Jump to ContentJump to Main Navigation
Biodiversity Response to Climate Change in the Middle PleistoceneThe Porcupine Cave Fauna from Colorado$

Anthony Barnosky

Print publication date: 2004

Print ISBN-13: 9780520240827

Published to California Scholarship Online: March 2012

DOI: 10.1525/california/9780520240827.001.0001

Show Summary Details
Page of

PRINTED FROM CALIFORNIA SCHOLARSHIP ONLINE (www.california.universitypressscholarship.com). (c) Copyright University of California Press, 2018. All Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a monograph in CALSO for personal use (for details see http://california.universitypressscholarship.com/page/535/privacy-policy-and-legal-notice). Subscriber: null; date: 21 June 2018

Age and Correlation of Key Fossil Sites in Porcupine Cave

Age and Correlation of Key Fossil Sites in Porcupine Cave

Chapter:
(p.64) Seven Age and Correlation of Key Fossil Sites in Porcupine Cave
Source:
Biodiversity Response to Climate Change in the Middle Pleistocene
Author(s):

Anthony D. Barnosky

Christopher J. Bell

Publisher:
University of California Press
DOI:10.1525/california/9780520240827.003.0007

Abstract and Keywords

This chapter assesses how key deposits in the cave (Pit, DMNH Velvet Room excavation, CM Velvet Room excavation, Badger Room) relate temporally to each other, to some other localities in the cave, and to the chronologic time scale. It describes the application of acid racemization for relative age placements of Porcupine Cave localities, and assesses the application of magnetostratigraphic, biochronologic, and biostratigraphic techniques to key fossil sites in Porcupine Cave.

Keywords:   Pit, Velvet Room excavation, Badger Room, acid racemization, Porcupine Cave, magnetostratigraphic technique, biochronologic technique, biostratigraphic technique

Establishing chronologic control in early and middle Pleistocene deposits is difficult, all the more so in cave deposits. In the absence of a lucky infusion of datable volcanic ash (which Porcupine Cave so far seems to lack), dating methods typically are restricted to magnetostratigraphic associations, biostratigraphic and biochronologic correlations, amino acid racemization, electron spin resonance, and uranium series dating. The latter two techniques have not been applied to samples from Porcupine Cave, although there is still opportunity for future attempts; therefore, all chronologic control comes from the former three. The chronologic information is augmented with sedimentological information that helps sort deposits into glacial and interglacial deposits.

Much of the information on dating the Porcupine Cave deposits, especially that from the Pit excavation, was presented in detail by Bell and Barnosky (2000). Other relevant information was detailed in Barnosky and Rasmussen (1988), Wood and Barnosky (1994), and Barnosky et al. (1996). Friedmann and Raynolds (chapter 6) discuss magnetostratigraphy in the DMNH Velvet Room excavation, and Shabel et al. (chapter 22) discuss age control in the Badger Room. This chapter integrates the earlier interpretations with some new information to hypothesize how key deposits in the cave (Pit, DMNH Velvet Room excavation, CM Velvet Room excavation, Badger Room) relate temporally to each other, to some other localities in the cave, and to the chronologic time scale. The conclusions are basically in accord with those of Bell and Barnosky (2000) but suggest that the age of the top of the Pit sequence is probably closer to circa 780 Ka than previously thought. (See chapter 2 for map locations and excavation information for each of the localities discussed in this chapter.)

The Pit as a Key Reference Section

The least complicated and possibly the temporally longest stratified sequence known in the cave is the CM/UCMP Pit excavation. It was excavated in 14 stratigraphic levels defined either by sedimentological breaks or as being less than 10 cm thick, whichever was thinner. Figure 23.1 illustrates the deposits and provides a schematic of the section. As detailed in Bell and Barnosky (2000), the sedimentological evidence suggests that at least three cool-warm cycles (probably glacial-interglacial cycles) are represented. The oldest stratigraphic levels—14, 13, 12, and 11—have been interpreted as interglacial deposits, although there is no firm basis for understanding levels 14 and 13. Level 10 has been interpreted as glacial; levels 9, 8, 7, and 6, as interglacial; levels 5 and 4, as glacial; and the youngest levels, 3, 2, and 1, as interglacial.

The sediments—dark brown clay and clay nodules for glacials, and less consolidated tan and light brown dust for interglacials—suggest that each glacial was moist and cool relative to its subsequent interglacial. Barnosky and Rasmussen (1988) and subsequent papers (Wood and Barnosky, 1994; Barnosky et al., 1996; Bell and Barnosky, 2000) assumed, based only on the upper glacial-interglacial transition, that all the glacials were moister than all the interglacials. However, the more complete set of information now available from the lower levels suggests that the lower interglacial represented by levels 9–6 was probably moister than the glacial represented by levels 5 and 4, based on the climatic implications of the small mammal remains found there. Palynological, sedimen-tological, and invertebrate fossil data support this interpretation for Hansen Bluff, Colorado, approximately 200 km south of Porcupine Cave (Rogers et al., 1992). The warmest, most arid of all the interglacials represented at Porcupine Cave appears to have been the youngest one (levels 3–1), as indicated by the character of the sediments (a distinctive loose, dry dust that occurs nowhere else in the sequence), the dominance of xeric taxa such as the ground squirrel Spermophilus, and the diversity of herpetofauna (Bell and Barnosky, 2000).

Magnetostratigraphy

Magnetostratigraphic studies in the Pit were undertaken by V. A. Schmidt, University of Pittsburgh Paleomagnetics Laboratory, just prior to his untimely death. Samples from levels (p.65) 14–8 recorded predominantly reversed polarities, although there is an admixture of intermediate and to a much lesser extent normal components as well. The data suggested to Schmidt (pers. comm., 1993) that levels 14–8 are older than 780 Ka, which marks the Brunhes-Matuyama boundary. They would fall in Chron 1r of Berggren et al. (1995). Samples from the base of level 5 provided equivocal magnetic signatures. Bell and Barnosky (2000) suggested that the presence of predominantly reversed samples with intermediate and normal components in levels 14–8 was consistent with sampling a time near the Brunhes-Matuyama transition from reversed to normal polarity. Viable magnetostratigraphic samples were impossible to obtain above level 8 because of the friable sediments.

Biochronology

Figures 7.17.3 show stratigraphic ranges of Pit taxa. Arvico-line rodents currently provide the most feasible link from biostratigraphic to biochronologic time scales (Repenning, 1987; Fejfar and Repenning, 1992; Bell, 2000; Bell et al., in press). At least 10 species of arvicolines range through the Pit (figure 7.2). Biochronologically informative species fall into two groups. One group, including Phenacomys gryci, Mimomys virginianus, and Allophaiomys pliocaenicus, is known elsewhere, primarily from sites older than 800 Ka. The second group is composed of species that elsewhere first appear 900–800 Ka ago: Microtusparoperarius, Lemmiscus curtatus (primitive four-triangle morphotype), and Microtus meadensis (see figures 7.17.3 and tables 10.110.13 for common names). The sympatric occurrence of all these species at Porcupine Cave suggests that levels 8–4 were deposited sometime between 900 and 800 Ka ago. Bell and Barnosky (2000) provided details of the species ranges for the relevant taxa and for species identifications. Bell and Barnosky (2000) believed that levels 8–4 dated to between 850 and 750 Ka. This slightly younger estimate assumed that the Brunhes-Matuyama boundary was captured in the sequence somewhere above level 8. However, based on the absence of viable magnetostratigraphic information above level 8 and the biostratigraphic correlations to the DMNH Velvet Room sequence explained later in this chapter, an estimate of 750 Ka for level 4 could be too young.

Stratigraphically within the Pit, A. pliocaenicus, M. virginianus, and P. gryci disappear at the level 4/3 transition. This probably does not reflect sampling bias because levels 4–1 produce the most specimens of any level. Given that these species disappear by 800 Ka ago elsewhere (by circa 1.5 Ma ago in the case of P. gryci; Repenning et al., 1995), it seems probable that levels 4 and below date to at least 800 Ka. The taxa remaining in levels 1 and 2 may have survived elsewhere to about 250 Ka in the case of M. paroperarius, M. meadensis, and Mictomys kansasensis/meltoni, and to the present for L. curtatus. However, most localities with Mictomys kansasensis/meltoni date to between 2 Ma and 600 Ka, with the most consistently reliable dates on M. meltoni being from 600 to 700 Ka. In view of this observation, plus the fact that levels 1–3 probably represent an interglacial that lasted no more than 100,000 years, it seems unlikely that level 1 would be younger than about 600 Ka. These age estimates based on biochronology are consistent with the independent magnetostratigraphic interpretations. Note, however, that the estimate is substantially older than that posited by most earlier papers on Porcupine Cave (Barnosky and Rasmussen, 1988; Wood and Barnosky, 1994; Barnosky et al., 1996). Dating the top of the sequence is still problematic. Bell and Barnosky (2000) estimated that Pit levels 1–3 were deposited sometime between 252 and 750 Ka ago. Based on the information presented in this chapter, we revise Bell and Barnosky’s (2000) estimate of the upper age of the Pit to be greater than 600 and very probably closer to 800 Ka.

Levels 9 and below contain two additional taxa that, like P. gryci, became extinct elsewhere by early Irvingtonian time (circa 1.5 Ma ago). These are the rabbit Hypolagus sp. and the ground squirrel Spermophilus (Otospermophilus) sp.; the latter was not named as a species but has morphological affinities to early Blancan species in Kansas (Rexroad and Fox Canyon faunas) (Goodwin, chapter 17). Unlike P. gryci, both disappear from the Porcupine Cave record by level 9 (figure 7.1). Sampling issues might explain their absence in level 9, where fossils are scarce, but not above that level because specimens become more abundant, and congeners are well represented in higher levels. Hypolagus and the Rexroad / Fox Canyon-like morph of Spermophilus (Otospermophilus) sp. co-occur with L. curtatus, but the other two arvicolines that suggest ages younger than 900 Ka do not appear until higher in the strati-graphic section (figure 7.2). This finding indicates either significant temporal range extensions for Hypolagus and the Rexroad / Fox Canyon-like morph of Spermophilus (Otosper-mophilus) sp., as was apparently also the case for P. gryci (Bell and Barnosky, 2000), or an even earlier first appearance for Lemmiscus curtatus than has previously been recognized. Based on its co-occurrence with P. gryci, M. virginianus, and A. pliocaenicus, Bell and Barnosky (2000) agreed with Repenning (1992) in extending the range of L. curtatus from circa 300 to 800 Ka, a decision that has been further substantiated by the association of L. curtatus with magnetically reversed sediments in SAM Cave, New Mexico (Rogers et al., 2000). In light of this, it seems more parsimonious to extend the range of one species (L. curtatus) downward rather than those of two species (Hypolagus and the primitive, unnamed Spermophilus [Otospermophilus] sp.) upward. Therefore, it would not be surprising if levels 10 and below in the Pit were substantially older than 800 Ka, even in view of the presence of L. curtatus. On these grounds the bottom of the Pit deposit is considered younger than 1 Ma.

Biostratigraphic Zones

Biostratigraphic changes within the stratified Pit sequence may prove useful for correlation of deposits within Porcupine Cave. Two zones based on first or last known stratigraphic appearance data are recognized (figures 7.17.3). These zones conceptually resemble the lowest known stratigraphic datum (LSDk) and highest known stratigraphic datum (HSDk), respectively, as described by Walsh (1998). The Hypolagus Zone is defined by the HSDk for Hypolagus. The overlying Marmota Zone is defined by the LSDk for Marmota. In practice, the approximate boundary between the zones is placed at the boundary between levels 8A and 9. That placement splits the difference between the last appearance of Hypolagus and the first appearance of Marmota in the section, and it takes into account the fact that fossils are rare enough in levels 8A and 9 that it is impossible to know if further sampling would produce either taxon. For correlation purposes, other localities in the cave that contain Hypolagus most likely correlate with Pit levels below 8A; those that contain Marmota have a high probability of correlating with levels above 9.

(p.66)

Age and Correlation of Key Fossil Sites in Porcupine Cave

Figure 7.1 Stratigraphic ranges of Amphibia, Reptilia, ochotonids, leporids, sciurids, and geomyids at the Pit. These ranges and those in figures 7.2 and 7.3 are used to define the biostratigraphic zones noted at the top and explained in the text. Levels 13 and 14 lacked fossils; thus the diagram extends back only to level 13. Abbreviations: G, glacial; IG, interglacial; LSDk, lowest known stratigraphic datum; HSDk, highest known stratigraphic datum; M. 5T Z., Microtus 5T Zone.

(p.67)

Age and Correlation of Key Fossil Sites in Porcupine Cave

Figure 7.2 Stratigraphic ranges of arvicolines and other rodents at the Pit. See figure 7.1 for further explanation.

One assemblage zone is recognized: the Allophaiomys Zone, which is defined on the co-occurrence of Allophaiomys plio-caenicus and Mimomys virginianus. A characterizing taxon is Phenacomys gryci (figure 7.2). Other localities within the cave that produce these taxa have a high probability of correlating with levels 4–10 if both of the definitive taxa are present.

Two relative abundance zones are also apparent, based on arvicoline rodent percentages detailed in chapter 23 (see figure 23.7). The Mictomys Zone (figure 7.2) is defined by Mictomys exhibiting a higher percentage of individuals (〉25%) within the arvicoline rodent component of the fauna than the combined total of Microtus 5T (i.e., Microtus characterized by five or more triangles on the first lower molar) and Micro-tus meadensis. The Microtus 5T Zone has relatively low numbers of Mictomys specimens (〈25%), but high percentages of Microtus 5T and/or M. meadensis specimens (〉25% combined). Ambiguity in differentiating the zones arises when the percentages of Mictomys versus combined Microtus 5T-M. mead-ensis are about equal. This situation occurs in level 3 and to a lesser extent in level 4. However, above and below those levels, the percentages provide a reasonable basis for biostratigraphic differentiation. The boundary between these relative abundance zones is thus recognized to be somewhat fuzzy and to encompass the level 4/3 transition. In general, the Mictomys Relative Abundance Zone overlaps broadly with the Allo-phaiomys Assemblage Zone, running from level 12 to level 5, before becoming hard to differentiate in level 4. The Microtus 5T Zone spans levels 1 and 2 before becoming hard to differentiate in level 3 (figure 7.4).

Goodwin (chapter 17) also recognized conceptual equivalents to relative abundance zones based on the sciurid (p.68) component of the fauna. Levels 6 and deeper contain ?Cyno-mys andersoni as the only prairie dog species; levels 4 and 5 yield ?C. andersoni and C. cf. C. leucurus in about equal abundances; and levels 1–3 produce C. cf. leucurus almost exclusively. Wood and Barnosky (1994), Barnosky et al. (1996), and Bell and Barnosky (2000) noted dramatic increases in the relative abundance of Cynomys and Spermophilus relative to Marmota at the level 3/4 transition.

Age and Correlation of Key Fossil Sites in Porcupine Cave

Figure 7.3 Stratigraphic ranges of carnivores, artiodactyls, and equids at the Pit. See figure 7.1 for further explanation.

Lemmiscus exhibits a population-level change that may be useful in correlation. Lower first molars with only four closed triangles are equal or greater in abundance than those with five closed triangles in levels 3–10 (see figure 23.8). The first six-triangle forms appear in level 1. Thus other Porcupine Cave localities in which Lemmiscus populations are characterized by a predominance of five-triangle forms correlate more likely with levels 1 and 2 than with lower levels of the Pit, and populations with six triangles probably are coeval with or younger than the uppermost Pit level.

The CM Velvet Room Excavation

Biostratigraphy

As noted in chapter 2, this excavation included five strati-graphic levels, the sediments of which very closely resembled the distinctive loose, dry dust of levels 1–3 of the Pit. Paleo-magnetic samples taken from beneath the lowest level were interpreted as questionably normal by Fred Luiszer and V. A. Schmidt (pers. comm., 1993). Most of the fossils have not yet been processed, but based on the small sample of arvicolines so far identified (see table 10.11), the deposits seem to fall within the Microtus 5T Relative Abundance Zone. The lowest fossiliferous level in the excavation, level 3, is the only one from which a reasonably large sample of arvicolines has been identified, and it shows the following percentages for minimum number of individuals: Mictomys, 19%; M. meadensis, 19%; and Microtus 5T, 23%.

(p.69)

Age and Correlation of Key Fossil Sites in Porcupine Cave

Figure 7.4 Correlation of the Pit sequence with the CM and DMNH Velvet Room excavations. Correlation of the sections is based on the biostratigraphic zones defined in the Pit and the local magnetostratigraphic evidence in each section. Correlation to the time scale is by biochronology and assumption that the reversed (white) magnetic interval indicates the Matuyama magnetic epoch. Dark gray shading denotes parts of the respective sections that fall in the Mictomys Zone. Climatic intervals are defined by sediment type. Numbers in the Pit Sequence column schematically represent the stratigraphic levels. Heavy dotted lines denote suspected unconformities. N in the Velvet Room columns indicates normally magnetized (black) parts of the section. Question marks in the Pit Paleomag. column indicate lack of data. The question mark at the base of the DMNH Velvet Room column expresses uncertainty about the age of the basal part of the section.

Of interest is one specimen of Allophaiomys pliocaenicus present in level 3. This finding hints that level 3 of the CM Velvet Room could fall within the Allophaiomys Assemblage Zone of the Pit, although the other requisite taxon, Mimomys virginanus, has not been recovered from the CM Velvet Room. The presence of A. pliocaenicus suggests that level 3 in the CM Velvet Room could be as old as level 4 in the Pit.

However, all other lines of evidence suggest that the CM Velvet Room deposits are younger than the Allophaiomys Assemblage Zone sensu stricto (i.e., also containing Mimomys virginianus) in the Pit. In the Lemmiscus sample from level 3 in the CM Velvet Room, approximately 31% of the specimens are four-triangle morphs; the rest are five-triangle forms. In the highest level of the CM Velvet Room, 100% of the Lemmiscus specimens exhibit either five or six triangles; six-triangle forms also occur in level 2A. This finding would suggest that levels 1 and 2 in the Velvet Room are younger than any levels in the Pit, possibly excepting Pit level 1 (see figure 23.8).

These lines of evidence suggest that the CM Velvet Room excavation is at least as young as levels 1–4 in the Pit. The presence of six-triangle forms of Lemmiscus in CM Velvet Room level 2A implies correlation with level 1 (or younger) in the Pit. The high percentage of five-triangle morphs of Lemmiscus in CM Velvet Room level 3, combined with the clear dominance of Microtus 5T and M. meadensis over Mictomys, suggests that even that level is younger than Pit level 4, despite the occurrence of Allophaiomyspliocaenicus. The equivocally normal paleomagnetic signature at the base of the excavation hints that the sediments in the CM Velvet Room excavation are younger than 780 Ka, but the paleomagnetic data are far from firm.

DMNH Velvet Room Excavation

The DMNH Velvet Room excavation is a stratified sequence excavated as described in chapter 2. Horizons are designated by letter, with A the highest and deeper horizons progressing through the alphabet. The sediments of horizons A–C are very similar to those of levels 1–3 of the CM Velvet Room excavation (chapter 2). All sediments are unconsolidated. The brown clays that intermittently characterize Pit levels below level 3 were not observed in the Velvet Room. However, DMNH Velvet Room horizons below C generally contain more clasts than horizons A-C (see figure 2.11).

The magnetostratigraphy of the site is described in chapter 6. The uppermost horizons are normally magnetized down to horizon C. Horizons D-I demonstrate reversed polarity. The Brunhes-Matuyama boundary at 780 Ka therefore appears to be located near the transition from horizon C to horizon D.

Biostratigraphy

The arvicoline and sciurid rodents are the only components of the fauna that have been studied sufficiently to allow biostratigraphic conclusions to be reached. Relative abundance calculations (table 7.1; see also tables 10.11, 10.12) place horizons C-F in the Microtus 5T Zone. The sample in horizon G is too small to record the pertinent taxa. Horizons H and I clearly fall within the Mictomys Zone, with Mictomys specimens composing 32% and 34% of the samples, respectively, and Microtus meadensis and Microtus 5T composing 0%.

(p.70)

Table 7.1 Percentages of Arvicoline Rodents Identified from the DMNH Velvet Room Excavation

 

Horizon

Taxon

A

B

C

D

E

F

G

H

I

Phenacomys sp. (not P. gryci)

4

33

Microtus meadensis

54

33

40

32

Lemmiscus (4 triangles)

8

7

4

23

33

Lemmiscus curtatus (sagebrush vole)

83

85

23

13

20

12

33

9

Microtus paroperarius

2

8

7

7

4

33

41

33

Microtus sp. (5+ triangles)

8

20

15

33

7

4

Mictomys kansasensi/meltoni

13

7

16

23

17

Mictomys sp.

7

28

9

17

Total minimum number of individuals

12

51

13

15

14

25

3

23

7

Horizons A-G are characterized by a majority of Lemmiscus specimens with at least five triangles, resembling levels 1 and 2 in the Pit. In fact, the proportion of five-triangle specimens in Velvet Room horizons A-C (treated as a lumped sample) is significantly greater than that in Pit levels 1 and 2 (p 〈 0.0001 for χ2 and Fisher’s exact test). Given the trend for populations of Lemmiscus to exhibit more five-triangle forms through time, horizons A-C in the Velvet Room probably are younger than levels 1 and 2 in the Pit. Horizons H and I yield more four-triangle forms, resembling Pit levels 3 and below. More four- than five-triangle forms characterize horizons H and I (p 〈 0.0001 for both χ2 and Fisher’s exact test) in comparison with levels A, B, and C in the DMNH excavation and levels 1 and 2 in the Pit, but not in comparison with levels below 2 in the Pit (p = 0.15 for Fisher’s exact test; p = 0.08 for χ2). This would be consistent with correlating horizons H and I somewhere below Pit level 2.

DMNH Velvet Room horizons D-G contain more five-triangle than four-triangle Lemmiscus. They show statistically more four-triangle forms than horizons A-C, but fewer four-triangle forms than horizons H and I (p 〈 0.005 for χ2 and Fisher’s exact test). Individually each of horizons D-G produced a very small sample of Lemmiscus (see tables 10.11, 10.12), but even so most horizons exhibit (horizons E and F) or approach (horizon D) statistically significant differences with the horizon H and I sample, but not with the horizon A-C sample (horizon E, with statistically more four-triangle forms, being the exception). Horizon G, with a single specimen, could not be statistically differentiated from either the A-C or H and I samples. Horizons D-F as a lumped sample could not be statistically distinguished from Pit levels 1–3 (p = 0.77 for χ2; p 〉 0.9999 for Fisher’s exact test). The Lemmiscus morpho-type percentages thus are consistent with correlating DMNH Velvet Room horizons D-F near Pit levels 1–3. However, consideration of the Mictomys/Microtus (44%/36%) abundances constrains placement of Velvet Room horizon F nearer to Pit levels 3 or 4—the transitional zone between the two relative abundance zones—than to Pit level 1.

Based on the derived condition of Cynomys cf. C. leucurus and Spermophilus cf. S. elegans, Goodwin interpreted horizons A and B as younger than the entire Pit sequence (chapter 17).

These considerations suggest that horizons A-C are younger than levels 1–3 in the Pit (figure 7.4). Horizons F-I probably correlate with Pit levels 4 and below. Interestingly, Baxter (chapter 15) reported one Hypolagus specimen from horizon G in the DMNH Velvet Room excavation, which would imply that horizons G and below extend into the Hypolagus zone, i.e., are at least as old as levels 9 or 10 in the Pit. If so, an unconformity or extremely low depositional rates between horizons F and G would be implied.

However, considerably more study of the complete DMNH Velvet Room fauna is needed before these conclusions should be regarded as firm. For example, the absence of taxa such as Allophaiomys pliocaenicus and Mimomys virginianus presents an inconsistency if DMNH Velvet Room horizons F-I correlate with levels 4 to as old as 10 in the Pit. However, these taxa are always found in low abundance in the Pit, which yielded only 23 specimens of A. pliocaenicus and 17 of M. virginianus in the sample of 1004 arvicolines. The DMNH Velvet Room sample comprises 254 arvicoline specimens, only 118 of which are from horizons where the two “missing” taxa would be expected. Therefore it is possible that the absence of these taxa (p.71) simply reflects the relatively lower sample size. Both taxa, as well as Phenacomys gryci, are present in DMNH Velvet Room excavation samples that could not be reliably assigned to a stratigraphic horizon (see table 10.13).

The Badger Room

Arvicoline rodents from the Badger Room clearly place it within the Allophaiomys Assemblage Zone and Mictomys Relative Abundance Zone (see table 10.1). The only arvicolines identified from the sample of 50 teeth are A. pliocaenicus, Mimomys virginianus, Phenacomys gryci, Mictomys kansasensis/ meltoni, and a single specimen of Microtus 5T. Also present in the fauna is Marmota, a finding that places the locality securely within time encompassed by the Marmota LSDk Zone in the Pit. These data indicate that the Badger Room sediments analyzed by Shabel et al. (chapter 22) are coeval with some part of the interval represented by levels 8–4 in the Pit. The absence of Lemmiscus curtatus in this assemblage is puzzling, given that Lemmiscus is represented in most Pit levels. Small sample size may explain this, as may the different taphonomic vector that apparently characterizes the Badger Room relative to the Pit (Shabel et al., chapter 22).

Amino Acid Racemization

Relative age placements of Porcupine Cave localities derived from amino acid racemization are consistent with the biostratigraphic and paleomagnetic conclusions noted previously. Bell and Barnosky (2000) described the relative age results obtained from the Badger Room, Pit level 2, Pit level 6, and CM Velvet Room level 1. The technique suggested that Pit levels 2 and 6 and the Badger Room samples were approximately similar in age, but that CM Velvet Room level 1 was substantially younger. A sample from the Gypsum Room was intermediate in age between the Pit and CM Velvet Room samples. Absolute age derivations from amino acid racemization were inconsistent with other age data from Porcupine Cave and were deemed unreliable because assumptions for applying the technique to absolute age determinations, such as constant temperature through time, clearly did not hold.

Additional Paleomagnetic Data

To assess the feasibility of magnetostratigraphic analyses in the cave, a series of samples collected in 1988 was analyzed by Fred Luiszer in the University of Colorado Paleomagnetism Laboratory; the results were interpreted by Luiszer and subsequently by V. A. Schmidt (pers. comm., 1993). The results suggested that sediments about 10 cm below the surface of the Gypsum Room were potentially normally magnetized. Additional samples were taken and interpreted by Schmidt in 1990. Those from Tobacco Road taken near the surface from sediments about 1 m below the ceiling showed some signs of reversal; this location was about two-thirds of the way toward the Velvet Room. Farther toward the Velvet Room, samples from near the surface exhibited normal polarity. None of these samples provided conclusive results, but the observations indicated that future work could prove fruitful.

Climatic Correlations

In the Pit, the contacts between glacial and interglacial strati-graphic levels are abrupt, with dramatically differing sediments on each side of the contact. Within each climatic interval, the sedimentary transitions across stratigraphic levels are much more subtle, or, in the case of arbitrary levels, lacking entirely (Barnosky and Rasmussen, 1988; Bell and Barnosky, 2000). This pattern suggests episodic deposition, with periods of nondeposition (unconformities or disconformities between climatic intervals) separating a sequence of discrete sediment packages, each of which samples some unknown slice of time within its respective climate interval. Which of the several middle Pleistocene glacial and interglacial stages are represented is unknown, although the dating considerations discussed earlier provide some constraints.

If the correlation expressed in figure 7.4 is correct, Pit levels 1–3 probably accumulated during interglacial stage 19 or 21, using the oxygen isotope chronology of Raymo et al. (1997) and Raymo (1998) (figure 7.5). Because the correlation of Pit to Velvet Room sediments suggests that levels 1–3 are older than 780 Ka, stage 21 is the most likely candidate. In that case, the climatic intervals that precede levels 1–3 may match sequentially with the glacial-interglacial cycles interpreted from the oxygen isotope curve (Raymo et al., 1997; Raymo, 1998), as indicated in figure 7.5 (climatic intervals 22, A, B, C). This would be consistent with the sedimentological and faunal evidence (chapter 23), which suggests that levels 1–3 represent the warmest and most arid of the warm periods, because the oxygen isotope excursion that precedes stage 22 (A in figure 7.5) was considerably less pronounced than that for stage 21. Alternatively, some other sequential combination of the glacials and interglacials noted earlier and those labeled D, F (glacials) and E (interglacial) in figure 7.5 could be represented. It does not seem likely that glacials or interglacials older or younger than these are represented, in view of the suspected age of the Pit and Velvet Room sequences.

This interval of time is of interest from a climatic point of view in that it spans at least part of the transition from a 41,000-year rhythm for glacial-interglacial cycles to a 100,000-year rhythm. Transition to the 100,000-year cycle began about 1 Ma ago. By about 800 Ka ago the 100,000-year cycle started to become dominant (Raymo, 1998), and it was firmly established by circa 640 Ka ago (Schmieder et al., 2000). Thus the upper cycle at Porcupine Cave (levels 1–5) very probably accumulated during the time of the 100,000-year periodicity, whereas the lower Porcupine Cave levels may sample the transitional climatic intervals as the 41,000-year rhythm switched over to 100,000-year periodicity.

(p.72)

Age and Correlation of Key Fossil Sites in Porcupine Cave

Figure 7.5 Potential correlation of the Porcupine Cave stratigraphic sequences with global climate changes indicated by the oxygen isotope curve. The oxygen isotope data are adapted from Raymo et al. (1997).

Conclusions

The combination of magnetostratigraphic, biochronologic, biostratigraphic, and amino acid racemization techniques that has been applied to Porcupine Cave sediments suggests that the Pit sequence has an age near 780 Ka at its top and is as old as 900 Ka or 1 Ma at its bottom. The DMNH Velvet Room appears younger than the uppermost Pit sediments at its top (levels A-C), with levels D and E potentially coeval with Pit levels 1–3, level F in the vicinity of Pit levels 3 or 4, and levels G-I potentially as old as Pit levels 9 or 10. The CM Velvet Room sediments appear at least as young, and probably younger, than level 1 in the Pit. The Badger Room seems coeval with the interval represented by levels 8–4 in the Pit. The correlations between the Pit, Velvet Room, and Badger Room localities are based primarily on a combination of LSDk-HSDk zones, relative abundance zones, and assemblage zones defined for arvi-coline rodents in the Pit, which are the only component of the fauna that has been well enough identified from all the pertinent localities and levels for meaningful biostratigraphic application. These zones presently are applicable only within Porcupine Cave and possibly in the immediate vicinity; they should not be assumed to apply over broader geographic regions. The correlations are supported by biostratigraphic evidence from squirrels and rabbits, internally consistent paleo-magnetic data, and the amino acid racemization results.

Nevertheless, important caveats should be kept in mind. The stratigraphy in the cave is complex and unconformities probably abound, as illustrated in the Pit and probably in the DMNH Velvet Room. Especially in the DMNH Velvet Room, the assignment of specimens from a particular excavation bag to a definite stratigraphic interval is sometimes difficult. Moreover, the arvicoline biochronology is correlated to an independent time scale primarily from sites at much lower elevations than Porcupine Cave. Thus the role of biogeographic differences in influencing interpretations of temporal distributions has been little studied; in fact, the Porcupine Cave data are among the most applicable in this regard. Finally, many additional specimens remain to be identified from both of the Velvet Room excavations, as well as from most of the other localities in the cave. Much of the material has not even been processed out of the matrix.

As the sample size from the Velvet Room and other localities in Porcupine Cave grows in the coming years, additional analyses and comparisons to other important sites—such as those at nearby Hansen Bluff and SAM Cave (Rogers et al., 1985, 1992, 2000; Rogers and Wang, 2002)—should provide valuable tests of the chronology proposed here. These further (p.73) tests are particularly crucial because, as the most diverse sample of high-elevation Irvingtonian fauna located in the heart of the Rocky Mountains, Porcupine Cave provides a unique perspective on faunal dynamics, faunal provinciality in the West, and attendant biochronologic questions.

Acknowledgments

Preparation of this chapter was partially supported by NSF grant BSR-9196082. This chapter is University of California Museum of Paleontology contribution 1810.