Abstract
Kiwi (Aves; genus Apteryx) are famous for laying an enormous egg in comparison with their relatively small body size. Considering the peculiar gait of this flightless bird, we suspected the existence of morpho‐functional trade‐offs between reproduction and locomotion. To understand how structural constraints, imposed by a large egg size, might influence the terrestrial locomotion of Apteryx, we analysed the anatomy of the limb osteomuscular system in two species of kiwi (Apteryx mantelli and Apteryx owenii). We performed detailed dissections and brought to light specific anatomical features of kiwi, in comparison with other ratites and neognathous birds. Our osteological study revealed a strongly curved pelvis, a rigid tail, and enlarged ribs. Our myology study showed an unusual location of the caudofemoralis muscle origin and insertion. The insertion of the pars pelvica along the entire caudal face of the femur, contrasts with the proximal insertion usually seen in other birds. Additionally, the pars caudalis originates along the entire tail, whereas it only inserts on the uropygium in the other birds. To interpret these specificities from a functional point of view, we built three‐dimensional osteomuscular models based on computed tomography scans, radiographies and our dissections. We chose three postures associated with reproductive constraints: the standing position of a gravid compared with a non‐gravid bird, as well as the brooding position. The 3D model of the brooding position suggested that the enlarged ribs could support the bodyweight when leaning on the huge egg in both males and females. Moreover, we found that in gravid females, the unusual shape of the pelvis and tail allowed the huge egg to sit ventrally below the pelvis, whereas it is held closer to the rachis in other birds. The specific conformation of the limb and the insertions of the two parses of the caudofemoralis help to maintain the tail flexed, and to keep the legs adducted when carrying the egg. The caudal location of the hip and its flexed position explains the long stance phase during the strange gait of kiwi, revealing the functional trade‐off between reproduction and locomotion in this emblematic New Zealand bird.
Keywords: bird morphology, egg incubation, gravid, hind limb, lower appendicular skeleton, myology, New Zealand, osteology
The kiwi is known to carry an enormous egg in relation to its size. Here, we combine traditional methods (quantified dissections) with cutting‐edge techniques (computed tomography scans and 3D modelling) to explore the potential trade‐offs between reproduction and locomotion. We found that the huge size of the egg, associated with a narrow pelvis, constrains the female to carry the egg below the pelvis, leading to specific anatomical adaptations, for example a unique architecture of the caudofemoralis muscle. The caudal location of the hip and its flexed position explains the long stance phase during the strange gait of kiwi, revealing the functional trade‐off between reproduction and locomotion in this emblematic New Zealand bird.
Introduction
Kiwi (Aves; genus Apteryx; five species) are outlandish in many aspects of their biology and life histories, when compared with other birds. Kiwi are flightless, nocturnal, ground‐dwelling birds that feed primarily on invertebrates. They belong to the paleognathous, an early avian taxon comprising flightless birds (the ratites) and the tinamous (47 species in nine genera). The current paleognathous include the ostrich (Struthionidae), the rhea (Rheidae), the tinamous (Tinamidae), the cassowary (Casuariidae), the emu (Dromaiidae), and the kiwi (Apterigidae; Mitchell et al. 2014; Yonezawa et al. 2017; Sackton et al. 2018). Previous studies showed that the ancestors of ratites were small enough to fly (Yonezawa et al. 2017) and that the loss of flight was convergent in this lineage (Sackton et al. 2018). With a size ranging from 23 kg for the smallest rhea to 111 kg for the ostrich (Wilman et al. 2014), most extant ratites are remarkable for their impressive sizes. Among their extinct representatives, the moa (Dinornis robustus) had an estimated weight of up to 240 kg, and the elephant birds (Aepyornis maximus) weighed more than 500 kg (Angst & Buffetaut, 2018). Surprisingly, kiwi are significantly smaller than their sister group, the elephant birds, and closer in size to the tinamous, a flying paleognath. They are only the weight of a chicken, with the female Rakiura tokoeka (Apteryx australis), being the largest species of kiwi, weighing up to 4 kg. Even more surprising, in spite of their small size, they have a relatively large egg, a feature also observed in the elephant bird (Yonezawa et al. 2017). Kiwi eggs can be as large as the egg of a rhea or an emu, birds that are 10 times heavier (Migeon, 2014). The kiwi egg weighs up to a quarter of the female body mass and takes a huge volume in the mother's abdomen (Calder, 1979). In Brown kiwi (Apteryx mantelli) and Little‐spotted kiwi (Apteryx owenii), the male broods the egg inside a nest built in a burrow for 75–84 days, one of the longest incubation periods among birds (Calder et al. 1978). From this large egg, a chick hatches that resembles an adult bird and that is nearly completely independent just a few hours after hatching (Calder et al. 1978). In addition to the above unusual characteristics, kiwi gait is different from that of other ratites: proportionally to their size, their steps are longer, they put their feet on the ground further forward and, to accelerate, they increase mainly the step length instead of the step frequency (Abourachid & Renous, 2000). Furthermore, compared with other ground‐dwelling birds of the same size, kiwi take more time to move their foot forward to complete a step (longer swing time; Abourachid & Renous, 2000).
In this paper, we wanted to investigate potential morphological adaptations to the strange reproductive and locomotive functions in kiwi, and understand whether and how the size of the kiwi egg influences kiwi body shape and posture. Therefore, herein we describe the anatomy of kiwi, with a focus on the pelvis and hind limb musculoskeletal system. Then, we compare the postcranial anatomy of kiwi with that of other birds.
To understand how carrying and brooding such a huge egg affects the female and male musculoskeletal system, we used three‐dimensional musculoskeletal models to compare the position of bones and muscles in different postures: standing with and without carrying an egg, as well as brooding. In light of the anatomical specificities of kiwi, we linked the musculoskeletal system shape to the functional constraints associated with carrying a huge egg. This allowed us to find causal links between the skeleton morphology, the reproductive traits, and the locomotion of kiwi.
Materials and methods
Specimens
Two frozen specimens of Little‐spotted kiwi (A. owenii), one female (no. 55790) and one male (no. 55789), and two specimens of male brown kiwi (A. mantelli) (nos. 54765 and 56040) submitted for postmortem to Wildbase, School of Veterinary Sciences, Massey University, were available for anatomical study. In this study, we only considered postcranial anatomy. The ribs were cut under the uncinate process and the viscera were removed for autopsy before freezing, but the limb muscles were intact on one side at least. After the study, the bodies were given back to Wildbase.
Bone anatomy
We identified the kiwi‐specific features using the character matrix provided by Livesey & Zusi (2007) who used the anatomical characteristics of bird skeletons to carry out a phylogenetic analysis. The authors defined the states of 2954 anatomical characters in 188 avian species, including several fossil species, and the genera Apteryx and Struthio. From the matrix, we extracted the number of occurrences of each state for the 2954 characters. We determined the most common states among the 188 species, and compared them with those of Apteryx and Struthio. We thus were able to identify the kiwi characters that differed from most of the other bird species, and those that differed from Struthio. We refer to the most common state of each of the characters when making comparisons in the description of the anatomical characters. We paid special attention to the specific characters in our description of the lower appendicular skeleton. We followed the Nomina anatomica avium osteological nomenclature (Baumel & Witmer, 1993).
Muscle dissection
We performed detailed dissections of the hindlimbs on the four specimens, using the accurate description of Beddard (1899) as a reference. Following dissection, the thigh and leg muscles were photographed, removed, and weighed. We carefully referenced the origin and insertions of the muscles during dissection. We did not dissect the tarsometatarsus and toes but verified the function of the digit flexors during dissections. We compared the muscle anatomy, and the muscle mass proportions to those available from ostriches (Smith, 2015), emus (Lamas et al. 2014), and turkeys (Abourachid, 1991).
Osteomuscular system 3D model reconstruction
Two specimens of Little‐spotted kiwi (A. owenii), one female (no. 55790) and one male (no. 55789), and one specimen of male brown kiwi (A. mantelli) (no. 56040) were scanned using the computed tomography facility (CT), at Massey University, New Zealand (power: 90 kV, 88 mA; resolution: 0.8 mm on each axis). The CT‐scan data are available for the three specimens (https://data.mendeley.com/datasets/gh6w4c6nrx/2). We used avizo (version 6.3; FEI Visualization Sciences Group) to reconstruct bone models from the CT scans and imported the models in obj file format into autodesk maya (2018, Autodesk Inc., San Rafael, CA, USA).
We observed mounted kiwi skeletons at Massey University at Palmerston North and at the Muséum National d'Histoire Naturelle (MNHN) osteological collection (specimen MNHN 2012‐194) to complement information on missing ribs. Radiographies of a gravid alive Brown kiwi (Calder, 1979) and of a non‐gravid alive Brown kiwi (Massey University), as well as our observations of a brooding kiwi bird were used to approximate the position of the brown kiwi bones in these three postures using autodesk maya software.
Although there were no significant differences between the Little‐spotted and the Brown kiwi anatomy, the proportions were different between the bones of these two species, making the 3D model construction more difficult using a female Little‐spotted kiwi bones to match Brown kiwi radiographies. Therefore, we scaled the bones of a male Brown‐kiwi to build our models in autodesk maya software.
Following our observations during dissections, we determined a point that corresponded to the geometrical centre of the origin, and of the insertion areas of each muscle on the bones. We represented each dissected muscle by a cylinder attached to the bone by these geometric centres. For a given muscle, the cylinder diameter corresponded to the percentage mass of the muscle in relation to the total limb muscle mass. As muscles were attached virtually (i.e. using the software) to the bones, they passively increased or decreased in length to fit the new 3D position in autodesk maya.
Results
There were no differences in the overall anatomy of the two dissected species (Little‐spotted and brown kiwi) or between sexes. Moreover, as the radiographies were available for Brown kiwi, we decided to focus our description on this species. We provide the quantification of the relative muscle sizes (Supporting Information Table S1) for each specimen and 3D bone models (Supporting Information Video S1) for a specimen of brown kiwi.
Osteology
Thorax and pectoral girdle
On the thoracic area, the ribs were wide, with large uncinate processes. The ventral ribs were small and attached on a small and flat sternum. The pectoral girdle was typical of flightless birds, with vestigial wing bones. The coracoid was very short and the scapula anchored on the muscles laterally on the ribs, at the level of the uncinate process of the first thoracic rib. The thoracic rib was closed rather anteriorly (Video S1).
Synsacrum and pelvis
Cranially, the synsacrum aligned with the thoracic vertebrae, which are virtually horizontal when standing. The synsacrum was curved ventrally (Fig. 1A). The curvature was slight in the preacetabular part and more marked in the fused caudal vertebrae. Caudally, the free caudal vertebrae were nearly vertical in a standing position (Fig. 1A).
Figure 1.
Osteology of the pelvis and caudal vertebrae of a brown kiwi (Apteryx mantelli), lateral view with synsacrum coloured in red (A). Dorsal view (B). The dotted line highlights the vertebral column. The synsacrum is highlighted in red on the lateral view.
The vertebrae spinal processes were as high on the synsacrum as on the thoracic vertebrae. The transverse processes were very short, even near the acetabulum and absent in the most caudal part, as on the free caudal vertebrae. There was no transverse lamina on the pelvis (Fig. 1).
The two iliac alae were projected cranially (Fig. 1) from the last thoracic vertebrae so that two ribs were lying under the pelvis (Video S1). On the cranial part of the pelvis the iliac ala were separated dorsally and curved laterally to cover the ribs. The two iliac alae joint dorsally above the synsacrum and the pelvis was compressed laterally. Caudal to the antitrochanter, the ilium was short and narrow, laterally compressed, so that it covered only the spinal process of the fused caudal vertebrae. The vertebral body extended ventrally to the ilium.
The ilium and ischium were only fused anteriorly, around the acetabulum (Fig. 1, Video S1). The ilioischiatic fenestra was very large. A membrane between the ilium and the ischium covered the lateral side of the vertebrae and closed the ilioischiatic fenestra. Cranially, this membrane did not join the corpus ischii, leaving space for nerves and vessels.
The ischium body formed the caudoventral wall of the acetabulum when standing. Its dorsal extremity joined the ventral part of the antitrochanter. A lateral process extended laterally under the antitrochanteric part of the ilium. The ischium corpus was oblique medially. The scapus formed a blade extending caudally, medial to the antitrochanter. It was oriented downward, almost vertically when standing.
The corpus pubis was small. It joined the ilium ventrally toward the acetabulum and joined the ilium corpus cranially. The scapus pubis was oriented dorsoventrally, almost vertical when standing. A membrane closed the puboischiatic fenestra. Both the pubis and the ischium were longer than the postacetabular part of the ilium.
Femur
The femoral diaphysis was incurved craniodistally on a lateral perspective (Fig. 2A). The distal extremity of the femur was asymmetrical, the medial condyle extended distally and ventrally (Fig. 2B). The condyle was sub‐equal on the dorsal side.
Figure 2.
Osteology of a Brown kiwi (Apteryx mantelli). Femur lateral view (A) and caudal view (B), tibiotarsus frontal view (C), tibiotarsus distal view (D), tarsometatarsus cranial view (E), and all the limb bones in a standing position with a frontal view (F).
Tibiotarsus
The proximal extremity of the tibiotarsus was virtually planar with no elongation of the cnemial crest, the fibula extended cranially (Fig. 2C). At the distal extremity of the tibiotarsus, the medial condyle was distinctly more cranially prominent than the lateral condyle. The lateral and medial margins were not parallel (Fig. 2D).
Tarsometatarsus
At the proximal extremity of the tarsometatarsus, the intercondylar eminence was well developed and the hypotarsal sulcus was open (Fig. 2E). The plantar face of the tarsometatarsus diaphysis was flat.
Myology
The muscle name synonymy between Vanden Berge (1982) and McGowan (1979) is given for the proximal (Table 1) and distal hind limb muscles (Table 2) with the description of origins and insertions. The location of the muscles on the hind limb is shown in Fig. 3. The location of the origins and insertions of the muscles is shown on the pelvis (Fig. 4), femur (Fig. 5), tibiotarsus (Fig. 6), and proximal part of the tarsometatarsus (Fig. 7).
Table 1.
Proximal hind limb muscles dissected, alternate names in other sources, and description of their origins and insertions
Abbreviation | Vanden Berge & Zweers | McGowan | Origin | Insertion |
---|---|---|---|---|
Iliotib cran | iliotibialis cranialis | M. sartorius | Cranio‐dorsal edge ilium | Medial part patellar tendon |
Iliotib lat | iliotibialis lateralis | M. iliotibialis | Dorsal edge ilium | Cranial and lateral part patellar tendon |
Iliotroch cran | iliotrochantericus cranialis | M. iliotrochantericus anterior | Ventral edge preacetabular ilium | Lateral proximal femoral epiphysis |
Iliotroch med | iliotrochantericus medius | M. iliotrochantericus medius | Ilium cranioventral to acetabulum | |
Iliotroc cau | iliotrochantericus caudalis | M. iliotrochantericus posterior | Depression of the preacetabular ilium surface | Proximal lateral femoral epiphysis |
Iliofib | iliofibularis | M. biceps femoris | Dorsal edge of the ilium from the antitrochanter area to most of the preacetabular part | Through a ligamentary loop on a lateral tubercle on the proximal 1/3 fibula |
Iliofem ext | iliofemoralis externus | M. gluteus medius et minimus | Dorsal ilium antitrochanter | Lateral femur proximal epiphysis |
Flex cru lat pelv | flexor cruris lateralis pars pelvica | M. semitendinosus | Caudal part of the ilium | Proximal tibio by flat thin tendon between gastrocnemius lateralis and intermedia |
Flex cru lat acc | flexor cruris lateralis pars accessoria | Distal caudal part of the femur | ||
Caudofem cau | caudofemoralis caudalis | M. piriformis pars caudofemoralis | All caudal vertebrae | Femur posterior proximal shaft |
Caudofem pelv | caudofemoralis pelvica | M. piriformis pars iliofemoralis | Synsacrum ventral to the ilium | Caudal part of the femur, just distal to the caudofemoralis caudalis insertion, medial to and all along the femorotibialis lateralis origin |
Ischiofem | ischiofemoralis | ischiofemoralis | Ilioischiatic membrane and ischium | Lateral proximal femoral epiphysis |
Obt int | obturatorius internus | M. obturator internus | Visceral side of the pelvis, on the ilioischiatic membrane | Lateral proximal femoral epiphysis |
Amb | ambiens | M. ambiens | Preacetabular part of the pubis | Patella (F) tendon through the patella to the tibiotarsus (M) |
Pub isch fem | Pubo‐/ischio‐femoralis | M. adductor longus et brevis | Entire length of the ischium (F) and pubis (M) | Femur along lateral face to the medial trochlea |
The left‐most column shows abbreviations of the muscles used in this study with their names. The second and third columns identify alternate names used in major avian myology works (Baumel & Witmer, 1993). The fourth and fifth columns provide a description of the origin and insertion of the muscles, respectively.
Table 2.
Distal hind limb muscles dissected, alternate names in other sources, and description of their origins and insertions
Abbreviation | Vanden Berge & Zweers | McGowan | Origin | Insertion |
---|---|---|---|---|
Femtib lat | femorotibialis lateralis | M. femorotibialis externa | Lateral and cranial surface of femoral diaphysis | Mediocranial, cranial and lateral face of the patella |
Femtib int | femorotibialis internus | M. femorotibialis internus | Femur, mediocaudal surface | On the tibial crest by a tendon below the patellar ligament, and by fibers on gastrocnemius medialis |
Femtibmed | femorotibialis medius | M. femorotibialis medius | ||
Fib long | fibularis longus | M. peroneus longus | Medial part of the patellar ligament, and by a thin profound aponeurosis on the proximal part of the tibialis cranialis and extensor digitorum longus. | Long tendon lateral that cross the tarsometatarsal joint and inserts on the proximo lateral part of the iliotarsus. An accessorial part of the tendon inserts on the tibila cartilage laterally |
Ext dig long | Extensor digitorum longus | M. extensor digitorum longus | From the sulcus intercnemial to the 1/3 proximal of the frontal surface of the tibiotarsus | By a tendon restrained by a fibrous reticunaculum with the tibialis cranialis then pass in the osseous canal distal on the tibiotarsus and cross the tarsometatarsal joint and insert dorsally on digiti II,III and IV |
Tib cran | tibialis cranialis | M. tibialis anterior | By a tendon on the lateral condyle of the femur and by fleshy fibres on the cnemial crest and a thin caput on the proximal third of the tibiotarsus | By a tendon restrained by a fibrous reticunaculum with the extensor digitorum longus and inserts on the tarsometatarsus |
Gast med | gastrocnemius medialis | M. gastrocnemius pars interna | Craniomedial proximal part of the tibiotarsus, below the cnemial crest | Long tendon merged with the intermedia tendon to form the Achilles tendon |
Gast int | gastrocnemius intermedia | M. gastrocnemius pars media | Deeply associated with the flexor cruris medialis pars accessoria by an aponeurosis from the lateral surface of the tibiotarsus, | Long tendon merged with gastrocnemius medialis tendon then to the gastrocnemius lateralis tendon to form the Achilles tendon |
Gast lat | gastrocnemius lateralis | M. gastrocnemius pars externa | By a tendon, from laterodistal part of the femur diaphysis | |
Flex ppd II | Flexor perforans et perforatus digiti II | M. flexor perforans et perforatus digiti II | By a tendon on the laterodistal part of the femur, and to the tibial part of the iliofibularis loop | |
Flex ppdiii | Flexor perforans et perforatus digiti III | M. flexor perforans et perforatus digiti III | On the fibula and in the male, to the tibial part of the iliofibular loop | |
Flex pdii | Flexor perforatus digiti II | M. flexor perforatus digiti II | The fibres are threaded together and share deep aponeuroses. Origin proximal on the popliteal depression on the distocaudal part of the femur | Plantar on the digit II |
Flex pdiii | Flexor perforatus digiti III | M. flexor perforatus digiti III | Plantar on the digit III | |
Flex pdiv | Flexor perforatus digiti IV | M. flexor perforatus digiti IV | Plantar on the digit IV | |
Fib brev | fibularis brevis | M. peroneus brevis | Plantar surface of the digits | |
Flex dig long | Flexor digitorum longus | M. flexor digitorum Iongus | Caudolateral part of the tibiotarsus and on the fibula. | On the plantar face of the digits |
Plant | plantaris | M. plantaris | Caudomedial proximal part of the tibiotarsus | By a tendon on the medial part of the tibial cartilage |
The lefthand column shows the abbreviations used in this study. The second and third columns identify alternate names used in major avian myology works (Baumel & Witmer, 1993). The fourth and fifth columns provide a description of the origin and insertion of the muscles, respectively.
Figure 3.
Location of the hind limb muscles on a specimen of a brown kiwi (Apteryx mantelli). The different muscle layers are represented on the lateral view (A–D) and on the medial view (E, F), from the more superficial muscle layer on the left, to the deepest layer on the right. Abbreviations follow Tables 1 and 2.
Figure 4.
Muscle origins and insertions on the pelvis of a brown kiwi (Apteryx mantelli), dorsal view (A) and lateral view (B). Abbreviations follow Tables 1 and 2.
Figure 5.
Muscle origins and insertions on the femur of a brown kiwi (Apteryx mantelli) on the cranial (A), lateral (B), latero‐caudal (C), caudal (D), and medial views (E). Abbreviations follow Tables 1 and 2.
Figure 6.
Muscle origins and insertions on the tibiotarsus of a brown kiwi (Apteryx mantelli), on the lateral (A), caudal (B), medial (C), and cranial views (D). Abbreviations follow Tables 1 and 2.
Figure 7.
Position of the muscle tendons through the tarsometatarsal joint, and insertion of the tibialis cranialis muscle on the cranial face of the tarsometatarsus. Abbreviations follow Tables 1 and 2.
The thigh muscle caudofemoralis has a remarkable shape. The insertion of its pars pelvica is located along the entire femur caudal face (Figs 3B and 5D), whereas when present in other birds, it inserts proximally on the femur (Baumel, 1993; Smith et al. 2006; Lamas et al. 2014).
Its pars caudalis originates all along the free caudal vertebrae, i.e. all along the tail (Figs 3C and 5D), whereas it usually inserts on the uropygium (Baumel & Witmer, 1993).
The proportion and orientation of the other muscles vary in comparison with other species, but the bone attachments are similar. The relative mass of this muscle is similar in the kiwi and the turkey, but it is relatively smaller in the ostrich (Table S1).
3D model reconstruction
The static 3D models of a Brown kiwi osteomuscular system in a standing position with and without an egg and in the brooding position are given Figs 8 and 9, respectively.
Figure 8.
X‐ray pictures of the lateral view of a standing Brown kiwi (Apteryx mantelli) (A): bird with no egg on the left and with egg in the middle (modified from Calder, 1979), mounted‐skeleton on the right (MNHN 2012‐194). Models of the hind limb musculoskeletal system of a Brown kiwi (A. mantelli) without (on the left) and with an egg (on the right), lateral view (B), frontal view (C), and dorsal view (D).
Figure 9.
Model of the musculoskeletal system of a brown kiwi (Apteryx mantelli) (The sternal ribs are missing due to the autopsy) during brooding on the lateral view (A), frontal view (B), and dorsal view (C). The bottom line shows the position of the body shape during brooding.
Discussion
Anatomical specificities of the kiwi osteology
Pectoral girdle
The kiwi skeleton presents several unusual features. Their vestigial wings are the most reduced among extant birds. Moreover, the width of the vertebral ribs are larger than the intercostal spaces, which is a trait only shared with the hoatzin (Opisthocomus hoatzin) (Livesey & Zusi, 2007). The hoatzin spends most of its time perched on branches, lying on its sternal keel (Strahl, 1988). Considering the anatomical similarity with that bird, we can assume that the ribs of kiwi are enlarged to support bodyweight as well. Interestingly, during the brooding period, the male Brown kiwi develops well‐defined brood patches (Colbourne, 2002), an area of featherless skin on the ventral side of an incubating bird. The avian brood patch is supplied with superficial blood vessels, making it possible for the incubating parent to transfer heat to the egg. In brown kiwi, the brood patch usually lies to the back of a ventral line between the vestigial wings and covers the ventral body to about 50 mm from the cloaca. It is usually about 10 cm wide × 12 cm long (RM Colbourne, pers. comm.). The brown kiwi egg length is on average 12.8 ± 0.5 cm (n = 15) (I. Castro, unpubl. data). Thus, to be able to cover such large egg with the brood patch, the incubating parent (male or female) needs to lean its body on the egg, as shown by our 3D musculoskeletal model (Fig. 9). Therefore, we can assume that the kiwi ribs are enlarged to support the bodyweight when leaning on the egg, as it is for the hoatzin leaning on its branch.
Pelvis
The ratite pelvis is laterally compressed in both its preacetabular and postacetabular parts (Baumel & Witmer, 1993; Fig. 1, Supporting Information Fig. S1). In most neognaths, except diving neognaths (Anten‐Houston et al. 2017), the ilia is usually separated by the synsacrum, which is enlarged at the postacetabular region, leading to the two ilia being widely separated (Fig. S1; Baumel & Witmer, 1993). Moreover, in ratites, the ilioischiatic fenestra is open and the ischium is narrower than in most neognaths. In ratites, ilium and ischium are not fused, and the ischium is longer than the ilium. Considering these features, the overall shape of the kiwi pelvis is typical of ratites.
However, the kiwi synsacrum is bent downward, whereas it is straight in both the ostrich and the turkey. In kiwi, the free caudal vertebrae are oriented downward and the tailbones form a rigid stick, whereas the tail of other birds is usually mobile. This specific caudal part of the kiwi pelvis could help to maintain the egg in gravid females, and both sexes may rest on the rigid stick‐like tail when roosting.
Moreover, kiwi have a pronounced lateral compression of the postacetabular part of the pelvis. The two iliac crests were in contact medially, in contrast to the ostrich, where the postacetabular iliac alae are separated by the synsacrum and the dorsal part of the ilia is dorsally enlarged (Figs 1 and S1). This narrow pelvis does not allow the large egg to be carried between the ilia and pubis, close to the rachis, as is the case in other birds (Shatkovska et al. 2018). Therefore, in kiwi, the egg is carried ventrally below the pelvis. The last thoracic vertebrae allows a slight dorsal flexion of the pelvis, clearly seen on radiography (Calder, 1979; Fig. 8A). All the viscera had to be contained in the cranial part of the ribcage, thus the sternal ribs and the sternum seem to be pushed forward.
The narrow pelvis is a feature usually associated with an abducted femur, as observed in diving (Hertel et al. 2007; Anten‐Houston et al. 2017) and paddling birds (Provini et al. 2012, 2013). However, in contrast to most diving birds (except guillemots and penguins (Shufeldt, 1901) as well as crex, rallus, and phasianus (Bogdanovich, 2014), the kiwi postacetabular ilium is short, with a downward orientation of the synsacrum and a downward orientation of the ischium (Shatkovska et al. 2018). The cranial projection of the iliac ala, the short postacetabular ilium, the bending of the vertebral column, and the large ilioischiatic fenestrae contribute to the ventral bending of the pelvis and the posterior position of the hip on the body. The hip needs to be flexed and the femur abducted to bring the knees on each side of the body, even when carrying a large egg.
Limb
The overall shape of the limb bones looks like those of ground‐dwelling birds (Fig. 2F), although bone proportions are similar to neognath birds, and therefore different from the other ratites (Gatesy & Middleton, 1997). For example, the tarsometatarsus is short (25% leg length) in contrast to those of other ratites (40% of the leg length; Gatesy & Middleton, 1997). The short bones, associated with a flexed posture (Fig. 8; Abourachid & Renous, 2000), keep the centre of mass close to the ground. In addition, the toes are strong, with none of the reduction seen in the other ratites. These features contribute to increase the stability of kiwi, while walking in a complex habitat, and is shared by other burrowing animals (Kardong, 1995).
The long axis curvature of the kiwi femur is unique (Cracraft, 1974) and the length of the fibular crest is only shared with the emu (Dromaius) and the cassowary (Casuarius) (Livesey & Zusi, 2007). In kiwi, the long fibular crest together with the long axis curvature of the femur turns the rotation axis of the knee horizontally, in a standing posture when the femur is abducted (Fig. 2F, Video S1). The tibial plateau is flat in kiwi and the fibular head protrudes laterally onto the lateral face of the femoral condyle. This contributes to the cohesion of the knee joint, preventing the lateral dislocation of the tibiotarsus even when it is adducted, to bring the ankle joint closer to the sagittal plane (Fig. 2F). On the distal epiphysis of the tibiotarsus, the medial condyle is larger than the lateral one. Due to this asymmetry of the condyles, the rotation axis of the ankle joint is horizontal and the tarsometatarsus is vertical on standing position.
The above adjustments of the shape of each limb segment permit a standing posture with a flexed hip, an abducted femur, an adducted tibiotarsus, and a sub‐vertical tarsometatarsus. Thus, despite a very narrow pelvis, the knees can move without the distal part moving too laterally, keeping the feet close to the sagittal plane, under the body. Moreover, the limb features contribute to the rounded body morphology of the kiwi. Due to the backward curvature of the postacetabular pelvis, the hips are located at the caudal end of the body. The relatively long femur (31% leg length) on the much‐flexed hip brings the knee cranially and dorsally (Fig. 2F). The tibiotarsus proportion (44% leg length) is larger than in other ratites (35–45%; Gatesy & Middleton, 1997). It allows the ankle joint to be kept down on the side of the trunk and to assure enough displacement during walking. As is the case for other burrowing vertebrates (e.g. Emerson, 1976; Warburton, 2003; Ilyinsky, 2008) the tarsometatarsus is short (25% leg length) and the toes well developed and strong. This suggests kiwi could use their feet for excavating soil, for example when digging their nests.
The proportions of the leg segments and the high hip flexion have consequences for the walking gait. At the same speed relative to their size (hip height), the kiwi walks at lower frequencies (i.e. relatively slower motion) and longer strides relative to other ratites (Abourachid & Renous, 2000). The flexed hip limits the craniocaudal motion of the knee because the hip extension moves the knee downward rather than backward. This explains why the long tibiotarsus permits the foot to be brought forward but, as the knee is cranial, the foot cannot be brought caudally, which is also a specific trait of the kiwi gait (Abourachid & Renous, 2000).
Anatomical specificities of the kiwi myology
One striking muscle feature observed during dissection was the unique shape of a thigh muscle, the caudofemoralis. The insertion of its pars pelvica is located along the entire femur caudal face (Figs 3B and 5D), whereas when present in other birds, it inserts proximally on the femur (Baumel, 1993; Smith et al. 2006; Lamas et al. 2014). The pars pelvica thus participates to the femur extension. Moreover, when co‐contracted with antagonist extensors, it could form a second muscular layer parallel to the abdominal wall. Therefore, the caudofemoralis pars pelvica could help the infrapubic muscles to maintain the abdominal wall during gestation. This role is also played by the pelvic ilioischiatic and ischiopubic membranes, which increase the limb muscles insertion areas and hold the infrapubic muscles (Fechner et al. 2013).
Its pars caudalis originates all along the free caudal vertebrae, i.e. all along the tail (Figs 3C and 5D), whereas in the other birds it inserts on the uropygium (Baumel & Witmer, 1993). Interestingly, the kiwi uropygium wraps around the caudal vertebrae, in contrast to other birds where it is restricted to the pygostyle region (Reynolds et al. 2017). The pars caudalis pulls on the tail and brings it cranially under the trunk, potentially playing a role holding the egg in gravid females (Fig. 8). This specific insertion of the caudofemoralis pars caudalis could take part in the downward orientation of the caudal vertebrae. The downward orientation of the postacetabular pelvis and tail is likely to help the infrapubic muscles to surround the egg in the abdominal cavity. However, as previously seen, this pulls the hip backwards and causes the hip to flex and the femur to be abducted sufficiently to carry the knee forward. The proportion and orientation of the other muscles vary in comparison with other species, but the bone attachments are similar. The previously discussed distinctive features of the kiwi pelvis shape contribute to the differences in the topology of the muscle attachments. For instance, the ischiofemoralis originated from the ischium both in turkey and kiwi (Fig. S1), but the different bone position and proportions led to a different relative position of the origin between species. The relative mass of this muscle is similar in the kiwi and the turkey, whereas it is relatively smaller in the ostrich (Table S1). Surprisingly, despite the differences in the topology of the muscle attachments due to the specific shape of the ratite pelvis, the proportion of leg muscle mass differs less between the kiwi and the turkey (Abourachid, 1991) than between the kiwi and the ostrich (Smith et al. 2006). Although the sum of the differences between the relative muscle masses of the kiwi and of the ostrich is 52%, it is only 29% between the kiwi and the turkey (Table S1). The ostrich is not considered a ground‐dwelling bird but a cursorial bird (Abourachid & Renous, 2000). Specific adaptations can be observed, as the lengthening and the mass reduction of the distal part of the leg (Abourachid & Renous, 2000). In addition, the number of toes is reduced, as in horses, considered typical cursorial mammals (Alexander, 1982). Thus, despite a closer phylogenetic relationship between the ostrich and the kiwi, the functional demand for cursorial vs. ground‐dwelling locomotion is a strong constraint that can affect muscle proportions and explain the differences observed between the ostrich and the kiwi on one hand and the turkey and kiwi on the other hand. In addition, the lack of differences between male and female anatomy in the two analysed kiwi species also suggests that the selective pressure on locomotion is stronger than the selective pressure on egg transport, which provides important insights into evolutionary processes.
Conclusion
The kiwi is a small, burrowing bird. Its egg is disproportionally large and this reproductive feature imposes mechanical constraints on the musculoskeletal system. As the pelvis is not wide enough to hold the egg, it is carried in front of the pelvis. The postacetabular part of the body (pelvic and tail) is curved downward, which could help to maintain the egg. The hip is flexed and the femur abducted, holding the knee high on the side of the body. The tibiotarsus is rather long, participating in the motion of the legs. The tarsometatarsus is short, and the foot is strong, as usual in burrowing animals. The body shape, as well as the peculiar gait parameters of the kiwi, are a trade‐off between two functions: reproduction and locomotion.
Conflict of interest
The authors declare no conflict of interests.
Authors’ contributions
A.A., I.C., and P.P. participated in the design of the paper. I.C. provided the material and facility to acquire the data, A.A. and P.P. proceeded to the data acquisition (dissections), and P.P. built the musculoskeletal models. A.A. and P.P. interpreted the data. A.A., I.C., and P.P. drafted and made critical revisions of the manuscript. All authors approved the article.
Supporting information
Table S1. (A) Table presenting mass and percentage of the total dissected muscle mass of the four dissected specimens. (B) Mass proportion of the kiwi pelvic muscles, compared with ostrich (from Smith et al. 2006) and turkey (Abourachid, 1991). (C) Proportion of the functional groups of muscles in the kiwi the ostrich and the turkey. Ostrich data from Smith (DATE), turkey data from Abourachid (1990). Abbreviations follow Tables 1 and 2 of the present article.
Fig. S1. Osteology of the pelvis of a brown kiwi (Apteryx mantelli) (A), an ostrich (Struthio camelus) (B), and a turkey (Meleagris gallopavo) (C). Dorsal views at the top and lateral views at the bottom, the cranial part points to the left. The dotted line highlights the vertebral column. The synsacrum is highlighted in red on the lateral view of the brown kiwi . Muscle origins and insertions on the pelvis of a brown kiwi (A. mantelli) (D) and of a turkey (M. gallopavo) (modified from Abourachid, 1991) (E). Dorsal views at the top, lateral views at the bottom. Abbreviations follow Tables 1 and 2.
Video S1. Video of the 3D models of a standing gravid, non‐gravid and incubating kiwi. The associated maya scene is available at https://data.mendeley.com/datasets/gh6w4c6nrx/2.
Acknowledgements
We are thankful to the Action Transversale du Muséum (ATM) for providing financial support. Thanks to Nicola Moffatt and Brett Gartrell for their help with the CT‐scans.
Data accessibility
The CT‐scan data of two specimens of Little‐spotted kiwi (A. owenii), one female (no. 55790) and one male (no. 55789), as well as one specimen of male brown kiwi (A. mantelli) (no. 56040) are available at https://data.mendeley.com/datasets/gh6w4c6nrx/2.
The maya scene with the 3D musculoskeletal models is available at https://data.mendeley.com/datasets/gh6w4c6nrx/2.
References
- Abourachid A (1990) Etude morpho‐fonctionnelle de l'appareil locomoteur de deux souches de dindons domestiques: Recherche d'une explication fonctionnelle aux boiteries des dindons ultra‐lourds. Université Rennes 1.
- Abourachid A (1991) Myologie du membre pelvien du dindon domestique Meleagris gallopavo . Anat Histol Embryol 20, 75–94. [DOI] [PubMed] [Google Scholar]
- Abourachid A, Renous S (2000) Bipedal locomotion in ratites (Paleognatiform): examples of cursorial birds. The Ibis 142, 538–549. [Google Scholar]
- Alexander RM (1982) Walking, running and jumping In: Locomotion of Animals. (ed. Alexander RM.), pp. 81–113. Dordrecht: Springer Netherlands. [Google Scholar]
- Angst D, Buffetaut É (2018) Paléobiologie des Oiseaux Géants Terrestres. London, UK: ISTE Editions. [Google Scholar]
- Anten‐Houston MV, Ruta M, Charles D (2017) Effects of phylogeny and locomotor style on the allometry of body mass and pelvic dimensions in birds. J Anat 231, 342–358. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Baumel JJ (1993) Handbook of Avian Anatomy: Nomina Anatomica Avium. Cambridge, MA: Publ Nuttall Ornithol Club (USA) no 23. [Google Scholar]
- Baumel JJ, Witmer L (1993) Handbook of Avian Anatomy. Cambridge, MA: Nuttall Ornithological Club. [Google Scholar]
- Beddard FE (1899) Notes on the anatomy of the genus Apteryx . Novit Zool 6, 386–402. [Google Scholar]
- Bogdanovich IA (2014) Morphoecological peculiarities of pelvis in several genera of rails with some notes on systematic position of the coot, fulica atra rallidae, gruiformes. Vestn Zool 48, 249–254. [Google Scholar]
- Calder WA (1979) The Kiwi and egg design: evolution as a package deal. Bioscience 29, 461–467. [Google Scholar]
- Calder WA, Parr CR, Karl DP (1978) Energy content of eggs of the brown kiwi Apteryx australis; an extreme in avian evolution. Comp Biochem Physiol A Physiol 60, 177–179. [Google Scholar]
- Colbourne R (2002) Incubation behaviour and egg physiology of kiwi (Apteryx spp.) in natural habitats. N Z J Ecol 26, 129–138. [Google Scholar]
- Cracraft J (1974) Phylogeny and evolution of the ratite birds. Ibis 116, 494–521. [Google Scholar]
- Emerson SB (1976) Burrowing in frogs. J Morphol 149, 437–458. [DOI] [PubMed] [Google Scholar]
- Fechner R, Stratmann M, Gößling R, et al. (2013) The functional role of the ischiopubic membrane for the mechanical loading of the pubis in the domestic fowl (Gallus gallus). J Anat 222, 305–312. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Gatesy SM, Middleton KM (1997) Bipedalism, flight, and the evolution of theropod locomotor diversity. J Vertebr Paleontol 17, 308–329. [Google Scholar]
- Hertel F, Campbell KE Jr, James HF (2007) The antitrochanter of birds: form and function in balance. Auk 124, 789–805. [Google Scholar]
- Ilyinsky VA (2008) Locomotor adaptations in the hindlimbs of owls: the burrowing owl (Athene cunicularia), compared to the Little Owl (Athene noctua). Oryctos 7, 267–272. [Google Scholar]
- Kardong KV (1995) Vertebrates: comparative anatomy, function, evolution.
- Lamas LP, Main RP, Hutchinson JR (2014) Ontogenetic scaling patterns and functional anatomy of the pelvic limb musculature in emus (Dromaius novaehollandiae). PeerJ 2, e716. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Livesey BC, Zusi RL (2007) Higher‐order phylogeny of modern birds (Theropoda, Aves: Neornithes) based on comparative anatomy. Zool J Linn Soc 149, 1–95. [DOI] [PMC free article] [PubMed] [Google Scholar]
- McGowan C (1979) The hind limb musculature of the brown kiwi, Apteryx australis mantelli . J Morphol 160, 33–73. [DOI] [PubMed] [Google Scholar]
- Migeon A‐S (2014) Anatomie Comparée des Ratites. Toulouse: Université Paul Sabatier. [Google Scholar]
- Mitchell KJ, Llamas B, Soubrier J, et al. (2014) Ancient DNA reveals elephant birds and kiwi are sister taxa and clarifies ratite bird evolution. Science 344, 898–900. [DOI] [PubMed] [Google Scholar]
- Provini P, Goupil P, Hugel V, et al. (2012) Walking, paddling, waddling: 3D kinematics Anatidae locomotion (Callonetta leucophrys). J Exp Zool A Ecol Genet Physiol 317, 275–282. [DOI] [PubMed] [Google Scholar]
- Provini P, Simonis C, Abourachid A (2013) Functional implications of the intertarsal joint shape in a terrestrial (Coturnix coturnix) versus a semi‐aquatic bird (Callonetta leucophrys). J Zool 290, 12–18. [Google Scholar]
- Reynolds SM, Castro I, Alley MR, et al. (2017) Apteryx spp. (Kiwi) possess an uropygial gland: anatomy and pathology. Eur J Anat 21, 125–139. [Google Scholar]
- Sackton TB, Grayson P, Cloutier A, et al. (2018) Convergent regulatory evolution and the origin of flightlessness in palaeognathous birds. bioRxiv, 262584 10.1101/262584. [DOI] [Google Scholar]
- Shatkovska OV, Ghazali M, Mytiai IS, et al. (2018) Size and shape correlation of birds’ pelvis and egg: impact of developmental mode, habitat, and phylogeny. J Morphol 279, 1590–1602. [DOI] [PubMed] [Google Scholar]
- Shufeldt RW (1901) Osteology of the Penguins. [PMC free article] [PubMed]
- Smith DK (2015) Craniocervical myology and functional morphology of the small‐headed therizinosaurian theropods Falcarius utahensis and Nothronychus mckinleyi . PLoS One 10, 1–16. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Smith NC, Wilson AM, Jespers KJ, et al. (2006) Muscle architecture and functional anatomy of the pelvic limb of the ostrich (Struthio camelus). J Anat 209, 765–779. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Strahl SD (1988) The social organization and behaviour of the Hoatzin Opisthocomus hoazin in central Venezuela. The Ibis 130, 483–502. [Google Scholar]
- Vanden Berge JC (1982) Notes on the myology of the pelvic limb in kiwi (Apteryx) and in other birds. Auk 99, 309–315. [Google Scholar]
- Warburton N (2003) Functional morphology and evolution of marsupial moles (Marsupialia: Notoryctemorphia).
- Wilman H, Belmaker J, Simpson J, et al. (2014) EltonTraits 1.0: Species‐level foraging attributes of the world's birds and mammals: ecological archives E095–178. Ecology 95, 2027. [Google Scholar]
- Yonezawa T, Segawa T, Mori H, et al. (2017) Phylogenomics and morphology of extinct paleognaths reveal the origin and evolution of the ratites. Curr Biol 27, 68–77. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Table S1. (A) Table presenting mass and percentage of the total dissected muscle mass of the four dissected specimens. (B) Mass proportion of the kiwi pelvic muscles, compared with ostrich (from Smith et al. 2006) and turkey (Abourachid, 1991). (C) Proportion of the functional groups of muscles in the kiwi the ostrich and the turkey. Ostrich data from Smith (DATE), turkey data from Abourachid (1990). Abbreviations follow Tables 1 and 2 of the present article.
Fig. S1. Osteology of the pelvis of a brown kiwi (Apteryx mantelli) (A), an ostrich (Struthio camelus) (B), and a turkey (Meleagris gallopavo) (C). Dorsal views at the top and lateral views at the bottom, the cranial part points to the left. The dotted line highlights the vertebral column. The synsacrum is highlighted in red on the lateral view of the brown kiwi . Muscle origins and insertions on the pelvis of a brown kiwi (A. mantelli) (D) and of a turkey (M. gallopavo) (modified from Abourachid, 1991) (E). Dorsal views at the top, lateral views at the bottom. Abbreviations follow Tables 1 and 2.
Video S1. Video of the 3D models of a standing gravid, non‐gravid and incubating kiwi. The associated maya scene is available at https://data.mendeley.com/datasets/gh6w4c6nrx/2.
Data Availability Statement
The CT‐scan data of two specimens of Little‐spotted kiwi (A. owenii), one female (no. 55790) and one male (no. 55789), as well as one specimen of male brown kiwi (A. mantelli) (no. 56040) are available at https://data.mendeley.com/datasets/gh6w4c6nrx/2.
The maya scene with the 3D musculoskeletal models is available at https://data.mendeley.com/datasets/gh6w4c6nrx/2.