The origin of the bird's beak: new insights from dinosaur incubation periods

Abstract

The toothless beak of modern birds was considered as an adaption for feeding ecology; however, several recent studies suggested that developmental factors are also responsible for the toothless beak. Neontological and palaeontological studies have progressively uncovered how birds evolved toothless beaks and suggested that the multiple occurrences of complete edentulism in non-avian dinosaurs were the result of selection for specialized diets. Although developmental biology and ecological factors are not mutually exclusive, the conventional hypothesis that ecological factors account for the toothless beak appears insufficient. A recent study on dinosaur incubation period using embryonic teeth posited that tooth formation rate limits developmental speed, constraining toothed dinosaur incubation to slow reptilian rates. We suggest that selection for tooth loss was a side effect of selection for fast embryo growth and thus shorter incubation. This observation would also explain the multiple occurrences of tooth loss and beaks in non-avian dinosaur taxa crownward of Tyrannosaurus. Whereas our hypothesis is an observation without any experimental supports, more studies of gene regulation of tooth formation in embryos would allow testing for the trade-off between incubation period and tooth development.

1. Introduction

As often in evolutionary research, a comprehensive perspective that includes life forms only known from the fossil record helps to decide between alternative hypotheses that were established only based on the living survivors. Theropod dinosaurs crownward of Tyrannosaurus evolved smaller body sizes approaching those of birds and varied feeding preferences, particularly in near-bird dinosaurs [1,2]. Some were even herbivorous and beaked in the bird sense (toothless and with a rhamphotheca).

While such beaks are also present in turtles and some extinct reptile clades, birds are the most diverse and arguably most successful beaked amniote clade. The beak is one of many other characters (prime among them are feathers [3,4]) once considered unique to birds and a pivotal innovation that enabled birds to survive through the Cretaceous–Paleogene (K–Pg) extinction event [5]. The origin of the beak has been ascribed to both genetic (developmental economy) and ecological (food acquisition and flight) causes in the recent literature [1,6–13]. Based on a recent elegant study of dinosaur embryonic teeth [14], we hypothesise that incubation period is a key selection pressure for toothlessness.

2. The conventional hypothesis for the origin of edentulism and the beak

Numerous toothless non-avian dinosaurs and toothed Mesozoic birds (e.g. Hesperornis, Ichthyornis) reveal that edentulism evolved several times in dinosaurs (figure 1). Zhou et al. [15] assumed that birds lost their teeth for a reduction in body weight and active flight; however, this hypothesis is invalidated by numerous discoveries of airborne toothed Mesozoic birds such as pelagornithids and Ichthyornis. The conventional hypothesis for the origin of birds' toothless beak attributes this feature to ecological adaptations [16], and a surge of recent studies appears to provide support for this [5,9,17,18]. For instance, Cooney et al. [9] showed that ecological adaption led to the explosive diversification of birds' beaks after the mass extinction at the end of the Cretaceous, 66 million years ago. However, the K–Pg mass extinction event also wiped out some edentulous Mesozoic birds, indicating that ecological factors cannot be the only explanation [5]. Alternatively, edentulism and the beak have been hypothesised to strengthen the skull of some non-avian theropods and birds [1], but hypotheses rooted in development also have to be considered. The morphometric studies of Bright et al. [8] on living raptors and Felice & Goswami [13] indicated that bird beak shape is strongly coupled with body size and controlled by nondietary factors such as specific genes active during development. They also hypothesised that non-avian dinosaurs and birds faced similar constraints on beak shape and broke a developmental constraint to vary their beak shape.

Figure 1.

Toothlessness in dinosaurs. Consensus cladogram of Archosauria shows that edentulism frequently evolved independently in several lineages. All Ornithischia, Therizinosauroidea, Incisivosaurus and toothed Enantiornithes exhibit partial tooth loss. Ontogenetic edentulism has been reported for the theropod Limusaurus from China [10]. Complete edentulism is observed in some pterosaurs, Ornithomimosauria, Caenagnathoidea (including Oviraptoridae), some Mesozoic birds such as Confuciusornis and Gobipteryx, and all extant birds (Neornithes). Silhouettes were taken from public domain images on phylopic.org.

3. Developmental evidence from genetics and fossils

Numerous studies from developmental biology have proposed various genetic mechanisms to explain edentulism. Chen et al. [19] provided evidence for edentulism involving a mutation that inactivated the genetic pathways for tooth formation in deep evolutionary history. In fact, birds still retain some intact signalling pathways for tooth formation, consistent with the existence of tooth primordia [19]. Chen et al. [19]'s developmental biology experiments suggested that the absence of Bmp4 contributes to edentulism in the chicken. In another study, Harris et al. [20] hypothesised that the loss of direct apposition between an epithelial signalling centre and the mesenchyme is responsible for avian edentulism.

Combining evidence from fossils and experimental developmental biology revealed the genetic regulation of edentulism in modern birds and provided a potential genetic mechanism for its evolution [6,7]. A developmental explanation for the origin of edentulism was further supported by three very recent studies based on fossils: one on the Jurassic ceratosaurian theropod Limusaurus [10] and the other two on a caenagnathid oviraptorosaur [11,12] and the Early Cretaceous bird Sapeornis [11]. Limusaurus gradually lost teeth with age, indicating that its edentulism evolved through a shift toward earlier cessation of tooth replacement, which can be explained by heterochrony, specifically progenesis (figure 2) [10]. The studies on the caenagnathid and Sapeornis supported this interpretation of heterochrony [11,12].

Figure 2.

Schematic illustration of the hypothetical four stages of the evolution of edentulism. Stage 1, neither edentulism nor heterochronic truncation of tooth formation. Stage 2, ontogenetic edentulism, heterochronic truncation of tooth replacement occurs after the eruption of functional generations of teeth. Stage 3, complete edentulism, heterochronic truncation occurs at the cessation of embryonic tooth formation. Stage 4, complete edentulism, heterochronic truncation occurs at the onset of embryonic tooth formation, as indicated by the existence of tooth primordia in modern birds [19].

4. Incubation period as a selective factor for edentulism

Support for the importance of heterochrony in beak evolution now comes from an unexpected line of evidence: reproductive biology of non-avian dinosaurs. Once believed to be inaccessible in fossils, data on dinosaur incubation periods have become available from an ever-increasing sample of dinosaur embryos preserved at different stages of development. This unexpected line of evidence is brought to bear in an elegant new study based on such fossils by Erickson et al. [14].

All dinosaurs hatched from eggs, and incubation period is one of the crucial traits in reproductive biology of egg-laying (oviparous) amniotes. Incubation time is of great evolutionary significance [14,21,22]. Fast incubation reduces the risk of egg mortality from predation, diseases and natural disasters. In the past, dinosaur incubation periods were estimated using bird models as 45 to 82 days, but Erickson et al. [14] now provide direct evidence to the contrary. This evidence is von Ebner lines (daily incremental lines in tooth dentin) in embryonic teeth of two groups of ornithischian dinosaurs, ceratopsians and hadrosaurids. The von Ebner line counts in Late Cretaceous Protoceratops embryos from Mongolia and Hypacrosaurus embryos from Canada indicate that these dinosaurs had the long incubation periods of three to six months typical of modern reptiles.

Since there is a great selective advantage to fast incubation and to fast development in general, this raises the question of what limited incubation period in these ceratopsians and hadrosaurs to slow reptilian rates and how this limitation was overcome in derived theropods, i.e. on the line to modern birds. Erickson et al. [14] hypothesised the rate of tooth dentin formation in the embryo to be the limiting factor. They note that teeth do not appear earlier than around 40% through incubation time, and that tooth formation rate is constrained by the maximum thickness of 30 µm of the daily growth increments in the dentin (von Ebner lines). The authors posit accordingly that all toothed dinosaurs would have been restricted to long reptilian incubation periods.

This hypothesis is consistent with the observation that most dinosaurs built reptile-like nests by burying their eggs and providing little parental care [23]. Only in dinosaurs crownward of Tyrannosaurus is there evidence for the evolution of open nests and brooding, which would have made faster incubation possible than in buried eggs (figure 1) [23,24]. Evidence for brooding goes along with evidence for endothermy such as fast growth [25] and insulation [3]. If Erickson et al. [14] are correct, selection pressure for fast incubation would favour tooth loss and would explain the multiple evolution of beaks in non-avian dinosaurs crownward of Tyrannosaurus. The selection pressure for fast incubation also acted in early birds (figure 1) and is consistent with the slow development of the toothed urvogel Archaeopteryx [4].

However, independently of dinosaurs and modern birds, edentulism and a horny beak evolved in several other amniote lineages, such as turtles, some marine reptiles and pterosaurs. Turtles retain a long incubation period of approximately 42–112 days [24], which seemingly contradicts our hypothesis that edentulism results from selection for a shorter incubation period. This contradiction is also presented by megapode birds that have long incubation times and bury their eggs, yet it is generally agreed that megapode nesting biology is secondarily derived instead of being inherited from non-avian dinosaurs [26]. Given that the evolution of edentulism is a complex process, these contradictions suggest the influence of other constraints, some of which may be phylogenetic. This is hinted at by the new study of Baron et al. [27], which challenges dogma and re-divides dinosaurs into Ornithoscelida and Saurischia, doing away with the canonical Saurischia–Ornithischia dichotomy. As newly defined, Saurischia no longer includes theropods, but the Theropoda (excluding Herrerasauridae) and Ornithischia form a clade Ornithoscelida. All Ornithischia have beaks, but always in combination with teeth, as do some theropods. Among dinosaurs, beaks would be thus further restricted to a subclade of Dinosauria, the Ornithoscelida. Alternatively, the ability to evolve a beak may characterize a much more inclusive clade because some derived pterosaurs possessed edentulous beaks and it was recently suggested that sauropod dinosaurs also had a type of beak [28].

5. Potential tests of hypotheses of developmental economy

Reproductive biological factors, as suggested by our hypothesis, represent an improvement in developmental economy for the evolution of edentulism called for in an earlier review [29]. A preliminary report by Varricchio et al. [30] on Troodon embryonic teeth inferred an incubation period of 74 days, which is much shorter than the 113.93 days predicted for a reptilian egg of the same mass but significantly longer than the 49.01 days for a bird egg of the same mass [24]. Such an intermediate incubation period, in between the slow reptilian and fast avian modes, was possibly determined by the interplay of tooth formation and extensive parental care [31], which lends support to our hypothesis. It is hence crucial to investigate if edentulous fossil taxa exhibited accelerated incubation and if toothed basal birds had long incubation periods.

Apart from allometric analyses [21,22], counting growth marks in embryonic teeth is currently the only method for estimating incubation periods in fossil taxa [14,30]; however, this method obviously cannot be applied to those edentulous fossil taxa. Experimental evo-devo research, such as inhibiting specific genes responsible for tooth formation, may inform on the influence of toothlessness on incubation periods [6,7,19,20].

In addition to ecological factors, our hypothesis explains the observation of multiple appearances of edentulism and highlights a new selective advantage of edentulism, i.e. improving developmental economy and accelerating incubation periods. Developmental and reproductive biology thus play an important role in influencing many features such as the beak and feathers that were previously explained by feeding preferences and flight, respectively.

Authors' contributions

The Opinion piece was designed, written and drafted by both authors. Both authors approved the final version of this manuscript.

Competing interests

The authors declare no competing interests.

Funding

This study was funded by the Jurassic Foundation.