfensis are represented by unusually complete crania, almost all other taxonomically identified specimens in the 3-to 2-Ma interval are fragmentary maxillae or mandibulae or isolated teeth. On the basis of the latter specimens, these two taxa cannot easily be distinguished from A. boisei and H. habilis, respectively. In the range charts used for this study, H. rudolfensis was thus placed in H. cf. habilis, and, following Wood et al. (45), megadont hominid specimens older than 2.3 Ma, most of which could represent A. (P.) aethiopicus, were placed in A. (P.) cf. boisei. In Table 2, A. aethiopicus would represent an additional LAD within the Hominidae at about 2.3 Ma, when it is replaced by A. boisei, if it were not included in the taxon "A. (P.) cf. boiser" (maximizing) or in "A(P.) boiser" (minimizing) to facilitate analysis of sampling biases.
42. Taxonomic designations follow B. A. Wood [Koobi Fora Research Project, vol. IV, Hominid Cranial Remains (Clarendon, Oxford, UK, 1991); G. Suwa, T. D. White, and F. C. Howell [Am. J. Phys. Anthropol. 101, 247 (1996)]; and B. Wood et al. (45). Although A. (P.) aethiopicus and H. rudolfensis are represented by unusually complete crania, almost all other taxonomically identified specimens in the 3-to 2-Ma interval are fragmentary maxillae or mandibulae or isolated teeth. On the basis of the latter specimens, these two taxa cannot easily be distinguished from A. boisei and H. habilis, respectively. In the range charts used for this study, H. rudolfensis was thus placed in H. cf. habilis, and, following Wood et al. (45), megadont hominid specimens older than 2.3 Ma, most of which could represent A. (P.) aethiopicus, were placed in A. (P.) cf. boisei. In Table 2, A. aethiopicus would represent an additional LAD within the Hominidae at about 2.3 Ma, when it is replaced by A. boisei, if it were not included in the taxon "A. (P.) cf. boiser" (maximizing) or in "A(P.) boiser" (minimizing) to facilitate analysis of sampling biases.
43. Taxonomic designations follow B. A. Wood [Koobi Fora Research Project, vol. IV, Hominid Cranial Remains (Clarendon, Oxford, UK, 1991); G. Suwa, T. D. White, and F. C. Howell [Am. J. Phys. Anthropol. 101, 247 (1996)]; and B. Wood et al. (45). Although A. (P.) aethiopicus and H. rudolfensis are represented by unusually complete crania, almost all other taxonomically identified specimens in the 3-to 2-Ma interval are fragmentary maxillae or mandibulae or isolated teeth. On the basis of the latter specimens, these two taxa cannot easily be distinguished from A. boisei and H. habilis, respectively. In the range charts used for this study, H. rudolfensis was thus placed in H. cf. habilis, and, following Wood et al. (45), megadont hominid specimens older than 2.3 Ma, most of which could represent A. (P.) aethiopicus, were placed in A. (P.) cf. boisei. In Table 2, A. aethiopicus would represent an additional LAD within the Hominidae at about 2.3 Ma, when it is replaced by A. boisei, if it were not included in the taxon "A. (P.) cf. boiser" (maximizing) or in "A(P.) boiser" (minimizing) to facilitate analysis of sampling biases.
44. R. Tiedemann, M. Samthein, N. J. Shackleton, Paleoceanography 9, 619 (1994).
45. R. Bonnefille and A. Vincens, in Recherches Francaises sur le Quaternaire Hors de France (Supplementary Bulletin 50/1, Association Française pour l'Étude du Quaternaire, Paris, 1977); R. Bonnefille and R. Dechamps, in The Omo Group: Archives of the International Omo Research Expedition, J. de Heinzelin, Ed. (Musée Royal de l'Afrique Central, Tervuren, Belgium, 1983), pp. 191-207; R. Bonnefille, in (44), pp. 299-310.
46. R. Bonnefille and A. Vincens, in Recherches Francaises sur le Quaternaire Hors de France (Supplementary Bulletin 50/1, Association Française pour l'Étude du Quaternaire, Paris, 1977); R. Bonnefille and R. Dechamps, in The Omo Group: Archives of the International Omo Research Expedition, J. de Heinzelin, Ed. (Musée Royal de l'Afrique Central, Tervuren, Belgium, 1983), pp. 191-207; R. Bonnefille, in (44), pp. 299-310.
47. R. Bonnefille and A. Vincens, in Recherches Francaises sur le Quaternaire Hors de France (Supplementary Bulletin 50/1, Association Française pour l'Étude du Quaternaire, Paris, 1977); R. Bonnefille and R. Dechamps, in The Omo Group: Archives of the International Omo Research Expedition, J. de Heinzelin, Ed. (Musée Royal de l'Afrique Central, Tervuren, Belgium, 1983), pp. 191-207; R. Bonnefille, in (44), pp. 299-310.
48. T. E. Cerling, J. R. Bowman, J. R. O'Neil, Palaeogeogr Palaeoclimatol. Palaeoecol. 63, 335 (1988); T. E. Cerling, Global Planet. Change 5, 241 (1992).
49. T. E. Cerling, J. R. Bowman, J. R. O'Neil, Palaeogeogr Palaeoclimatol. Palaeoecol. 63, 335 (1988); T. E. Cerling, Global Planet. Change 5, 241 (1992).
50. R. Potts, Humanity's Descent: The Consequences of Ecological Instability (Morrow, New York, 1996), pp. 114-124.
51. C. S. Feibel, J. M. Harris, F. H. Brown, in (25), pp. 321-346.
52. R. Bobé, J. Vertebr. Paleontol. 16, 23A (1996); thesis, University of Washington, Seattle (1997).
53. F. E. Grine, Ed., The Evolutionary History of the "Robust" Australopithecines (Aldine, New York, 1988).
54. E. S. Vrba, G. H. Denton, T. C. Partridge, L. H. Burckle, Eds., Paleoclimate and Evolution, with Emphasis on Human Origins (Yale Univ. Press, New Haven, CT, 1995).
55. B. Wood, C. Wood, L. Konigsberg, Am. J. Phys. Anthropol. 95, 117 (1994).
56. In plotting the pattern of FADs and LADs through time, we used the midpoint of each member (Fig. 2) and the midpoint of each 200,000-year interval (Fig. 4) as the most objective representation of summed data that are scattered throughout the time represented by these intervals. This convention compensates in part for the clustering of FADs and LADs at the beginnings and ends of members, which results from the use of member boundaries as estimates on species ranges. The total number of species for each member is based on counts of the continuing species plus any that appear or disappear in that member; species with a FAD and a LAD in the same member are counted once. The same method applies to the 200,000-year intervals in Fig. 4. Total turnover was calculated with the cumulative records of FADs and LADs compared with the total number of species for the entire time interval under consideration (3.0 to 1.8 Ma or 3.0 to 2.0 Ma) (Table 2).
57. The FAD for Equus sp. is based on a lower first or second molar from Area 102, East Turkana, which is reported to be from "Zone C," between the Hasuma Tuff (2.85 ± 0.08 Ma) and the Lokalalei Tuff (2.52 ± 0.05 Ma) [V. Eisenmann, in (24), pp. 156-214]. This early date for Equus disagrees with other evidence that places the FAD at about 2.3 Ma [Member G, Omo sequence (22)].
58. The FAD for Equus sp. is based on a lower first or second molar from Area 102, East Turkana, which is reported to be from "Zone C," between the Hasuma Tuff (2.85 ± 0.08 Ma) and the Lokalalei Tuff (2.52 ± 0.05 Ma) [V. Eisenmann, in (24), pp. 156-214]. This early date for Equus disagrees with other evidence that places the FAD at about 2.3 Ma [Member G, Omo sequence (22)].
59. Member boundaries in each area do not fall at 200,000-year intervals or at the same absolute times, and to construct the running fossil abundance totals on the basis of numbers of localities, we used the proportion of total time spanned by a member for each 200,000-year interval that it crossed. Thus, a member with 30 localities that covers 1/3 of one 200,000-year interval and 2/3 of the adjacent interval contributes 10 localities to the first and 20 to the second of these intervals. The total for each interval is then the sum of localities contributed by each of the three regions.
60. We thank P. San Miguel for assistance with the data input; A. Cutler and R. Chapman for help with the statistical tests; J. Alroy and C. Feibel for advice and access to unpublished data; J. Harris, B. Cooke, E. Weston, and L. Bishop for taxonomic information on the Turkana Basin faunas; R. Bernor, R. Bobe, A. Gentry, E. Vrba, B. Wood, S. Suter, and several anonymous reviewers for providing comments on the manuscript; the governments of Kenya and Ethiopia; research project leaders Y. Coppens and F. C. Howell (Omo) and R. E. Leakey and M. G. Leakey (East and West Turkana); and the many skilled and dedicated individuals who contributed to the documentation, collection, curation, and publication of the fossil specimens on which this study is based. We gratefully acknowledge the National Museum of Natural History and the Smithsonian Institution for support of the Evolution of Terrestrial Ecosystems Program (ETE). This is publication 59 of the ETE Program. Information on the ETE Database, including data relevant to this report, is available at eteweb. lscf.ucsb.edu and www.sciencemag.org/feature/ data/973775.shl