Eggs of extinct dwarf island emus retained large size

1. Introduction

Two dwarf and a smaller subspecies of the mainland emu (Dromaius novaehollandiae) once occurred on islands off southern Australia (figure 1). Relative to the mainland form, with nomenclature following [4], the King Island emu (D. n. minor), extinct by 1822, was around 44–45% smaller in size (physical height and body mass). The Kangaroo Island emu (D. n. baudinianus), extinct ca 1830, was approximately 25% smaller than mainland emu and Tasmanian emu (D. n. diemenensis), extinct ca 1850, was approximately 10% smaller [5–9] (table 1). All were victims of over-hunting by European colonists [9]. Island emus became isolated from the mainland in comparatively recent times after the separation of Tasmania around 14 kya, King Island at 11 kya and Kangaroo Island at 10 kya [12], but the Kangaroo Island emu, or a related population on nearby parts of the mainland, may have been isolated for much longer [1]. Dwarfism appears to have evolved rapidly [2,7], with a direct correlation between the extent of dwarfing and island size [7]: King Island, with an area of 1100 km², had the smallest species, followed by Kangaroo Island (4400 km²) and Tasmania (62 400 km²), respectively. Such was the rapidity of their extinction that the life histories of dwarf emus are virtually unknown; however, François Péron in 1804 provided details of the King Island emu, based on his questioning of a resident sealer, Daniel Cowper [3,13]. Cowper was evidently a careful observer, but not a trained naturalist, so interpretation of these details must be approached with caution. Cowper stated among other things that males were slightly larger than females, weighing 45–50 lb (20.4–22.6 kg), both shared incubation, and the clutch size ranged from 7 to 9. The life history of Kangaroo Island emu is unknown, whereas Tasmanian emu had a clutch size of 8–9 and shared incubation [9,14]. The eggs when fresh of both Tasmanian and Kangaroo Island emu were also finer grained and darker green in colour than those of mainland birds [10,14]. Mainland emu clutch size averages 6.7 per female, but one nest can contain the eggs of several females [15,16]. Eggshell fragments collected in 1917 on Flinders Island, and thought to represent a new taxon [17], are referable to a Late Pleistocene deposit of mainland emu [18].

• Download figure
• Open in new tab
• Download PowerPoint

To test body size reduction in relation to egg size in each of the island emu taxa, in particular the extreme dwarfism in D. n. minor, we obtained measurements of femora to calculate body mass. We used measurements of museum and subfossil intact eggs and subfossil eggshell to ascertain egg size and volume and compared them with eggs of mainland emu.

2. Material and methods

3. Results and discussion

Male mainland emus are smaller than females, with body masses averaging 32 kg and 37 kg for each sex, respectively [11]. Cowper's weight for King Island emu (20.4–22.6 kg) is within or slightly less than our data (20.2–23.3 kg) (table 1); therefore, body masses of mainland (32–37 kg) and King Island emu (20.2–23.3 kg) show a difference of 44.2% or 45.4% (table 1). Tasmanian and Kangaroo Island emus' body masses lie between these two measurements, with Kangaroo Island emu estimated at 23.8–27.4 kg and Tasmanian emu at 30.3–34.2 kg (table 1). Observations of living birds and other skeletal element measurements confirm this decreasing size trend from mainland emu, the largest, through to King Island emu [5,7], the smallest of all emu taxa.

We refute Cowper's statement that King Island emu males were reputedly slightly larger than females and suggest that the heaviest birds weighing 50 pounds (22.6 kg) [13] were female. Cowper most likely mistook brooding birds to be female, as in all other ratites except ostrich, males incubate the eggs and females are larger [11,15]. Mainland emus have an average recorded egg mass of 595 g [11,15,23], which accounts for 1.6% of its body mass. With a body mass of 23.3 kg for females, and the egg mass around 8% less than mainland emu eggs (table 1), a King Island emu egg of approximately 547 g would represent 2.3% of the body mass, yet the bird was around 44% lighter than the female mainland emu. Relative to length, width and volume percentages of mainland birds, eggs of dwarf emus are slightly shorter in greatest length in each taxon (97.7–99.2%), are narrower in greatest width in King Island and Tasmanian emus (90.0–95.7%) and have a more reduced volume in each taxon (86.2–93.8%) (table 1). However, the mean values for the egg measurements of each island emu fall within the size ranges of mainland emu, excepting only that the egg of the King Island bird is slightly narrower (80.0 mm) and has a smaller volume (465 ml) than the smallest mainland emu egg (80.8; 487) at 4.6% difference (table 1). T-test analysis showed a significant difference in volume only between mainland and Tasmanian emu egg size, with a p-value below 0.05 (electronic supplementary material, tables S4, S5 and S6), which supports our results. The eggshell thickness of the historically collected Kangaroo Island and Tasmanian museum eggs is thinner (6.8%/4.8%) than mainland emu, with subfossil eggshell of King and Kangaroo Island birds thinner still (table 2).

Our data thus demonstrate that the eggs of the island emus, despite the smaller dimensions of adults, were within or close to the linear size, and smallest volume and mass ranges of mainland birds (figure 2), and the eggshell was seemingly slightly thinner (table 2), but see below. What is less clear are the reasons why these phenomena occurred.

• Download figure
• Open in new tab
• Download PowerPoint

Emus, along with other ratites, have precocial young (juveniles that are relatively mature and mobile at the point of hatching) [11]. Péron was informed by Cowper that all King Island emus synchronized laying to 2 days of the year, on July 25th–26th, clutch size was large, and that chicks left the nest within 3 days of hatching [13]. Although Cowper's egg-laying observation appears to be an exaggeration, mainland emu egg clutches, although laid sequentially, tend to hatch within two days of each other [23], which probably accounts for the confusion. Mainland emus also lay during the coolest time of the year, generally in July, and chicks leave the nest after 2–3 days of hatching [23–25]. Furthermore, in mainland emus, where body size is large relative to egg size, a single adult can incubate up to 15 eggs [15]. Thus, multiple laying by a number of females into one nest allows accumulation of a large clutch in a short period, as a single female emu can only lay one egg every 2 days [26]. This synchronized laying prevents embryo death from over-heating and exposure, and simultaneous hatching increases survival by dilution effects against predators [24].

Mainland emu eggs hatch after 56 days [10,15], so if the same duration applied to King Island emus, particular environmental conditions or food resources abundant at the time of hatching may have caused synchronized breeding. July has the coolest annual temperature and highest rainfall on King Island [27], but the 56-day period of incubation would see chicks hatch at the start of spring in September. Because dwarfing of island emus occurred within at least the last 14 000 YBP [12], it appears that the evolutionary advantage for retaining a large egg size and precocial chick was primarily driven by limited food resources on their respective island homes. The chicks had to be sufficiently large after hatching to be able to consume the food that was seasonally available independently, and perhaps to develop adequate thermoregulation to enable them to cope with cool external temperatures, as in kiwi Apteryx sp. and mainland emu (see [28]). Although some aspects of ratite life history are r-selected, e.g. large clutch size, retention of large-sized eggs in dwarf emus appears to have been an evolutionary step towards an extreme k-selected life history, e.g. smaller clutch size, large egg, appearing to mirror that of the kiwi. However, the correlation between body mass and egg mass shows kiwi as a distant outlier (see also [29]) (figure 2b); therefore, kiwi egg size is a result of its atypical life history and environmental pressures [11]. Emu hatchlings also have heavier and more residual yolk than many other avian species [30,31], and in studies of other precocial species, an increase in residual yolk correlated with increased egg size, resulting in increased hatchling survival under limited food resource conditions (see [24] and references therein). In Greater Rhea (Rhea americana), reduction in egg size may negatively affect embryo development and post-fledgling survival [24].

If predation was also a driving force for precocial adaptation on the southern islands, Tasmania had two large predators, Thylacine (Thylacinus cyanocephalus) and Tasmanian Devil (Sarcophilus harrisii), while thylacine formerly occurred on Kangaroo Island [7]. King Island reputedly lacked a large, terrestrial carnivore [7,32], but an extremely robust, but now extirpated Tiger Quoll (Dasyurus m. maculatus) was once resident, and formerly considered specifically distinct due to its large size [33]. It was certainly capable of taking emu chicks, as they prey on vertebrates many times larger than themselves [34].

Our data indicate that island emus had thinner eggshell than mainland birds, but this can occur intra-specifically due to a number of factors, including environmental conditions and calcium availability for eggshell development [35–37]. In addition, smaller eggs have thinner eggshells [38], so a 14% decrease in egg volume (table 1) is likely to result in a similar level of decrease in shell thickness. T-test analysis also showed no significant differences (electronic supplementary material, tables S7).

As small island size is clearly selecting for smaller dwarf emu adults, thus egg, and consequently chick size, presumably decline proportionately much less. However, this depends on dwarf emus producing a similar clutch size, as there is no certainty of how many eggs individuals of the different taxa actually laid themselves. If the clutch size remained large and was not the product of more than one female, this suggests that dwarf emus, relative to body weight, must have devoted a proportionately higher amount of energy to reproduction than mainland birds.

Compared with extant ratites, King island emu is morphometrically similar to rhea Rhea sp. (figure 2b), with an analogous breeding strategy. Rhea females lay an average of 20–30 eggs in one nest, incubated solely by a male, and piping stimuli from other eggs results in synchronized hatching [39]. In response to high predation, hatching of the entire clutch can occur in 24–28 h and chicks leave the nest within 2 days [40,41]. Female greater rhea have a body mass of 25 kg (King Island emu 23.3 kg) and an egg mass of approximately 618 g (approx. 547 g in King Island emu) [42], which equates to 2.4% of body mass (2.3% in King Island emu); thus, proportionally, the rhea averages a larger egg. However, egg size variation can be related to clutch size, seasonality and female body mass, and age at the time of laying [42,43]. Tasmanian emu is more similar to Cassowary (Casuarius casuarius) in morphometrics (figure 2b), but cassowary breeding strategy is not comparable [15].

Our study has shown that dwarf emus had a comparable breeding strategy to mainland emu that included a large clutch size, synchronized hatching of young to counter predator effects and thermos-regulation in hatchlings to provide warmth [10,21,23]. It was only on the southern Australian islands that limited resources resulted in rapid dwarfing and retention of a large egg. This scenario provides an interesting evolutionary response to island size, insular population and morphological plasticity in dwarf emus and warrants further study. However, due to their complete and rapid extinction, the true extent of these adaptations to a rapidly changing environment brought on by fluctuating sea levels is now impossible to determine.

Ethics

All research material was either fossil or from museum collections.

Data accessibility

Specimens used in this study are available at the palaeontological collections of TMAG, QVMAG and SAM, Australia and NHMUK, UK. The datasets supporting this article have been uploaded as part of the electronic supplementary material [44].

Authors' contributions

Conceptualization of the paper was by J.P.H. and C.R.; data curation and formal analysis were equally completed by J.P.H. and C.R.; funding acquisition was by J.P.H.; investigation and methodology by J.P.H. and C.R; project administration by J.P.H.; resources by J.P.H. and C.R.; software by J.P.H.; supervision, where necessary, by J.P.H.; validation and visualization by J.P.H. and C.R.; writing—original draft by J.P.H.; writing—review and editing by J.P.H. and C.R. In addition, both authors gave final approval for publication and agree to be held accountable for the content therein.

Competing interests

The authors have no competing interests.

Funding

The NHMUK Department Investment Fund (DIF) awarded to J.P.H. supported this study.