Aims and Methods of the Dromornithid Study
Ken Ashwell
In the present study I was interested in determining how neurologically significant features of the crania of the Miocene dromornithids (as examples of giant anseriforms, and the largest birds ever to have lived) scaled with body weight and ECV and whether they showed evidence of neurological specializations such as might be required for adaptation to the semi-arid environment of Miocene/Pliocene Central Australia.
In the present study, quantitative analysis of extant anseriform (ducks, geese) and dromornithid skulls, as well as computerised tomography of the fossils, were used to evaluate the size of cranial features with neurological significance. Cranial features associated with olfactory sensation (olfactory bulb volume), visual function (optic canal area, orbital diameter, orbital area, orbital convergence angle, Wulst area), trigeminal sensation (maxillomandibular canal area), auditory function (external otic recess area); as well as evidence for specialized control of the extraocular muscles (oculomotor canal area) and the tongue (hypoglossal canal/s area), and postcranial musculature (foramen magnum area) were also analysed.
The working hypothesis was that these large extinct anseriforms would exhibit a disconnection between brain and body expansion (as has previously been found for the Haast eagle, Scofield and Ashwell, 2009), meaning that the body has expanded much more than the brain. It was also hypothesised that adaptation to the arid environment of Miocene Central Australia and consumption of tough dry seeds would have driven expansion of the tongue musculature and hypoglossal nerve size, as reflected in the hypoglossal canal/s area.
Materials and Methods
Specimens
The present study was based on 11 dromornithid fossils (7 Dromornis planei, 4 Dromornis stirtoni) held at the Museum of Northern Territory (Alice Springs) Australia. Note that Dromornis planei was formerly placed in the genus Bullockornis(Murray and Vickers-Rich, 2004) but it has been included with Dromornis in the present study based on a recent revision (Worthy et al., 2015). The specimens used were NTM-P907-6, NTM-P907-29, NTM-P9464, NTM-P9464-111, NTM-P9484-106, NTM-P9612-1 and NTM-P9973-6 for Dromornis planei; and NTM-P3249, NTM-P3250, NTM-P3251, NTM-P9810 for Dromornis stirtoni. Specimens of Dromornis planei come from the Middle Miocene, Bullock Creek local fauna, Camfield station, Central Australia; whereas specimens of Dromornis stirtoni come from the Late Miocene, Alcoota local fauna, Central Australia. Descriptions of the most complete fossil specimens are available in Murray and Megirian (1998), Murray and Vickers-Rich (2004) and Worthy et al. (2015).
The crania of 37 species (49 specimens) of modern anseriforms (ducks and geese) were used to obtain measurements of cranial features with neurological significance. The relationship between Wulst area and ECV was based on extrapolation from 27 species (35 specimens: ratites, galloanseriforms, passerines, podargiforms, caprimulgiforms; Ashwell and Scofield, 2008, Brauth et al. http://www.brauthlab.umd.edu/atlas.htm; Ebinger, 1995; Ebinger and Löhmer, 1987; Edinger and Röhrs, 1995; Izawa and Watanabe, 2007; Igarashi and Kamiya, 1972; Iwaniuk and Wylie, 2006; Karten http://avianbrain.org/nomen/Pigeon_Atlas.html; Nixdorf-Bergweiler and Bischof, 2007; Puelles et al., 2007; Rehkämper et al., 1991).
The dromornithid crania used in the present study cannot be definitively linked with the long bones usually used for body mass estimation (Murray and Vickers-Rich, 2004; Handley et al., 2016). Wherever approximations of body weight were required for estimating expected size of osteological features with neurological significance, the mean body mass estimates for the species as a whole were used. This is 300 kg for D. planei (Murray and Vickers-Rich, 2004) and 490 kg for D. stirtoni, the mean of the male and female adult body weight estimates based on tibiotarsi from the most recent analysis of body size and sexual dimorphism in this species (Handley et al., 2016).
CT scanning
Four specimens of dromornithid skulls (Dromornis planei – NTM-P9464, NTM-P9484-106, NTM-P9973-6; Dromornis stirtoni – NTM-P3249) were CT scanned using a medical scanner at the Alice Springs Hospital (Toshiba Aquilon 16 slice; 0.3 mm slice thickness).
The skulls were reconstructed from the DICOM files and estimates of the absolute and proportional area of the olfactory bulb, combined forebrain and midbrain, pituitary fossa and cerebellum/pons/medulla oblongata were calculated from serial sections measured with OsiriX. Olfactory bulb volume was estimated from analysis of frontal sections where the cribriform plate defines the medial and lateral extent of the bulb. The ECV and volumes of subregions were calculated from the product of the sum of regional areas in serial sections and the between-section intervals. The surface area of the Wulst was also estimated from measurements made in serial frontal sections, where the lateral border of the hyperpallium apicale is clearly demarcated in the dromornithids (as in other birds) by the prominent ridge of the vallecula on the inside of the skull and the medial extent is marked by the dorsomedial border of the telencephalon. The sum of the transverse dimensions of the Wulst in regularly spaced sections was multiplied by the between-section interval to obtain the Wulst area for both sides of the forebrain.
Endocranial volume measurement
Endocranial volume (ECV) was measured by either the dried mustard seed technique (Stewart, 1947) for modern anseriforms or from pixel counting of high-resolution spiral CT (OsiriX) on four selected dromornithid specimens (i.e. Dromornis planei – NTM-P9464, NTM-P9484-106, NTM-P9973-6; Dromornis stirtoni – NTM-P3249)(see above). Re-measurement of ECV by the mustard seed technique in modern birds shows that estimates are reproducible within +0.25 ml (Ashwell and Scofield, 2008). In the present study, ECV was considered to be equivalent to brain weight (i.e. an assumption that brain density is 1 g/ml), as has been used by previous authors for avian skulls (Jerison, 1973; Iwaniuk and Nelson, 2002). The encephalization quotient for a given specimen i (EQi) was calculated using EQi = Ei/0.1371Wi0.568using allometric constants for brain volume (Ei) versus body mass (Wi) derived from measurements of 1,482 species of birds (r = 0.939)(Iwaniuk and Nelson, 2003).
Endocranial volume of the largest elephant-bird species (Aepyornis maximus) is not available in the literature, so ECV was estimated on the basis of stated endocranial cavity length (Edinger, 1942). This was then used to deduce brain volume assuming a similar proportional shape in the brains of the nearest living ratite relatives (Dinornis robustus and Anomalopteryx didiformis). Body weight estimates were taken from publications (reviewed in Deeming and Birchard, 2008).
Cranial canal measurement
Measurement of the area of cranial canals that transmit neurologically significant structures was made using photographs taken with a Canon EOS 400D digital camera equipped with an EF-S60mm f/2.8 Macro USM lens. These canals are illustrated in Worthy et al. (2015) and include the foramen opticum (optic canal - transmitting the optic nerve, CNII), the canal for the oculomotor nerve (CNIII), the ostium canalis carotici (carotid canal - transmitting the internal carotid artery), the maxillomandibular canal (transmitting the maxillary and mandibular divisions of the trigeminal nerve, CNV2 and CNV3), external otic recess rim (as an indicator of potential maximal tympanic membrane area), the foramina hypoglossi (hypoglossal canal/s, transmitting the hypoglossal nerve/s) and the foramen magnum (transmitting the spinal cord and vertebral arteries). Note that large anseriforms may have two (and even sometimes three) hypoglossal canals on each side (Worthy et al., 2015) so (where multiple canals were present) the sum of areas was used in analysis. The photographs of foramina or canals were taken with a calibrated scale at the same focal distance as the structure of interest. Images were exported to ImageJ for measurement of the area of the foramen or canal.
Orbital dimensions and orientation
The width of the orbit was measured in fossils and dried skulls of extant anseriforms with the aid of calibrated photographs (Figure 2). These photographs were of the orbit from a direction along the orbital axis for the skulls of modern anseriforms, but were of the inferior surface of the orbital roof for dromornithids, because the quadratojugal bar that defines the inferior orbital margin in extant birds is missing from the fossils. Width of the orbit was measured from the anterior orbital margin (OA - orbital anterior, the point of the orbital margin most distant from the inion, Iwaniuk et al., 2008) to the posterior orbital margin (OP - orbital posterior, the point of the orbit closest to the inion, Iwaniuk et al., 2008) in both extant anseriforms and fossils.
Photographs of the orbital roof in dromornithid fossils were used to estimate the angle between the left and right orbital axes (line bisecting the angle between medial and lateral orbital walls with the apex at the optic canal) and the midline, in the fossils (see Figure 2). Orbital convergence was also measured on photographs of the orbit in dromornithid fossils to make comparisons with published data in modern birds (Iwaniuk et al., 2008). The quadratojugal bar was absent from the fossils, so the plane of the orbital rim (i.e. line from the OA to OP) was used as the orbital plane. The angle between this plane and the midline was measured for both sides and summed to give the dihedral angle of convergence (Iwaniuk et al., 2008), such that the higher the angle, the greater the degree of ocular convergence. Orbital area was measured as the ventral surface of the orbital roof and compared between dromornithids and modern anseriforms (see below).
Comparisons with modern anseriforms and other birds and prediction of expected neurological variables
The values for neurological variables measured on the dromornithid fossils were compared with those that would be expected for a modern anseriform scaled to the same body weight or ECV. The skulls of 37 modern species of anseriforms (total of 49 specimens) held at the Australian Museum in Sydney were used to measure orbital diameter, orbital area, optic canal area, oculomotor canal area, maxillomandibular canal area, carotid canal area, external otic recess area, hypoglossal canal/s area and foramen magnum area. These measurements were made from photographs of the skulls with a calibrated scale positioned at the same distance from the lens as the structure of interest. Images were exported to ImageJ (1.37v software) for measurement of the area of the canal.
Wulst area was determined by different means depending on the nature of the material, because the data had to come from diverse sources. For the dromornithid specimens, Wulst area (both sides of the brain) was calculated from the transverse diameter of the Wulst as measured in serial sections spaced at 5 mm intervals through the endocranial cavity. The sums of these were multiplied by the inter-section interval. A similar analysis had been made in our previous study of moa and ratites (Ashwell and Scofield, 2008) and these data were available for comparison. Data were gathered from the literature for a variety of other birds (27 species, 35 specimens: ratites, galloanseriforms, passerines, podargiforms, caprimulgiforms) either from stated values or measurements of illustrations in atlases and papers. For some of the latter (e.g. Igarashi and Kamiya, 1972) the area of the Wulst could be directly measured from photographs of the dorsum of the brain. For others (e.g. Ebinger and Löhmer, 1987; Izawa and Watanabe, 2007) analysis of serial sections was the only possible approach and was performed as described above for the dromornithids.
Statistical analysis was performed using Excel (14.0.0 for Mac 2011). To analyse the correlation of the size of the neurological variable with body weight and ECV, measurements were firstly converted to a single spatial dimension by taking the square root of areal measurements (e.g. canal, foramen or orbital area), or the cube root of volumes (e.g. ECV or body weight). To accommodate the large range of values, measures were plotted on logarithmic scales, i.e. the log10of a square root of given variable was plotted against log10 of the cube root of body weight or ECV. Expected values of variables for dromornithids body weights and ECV were calculated from the linear regression equations. Prediction intervals (95%) were calculated using the predicted value for that body weight or ECV + the product of standard error of predicted y for x (STEYX) and TINV (0.05, df).
Figure 2
Photograph of the ventral aspect of the cranium of NTM-P9464-111 Dromornis planei, showing the way that the orbital axes and angle of convergence were determined. The midline sagittal plane is indicated by a thick black dashed line. The anterior and posterior margins of the orbit (OA and OP, respectively) were determined as the parts of the orbital rim furthest and closest to the inion (out of view to bottom of image), respectively. The lines between OA and OP define the orbital planes of each orbit and the sum of the angles between the orbital planes of each side and the midline (A and B) is the dihedral angle of convergence. The orbital axis of each orbit was defined as the line that passes through the optic canal and bisects the angle OA-OC-OP. The angle formed between this line and the midline was measured.
References
Ashwell KWS, Scofield RP (2008) Big birds and their brains: paleoneurology of the New Zealand moa. Brain Behavior and Evolution 71, 151-166.
Brauth SE, Heaton JT, Roberts TF, Liang W (2016) Budgerigar brain atlas. http://www.brauthlab.umd.edu/atlas.htm. Accessed March 13 2016.
Deeming DC, Birchard GF (2008) Why were extinct gigantic birds so small? Avian Biological Research 1, 187-194.
Ebinger P (1995) Domestication and plasticity of brain organization in mallards (Anas platyrhynchos). Brain Behavior and Evolution 45, 286-300.
Ebinger P, Löhmer R (1987) A volumetric comparison of brains between greylag geese (Anser anser L.) and domestic geese. Journal für Hirnforschung 28, 291-299.
Edinger P, Rohrs M (1995) Volumetric analysis of brain structures, especially of the visual system in wild and domestic turkeys (Meleagris gallopavo), Journal of Brain Research 36, 219-228.
Handley WD, Chinsamy A, Yates AM, Worthy TH (2016) Sexual dimorphism in the late Miocene Mihirung Dromornis stirtoni (Aves: Dromornithidae) from the Alcoota local fauna of Central Australia. Journal of Vertebrate Paleontologye1180298.
Igarashi S, Kamiya T (1972) Atlas of the Vertebrate Brain. Morphological Evolution from Cyclostomes to Mammals. University of Tokyo Press, Tokyo.
Iwaniuk AN, Nelson JE (2002) Can endocranial volume be used as an estimate of brain size in birds? Canadian Journal of Zoology 80, 16-23.
Iwaniuk AN, Nelson JE (2003) Developmental differences are correlated with relative brain size in birds: a comparative analysis. Canadian Journal of Zoology 81, 1913-1928.
Iwaniuk AN, Wylie DRW (2006) The evolution of stereopsis and the Wulst in caprimulgiform birds: a comparative analysis. Journal of Comparative Physiology A 192, 1313-1326.
Iwaniuk AN, Heesy CP, Hall MI, Wylie DRW (2008) Relative Wulst volume is correlated with orbit orientation and binocular visual field in birds. Journal of Comparative Physiology A 194, 267-282.
Izawa E, Watanabe S (2007) A stereotaxic atlas of the brain of the jungle crow (Corvus macrorhynchos), In: Watanabe, S., Hofman, M.A. (Eds.) Integration of Comparative Neuroanatomy and Cognition. Keio University Press, Keio, pp.215-273.
Jerison H (1973) Evolution of the Brain and Intelligence. Academic, New York.
Karten H (2016) Pigeon brain atlas. http://avianbrain.org/nomen/Pigeon_Atlas.html. Accessed 13 March 2016.
Murray PF, Megirian D (1998) The skull of dromornithid birds: anatomical evidence for their relationship to Anseriformes. Records of the South Australian Museum 31, 51-97.
Murray PF, Vickers-Rich P (2004) Magnificent Mihirungs: The Colossal Flightless Birds of the Australian Dreamtime. Indiana University Press, Bloomington & Indianapolis.
Nixdorf-Bergweiler BE, Bischof HJ (2007) A Stereotaxic Atlas of the Brain of the Zebra Finch, Taeniopygia Guttata. With special emphasis on telencephalic visual and song system nuclei in transverse and sagittal sections.
http://www.ncbi.nlm.nih.gov/books/NBK2348/
Puelles L, Martinez-de-la-Torre M, Paxinos G, Watson C, Martinez S (2007) The Chick Brain in Stereotaxic Co-ordinates. An Atlas Featuring Neuromeric Divisions and Mammalian Homology. Academic Press, San Diego.
Rehkämper G, Frahm HD, Zilles K (1991) Quantitative development of brain and brain structures in birds (Galliformes and Passeriformes) compared to that in mammals (insectivores and primates). Brain Behavior and Evolution 37, 125-143.
Scofield RP, Ashwell KWS (2009) Rapid somatic expansion causes the brain to lag behind: The case of the brain and behaviour of New Zealand’s Haast’s eagle (Harpagornis moorei). Journal of Vertebrate Paleontology 29, 637-649.
Stewart TD (1947) Hrdlicka’s Practical Anthropometry. third ed.. Wistar Institute of Anatomy and Biology, Philadelphia.
Worthy TH, Handley WD, Archer M, Hand SJ (2015) The extinct flightless mihirungs (Aves, Dromornithidae): cranial anatomy, a new species, and assessment of Oligo-Miocene lineage diversity. Journal of Vertebrate PaleontologyDOI:10.1080/02724634.2015.103134.