Structure and biomechanical properties of the trachea of the striped dolphin Stenella coeruleoalba: Evidence for evolutionary adaptations to diving



This study analyzes the structure and mechanical properties of the trachea of the striped dolphin Stenella coeruleoalba, one of the most common cetacean species. The cetacean trachea is made up of closed or semiclosed cartilaginous rings without a paries membranaceus. Our results indicate that the inner lining of the trachea contains erectile tissue in which several venous lacunae permeate the mucosa. We also observed and described the presence of peripheral neurons containing nitric oxide along the rim of the venous lacunae. Data obtained from compression and tensile tests and comparison with the pig and goat tracheas indicate a higher stiffness and a different, higher breaking point for the dolphin trachea. On the whole, our data suggest that the trachea of the striped dolphin possesses structural properties that allow rapid filling with blood, possibly in relation to dive activities, and also allow modifications due to increased pressure and immediate return to the original shape without risks of permanent bending or rupture, as would happen in a terrestrial mammal. As the organ undergoes intense pressure difference during descent to optimal foraging depth and subsequent rapid ascent to surface, especially in deep dives of hundreds of meters, the specific structural and biomechanical peculiarities of the trachea of the striped dolphin may represent an evolutionary adaptation to life in the water and to diving. © 2005 Wiley-Liss, Inc.

The trachea of mammals is a semirigid tube that allows constant passage of air for gas exchange in the lungs. Shape and number of the cartilaginous rings vary with species. Cetaceans have an extremely hydrodynamic body with no externally recognizable neck (with very few exceptions, such as the beluga whale Delphinapterus leucas). The blow hole that opens on the top of the head leads air through elaborate passageways into a duck bill-shaped larynx (Sukhovskaya and Yablokov, 1979; Reidenberg and Laitman, 1987, 1988; Smith et al., 1999). The special morphology of the cetacean larynx forces air flow to bend at almost 90° to enter the short and large trachea.

Although the general structure of the trachea and bronchial tree of cetaceans is basically the same as other mammals, a few specific features make it unique and peculiar both macro- and microscopically (Fanning and Harrison, 1974; Drabek and Kooyman, 1983; Kida, 1990, 1998; Nakakuki 1994; Endo et al., 1999). The trachea of the striped dolphin Stenella coeruleoalba is composed of semicomplete rings. In fact, the tracheal rings follow a spiral pattern along the axis of the organ, and their free dorsal extremities overlap to form a de facto continuous tunnel. A similar structure has been described also for other delphinids, such as the bottlenose dolphin Tursiops truncatus (Fanning and Harrison, 1974). Therefore, the trachea of dolphins lacks a paries membranaceus and a distinct musculus trachealis. The relative shortness and strength derived from the peculiar disposition of the rings give the whole structure specific architectural characteristics.

We are aware that dolphins dive to depths varying from a few meters below the surface to several hundreds of meters to hunt, depending on the nature of the prey. During this activity, dolphins must withstand significant increases of environmental pressure on the way down, followed by rapid opposite gradients (Hooker and Baird, 2001). The rib cage of the animals gives way, and the thorax literally caves in, with consequent reduction of lung volume to a minimum (Olsen et al., 1969; Ridgway and Howard, 1979; Ponganis et al., 2003). At the same time, blood squeezed out of the thorax is mostly shifted to the huge rete mirabiles of the vertebral column and spinal cord, from where originate the vertebral vessels that supply the brain (Snyder, 1983). Several aspects of the diving physiology of dolphins remain unexplained. Our attention has been focused on the trachea, possibly the largest dead space of the dolphin body. Although protected by the base of the skull and the vertebral column dorsally, and by the ribs and the powerful musculature of the neck and thorax on the other sides, external pressure effects on the wall of the trachea must be of the same magnitude as those directed on the lungs. Although a few studies have analyzed the structure of the trachea in diving mammals, to the best of our knowledge, no study has ever been performed on its mechanical properties. We focused our attention on the trachea of one of the most common marine mammals in the world, the striped dolphin Stenella coeruleoalba, aiming to understand what is different in the general architecture of the organ if compared to terrestrial mammals, and how eventual structural differences may help sustain increased pressures at depth.

For this purpose, we analyzed the biomechanical behavior of the isolated trachea of dolphins when subjected to classical mechanical structural tests (Fung, 1993), as described in the literature by other authors who performed tests on tracheal cartilage (Lambert et al., 1991; Rains et al., 1992) or smooth muscle (Wang et al., 2000) of human beings (Ogawa, 1959; Meyers et al., 1980; Begis et al., 1988) or terrestrial animals (Penn et al., 1989; Codd et al., 1994). To check for possible differences, we also tested tracheae of terrestrial mammals such as the pig and the goat and compared the data with the ones obtained by testing dolphin tracheal samples. Comparison with the pig allows a general evaluation of the biomechanical properties of the trachea of a land mammal, while comparison with the goat has been performed because the structure of the trachea of this species closely resembles the human (Barone, 1981), with a wide gap between the free dorsal extremities of the cartilaginous rings and therefore a large dorsal paries membranaceous. As well known worldwide, cetaceans are protected almost everywhere, so we relied on stranded animals for tissue sampling.


Animals and Tissue Sampling: Striped Dolphins

For the present investigation, tracheas from specimens of striped dolphin Stenella coeruleoalba were obtained from the Mediterranean marine mammal tissue bank located at the University of Padua in Italy. The larynx, the trachea, and the first bronchial generation were excised from each dolphin.

The organs were sampled from wild cetaceans that died shortly after stranding along the Italian coasts of the Mediterranean Sea, notwithstanding veterinary medical aid. Causes of death as determined at necropsy never involved pathologies of the trachea. The organs were sampled and immersion-fixed (four animals; three adult and one juvenile) or frozen for mechanical tests (six animals; two adult and four juvenile). The trachea of one animal was used for postmortem endoscopy prior to immersion fixation. The time interval occurring between death and tissue sampling varied between 2 to 36 hr after death.

Samples for mechanical tests were not fixed to avoid alterations of the tissue structure and to preserve mechanical properties of the tissues (Bradley and Wilkes, 1977; Dobrin, 1996; Wilke et al., 1996). After sampling, the airways were immersed in a sterile buffered ionic solution specific for organ transplant (EuroCollins, Monico, Venice, Italy) (Adam et al., 1996; Adham et al., 1996). Once immersed in the solution, the tracheas were rapidly frozen and maintained at −20°C until the moment of the experiment (Goh et al., 1989; Adham et al., 1996). Samples utilized for histological investigations were immersion-fixed in 4% paraformaldehyde in phosphate buffer 0.1 M at pH 7.6.

Animals and Tissue Sampling: Pigs and Goats

Tracheas from seven pigs were sampled at a commercial slaughterhouse in northeastern Italy. Animals were young adults of approximately 100 kg of weight. The tracheas were excised immediately after slaughtering and washed and immersed in EuroCollins solution. Three organs were immediately used for the experiment, while four organs were treated as the dolphin tracheas by freezing in EuroCollins solution at −20°C to allow comparison with the fresh ones. Tracheas from two adult crossbred goats were sampled at a small local abattoir and treated as the dolphin tracheas by freezing in EuroCollins solution at −20°C.

Conventional Histology

For conventional histology, immersion-fixed samples of tracheas from striped dolphins were dehydrated, embedded in paraffin, and cut in 4 μm thick sections, then processed and stained with either haematoxylin-eosin, Azan, or Mallory's trichrome stains.


Immunohistochemical analyses were performed with streptavidin-biotin technique following a validated protocol (Cozzi et al., 1994). Shortly after rinsing in 0.01 M PBS (pH 7.2–7.4), 4 μm thick sections were incubated in 1% hydrogen peroxide in 0.01 M phosphate-buffered saline solution (PBS) for 10 min to inhibit endogenous peroxide activity, preincubated in 4% normal swine serum in PBS with 1% bovine serum albumin (BSA fraction V Sigma A4503) and 0.4% Triton-X100 for 20 min at room temperature (RT) to lower unspecific tissue reactions, and then incubated with primary policlonal anti-NOS antisera raised in rabbit (EuroDiagnostica, Arnhem, The Netherlands) diluted 1:500 in a solution of PBS with 20% sodium azide, 1% BSA, 0.4% Triton-X100, and 1% normal serum. Incubation was carried out for 1 hr at RT and then at 4°C overnight. After washing in PBS with 0.25% BSA and 0.02% Triton-X100, the sections were incubated for 1 hr in biotinylated swine antirabbit immunoglobulins (Dakopatts E353, Copenhagen) diluted 1:200 in PBS with 0.02% Triton-X100. The sections were then washed in PBS and 0.25% BSA and incubated for 1 hr in streptavidin-biotin-horseradish peroxidase complex (Dakopatts K377) diluted 1:125 in PBS and 1% BSA. After washing in PBS and in 0.05 M Tris-HCl (pH 7.6; Tris), peroxidase activity was revealed with a solution containing 0.04% diaminobenzidine (DAB) and 0.03% hydrogen peroxide in Tris for 20 min. After some brief sequential washes, the sections were dehydrated and coverslipped with balsam.

Specificity controls were performed by absorption of 1 ml diluted antiserum with 10–100 μg immunogen and by replacing the primary antibody with normal swine serum. Both procedures abolished the staining.


Endoscopy was performed on a whole trachea and lungs of one specimen of Stenella coeruleoalba using a standard 10 mm 0° angle-of-view laparoscopic endoscope (Storz, Tuttlingen, Germany) attached to a 3 ccd endoscopic camera (Storz). The images were recorded on a videotape.

Mechanical Tests

Biomechanical tests were performed on whole tracheas and first bronchial generation of six Stenella coeruleoalba (named A to F in rostrocaudal direction) and on whole tracheas of seven pigs (named A to G) and two goats (named A and B). Circular samples made of two to four cartilaginous rings were cut from the airway segments (Fig. 1a) and numbered from the larynx to the carina with increasing order. The number of specimens from each animal depended on the length and conditions of the airways. After preconditioning the samples, tensile and compression tests were performed by a Synergie 200 MTS Axial machine (100 N or 1,000 N load cell; 950 mm span; 2003 MTS Systems, Eden Prairie, MN) in order to obtain stress-strain relationship (Fung, 1993; Pennati, 2001; Quaglini et al., 2002). Stress (σ) and strain (ε) were obtained from the cross-head force (F) and displacement (ΔL) measured by the test machine as σ = F/A and ε = ΔL/L0, where A is the area over which the force is applied and L0 is the original length of the sample.

Figure 1.

a: Scheme of the morphometrical parameters of a cylindrical sample. Pictures of a circular specimen on the grips during tensile test: (b) axial view; (c) lateral view.

Immediately before the mechanical tests, the characteristic morphometrical dimensions of each sample were measured (Fig. 1a): thickness (t), height in the direction perpendicular to the applied load (h), and length in the load direction (L0c for the compression tests and L0t for the tensile tests). Average and standard deviation (SD) values of these three morphometrical parameters were t = 1.14 ± 0.38 mm, h = 6.21 ± 0.99 mm, L0c = 16.8 ± 1.31 mm, and L0t = 17.07 ± 2.92 mm for the tracheal samples of neonatal animals; and t = 2.38 ± 0.25 mm, h = 16.25 ± 6.45 mm, L0c = 29 ± 0.01 mm, and L0t = 33.17 ± 2.36 mm for adult dolphins. The same values were measured for the first bronchial generation (neonates: t = 0.97 ± 0.08 mm, h = 7.21 ± 1.29 mm, L0c = 12.75 ± 0.05 mm, L0t = 14.66 ± 2.52 mm; adults: t = 1.25 ± 0.27 mm, h = 11.17 ± 5.04 mm, L0c = 22.25 ± 1.06 mm, L0t = 24 ± 0.82 mm). The loaded area of each sample was evaluated as A = 2 × t × h (Fig. 1a).

Compression tests.

During nondestructive compression tests, each specimen was placed between two steel plates fixed at the test machine grips, with the axis perpendicular to the compression force direction. The specimens were compressed with a 3 mm/min strain rate until complete occlusion. Cyclic loading was applied on some samples to quantify tissue recovery capacity after complete occlusion. Photographs of the sample were taken during tests and the stress-strain curves were acquired.

Tensile tests.

The circular specimens were placed on two U-shaped 2 mm diameter steel wires, which were fixed at the test machine grips (Fig. 1b and c). Each sample was preconditioned, then a circumferential tensile load was applied; the test was terminated when the specimen reached breaking point.

Preconditioning was mandatory as biological tissues are viscoelastic and some loading cycles are necessary to obtain repeatable results and to ensure that no further viscous effects affect mechanical behavior (Fung, 1993). At least six preconditioning cycles were performed on each sample at 20 mm/min strain rate, loading the sample until a maximum load (Fmax), then unloading to zero strain. The maximum load value (Fmax = 15 N) for tracheal samples was chosen based on the results of a preliminary tensile test on a sample from dolphin A trachea. This tensile test, performed without preconditioning the sample, made it possible to evaluate the magnitude of the sample ultimate stress, that is, the stress at which sample starts to break (Smax; Fig. 2). The bronchial specimens were preconditioned at a lower maximum load (Fmax = 10 N) to curtail or avoid breaking risks during preconditioning. Tensile tests with 50 mm/min strain rate were applied on each sample recording the stress-strain relationship until sample rupture.

Figure 2.

Characteristic mechanical parameters evaluated from the stress-strain curve. Left: Tensile test. Ultimate stress (Smax) and slope of the stress-strain curves in the linear segments (Et1, Et2). Right: Compression test. Slope of the stress-strain curves in the linear segment (Ec).

Similar tensile and compression tests were performed on tracheae of seven pigs and two goats to highlight if the specific structure of the cetacean trachea may influence its mechanical behavior. On the whole, 81 specimens from pig tracheae and 36 specimens from goat tracheae were tested.

The characteristic morphometrical dimensions of each sample were measured before mechanical tests; average and SD values of the thickness of the rings for goats and pigs were, respectively, 1.9 ± 0.10 mm and 2.57 ± 0.35 mm, while average and SD height of the samples were 11.87 ± 0.63 mm and 19.05 ± 2.06 mm. For goats and pig, the averaged and SD lengths of the specimens, measured in the load direction, were, respectively, 20.63 ± 1.24 mm and 26.50 ± 3.24 mm for the compression tests and 16.67 ± 3.65 mm and 22.60 ± 2.8 mm for the tensile tests.

Statistical Analyses

The ultimate stress (Smax) and the slope of the stress-strain curves in the linear segments (Et1, Et2, Ec) were derived (Fig. 2) from the stress-strain curve for each specimen (Fung, 1993), and a statistical analysis was performed on Smax and Et2 parameters.

Comparisons among specimens that belonged to the same trachea and were analyzed for both compression and tensile tests or for tensile tests only were performed for each animal using F-tests and two-sided Student's t-tests. ANOVA was used to compare striped dolphins, pigs, and goats. A value of P < 0.05 was taken as a significant difference.


Three of the pig tracheae (A, B, C) were tested immediately after animal sacrifice while the other pig tracheas (four) were immersed in EuroCollins solution and frozen at −20°C until the moment of the experiment to evaluate if the freezing treatment, mandatory for the stranded dead dolphins, may damage tracheal tissues. The goat tracheae were tested after freezing at −20°C to reproduce the preservation and storage conditions of the dolphin samples as mentioned above.


External Appearance

The general shape of the trachea in cetaceans differs from that of terrestrial mammals (Fig. 3). The first cartilaginous rings appear closely packed and complete, without free extremities, while the more caudal rings appear open dorsally and interdigitated with the adjoining ones. The depression that can be easily observed in the dorsal plane of the trachea of terrestrial mammals, corresponding to the paries membranaceus, was substituted by a shallow curvature, so that the whole structure appeared elliptical in section, making it still possible to distinguish the ventral from the dorsal plane at first sight even if the larynx was completely removed from the body.

Figure 3.

Whole trachea and bronchi of Stenella coeruleoalba.

Endoscopic Observations

Endoscopic examination revealed spiral ridges protruding into the lumen of the entire trachea and bronchi. Our images showed that the distance between adjacent ridges diminished as the camera moved from the laryngeal to the bronchial extremity of the organ.

Microscopic Observations

The structure of the dolphin trachea reveals nothing peculiar in the outer connective lining. The cartilaginous structure of the rings is similar to that of terrestrial mammals, if not for the shape of the ring itself, as described above (Fig. 4A). We observed no musculus trachealis. The submucosa revealed the presence of a huge venous vascular plexuses (Fig. 4B–D). Several venous lacunae were scattered all along the inner surface of the cartilaginous rings, lined and interspersed by connective tissue. The mucosa revealed the typical respiratory epithelium.

Figure 4.

A: Section of trachea enclosed in paraffin. B: Section of the trachea stained with hematoxylin-eosin, showing presence and distribution of vascular lacunae (black arrows). C and D: Details of the vascular lacunae in the submucosa of the trachea: (C) hematoxylin-eosin; (D) Mallory's trichrome stain. Scale bars = 1 cm (B); 200 μm (C and D).


Immunostaining with anti-NOS antibodies revealed positive neurons along the endothelial lining of some lacunae and in the interstitial connective spaces between adjacent lacunae (Fig. 5). Nitric oxide-immunoreactive neurons showed finely positive granular material in the cytoplasm, sometimes extending into the slender neurites. Negative controls failed to identify positive cells of any sort.

Figure 5.

Black arrows indicate nitric oxide-containing neurons (intramural ganglion) in the submucosa of the trachea. Asterisks are located in vascular lacunae. Scale bar = 100 μm.

Mechanical Tests

Compression tests.

Stress-strain curves obtained during compression tests on D dolphin specimens are shown in Figure 6a as an example of the results we obtained. All the curves, recorded during compression tests, display a similar trend, although with slightly different values of stress at a given strain. Stress increases slowly with increasing strain until the sample inner surfaces do not reach contact; when lumen occlusion is reached, sample stiffness increases abruptly. The values of the slope (Ec) of the linear segment of the σ-ε curve for each dolphin specimen are summarized in Table 1.

Figure 6.

Stress-strain relationship during compression tests. a: Curves recorded for D dolphin specimens. b: Curves recorded during three subsequent compression tests on E4 sample. Photographs taken during compression test on C2 sample: (c1) 0 N load; (c2) 0.005 N load; (c3) 0.01 N load; (c4) 0.02 N load; (c5) 0.03 N load.

Table 1. Characteristic mechanical data for each tested dolphin specimen*
SampleAnatomical specimenSmax (MPa)Et1 (MPa)Et2 (MPa)Ec (MPa)
  • *

    Specimen name (each dolphin was named by an alphabet from A to F); anatomical site; ultimate stress (Smax); slope of the stress-strain curves in the linear segments (Et1, Et2, Ec); test applied on each specimen (t = tensile test, c = compression test).

A3Left bronchus1.8460.02510.800 
A4Right bronchus2.0040.01412.539 
C1Left bronchus1.1410.03215.5040.029
C3Right bronchus3.1970.04022.6920.0123
D3Left bronchus2.0790.02514.0810.020
D4Right bronchus0.7810.0347.5770.026
D5Right bronchus1.9860.09716.269 
E1Trachea   0.044
E4Right bronchus   0.051
E5Right bronchus3.3180.20327.607 
E6Right bronchus4.1360.20329.260 
E7Right bronchus   0.050
E8Right bronchus3.5110.15421.551 
E9Right bronchus3.5690.21326.408 
 Mean ± SDTrachea2.444 ± 0.7710.065 ± 0.05813.389 ± 4.2300.044 ± 0.019
 Mean ± SDLeft bronchus2.813 ± 1.1130.120 ± 0.08420.488 ± 7.7120.035 ± 0.019
 Mean ± SDRight bronchus1.689 ± 0.4880.027 ± 0.00413.462 ± 2.4120.024 ± 0.006

Data dispersion mainly depends on the interindividual anatomical variability of the tracheal samples. However, additional dispersion factor due to little geometrical differences among the cut samples, due to unavoidable cutting tolerances, cannot be excluded. Pictures taken during the compression tests (Fig. 6c) show the evolution of C2 tracheal lumen occlusion.

Cyclic compression test were performed on some samples from D and E dolphins to quantify recovery capacity after complete occlusion. The curves obtained for sample E4 relating to three subsequent cycles are shown in Figure 6b. The stress-strain curve representing the second compression cycle is quite superimposed to the first compression curve, witnessing structural recovery after complete occlusion, due to elastic properties of biological tissues. Similar results were obtained for each sample used in the multiple compression experiments.

Tensile tests.

An example of the preconditioning cycle curve morphology is shown in Figure 7a with reference to sample D1. The histeresis loop shrinks and tightens at increasing number of cycles up to complete superimposition of the loops, indicating complete sample preconditioning. Curves with the same morphology were obtained for each sample.

Figure 7.

Stress-strain relationship during tensile tests: (a) preconditioning test on D1 sample; (b) stress-strain curves recorded during tensile tests for A dolphin specimens.

The stress-strain curves acquired, after preconditioning, during tensile tests from specimens from dolphin A are shown in Figure 7b. The data are rather dispersed but show the same trend as it was reported for the outcomes of the compression tests. The curve shows increasing structural stiffness of the tissue with increasing strain until ultimate stress is reached. Then the sample fibers reach progressive breaking, causing a sharp reduction of the stress recorded.

Statistical analysis showed that samples that underwent both compression and tensile tests and samples that underwent tensile tests only did not show significant statistical difference. Therefore, the differences among some of the curves should most probably be ascribed to the anatomical and geometrical variability between the couples of samples compared and not to the loading sequence.

Statistical Analyses on Biomechanical Tests

Mean and the SD values for the characteristic parameters evaluated for each tensile σ-ε curves are summarized in Table 1 for dolphins and Table 2 for terrestrial mammals. The statistical analysis performed on the pig samples shows that the fresh and frozen samples are not significantly statistically different; thus, freezing caused no significant damage to the tracheal tissue or at least to its biomechanical properties.

Table 2. Characteristic mechanical parameters for goat and pig tracheae*
SpeciesSmax (MPa)Et1 (MPa)Et2 (MPa)Ec (MPa)
  • *

    Values are mean ± SD; specimen name (each trachea was named by an alphabet from A to B for goats and from A to G for pigs), ultimate stress (Smax); slope of the stress-strain tensile and compression curves in the linear segments (Et1, Et2, Ec).

Goat A1.04 ± 0.420.06 ± 0.044.67 ± 1.620.08 ± 0.01
Goat B1.34 ± 0.450.01 ± 0.016.06 ± 1.590.06 ± 0.02
Pig A0.93 ± 0.430.07 ± 0.044.57 ± 1.610.12 ± 0.05
Pig B0.75 ± 0.150.10 ± 0.023.80 ± 0.570.51 ± 0.58
Pig C1.07 ± 0.440.07 ± 0.045.33 ± 1.270.09 ± 0.06
Pig D0.69 ± 0.350.02 ± 0.013.58 ± 0.920.04 ± 0.02
Pig E0.84 ± 0.320.02 ± 0.014.19 ± 1.110.03 ± 0.01
Pig F0.53 ± 0.160.05 ± 0.022.08 ± 0.780.08 ± 0.00
Pig G0.56 ± 0.220.03 ± 0.022.58 ± 0.730.06 ± 0.03

The statistical analysis performed on Smax and Et2 shows that all three different species (pigs, goats, and striped dolphins) belong to different populations. A comparison between mean values of the biomechanical parameters of dolphins and terrestrial mammals is summarized in Table 3, while Figure 8a shows the σ-ε tensile curve for all the tested tracheal samples and Figure 8b summarizes the analyses of the compression tests results.

Table 3. Characteristic parameters of the σ-ε curves for terrestrial mammals and dolphins*
SpeciesSmax (MPa)Et1 (MPa)Et2 (MPa)Ec (MPa)
  • *

    Values are mean ± SD; Smax is the ultimate tensile stress; Et1 and Et2 are the slope of the stress-strain tensile curve in the linear segments; Ec is the slope of the stress-strain compression curve in the linear segment.

Pigs0.768 ± 0.3570.052 ± 0.0393.735 ± 1.4330.133 ± 0.227
Goats1.189 ± 0.4570.035 ± 0.0345.364 ± 1.5910.069 ± 0.019
Terrestrial mammals0.861 ± 0.4420.048 ± 0.0384.097 ± 1.6480.119 ± 0.202
Striped dolphins2.444 ± 0.7710.065 ± 0.05813.389 ± 4.2300.044 ± 0.019
Figure 8.

σ-ε tensile curves (a1) and Smax and Et2 values (a2) for all the tested tracheal samples: dolphins in blue, goats in red, pigs in yellow. σ-ε compression curves (b1) and Ec values (b2) for all the tested tracheal samples (same colors). Standard deviation bars are shown.


Our morphological observations suggest that the trachea of the striped dolphin possesses a more rigid anatomical structure than that of land mammals. The presence of closed or semiclosed cartilaginous rings leads to a general increase in the rigidity of the trachea. However, our study shows also biomechanical evidences of a greater stiffness of the dolphin samples, as all the curves show higher slope and higher Smax than in the terrestrial mammals (Fig. 8a). This mechanical behavior is due to cartilaginous rings that are semicomplete in the dorsal region and do not present a soft paries membranaceus as the terrestrial mammals do. Thus, during tensile test, the mechanical load is distributed over the whole ring, while in the rings belonging to terrestrial mammals, the greatest deformation is directed toward the soft tissues. In specimens of pig and goat tracheas, rupture always takes place on the paries membranaceus that is the softer and less stiff structure, while in dolphin specimens, rupture took place in no preferential position. Differences between pig and goat tracheas may be due to their different anatomy, the goat tracheas having a wider paries membranaceus like in humans. In any case, differences between goats and pigs are smaller than those between dolphins and both terrestrial species.

These tests show that the anatomical structure of the rings provides them with a stiffer mechanical behavior, but they do not show whether if this greater stiffness can be sufficient to keep the trachea open during diving. Analyses of the compression test results, shown in Figure 8b, seem to contrast this assumption: the σ-ε curves for the terrestrial and marine mammals totally overlap. This result, even if obtained by applying a load different from the hydrostatic uniform pressure load, may suggest that the dolphin tracheal structure, even if stiffer than that of terrestrial mammal, is not absolutely resistant to pressures. Further studies on whole bodies of dead stranded animals may clarify whether it is sufficient to maintain an open lumen at simulated great sea depths. Furthermore, our assessments of the different mechanical stiffness of the cartilaginous and soft structures suggest specific mechanical behavior different from those of terrestrial mammals. We recognize that our compression and tensile strength tests are limited because they only address uniaxial forces. We realize that a live dolphin would be subjected to circumferential pressure changes during diving. Thus, these uniaxial tests, classical engineering tests commonly performed to analyze the mechanical behavior of both artificial and biological structures, were chosen to provide a controlled comparison between the tracheas of marine and terrestrial mammals. Further compliance tests may better evaluate the applied effects of uniform positive or negative hydrostatic pressure and show that the unique structure of the dolphin trachea (e.g., the nearly complete circular cartilaginous support) may provide additional strength to withstand rupture or collapse.

We demonstrated here for the first time the peculiar structural characteristics of the mucosa of the trachea of the striped dolphin Stenella coeruleoalba. The submucosal lacunae correspond to the longitudinal ridges seen during endoscopic observations. These ridges were already described in the trachea of Tursiops truncatus (Fanning and Harrison, 1974). The huge venous submucosal lacunae indicate the presence of a vascular device morphologically similar to that typical of the erectile tissues of the mammalian body. The vascular venous spaces could be filled and voided according to specific physiological necessities of the animal. Although our data allow no direct indication of which mechanisms activate filling or voiding of the vascular spaces, the presence of nitric oxide-containing neurons along the rims of the lacunae further supports the hypothesis of an analogy with the mammalian penile corpora cavernosa, thus suggesting an active nitric oxide-induced mechanism of vascular engorgement.

But what would be the final physiological advantage of erectile tissue in the inner lining of the trachea? Dolphins feed at various depths (Hooker and Baird, 2001). Striped dolphins generally hunt at depths comprised between 5 and 100 meters with short 1- to 2-min apnea, while other common species such as the bottlenose dolphin Tursiops truncatus may prefer shallower waters but may extend breath-holding time to 15 min. Intense foraging activities requires a body structure apt to resist the formidable continuous atmospheric and water pressure of several bars. The general theory for cetacean diving is that as pressure increases during a dive, the lungs are compressed to collapsing point, and air escapes into rigid air spaces (bronchi and trachea) (Ridgway and Howard, 1979; Skrovan et al., 1999; Williams et al., 1999, 2000; Williams, 2001). The airways must be stiff enough to resist transmural pressure, even if at depths reached by striped dolphins the gas in the respiratory tract is reduced in volume to a considerable extent. So if it is surprising that the cetacean trachea could remain a dead space, with large quantities of precious air left to no use, a denser and larger vascular network in the submucosa may play a key functional role. Space seclusion may be meaningful in the lower respiratory tract, as in the bronchial tree of dolphins, where several cartilaginous rings even in the smaller ramifications create a structure able to divide the lower airways into tight compartments where air is locked during descent (Engel, 1966; Ito et al., 1967; Yamasaki et al., 1977; Henry et al., 1983; Henk and Haldiman, 1990; Haldiman et al., 1998). But in the middle ear of marine mammals, where a similar situation takes place, vascular sinuses in the mucosa reduce the volume of the dead space when filled, presumably at depth (Welsch and Reidel-Scheimer, 1997). We believe that the presence of erectile tissue in the subcumosa of the dolphin trachea may represent a diving device per se. Active filling of the submucosal vascular lacunae could lead to a reduction in the diameter of the trachea, with a mechanism specific to resist pressure (blood, as any liquid, cannot be compressed by environmental pressure) and curtail waste space within the respiratory system. Furthermore, blood filling of the vascular lacunae and decrease of the internal diameter of the organ may accommodate for reduction of gas volume at depth and prevent leakage of blood from vessels into the airway space due to pressure gradients, a hypothesis considered for the airways of sperm whales Physeter macrocephalus (Leith, 1989). On the other hand, congestion of vascular lacunae may contribute to rigidity and assist in maintaining constant shape and diameter of the organ. The presence of nitric oxide-containing neurons points to an active on-off mechanism of vascular filling. Alternative hypothesis, based on the elastic properties shown by the organ and by the presence of neuroactive substances able to induce venous repletion, may also suggest the possibility that the trachea becomes flattened under pressure and then eventually snaps back to its former shape due to its high elastic properties and with the important aid of blood rapidly entering the vascular lacunae during ascent.

Our evidences suggest that the trachea of dolphin contains huge submucosal vascular spaces that can be filled either to reduce the inner volume of the organ while maintaining an external increased rigidity or help the trachea elastically regain its former shape when external pressure decreases. Since the structure of the dolphin trachea seems very different from that of terrestrial mammals, possible explanations for its morphology must obviously include intermittently increased environmental pressure at depth. Our data indicate that the structure of the dolphin trachea may possess a physiological mechanism that can be explained on the basis of evolutionary adaptations to deep diving.


The authors thank Dr. Lara Papini and Dr. Fiorenza Anfuso for help in sampling tissues, Dr. Giustina Casagrande for assistance with statistical analysis of mechanical test data, and Mr. Antonio Bafunno, Mr. Giovanni Caporale, and Mr. Emanuele Zanetti for technical assistance. Supported by University of Padua Grant 60A08-4245/03 (to B.C.).