The complex, multicameral lungs of the Nile crocodile are characterized by rows of tubular chambers, which in cranial and ventral lung regions are broad and sac-like. The inner surface of the chambers is enhanced by cubicles (ediculae), the capillary-bearing walls of which are often perforated. Extrabronchial communication among chambers is infrequent. The ediculae end in a network of myoelastic trabeculae, which face the central lumen of the chambers. The trabecular epithelium is similar to that of mammalian bronchi and contains isolated endocrine-hke cells basally, whereas the edicular epithelium is similar to that of other reptiles and of mammals. The distribution of non-vascular smooth muscle, 64% in trabeculae and 36% in interedicular walls, is consistent with the hypothesis that these two antagonistically oriented muscle groups interact to effect lung patency. The volume-specific lung compliance is similar to that of much simpler, unicameral gekko lungs, implying that lung compliance is a function of parenchymal structure and not of primary structural type.
Numerous studies have dealt with the morphology of the epithelium in the respiratory partitions of reptilian lungs (Okada et al. 1964; Meban, 1977, 1978a,b; Welsch, 1979; Klemm, Gatz, Westfall & Fedde, 1979; Luchtel & Kardong, 1981; Perry, 1972, 1978, 1983; Pohunková & Hughes, 1985). Little attention, however, has been given to those structures which bear upon the mechanical properties of such lungs and allow the respiratory surfaces to remain exposed to ventilatory air movement during and between breaths (Ogawa, 1920; Klemm et al. 1979; Perry, 1983). This applies particularly to crocodilian lungs, for which only general, gross anatomical or developmental information is available (Milani, 1897; Broman, 1939; Duncker, 1978b).
The present paper describes the gross and fine anatomy in one of the most complex structural types of reptilian lung with respect to its dynamic function, and presents hypotheses for the coupling of structure and function with respect to the support of respiratory surfaces during breathing.
MATERIALS AND METHODS
The lungs of two zoo-born, juvenile Nile crocodiles, Crocodylus niloticus Laurenti (mass = 763 g and 325 g), which had recently died were fixed in situ by inflation with 10% formalin, removed intact and air dried from ethanol or xylol under constant intratracheal pressure with compressed air. The dried lungs were opened laterally and dissected by removal of lung tissue with forceps to reveal the lung chambers and their connection to the intrapulmonary bronchus. The results were recorded as stereo-pair photographs and as drawings. Using these data, a wire model of a lung was constructed, which served as a basis for Fig. 1.
Microscopic anatomy and lung mechanics
Four Nile crocodiles of body mass W = 3·38 kg, 3·59 kg, 3·69 kg and 5·68 kg were obtained through legal, commercial channels and maintained in aqua-terraria on a diet consisting mainly of fish. They were killed by an intraperitoneally administered overdose of sodium pentobarbital (60-120 mg kg−1). Immediately after death, 50 ml of blood was removed for other studies and substituted with an anticoagulant-vaso-dilating solution of 2 ml of heparin (5000 units ml−1), 2·5 ml of nitroprusside sodium (2·4 % in 0·9 % NaCl), 39·5 ml of dextran (0·96g in 0·9 % NaCl) and 6 ml of distilled water. Before the body was opened, two volume—pressure diagrams were obtained using the method of Perry & Duncker (1978, 1980) to determine lung volumes. Residual lung volume was determined as the amount of air which could be withdrawn from the lungs of the anaesthetized, supine animal after equilibration to atmospheric pressure without exceeding an intratracheal pressure of — 10cmH2O (1 cmH20 = 98·1 Pa). Maximal lung volume (VLM) was the amount of air required to raise the intratracheal pressure from —10 to +20cmH2O. These arbitrary limits were set to avoid damaging the lungs for subsequent electron microscopic examination.
The body cavity was opened ventrally and two volume-pressure diagrams were obtained for the exposed lungs of three specimens. The body cavity was then sutured and the lungs were fixed in situ by intratracheal instillation of 0·75 VLM of cold (4°C) glutaraldehyde (3 % glutaraldehyde in 0·05 mol 1−1 phosphate buffer, pH 6-8). After 3 h of fixation, the lungs were removed intact and their length and displacement volume determined according to the method of W’eibel (1970/71). Each lung was then cut transversely into 10—12 slices. Samples for transmission electron mi-croscopy were taken from centrally (proximally) and peripherally (distally) located sites in either the right or the left lung. These samples were maintained overnight in cold (4°C) phosphate buffer (pH7 0, 350mosmolI−1) for postfixation in 1% 0·1 moll−1 phosphate-buffered OsO4 at 0°C, alcohol dehydration and Epon embedding the following day. The remaining tissue was postfixed in Bouin’s fluid, dehydrated through graded alcohols and embedded in paraffin wax. These preparations were planed using a sliding microtome and the blocks were exposed to alcoholic methylene blue, revealing the cut surfaces of the embedded tissue (Perry, 1981a).
Morphometric evaluation of the paraffin preparations was carried out with the aid of a dissecting microscope and using stereological methods for multicameral lungs (Perry, 1981b). Volumetric analysis of tissue components employing stereological point- and intersection-counting techniques (Weibel, 1979) was performed on photomontages of toluidine-blue-stained, semi-thin sections at a final magnification of 625 X.
The volumes of all tissue elements were calculated separately for proximal and distal sampling sites of each lung slice. The sums of these values, calculated for both lungs of each test animal, are presented in Table 1.
Electron micrographs were obtained from uranyl acetate, lead citrate contrasted thin sections using a Zeiss EM 9 or Zeiss EM 109 electron microscope.
Symbols and definitions
The hierarchy and definition of symbols used in the designation of anatomical structures are as follows. The symbols are used as subscripts.
L, lung; L consists of
NP, tissue-free central lumen of lung or of lung chambers; the central lumen is arbitrarily separated from the parenchyma by a line connecting the larger trabeculae.
P, parenchyma: that portion of the lung in which air spaces are surrounded by partitions and their associated trabeculae; P consists of
Nt, air spaces.
t, tissue component; t consists of:
B, large blood vessels.
A, septa: wall-like, flattened structures connecting trabeculae (NA) with the general inner surface of the lung or of intercameral septa. The central leaflet consists of a collagenous matrix in which non-vascular smooth muscle, blood vessels and nerves are found. A network of capillaries in which gas exchange occurs covers the surface; there are separate networks on opposing surfaces. A consists of c, m, e and u (see below).
NA, trabeculae: structures supporting free edges of septa. They are composed of a central core of smooth muscle and elastic tissue. Large trabeculae possess a non-respiratory, ciliated epithelium facing the central lumen (NP) and a respiratory epithelium on the abluminal surface. NA and A consist of
c, content of capillaries on partitions or trabeculae;
m, non-vascular smooth muscle in partitions or trabeculae;
e, non-respiratory epithelium;
u, other components (not c, m or e) of partitions or trabeculae.
The lungs, together with the heart (including pericardium and the great vessels) and the oesophagus occupy the cranial half of the body cavity. Each lung lies in a separate, closed pleural space, bordered laterally by the dorsal ribs and intercostal musculature, dorsally by the vertebral column, medially by the mediastinum, and ventrally by the sternal ribs. The lungs, particularly in the dorsal half, tend to fuse with the parietal pleural surface.
Right and left lungs are mirror images of each other reflected in the mediosagittal body plane. Each represents approximately a truncated cone, with its base lying against the liver and its cranial apex extending ventrally into the base of the neck, between the shoulder girdle and the oesophagus.
As in other crocodilians (Duncker, 1978b) the lungs are multicameral, consisting of a variable number of chambers (camera), each of which connects independently with an unbranched, intrapulmonary bronchus (Fig. 1). The cranial half of the intrapulmonary bronchus (solid outline in Fig. 1) is cartilage reinforced. It displays three rows of orifices which supply four dorsal chambers, four lateral chambers and three ventral chambers (Fig. 1). A medial row is lacking. The rows of orifices do not run parallel to the long axis of the intrapulmonary bronchus, but instead form a left-hand spiral in the left lung (right-hand in the right lung). Hence the first lateral chamber opens into the lateral aspect of the intrapulmonary bronchus, but the orifice of the fourth lateral chamber is dorsal (Fig. 1).
The first dorsal chamber is rudimentary. The remaining chambers in the dorsal row are long, tubular structures. Together with a cluster of medial chambers which originate at the end of the cartilage-reinforced portion of the intrapulmonary bronchus, they form a dorsomediai lung lobe (Fig. 1).
The cavernous first lateral chamber extends cranially from its bronchial entrance and, together with its large ventral and lateral branches, forms the lung apex (Fig. 1). Lateral chambers 2-4 make up the deep lateral and superficial dorsal portions of the lung along the middle third of its length.
In the middle third of the lung the three ventral chambers comprise the sac-like ventrolateral and ventromedial lung regions. The first two ventral chambers then double back dorsally and form the superficial lateral portion of the lung (Fig. 1).
The caudal half of the intrapulmonary bronchus lacks cartilage and its chambers are not continuous with the rows described above. In addition to the cluster of medial chambers mentioned above, five long, tubular chambers and the sac-like terminal chamber supply the superficial portions, while a large number of irregularly distributed outgrowths (bronchial niches) supply the deep lung parts (Fig. 1).
The inner surface of the chambers is elaborated with a system of cubicles (ediculae; Duncker, 1978b). The free edges of the ediculae face the central lumen of the chambers and are supported by a system of stout trabeculae, as described for amphibian and reptilian lungs (Goniakowska-Witalinska, 1986; Gräper, 1931). The interedicular septa, which bear the respiratory capillaries, are often perforated, whereas the intercameral septa, which separate adjacent chambers, rarely show such perforations.
The interedicular partitions in the Nile crocodile consist of a central leaflet of collagenous connective tissue and smooth muscle. They bear on each surface a separate network of capillaries which rarely communicate across the central leaflet (Fig. 2).
The trabecular epithelium is composed primarily of ciliated cells and serous secretory cells (Figs 2, 3). On the ciliated cells, short microvilli outnumber the kinocilia by 7 to 1 and at the base of the microvilli horseshoe-shaped invaginations of the plasmalemma are observed (Fig. 3). The cilia display prominent, striated rootlets; the longer, terminal rootlet extending at approximately 45° from the tip of the basal body and the shorter rootlet inserting perpendicularly on the middle of the basal body (Figs 3, 6).
The ciliated cells and secretory cells are joined at their apical surface by a junctional complex (Fig. 3) and exhibit laterally interdigitating processes. The microvilli and mitochondria of these two cell types are similar but more numerous in the ciliated cells. The serous secretory cells, however, are packed with large granules of moderate electron density (Fig. 3). Scattered glycogen deposits are observed in both cell types.
Endocrine-like cells (elc) are often found singly or in small groups at the base of the trabecular epithelium, where they form desmosomes with the ciliated cells (Fig. 6). Typical of thee/cs are dense-core vesicles (Fig. 7), large, oval mitochondria and abundant smooth-surfaced endoplasmic reticulum. The often dilated cisternae of the latter give these cells a lacey, bright appearance (Figs 6, 7). elcs have not been observed to contact the free trabecular surface in the Nile crocodile.
Beneath the basement lamina of the trabecular epithelium, capillaries and nerves are found in the connective tissue, which envelops the ‘myoelastic’ core of the trabecula (Fig. 2). The smooth muscle of the core (Fig. 4) is characterized by prominent fusiform densities. Amorphous elastic tissue deposits between the smooth muscle cells are embedded in a fine, fibrous or electron-lucent, multivesicular matrix (Fig. 5).
The connective tissue sheath of the trabecula is continuous with the central leaflet of the interedicular partition. It supports a double capillary net (Fig. 2) and contains smooth muscle (Fig. 2), which is ultrastructurally indistinguishable from that of the trabecula. The interedicular muscle bundles are oriented perpendicular to the trabecula (Fig. 8). At the base of the trabecula a subtrabecular vein and an accompanying perivascular lymph space are often present (not illustrated).
The results of the volumetric analysis of the lungs are summarized in Table 1. The average pulmonary displacement volume (VL), which is the reference volume for morphometric calculations, is 109 ml kg−1. In the three smaller specimens (W = 3·6kg) VL per kg body mass is greater than in the larger, 5·7-kg specimen. Also, the smaller animals tend to have a greater parenchymal volume in relation to body mass (48 ml kg−1) than the larger one (39 rnl kg ‘). Other parameters are similar for small and large specimens.
The volume of lung tissue is only approximately 4 ml kg−1. The remainder (105 ml kg−1) is air, of which 63 ml kg−1 is the central lumina of the chambers and the intrapulmonary bronchus, with the remaining 42 ml kg−1 in the parenchymal air spaces (ediculae).
The interedicular septa make up some 68% of the total tissue volume. The remaining lung tissue consists of trabeculae and large blood vessels (13 % and 19%, respectively). The major components of the septa are connective tissue (57 %) and blood (36%). This represents 93% of the connective tissue and 96% of the blood (exclusive of major vessels) in the lung.
Non-vascular smooth muscle makes up 7 % of the interedicular septa but 55 % of the trabeculae. However, since the absolute volume of the septa is five times that of the trabeculae, 36% of the total non-vascular pulmonary smooth muscle is in the interedicular septa (for specimen 3; 47 %). The largest specimen also has the lowest percentage of muscle in the septa: 24%.
The regional distribution of components of the septa is shown in Fig. 9. This indicates that only the cranial third of the lung (which consists primarily of the sac-like, first lateral chamber) appears to differ from the rest. Here blood makes up between 20 and 33 % of the total volume compared with 38-50 % in the caudal two-thirds.
In general, sampling sites distal to the intrapulmonary bronchus tend to yield relatively more blood and often less smooth muscle than proximal sites. The proportion of connective tissue is also greater distally than proximally. These differences tend to be most pronounced cranially (see Fig. 9), and are most clearly seen when the (perhaps variable) blood content is subtracted.
Lung volumes and mechanics
The compliance of the lungs and body wall together is 1·23 ml cmH20−1 100 g−1 (±0·26S.D.) or, standardized against VLr, 0·71 ml cmH20−1 ml−1 (±0·06s.D.)-The lungs, with a value of 7·39mlcmH2O−1 100g−1 (±2·06S.D.) are more than four times as compliant as the body wall. The VLr-standardized lung compliance is 4·32mlcmH2O−1ml1 (±1·18s.D.).
Studies of development and comparative anatomy of crocodilian lungs carried out during the nineteenth century (for references see Milani, 1897; Broman, 1939) tend to be fragmentary, but stress the basic similarity of lung structure in alligators, caymans and crocodiles. Only Broman (1939) traced the development of the lungs in a single species (Alligator mississippiensis) through a period sufficient to explain the origin of the chambers. He describes a row of large dorsal chambers which give rise at their bases to medial and lateral subdivisions. These subdivisions later communicate as separate chambers with the intrapulmonary bronchus.
During late foetal development the cranial portion of the lung rotates so that the chambers which were originally dorsal later communicate with the lateral aspect of the intrapulmonary bronchus. These chambers may represent the lateral row in the young Nile crocodile (Fig. 1). The medial and lateral chamber rows of the foetal alligator would then represent the dorsal and ventral rows, respectively, in the present study.
Broman (1939) reported that the caudal portion of the alligator lung develops much more slowly than does the cranial portion, thus explaining the lack of continuity between cranial and caudal rows of chambers. A similar process is expected in the Nile crocodile.
Unfortunately no developmental study similar to that of Broman (1939) exists for the Nile crocodile. Since it is not justifiable to apply nomenclature based upon lung development of an alligator to a crocodile, a provisional nomenclature for chambers based only on the present findings is employed here (see Fig. 1).
The only original gross anatomical work explicitly concerning the lungs of the Nile crocodile is a brief description by Lereboullet (1838), later paraphrased by Cuvier (1840). Both descriptions are included verbatim in Milani, 1897. Lereboullet counted only five chambers: an anterior chamber (LI in Fig. 1), three further chambers (probably D2 + L2 + VI, D3 + L3 + V2, D4 + L4 + V3) and a posterior chamber (the cartilage-free portion of the intrapulmonary’ bronchus and its associated chambers).
The presence of a small number of large chambers cranially is a common feature of most multicameral lungs. Similarly, the reduction of cartilagenous support in the caudal portion of the intrapulmonar)’ bronchus is common not only to crocodilians, but also to monitor lizards and to many chelonians (Milani, 1894, 1897; Gräper, 1931; Kirschfeld, 1970).
Peculiar to the crocodilian lung, however, is the tendency for monopodal — as opposed to dichotomous -branching of the chambers, as well as the tendency for this branching to occur at the bases rather than at the tips of the chambers (Milani, 1897; Broman, 1939). The closest affinity in these respects is seen in the pattern of formation of the secondary bronchi of the avian lung (Locy & Larsell, 1916; Perry, 1987). Further similarities between crocodilian and foetal avian lungs are the small number of cranial chambers (secondary bronchi in birds), their tendency to occur in a spiral row or rows, the lack of a medial row of cranial chambers and the large number of relatively loosely ordered caudal chambers. Furthermore, the tendency of crocodilian lung chambers to form arching, tubular structures is reminiscent of developing avian secondary bronchi and parabronchi (Duncker, 1978a). It is possible to construct an approximation of the avian lung—air-sac system from the crocodilian structural type: sac-like cranial, ventral and caudal chambers become air sacs, dorsal or medial chamber rows arch caudally (with shortening of the proto-avian thorax), their chamber walls deepen to parabronchi which meet terminally in the plane of anastomosis with their counterparts from the caudal lung regions, and the parabronchial lumina become contiguous through perforations. Only the blood-air—capillary net remains exclusively avian.
Although the present data are not suitable for construction of an allometric regression curve, the displacement volume of the lungs in the three small specimens is 15% greater per unit body mass than that of the large specimen . The proportion of total lung volume devoted to parenchyma (38—44 %) is similar in all specimens, suggesting that the fixation conditions were the same and that the tendency towards smaller lungs in larger animals is real. Tenney & Tenney (1970) reported that the lung volume in reptiles in general increases in proportion to W0·75 whereas studies on single species indicate that the exponent of this relationship is closer to 1·0 (Hughes, 1977; Perry, 1978, 1983).
Compared with terrestrial lizards, the Nile crocodile has very small lungs: V Lr =1·8 ml 100g−1, as opposed to 4·4, 6·3, 16·0 and 24·6 ml 100g−1 in the teju, the tokay gekko, the savanna monitor and the European chameleon, respectivelv (Perry & Duncker, 1978; Milsom & Vitalis, 1984). These differences may relate to the possibly species-dependent relevance of VLr to survival.
VLr represents that lung volume at which the elastic contraction of the lungs is in equilibrium with the relaxed condition of the skeletomuscular pulmonary encase-ment. Although this condition is useful as a reproducible standard for comparison of lung volumes in different species, its functional significance is unclear.
Most reptiles - including crocodilians — display an intermittent breathing pattern (Glass & Johansen, 1979; Glass & Wood, 1983; Naifeh, Huggins & Hoff, 1970, 1971a,b,c) in which the breathing phase begins with expiration. VLr could represent a minimal respiratory pause volume, below which an initial expiration is no longer effective. Inspection of Fig. 10 reveals that the equilibrium point from which the static, closed-chest volume-pressure diagram begins is at the end of the linear portion of the deflation curve, and that ’Lr is only a small portion of the total lung volume. The constancy of the Nile crocodile’s lung compliance over long portions of the volume-pressure curve, however, suggests that this species may be capable of breathing over a wide range of residual volume states, depending on the mean buoyancy desired. Thus, although VLr in the Nile crocodile lies well below that of non-crocodilian reptiles, the range of lung volumes available for breathing overlaps with that of many lizards (Perry, 1983).
In spite of marked differences in lung structure in crocodiles and lizards, the static lung compliance in the crocodile (3·9 ml cmH20−1 ml VLr−’) is very similar to that of the tokay gekko and the savanna monitor (4·1 and 3·2—3·3 ml cmH20−1 ml VLr−1, respectively) (Perry & Duncker, 1978; Milsom & Vitalis, 1984). The highly parenchymatous unicameral lungs of the teju and the emerald lizard show a lower compliance (2·4 and l·8mlcmH2O−1 ml VLr−1, respectively) (R. M. Jones, unpublished results; Perry & Duncker, 1978) and the sac-like chameleon lung is more compliant (5·7 mlcmH2O−1 ml VLr−1) (Perry & Duncker, 1978).
The importance of these static compliance values in total work of breathing in these species awaits direct measurement of dynamic compliance in living specimens. Recent studies (Milsom & Vitalis, 1984; Bartlett, Mortola & Doll, 1986) indicate that the major resistance to breathing in reptiles is in the body wall or in the extrapulmonary airways. The body wall compliance in the crocodile is calculated as 1/CB = 1/CLB— 1/CL to be 0·85 ml cmH20−1 ml VLr−1. This is similar to the value reported by Perry & Duncker (1978) for the gekko, but nearly four times greater than values based upon dynamic measurements in the same species (Milsom & Vitalis, 1984). The role of body wall compliance in crocodilian breathing mechanics, however, is uncertain because of their unique breathing mechanism discussed below.
Gans & Clark (1976) reported that in Caiman crocodilus both external and internal intercostal muscles tend to be simultaneously active during breathing. The intercostal musculature thus serves to stiffen the body wall while the m. diaphragmaticus affects inspiration by retracting the liver, thus displacing the lungs caudally. Preliminary data from electromyographic and X-ray cinematographic studies in the Nile crocodile imply a similar breathing mechanism in this species (B. Jutsch, personal communication). Further studies will elucidate details of this breathing mechanism.
Of particular relevance to the present study are the anatomical possibilities for gross lung movement and thus for ventilation of respiratory surfaces. The presence of deep costal impressions, as well as the attachment of the lungs to the parietal pleura, suggests that the lungs do not slide, but rather stretch to accompany liver movement. The stretching of the lungs into the broad, caudal region of their conical, pleural cavities would widen the thin-walled caudal and ventral regions and draw air into the arching distal regions of the tubular chambers, as in the mammalian lung (Loring & de Troyer, 1985).
Air can escape from the chambers during expiration by reversed flow, or possibly through intercameral perforations to neighbouring chambers. Patency of the chamber walls during expiration may be ensured by dynamic interplay of antagonistic smooth muscle in the trabeculae and in the interedicular walls.
In the Nile crocodile 14% of the lung tissue is non-vascular, smooth muscle, of which 64% is located in the trabeculae (see Fig. 9). Contraction of the trabeculae would tend to raise the intrapulmonary pressure and deepen the parenchyma at the expense of the central lumina of the chambers. The remaining 36% of the intrapulmonary smooth muscle is oriented approximately perpendicular to the trabeculae and lies in the central leaflet of the interedicular walls (see Fig. 8). Its contraction would thus lower the parenchyma and widen the central lumina.
The mechanism controlling and coordinating the activity of intrapulmonary smooth muscle remains to be physiologically demonstrated. We were unable to demonstrate direct innervation of the smooth muscle, and recent observations in the red-eared turtle imply the involvement of endocrine-like cells (e/c) in that species (Scheuermann, de Groodt-Lasseel, Stilman & Meisters, 1983; Scheuermann, de Groodt-Lasseel & Stilman, 1984). It is proposed that these cells release indolamines or catecholamines in direct response to hypoxia or to nervous stimulation. In contrast to the red-eared turtle, the elcs of the crocodilian trabecular epithelium do not contact the free surface. In addition, although abundant in the trabecular epithelium, they have not been observed in the interedicular tissue. In the teju lung, in which non-trabecular and trabecular smooth muscle are equally abundant, elcs have been demonstrated only in the equivalent of interedicular tissue (S. F. Perry & U. Aumann, in preparation), whereas in the water snake, Nerodia sipedon, they occur in domed structures on the trabeculae (S. F. Perry, unpublished observations).
It is tempting to suggest that the microvilli-rich, ciliated epithelial cells of the trabeculae may serve in more than lung clearance. The presence of a lateral labyrinth is consistent with the hypothesis that these cells may be active in fluid resorption, as proposed for similar cells in the mammalian airway epithelium (Rhodin, 1966; Wilson, Plopper & Hyde, 1984). elc secretions could be carried with the fluid flow to subepithelial capillaries or across the smooth muscle to subtrabecular lymph spaces.
The author gratefully acknowledges the technical assistance of Iris Zaehle, Sabine Willig and Christine Ivanisi, the dactylographic assistance of Angelika Sievers and the generous financial support of the Deutsche Forschungsgemeinschaft through grants Pe 267/1 and Pe 267/2-3.