Bone Density Distribution Patterns in the Rostrum of Delphinids and Beaked Whales: Evidence of Family-Specific Evolutive Traits



Toothed whales have undergone a profound telescopic rearrangement of the skull, with elongation of facial bones and formation of a hollow rostrum, filled in vivo by the mesorostral cartilage. In most species of the family Ziphiidae, this latter cartilage becomes secondarily ossified, producing in some cases the densest bone existing in nature. Starting from this observation, we wanted to investigate the patterns of distribution of bone mineral density (BMD) in the rostrum of two families of toothed whales with different ecological and behavioral traits: Delphinidae and Ziphiidae. We analyzed BMD non invasively by means of the dual energy X-ray absorptiometry technology, and found two different density distribution patterns that distinctly set the two families apart. Namely, BMD values decrease from the proximal to the distal region of the rostrum in delphinids, whereas the beaked whales show a BMD peak in the central region. Possible functions such as ballast or protection against clashes might be likely, although more data about the species of both families is needed to give better evidence. Anat Rec, 2010. © 2009 Wiley-Liss, Inc.

Cetacean evolution has prompted specific adaptations for a fully aquatic life. Structural modifications of the body include a radical telescopic transformation of the skull (Miller,1923), consisting principally of an elongation of some bones of the face. Specifically, in toothed whales (odontocetes), the premaxillary bones, the maxillary bones, and the vomer have grown in length and formed a rostrum, with enlarged palatine and pterygoid bones at its posterior region. Furthermore, the bony nasal passages have rotated vertically so that the nares have moved on the top of the head, and the large ascending processes of premaxillary and maxillary bones overlap and conceal part of the enlarged frontal bones. The adaptive significance of the development of a rostrum could be related to morphological and behavioral aspects. It provides a highly hydrodynamic shape to the body, can bear teeth (or baleen plates in the case of baleen whales or mysticetes), and is associated with the development of the “melon,” a specialized fatty organ located in the forehead region of toothed whales and involved in sound emission and echolocation (Cranford et al.,1996). The rostrum contains a narrow groove that in vivo is filled with the mesorostral cartilage, homologous to the cartilaginous nasal septum of all mammals. In most species of the family Ziphiidae (or beaked whales), it begins to ossify with the attainment of sexual maturity. It is unclear whether bone formation arises from ossification of the cartilage (McCann,1965; MacLeod,2002), or derives from a dorsal growth and intrusion of the vomer (Fraser,1942; Heyning,1989). In beaked whales, this ossification may produce in some species a very dense and compact bone (Raven,1942; Heyning,1989; de Buffrènil and Casinos,1995; de Buffrènil et al.,2000; Zotti et al.,2009). This unique feature is one of the anatomical characteristics that clearly distinguish the skull of beaked whales from that of the family Delphinidae, including an elevated wide vertex and reduced dentition. The two families are also set apart by very different ecological traits. Beaked whales are typically pelagic animals, swim alone or in very small groups, feed mostly on deep-sea cephalopods and, for that reason, dive to extreme depths for very long times. On the contrary, although many odontocete species are pelagic as well, delphinids are highly gregarious, prey principally upon fish and dive mostly in middle-shallow waters (Jefferson et al.,2008). Given the evident external morphological distinction between the skulls of these two families, we wanted to investigate whether anatomical differences are associated with variations in bone density, and in particular, in its distribution along the rostrum. In this study, we analyzed and compared biometrically the proportions of the rostrum of the two families, and measured bone mineral density (BMD) by using the dual-energy X-ray absorptiometry (DXA), a non invasive technology nowadays regarded as the “golden standard” for this purpose in clinical medicine (Adams,1997; Fogelman and Blake,2000).


Specimens and Bone Density Measurements

The skulls of 13 delphinids and five beaked whales were collected from four Italian museums (Table 1). Each specimen was scanned laying horizontally, in a caudo-rostral direction and in a dorso-ventral projection, by means of a DXA device (Hologic QDR-1000™, Hologic, Waltham, MA) (Fig. 1). All the scans were performed by the same operator (AZ); the calibration of the machine was checked on a regular basis by means of its calibration phantom (Hologic Calibration Phantom™, Hologic, Waltham, MA). The rostrum of each skull was subdivided into three Sub-Regions of Interest of the same length [proximal (SRoI-Prox), medial (SRoI-Med), and distal (SRoI-Dist)], starting from an imaginary line connecting the two infraorbital foramina in the maxillary bones (Fig. 2). The Subregion Analysis Lumbar Spine Software™ (version 6.2 D, Hologic, Waltham, MA) was used to calculate the BMD of each SRoI.

Figure 1.

Skull of Ziphius cavirostris on a DXA scanner.

Figure 2.

On the left, dorsal view of two skulls, with Sub-Regions of Interest marked by dotted lines and maxillary infraorbital foramina indicated by arrows; on the right, the corresponding digital images of each sub-region. Panel A, skull of the beaked whale Z. cavirostris MSNM 472-4904; Panel B, skull of the delphinid G. griseus MSNM 6135. Dist, distal Sub-Regions of Interest; Med, medial Sub-Regions of Interest; Prox, proximal Sub-Regions of Interest.

Table 1. BMD values (g/cm2) for each sub-region of interest of the rostrum of each specimen, whose respective ID is given in brackets
SpecimenRostrum sub-region
  1. FMV, Museum of the Faculty of Veterinary Medicine, University of Padua; MZ, Zoological Museum of the University of Padua; MSNM, Museum of Natural History of Milan; MSNG, Museum of Natural History “G. Doria” of Genua; SRoI-Prox, proximal Sub-Region of Interest; SRoI-Med, medial Sub-Region of Interest; SRoI-Dist, distal Sub-Region of Interest.

 Tursiops truncatus (FMV 2)0.9360.9200.742
 Tursiops truncatus (FMV 3)1.0180.8490.570
 Tursiops truncatus (FMV 4)0.8970.8120.676
 Tursiops truncatus (FMV 5)1.1941.0850.753
 Tursiops truncatus (MZ 2)1.1411.0220.695
 Stenella coeruleoalba (FMV 1)1.0330.8750.488
 Grampus griseus (MSNM 6135)1.4471.0870.931
 Grampus griseus (MSNM 7277)1.3581.1250.835
 Grampus griseus (MZ 1)1.681.2230.951
 Pseudorca crassidens (MSNM 4784)1.9971.3040.895
 Delphinus delphis (MZ 3)0.8800.7800.494
 Globicephala melas (MSNM 469)1.2111.1941.066
 Globicephala melas (MSNM 4883)1.4161.0710.972
 Ziphius cavirostris (FMV 6)0.8531.3070.956
 Ziphius cavirostris (MSNM 472-4904)1.1481.4610.854
 Ziphius cavirostris (MSNM 4903)0.8701.1410.715
 Mesoplodon europaeus (MSNM 7275)1.1831.3240.734
 Mesoplodon densirostris (MSNG 16)3.2454.4813.543

Biometric Measurements

Biometric measurements were taken from the rostra of three Z. cavirostris (representative for the Ziphiidae family) and four T. truncatus (representative for the Delphinidae family). Each rostrum was subdivided into 40 parts, corresponding to 2.5% of the total length, starting at the imaginary line connecting the two infraorbital foramina. The width and vertical thickness of the rostrum were measured by means of a Vernier Calipers in correspondence of each percentile.

Statistical Analysis

The mean bone density of single SRoIs between the two families was compared and analyzed by Student's t-test for independent samples using the software Minitab (version 15.1.1, Minitab, State College, PA) and setting the significance level for P < 0.05.


Bone Mineral Density Measurements and Statistical Analysis

The results of BMD measurements are displayed in Table 2, and the corresponding plot showing the trend of the values along the rostrum of all specimens is represented in Fig. 3. For the delphinids, BMD values range from 0.488 g/cm2 (SRoI-Dist of S. coeuruleoalba FMV 1) to 1.997 g/cm2 (SRoI-Prox of P. crassidens MSNM 4784). Among the beaked whales, M. densirostris MSNG 16 shows BMD values ranging from 3.245 g/cm2 (SRoI-Prox) to 4.481 g/cm2 (SRoI-Med). These are much higher than what was observed for the other con-familiar specimens, whose values range from 0.715 g/cm2 (SRoI-Dist of Z. cavirostris MSNM 4903) to 1.461 g/cm2 (SRoI-Med of Z. cavirostris MSNM 472-4904). Since the inter-familiar comparisons would be too unbalanced including also the exceptional bone density of M. densirostris, we have chosen to consider only the other beaked whales for comparison with delphinids.

Figure 3.

Different patterns of BMD along the rostrum in the two families. Dotted lines represent delphinid specimens, solid lines represent beaked whales. The central gray zone marks the medial BMD peak in the rostrum of beaked whales. Notice the extremely high density values for M. densirostris MSNG 16, represented by the upper solid line.

Table 2. Mean BMD (g/cm2) of each sub-region of interest, within each family
FamilyRostrum sub-region
  • Values are Mean ± SD.

  • *

    Total Ziphiidae, all the beaked whales specimens; Ziphiidae*, all the beaked whales specimens but M. densirostris SNG; SRoI-Prox, proximal Sub-Region of Interest; SRoI-Med, medial Sub-Region of Interest; SRoI-Dist, distal Sub-Region of Interest.

  • **

    **P < 0.05 Student's t-test.

Delphinidae1.247 ± 0.329**1.027 ± 0.1670.774 ± 0.187
Ziphiidae*1.014 ± 0.176**1.308 ± 0.1310.815 ± 0.112
Total Ziphiidae1.460 ± 1.0101.943 ± 1.4231.360 ± 1.224

The mean BMD of the SRoI-Prox is higher for Delphinidae (1.247 ± 0.329 g/cm2) than for Ziphiidae (1.014 ± 0.176 g/cm2), although this difference is not significant (t = 1.840, P = 0.096). On the contrary, the mean bone density of the SRoI-Med is significantly higher (t = −3.510, P = 0.013) for the beaked whales (1.308 ± 0.167 g/cm2) than for the delphinids (1.027 ± 0.131 g/cm2). Finally, mean BMD of SRoI-Dist are only slightly lower for the delphinids (0.774 ± 0.187 g/cm2) than for the beaked whales (0.815 ± 0.112 g/cm2), again with no significant difference (t = −0.53, P = 0.612). Notably, the mean BMD of SRoI-Prox of delphinids is not statistically different from the mean BMD of the SRoI-Med of beaked whales (t = −0.550, P = 0.593); vice versa, SRoI-Prox mean BMD of Ziphiidae and the SRoI-Med mean BMD of Delphinidae are not statistically different (t = 0.130, P = 0.901).

Biometric Measurements

The distribution of width and vertical thickness values of the rostrum of the four T. truncatus and the three Z. cavirostris is illustrated in Fig. 4. Although curves of both families have similar shapes (especially in the middle segment and in the rostralmost part), both parameters, from base to tip, are higher and decrease more rapidly in beaked whales than in delphinids.

Figure 4.

Width (Panel A) and vertical thickness (Panel B) distribution pattern along the rostrum of three Z. cavirostris (solid lines) and four T. truncatus (dotted lines); values are given as Mean ± SD. Upper traces represent the distribution pattern of mean BMD along the rostrum in beaked whales specimens [except M. densirostris MSNG 16 (solid lines) and in delphinids (dotted lines)]. *P < 0.05 Mean BMD of SRoI-Med of delphinids compared with Mean BMD of SRoI-Med of beaked whales.


The DXA technology is regarded as the “golden standard” for bone density measurement both in human and veterinary medicine (Grier et al.,1996; Fogelman and Blake,2000). Till now, the only studies addressing cetaceans in which this technique has been used concerned the flippers of S. coeruleoalba (Guglielmini et al.,2002) and of T. truncatus (Butti et al.,2007), and the rostrum of M. densirostris (Zotti et al.,2009). To the best of our knowledge, this is the first systematic application of DXA to the comparative study of BMD of the rostrum of toothed whales. Other studies analyzing the rostrum used the Archimede's principle (Raven,1942; Heyning,1984; de Buffrènil and Casinos,1995), ultrasound propagation (de Buffrènil et al.,2000), radiographic analysis (de Buffrènil et al.,1986), or CT-imaging (Cranford et al.,2008). All these latter techniques provide bone density in a volume (g/cm3), whereas DXA measures density in an area (g/cm2), so the two types of data are hardly comparable. Nevertheless, a high correlation has been found between planar BMD and volumetric density values obtained with Archimede's principle, in rodents (Keenan et al.,1997) and bovines (Toyras et al.,1999), as well as between BMD and X-ray CT values in human beings (Grampp et al.,1997).

Our study confirms the former findings that the beaked whale M. densirostris (MSNG 16) is characterized by an extremely high BMD, ranging from 3.245 g/cm2 to 4.481 g/cm2. In literature, measured bone density for the rostrum of this species ranges from 2.3 g/cm3 (Raven,1942) to 2.87 g/cm3 (Cranford et al.,2008); although not volumetric, we can reasonably suppose that our data surpass any bone density recorded in literature. The role of such an extreme density for a structure not bearing weight is still debated (Heyning,1984; Zioupos et al.,1997; Currey,1999; de Buffrènil et al.,2000; MacLeod,2002).

The data that we report here (Fig. 3) show that the distribution of BMD follows two distinct patterns in Delphinidae, where it always decreases from the SRoI-Prox toward the SRoI-Dist, and in Ziphiidae, where it increases toward SRoI-Med and diminishes toward the SRoI-Dist. Delphinidae have a higher inter-individual variability than Ziphiidae, possibly due to the composition of the experimental series, but we also emphasize that, despite having two different patterns, the two families have quite similar BMD ranges (except for M. densirostris MSNG 16). The way bone density is spread in the rostrum could be a new discriminating factor between Delphinidae and Ziphiidae. The width and vertical thickness of the rostrum are much higher toward the base in beaked whales (Fig. 4), but the BMD data we gathered show that bone density in the SRoI-Prox is higher for delphinids than for beaked whales (Table 2). Our main finding is the significantly higher BMD in the SRoI-Med of Ziphiidae. The apparent “shift” of the BMD peak toward the center in the rostrum in the Ziphiid family may reveal an important evolutionary and functional trait. Three main hypothesis may be proposed to explain the possible adaptive significance of the peculiar BMD pattern here reported for the beaked whales: (1) muscle insertion, (2) aggressive intra-specific interactions, and (3) diving behavior.

Muscle Insertion

Mead (1975) and Purves and Pilleri (1978) described the facial anatomy of several delphinid species, and Heyning (1989) of many beaked whales. These observations indicated a very similar general organization of the rostral muscles in all the species studied. Namely, these are the “medial rostral muscle” and the “lateral rostral muscle,” whose function is not yet fully understood. They are probably highly derived portions of the mammalian maxillonasolabialis muscle (Mead,1975). Since the anatomical organization of the rostral muscles seems to be very similar between the two families, this factor alone cannot explain the distinct bone density patterns described here, although Heyning (1989) reported that in a specimen of M. densirostris, the rostral muscles were more developed than in other congeneric species.

Aggressive Intra-Specific Interactions

The aggressive interactions in Ziphiidae have not yet been observed and can only be inferred analyzing the external scarring patterns on the body of sighted or stranded specimens. Unlike Delphinidae (for description of fighting in this family refer Scott et al.,2005), the dentition of most beaked whales is reduced to two mandibular teeth unsuitable for biting. The scars observed on beaked whales are mainly linear, single or double and parallel, focused around the blowhole and along the back. The scars may be most likely the result of competitors swimming toward each other and raking the opponent with the teeth as they encounter (Heyning,1984; MacLeod,2002). In these clashes, heavy compression forces might be exerted either longitudinally along the rostrum and from above. A higher BMD around the central region of the rostrum (SRoI-Med) could, therefore, help cushioning the impact and preventing serious damages. However, a series of observation on the bone ultrastructure in M. densirostris (de Buffrènil and Casinos,1995; Zioupos et al.,1997; Zylberberg et al.,1998; Currey,1999; de Buffrènil et al.,2000) suggest that the osteone distribution may indicate an increased fragility of the rostrum. We instead agree with MacLeod (2002), who suggested that the peculiar ultrastructural organization would prevent serious transverse fractures to occur.

Diving Behavior

Delphinidae typically dive to maximum depths of a few hundred meters and feed mainly on fish or small invertebrates that live in the upper layers of the water column (Jefferson et al.,2008). On the contrary, Ziphiidae are mainly teuthophagous, feed principally upon meso- or bathy-pelagic cephalopods, and therefore, have to dive to great depths. For example, Tyack et al. (2006), for the con-familiar Z. cavirostris and M. densirostris, observed a maximum depth of 1888 m for the former, and of 1251 m for the latter. In both species, descents were faster than ascents, and the angle to the surface varied between 60° and 83° in both directions. A higher BMD at the center of the rostrum in Ziphiidae could possibly function as a ballast (Reidenberg,2007), displacing forward their center of gravity and helping these animals to rapidly reach deep foraging grounds. Delphinidae, which instead dive to a maximum of few hundred meters, would not need such a ballast. Notably, de Buffrènil and Casinos (1995) pointed out that beaked whales have a very low relative weight of the whole skeleton (2.37%), lower than most cetaceans (3.5%–5%). This, together with the higher BMD in the centre of the rostrum of Ziphiidae, could bring about an unbalancing forward, facilitating a passive sinking during the angular descent. MacLeod (2002) emphasized that the beaked whale genus Hyperoodon, able to reach depths of more than 1400 m, lacks mesorostral ossification. However, we do not know whether the rostrum of the genus Hyperoodon has a BMD distribution pattern like Ziphiidae analyzed here, and this could be the aim of a future study.

In conclusion, we consider that the diversity in the pattern of bone density distribution that we report here in the two toothed whale families may be related to other structural traits that set them apart. Delphinidae and Ziphiidae seem to share a very similar organization of the facial muscles, a fact that does not account for different BMD patterns. The peculiar modalities of intra-specific aggressive interactions and the diving patterns may be relevant for the evolution of diverse distribution of calcium salts in the rostrum. Our results do not allow us to state positively which factor most influenced the shape and the bony structure of the face. However, a higher BMD in the center of the rostrum of beaked whales could serve both as protection during clashes with rivals and as ballast to aid deep diving. Future broader studies could ascertain whether other Ziphiidae (including large members of the genera Hyperoodon and Berardius) show the same bone density pattern as the con-familiars, thus possibly further supporting one or more of the hypotheses presented here. Another aspect to be cleared concerns the possible sexual dimorphism of the BMD patterns: the data we gathered are not sufficient to state clearly whether gender affects or not the BMD pattern. Future studies could further investigate this issue, besides possible relationships with teeth position or length/thickness of the rostrum, relying on a higher number of specimens belonging to both families.


The authors thank Dr. Roberto Poggi, Director of the “G. Doria” Museum of Natural History of Genova (Italy), for allowing them to use the precious specimens of the collections. They also thank Emanuele Zanetti, Giuseppe Palmisano, and Maristella Giurisato of the University of Padova; and Ermano Bianchi, Giorgio Bardelli e Ermanno Montanara of the Museum of Natural History of Milan, for their qualified and tireless technical assistance. They thank Dr. Beniamino Fiore and Dr. Monica Soncini of the Politecnico di Milano for the useful insights during the initial planning of this study.