OUP user menu

Elemental Analysis of Soricine Enamel: Pigmentation Variation and Distribution in Molars of Blarina brevicauda

Suzanne G. Strait, Staci C. Smith
DOI: http://dx.doi.org/10.1644/05-MAMM-A-265R4.1 700-705 First published online: 24 August 2006


Soricine shrews are typified by the presence of reddish pigment in their tooth enamel. This coloration is caused by iron and is thought to make the enamel more resistant to wear. Ten specimens of Blarina brevicauda were examined for percentage iron concentration using a scanning electron microscope in conjunction with an energy-dispersive spectrometer. Substantial iron variation was found between individual molar cusps and between different tooth positions. Specifically, the cusps associated with crushing and grinding, as opposed to shearing, had more iron incorporated into their enamel. In addition, the enlarged ml exhibits higher iron density than the more-posterior teeth. These results suggest that increased iron density is expressed on teeth, and parts of teeth, that are subject to greatest stresses and are most prone to fracture and excessive wear.

Key words
  • Blarina
  • energy-dispersive spectrometer
  • iron pigmentation
  • molars
  • shrews
  • Soricinae
  • teeth

Soricids have the widest geographic distribution of any extant insectivoran clade and are represented by 20 living genera and more than 260 species (Churchfield, 1990). A synapomorphy of the subfamily Soricinae is dental enamel that is reddish brown in color. The degree and distribution of this pigmentation has been used in soricine lower-level taxonomic analyses (Dannelid 1994). As with the pigmentation of rodent incisors, coloration in shrew teeth is caused by iron (Akersten et al. 2001, 2002; W. A. Akersten, in litt.; DÖtsch and Koeningswald 1978; Kozawa et al. 1988a, 1988b; Lunt and Noble 1975; Miles 1963; Soderlund et al. 1992). However, unlike in rodents, shrew tooth pigmentation is not limited to the incisors and is found throughout the toothrow. Even though pigmentation in shrews is limited to the superficial aprismatic enamel layer (DÖtsch and Koenigswald 1978; Koenigswald 1997), the pigmented layer in shrews is reported to be substantially thicker than in rodents (Kozawa et al. 1988a, 1988b). Although pigmented enamel is generally considered to increase wear resistance of teeth, there are no experimental data to support this assumption. Here we evaluate the hypothesis that iron pigmentation is an adaptation to increase the resistance of teeth to abrasive wear (by contact with either food or exogenous grit) by examining 2 predictions that are based on the biomechanics of chewing and principles of tooth wear.

Dental pigmentation in mammals is thought to make the teeth harder or more wear resistant (Adamczewska-Andrzejewska 1966; Churchfield 1990; W. A. Akersten, in litt.), Iron increases the acid resistance of rodent incisors (Halse 1974; Selvig and Halse 1975; Stein and Boyle 1959) and is thought to make enamel “harder” (Selvig and Halse 1975). However, only 1 study has quantified and compared the physical properties of pigmented and unpigmented shrew enamel. Soderlund et al. (1992) compared the indentation hardness of pigmented and unpigmented enamel on shrew incisors, and found no evidence that pigmented enamel is harder. However, the relationship between wear and material properties is complex (Strait 1997), and indentation hardness may not be the most appropriate value to compare. Soderlund and colleagues (1992) correctly noted that a more appropriate test for evaluating wear resistance is fracture toughness, a property that has yet to be examined and is more complex to study.

Even though studies have yet to demonstrate why pigmented enamel is more wear resistant, observations of wear pattems on soricine teeth indicate that pigmented enamel is more wear resistant than unpigmented enamel. The slower wear rate of soricine pigmented enamel creates sharp cutting edges where it contacts the unpigmented enamel (DÖtsch and Koenigswald 1978; Vogel 1984). Rodent incisors also demonstrate differential wear, with the pigmented enamel wearing at a slower rate than the underlying unpigmented enamel, creating a sharp chisellike tooth. Resistance to excessive tooth wear is especially important considering some of the behavioral and physiological specializations of soricines. First, the deciduous teeth of shrews are shed in utero and animals are bom with a full set of permanent teeth (Churchfield 1990; Kindahl 1959). Thus, this single set of teeth must be utilized during their entire life span. Experimental studies on mammals have shown that individuals with extreme tooth wear show increased food particle size after mastication, thereby decreasing the surface area that digestive enzymes have to act upon (Lanyon and Sanson 1986; Sheine and Kay 1977). Consequently, older mammals with very worn teeth must increase their food intake to maintain adequate nutrition (Lanyon and Sanson 1986). This is especially relevant for shrews because field studies have demonstrated that shrews have voracious appetites (Churchfield 1994), consuming 80–125% of their body weight in food daily (Churchfield 1990). Vogel (1980) suggested that pigmented enamel is an adaptation to the high metabolic requirements of this group. This is supported by the fact that red–toothed shrews (soricines) have higher metabolic rates than white–toothed shrews (crocidurines— Genoud 1988; Vogel 1980). Other specializations of the masticatory apparatus of soricines also support the contention that pigmented enamel is an adaptation to immense dietary pressures. For example, soricines exhibit higher chewing rates, more efficient comminuation of food, and the structures of the jaw joint, masticatory muscles, and tooth occlusal surfaces permit a greater variety of jaw movements than is possible in crocidurines (DÖtsch 1994).

Although researchers have analyzed the histological stmcture and development of pigmented enamel (Akersten et al. 2001, 2002; Koenigswald 1997; Kozawa et al. 1988a, 1988b; W. A. Akersten, in litt.), few studies have evaluated its functional implications. This study was designed to test 2 predictions. The 1st is that iron density will vary on different parts of individual molars. Specifically, cusps that are subject to the most abrasion will have the greatest concentration of iron. Microwear studies have shown that crushing and grinding facets experience more abrasive wear (higher feature density and more pitting) than shearing facets (Gordon 1982, 1984). Therefore, we predict that cusps associated with crushing and grinding (hypoconid) will have more iron than cusps associated with shearing (i.e., metaconid, paraconid, protoconid, and entoconid).

The 2nd prediction is that iron density will vary along the toothrow. A study of Sorex araneus by DÖtsch and Koeningswald (1978:67, figure 2) showed that iron content decreased from ml to m3. Here we investigate whether this pattem of decreasing molar iron density posteriorly is evident in the molars of another shrew, Blarina brevicauda.

Materials and Methods

Study animals.—Relatively unworn lower molars on 10 specimens of B. brevicauda from the Marshall University Vertebrate Museum (MU) were examined for percentage iron. B. brevicauda is the most– fossorial North American shrew and its diet includes large quantities of earthworms, millipedes, snails, slugs, and insects (George et al. 1986; Hamilton 1930; Linzey and Linzey 1973; Mumford and Whitaker 1982; Whitaker and Ferraro 1963; Whitaker and French 1984).

Microscopy.—iron density was quantified on surface enamel with a scanning electron microscope (5310LV; JOEL, Peabody, Massachusetts) in conjunction with energy-dispersive X-ray analysis (energy-dispersive spectrometer with ISIS 300; Oxford Instruments, Concord, Massachusetts). As this method was applied here, an electron beam scanned the tooth surface and various signals were emitted as the beam interacted with the enamel. In this study, the relevant signals were X-rays. The basic principle of this method is that when excited, different elements can emit X-rays of characteristic energy based on their electronic structure. These characteristic X-rays are collected by a solid-state detector and their energy is dispersed and measured. These data are then binned into channels, which when viewed together form an energy spectrum.

Energy-dispersive spectrometry is a nondestructive technique, which is important when examining material from museum collections. Lower jaws were simply adhered to scanning electron microscope stubs with double-sided carbon tape; no other specimen preparation was necessary. All measurements were made at 15-kV, 48-μA load current (an approximation of beam current at the sample), and at a working distance of 20 mm. The relative amount of iron present was recorded for each molar cusp (shearing facets for the protoconid, metaconid, paraconid, and entoconid and a crushing and grinding facet for the hypoconid) for ml-3 (Fig. 1). Dehydrated enamel is more than 99% hydroxyapatite (Ca10(PO4)6(OH)2Osborn 1981). The energy-dispersive spectrometer detector used in this study is ill–suited for detecting light elements (i.e., lighter than sodium), and thus we ignored the atomic contributions of hydrogen and oxygen and restricted the analysis to iron, calcium, and phosphorus. Furthermore, rather than absolute atomic percentages, these data represent the percentage of iron, relative to the calcium phosphorus in hydroxyapatite. Because of the variation in absolute cusp size, magnification varied from 200× to l,000× depending on the enlargement necessary to fill the frame with the cusp tip.

Fig. 1

Lower jaw of Blarina brevicauda in a) lateral and b) occlusal view illustrating major cusps, the distribution of pigmentation, and the relative sizes of molars examined in this study (redrawn from Merriam 1895).

Because analysis of density can be affected by slight variations in specimen position and tilt, an error study was performed to quantify the amount of the intraobservational variation in data collection. For 2 randomly chosen specimens (MU 7570 and MU 7557), each cusp on ml-3 was quantified 5 times each (on nonconsecutive days) by the same worker (SCS).

Statistical analyses.—To test the prediction that there is variation within a single tooth at different cusp locations, 1-way analysis of variance (ANOVA; or a nonparametric Kruskal-Wallis test) was used to compare among cusps at each tooth position. Subsequently, pairwise multiple comparison procedures (Holm-Sidak method or Student-Newman-Keuls method) were used to identify pattems of similarities and differences in iron content among cusp positions. The same statistical procedures were used to compare individual cusps across different molar positions. Overall comparisons between m1, m2, and m3 (ANOVA or Kruskal-Wallis test) were followed by pairwise comparisons (Holm-Sidak method or Student-Newman-Keuls tests). All statistical analyses where calculated with SigmaStat 3.0 (Systat Software Inc., Richmond, California).


Error study.—The results from the error study indicate that, although variation does exist, it is minimal and should not detract from other conclusions of this study (Table 1). As the example in Fig. 2 illustrates, variation within cusp positions is slight relative to intercusp variation.

View this table:
Table 1

Error study of percentage iron concentration. Includes means (and SEs) from 5 samplings of 2 specimens of Blarina brevicauda. Also included are calcium and phosphoms concentrations for comparison. Cusps sampled included protoconid (proto), paraconid (para), metaconid (meta), hypoconid (hypo), and entoconid (ento).

Specimen number and elementm1m2m3
MU 7557
MU 7470
Fig. 2

An example of the precision of measurement of iron concentration based on 5 repeated measurements of ml of 1 specimen (MU 7557) of Blarina brevicauda.

Comparison of iron concentration in individual cusps.— Mean percentage of iron incorporated into the hypoconid, entoconid, protoconid, metaconid, and paraconid of B. brevicauda varied in ml, m2, and m3 (Fig. 3). On ml there was substantial variation in iron content among cusps (Kruskal-Wallis, H = 36.59, df = 4, P < 0.001). Specifically, in all 10 samples the hypoconid contained substantially more iron than any of the other cusps (Student-Newman-Keuls, P < 0.05; Table 2). Additionally, there was a general pattem of decreasing iron incorporation from the entoconid, protoconid, and paraconid, to the metaconid. With the exception of the paraconid and metaconid comparison, all the other pairwise comparisons between cusps showed significant differences (Student-Newman-Keuls, P < 0.05).

Fig. 3

Percentage iron concentration (mean and SE) in cusps of m1, m2, and m3 of Blarina brevicauda.

View this table:
Table 2

Percentage iron in lower 1st through 3rd molars on each cusp (cusps as in Table 1) for 10 specimens of Blarina brevicauda.

Specimen numberm1m2m3
MU 75573.551.271.425.683.771.492.320.418.862.760.190.300.164.222.16
MU 75632.601.350.526.684.822.590.810.284.802.241.840.450.161.981.59
MU 75674.631.382.5212.295.512.270.971.115.093.781.130.820.943.081.90
MU 75683.732.723.6811.265.821.891.612.555.683.621.110.531.892.121.57
MU 75693.500.470.286.364.682.610.280.164.983.711.740.170.092.662.09
MU 75704.352.884.655.905.380.290.821.652.712.651.130.970.470.471.45
MU 75712.221.862.144.962.981.750.981.364.271.751.030.230.651.561.23
MU 75722.582.511.884.772.891.431.641.052.781.810.780.660.631.220.87
MU 75734.332.883.679.474.572.781.372.133.803.231.630.731.752.091.81
MU 75753.241.951.847.664.862.050.661.284.273.531.860.410.772.342.27
X̄ (SE)3.471.942.557.644.531.901.301.194.752.901.240.660.752.181.70

Second molars demonstrate a similar pattem of intra-tooth cusp variation (F = 21.61, df = 4, 45, P < 0.001). Here too, the hypoconid had greater iron density than the other cusps (Holm-Sidak method, P < 0.05). Distinct variation was again notable among the other cusps with decreasing iron concentration from the entoconid, protoconid, and paraconid, to the metaconid. However, these differences were not always significant; only the iron content of the entoconid was statistically distinct from that of the paraconid, metaconid, and protoconid (Holm-Sidak method, P < 0.05).

Significant differences also were discovered between cusps on m3s (F = 11.37, df = 4, 45, P < 0.001). Again, the hypoconid had more iron than the other cusps and these differences were statistically evident in comparisons with the protoconid, metaconid, and paraconid, but not with the entoconid (Holm-Sidak method, P < 0.05). The m3s demonstrated the same pattern of decreasing iron concentration seen in m1s and m2s from the entoconid, protoconid, and paraconid, to metaconid. Among these other pairwise comparisons between cusps only the entoconid and paraconid, and entoconid and metaconid pairs could be distinguished statistically (Holm-Sidak method, P < 0.05).

Comparison of iron concentration along the toothrow.— Comparison of iron concentration across m1–3 protoconids demonstrated substantial variation among tooth positions (F = 26.04, df = 2, 27, P < 0.001; Fig. 3). Specifically, the m1 had relatively more iron, on average, than either the m2 or m3 and m2 contained more iron than m3 (Holm-Sidak method, P < 0.05). The percentage of iron on the metaconid from ml to m3 also was highly variable (F = 6.04, d.f. = 2, 27, P = 0.007). The ml metaconid contained more iron than this cusp did on either m2 or m3 (Holm-Sidak method, P < 0.05). Paraconid iron concentration along the toothrow also varied significantly (Kruskal&-Wallis, H = 15.10, df = 2, P < 0.001). As with the other trigonid cusps, the ml paraconid has more iron than the more posterior molars, but additionally, the m2 paraconid had relatively more iron than the m3 (Student-Newman-Keuls method, P < 0.05). Hypoconid iron concentration variation from ml to m3 also is substantial (Kruskal-Wallis, H = 21.20, d.f. = 2, P < 0.001). Again, the ml has relatively more iron than m2 and m3, and m2 has more than m3 (Student-Newman-Keuls method, P < 0.05). The 5th cusp sampled, the entoconid, demonstrated the exact same pattern of variation within the toothrow (F = 33.17, df = 2, 27, P < 0.001); the ml entoconid had more iron than either m2 or m3 and m2 had more than did m3 (Holm-Sidak method, P < 0.05).


Incorporation of some iron oxide compounds into dental tissues has been observed in fish, salamanders, and snakes (Akersten et al. 2002), but few mammalian groups have developed this adaptation. Rather, other adaptations for eating hard foods or foraging in abrasive environments are more common in mammals. For example, many ungulate taxa have uniquely high-crowned teeth and hard-object-feeding primates and bats thicken their enamel (e.g., Dumont 1995; Jams 1988; Janis and Fortelius 1988; Kay 1981). Among mammals, iron incorporation only occurs in shrews, rodents, and members of 2 extinct clades (Multituberculata and Aptemodontidae). Pigmented enamel evolved independently in these diverse lineages, and at least 2 main functions can be discerned in the modem taxa. One function is to prolong the life of the tooth, and therefore the organism, as in soricines. Pigmentation may be a primitive characteristic of soricids and is preserved in forms such as Domnina gradata from the early Oligocene of North America (Repenning 1967). A 2nd function of pigmented enamel is the primary development of cutting edges, as is the case of rodent incisors. Unlike shrews, pigmented enamel in rodents is found only on the labial surface of incisors. Rodents have evergrowing incisors so that food or foraging related abrasion is not life-limiting. Instead the primary function of the pigmentation is to harden the enamel and maximize the differential wear between the enamel and the dentine to produce a sharp chisellike tooth. The early Paleogene taenio-labidoid multituberculate Lambdopsalis bulla had a band of pigmented enamel on the ventrolateral surface of il and the crowns of m2 (Desui 1986). Multituberculates were morphologically, and probably ecologically, similar to rodents (Krause 1986), and perhaps iron concentrations in the enamel of Lambdopsalis functioned to increase cutting surfaces as in rodents. Pigmentation on cusp apices throughout the dentition, like that seen in sorcines, is only known for several species of Eocene insectivores belonging to the genus Apternodus (Asher et al. 2002). Aptemodontids were very small insectivorous mammals with exceptionally high-crowned teeth and it is possible that, like soricines, their pigmented enamel functioned to resist abrasion. Although pigmentation in dental tissues has evolved independently in several clades, only red-toothed shrews have demonstrated throughout their history an abrasion-resistant toothrow adapted for their unique physiological demands.

This study demonstrated that the teeth of B. brevicauda, and by inference other soricines, are modified by the incorporating of iron into their enamel in a functionally significant pattern. As predicted, the hypoconids consistently demonstrated higher concentrations of iron than any other cusp. This is bio-mechanically relevant because the hypoconid includes the primary cmshing and grinding surfaces and therefore is more prone to excessive wear than the shearing surfaces. Furthermore, a pattem of relative iron concentration reveals a ranked pattem on m1–3 when the other cusps also were explored. In all tooth positions the entoconid always had the most iron, followed then by the protoconid, paraconid, and metaconid. Even though these differences are not always statistically significant, this pattem is intriguing and warrants further study. Iron content is not a correlate of absolute cusp size; in B. brevicauda the protoconid is substantially larger than the entoconid. Rather, iron content is highest in the talonid cusps (hypoconid and entoconid), which are subject to greater abrasive forces (crushing and grinding) than those of the trigonid.

Comparisons across the molar row showed that m1 had greater iron density than did the more-posterior molars. Additionally, each m2 cusp contained more iron than did the homologous cusps on the smaller m3. This pattern had been previously reported in S. araneus (DÖtsch and Koeningswald 1978). We suggest 2 possible functional explanations for this pattem. The 1st is that bite force in shrews may be greater at ml than m3, and therefore increased strength in the anterior molars is associated with greater stresses in those teeth. Although bite force at each molar position has not been measured in shrews, there is both theoretical and empirical evidence from nonprimate mammals that bite force is either constant across the molar toothrow, or is greatest at the anterior molars and then decreases posteriorly (Greaves 1978; Spencer 1999; Thompson et al. 2003).

A 2nd potential functional explanation concerns the relative size of the molars in shrews. In many mammals molar size increases posteriorly; however, the inverse relationship is evident in shrews. This m1–3 size relationship is possibly an adaptation for relatively large prey size for an individual. It would be difficult for a shrew to begin mastication more posteriorly in their mouth, because of large prey size (e.g., earthworms) compared to their absolute gape. Additionally, because of the enlargement of m1, compared to the more-posterior molars, this tooth is the 1st masticatory tooth to contact foods during the initial stages of mastication. If the more-anterior molars are more apt to 1st contact foods during ingestion and mastication they also would be more prone to fracture and excessive wear. In this situation, increased iron concentration on ml could represent an adaptation for increased strength and resistance to abrasion.

Further examination of other soricines will determine whether these pattems of iron incorporation in cusps and tooth positions we identified in Blarina exist in other taxa. Additionally, there is enormous variation in the degree of pigmentation between soricine species (Dannelid 1994; Repenning 1967), and detailed quantitative interspecific studies of iron distribution and density could add new functional insights to these pattems of interspecific differences. It has been suggested that the variation in pigmentation among soricine species could reflect adaptations to diets with different physical properties. For example, Dannelid (1994) suggested that the lighter pigmentation in some sorcines could be linked to a diet of less-sclerotized food. However, it also might be important to consider the selective impact of abrasion caused by grit (soil) that is associated with common prey items such as earthworms (Akersten et al. 2001; Silcox and Teaford 2002). Some shrews dig and burrow and this behavior also may introduce soil and grit into the oral cavity causing selection for more-resistant dental tissues (Dannelid 1994). Finally, some species forage in abrasive environments where exogenous soil also may be ingested.


We thank National Aeronautics and Space Administration and Department of Biological Sciences, Marshall University, for the scholarship and undergraduate tuition wavier that SCS received during this study. The manuscript was improved by the comments of B. Dumont, D. Neff, and 2 anonymous reviewers. We also thank the National Science Foundation (Course, Curriculum, and Laboratory Improvement Program) and Marshall University for support of the scanning electron microscope facility and Farrah Boggess for microscope assistance. M. E. Hight also is thanked for the loan of the specimens from the Marshall University Vertebrate Museum.


  • Associate Editor was Nancy G. Solomon.

Literature Cited

View Abstract