OUP user menu

Variation and Integration of the Simple Mandibular Postcanine Dentition in Two Species of Phocid Seal

Edward H. Miller, Ha-Cheol Sung, Valerie D. Moulton, Gary W. Miller, J. Kerry Finley, Garry B. Stenson
DOI: http://dx.doi.org/10.1644/06-MAMM-A-243R.1 1325-1334 First published online: 18 October 2007


Pinnipeds generally swallow prey whole, and most have simple, homodont, nonoccluding cheek teeth. We investigated whether cheek teeth in seals are more variable and weakly integrated than in terrestrial Carnivora. We measured mandibular length and crown length of mandibular postcanines (PCs) in ringed seals (Pusa hispida; n = 912) from the Canadian Arctic, and harp seals (Pagophilus groenlandicus; n = 636) from Newfoundland and Labrador. PC size was uncorrected or only weakly correlated with adult mandibular length. PC length and mandibular length were strongly bilaterally symmetrical (r ≥ 0.8 between left and right sides). PC size was moderately variable (coefficients of variation [CVs] ∼ 7–10%), and CVvaried with position in the toothrow. Adjacent PCs were correlated more strongly in size (to r > 0.8) than PCs more distant from one another. In summary, PC size in ringed and harp seals was slightly more variable than cheek teeth in complex dentitions of fissipeds, and the 2 seals were similar to fissipeds in strong bilateral symmetry in mandibular and PC size, patterned variation along the toothrow, and correlated size between adjacent PCs.

Key words
  • bilateral symmetry
  • Canadian Arctic
  • dental variation
  • dentition
  • harp seal
  • Labrador
  • morphometrics
  • Newfoundland
  • ringed seal

Functional, developmental, and evolutionary interdependence of teeth within mammalian dentition has long been appreciated (Butler 1937, 1939). Tight integration serves diverse functions, including food procurement, mastication, tooth sharpening, or social activities such as fighting or display (Evans and Sanson 2003, 2006; Every et al. 1998; Ewer 1973). Integration is reflected in specific movements used in mastication or tooth sharpening, and is shown morphologically in dental orientation, surface features, wear patterns, or size (Crompton and Hiiemae 1970; Evans 2005; Every et al. 1998; Popowics 2003). For example, teeth that interact in occlusion or mastication, neighboring teeth, and corresponding teeth on left and right sides often vary little and are morphologically complementary and similar in size (Gingerich and Schoeninger 1979; Gingerich and Winkler 1979; Kurten 1953, 1967: Pengilly 1984; Prevosti and Lamas 2006). Teeth that are not integrated within the dentition for food processing may be variably present or variable in morphology or size; examples are the anteriormost premolars of some ursids and m2 (lowercase letters signify lower teeth) of Eurasian lynx (Lynx lynxKurtén 1953, 1963, 1964; Lanyon and Sanson 2006; Rui and Drehmer 2004; however, nonoccluding teeth occasionally are positively correlated in size—Szuma 2000). Apart from functional requirements operated on by natural selection, intraspecific morphological variability in teeth also may reflect tooth-specific levels of variability or underlying developmental processes (Evans and Sanson 2003; Gingerich 1974; Jernvall 2000; Kangas et al. 2004; Pengilly 1984; Polly 1998a).

Pinnipeds evolved from carnivores with complex dentition, and the postcanine (hereafter, PC) dentition of pinnipeds became secondarily simplified in several ways (as in aquatic mammals generally), because little preparation of food takes place orally aside from holding and puncturing (the widespread and generalized phases of shearing and grinding do not occur-Gingerich 1973): the PC dentition is homodont; PCs do not occlude; and PCs are anatomically simple in most species, often being just more-or-less pointed for grasping active slippery prey (Adam and Berta 2002; Chapskii 1955; Eastman and Coalson 1974; Frechkop 1955; Howell 1930; King 1983; Wang et al. 2005). The mandibular PCs of phocids represent p1 to p4 plus m1 (Eastman and Coalson 1974; Meyer and Matzke 2004; Stewart and Stewart 1987b; Stewart et al. 1998; Weber 1928). Secondarily simplified structures often are phenotypicaily variable (e.g., in presence, morphology, or size—Dayan et al. 2002; Fong et al. 1995; West-Eberhard 2003; Yablokov 1974) and exhibit weakened morphological integration (Gould and Garwood 1969; Tague 1997). Scattered observations suggest that these characteristics apply to the dentition of pinnipeds; for example, supernumerary PCs are fairly common (Odobenidae—Fay 1982; Otariidae—Braunn and Ferigolo 2004; Drehmer et al. 2004; Drehmer and Ferigolo 1996; Tedman 2003; Phocidae—Eastman and Coalson 1974; Könemann and Van Bree 1997; Stewart and Stewart 1987a; Suzuki et al. 1990). Variation in number of PCs is highest in northern elephant seals (Mirounga angustirostris, >30% of specimens are bilaterally asymmetric—Briggs 1974), and bearded seals (Erignathus barbatus—Chapskii 1955; Manning 1974). However, few supernumerary teeth have been reported in the complex specialized filter-feeding PC dentition of crabeater seals (Lobodon carcinophagusAdam 2005; Eastman and Coalson 1974). Fine-scale dental variation also has been detailed for several pinniped species (Briggs 1974; Briggs and Morejohn 1976; Burns and Fay 1970; Chapskii 1955; Jernvall 2000; Scheffer 1960; Scheffer and Kraus 1964).

Mensural dental variation and the correlation structure of tooth size within the dentition have been investigated in detail for a number of fissipeds and other mammal species (Polly 1998b; Meiri et al. 2005; Szuma 2000). We use those studies as a basis for comparison, while acknowledging fissiped paraphyly and the superficial nature of many differences between pinnipeds and their terrestrial relatives (Bininda-Emonds and Gittleman 2000; Bininda-Emonds et al. 2001). In this paper we present the 1st analysis of mensural variation in tooth size for pinnipeds, based on collections of lower jaws from animals of known sex and age for 2 species of Phocidae that fall within the “pierce-feeding” marine mammal guild recognized by Adam and Berta (2002) and Deméré and Berta (2005): ringed seals (Pusa hispida) and harp seals (Pagophilus groenlandicus). Members of the pierce-feeding guild use a piercing bite, with prey captured in the mouth and held in place by small sharp teeth. In pinnipeds, this guild is characterized morphologically by nonoccluding upper and lower cheek teeth with lack of occlusal wear facets, m1 approximately midway along dentary, and homodonty with morphologically similar premolars and molars (Adams and Berta 2002). Many prey of these species are small and are swallowed whole; for example, pelagic amphipods and other crustaceans (particularly by young seals and adults in parts of the range, e.g., the offshore population of Canadian Arctic ringed seals—Finley et al. 1983) and small fish such as Arctic cod (Boreogadus saida), capelin (Mallotus villosus), polar cod (Arctogadus glacialis), and sand lance (Ammodytes species—Chapskii 1955; Frost and Lowry 1981; Holst et al. 2001; McLaren 1958; Svetocheva 2004; Vikingsson and Kapel 2000; Wallace and Lawson 1997). However, both species take large prey in some parts of the range or at some seasons (Lowry et al. 1980; Wallace and Lawson 1997), and these must require use of the dentition for gripping, subduing, or dismembering. Some interspecific differences in size and complexity of PCs reflect gross dietary differences, but the trophic significance of most interspecific variation is unknown (Chapskii 1955; Fig. 1).

Fig. 1

Mandibular postcanines of phocid seals vary in size and complexity: mandibles and lower teeth of some Northern Hemisphere phocids (right lateral view). The scale bar is based on a mean mandibular length (as defined in this paper) of 130 mm, for harp seals (Pagophilus groenlandicus) ≥8 years of age (after Chapskii 1955:162, figure 1).

We hypothesized that lower jaws would be bilaterally symmetrical in size for general functional purposes (Chapskii 1955), although less so than in fissipeds, and that size of the relatively simple mandibular PCs would be fairly variable and with weaker levels of intercorrelation than the more morphologically complex, functionally integrated PCs of many fissipeds and other mammals.

Materials and Methods

Ringed seals.—Lower jaws (mandibles) were purchased from Inuit hunters in Grise Fiord, Pond Inlet, and Clyde River, Nunavut, Canada, from June 1978 to September 1980, by LGL Ltd., Environmental Research Associates. Specimens were frozen until processing. Jaws were boiled for about 1 h until the canines could be extracted easily by hand, and then were refrozen. After thawing, jaws were soaked in tap water at room temperature for a day or so, until the central PC (i.e., PC3) could be removed by hand from left and right sides. Jaws then were cleaned of tissue and dried at room temperature. Jaws were measured approximately 1 year later. Only PC3 was available for study.

The technique of ageing ring seals by counting dentinal annuli was described by McLaren (1958) and Smith (1973). Thin (0.3-mm) cross sections of both lower canines were cut on a rotating saw approximately one-third of the distance from the root end to the crown. Extracted teeth and the thin sections were stored in a mixture of equal parts of tap water, absolute ethanol, and 10% glycerin. Tooth sections were examined under transmitted light with a dissecting microscope (magnification 20–30 ×). Each section was read in blind replicates until 2 identical readings, or a maximum of 3 readings, were completed. When no 2 readings were identical, the mean of the 3 readings was calculated and rounded to the nearest integer. Seals were aged relative to their month of birth, which was assumed to be April in the study area. This technique slightly underestimates age up to about 15 years of age; underestimates can be substantial for seals of that age or older (Stewart et al. 1996). Therefore, data on seals ≥15 years of age were combined.

Information on body size was not available, so we used size of the lower jaw as a proxy for body size. Greatest length (or “size”) of the lower jaw (mandible) was measured to 0.1 mm with calipers, from the anteriormost point on the jaw to the posteriormost point on the articular condyle. Length of toothrow was measured (to 0.1 mm) from the anterior margin of the PC1 alveolus to the posterior margin of the PC5 alveolus. Mesiodistal extent (“length” or “size”) of PC3 was measured in lingual aspect to 0.124 mm, using a dissecting microscope with an ocular micrometer (measurement precision reflects the micrometer's scale). Measurement error was assessed by measuring variables 10 times for 10 seals, blind and in a random sequence over several days. The coefficient of variation (CV) for these repeated measurements was very small (<0.5% for jaw length and ∼0% for PC length), so we ignored measurement error in statistical analyses (Bailey and Byrnes 1990; Polly 1998b). For analyses, we used mean values of left ami right jaws and teeth for complete specimens, unless indicated otherwise.

Most specimens were suitable for measurement, although many had been damaged by shooting. The final sample contained 520 males and 392 females. Males ranged up to 23 years in age, and females to 20 years, though the sample of males had a slightly lower median age (5.38 versus 6.23 years, respectively).

Harp seals.—Seals were shot with high-powered rifles by professional sealers or by employees of the Department of Fisheries and Oceans around the island of Newfoundland in September 1994, November 1994–July 1995, and November 1995–June 1996. Lower jaws were removed and frozen until processing. After thawing, jaws were boiled for about 1 h, and a canine was removed for ageing, then jaws were cleaned of tissue and dried at room temperature. They were measured about 1 year later. Variables and measurement protocols followed those for ringed seals, except all mandibular PCs (i.e., PC1–PC5) were available for measurement, and teeth were not removed from jaws for measurement. We removed and remeasured 10 teeth 3 times to determine whether the different measurement procedure affected our estimates; measurements were identical (all differences < 1%). This results from the wide spacing of the cheek teeth of seals (unlike terrestrial Carnivora—Dayan et al. 2002), which permits ready measurement of teeth. Some of these specimens were used by Miller et al. (1998) and Miller and Burton (2001).

As for ringed seals, many specimens had been damaged by shooting, so data were not complete for all specimens. The final sample contained 323 males and 313 females. Males ranged up to 28 years in age and females to 31 years; age distributions were similar between the sexes and were dominated by young age classes (median ages 2.80 versus 3.05 years, respectively). Sampling of both species accorded with guidelines of the American Society of Mammalogists (Animal Care and Use Committee 1998).


Tooth size in relation to position, sex, age, and jaw size.— The PCs of Pagophilus differed in size according to position in the toothrow, with PC1 being the smallest, PC2 about 40% larger, and PC3–PC5 about 7% larger again (Table 1; Fig. 2). For seals ≥ 8 years old, males were ∼2–3% larger than females in jaw size in both species, and were relatively larger in PC size: ∼8% in male Pusa and 5% in male Pagophilus (Table 1). When age classes were pooled, PC3 was ∼8% larger for male Pusa and ∼3% larger for male Pagophilus. Other PCs of Pagophilus differed less between the sexes: PC2 and PC4 were ∼2% larger in males, and PC1 plus PC5 were only ∼1% larger in males.

Fig. 2

Size (mesiodistal length) of mandibular postcanines varied with position in the toothrow and between sexes in harp seals (Pagophilus groenlandicus) from Newfoundland and Labrador. Mean ± 95% confidence intervals (= ±1.96 SE) are shown.

View this table:
Table 1

Size (mesiodistal length in mm) of mandibular postcanines (PCs) differed between sexes in ringed seals (Pusa hispida) from the Canadian Arctic, and between sexes and with position in the toothrow in harp seals (Pagophilus groenlandicus) from Newfoundland and Labrador. Mean ± SD are shown with sample size in parentheses. See Fig. 2.

Species and sexPC1PC2PC3aPC4PC5
Ringed seal
Males5.26 ± 0.52 (509)
Females4.87 ± 0.45 (361)
Harp seal
Males4.75 ± 0.40(311)7.12 ± 0.51 (317)7.74 ± 0.50 (319)7.67 ± 0.49 (320)7.80 ± 0.58 (313)
Females4.69 ± 0.38 (304)6.97 ± 0.48 (306)7.55 ± 0.53 (306)7.51 ± 0.53 (303)7.70 ± 0.61 (301)
  • a For seals ≥8 years old, jaw and PC3 length means were: male Pusa, 103.8 ± 5.1 mm (103) and 5.21 ± 0.52 mm (167); female Pusa, 102.3 ± 4.9 mm (97) and 4.84 ± 0.44 mm (161); male Pagophilus, 133.3 ± 4.4 mm (19) and 7.80 ± 0.49 mm (39); and female Pagophilus, 128.8 ± 4.0 mm (29) and 7.43 ± 0.60 mm (65).

Jaw length increased in size up to about 8 years of age in both species; toothrow length reached an asymptote slightly earlier (E. H. Miller, in litt.). In contrast, PC length did not vary with age in either sex of either species (slopes in simple linear regression ranged from positive to negative, all were statistically inseparable from 0, and all had extremely low r2 values [maximum, 3%]). All estimates of r between jaw length and PC length were slightly positive but only 1 was significant (for female Pusa ≥ 8 years old; r2 = 12%) after a was corrected for multiple tests (Table 2). Several other relationships (PC2, PC3, and PC5 of male Pagophilus ≥ 8 years old; r2 = 20–22%) were significant at an uncorrected α = 0.05; however, these 3 relationships were based on the same sample of males so were not independent of one another. In summary, PC size was unrelated to or correlated with mandibular size only very weakly within species.

View this table:
Table 2

Size (mesiodistal length in mm) of mandibular postcanines did not vary with mandibular length in ringed seals (Pusa hispida) ≥8 years old from the Canadian Arctic, or harp seals (Pagophilus groenlandicus) ≥8 years old from Newfoundland and Labrador. Values for Pearson's product-moment correlation coefficient (r) are shown; P (unconnected) and sample size are given in parentheses. Significant correlations at an uncorrected significance level of α = 0.05 are in bold (following Sidak's correction to α = 0.05 for multiple tests [2 for ringed seal and 5 for each sex of harp seal]; only the value for female ringed seals was significant).

Species and sexPC1PC2PC3PC4PC5
Ringed seal
Males0.119(0.23; 102)
Females0.343 (0.001; 89)
Harp seal
Males−0.042 (0.86; 21)0.448 (0.04; 21)0.468 (0.03; 21)0.221 (0.35; 20)0.468 (0.04; 20)
Females−0.107 (0.59; 28)−0.038 (0.85; 29)0.130 (0.50; 29)0.084 (0.67; 28)0.122 (0.54; 28)

Character variation and correlations.—Left and right sides were similar in size, with bilateral differences within seals averaging ∼2–4% for teeth, ∼1% for toothrow length, and <1% for mandibular length (Table 3). High correlations between sides also were found for mandibular length (r = 0.995 for Pagophilus and r = 0.944 for Pusa) and PC size (mean r = 0.876 for Pagophilus and mean r = 0.944 for Pusa;Table 3)

View this table:
Table 3

Size (mesiodistal length in mm) of mandibular postcanines and mandibular length were similar, and were closely correlated between right and left sides in ringed seals (Pusa hispida) from the Canadian Arctic and harp seals (Pagophilus groenlandicus) from Newfoundland and Labrador. Values for percent left-right differencesa are shown as mean ± SD and (below), Pearson's product-moment correlation coefficient (r) between left- and right-side measurements (sample size is given in parentheses). All correlations were highly significant (P < 0.001).

Ringed sealHarp seal
Jaw length0.71 ± 0.550.57 ± 0.490.74 ± 0.650.79 ± 0.64
0.991 (160)0.994 (109)0.995 (50)0.995 (70)
Toothrow length1.3 ± 1.91.2 ± 1.01.2 ± 1.11.0 ± 0.91
0.983 (370)0.960 (306)0.982 (152)0.986 (172)
PC1 length4.0 ± 3.44.3 ± 3.8
0.821 (290)0.772 (286)
PC2 length2.8 ± 2.33.2 ± 2.9
0.883 (300)0.828 (299)
PC3 length2.2 ± 2.42.3 ± 2.62.6 ± 2.22.5 ± 2.4
0.950 (449)0.937 (323)0.883 (289)0.894 (292)
PC4 length2.1 ± 1.82.4 ± 2.2
0.920 (292)0.905 (285)
PC5 length2.5 ±2.12.7 ± 2.2
0.931 (282)0.918 (275)
  • a Percent difference between left (L) and right (R) sides computed per specimen as 100(L − R)/[2(L + R)].

In Pusa, CVs for length of PC3 (left side, for purposes of exposition) were 9.9% (males) and 9.5% (females). Slightly lower variation occurred across the toothrow of Pagophilus (left PC1–PC5): 7.4%, 6.6%, 8.8%, 6.8%, and 7.4% in males; and 7.4%, 7.4%, 8.4%, 7.1%, and 8.0% in females. In this species, CVs had a fairly small range (6.4–8.9% across all teeth and between the sexes); nevertheless, teeth in different positions had characteristically different levels of variation. The evidence for this is 2-fold; 1st, CVs of corresponding teeth on left and right sides within sexes were correlated (males: r = 0.988, n = 5, P < 0.002; females: r = 0.978, n = 5, P = 0.004). Second, CVs of corresponding teeth in males and females were correlated: r = 0.881 (left teeth only; N = 5, P < 0.05), 0.937 (right teeth only; n = 5, P = 0.02), and 0.906 (all teeth; n = 10, P < 0.001): PC2 varied least (between sexes, mean CV = 7.0%); PC1 and PC3 varied most (mean CVs = 8.1% and 8.2%, respectively); and PC4 and PC5 varied at intermediate levels (mean CVs = 7.1% and 7.3%, respectively).

The PC size was not correlated with mandibular size in either species (Fig. 3). In Pagophilus, tooth size was correlated most strongly between adjacent PCs, and strength of correlations weakened with the distance between teeth (Table 4; Fig. 4). In both sexes, PC3 size was correlated most strongly with size of adjacent teeth, and PC1 size was correlated most weakly with size of other teeth.

Fig. 3

Size (mesiodistal length) of the central lower postcanine tooth (PC3) was not correlated with mandibular size (length) in ringed seals (Pusa hispida) from the Canadian Arctic or harp seals (Pagophilus groenlandicus) from Newfoundland and Labrador. The 95% confidence ellipses are shown for specimens of all ages (for data on seals ≥8 years old; see Table 2).

Fig. 4

Size (mesiodistal length) of postcanine teeth was correlated most strongly between adjacent teeth, and declined with distance between teeth, in harp seals (Pagophilus groenlandicus) from Newfoundland and Labrador (see Table 4).

View this table:
Table 4

Size (mesiodistal length in mm) of mandibular postcanines (PCs) was correlated most strongly between PCs closest to one another, in both sexes of harp seals (Pagophilus groenlandicus) from Newfoundland and Labrador (see Fig. 3). Estimates for Pearson's product-moment correlation coefficient (r; using means of left and right PC measurements) are shown (sample size is given in parentheses; males above diagonal, females below diagonal). All correlations were highly significant (P < 0.0001).

PC10.557 (310)0.464 (309)0.470 (310)0.346 (305)
PC20.549 (304)0.789 (316)0.677 (316)0.479(310)
PC30.453 (303)0.758 (305)0.821 (318)0.590(311)
PC40.447 (300)0.696 (302)0.855 (303)0.632(312)
PC50.426 (298)0.488 (300)0.590 (301)0.661 (299)


Tooth size in relation to position, sex, age, and jaw size.— The PC size varied with position in the toothrow in the harp seal, with PC3 being (slightly) the largest tooth, as in the ringed seal (Belobdorov 1975; Chapskii 1955; Doutt 1942). Size ordering of PCs is ubiquitous in pinnipeds (e.g., Phocidae— Briggs and Morejohn 1976; Burns and Fay 1970; Doutt 1942; Scheffer 1960).

Sexual size differences in pinnipeds have been investigated extensively (mainly in relation to social systems), but most analyses pertain to overall body size (Alexander et al. 1979; Lindenfors et al. 2002). Yablokov and Sergeant (1963) quantified cranial sexual size differences in adult harp seals similar to levels we observed: ∼5% in mandibular length and ∼7% in distance between PC5 and anterior extreme of cranium or mandible (Yablokov and Sergeant 1963:1859–1860, tables 1 and 2). Ringed seals show extensive ecogeographic variation in size and sexual size differences (Amano et al. 2002; Fedoseev 1975; Finley et al. 1983; Hyvärinen and Nieminen 1990; McLaren 1993; Reeves 1998), so we do not discuss our limited findings further, aside from suggesting that the apparently greater sexual size difference in PC3 than mandibular length merits further study.

The PC toothrow lengthens substantially with age in phocids, because PCs are crowded in young animals and become less crowded with growth (Belkin et al. 1969; Doutt 1942; Harington and Sergeant 1972; Ooe and Esaka 1981); however, PC size does not change with age because of wear after full eruption of the crown in ringed or harp seals. Therefore, sample size can be increased in future studies by pooling samples across ages within sexes. In ringed seals, other measurement variables on PC3 decline with age, indicating wear over time (E. H. Miller, in litt.). Tooth wear, breakage, and associated problems occur in all pinniped species but with different patterns, according to diet, intraspecific strife, and other factors (Braunn and Ferigolo 2004; Chapskii 1955; Drehmer and Ferigolo 1996; Fay 1982; Stirling 1969).

Tooth size has been used widely to predict body size interspecifically, especially in paleobiology (Creighton 1980; Gingerich et al. 1982; Pan and Oxnard 2003; Van Valkenburgh 1990). Some significant intraspecific correlations between size of lower cheek teeth and cranial or mandibular size also have been reported for terrestrial Carnivora (Kurten 1953, 1967; Meiri et al. 2005). Significant relationships occur in other taxa, including rodents and primates (Dayan et al. 2002; Gould and Garwood 1969; Moyer et al. 1985; Olson and Miller 1958), but are not universal even within Carnivora (e.g., brown bear [Ursus arctos]—Kurtén 1953; jaguar [Panthera onca]—Turner and O'Regan 2002), and we detected little or no relationship of tooth size to mandibular size in our study.

Character variation.—Carnassial dentition has evolved independently several times in mammals (Butler 1946). In Carnivora, carnassial teeth are P4 (uppercase letters signify upper teeth) and ml, and these have been reported to vary little intraspecifically in species with complex dentition, presumably because of high morphological complexity related to precise integrated functions in food processing. Hence, CV estimates for mesiodistal length in carnassials provide a basis for comparison with our results. In our study, CVs for PC size were ∼9–10% in ringed seal and ∼7–9% in harp seal, very close to estimates for M2 of the polar bear (Ursus maritimus)— a reduced structure in this species that would be expected to be more variable than in other bears (CVs = 7.2–9.6%—Kurté n 1964). Reported values for carnassials in other fissipeds invariably are lower than those values. Some examples of CVs for P4 and ml (respectively) are 4.7% and 4.3–4.8% for red foxes (Vulpes vulpesSzuma 2000, 2003); 4.1% and 4.3% for Arctic foxes (Vulpes lagopusPengilly 1984); 6.4% and 5.1% for cave bears (Ursus spelaeusBaryshnikov et al. 2003; computed from weighted means of data in their tables 3 and 4); and 6.1% and 5.4% for Eurasian badgers (Meles melesBaryshnikov et al. 2002; computed for “cluster 1” in their tables 4 and 5). Other cheek teeth are similarly or slightly more variable (Kurten 1964; Wolsan et al. 1985). Similarly low levels of variation characterize complex cheek teeth in other fissipeds as well as primates (Gingerich 1974; Polly 1998b).

High variation in dental traits has been related to evolutionary factors and processes other than the weakening of stabilizing selection, including directional selection and geographic isolation of small populations (Juste and Ibáñez 1993; Tomow et al. 2006). We interpret high CVs for PC size in ringed and harp seals to indicate increased variation due to evolutionary simplification of morphology via selective release, in connection with evolution from complex specialized to simple generalized food-processing (Adam and Berta 2002; Briggs 1974; Chapskii 1955; Gould and Garwood 1969).

Character correlations.—Close functional relationships between neighboring or occluding teeth, or between corresponding teeth on left or right sides, are reflected in complementary size and morphology (Gould and Garwood 1969; Kurtén 1953, 1967; Prevosti and Lamas 2006). High bilateral symmetry is expressed jointly through similarity in size and high correlations, and we found both in our study. Similarity in size between left- and right-side measurements was strong overall, but particularly for mandibular length (left–right differences averaged <1%) and toothrow length (differences ∼1%); differences averaged ∼2–4% for PCs. Similarly, bilateral correlations were high overall but especially (r > 0.98) for mandibular length and toothrow length; r was 0.77–0.95 for PCs. In the complex dentition of pine martens (Martes martes), r between linear dental measurements on left and right sides was similarly high, averaging 0.88 (maximum, 0.92— Wolsan et al. 1985).

We predicted relatively weak size correlations among teeth, but again our estimates were fairly high (e.g., for Pagophilus, r averaged ∼0.60 with maxima > 0.82). These values are surprisingly similar to some for lower cheek teeth of fissipeds; for example, r averaged 0.57–0.65 (maxima, 0.81–0.87) for mesiodistal tooth lengths in several studies on red foxes (Szuma 2003), and 0.47 (maximum, 0.78) for the same variables in Arctic foxes (Pengilly 1984). Correlations are similar or weaker in other mammals with complex integrated dentitions (cuscuses [Ailurops ursinus]—Kurtén 1953; wild house mice [Mus musculus]—Wallace 1968; and white-footed mice [Peromyscus leucopus]—Van Valen 1962).

We conclude that evolutionary simplification of the PC dentition of ringed and harp seals was not accompanied by marked weakening of bilateral asymmetry or of size correlations among mandibular PCs. Our observations may be explained by evolutionarily conserved developmental programs for mammalian dentition (Butler 1937, 1939,1946; Kangas et al. 2004). Other observations that support this possibility for pinnipeds concern the nature of dental variation, which can be high but is bounded. For example, occasionally teeth of ribbon seals (Histriophoca fasciata) are indistinguishable from those of ringed or harp seals, or common seals (Phoca vitulinaBurns and Fay 1970). Furthermore, dental variants often are expressed bilaterally in ribbon seals and in otariids (Drehmer et al. 2004; Tedman 2003). Other observations along these lines are reported by Bateson (1894), Briggs (1974), Briggs and Morejohn (1976), Burns and Fay (1970), Chapskii (1955), Drehmer et al. (2004), Fay (1982), Könemann and Van Bree (1997), and Scheffer and Kraus (1964).

Concluding comments.—Quantitative aspects of variation and integration of pinniped dentition have not been reported previously. Compared with complex cheek teeth of most fissiped carnivores, PCs of ringed and harp seals express slightly higher variation in size, but similar levels and patterns of size correlations and bilateral symmetry. Whether the same patterns occur in other pinnipeds is an open question, because no other studies have been conducted. Yet PCs of phocids, in particular, vary greatly across species in size, morphology, and spacing (Bininda-Emonds and Russell 1996; Burns and Fay 1970; Chapskii 1955; Doutt 1942; Eastman and Coalson 1974; King 1983; Fig. 1). Cranial and masticatory anatomy of phocids varies in relation to diet (Endo et al. 1998a, 1998b, 2002; Howell 1929; King 1972; Kosygin and Shustov 1971), and geographic variation in PCs and mandibles is known (Chapskii 1955; Doutt 1942). Therefore, this group seems to provide many opportunities for morphometric studies, to complement the extensive work that has been carried out on terrestrial Carnivora.


Research on ringed seals was funded by Petro-Canada Explorations Inc. (Calgary, Alberta, Canada) as part of the Eastern Arctic Marine Environmental Studies research program. N. Snow (then of Department of Indian and Northern Affairs, Ottawa, Ontario, Canada) was the driving force behind this program; R. Davis and B. Koski managed the project on behalf of LGL Ltd., Environmental Research Associates. G. Glazier, G. Koening, W. Speller, and B. Veldhoen of Petro-Canada gave invaluable assistance and support to the work. The study could not have been accomplished without the assistance of Pond Inlet hunters, particularly P. Aglak. G. Sleno of the former Department of Fisheries and Oceans Arctic Biological Station (Ste. Anne de Bellevue, Quebec, Canada) assisted in preparation of tooth sections of ringed seals. Collections of harp seals were made as part of the research program of the Marine Mammal Section, Department of Fisheries and Oceans (St. John's, Newfoundland and Labrador, Canada); we thank D. Wakeham and D. McKinnon for dissecting out harp seal jaws, extracting teeth, and entering data, and W. Penney for ageing teeth. A. Beltane, L. Burton, J. Hamilton, J. Hinchey, and A. Hussey helped to clean and organize specimens and data. For commenting on manuscript drafts, we thank P. J. Adam (Department of Ecology and Evolutionary Biology, University of California Los Angeles), C. J. Drehmer (Universidade Federal de Pelotas, Pelotas, Brazil), J. M. Lawson (Department of Fisheries and Oceans, St. John's), and W. J. Richardson (LGL Ltd., Environmental Research Associates). Financial support specifically for costs related to this research project was provided to EHM by York University (Toronto, Ontario, Canada), Memorial University, and the Natural Sciences and Engineering Research Council (Discovery Grant to EHM).

Literature Cited

View Abstract