Next time you're cutting up a fresh bird, try looking for the lungs. They're about where you'd expect them to be, but they're nestled up dorsally against the ribs and vertebrae, and they're surprisingly small. If you think about it, the the thorax of a bird is a fairly rigid box, with that large sternal keel up front and short ribs—it's a wonder that they are able to get enough air from those tiny organs with relatively little capability for expanding and contracting the chest.

How they do it is an amazing story. Birds have a radically effective respiratory system that works rather differently than ours, with multiple adaptations working together to improve their ability to take in oxygen. There is also a growing body of evidence that dinosaurs also shared many of these adaptations, tightening their link to birds and also making them potentially even more fierce—they were big, they were active, and their lungs were turbocharged.

Inhalation: inflate lungs and posterior air sacs

Exhalation: empty the anterior air sac, fill lungs with air from the posterior sac

Inhalation: inflate posterior sac with fresh air, fill anterior sac with air from the lung

Exhalation: empty anterior sac, fill lung with air from the posterior sac

I'm going to give a very brief overview of the main properties of the avian lung; if you want more details, I'll recommend this fine class site. But first, a quick review of how your lungs work. Your lungs are a kind of blind sac with one opening; muscles expand your rib cage or depress the diaphragm, decreasing the pressure in the lungs and sucking fresh air into them. The air inflates tiny epithelial balloons called alveoli, which are rich in blood vessels. Your blood extracts the oxygen from the inhaled air and dumps carbon dioxide into the alveoli…and your muscles then compress the lungs, forcing the stale air out. And the cycle is repeated on and on, alternating filling with rich, fresh air and expelling old, stale air.

Birds have a couple of other features to their respiratory system: air sacs. Many of their bones are perforated and hollow, and contain air-filled membranes connected to the lungs and trachea. There is a set of anterior air sacs, and another set of posterior sacs, with the lungs located between them. One function, perhaps the primitive function, of these sacs is to simply lighten the structure of the animal—important in flyers, of course, but also useful in large animals that just need to reduce the burden of all that bony weight.

The sacs have also been couple to breathing. Instead of inflating their lungs directly, birds inflate the air sacs first, and then the sacs are compressed like a bellows to drive air through the lungs. There is a fairly complex, continuous cycle of respiration, illustrated in the diagrams to the right. One set of sacs is being inflated by inhalation, and then air is expelled from the sacs through the lungs and to the other set, and then expelled from the sacs to the trachea. It actually takes two breaths to move an intake of air through the complex of sacs and lungs.

So what's the advantage? One is that there isn't any dead air: the lungs aren't ever filled with stale carbon dioxide rich air that needs to be expelled before more can be taken in. Fresh air flows continuously through the lungs. Another is that the air always moves in the same direction; the blue arrows in the diagram show that the flow of air through the lungs is always from left to right. A bird's lungs do not contain the collection of tiny balloons ours do, but instead contains slim tubes that carry out the same function, and that are invested with blood vessels organized to most efficiently extract oxygen. It's very impressive, and I've got to admit, if there were a designer, he should have used this design in us mammals, too. We were robbed.

You can see just how pneumatic birds are in this latex-injected duck—everything that's blue is part of the system of air sacs.

a, Latex injection (blue) of the pulmonary system in a duck (Anas crecca), highlighting the extent of air sacs throughout the body. b, Main components of the avian flow-through system (ribs have been illustrated in their proper anatomical positions). Abd, abdominal air sac; Cdth, caudal thoracic air sac, Cl, clavicular air sac; Crth, cranial thoracic air sac; Cv, cervical air sac; Fu, furcula; Hu, humerus; Lu, lung; Lvd, lateral vertebral diverticula; Pv, pelvis; Tr, trachea.

What you can't quite see in that picture is how pervasive the sacs are. Some of the vertebrae, the ribs, the sternum, and some long bones have openings called pneumatic foramina, and diverticula of the sacs infiltrate right down into the core of the bones. Here on the left, for instance, are some vertebrae from a crane and most of the holes (NaP and CeP) are places where the air sacs slip in.

Comparisons between a bird (a, b) and theropod dinosaur (c, d) in caudal (a, c) and right lateral (b, d) views, illustrating the topological similarity of pneumatic features. a, b, Cranial thoracic vertebra of a sarus crane (Grus antigone, SBU AV104063). c, d, Mid-cervical (c) and cervicothoracic (d) vertebra of an abelisauroid theropod (Majungatholus atopus, UA 8678). Scale bar, 1 cm (a, b) and 3 cm in (c, d). CeP, central pneumatic foramen; NaP, neural arch pneumatic foramen; Nc, neural canal; Ns, neural spine; Pp, parapophysis.

The vertebrae on the right are fossils from a theropod, Majungatholus atopus. Notice any similarity with the crane?

What the investigators did in this study was analyze the location of these foramina in a specimen of Majungatholus and reconstruct the likely position of the air sacs (which were not preserved, unsurprisingly—they would consist of thin membranes in the living animal). What they found is diagrammed below: the animal had both an anterior set of air sacs (in green) and a posterior set (in blue), with the lungs (in orange) between them. In the absence of soft tissues, it is not a conclusive demonstration…but it is very suggestive that the theropods had a flow-through respiratory system like modern birds.

a, Pulmonary components based directly on vertebral morphology in UA 8678 include the following: cervical system (green), lung (orange) and abdominal system (blue). In modern birds, clavicular and thoracic air sacs are also present (see Fig. 1); their placement here (indicated in light grey), combined with the overlapping nature of other components, represents tertiary-level inferences emphasizing the uncertainty surrounding the reconstruction of soft tissues not constrained by osteological evidence. Unknown skeletal elements are indicated by dark grey shading. b−f, Vertebrae (b−e) and rib (f) of UA 8678 illustrating pneumatic features. b, Second cervical; c, first thoracic; d, ninth thoracic (reversed); e, sacral complex, left lateral view; f, left ninth cervical rib, medial view (reversed). CeP, central pneumatic foramen; CoP, costal pneumatic foramen; Cp, capitulum; IvF, intervertebral foramen; NaP, neural arch pneumatic foramen; Ns, neural spine; Pp, parapophysis; Tb, tuberculum.

The observations suggest further links between bird and saurian anatomy and physiology, and also support the idea of high metabolic activity in dinosaurs.

Recent studies of non-avian theropod dinosaurs have documented several features once thought solely to characterize living birds, including the presence of feather-like integumentary specializations, rapid, avian-like growth rates, 28, and even bird-like behaviours captured in the fossil record. Either implicitly or explicitly, these studies have linked anatomical, physiological or behavioural inferences with an increased metabolic potential, suggesting that if not bird-like in metabolism, theropods were at least 'more similar' to birds than to reptiles. Our study indicates that basal neotheropods possessed the anatomical potential for flow-through ventilation of the pulmonary system, emphasizing the early evolution of respiratory adaptations that are consistent with elevated metabolic rates in predatory dinosaurs.

O'Connor PM, Claessens LPAM (2005) Basic avian pulmonary design and flow-through ventilation in non-avian theropod dinosaurs. Nature 436:253-256.