Rorqual whale nasal plugs: protecting the respiratory tract against water entry and barotrauma

Rorqual whales spend most of their lives holding their breath underwater. Expiring and inspiring at the interface of air and water, and spending much of the time at depths where ambient pressures are high generate constant risks to the respiratory system. Air exits and enters the respiratory tract through the blowholes (nostrils), and the blowhole margins have been described as the first line of defence against water incursion (Maust-Mohl et al., 2019). The actions of the margins are clearly visible with aerial photography and animal cameras at the surface (Fig. 1). Prior to surfacing, the blowhole margins remain closed and the rostrum appears flat. Exhalation starts once the blowholes break the water surface and open. As this occurs, the blowhole margins rise up, and a bulge is created rostrolateral to each blowhole. Inhalation follows, with a widening of the blowholes (Racovitza, 1904; Scholander, 1940) and a slight enlargement of the rostrolateral bulge. Once inhalation is complete, the blowhole margins close and the rorqual dives.

While viewing footage of rorquals surfacing and breathing, we have made observations that indicate that protection of the respiratory tract involves more than just closure of the blowhole. First, rorquals occasionally submerge their rostrum before the blowhole margins are closed, allowing water to flow over and into the open blowholes (Fig. 2A,B). Second, bubbles can be released from the blowholes prior to the rorqual surfacing while the blowhole margins remain closed (Fig. 2C,D). These occurrences show that the blowhole margins do not always prevent water entry, nor are they airtight while submerged.

The respiratory system is protected by two groups of nasolabial muscles: those distributed superficially that link to and control the cartilaginous blowhole margins (see Maust-Mohl et al., 2019, for details), and those on a deeper plane that form paired, cylindrical nasal plug muscles that extend from the rostrum to halfway down the nasal cavities in the skull (Maust-Mohl et al., 2019). Nasolabial muscles are described anatomically as a single functional group; however, we argue here that their different morphologies and different positions relative to the skull indicate they form two, separate functional groups.

Rorqual breathing is explosive, lasting only 1–2 s. Achieving high air flow rates requires minimizing resistance through the nasal cavities and blowholes (Kooyman, 1973), but the size and tethering of the nasal plug muscles do not appear to be optimized for facilitating this. The plugs are very large; considerable effort must be exerted to move them quickly for each breath, and their attachments deep in the nasal cavities would appear to prevent their complete removal when breathing. This incongruity raises the possibility that the nasal plug muscles are designed for another function that necessitates their large size and deep attachment in the nasal cavity. High ambient pressures expose diving mammals to two potential problems: decompression sickness from excess nitrogen absorption at depth and barotrauma from mechanical distortion of the tissue. Shifting air out of gas exchange regions prevents nitrogen absorption (Bostrom et al., 2008; Fahlman et al., 2017; Kooyman and Ponganis, 2018; Leith, 1970; Scholander, 1940), and both the high compliance of the alveoli and chest wall and the low compliance of the upper respiratory tract are essential to achieve this (Bostrom et al., 2008; Cozzi et al., 2005). The risk of nitrogen absorption ends once the lungs have collapsed and gas exchange has ceased, but the danger of barotrauma to the stiffened portions of the respiratory tract remains a problem.

We argue that the deep attachment of the nasal plug muscles is a compromise that allows them to accomplish two disparate tasks, control of respiratory tract opening and prevention of barotrauma at depth (Fig. 3). We hypothesize that the nasal plug muscles, rather than the blowhole margins, form the main valve controlling the opening to the respiratory tract. The blowhole margins are not competent on their own at depth, and instead they streamline the rostrum. The position of the nasal plug muscles in the nasal cavities presents large masses of tissue that prevent water from being forced into the nasal cavities at high ambient ocean pressures when at depth. We suggest that the nasal plug muscles actively withdraw from their position in the nasal cavities before the blowhole margins open, and that the nasal plug muscles are back in position occluding the nasal cavities before the blowhole margins close. This timing would account for air escaping the blowhole while the margins remain closed and would also allow water to flow over the open blowhole margins without incursion into the respiratory tract. We also hypothesize that the deep attachment of the nasal plug muscles in the nasal cavities is necessary to prevent barotrauma at depth. High ambient ocean pressures force the nasal plug muscles deeper into the skull, partially filling and hence raising the pressure inside the non-collapsible nasal cavities. We investigated these hypotheses by examining videos of rorquals surfacing, morphometry of skulls, performing dissections, and making model calculations. We also examined the relevant anatomy in an MRI scan and evaluated the histology of selected regions of the nasal plug muscles.

Photos and videos from animal cameras and aerial cameras of blue and humpback whales were obtained from Duke Marine Robotics and Ved Chirayath, NASA Ames Research Center [National Science Foundation (NSF) IOS 1656691 and National Marine Fisheries Service (NMFS) permit 16111]. From these, we were able to observe and describe the timing of respiration, extent of blowhole opening, anatomical features surrounding the blowhole, and general movements of tissue surrounding the blowholes.

Skulls from six rorqual whales in museums were examined in their respective locations for morphological measurements including length and diameter of the nasal cavities. Details of all skulls and heads used are listed in Table 1. A sub-adult minke whale skull was used to estimate nasal cavity volume by filling the nasal cavities with rice and measuring the volume of rice used.

MRI images of a fetal minke whale head from a previous anatomical study (Pyenson et al., 2012) were viewed on OsiriX Lite (Rosset et al., 2004). We observed the position of the nasal plug muscles in the nasal cavities and determined the volume of both the nasal cavities and nasal plug muscles from slice thickness and area using Fiji (Schindelin et al., 2012).

Fin whale specimens were collected postmortem from the commercial whaling operation at Hvalfjörður, Iceland, in the summers of 2015 and 2018. All specimens were examined fresh at the station within 24 h of death. A total of 14 animals were surveyed, with 8 examined in detail. One dissected animal was a fetus, 3 m long. All others were adults with an average total length of 17–21 m. Tissue samples were imported to Canada under Convention on International Trade in Endangered Species of Wild Fauna and Flora (CITES) permits.

Nasal plug muscles in fin whales were examined by removing blubber surrounding and directly dorsal to the blowhole and along the rostrum. Nasal plug muscle was revealed by progressive horizontal slicing through superficial muscle until it was visible. At various stages in the dissection, the nasal plug muscles were physically manipulated to understand how they could function in a living whale. Hooks were placed in nasal plug muscle to pull rostrally. The extent to which we were able to open the vestibule (the space created by the open blowhole) was recorded and compared with aerial photos. Likewise, the nasal plug muscles were forced ventrocaudally into the nasal cavities to determine the position the plugs could achieve under high ambient pressure while diving. In both fetal and adult specimens, sagittal and transverse slices were made through the skull to examine the attachments and location of nasal plug muscle and the proportion of the nasal cavities occupied by nasal plug tissue. Photos of transverse slices of the skull were used to estimate nasal cavity volume and nasal plug volume using Fiji.

Nasal plug tissue from a fin whale was sampled from five locations in a single nasal plug: medial and lateral aspects of the nasal fossa/vestibule region, midway between the vestibules and nasal septum attachment, and at two locations along the attachment to the nasal septum. Samples were fixed in 10% neutral buffered formalin, and then processed by Wax-It Histology Services Inc. (University of British Columbia), following standard techniques. Samples were sliced at 5 µm thickness and stained with Verhoeff–Van Gieson stain to display elastin in black, collagen in pink and muscle in yellow/beige. Slide scans were viewed using QuPath (Bankhead et al., 2017) to determine relative proportions of elastin, collagen and muscle at each location sampled.

Exhalation and inhalation in blue and humpback whales followed the same patterns as previously described in rorquals – exhaling upon surfacing and inhaling before diving (Fig. 1). During exhalation, the superficial muscles contracted to open the blowholes and lift the blowhole margins to create a raised barrier. Additionally, a bulge was created just rostral and lateral to the blowhole (rostrolateral bulge) and appeared to divert water from the blowhole. We found that the rostrolateral bulge projected above the flat surface of the rostrum rather than solely fitting in the skull depression/nasal fossa as previously described (Maust-Mohl et al., 2019). When the blowhole margins opened, the dark epithelial lining of the vestibules became visible. During inhalation, the margins opened slightly further in a buckling-like motion. Muscular contraction to open the blowholes probably further pulls on the cartilage lateral and caudal to the blowholes to cause this buckling. The rostrolateral bulge increased slightly in size during inhalation. The superficial muscles relaxed to close the blowholes and streamline the rostrum by smoothing its dorsal surface when submerged.

The rorqual skull possesses a large rostrodorsal depression termed the nasal fossa (Buono et al., 2015). The nasal fossa extends from the tip of the rostrum to the external bony nares at the top of the nasal cavities. The nares are bounded caudally by the nasal bones and rostrolaterally by the premaxilla/maxilla. Below this, the nasal cavities curve caudoventrally towards the internal bony nares. The nasal cavities were elliptical in cross-section and did not change significantly in size throughout their length.

In minke and fin whales, the superficial muscles formed the blowhole margins surrounding the external nares. The nasal plugs were deep to the more superficial muscle surrounding the nares, and extended caudoventrally from rostral origins in the nasal fossa into the nasal cavities (Figs 4, 5 and 6). The nasal plugs were attached rostrally, approximately two-thirds to one-half of the way down the rostrum from the anterior tip to the mesorostral cartilage. This cartilage lies in the midline of the nasal fossa and separates it into left and right sides (Fig. 5A,C). The caudal attachment point of each nasal plug was also medial, on the nasal septum, approximately halfway down the nasal cavities (Figs 4 and 5). Each side of the nasal fossa was filled by a cylindrical nasal plug. The nasal plugs enlarged towards the blowholes, acquiring a bulbous appearance and filled the vestibules. From this position, the nasal plugs extended down into and completely filled the upper half of the nasal cavities. The plugs narrowed toward their attachment on the nasal septum, leaving some unoccupied air space lateral to the plugs when relaxed (Figs 4 and 6).

Removing the skin, blubber and superficial muscle surrounding the blowholes in fin whales revealed the nasal plugs. We were able to pull the nasal plugs rostrally (Fig. 5B,D), a manoeuvre that probably mimics the creation of the rostrolateral bulge visible in videos of rorquals breathing. The plugs are attached to the skull at both ends; it was the fatty bulbous portion in the vestibule that we pulled rostrally from the upper nasal cavities to create the bulge. Shifting the plug formed an open space through the dorsal portion of the nasal cavity and revealed the black epithelial lining of the nasal cavities that was visible in videos. With more force, the entire bulbous portion could be cleared from the nasal cavity to maintain an unobstructed airway for breathing.

Sagittal slices of the adult fin whale skulls made slightly off the midline and through one of the nasal cavities revealed the position of nasal plug tissue in the nasal cavity and the remaining air space below this (Fig. 6). Transverse slices of multiple skulls also yielded an average estimate of nasal plug volume and air space volume. For an average 19 m adult fin whale, the volume of the two nasal plugs combined within the nasal cavities was 16 l, and the air space volume in the nasal cavities was 16 l with the plugs in place and 32 l with the plugs removed. The total length of the plugs from attachment in the rostrum to attachment in the nasal cavities was approximately 1.3 m. The total length of the nasal cavities alone was approximately 1 m.

There were visibly distinct differences in the composition of the fin whale nasal plugs along their length (Fig. 6). The rostral nasal fossa region was almost entirely skeletal muscle. Near the blowhole and vestibules, this transitioned to adipose tissue (fat) with much less muscle. A slice through the plugs in the vestibule revealed there was some regionality to the muscle and fat distribution, laterally presenting more muscle, while medially there was more fat. Collagen fibres were visible throughout the fat and muscle in this region (Fig. 7A,B). There was very little muscle beyond the vestibules as the plugs descended towards their attachment points.

Histology confirmed this difference in tissue composition along the fin whale nasal plugs (Fig. 7). No samples were collected from rostral to the blowhole, but our gross observations demonstrate that this region is almost strictly muscle, clearly making it the contractile region of the nasal plugs (Figs 5–7). The bulbous vestibule region and the upper nasal cavity region were mainly fat with only some muscle (Fig. 7A–D). As fat is incompressible and easily deformed, this region of the plug can readily seal the nasal cavity. Collagen was distributed around the muscle bundles and throughout the adipose tissue (Fig. 7C,D). There was almost no muscle present caudal to this region, except for a few small ‘islands’ of muscle near the rostral attachment region to the nasal septum (Fig. 7E). Along the attachment region to the nasal septum, the plug was composed almost completely of elastin and collagen, especially at the caudal-most part (Fig. 7F,G). We thus refer to this region as the ‘tendon’, because of its composition and its role in attaching the nasal plug muscle, through collagen, to the bony nasal septum at the posterior of the vomer bone. The very high elastin content of the tendon would allow it to readily stretch and store elastic energy. Because of the wide distribution of tissue types seen along their length, we have chosen to use the term ‘nasal plugs’ to refer to the entire structure rather than ‘nasal plug muscles’ as this more accurately reflects their heterogeneous composition and specific function(s). The heterogeneous composition of the plugs is summarized in Fig. 8.

Actions of the blowhole margins are easily observable in breathing rorquals, so it is not surprising that the blowhole margins are currently regarded as the primary mechanism preventing water incursion. However, based on observations of videos and photographs taken from directly above living rorquals at the surface and the anatomy of fresh specimens, our results indicate that the nasal plugs are the primary valves preventing water incursion. Closing the respiratory tract and carrying a volume of air into a high-pressure environment poses challenges to marine mammals: preventing water entry, preventing decompression sickness and preventing the formation of damaging pressure differentials. The nasal plugs are a novel protective feature in rorquals, being the final step in the graded compressibility chain of the respiratory tract, and respond to the creation of pressure differentials to prevent barotrauma. Much of our data was collected from fin and minke whales, but we believe our results are representative of all rorquals as they possess a similar skull structure, and exhibit similar diving patterns with explosive breathing. We propose the following model (Figs 8 and 9) for how the nasal plugs function at the surface in preventing water incursion and at depth in preventing barotrauma.

The superficial muscles and nasal plugs are activated asynchronously throughout the breathing cycle and remain in a relaxed state while submerged. These muscles are probably innervated by sensory nerves of the maxillary division of the trigeminal nerve (CN V₂) and controlled by the facial nerve (CN VII), as nasolabial muscles are modified muscles of facial expression. When surfacing, before the blowholes have breached the water surface, air bubbles visibly escape the passively closed blowhole margins, indicating the muscular nasal plugs have begun to contract to allow passage of air through the nasal cavities (Fig. 2). Beginning exhalation while still underwater minimizes the time spent at or near the surface where drag forces can be substantially higher (Hertel, 1966). As a rorqual's blowholes break the water surface, the nasal plugs are pulled fully rostrally and the superficial muscles spread the blowhole margins. This withdraws the bulk of the plug tissue from the nasal cavities, clearing the passages for nearly unobstructed air flow. The rostral movement of the nasal plugs stretches the elastin-dominated tendons that attach the plugs to the nasal septum. After inhalation, but just before the rostrum dips underwater, the nasal plug muscles relax, releasing the tension on the tendons, which then pull the nasal plugs back into their rest position through elastic recoil. Given the high elastin composition of the tendon (Fig. 7), this recoil is likely to be rapid, which minimizes the time that the upper respiratory tract is open and at risk of water incursion. Importantly, the plugs are at rest when they are within the nasal cavities and therefore require no work to keep them in the ‘plugged’ or default position. In this position, the nasal cavities are occluded by the fatty bulbous part of the nasal plugs (Fig. 6). Occasionally, when near the end of a breathing cycle, the whale's rostrum dips below the water surface before the superficial muscles have relaxed to close the blowhole margins, and water flows into the vestibules (Fig. 2). This demonstrates that the superficial muscles relax after the nasal plug muscles. The relaxed nasal plug muscles prevent water from entering the respiratory tract beyond the vestibules during this interval.

The regional heterogeneous composition of the nasal plugs is necessary for them to be withdrawn from the nasal cavities during breathing. The nasal plugs have been referred to as ‘nasal plug muscles’; however, if these structures were entirely composed of muscle, it would be impossible to remove them from the nasal cavities. When muscles contract, they increase in cross-sectional area; when fat is pulled on, it decreases in area. If the plugs were strictly muscular and sealed the nasal cavities in their relaxed state, then muscular contraction could only further plug the nasal cavities. Instead, the plugs must have a contractile region to withdraw them from the nasal cavities; an elastic region to anchor them, allow them to be withdrawn and pull them back into place; and a fatty region that lengthens and thins when pulled, and settles to fill out and seal the nasal cavities as the tendon recoils.

The air-filled respiratory tract is subject to ambient ocean pressures (P_amb) that increase by 1 atm (∼101.3 kPa) for every 10 m of depth. Depending on the compliance of the surrounding structures, these external pressures progressively reduce the air volume and consequently raise the pressure inside the respiratory tract. The resultant internal pressure can be calculated from Boyle's law: P_d=(P_sV_s)/V_d, where P_d and V_d are the pressure and volume at depth, and P_s and V_s are the values at the surface. If the surrounding structures are stiff and resist compression, there will be little reduction in volume and hence little rise in internal pressure. This will result in a pressure differential between the respiratory tract and P_amb. A pressure differential may not damage the walls of the structure if they are strong, as in submarines or the bones of the skull that surround the nasal cavity, but the soft tissue lining the bony nasal cavities or in contact with the cavities could swell and rupture, leading to haemorrhaging. To avoid barotrauma, the respiratory tract must be sufficiently compliant to prevent pressure differentials from developing at any depth to which the animal dives.

The cetacean respiratory tract is composed of upper and lower components, defined by developmental origins. The upper respiratory tract (URT) resides in the skull and is composed of the bony nasal cavities and nasopharynx. Below this, the larynx, trachea, bronchi and lungs form the lower respiratory tract (LRT). The main structural component of the URT is bone, whereas cartilage dominates the LRT. The different structural materials of these regions will determine their response to pressure. ‘Lung collapse’, the compression and reduction of the volume of the lungs and thoracic cavity, requires the relatively compliant tissue of the LRT. Estimates of lung collapse depth in marine mammals range from 70 to 225 m, with the deeper estimates often corresponding to deep-diving pinnipeds (Kooyman, 1973; McDonald and Ponganis, 2012; Moore et al., 2014; Scholander, 1940). There is a wealth of knowledge on how the compressible LRT responds to pressures at depth, but no equivalent information on how the URT responds, or what happens to the URT once the LRT reaches lung collapse and is no longer capable of deforming under pressure.

We considered two scenarios for how the URT and LRT might work together at depth to accommodate pressure and volume changes to avoid barotrauma in the URT. The first follows the ideas of Scholander (1940) – that the entire respiratory tract remains connected, allowing air to be forced rostrally into the bony nasal cavities during a dive. The second integrates the movement of laryngeal cartilages during sound production and the larynx's potential to compress under pressure, effectively dividing the respiratory tract into its anatomically defined upper and lower regions. All of the morphological measurements provided here are representative of an average 19 m long adult fin whale.

Scholander (1940) stated that during a dive, air from the lungs is forced rostrally into the rigid airways and nasal cavities, with the air ‘safely stored’ in these rigid spaces until surfacing. Our underlying assumption of the term ‘safely stored’ is not only that the air has a place to go once lung compression occurs but also that the air fills any incompressible space and prevents barotrauma. This assumption follows from the graded compliance of the respiratory tract whereby the distal-most structures (alveoli) are most compliant and lead proximally to decreasingly compliant structures (bronchi, trachea, larynx) and eventually the rigid nasal cavities (Bagnoli et al., 2011; Davenport et al., 2013; Scholander, 1940). With depth, as a pressure differential begins to form between the respiratory tract and P_amb, the thorax (including the diaphragm) responds first and forces air rostrally. The response of the respiratory tract essentially follows Boyle's law – air volumes in compressible spaces decrease as ambient pressure increases up to the point where further tissue deformation is no longer possible.

We found the total volume of the two nasal cavities combined was ∼32 l. Half of this (∼16 l) is occupied by the nasal plugs. The other half (∼16 l) is air space below the nasal plugs. Total lung capacity of an average adult fin whale is around 1440 l (Scholander, 1940). Assuming the air in the semi-rigid trachea is largely diminished (Bostrom et al., 2008; Cozzi et al., 2005), a fin whale could dive to ∼900 m before the volume of respiratory tract air compresses down to 16 l. Feeding depths for fin whales average 300 m (Croll et al., 2001), and the deepest recorded dive is 470 m (Panigada et al., 1999). Therefore, in this scenario, a fin whale can compensate for a pressure differential at all depths encountered while diving, and barotrauma would not occur.

Scholander's assumption does not account for the action of the larynx in rorquals. The larynx acts as a valve for breathing and has been established as the sound-producing organ in rorquals. Sound production requires movement of the laryngeal cartilages and inflation/deflation of the laryngeal sac (Aroyan et al., 2000; Damien et al., 2019; Gandilhon et al., 2015; Reidenberg and Laitman, 2007, 2008). These movements can isolate portions of the respiratory tract. When the blowholes and larynx are closed, the respiratory tract is functionally separated into two sections – the lungs and trachea form one section, and the laryngeal sac, larynx and nasal cavities form another (Damien et al., 2019; Gandilhon et al., 2015). The laryngeal sac can also be pushed upward and into the caudal portion of the larynx and rostral portion of the trachea, and this physically separates the URT and LRT (Reidenberg and Laitman, 2008). If separation occurs during a dive, it would prevent any further shift of air into the URT, and alter the pressures in the now isolated URT.

Separation of the respiratory tract by the larynx introduces a potentially damaging situation where the compressible parts of the respiratory tract (i.e. lungs, bronchi, trachea) are no longer connected to the incompressible parts (i.e. nasal cavities). In this scenario, we assume that the larynx collapses when the lungs collapse and, therefore, lung collapse dictates the separation depth. Prior to separation, the respiratory tract response follows Scholander (1940), with the graded compliance accounting for pressure and volume changes that prevent the development of pressure differentials (Fig. 9). After separation, we are concerned only with the rigid URT. If no URT tissue can deform, air volume and hence internal pressure will remain at the value at separation, and a pressure differential between P_amb and nasal cavity pressure (P_nasal) will form and grow progressively with greater depth. Although ambient ocean pressure pushes in on the nasal plugs at all depths up to the point of separation, the internal pressure has been maintained at ambient by thorax deformation. Thus, up to the point of separation, there is no pressure differential across the plugs. Only after separation does a pressure differential form across the nasal plugs if depth continues to increase. Ambient ocean pressure pushes on the muscular region of the nasal plugs in the rostrum, perpendicular to the surface, forcing the plugs into the only available space – the nasal cavities. Below separation depths, the nasal plugs are the last remaining compliant tissues in the URT that can respond to P_amb. Movement of the nasal plugs decreases the volume of air space in the nasal cavities to equilibrate pressure between P_amb and P_nasal. From our assessment, the fatty caudal region of the plugs can be easily forced deeper into the nasal cavities to account for half of the air space. Thus, for our estimate of 16 l total air space in an average adult fin whale skull, perhaps 8 l could be accounted for by the forced inward movement of the plugs. The remaining space under the nasal plugs would be 8 l. Inward movement of the nasal plugs raises the internal pressure within the nasal cavities and prevents the formation of a significant pressure differential up to the point where they can effectively move no farther. Assuming a lung collapse depth of 150 m, with the nasal plugs responding to pressure beyond this depth, a fin whale could dive to 300 m without pressure differential formation. This places a fin whale within normal feeding depths (Croll et al., 2001). Importantly, the depth at which the nasal plugs can help alleviate pressure differentials depends on the depth of lung collapse. Our determination that the nasal plugs can decrease air space volume by half means that, from Boyle's law, they can double the internal pressure; this allows the nasal plugs to manage pressure in the nasal cavities between the collapse depth and double the collapse depth (150 m and 300 m in this scenario). Lung and larynx collapse depth is thus an important variable for determining the maximum depth to which the nasal plugs contribute to pressure management.

We do not know the depth at which the nasal plugs stop moving, but it appears likely that fin whales can descend below that limit. Other anatomical structures must compensate for the increase in pressure required to reach depths below 300 m. Although not observed in our own study, swelling of vascular tissue (such as retia) in the nasal cavities could decrease air volumes, as in the Southern right whale (Buono et al., 2015). Beyond 300 m, the remaining air volumes are small (8 l) and require minimal further volume changes to effect large changes in pressure (Fig. 9B).

Although the term ‘nasal plug’ is used to describe anatomical structures in other cetaceans, these plugs do not share the same morphology or proposed barotrauma function as the rorqual nasal plugs. Odontocetes possess a single nasal plug directly ventral to the singular blowhole opening and dorsal to the external nares (Maust-Mohl et al., 2019). The odontocete nasal plug resides completely above the skull with no extensions into the nasal cavities. It is contiguous with the dorsal surface of the premaxillary air sacs used in sound production (Reidenberg and Laitman, 2008). The position of the plug implies it could block the dorsal-most portion of the fleshy nasal passage. Southern right whales possess a bulbous protrusion in each vestibular lumen termed the nasal plug (Buono et al., 2015), but the plugs do not extend into the bony nasal cavities. Bowhead whales possess a valvular mass that arises rostrolaterally from the vestibular walls and passively occludes the blowholes by relaxing medially (Henry, 1983). It is unclear how deep these valvular masses extend as the nasal cavity was not examined. Nasal plug anatomy was described for a fin whale calf and minke whale calf (Maust-Mohl et al., 2019); however, the descriptions were based on older tissue that was detached from the skull, which may have led to an incomplete assessment of the size and composition of the nasal plugs.

Nasal plugs are a unique innovation in rorqual whales, and we argue that the plugs are the true valves protecting the respiratory tract both from water incursion and from barotrauma. Protecting the respiratory tract requires a compromise between two opposing actions – opening the respiratory tract at the surface and filling it when diving. These functions are made possible by the plugs' heterogeneous composition – muscle at the rostral end, fat in the part that fills the nasal cavity, and elastin and collagen in the tendon that attaches each plug to the nasal septum (Fig. 8). The deep attachments of the nasal plugs in the nasal cavities – seemingly confounding when considering the function of the plugs at the surface, is explained by their function at depth. Our study demonstrates the necessity of compliance in nasal plug tissue to protect rorqual nasal cavities from water entry and barotrauma in a high-pressure environment. The nasal plugs demonstrate their importance and novelty during diving, where pressure becomes as important an issue as the danger of water entry.

The authors thank Kristján Loftsson and the staff at Hvalur hf, Iceland, and Daníel Halldórsson at the Marine and Freshwater Research Institute, Reykjavík, Iceland. We also thank Jeremy Goldbogen for his contributions, and Duke Marine Robotics and Ved Chirayath for their permission to use photographs and videos. We thank Stephen Raverty for helping to obtain permits.