Aptenodytes patagonicus, Miller, 1778

Peculiar traits of A. patagonicus View in CoL somatic development

Among modern penguins, A. patagonicus has a unique developmental pattern, with two growth phases separated by a fasting period during the austral winter. The formation of a LAG (already documented in the femur of one king penguin individual by Castanet 2006) during ontogeny is probably unique to A. patagonicus among modern Spheniscidae (although a LAG has been observed in the outer circumferential layer of a Spheniscus individual; see Ksepka et al. 2015). Even A. forsteri , the closest relative to A. patagonicus , shows a continuous growth during ontogeny without any interruption ( Stonehouse 1953, 1960). The formation of LAG(s) is rare in Neornithes and has only been observed in a handful of species, because in most modern birds, skeletal growth is continuous and somatic maturity reached within less than 1 year ( Chinsamy and Elzanowski 2001). Only a few extinct insular and/or giant flightless birds showed protracted skeletal growth over several years, associated with a series of LAGs in their limb bone cortices ( Turvey et al. 2005, Chinsamy et al. 2020, 2023, Canoville et al. 2022). The extant kiwis ( Apteryx ) grow slower than most relatives, and over several years, thus exhibiting LAGs in their cortices ( Bourdon et al. 2009, Heck and Woodward 2021). Recent studies reported on the presence of LAGs in the cortices of some extant ratites ( Canoville et al. 2022). In most of these cases, the formation of LAGs is associated with a cyclical and yearly stop in growth during the unfavourable season until attainment of somatic maturity (e.g. Köhler et al. 2012). In A. patagonicus , however, the stop in growth is driven by the restricted feeding period of the austral winter ( Barrat 1976, Castanet 2006).

Osteosclerosis, and adaptation to aquatic life in A. patagonicus

The main limb bones of A. patagonicus (humerus, radius/ulna, femur, tibiotarsus) are very compact in adult individuals when compared to other flying or strictly terrestrial birds (e.g. Habib and Ruff 2008, Canoville et al. 2022). Two major mechanisms leading to osteosclerosis in clades adapted to an aquatic or semi-aquatic lifestyle are currently documented (reviews in Ricqlès and Buffrénil 2001, Houssaye and Buffrénil 2021). In some groups the sequence of endochondral ossification during skeletal development is incomplete: the medullary region of the bones retains large amounts of calcified cartilage and a medullary cavity does not differentiate. This situation is best exemplified by pachypleurosaurids ( Hugi et al. 2011), sirenians ( Buffrénil et al. 2010), archaeocetes ( Buffrénil et al. 1990, Gray et al. 2007), and mesosaurs ( Klein et al. 2019). Conversely, in other tetrapods (e.g. some sauropsids, mammals, etc.), the sequence of endochondral ossification is complete, with a total resorption of calcified cartilage and an initial differentiation of the medullary cavity. High bone compactness is then achieved by subsequent dense deposits of endosteal bone tissue, which fill up the medullary territory, in part or in totality ( Ricqlès and Buffrénil 2001, Dewaele et al. 2022).

The latter process is at work in king penguins, with a characteristic imbalance during bone remodelling between resorption and re-deposition. Secondary deposits are more abundant than the bone previously eroded, which results in a compaction of the trabecular network and a closure of the medullary region (e.g. Houssaye and Buffrénil 2021). In the long bones of the penguins, the CCCB thus created includes trabeculae with a core of woven tissue (former primary cortical bone tissue), covered by successive layers of parallel-fibred or lamellar bone. In all the adults sampled, the bones most affected by osteosclerosis are stylopodials and, to a lesser extent, zeugopodials. Autopodial elements seem little affected by inner bone compaction.

Inner bone compaction is common among extinct and extant tetrapods incipiently adapted to the aquatic environment or feeding in coastal, shallow waters upon sea grass or sessile preys ( Ricqlès and Buffrénil 2001, Canoville and Laurin 2009, 2010, Houssaye 2009, Buffrénil et al. 2010, Houssaye and Buffrénil 2021). This adaptive convergence has been interpreted as an ecological specialization to optimize sub-aquatic locomotion, body trim, and buoyancy control ( Kaiser 1960, Buffrénil et al. 1990, Taylor 1993, 1994, 2000, Sato et al. 2002, Cook et al. 2010). In groups completely readapted to an aquatic way of life and that further invaded the pelagic environment, a reversion towards a more osteoporotic-like condition occurred in the microstructure of the axial and appendicular skeleton ( Ricqlès and Buffrénil 2001, Canoville et al. 2021).

The earliest Sphenisciformes appear in the fossil record in the Lower Palaeocene ( Slack et al. 2006, Mayr 2022) and are already physiologically and anatomically highly specialized for sub-aquatic locomotion ( Hinic-Frlog and Motani 2010). Comparative data by Ksepka et al. (2015) in the humerus, femur, and tibiotarsus of three Eocene taxa revealed that osteosclerosis was already present in sphenisciform stem lineages, some 25 Myr after the loss of flight. Although penguins are agile and fast swimmers, no reversion process involving an osteoporotic condition ever occurred during their evolutionary history, contrary to what is observed in cetaceans or extinct pelagic marine reptiles ( Buffrénil and Schoevaert 1988, Houssaye et al. 2015, Canoville et al. 2021, Houssaye and Buffrénil 2021). This situation is possibly related to the fact that penguins retained a semi-aquatic lifestyle. Ksepka et al. (2015) also proposed that penguins’ retention of an osteosclerotic condition might be related to the inability of these animals to collapse their lungs, which limits their flexibility for dynamic buoyancy reduction. Whatever the explanation is, the common hydrostatic interpretation of osteosclerosis in tetrapods ( Taylor 1994, 2000, Ricqlès and Buffrénil 2001) is hardly applicable to Spheniscidae , and does not convincingly explain the causes and functional involvement of such a skeletal specialization in this taxon (except that it could contribute to somewhat reduce buoyancy, and therefore facilitate diving). Osteosclerosis is nevertheless by far less extensive in their skeleton than in most other aquatic tetrapods, including sauropsids, mammals, etc. (review in Houssaye and Buffrénil 2021).

Remarks on palaeobiological inferences

The developmental pattern of A. patagonicus is unique, even among modern penguins ( Barrat 1976), a situation that constrains the potential use of this species in palaeobiological inferences relative to Sphenisciformes . However, the present study sets a significant comparative framework for the study of bone microstructural variability in extant and extinct penguins, and for the ontogenetic development of their osteosclerotic limb bones. Fast growth rates of limb bones, with attainment of nearly adult size within a few weeks or months, is adaptive in all penguin species (see Volkman and Trivelpiece 1980 and references therein). Indeed, while at crèche, chicks need to be able to escape predators and/or chase their parents to beg for food. As in A. patagonicus , a decoupling between the growth (in diameter and length) of limb bones and the inner compaction of their shaft (a process that increases skeletal weight and inertia and thus reduces agility), is thus to be expected for all Sphenisciformes . Therefore, limb size is not a good proxy to infer ontogenetic stages in this clade. Similarly, we predict a time offset in the development of hindlimbs and forelimbs in most penguin species because chicks adopt a strictly terrestrial locomotion during their early development, before becoming semi-aquatic after fledging. This observation has been quantified in species other than the king penguin, such as Spheniscus demersus , in which the post-hatching growth of the tarsometatarsus and foot outpace the growth of the flipper during a few weeks ( Cooper 1977). Finally, all penguin chicks go through a juvenile moult before fledging ( Janes 1997, Dégletagne et al. 2013) and are thus prone to form secondary osteons in their limb bone cortices before adulthood.

All these observations allow to critically reassess previous palaeoecological inferences about extinct Sphenisciformes . For instance, Cerda et al. (2015) investigated the bone microstructure of eight Eocene stem penguin species and interpreted the observed variation in bone cortical thickness as differential degrees of adaptation to the aquatic environment between taxa. However, they used fragmentary tarsometatarsi to draw their palaeoecological inferences. The present study (as well as previous works e.g. Buffrénil and Schoevaert 1989, Canoville et al. 2021) has shown that there is usually a decreasing proximo-distal gradient in bone compactness in the limbs of osteosclerotic species: stylopodial and zeugopodial elements are often more compact than autopodials ( Buffrénil and Schoevaert 1989; the present study) and bear overall a stronger ecological signal ( Kriloff et al. 2008, Canoville and Laurin 2010, Quemeneur et al. 2013). Besides, despite being efficient divers, adult king penguins exhibit a somewhat spongious tarsometatarsus (Supporting Information, Figs S8, S9), with a global compactness that varies along the shaft of this element (see Supporting Information, Fig. S9). Hence, looking at the bone microstructure of tarsometatarsi at different levels of the shafts, as done by Cerda et al. (2015), does not yield directly comparable data. Although we recognize that the tarsometatarsus might be the most diagnostic skeletal element in Sphenisciformes in a taxonomical point of view, the use of its microstructure for palaeoecological inferences at least requires a sampling strategy based exclusively on precise midshaft cross sections. Cerda and colleagues further assumed that all sampled elements belonged to adult individuals, based on linear measurements and the presence of secondary osteons in the cortices ( Cerda et al. 2015). Once again, our investigations reveal that in penguins, the adult size of limb elements is attained early in development even though their microanatomy continues to change drastically until adulthood. Cerda et al. ’s sample might thus include individuals close to adult size but still juvenile, especially because some specimens do not exhibit an outer circumferential layer (a feature observed in somatically mature birds; Ponton et al. 2004). This could further explain the microanatomical variability observed in their sample. Moreover, the bone remodelling observed in their specimens could be the outcome of other factors than individual age, such as the juvenile moult or episodes of physiological stress (e.g. periods of low food intake). As seen in the present study, young penguin individuals already present a high mineral turnover in their limb bones, contrary to what is commonly seen in juvenile individuals of most tetrapods, in which the formation of Haversian tissue is milder, because it mainly results from normal bone repair and maintenance, a process that extends throughout life ( Buffrénil and Quilhac 2021). Finally, the authors might have overlooked the extent of inter-individual variability in their study, since they only sampled one individual per species. In view of all these considerations, their palaeoecological interpretations have to be considered carefully and are not yet properly supported by their data.

In light of our results, similar criticism can be made towards the palaeoecological interpretations of Ksepka et al. (2015). These authors observed important differences in the humeral microstructure between Eocene stem Sphenisciformes and extant penguin species. Indeed, the humeri of the extinct stem taxa show a much more spongious internal structure as compared with extant species, whereas their hindlimb bones (femora and tibiotarsi) are more compact. The authors considered that an ontogenetic explanation is unlikely to account for the difference in humeral bone microstructure in stem penguins and considered that all sampled specimens were most likely adults based on the size of skeletal elements. They, in turn, proposed that the stem penguin humeral microstructure represented an early stage predating the future differentiation of the osteosclerotic state displayed by extant forms. However, their conclusions have to be considered carefully, since their study does not take into account intra-specific variability, as they only sampled one specimen per species. Moreover, our study has shown that (i) most skeletal elements reach a nearly adult size in length and diameter very quickly, although the bone microanatomy and overall compactness still fluctuates during ontogeny. Hence, diameter and length of limb bones are not good proxies for ontogenetic stage in penguins; (ii) hindlimb bones tend to mature before forelimb bones (or at least become more compact more quickly) and this trend should be expected in all penguin species that show a strictly terrestrial lifestyle in chicks; (iii) overall, the humerus and radius/ulna have a more spongious organization than the femur and tibiotarsus throughout ontogeny and at the adult stage; (iv) the global compactness of the humerus and radius/ulna is lower than in the tibiotarsus and femur until adulthood.