Structure and function of respiratory turbinates in phocid seals
- 188 Downloads
In terrestrial mammals, the respiratory turbinate bones within the nasal cavity are employed to conserve heat and water. In order to investigate whether environmental temperature affects respiratory turbinate structure in phocids, we used micro-computed tomography to compare maxilloturbinate bone morphology in polar seals, grey seals and monk seals. The maxilloturbinates of polar seals have much higher surface areas than those of monk seals, the result of the polar seals having more densely packed, complex turbinates within larger nasal cavities. Grey seals were intermediate; a juvenile of this species proved to have more densely packed maxilloturbinates with shorter branch lengths than a conspecific adult. Fractal dimension in the densest part of the maxilloturbinate mass was very close to 2 in all seals, indicating that these convoluted bones evenly fill the available space. The much more elaborate maxilloturbinate systems in polar seals, compared with monk seals, are consistent with a greater need to limit respiratory heat loss.
KeywordsPinnipeds Phocids Monk seals Turbinates Thermoregulation
Functionally, the respiratory turbinates are used for heat and water exchange (Cole 1954; Jackson and Schmidt-Nielsen 1964). By the time it reaches the lungs, inspired air is warmed to body temperature and saturated with water vapour. Exhalation in the absence of any kind of water conservation mechanism would lead to all this water vapour simply being lost. However, when an animal breathes in through its nose, the respiratory turbinate mucosa over which the air passes is cooled in the process. When breathed back out across the same surface, the exhaled gas is cooled to a value below core body temperature, reducing heat loss. This also leads some of the water vapour, which otherwise would have been exhaled, to condense on the turbinate mucosa, whence it can be absorbed back into the body. Given their high ventilatory rates, respiratory heat and water conservation is particularly important in endotherms, even for those from mesic environments (Hillenius 1992). Some mammals such as domestic cats, rats and rabbits have an inflatable region on either side of the nasal septum called the ‘swell body’. Changes in the level of inflation of swell bodies, which may be cyclic, will change the patterns of airflow through the nose and might therefore be used to exert some control over this heat-exchange function (Negus 1958; Bojsen-Moller and Fahrenkrug 1971).
Phocid seals (Carnivora: Phocidae) are typically marine mammals, found throughout the world but most abundant in colder waters including polar regions. Seals have long been known to have relatively small olfactory turbinates, linked to a poor sense of smell, but branched-type (‘dendritic’) maxilloturbinates with a high surface area (Scott 1954; Negus 1958; Folkow et al. 1988; Mills and Christmas 1990; Van Valkenburgh et al. 2014b). The lack of a ‘swell body’ in seals implies that air must always pass through the narrow maxilloturbinate passageways (Negus 1958). Turbinate surface area measurements have been made from physical sections of seal noses (Huntley et al. 1984; Folkow et al. 1988; Schroter and Watkins 1989) and more recently from CT data (Van Valkenburgh et al. 2011). Van Valkenburgh et al. (2011) made detailed measurements of turbinate area from five species of aquatic carnivores: the phocid seals Hydrurga leptonyx, Mirounga angustirostris and Neomonachus tropicalis, the sealion Zalophus californianus and the sea otter Enhydra lutris. Relative to body mass or skull length, these animals were found to have significantly lower olfactory turbinate surface areas than terrestrial carnivores, while their respiratory turbinate surface areas tended to be larger. Exhaled gas in seals can be 20 °C below body temperature, depending on the environmental air temperature (Huntley et al. 1984; Folkow and Blix 1987). The resulting savings in respiratory water loss can be very impressive, with recoveries of over 80% recorded in the grey seal Halichoerus grypus (Folkow and Blix 1987) and over 90% in the northern elephant seal, Mirounga angustirostris (Lester and Costa 2006), again depending on environmental conditions.
Intriguingly, while Van Valkenburgh et al. (2011) found the maxilloturbinates of the phocid seals Hydrurga and Mirounga to be densely packed with high surface areas, the Caribbean monk seal Neomonachus tropicalis had much less elaborate maxilloturbinates, representing “perhaps the least respiratory surface area for its size in the entire sample” of 16 carnivore species. All marine seals should benefit from reducing respiratory water loss, given that they lack fresh water to drink. If anything, monk seals might be under more intense pressure in this regard than seals from colder environments, given that their cutaneous water losses would be expected to be higher when hauled out. However, if thermoregulation rather than water conservation represents the major selective pressure on respiratory turbinate structure in phocids, seals from higher latitudes would be expected to have better-developed respiratory turbinates than monk seals. The relatively low maxilloturbinate surface area of N. tropicalis, which became extinct probably in the 1970s or 1980s (LeBoeuf et al. 1986), suggests that this might indeed be the case. Whether the two living species of monk seals, the Hawaiian monk seal (N. schauinslandi) and the Mediterranean monk seal (Monachus monachus), have similarly low maxilloturbinate surface areas has not previously been examined.
The aim of this study was to investigate differences in turbinate morphology, with regard to the hypothesis that heat conservation rather than water conservation is the main selective pressure on respiratory turbinate structure in seals. As in previous studies of respiratory turbinate surface areas, we chose to examine the maxilloturbinates only, since these bones are relatively easy to distinguish from the surrounding bones of the skull and represent the great majority of ‘respiratory’, as opposed to ‘olfactory’, turbinate surface area (Van Valkenburgh et al. 2011; Green et al. 2012). We measured maxilloturbinate surface areas in all three species of monk seals, two Antarctic seals (Ross and leopard), two Arctic seals (hooded and bearded) and the grey seal. The grey seal, Halichoerus grypus, is a cold-water species from the North Atlantic. Although it occurs in Greenland, Iceland, northern Norway and Russia (Boehme et al. 2012), it is not regarded as a true Arctic species, unlike hooded and bearded seals which may spend their entire lives within the Arctic region and breed on sea ice (Blix 2005).
Materials and methods
Measurements of condylobasal length (CBL) and total maxilloturbinate surface area (MSA) in eleven seal specimens
Caribbean monk seal
Caribbean monk seal
Hawaiian monk seal (juv)
Mediterranean monk seal
Mediterranean monk seal
Grey seal (juv)
Grey seal (adult)
ASB, pers. coll
ASB, pers. coll
ASB, pers. coll
ASB, pers. coll
One skull of each of the following four polar phocids was examined: bearded seal (Erignathus barbatus; female), hooded seal (Cystophora cristata; male), leopard seal (Hydrurga leptonyx; male) and Ross seal (Ommatophoca rossii; male), the first two being Arctic species, the second two Antarctic. These skulls came from the personal research collection of ASB. The leopard seal had weighed only 178 kg and was therefore deemed to be immature, but the other skulls came from adults. Two skulls of the grey seal (Halichoerus grypus) were examined. One was an adult female skull from the Natural History Museum, University of Oslo (NHMO 7367). The other came from a 14-kg female neonate, a casualty from the Norfolk Wildlife Hospital, UK, which had been prepared in 1998 and is housed in the Veterinary Anatomy Museum, University of Cambridge (SE1). Two Mediterranean monk seal (Monachus monachus) specimens were examined: an unsexed skull from the University Museum of Zoology, Cambridge (UMZ K7781) and a male skull from The Natural History Museum, University of Oslo (M115). Two Caribbean monk seal (Neomonachus tropicalis) specimens were also examined: an unsexed skull from the University Museum of Zoology, Cambridge (UMZ K7801) and a male skull from The Natural History Museum, London (NHMUK 1822.214.171.124). The single examined skull of a Hawaiian monk seal (Neomonachus schauinslandi) was from a juvenile male specimen belonging to The Natural History Museum, London (NHMUK 19126.96.36.199). The condylobasal lengths of these specimens are given in Table 1.
Our sample size was unfortunately limited by specimen availability, scanning resources and the extremely labour intensive image processing described later. Comparable studies have also been based, by necessity, on a very restricted number of specimens (Craven et al. 2007; Van Valkenburgh et al. 2011; Xi et al. 2016).
MicroCT scanning procedure
The skulls of the London and Cambridge specimens were CT scanned at the Cambridge Biotomography Centre. Each skull was separately placed within a Nikon XT H 225 CT scanner on radiotranslucent material in such a way that the tomograms produced were approximately in the transverse plane and covered the anterior skull. Tomograms were reconstructed using CT Agent XT 3.1.9 and CT Pro 3D XT 3.1.9 (Nikon Metrology, 2004–2013). Cubic voxel side lengths were 57–100 μm. The remaining specimens were scanned at The Natural History Museum, University of Oslo, Norway, also using a Nikon Metrology XT H 225 ST CT scanner. Cubic voxel side lengths were 60–75 μm. Further details of the scans are provided as Online Resource 1.
Image processing and measurement
Image processing for the purpose of maxilloturbinate surface area measurement was performed on a subset of equally spaced tomograms. The tomograms chosen ranged from one in three to one in five sections, depending on specimen, so as to obtain a comparable number of processed sections (between 149 and 203) encompassing the whole maxilloturbinate region.
Although the processes described above allowed the maxilloturbinates to stand out well from the background they also introduced noise, especially in the wider regions of air-space found between turbinate bones and between turbinates and nasal cavity wall. Pixels considered to represent noise were painstakingly and manually deleted from each tomogram, with visual reference to pre-threshold scans. Some of the scans, for example those of Cystophora and the adult Halichoerus, proved to require more manual editing than others, owing to poorer signal to noise ratios. Residual soft tissue within the nasal cavity was most evident in the Hawaiian and one of the Caribbean monk seals: this was also manually removed from the tomogram sections, wherever identifiable. All skull elements with the exception of the maxilloturbinate bones, including the nasal septum, were also cropped out, and marrow spaces within thicker regions of the remaining maxilloturbinates were filled in so that they would not contribute to measured turbinate surface area. These processes were performed using GIMP 2.10 and Adobe Photoshop CS 8.0 (Adobe Systems Inc., 2003).
Black pixels with four edges exposed to ‘air’ or four edges exposed to ‘bone’ were not counted at all.
A black pixel with three edges exposed to ‘air’ was counted as 1 unit.
A black pixel with two adjacent edges exposed to ‘air’ (i.e. a corner) was counted as √2 units.
A black pixel with two opposite edges exposed to ‘air’ was counted as 2 units.
A black pixel with one edge exposed to ‘air’ was counted as 1 unit.
The total number of pixel units representing the turbinate perimeter in a given section was then converted to millimetres using the appropriate scaling factor, then multiplied by the distance between adjacent sections. The areas thus obtained were added together to generate an estimate of total (left plus right) maxilloturbinate surface area for that seal specimen.
In order to assess the repeatability of the procedure, a tomogram section which required relatively aggressive noise reduction was reprocessed twelve times on different days, using a range of CLAHE and unsharp mask settings following manual removal of noise. Of the twelve processed sections obtained, the maximum and minimum perimeters differed by under 3.5%. A second test was to compare surface areas of the right and left maxilloturbinate masses, within each specimen, the expectation being that these areas should be very close if the method is reliable. This was indeed the case in all but one specimen (see “Results”).
Nasal cavity cross-sectional areas were measured from the same transverse section tomograms. These were total areas, including not just the air-space but also the areas occupied by the maxilloturbinate bones. Right and left nasal cavities were considered together as one combined area. Areas could only be measured in the region where the bony perimeter of the nasal cavity was intact, or nearly so. The region measured extended from the rostral end of the maxilloturbinate zone where the skull bones separate to create the external nasal opening to the caudal end where the orbital vacuities open.
Measurements from re-oriented sections
This reorientation process was performed for every specimen except for the two Neomonachus tropicalis. It was concluded that the maxilloturbinates were damaged in these specimens (see “Discussion”) and so the results obtained from re-oriented images in these two cases would not be meaningful.
Hydraulic diameter and complexity
If the nasal cavities were perfectly circular and empty of turbinates, Dh (which has dimensions of length) would equal the diameter and the dimensionless Cx would equal 1. The more branched and complex the turbinates occupying the nasal cavity, the more the bony perimeter would increase relative to total airway area, resulting in Dh becoming smaller and Cx larger.
Dh and Cx were calculated from the re-oriented sections made through the densest regions of the maxilloturbinates. In each of these sections, the airway area Ac was calculated by measuring the total area occupied by the left and right maxilloturbinate masses, which includes the air-spaces between the bones, and subtracting from this the total area occupied by bone (i.e. black pixels). The air-space separating the maxilloturbinate mass from the surrounding nasal cavity, nasoturbinate and septum walls was not included in the calculation of Ac, nor were the perimeters of these surrounding structures included in P. These calculations also ignore the area occupied in vivo by soft tissue, i.e. the nasal mucosa, which was missing in the prepared skulls examined. The resulting measurements therefore refer to the maxilloturbinate mass only. Hydraulic diameter and complexity were calculated in this way for each of the three oblique cross-sections constructed from each seal specimen, and the mean values were taken.
The maxilloturbinates of Neomonachus schauinslandi (Fig. 3B, 4B) branch less than in Monachus, and the bones have wider spaces between them. The maxilloturbinates fill the nasal cavity much less fully, but the rounded contours of the right and left maxilloturbinate masses are symmetrical and match the shape of the nasal cavity around them. The maxilloturbinates in the two N. tropicalis specimens (Fig. 3A) have few branches and each occupy only a small, lateral part of the nasal cavity. It appeared that the finer divisions of the maxilloturbinates in these Caribbean monk seals had been broken off (see “Discussion”).
In all seals examined, the maxilloturbinates on each side are connected to the lateral nasal cavity wall through a single, dorsolateral ridge, in the rostral region. However, in species other than the monk seals, this ridge shortens and divides in the posterior part of the nasal cavity, such that transverse sections through this region showed the maxilloturbinates branching from multiple ‘trunks’ emerging from the lateral cavity wall. There were two such trunks in our specimen of Leptonyx, three in Halichoerus, four or five in Erignathus and Ommatophoca and around fifteen in Cystophora (Fig. 7).
Beginning within the olfactory turbinate mass, the short and unbranched nasoturbinates extend downwards in all seals from the roof of the nasal cavity, on either side of the septum (Fig. 7). Between each nasoturbinate and the septum is an open air channel. Right and left channels run adjacent to each other along the top of the nasal cavity, becoming progressively more open rostrally as the nasoturbinates shorten and diverge from the septum.
Left and right maxilloturbinate surface areas were found to be within 5% of each other in all but one specimen. In one N. tropicalis specimen (NHMUK 18188.8.131.52), one side was only 69% the area of the other. No consistent trend was observed across specimens in whether the right area or the left area was the larger.
Measurements of hydraulic diameter (Dh), complexity (Cx) and fractal dimension (D) of the maxilloturbinate mass in nine seal specimens
Hydraulic diameter (mm)
Hawaiian monk seal (juv)
1.70 ± 0.03
1.50 ± 0.00
2.10 ± 0.29
Mediterranean monk seal
1.17 ± 0.07
1.60 ± 0.01
2.01 ± 0.12
Mediterranean monk seal
1.44 ± 0.04
1.54 ± 0.00
1.97 ± 0.12
Grey seal (juv)
0.86 ± 0.05
1.59 ± 0.01
1.98 ± 0.11
Grey seal (adult)
1.06 ± 0.04
1.59 ± 0.01
2.06 ± 0.17
ASB, pers. coll
0.69 ± 0.03
1.67 ± 0.01
2.01 ± 0.09
ASB, pers. coll
0.68 ± 0.04
1.68 ± 0.01
2.00 ± 0.06
ASB, pers. coll
0.89 ± 0.01
1.63 ± 0.00
2.01 ± 0.09
ASB, pers. coll
0.72 ± 0.05
1.67 ± 0.01
2.00 ± 0.07
The log–log plots constructed in the calculation of fractal dimension values (30 per specimen: see “Materials and methods”) all showed tight, linear relationships, the lowest R2 value for any relationship being 0.934. Fractal dimension, calculated by averaging the slopes of these lines, was very close to 2 in all seal specimens examined. Values ranged between 1.97 and 2.10 according to species (Table 2).
Limitations of this study
The methods used here to process the CT data clearly involved an element of subjectivity, but were chosen because the resulting black-and-white images were much better representations of turbinate structure than anything that we could achieve in an automated way, when compared with the original, greyscale tomograms. Manual editing and segmentation has been used in comparable studies of turbinate structure, when considering poorly resolved regions (Craven et al. 2007; Xi et al. 2016). Three findings attest to the internal consistency of our processing methods: (1) repeated processing of the same section yielded very similar measurements (see “Materials and methods”); (2) right and left maxilloturbinate areas were within 5% of each other in all but one specimen; and (3) total areas for our two Monachus specimens, one scanned in Cambridge and one in Oslo, were within 10% of each other (Table 1). As a result, we are confident that the striking differences in species turbinate density documented here, clearly visible by eye from Figs. 3 and 4, cannot be put down to measurement error. The area that we calculated for our immature Hydrurga lay in-between the two values given for larger specimens of this species by Van Valkenburgh et al. (2011), so it may be that our methods generate larger area estimates than those used in the earlier study.
The maxilloturbinate surface areas of our two Neomonachus tropicalis specimens were both extremely low (Table 1; Fig. 10A). Although our two specimens had similar turbinate morphology, we believe that the maxilloturbinates of both specimens were in fact damaged, the more delicate, peripheral bone being broken away and only the thicker structural elements remaining. The fact that one of these specimens had very different right and left maxilloturbinate surface areas is consistent with this conclusion. Although residing in different UK museums, both specimens had been collected by F.D. Godman in the Gulf of Mexico in the 1880 s, and it is therefore likely that the two skulls had been prepared, and damaged, in a similar way. We were unable to access any other usable specimens of this extinct species, but Van Valkenburgh et al. (2011) found maxilloturbinate areas which were twice as high in two specimens from the United States National Museum. The cross-section through the nasal cavity of N. tropicalis illustrated in their paper shows more extensive and elaborate maxilloturbinates than what we observed in our specimens, most closely resembling what we observed in N. schauinslandi. We believe that the Van Valkenburgh et al. (2011) values for this species are more accurate than ours, which should be discounted as coming from damaged specimens.
Maxilloturbinate surface area differences
This study has shown that the maxilloturbinate surface areas in polar seals are much larger than those of the monk seals. This is a result of (1) a nasal cavity of increased cross-sectional area, (2) a longer region of nasal cavity occupied by maxilloturbinates and (3) more complex and densely packed maxilloturbinates. This difference is clear when comparing Monachus to Erignathus and Ommatophoca, which have similar condylobasal lengths (Fig. 6; Table 1). Our Neomonachus tropicalis specimens, which were of comparable size, also had low nasal cavity cross-sectional areas, and maxilloturbinate areas were evidently very low in this species too (Van Valkenburgh et al. 2011). The only N. schauinslandi skull that we were able to examine was from a juvenile specimen, but this alone cannot account for its very low maxilloturbinate surface area because this area was considerably higher in both our Halichoerus pup, which was a much smaller animal, and in our immature Hydrurga.
In order to examine whether differences in maxilloturbinate surface area are simply a result of body size, an allometric scaling approach was taken (Fig. 12). Data points representing our seals were plotted on the regression of maxilloturbinate area on skull length, calculated for non-pinniped carnivores using data from Van Valkenburgh et al. (2011). Van Valkenburgh et al. showed that N. tropicalis has a smaller maxilloturbinate surface area than expected for a carnivore of its size. Our juvenile N. schauinslandi specimen had a surface area very close to expected, while Monachus has maxilloturbinates of relatively higher surface area than Neomonachus species. Halichoerus, from more northernly waters, has a still higher maxilloturbinate area for its size, while the polar seals Erignathus, Cystophora and Ommatophoca have the highest of all. The leopard seal Hydrurga has a lower maxilloturbinate area for its skull length than the other polar seals, but it should be noted that this species has a relatively long skull. No significant correlation of turbinate surface area with latitude was found among terrestrial canid and arctoid carnivores in a previous study, but comparisons between sister taxa (Canis lupus arctos vs. C. l. baileyi; Vulpes lagopus vs. V. macrotis; Ursus maritimus vs. U. arctos) suggested that the three Arctic species/subspecies have enlarged turbinate surface areas compared with their lower-latitude relatives (Green et al. 2012).
The widths of the airways between adjacent turbinate bones can assessed through calculation of hydraulic diameters. Hydraulic diameter in the maxilloturbinate region was found to take a minimum value of around 1 mm in a Labrador retriever mixed-breed dog (Craven et al. 2007) and around 0.7 mm in a New Zealand white rabbit (Xi et al. 2016). Both the dog and rabbit studies were performed using MRI scans on cadavers, which allowed imaging of the nasal mucosa. This was not possible in our study in which only the maxilloturbinate bones were considered. As a result, our airway cross-sectional area Ac is enlarged, leading to our calculated hydraulic diameter values being higher and complexity values lower than they should be in vivo. Xi et al. calculated complexity values between around 1.7 and 2.3 in the rabbit maxilloturbinate region, higher than in any of our seals. However, hydraulic diameters in our polar seals, taken from sections oriented to pass through the densest region, were lower than in the dog and comparable to that of the rabbit. Being much larger animals, this attests to the unusually dense packing of the maxilloturbinate bones in these seals.
The re-oriented cross-sections considered here had been chosen by eye such that the maxilloturbinates filled the available space as evenly and densely as possible. Maxilloturbinate fractal dimensions were close to 2 in all seals, indicating that these bones did indeed fill the nasal cavity evenly in the positions selected. The density of packing, and hence maxilloturbinate surface area for any given volume of cavity, varies according to species. Fractal dimension values slightly greater than 2.0 in Table 2 likely arise from the effects of noise on the regressions, in particular if the turbinate perimeters for the lowest areas sampled happened to be small.
The juvenile Halichoerus was included in this study mainly for comparison with our Neomonachus schauinslandi which was also a juvenile specimen. However, it is also interesting to recognize the differences between the juvenile and adult Halichoerus skulls. It is noteworthy that although complexity values did not differ between these specimens, hydraulic diameter was lower in the juvenile when compared to the adult. The highly branched maxilloturbinates in the juvenile filled the available space more densely than in the adult, in which the terminal divisions were longer. This suggests that the branching occurs early in development, with branch lengths increasing postnatally. The lower hydraulic diameter might help to compensate for the smaller overall maxilloturbinate area in the juvenile, which has a much smaller skull. Of course, comparing just two specimens can only be considered indicative: it would be worthwhile investigating ontogenetic changes in seal maxilloturbinate morphology in more detail.
Heat and water conservation by respiratory turbinate bones
Respiratory turbinates in seals, as in other mammals, reduce respiratory heat and water losses (Folkow and Blix 1987). This heat-exchange function will be improved if the total surface area exposed to the air stream is greater, the distance between the centre of the airstream and the surfaces is narrower, and the velocity of airflow is lower (Schmidt-Nielsen et al. 1970). The time required to achieve thermal equilibrium between the expired gas and turbinate surface is expected to depend strongly on the width of the air passageways between the turbinates, which can be varied by inflation of mucosal tissue overlying the bones (Schroter and Watkins 1989). There are substantial venous sinuses between the surface epithelium and the maxilloturbinate bone in seals (Boyd 1975; Folkow et al. 1988), but these were not taken into account in the measurements made in the present study.
Polar seals and the adult Halichoerus were found to have greater maxilloturbinate surface areas than monk seals (Fig. 10A), achieved in part by a greater packing density which is manifested in the low hydraulic diameter values (Fig. 10B). Both factors would be expected to improve heat exchange. In addition, the wider nasal cavities of these seals (Fig. 11) would lower airflow velocity for any given tidal volume, again augmenting heat exchange. Having wider nasal cavities would also help to offset the increased resistance to breathing conferred by the higher packing density.
While the polar seals live in near-freezing waters all year long, this is not the case for monk seals living around Hawaii, the Caribbean and the Mediterranean. Differences in air temperature, to which the seals are exposed when ashore, would exceed differences in water temperature. Neomonachus schauinslandi is known to feed at night and haul out during the day (Kenyon and Rice 1959; Parrish et al. 2002); animals may remain ashore for 2–3 days (Whittow 1978). When on land, this species could be subject to heat stress, as suggested by observations of seals seeking shady areas, moving to wet sand, changing posture or digging shallow wallows (Kenyon and Rice 1959; Whittow 1978). Like other pinnipeds (Matsuura and Whittow 1974) they do not pant, but may alternate periods of apnoea with periods of ventilation (Whittow 1978). This may contribute to reducing respiratory water losses: Lester and Costa (2006) observed the same in fasting northern elephant seals, Mirounga angustirostris. Seals have the ability to reduce their body temperature when at sea (Hill et al. 1987; Meir and Ponganis 2010), so it is possible that monk seals store heat when hauled out and subsequently lose the excess when they return to water.
Turning to a consideration of water balance, all the seal species examined live in marine environments without access to fresh water to drink. As well as the ‘free’ water obtained from their food, seals generate water from metabolism: Nordøy et al. (1992) have demonstrated that Halichoerus pups can cope with a 52-day period without food or water, meeting their needs through catabolism of body fat and protein. Arctic harp seals (Phoca groenlandica) are able to restore water balance after profound dehydration by drinking seawater (Hov and Nordøy 2007). For Halichoerus on land, Folkow and Blix (1987) have shown that the savings of respiratory water at high ambient temperature, when expired air temperature is also high, are minor. Expired air temperature does not change in response to experimental manipulations of water balance in this species (Skog and Folkow 1994). The need to conserve respiratory water might be expected to be greater for seals in warmer climates, but these studies together with our findings suggest that it may not be of paramount importance. However, in colder regions, seals would experience an increased need to conserve heat. We therefore conclude that heat rather than water conservation represents the more important selective pressure behind the much more elaborate turbinate structures in polar seals, in comparison with monk seals.
Phylogenetic and biogeographic implications
It has been observed that some phocine seals have a posterolateral expansion of the maxilla which might accommodate expanded maxilloturbinates, while lobodontine seals are said to achieve the same through dorsoventral expansion of the nasal cavity (de Muizon and Hendey 1980). Consistent with this, we found Halichoerus, Cystophora and Erignathus to have maxilloturbinates which were relatively more expanded laterally than those of the lobodontines Ommatophoca and Hydrurga, giving them a more diamond-shaped profile in some cross-sections (e.g. Figs. 4, 6). While this morphological difference might appear to support the hypothesis of independent evolution of expanded maxilloturbinates, one population of the Pliocene lobodontine Homiphoca, found in South Africa, also had a posterolateral maxillary expansion (de Muizon and Hendey 1980). There was no obvious difference in the detailed morphology of the maxilloturbinate branches between Arctic and Antarctic seals considered collectively. Interestingly, the looped pattern observed in Hydrurga, which has the least dense and complex turbinates of the polar seals studied, resembles that of Monachus.
Whether the much simpler turbinate structures of monk seals are primitive for crown-group phocids or represent a reduction from an intermediate condition remains unclear. It should be noted, however, that monk seal maxilloturbinate surface areas are relatively close to the expected values for their skull length, by comparison with the general carnivore regression line of Fig. 12.
The clear associations we have found here between nasal structures and latitude offer potential insights into the climatic environments inhabited by fossil seal species, if the nasal regions of these animals are preserved.
The authors would like to thank Richard Sabin and Roberto Portela Miguez of The Natural History Museum, London, Matt Lowe of the University Museum of Zoology, Cambridge and Professor Øystein Wiik for facilitating the loan of their specimens. The Norfolk Wildlife Hospital kindly donated the juvenile grey seal specimen. We thank the Cambridge Biotomography Centre and the Natural History Museum, University of Oslo, for use of their CT scanners. All images reconstructed from Natural History Museum specimens are shown courtesy of the Trustees of the Natural History Museum, London. We are grateful to Clifford Sia and Sergei Taraskin for advising us on our methodology. Finally, we would like to thank our three reviewers for their very constructive comments.
ASB and MJM designed the study. MJM and ØH made the CT scans. LW wrote the software for calculating areas and perimeters. MJM and LW processed the scans and analysed the data. ASB and MJM led the writing of the paper; the other authors contributed to this.
Some of the CT scans were paid for by a grant to MJM from the Department of Physiology, Development & Neuroscience, University of Cambridge.
Compliance with ethical standards
Conflict of interest
The authors declare no competing or financial interests.
- Blix AS (2005) Arctic animals and their adaptations to life on the edge. Tapir Academic Press, TrondheimGoogle Scholar
- Boehme L, Thompson D, Fedak M, Bowen D, Hammill MO, Stenson GB (2012) How many seals were there? The global shelf loss during the last glacial maximum and its effect on the size and distribution of grey seal populations. PLoS ONE. https://doi.org/10.1371/journal.pone.0053000 CrossRefPubMedPubMedCentralGoogle Scholar
- de Muizon C, Hendey QB (1980) Late Tertiary seals of the South Atlantic Ocean. Ann South Afr Mus 2:91–128Google Scholar
- Kenyon KW, Rice DW (1959) Life history of the Hawaiian monk seal. Pac Sci 13:215–252Google Scholar
- Koretsky IA, Barnes LG (2006) Pinniped evolutionary history and paleobiogeography. In: Csiki Z (ed) Mesozoic and Cenozoic vertebrates and paleoenvironments. Ars Docendi, Bucharest, pp 143–153Google Scholar
- Lønø O (1970) The polar bear (Ursus maritimus) in the Svalbard area. Norsk Polarinst Skri 149:1–103Google Scholar
- Moore WJ (1981) The mammalian skull. Cambridge University Press, CambridgeGoogle Scholar
- Negus V (1958) The comparative anatomy and physiology of the nose and paranasal sinuses. E. & S. Livingstone Ltd., EdinburghGoogle Scholar
- Nilssen KT, Haug T, Grotnes PE, Potelov V (1996) Seasonal variation in body condition of adult Barents Sea harp seals, Phoca groenlandica. NAFO Sci Coun Stud 26:63–70Google Scholar
- Saalfeld S (2010) https://imagej.net/Enhance_Local_Contrast_(CLAHE).
- Whittow GC (1978) Thermoregulatory behavior of the Hawaiian monk seal (Monachus schauinslandi). Pac Sci 32:47–60Google Scholar
Open AccessThis article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.