1051 Evidence for Endothermic Ancestors of Crocodiles at the Stem of Archosaur Evolution Roger S. Seymour1 Introduction Christina L. Bennett-Stamper2 The distinction between endothermy and ectothermy is one of Sonya D. Johnston1 the central characteristics that divides vertebrate animals. Al- David R. Carrier3 though there is variability in the timing, precision, and level of Gordon C. Grigg4 the endothermic state among birds and mammals, both groups 1 Department of Environmental Biology, University of are clearly distinguished from the majority of ectothermic rep- 2 Adelaide, Adelaide, South Australia 5005, Australia; Center tiles, amphibians, and fish. Birds and mammals typically have for Biological Microscopy, University of Cincinnati, 3125 considerably higher body temperatures, metabolic rates, and Eden Avenue, P.O. Box 670521, Cincinnati, Ohio 45267- stamina. Their body temperatures are usually physiologically reg- 3 0521; Department of Biology, University of Utah, Salt Lake ulated within a narrow range, and their metabolic pathways are 4 City, Utah 84112; Department of Zoology, University of generally aerobic during rest and activity. Ectotherms are char- Queensland, Brisbane, Queensland 4072, Australia acterised by environmentally dependent body temperatures, low metabolic rates, and reliance on anaerobic pathways for intense Accepted 11/19/04 but unsustainable activity. Probably because of the perceived di- chotomy between living endotherms and ectotherms, the ques- tion of the metabolic status of extinct vertebrate groups has been extensively debated (Thomas and Olson 1980; Bakker 1986; Far- ABSTRACT low et al. 1995; Feduccia 1999). Physiological, anatomical, and developmental features of the One line of evidence concerns the relationship between en- crocodilian heart support the paleontological evidence that the dothermy, metabolic rate, and a four-chambered heart. Loris ancestors of living crocodilians were active and endothermic, S. Russell (1965), and possibly others before him, recognised but the lineage reverted to ectothermy when it invaded the that the four-chambered heart of archosaurs perfectly separates aquatic, ambush predator niche. In endotherms, there is a func- oxygenated and deoxygenated bloods, which should optimise tional nexus between high metabolic rates, high blood flow gas transport by sending only deoxygenated blood to the lung rates, and complete separation of high systemic blood pressure and oxygenated blood to the body. Webb (1979) also drew from low pulmonary blood pressure in a four-chambered heart. attention to the significance of the four-chambered heart and Ectotherms generally lack all of these characteristics, but croc- elevated activity in the ancestors of birds and crocodiles. The odilians retain a four-chambered heart. However, crocodilians inference is that higher metabolic rates of endotherms would have a neurally controlled, pulmonary bypass shunt that is select for the perfect separation of bloods. This argument was functional in diving. Shunting occurs outside of the heart and used in connection with the discovery of the disputed fossil involves the left aortic arch that originates from the right ven- four-chambered heart in an ornithischian dinosaur (Fisher et tricle, the foramen of Panizza between the left and right aortic al. 2000). However, despite an anatomically incomplete sepa- arches, and the cog-tooth valve at the base of the pulmonary ration of the chambers of noncrocodilian reptilian hearts, some artery. Developmental studies show that all of these uniquely snakes and lizards are able to reduce mixing to almost zero (Hicks 1998). This shows that a four-chambered heart is not crocodilian features are secondarily derived, indicating a shift necessary merely to separate bloods. from the complete separation of blood flow of endotherms to We believe that the primary role of the four-chambered heart the controlled shunting of ectotherms. We present other evi- is to separate systemic and pulmonary blood pressures, rather dence for endothermy in stem archosaurs and suggest that some than just blood oxygenation states. To achieve high metabolic dinosaurs may have inherited the trait. rates, endotherms require a greater cardiovascular oxygen trans- port capacity, which they realise with higher blood flow rates and hemoglobin levels. Elevated cardiac output in endotherms is associated with markedly higher systemic arterial blood pres- Physiological and Biochemical Zoology 77(6):1051–1067. 2004. ᭧ 2004 by The sure (Rodbard et al. 1949; Johansen 1972; Fig. 1). The expla- University of Chicago. All rights reserved. 1522-2152/2004/7706-3066$15.00 nation for their high systemic blood pressure and appreciable This content downloaded from 143.107.245.005 on November 01, 2018 05:43:52 AM All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c). 1052 R. S. Seymour, C. L. Bennett-Stamper, S. D. Johnston, D. R. Carrier, and G. C. Grigg and Blaylock 2000). Heart mass averages about 0.4%–0.7% of body mass in mammals and 0.8%–1.2% in birds (Poupa and Osta´dal 1969; Bishop 1997; Seymour and Blaylock 2000), but it is only 0.19%–0.32% in most reptiles (Poupa and Lindstro¨m 1983; Seymour 1987; Farrell et al. 1998). In alligators, heart mass decreases from 0.25% at 1 kg body mass to 0.15% at 70 kg and 0.125% at 124 kg (Coulson et al. 1989). The level of systemic arterial blood pressure is related not only to metabolic rate and heart size but also to the size of the animal. More specifically, although all of the heart’s energy is ultimately lost to frictional resistance in the circulation, one immediate requirement of central systemic arterial blood pres- sure is to support the vertical blood column above the heart Figure 1. Systemic arterial blood pressures in relation to mass-indepen- dent standard metabolic rate among vertebrate groups. Statistics are (Seymour et al. 1993). This requirement explains why the giraffe means and 95% confidence intervals. In fish, amphibians, reptiles, birds, has an arterial blood pressure about twice the mammalian norm and mammals, blood pressure is based on 23, 6, 25, 12, and 23 species, (Hargens 1987) and why arterial blood pressure increases sig- respectively; metabolic rate is based on 8, 9, 69, 398, and 639 species, nificantly in larger mammals (Seymour and Blaylock 2000) and respectively, from the literature. The range of pulmonary arterial blood longer terrestrial snakes (Seymour 1987). Thus, it is possible pressures of air-breathing vertebrates is shown for comparison (Johansen 1972; Hicks 1998). to calculate the minimum arterial blood pressure of an animal from a skeletal reconstruction, assuming that the heart was in the sternal area and the blood column was the vertical distance peripheral resistance is not entirely clear, but we will present above it. Estimates of systemic arterial blood pressures lie be- possible functional roles below. At this point, however, we need tween 10 and 25 kPa in some ornithopod and theropod di- only to observe the strong correlation. Pulmonary arterial blood nosaurs (Seymour 1976) and possibly higher in some sauropods pressures, on the other hand, remain low in both ectotherms (Seymour and Lillywhite 2000). These measurements show that and endotherms (Fig. 1). In these cases, it is generally accepted the hearts of dinosaurs were capable of producing systemic that high pulmonary blood pressures would cause excess fluid pressures well within the endothermic range, which would have filtration into the air spaces (pulmonary edema) and inhibit required the functional separation of systemic and pulmonary gas exchange (Wang et al. 1998). Although the anatomically blood in a four-chambered heart. However, the analysis cannot undivided hearts of some exceptional reptiles (e.g., monitor determine what led to the evolution of such a heart and high lizards and large terrestrial snakes) can generate apprecia- blood pressure; these features could have evolved initially in ble pressure separation and moderately high pressures (Burg- support of endothermy, large body size, or both. If it can be gren and Johansen 1982; Seymour 1987; Wang et al. 2002, shown that they occurred first in short animals, it would be 2003), most reptiles have systemic systolic pressures that fall consistent with endothermy as the primary correlate. considerably short of the mammalian or avian norms (cf. Hicks This report expands an earlier presentation and provides 1998 for reptiles and Seymour and Blaylock 2000 for birds and evidence that the small ancestors of crocodiles possessed four- mammals). Hearts of noncrocodilian reptiles rely on muscular force to keep opposing walls and ridges pressed together to chambered hearts that were capable of generating high systemic achieve separation during systole; avian and mammalian four- blood pressures consonant with endothermy (Seymour 2001b). chambered hearts do not. Thus, a four-chambered heart is Ironically, the hearts of living ectothermic crocodilians provide the best solution for separating pressures. Despite a four- compelling clues concerning the history of endothermy in the chambered heart and good separation of systemic and pul- group. Although the hearts are relatively small and produce monary pressures, the mean systemic blood pressures of con- systemic blood pressures in the same range as other ectothermic scious crocodilians are well within the ectothermic range: values reptiles, their structure and development testify to endothermic for Alligator mississippiensis