Microtus von Schrank, 1798
Material. 11 Lm1s, FHSM VP-13771, 15637, 15644, 15647, 15648, 15654, 15710, 15711, 15734, 15749, 16382; 10 Rm1s, FHSM VP-15635, 15643, 15645, 15652, 15662–15664, 15668–15670.
Description. Twenty-one m1s of Microtus are characterized by five closed triangles and are assigned to Microtus sp. The teeth are ever growing and exhibit strong development of cementum in buccal re-entrant angles one through three and lingual re-entrant angles one through four. Presence and development of cementum varies from absent to strong in buccal re-entrant angle four and lingual re-entrant angle five. The teeth exhibit positive enamel differentiation (Martin, 1987).
A slight majority (n = 12, 57.1%) has an anterior cap with shallow re-entrants in buccal angle four and lingual angle five. This shallowness provides for a wider A–A' (observed range = 0.33–0.63 mm) and the appearance of slightly less developed triangles six and seven (Figure 3.1). The remainder (n = 9, 42.9%) have deeper re-entrants that allow for a narrower A–A' (observed range = 0.20–0.30 mm) and the appearance of better-developed triangles six and seven (Figure 3.2). The buccal re-entrant angle four and lingual re-entrant angle five of Microtus sp. are deeper than those of Variants A and C of M. paroperarius and M. llanensis. The opening between triangles five and six (AC) is narrower than in M. paroperarius and M. llanensis, and the opening between triangles four and five (C) is narrower than in M. llanensis and M. meadensis. The length of the anteroconid complex (A), total length (L), and length across triangles four and five (W) in Microtus sp. is greater than in M. llanensis (Table 1).
Two of the teeth assigned to Microtus sp. have a relatively square-shaped sixth triangle that includes a slight indentation at the approximate midpoint of its buccal edge (Figure 3.3). This indentation, often called a prism fold (Zakrzewski, 1967, figure 1A), occurs infrequently in extant species of Microtus (Zakrzewski, 1985). It occurs more commonly in some species of extinct, presumably more basal, arvicolines (e.g., Ogmondontomys, Cosmys). When present in these forms, it is found on triangle four.
The five closed triangle morphotype was considered a variant of Microtus paroperarius by
Paulson (1961) in his study of the Cudahy arvicolines; however, we follow the taxonomic conclusions of
Bell and Repenning (1999) who assigned this morphotype to an unidentifiable species of Microtus. The m1s from Fiene are the same size as those from Cudahy but smaller than those from Irvington (Table 2).
Comments. An m1 characterized by five closed triangles may represent the plesiomorphic condition for the most recent common ancestor of all North American Microtus (Bell and Bever, 2006), and apparently this morphotype was widespread across North America during the middle Pleistocene. However, with the exception of Fiene where m1s with five closed triangles constitute 43.8% of a combined sample of four- and five- triangle forms (n = 48) and Cathedral Cave (Jass, 2007), the five-triangle morphotype constitutes a very low percentage of the total sample relative to the four-triangle morphotype at other sites. For example, the five-triangle morphotype constituted 2.5% of the M. paroperarius sample at Cumberland Cave (n = 152) and 8–9% of the M. paroperarius sample from Sunbrite Ash Pit (n = 59). The five-triangle morphotype represented 17% of the combined five- and four-triangle sample of Trout Cave #2 (n = 18;
Pfaff, 1990). The samples at Cudahy and Hansen Bluff historically referred to M. paroperarius contained five closed triangles with frequencies of 7.3% and 8.1%, respectively (Bell and Repenning, 1999). Although unreported by
Eshelman and Hager (1984), the sample of m1s allocated to M. paroperarius from Hall Ash Pit (n = 13) contains a single five triangle, right m1 (7.7%).
These low percentages could be interpreted as supporting the conclusions of
Paulson (1961) that the five-triangle morphotype is a variant of M. paroperarius.
However, without a broadened understanding of variability in the development system producing these morphotypes and the phylogenetic pattern of that variability (sensu
Wagner and Altenberg, 1996; see
Bever 2009), the relative abundance of a variant morphology provides little empirical support for a taxonomic conclusion. We do not assign our five-triangle sample to a species following the recommendation of Bell and his coauthors, who in a series of recent papers (Bell and Repenning, 1999;
Bell and Barnosky, 2000;
Bell et al., 2004a;
Bell and Bever, 2006) presented cogent arguments, based on variation observed in extant populations, to refrain from assigning isolated teeth to the species level when the morphological data alone do not support such a refined identification. Historically, size, gestalt assessments of morphological uniqueness, and both stratigraphic and geographic distributions played an important role in refining taxonomic assignments to the species level (Bell and Bever, 2006; see
Bever, 2005 for a more general discussion of this methodology and its potential consequences). Discriminant function analyses of morphometric data (Smartt, 1977;
Wallace, 2006) and observations of enamel microstructure (Schmelzmuster;
Wallace, 1999) have demonstrated initial success in distinguishing extant species of Microtus. The results of these studies, with regards to their general application to taxonomic problems, however, suffers from sampling issues, and many additional species need to be studied to demonstrate the taxonomic distribution and phylogenetic polarity of these morphological data and thus the range of their usefulness.
Material. Variant A, 6 Lm1s, FHSM VP-13796, 15061, 15639, 15641, 15735, 15748; 9 Rm1s, FHSM VP-13795, 15636, 15640, 15649, 15651, 15653, 15655, 15657, 15659. Variant B, left dentary with m1–m2, FHSM VP-15674; 2 Lm1s, FHSM VP-13821, 15638; 3 Rm1s, FHSM VP-15642, 15656, 15767. Variant C, 4 Lm1s, FHSM VP-15660, 15658, 15671–15672; 2 Rm1s, FHSM VP-15634, 15648.
Description. The m1s (n = 27) of M. paroperarius are characterized by four closed triangles. The teeth are ever growing and exhibit strong development of cementum in buccal re-entrant angles one through three and lingual re-entrant angles one through four. Presence and development of cement varies from absent to strong in buccal re-entrant angle four and lingual re-entrant angle five. The teeth exhibit positive enamel differentiation.
Three morphotypes are recognized within the Fiene sample. The majority (n = 15, 55.6%; Variant A for discussion) has the pattern that characterizes the species wherein triangle five opens broadly (b > 0.13 mm) into the anterior cap on which a shallow buccal re-entrant angle four and lingual re-entrant angle five are present (Figure 3.4). The other morphotypes are divided evenly (n = 6 each; 22.2%) between specimens that show intermediate closure (Figure 3.5) of triangle five (b < 0.13 mm, Variant B for discussion) and those in which triangles five and six are equivalent (Variant C for discussion). Our Variant B is equivalent to the two morphotypes of
Bell and Repenning (1999) showing intermediate closure; i.e., 1 enamel band width of opening (B1) and < 2 and > 1 enamel band width opening (B2). In three of the six specimens exhibiting Variant C, the confluent triangles five and six are closed off from the anterior cap (Figure 3.6).
Bell and Repenning (1999) reported that 60 of the m1s from Cudahy have confluent triangles one and two. This morphotype was not found in the Fiene sample.
In statistical analyses, Variants A, B, and C were treated as independent samples and as a combined sample. Results obtained from the independent samples may not be meaningful because of small sample size, which tended to skew the measurements to one end of the range of variation observed in the non-paroperarius morphotypes. This skewness, on occasion, not only led to results suggesting significant differences between the individual paroperarius variants and the non-paroperarius samples but also between the combined sample of paroperarius and the latter samples. Keeping the previous in mind, a summary of the significant differences follows.
When the three variants are compared to each other, Variant B, as expected based on the way it was defined, is narrower between triangles five and six (B) than in Variants A and C, and Variant A is narrower between triangles five and six (B) than in Variant C (Table 1). The latter result is due to the confluence of triangles five and six in Variant C. Significant differences with Microtus sp. are listed in that section.
The combined and Variant A samples of M. paroperarius have a deeper buccal re-entrant angle four and lingual re-entrant angle five than M. llanensis. The constriction between triangles five and six (B) is wider in all M. paroperarius samples than in M. meadensis; wider in Variant C and narrower in Variant B than in M. llanensis. The constriction between triangles four and five (C) in all samples of M. paroperarius is narrower than in either M. llanensis or M. meadensis. This feature results from the former taxon having four closed triangles and the latter two taxa having triangles four and five confluent. The length across triangles six and seven (W') of the combined and Variant A samples and the length across triangles four and five (W), length of anteroconid complex (A), and total length (L) in all samples of M. paroperarius is greater than in M. llanensis. Except for Variant C, the m1 of M. paroperarius is wider (OW) than that of M. llanensis and is longer across triangles four and five (W) than M. meadensis. Variant A has a shorter anteroconid complex (A) than M. meadensis (Table 1).
The lengths of the Microtus paroperarius m1s (L) from Fiene fall within the range of those from Cudahy and encompass the range of those from Hall Ash Pit (Table 3). A similar relationship for length and ratio data exists between the m1s from Fiene and those from other selected Irvingtonian sites (Table 4). No significant differences were found in the lengths of Variant B m1s from Cudahy, whereas the m1 of Variant A in Cudahy was significantly shorter than that of Microtus sp. (Bell and Repenning, 1999). A mean length of 2.64 mm (n = 22) was reported for m1s of M. paroperarius from the Pit locality at Porcupine Cave (Bell and Barnosky, 2000). This mean value is smaller than that from any of the other sites considered herein and near the lower end of many of the observed ranges (Table 3 and
Comments. In their discussion of Microtus paroperarius,
Bell et al. (2004a) suggested the relationships of this taxon to M. operarius (a junior synonym of M. oeconomus) be carefully reconsidered. They pointed out that the close morphological similarity between the two taxa was implied by
Hibbard (1944) in his choice of names, by
Paulson (1961) in his study of the Cudahy material, and by
Van der Meulen's (1978)
suggestion that the two taxa could not be separated on the basis of dentition. In addition, all the examined specimens of M. oeconomus operarius in the Frank Pitelka Alaska collection at the Museum of Vertebrate Zoology, University of California, Berkeley, have the Variant A morphology of M. paroperarius (Bell, personal commun., 2010).
A summary of the geographic and temporal distribution of Microtus paroperarius was provided by
Bell et al. (2004a), who noted that recent finds of M. paroperarius west of the Rocky Mountains bring both the geographic and temporal range of the extinct species closer to that of the extant species. Although the Fiene specimens do not significantly alter our current understanding of the geographic and temporal range of M. paroperarius, the Fiene record, with its multiple morphotypes, does reinforce the need to consider a wider range of variation when drawing conclusions regarding both the taxonomic identification and secondary biological properties of M. paroperarius.
Material. 3 left dentaries with m1–m2, FHSM VP-13797, 15675, 15676; 19 Lm1s, FHSM VP- 13774, 13798, 13822, 15677, 15680–15682, 15684, 15685, 15690, 15691, 15693, 15697, 15699–15702, 15706, 15707; 14 Rm1s, FHSM VP-13935, 15678, 15683, 15886–15689, 15695, 15696, 15698, 15703–15705, 15766.
Description. The m1s (n = 36) of Microtus llanensis are characterized by three closed triangles and a confluent triangle four and five (= primary wings of
Repenning, 1992) that open into the anterior cap (Figure 3.7). The teeth are ever growing and exhibit strong development of cementum in buccal re-entrant angles one through three and lingual re-entrant angles one through four. Presence and development of cementum varies from absent to strong in buccal re-entrant angle four and
lingual re-entrant angle five. The teeth exhibit positive enamel differentiation.
Based on the m1s, Microtus llanensis is the most abundant and least variable of the arvicoline morphotypes at Fiene. With few exceptions, it is also the shortest (L), narrowest (OW), has the shortest anteroconid complex (A), and widest openings between triangles (B, C;
Table 1). These differences separate it from M. meadensis. Details of the differences with M. paroperarius and Microtus sp. are discussed under those sections.
The Fiene sample of Microtus llanensis has the lowest mean m1 length of any Irvingtonian locality for which mensural data are available (Table 5). The median m1 length for the Fiene sample is 2.63 mm, which is at or near the lower end of the observed ranges of two of the three other samples for which 30 or more specimens were available. Whether this difference is the result of sampling or has some temporal or ecological significance is currently unclear.
Comments. An existing hypothesis suggests that Microtus llanensis is part of a phyletic line that begins with Allophaiomys sp. and ends in M. ochrogaster (Van der Meulen, 1978), and the morphology that characterizes Allophaiomys sp. may indeed be the ancestral condition for M. llanensis. There are sites in addition to Fiene, such as the Kentuck (Zakrzewski, 1985), where both morphotypes are present, and it is difficult to separate them from each other. However, the phylogenetic analysis of
Conroy and Cook (2000) based on mtDNA sequence data supported a monophyletic North American Microtus in which the taxa characterized by three closed triangles were derived from five closed-triangle forms. These data complicate many historical hypotheses of dental evolution in arvicolines, including that of
Van der Meulen (1978), and emphasize the need to integrate fossils into explicit phylogenetic hypotheses. Except for the occurrence at Trout Cave #2 (Pfaff, 1990), other records of M. llanensis are restricted to west of the Mississippi River, east of the Rockies, and south of the Nebraska/Kansas border.
Microtus meadensis (Hibbard, 1944)
Material. 2 Lm1s, FHSM VP-15679, 15709; 5 Rm1s, FHSM VP-15692, 15694, 15708, 15733, 15750.
Description. The m1s (n = 7) of Microtus meadensis are characterized by three closed triangles and a confluent triangle four and five (primary wings of
Repenning, 1992) that is closed off from, or opens very slightly into, the anterior cap (Figure 3.8). The latter character distinguishes M. meadensis from M. llanensis, wherein the confluent set of triangles exhibits a wide opening into the anterior cap. Other differences between M. meadensis and other Fiene morphotypes are discussed in previous sections. Two of the seven m1s assigned to M. meadenesis exhibit a second set of confluent triangles—triangles six and seven (secondary wings of
Figure 3.9). The teeth are ever growing and exhibit strong development of cementum in buccal re-entrant angles one through three and lingual re-entrant angles one through four. Presence and development of cementum varies from absent to strong in buccal re-entrant angle four and lingural re-entrant five. The teeth exhibit positive enamel differentiation. The length of the m1s from Fiene fall within the range exhibited by the M. meadensis samples from the Cudahy and Sunbrite localities in the Meade Basin (Table 6).
Comments. A recent summary of the taxonomic status and geographic and temporal distribution of Microtus meadensis can be found in
Bell et al. (2004a).
Allophaiomys Kormos, 1932
Material. Lm1, FHSM VP-15712.
Description. The remaining m1 consists of three closed triangles with a confluent triangle four and five that opens broadly into the anterior cap (Figure 3.10). No additional re-entrants are present on the anterior cap, which nominally is indicative of Allophaiomys, and we allocate this specimen to Allophaiomys sp. The length of the Fiene m1 falls within the range of Allophaiomys from the Kentuck, Wathena, and Java localities (Table 7). The samples from Kentuck and Java originally were assigned to A. cf. A. pliocaenicus (Martin, 1975). Although a close resemblance between these two samples and A. pliocaenicus was acknowledged by
Van der Meulen (1978), he also recognized that the mean values for the ratio of the width of the opening between triangles five and six to the distance across triangles four and five
(B/W) were intermediate between those of A. pliocaenicus and A. deucalion and therefore assigned the samples to Allophaiomys sp. The B/W value for FHSM VP-15712 from the Fiene falls within the range of the values for the Kentuck and Wathena (Table 7) sample. Based on the distribution of these values and the lack of discrete characters that support a more refined identification, we follow Van der Meulen in restricting our taxonomic assignment.
Comments. The occurrence of a single specimen with a primitive morphotype in a sample of slightly less than 100 teeth presents us with at least two possible explanations. The first is to conclude that the specimen represents an individual of Allophaiomys. There are no specimens of Microtus llanensis in the Fiene sample that contain any hint of an Allophaiomys pattern as seen at some other localities. Allophaiomys, when present, is not found in great abundance at Kansas sites (Table 7). There is a minimum of four individuals with an Allophaiomys morphology at Courtland Canal (some 58 km northeast of Feine)—a site that contains no Microtus morphotypes (personal observation). However, in the nearby Hall Ash Pit, we found a specimen with five closed triangles in the sample reported as M. paroperarius by
Eshelman and Hager (1984)—but no Allophaiomys. Allophaiomys also is absent at Cudahy. There is a minimum of one individual of Allophaiomys in the Nash locality (Eshelman and Hibbard, 1981), four in Rick Forester, and two each in Aries A and Short Haul (Martin et al., 2003). It may be of importance that these Meade Basin records of Allophaiomys are all in the Borchers Badlands. If these Allophaiomys-like tooth records actually do represent the presence of Allophaiomys in the central Great Plains and are not a function of some taphonomic bias, they may reflect remnants of small populations either of early migrants moving into the area or stragglers that were leaving. As noted above, the relative scarcity of Allophaiomys records in Kansas is in contrast to the relative abundance of this taxon at sites in the Rocky Mountains, northern Great Plains, and Appalachians.
A second explanation is that FHSM VP-15712 belonged to an individual of Microtus llanensis that exhibited an Allophaiomys morphotype as an atavistic variant. Such a variant, however, has not been documented in any modern population of Microtus. The ability of a development system to produce this morphology as a population-level variant may have become extinct with M. llanensis or some other lineage, but this is difficult to support empirically. Another example of this type of intraspecific variation from the Fiene would be the two specimens mentioned under Microtus sp. that exhibit a prism fold. The difference in this case, as noted above, is that the presence of prism folds as a polymorphism in extant populations is documented.
Upper Third Molars
Material. Morphotype A, 17 LM3s, FHSM VP-15770–15772, 15774, 15782, 15783, 15787–15789, 15791–15794, 15796, 15798, 15800, 15803; 21 RM3s, 15747, 15768, 15769, 15773, 15775–15781, 15784–15786, 15790, 15795, 15797, 15799, 15801, 15817, 15818; Morphotype B, 10 LM3s, FHSM VP-15805–15807, 15809–15815; 8RM3s, 15802, 15804, 15808, 15816, 15819–15822.
Description. Fifty-six isolated M3s were recovered from Fiene. Two morphotypes are present (referred to here as A and B for discussion). Morphotype A (n = 38) consists of an anterior loop, three triangles, and a complex posterior loop. The posterior loop consists of a triangle four that opens into a posterior dentine field that varies from an oval to one that contains a potential triangle six (called a hook by
Figure 2.2). A sequence of specimens showing the possible development of the hook is figured beginning with a oval posterior field (Figure 4.1), followed by anterior elongation of the posterior field (Figure 4.2), formation of a shallow lingual re-entrant angle four (Figure 4.3), and a deepening of lingual re-entrant angle four to isolate the hook (Figure 4.4). Twenty-one M3s of morphotype A lack evidence of a lingual re-entrant angle four. The remaining 17 teeth exhibit a slight to well-developed lingual re-entrant angle four. The posterior loop is longer in morphotype A (Table 8).
The relation of triangle three to the posterior loop is variable in morphotype A. Triangle three can be closed (n = 14) from the loop or it can have a narrow (n = 13), intermediate (n = 9), or wide (n = 3) opening into the loop. Likewise, triangle three can be anterior to (n = 24) or opposite (n = 15) triangle four. In the latter case, triangles three and four often have a common opening into the posterior loop. The anterior loop is always closed from triangle one. Sixteen specimens of morphotype A have triangle one closed off from triangle two, 16 exhibit some degree of openness, and in seven specimens of A, triangles one and two are confluent (Figure 4.1, 4.3; = rhomb of
Bell and Repenning, 1999). Two specimens of morphotype A exhibit a slight opening between triangles two and three. One specimen of morphotype A exhibits a disrupted enamel pattern on the leading edge of triangle two.
Morphotype B (n = 18) consists of an anterior loop, two alternating triangles, a variable triangle three, and a simple posterior loop. Triangle three can be absent (Figure 4.5), poorly developed, and possibly considered part of the posterior loop (Figure 4.6), or better developed and opening into the posterior loop (Figure 4.7). A slight or moderately developed lingual re-entrant angle three often allows triangle four to form a hook (Figures 4.5–4.7). There is no evidence of triangles five and six. The posterior dentine field can be short (Figure 4.6) or elongate (Figure 4.7).
Three specimens in morphotype B, exhibit a slight opening between the anterior loop and triangle one, the remainder exhibit a closed condition. Ten specimens have triangle one closed off from triangle two, 10 specimens exhibit some degree of confluence, and nine specimens exhibit a slight opening between triangles two and three.
Both morphotypes exhibit positive enamel differentiation. Cementum is strongly developed in all re-entrants, except for the most posterior on each side, which often are not as deep as the others; therefore, cementum can be absent to strongly developed.
Comments. Previously, most workers, when presented with the two morphotypes described above, would probably assign morphotype A to one of the taxa with four or five closed triangles on m1 and morphotype B among the taxa with three closed triangles on m1. Indirect support for this approach is suggested by the presence of four M3s with morphotype A in Hall Ash Pit in conjunction with Microtus sp. and M. paroperarius. However,
Repenning (1992, figure 12A) described an associated skull and jaws of Microtus paroperarius from Hansen Bluff, wherein the M3 is more similar to morphotype B, the type associated with taxa possessing three closed triangles on m1. The third triangle opens into the posterior loop, triangle four forms a hook, and there is no evidence that a six triangle would develop. Therefore, because the Fiene M3s are all isolated and seven morphotypes of m1 are present, assignment of the M3s to a more refined group species is not feasible at this time.
Upper Second Molars
Material. 43 LM2s, FHSM VP-13773, 13794, 13799, 13819, 15738, 16878–16915; 44 RM2s, FHSM VP-15739, 16916–16958.
Description. All (n = 87) of the M2s from Fiene consist of an anterior dentine field or loop that is closed off from the first triangle. Triangle one is closed off from triangle two, and triangle two is closed off from the posterior dentine field. In the majority of M2s (n = 62, 71.3%) the posterior dentine field is composed of a confluent potential triangle three and four (Figure 5.1). The posterior end of the field can be straight or round. It may be oriented posteriorly or at a slight angle. In the remaining M2s (n = 25; 28.7%), triangle three is better defined and 'triangle four' consists of an incipient (n = 15,
Figure 5.2), intermediate (n = 4,
Figure 5.3), well-developed (n = 5,
Figure 5.4), or closed (n = 1, observed at base of tooth,
Figure 5.5) posterior lingual dentine field. This field once was thought to diagnose Microtus pennsylvanicus in North America but is now known to occur at varying frequencies in other extant and fossil populations of Microtus (Zakrzewski, 1985;
Bell and Repenning, 1999;
Bell and Bever, 2006).
The relative size of the first and second triangles exhibits some variation in the Fiene sample. The two triangles appear to be of equal size in the majority of cases, but in 10 specimens (11.5%) triangle one appears to be significantly smaller than triangle two. These specimens appear to be more robust than others in the sample. Four specimens (4.6%) display a disrupted enamel pattern on the leading edge of triangle two (Figure 5.6).
Bell and Repenning (1999) stated that this type of abnormality is caused by a deformity of the enamel organ. They provide a limited summary of abnormal occurrences in their table 2.
The enamel of the occlusal surface exhibits positive differentiation. Cementum is present in the re-entrant angles except for buccal re-entrant angle three where cementum may be absent (n = 32, 36.8%), present in trace amounts (n = 23, 26.4%), or well developed (n = 15, 17.2%). The remaining M2s (n = 17, 19.5%) exhibit an intermediate stage of cementum development in buccal re-entrant angle three.
Comments. The M2s cannot be associated with any of the m1 morphotypes. Hopefully the discovery of associated materials or the application of new morphometric techniques (e.g.,
Polly and Head, 2004) will help to rectify this situation in the future.
The observed differences between the Fiene and Cudahy M2s may be due to the smaller sample size in the former (87 vs. 1,170). It does seem of some interest that the Fiene sample has a higher percentage than the Cudahy of M2s with a posterior dentine field (28.7% vs. 4.4%), including one specimen (1.1% vs 0%) with a closed field and more specimens (4.6% vs. 0.9%) with a disrupted enamel pattern. A disrupted enamel pattern can occur on any triangle (Bell and Repenning, 1999), however, all the disruptions in the Fiene M2s are on the leading edge of T2.