Strömberg. 10.1073/pnas.0505700102. |
Table 1. Sample and locality information for Cenozoic sediment samples. An asterisk next to the sample number indicates that the sample was not included in the quantitative analysis. "Matrix off specimen" in the "Location in section" column indicates samples collected from museum specimens. AMNH, American Museum of Natural History; CUM, University of Colorado Museum, Boulder; DMNS, Denver Museum of Nature and Science; UCMP, University of California Museum of Paleontology; UM, University of Montana Museum of Paleontology, Missoula; UNSM, University of Nebraska State Museum, Lincoln; LWA, Lower Whitney Ash; UWA, Upper Whitney Ash; NALMA, North American Land Mammal Age.
a
Evanoff, E., Graham, R. W. & Tedford, R. H. (1999) Society of Vertebrate Paleontology Guidebook (Denver), pp. 1-52.b
R. Graham, personal communication.c
E. Evanoff, personal communication.d
University of California Museum of Paleontology collections.e
Tedford, R. H. (1999) in The Tertiary Record of Weld County, Northeastern Colorado, eds. Evanoff, E., Graham, R. W. & Tedford, R. H. (Denver), pp. 31-47.f
American Museum of Natural History collections.g
D. Hanneman, personal communication.h
University of Montana Museum of Paleontology, Missoula, collections.i
MIOMAP database (www.ucmp.berkeley.edu/miomap).j
H. LaGarry, personal communication.k
LaGarry, H. E. & LaGarry, L. A. (1999) Geol. Soc. Am. Abstracts with Programs 31, A-293.l
LaGarry, H. E. & LaGarry, L. A. (2000) Symposium: Nebraska Stratigraphy. Program and Proceedings of the 10th Annual Meeting of the Nebraska Academy of Sciences (Nebraska Academy of Sciences, Lincoln, NE), p. 67.1. Janis, C. M., Scott, K. M. & Jacobs, L. L. (1998) Evolution of Tertiary Mammals in North America: Terrestrial Carnivores, Ungulates and Ungulatelike Mammals (Cambridge Univ. Press, Cambridge, U.K.), Vol. 1.
2. Galbreath, E. C. (1953) Univ. Kan. Paleontol. Contrib. 4, 1-120.
3. Kron, D. G. (1988) Ph.D. thesis (Univ. of Colorado, Boulder), p. 377.
4. Tedford, R. H., Skinner, M. S., Fields, R. S., Rensberger, J. M., Whistler, D. P., Galusha, T., Taylor, B. E., Macdonald, J. R. & Webb, S. D. (1987) in Cenozoic Mammals of North America, ed. Woodburne, M. O. (Univ. of California Press, Berkeley), pp. 153-210.
5. Fields, R. W., Tabrum, A. R., Rasmussen, D. L. & Nichols, R. (1985) in Cenozoic Paleogeography of West Central United States: Rocky Mountain Paleogeography Symposium No. 3, Rocky Mountain Section, eds. Flores, R. M. & Kaplan, S. S. (SEPM, Denver), pp. 9-36.
6. Tabrum, A. R., Prothero, D. R. & Garcia, D. (1996) in The Terrestrial Eocene-Oligocene Transition in North America, eds. Prothero, D. R. & Emry, R. J. (Cambridge Univ. Press, New York), pp. 278-311.
7. Nichols, R., Tabrum, A. R., Barnosky, A. D. & Hill, C. L. (2001) in Society of Vertebrate Paleontology 61st Annual Meeting: Mesozoic and Cenozoic Paleontology in the Western Plains and Rocky Mountains - Guidebook for the Field Trips, ed. Hill, C. L. (Museum of the Rockies, Bozeman, MT), Vol. 3, pp. 77-144.
8. Kuenzi, D. W. & Fields, R. W. (1971) Geol. Soc. Am. Bull. 82, 3373-3394.
9. Tedford, R. H., Albright, L. B. I., Barnosky, A. D., Ferrusquia, I. V., Hunt, R. M., Jr., Storer, J., Swisher, C. C. I., Webb, S. D. & Whistler, D. P. (2004) in Late Cretaceous and Cenozoic Mammals of North America: Geochronology and Biostratigraphy, ed. Woodburne, M. O. (Columbia Univ. Press, New York), pp. 169-231.
10. Rasmussen, D. L. (1989) in Montana Geological Society 1989 Field Conference Guidebook: Montana Centennial Edition, Geologic Resources of Montana, eds. French, D. E. & Grabb, R. F. (Montana Geological Soc., Bozeman, MT), Vol. 1, pp. 205-215.
11. Rasmussen, D. L. (2004) in Cenozoic Systems of the Rocky Mountain Region, eds. Raynolds, R. G. & Flores, R. M. (RMS-SEPM, Denver), pp. 459-477.
12. Terry, D. O., Jr., & LaGarry, H. E. (1998) in Depositional Environments, Lithostratigraphy, and Biostratigraphy of the White River and Arikaree Groups (Late Eocene to Early Miocene, North America), eds. Terry, D. O., LaGarry, H. E. & Hunt, R. M., Jr. (Geological Soc. of America), Special Paper 325, pp. 117-142.
13. Terry, D. O., Jr. (1998) in Depositional Environments, Lithostratigraphy, and Biostratigraphy of the White River and Arikaree Groups (Late Eocene to Early Miocene, North America), eds. Terry, D. O., LaGarry, H. E. & Hunt, R. M., Jr. (Geological Soc. of America), Special Paper 325, pp. 15-37.
14. Schultz, C. B. & Stout, T. M. (1955) Bull. Univ. Neb. State Mus. 4, 17-52.
15. LaGarry, H. E. (1998) in Depositional Environments, Lithostratigraphy, and Biostratigraphy of the White River and Arikaree Groups (Late Eocene to Early Miocene, North America), eds. Terry, D. O., LaGarry, H. E. & Hunt, R. M., Jr. (Geological Soc. of America), Special Paper 325, pp. 63-92.
16. Tedford, R. H., Swinehart, J., Hunt, R. M. & Voorhies, M. R. (1985) Dakoterra 2, 334-352.
17. Swisher, C. C., III, & Prothero, D. R. (1990) Science 249, 760-762.
18. Schultz, C. B. & Falkenbach, C. H. (1968) Bull. Am. Mus. Nat. Hist. 139, 1-498.
19. Hatcher, J. B. (1902) Proc. Am. Philos. Soc. 41, 113-131.
20. MacFadden, B. J. & Hunt, R. M., Jr. (1998) in Depositional Environments, Lithostratigraphy, and Biostratigraphy of the White River and Arikaree Groups (Late Eocene to Early Miocene, North America), eds. Terry, D. O., LaGarry, H. E. & Hunt, R. M., Jr. (Geological Soc. of America), Special Paper 325, pp. 143-165.
21. Bailey, B. E. (2004) Paludicola 4, 81-113.
22. Yatkola, D. A. (1978) Neb. Geol. Survey Paper 19, 1-66.
23. Bailey, B. E. (2001) M.Sc. thesis (Univ. of Nebraska, Lincoln), pp. 92.
24. Hunt, R. M., Jr. (1990) Geol. Soc. Am. Special Paper 244, 69-111.
25. Hunt, R. M. (2002) Am. Mus. Novitates, 1-41.
26. Hunt, R. M., Jr. (1978) Palaeogeogr. Palaeoclimatol. Palaeoecol. 24, 1-52.
27. Coombs, M. C. & Coombs, W. P. (1997) Palaios 12, 165-187.
28. Peterson, O. A. (1909) Memoirs Carnegie Mus. 4, 41-158.
29. Cook, H. J. (1965) Am. Mus. Novitates 2227, 1-8.
30. Galusha, T. (1975) Bull. Am. Mus. Nat. Hist. 156, 1-68.
31. Skinner, M. F. & Johnson, F. W. (1984) Bull. Am. Mus. Nat. Hist. 178, 215-368.
32. Skinner, M. F., Skinner, S. M. & Gooris, R. J. (1977) Bull. Am. Mus. Nat. Hist. 158, 265-371.
33. Rich, T. H. V. (1981) Bull. Am. Mus. Nat. Hist. 171, 1-116.
Table 2. Cenozoic phytolith assemblages from Colorado, Montana/Idaho, and Nebraska/Wyoming. For explanation of morphotype classes, see text; for explanation of statistical procedures, see "Analysis of GSSC Assemblages" in Supporting Methods. GSSC morphotypes: Pooid, diagnostic, GSSC morphotypes that are generally diagnostic of pooid grasses; Pooid, nondiagnostic, GSSC morphotypes that are commonly abundantly produced by pooid grasses but also by other grasses. POO-1, crenate (pooids; Fig. 2E); POO-2, decorated Stipa-type bilobates (stipoid pooids; Fig. 2D); PAC-1, inverted bilobate (PACCAD grasses; Fig. 2B); PAC-2, true saddles or assemblage of saddle-like GSSC morphotypes (chloridoids; Fig. 2C). Parentheses around GSSC morphotypes signify some uncertainty in the identification. Vegetation inference, including interpretation of grass community composition using information from statistical comparisons with the modern reference collection (in Tables 4 and 5) and the presence of specific GSSC morphotypes: CF, closed forest; w CG, understory of closed-habitat grasses; (OG), low frequencies of open-habitat grasses present; OH, open, grass-dominated habitat; (CG), low frequencies of closed-habitat grasses present). Assemblages not included in quantitative analysis (marked with an asterisk): n.o., not observed; p, present; ab, abundant; vab, very abundant.
Table 3. Hypotheses that are tested regarding GSSC assemblage composition in Cenozoic samples from Colorado, Montana/Idaho, and Nebraska/Wyoming. The various urns, or background universes, against which the fossil assemblages are statistically compared, are based on the ratios of relevant GSSC compound variables in a reference collection of phytoliths from modern grasses (MPR) (unpublished work). The maximum MPR ratio is used in each case, both for leaf and reproductive material (which tend to differ). CH TOT, closed-habitat GSSC morphotypes; OH TOT, Pooid, diagnostic + Pooid, nondiagnostic + PACCAD GSSC morphotypes; POOID TOT, Pooid, diagnostic + Pooid, nondiagnostic GSSC morphotypes; PACCAD TOT, PACCAD GSSC morphotypes. LE, leaf; RE, reproductive material. See "Analysis of GSSC Assemblages" in Supporting Methods for further explanation.
Table 4. Hypothesis testing concerning the composition of fossil GSSC assemblages in Cenozoic samples from Colorado and Nebraska/Wyoming: GSSC assemblages with high frequencies of closed-habitat GSSC morphotypes. Note that in Montana/Idaho, there were very few closed-habitat grasses; hence, this test was not performed on these samples. Rejection of H0: R, H0 rejected using both leaf and reproductive material urns (when used); NR, H0 not rejected using both leaf and reproductive material urns; R/NR, H0 rejected using leaf material urn but not using reproductive material urn. For further explanation, see Table 3 and "Analysis of GSSC Assemblages" in Supporting Methods.
Table 5. Hypothesis testing concerning the composition of fossil GSSC assemblages in Cenozoic samples from Colorado, Montana/Idaho, and Nebraska/Wyoming: GSSC assemblages with high frequencies of open-habitat GSSC morphotypes. For further explanation, see Tables 3 and 4 and "Analysis of GSSC Assemblages" in Supporting Methods.
Supporting Methods
Analysis of GSSC Assemblages. Studies (ref. 1 and unpublished data) show that closed-habitat grasses can produce fair amounts of phytoliths typical of open-habitat grasses, and, conversely, open-habitat grasses can contain a certain abundance of morphotypes typical of closed-habitat grasses. The percentage of closed-habitat GSSC morphotypes is therefore a very coarse measurement of the abundances of closed vs. open-habitat grasses. For this reason, the composition of the fossil GSSC assemblages was also investigated by comparing it to the assemblage composition of modern grasses. This was done through hypothesis testing using bootstrapping (2). The hypotheses that were tested differed depending on the overall composition of the short cell assemblages; the various cases are summarized in Table 3. In an assemblage with high relative frequencies of closed-habitat GSSC morphotypes (CH TOT), presumably reflecting a high abundance closed-habitat grasses, the first hypothesis to be tested was whether the assemblage likely reflects the presence of open-habitat grasses in addition to closed-habitat grasses, based on the abundance of morphotypes typical of open-habitat grasses (OH TOT) (total sum of open habitat morphotypes = Pooid, diagnostic + Pooid, nondiagnostic + PACCAD). The null hypothesis is that the variation observed in the GSSC assemblage can be entirely explained by the phytolith production of closed-habitat grasses. The relationship between morphotypes typical of open-habitat grasses (OH TOT) and closed-habitat morphotypes (CH TOT) expressed as a ratio ((CH TOT)/(OH TOT)) served as the metric for this test.
The background universe (or urn) against which this ratio was tested was the maximum ratio of OH TOT and CH TOT morphotypes obtained from the modern reference collection (unpublished data), because this represents the "worst-case scenario." Note that this is a very conservative approach. For each ratio, separate background universes were constructed for leaf and reproductive material (using maximum ratios), because these materials often have very different GSSC composition (refs. 1 and 3 and unpublished data). Thus, each fossil assemblage was compared to two different urns. In this case, the maximum ratio between OH TOT and CH TOT is 0.954 for leaf material (Otatea) and 1.08 for reproductive structures (Pharus; Table 3). The 95% confidence intervals for the expected ratios (urn) were calculated for each fossil sample, and the upper limit of these confidence intervals was compared to the actual ratio of morphotypes observed in the fossil sample [this corresponds to a one-tailed test at a = 0.025 (2)]. Bootstrapping with 1,000 replicates using RESAMPLING STATS 5.0 (available at www.resample.com) were used for these calculations (Table 4). If the test metric in the actual fossil GSSC assemblage exceeds the upper 95% confidence limit for both leaf and reproductive material of the background universe, the null hypothesis can be rejected (i.e., in this case, it is likely that open-habitat grasses were present in the grass community in addition to closed-habitat grasses). If it does not exceed either of the upper 95% confidence limits, the null hypothesis cannot be rejected (i.e., in this case, it is not likely that open-habitat grasses existed in the grass community in addition to closed-habitat grasses, but see below). If the test metric exceeds the upper 95% confidence limit for the background universe of leaf material, but not reproductive material, it is somewhat equivocal, and more careful judgment is necessary (see further below).
The ratio of diagnostic pooid types (POOID-D) to CH TOT phytoliths and the ratio of PACCAD morphotypes (PACCAD TOT) to CH TOT phytoliths were analyzed in the same manner to further evaluate the contribution from pooid (hypothesis 2 in Table 3) and PACCAD (hypothesis 3) grasses, respectively (Table 4). The reason for excluding nondiagnostic and unknown GSSC (contained in "Other GSSC" in Table 2) from these calculations is that it is not clear whether this compound variable encompasses hitherto undescribed closed-habitat morphotypes. It should be mentioned that the sample from Otatea culm, which only contains GSSC morphotypes that are not typical of closed-habitat grasses (unpublished data), was not considered for the calculations of background universes. Note also that the background universe ratios are based on a very limited data set; this analysis is therefore an initial attempt at quantifying the relative contribution of closed- and open-habitat grasses.
In addition to the various ratios, a more detailed study of particular morphotypes was undertaken. Hence, crenate morphotypes (diagnostic of pooids), large, decorated Stipa-type bilobates (stipoid pooids), which are only rarely found in modern closed-habitat grasses, and inverted bilobates (PACCAD grasses), which are so far not found in closed-habitat grasses (1) (unpublished data), were used to indicate the presence of open-habitat grasses. This more detailed information was used in conjunction with the statistical tests described above to evaluate the presence and/or relative contribution from pooid and PACCAD open-habitat grasses in each fossil GSSC assemblage. Thus, even if the bootstrap tests failed to reject the hypothesis that the assemblages reflect the presence of open-habitat grasses in addition to closed-habitat grasses for both leaf and reproductive material, observation of highly diagnostic morphotypes allowed the recognition of open-habitat grasses. In Fig. 4, open-habitat grasses are only marked when they were judged to have been clearly present.
In assemblages with high frequencies of morphotypes typical of open-habitat grasses (OH TOT), the contribution from closed-habitat grasses was estimated by using a ratio between CH TOT and OH TOT forms. This ratio was compared to a background universe based on (maximum) ratios derived from open-habitat grasses in the modern reference collection (hypothesis 4 in Table 3) using bootstrapping (Table 5).
1. Piperno, D. R. & Pearsall, D. M. (1998) Smithsonian Contrib. Bot. 85, 1-40.
2. Simon, J. L. (1997) Resampling: The New Statistics (Duxbury, Belmont, CA).
3. Mulholland, S. C. (1989) J. Archaeol. Sci. 16, 489-511.