Art. XII.—On Harpagornis, an Extinct Genus of Gigantic Raptorial Birds of New Zealand.
[Read before the Philosophical Institute of Canterbury, 4th June, and 2nd July, 1873.]
Plates VII., VIII., IX.
In a paper read before the Philosophical Institute of Canterbury, in 1871, † I offered the first account of the discovery of a few bones belonging to a gigantic bird of prey, which were obtained with a considerable quantity of Moa bones in the turbary deposits of Glenmark, a locality which will ever be celebrated in the scientific annals of New Zealand as the spot which, doubtless, has furnished the largest quantity and variety of bones available for the elucidation of the anatomy of the wonderful, wingless, struthious birds of this country.
The bones described in that paper consisted of a left femur, two ungual phalanges, and a rib, all belonging to the same specimen.
Since the publication of those first notes, further excavations were undertaken in the same locality; and in following down the swampy water-course from which these few remains of Harpagornis were previously obtained a further series of bones was discovered, which, on examination, proved to be another portion of the same skeleton described in that first memoir.
The bones recently obtained were scattered over the bottom of the turbary deposit along the old water-course, 6ft. to 7ft. below the surface, amongst the remains of decaying swampy vegetation. They were mixed up with pieces of drift timber, and with a considerable number of Moa bones, several of them belonging to the larger species (Din. giganteus var. maximus, and Din. robustus).
The bones obtained during these latter excavations consisted of the following:—right and left metatarsus, right and left tibia, right and left fibula, right and left ulna, right and left radius (one fragmentary), right and left scapula, one rib, five phalanges, four ungual phalanges.
[Footnote] * At the request of the author, the publication of this paper at full length has been deferred until all the illustrations can be published of natural size, in quarto form.—Ed.
[Footnote] † Trans. N.Z. Inst., Vol. IV., p. 192.
Our search after the pelvis, sternum, and cranium, was in vain, so that I shall not be able to offer a description of these important parts of the Glenmark skeleton; but, as will be seen in the sequel, I can at least do so as far as the pelvis of the species is concerned, Dr. Hector having kindly handed over to me, for such purpose, a well-preserved specimen of that compound bone, found in one of the Otago caves. *
This list does also not contain any humerus, but we possess at least a fragmentary one, without doubt belonging to this species, which was obtained about a mile above Glenmark, from the banks of the Glenmark Creek. These banks rise in some places about 100ft. above the water-line, in nearly perpendicular cliffs, and consist of postpliocene alluvium, formed by large beds of shingle, with which smaller deposits of sand and turbary deposits are interstratified.
We obtained also the lower portion of a metatarsus, from a similar older postpliocene bed situated close to Glenmark, so that there is sufficient evidence to show that this diurnal raptorial bird existed, like the Dinornis and Palapteryx species, during a long period in New Zealand.
Some time after having made the discovery of the further portion of the skeleton of Harpagornis moorei, in continuing our excavations on the Glenmark property, on the left bank of the Glenmark Creek, and opposite the spot previously alluded to, we obtained, amongst a considerable quantity of Moa bones, a large portion of another skeleton of a raptorial bird, which, although of smaller size than the first-named species, is still of remarkable dimensions. These bones were found not far apart, and near the bottom of the swamp, close to a layer of clay, 7ft. to 8ft. below the surface.
This new find consisted of the following bones: pelvis (fragmentary), right and left metatarsus, right and left tibia, right and left femur, right humerus, right and left ulna, left metacarpal, left scapula, one rib, four phalanges, one ungual phalanx.
In comparing these with the bones of Harpagornis moorei, it became at once evident that they belonged either to a closely allied form, or, making allowance for sex, to the former species.
The disproportion in size of our recent diurnal raptorial birds is so great, that even at the present time the question as to the existence of one or two species of Hieracidea is not yet definitely settled. This remarkable difference in size is also observable in the New Zealand Harrier, where the female is
[Footnote] * This is one of the bones referred to in Trans. N.Z. Inst., Vol. IV., p. 114 (footnote), as having been forwarded by Mr. W. A. Low, which were found in the surface soil under an overhanging rock, and not in a proper cave. This particular bone is in wonderful preservation, and is still covered with periosteum and has the capsular and some other ligaments adherent, while the osseous substance has lost hardly any of the original animal matter which it contained. —J. Hector.
generally much larger than the male bird. Moreover, when comparing the male and female skeletons of Circus with each other, there are some slight sexual differences easily discernible, which might suggest that they belonged to two nearly allied species, did we not know their real relations to each other.
As I shall show further on, the bones of both specimens of Harpagornis belong to adult birds, of which the largest died at a more mature age than the smaller one. Thus the smaller specimen might possibly be the male of H. moorei, assuming the latter to be the female. However, as I am not able to settle this point at present, I shall propose for the second and smaller specimen the specific name of H. assimilis, in order to point out the close relationship of both.
Dr. Hector suggested * to me that the Harpagornis might possibly be the Hokioi of the Maoris, which, however, according to Buller, is the Great Frigate Bird (Fregata aquila), obtained repeatedly in New Zealand, and of which he gives several instances in his work on the birds of New Zealand. †
What the large bird of prey is that I have met several times during my explorations amongst the snow-clad ranges constituting the Southern Alps, without being able to secure a specimen, is a question which I hope future and more fortunate explorers of those regions will one day solve.
Before offering a description of the extremities of Harpagornis, I wish to draw attention to the following table of measurements, in which I have placed in juxtaposition the length of the principal leg and wing bones of all the diurnal birds of prey of which I had material for comparison.
[The section below cannot be correctly rendered as it contains complex formatting. See the image of the page for a more accurate rendering.]
|Harpagornis moorei.||Harpagornis assimilis.||Aquila audox, Australia.||Circus assimilis, New Zealand.||Hieracidea novæ zealandiæ, New Zealand.|
In comparing, in the first instance, the length of the femur with the
[Footnote] * On the authority of Sir George Grey. Trans. N.Z. Inst., Vol. V., p. 435.—Ed
[Footnote] † Buller's “Birds of New Zealand.” 4to., 1873. P. 340.—Ed.
metatarsus, it will be seen that in Harpagornis the former is longer than the latter limb bone, in this respect resembling Aquila, whereas in Circus the opposite is the case, the metatarsus being longer than the femur. This, to a minor extent, we observe also in Hieracidea. However, when we take the united length of the three principal leg bones into consideration, and compare them with the three principal wing bones, the result is quite different.
Thus, whilst the wing bones of H. assimilis are only 1.52 inches longer than the leg bones (20.88 inches to 22.40 inches) in Aquila, they are, notwithstanding their smaller dimensions, 5.28 inches longer (16.57 inches to 22.03 inches); Harpagornis here again agreeing more with Circus (10.52 inches to 11.34 inches).
According to their different proportions, the wing bones of H. assimilis, when compared with Aquila audax, ought to be 27.53 inches, instead of 22.40 inches their actual size; and, with Circus, 22.50 inches, a result which closely agrees with the above measurement. Of H. moorei we possess only the ulna, the length of which, 10.06 inches, compared with the same bone in the smaller H. assimilis, 9.35 inches, would give for the whole wing bones a total length of 24.10 inches, instead of 29.62 inches, as calculated according to the measurements of Aquila audax.
I wish also to point out that in Hieracidea the united length of the wing bones is actually less than that of the leg bones (7.38 inches to 6.61 inches), although this little bird is remarkably strong on the wing.
Harpagonis moorei. Trans., Vol. IV., Pl. X., Fig. 1.
In my former notes on Harpagornis I offered a short description of the femur (vol. iv., p. 193), comparing it at the same time with the corresponding limb bone in the skeleton of Palioaëtes leucogaster, the white-bellied Sea-Eagle of Australia, and of Circus assimilis, the New Zealand Harrier; but I shall, in the following notes, compare all the principal bones of Aquila audax, the largest Australian species of Diurnal Raptores, with those of the extinct New Zealand bird.
|Total length of femur||6.09|
|Circumference of proximal extremity||4.10|
|Circumference of distal extremity||4.83|
|Circumference of shaft where thinnest||2.22|
This bone, besides being of smaller dimensions, is somewhat slighter in its form, otherwise the description as given of that of Harpagornis moorei closely corresponds in all its principal points. There is no doubt, judging from the insertion marks of the muscles and the intermuscular linear ridges, that this
species was also very powerful. Of the latter the linea aspera is not quite continuous, being repeatedly interrupted at more or less considerable intervals. The form of the proximal orifice is somewhat different from that of H. moorei, it being more rounded; however, this may be a sexual or even individual peculiarity, and of no specific value.
Examining the femora of a male and a female Circus, I observe that this proximal orifice in the larger female is also oval, and in the smaller male more rounded off.
I have already alluded to the probability that the portions of the two skeletons of extinct birds under consideration might belong to the male and female of the same species, owing to a resemblance in their principal osteological features and to the great disparity of size of many of the recent Diurnal Raptores.
In order to illustrate this more fully, I would suggest a comparison of the femora of Circus assimilis of both sexes, both belonging to full-grown and mature birds, obtained under similar conditions.
Certainly, if these two bones had been found in a fossil state, one would not deem it expedient to place them in the same species, owing to their remarkable difference in size.
Harpagornis moorei. Pl. VII., Figs. 1 and 2.
|Circumference at proximal end||5.48|
|Circumference at distal end||4.60|
|Circumference of shaft where thinnest||2.22|
The same pachydermal character, if I may thus express myself, distinguishes also this bone, like all those of the posterior limb of this gigantic species, from any bird of prey known to inhabit New Zealand at present.
Even in comparing the same with that of Aquila audax, of Australia, with which it has otherwise many features in common, this character is well exhibited.
The form of the surface of the proximal end agrees well in both species, with the exception that the proximal ridge is more rounded off, and the intercondylar tuberosity stands higher in Harpagornis, in which two features the fossil bone agrees more with Circus.
Two narrow and low intermuscular ridges are well marked, the first of which begins at the base of the procnemial process and extends to the inner side of the extensor tendinal canal, above the bony bridge spanning over the precondylar groove; the other at the termination of the vertical fibular ridge, descending the shaft in a transverse line till it has crossed two-thirds of its
breadth, within one inch above the bridge, then retreating again with a gentle curve. After forming the outer boundary of the groove, it then terminates on the outer side of the canal for the extensor tendon; thus differing from Aquila, where this second ridge reaches only to the middle of the shaft, and does not describe such a considerable arc as we observe upon the fossil bone.
The distal condyles are well curved at their anterior ends, and have a more rounded form (which the outer condyle shows most conspicuously) than either Aquila or Circus, in which they are more oblong. Moreover, those of the recent species stand more in advance of the shaft.
The inner distal condyle is also more developed in a transverse extent than the outer one—a feature also exhibited by Aquila.
The shaft of the bone, although slightly bent backwards near its proximal end, is, however, straighter than in Aquila, but not so straight as that of Circus. The fibular ridge is strongly developed.
|Circumference at its proximal end||(partly broken away).|
|Circumference at its distal end||3.80|
|Circumference of shaft where thinnest||1.91|
This tibia, although possessing all the main characteristics of the larger species, is, when considering its total length, of a somewhat slighter form. I observe, however, that the distal condyles are more oblong, agreeing more in their shape with the recent species hitherto used for comparison. This is best seen in the outer condyle.
Might this peculiarity not be traced to age, the skeleton of Harpagornis moorei doubtless having belonged to a more aged bird than the smaller species? Thus the texture of the extremities of the tibia of the former is far more compact than in the latter, in which, although well anchylosed, a want of solidity is observable.
Fibula. Pl. “VII., Figs. 3 and 4.
Amongst the smaller bones obtained from the locality where the principal portion of the skeleton of Harpagornis moorei was excavated are a pair of fibulæ, which, on closer examination, proved to belong to that skeleton. Of these the right one is the most perfect. It is 4.27 inches long, the distal point being broken off.
The articular head, 0.80 inch long and 0.31 inch broad, is very large and posteriorly slightly convex, its anterior edge sloping down at a considerable angle, far more than in Aquila or Circus, in which the articulating surface is nearly plane, and stands at a right angle to the shaft. The head is also far more hollowed out on the inner side than Aquila.
The shaft in its upper portion is considerably bent backwards, and very broad where it is attached to the tibia, after which it decreases rapidly in size. Two shallow pits for the insertion of tendons are well marked.
Harpagornis moorei. Pl. VII., Figs. 5 and 6.
The following are the measurements of this important bone:—
|Circumference at its proximal end, the calcaneal ridges included||4.09|
|Circumference at its distal end||4.28|
|Circumference of shaft where thinnest||1.90|
In its general form also, this bone resembles in its main features that of Aquila, except being somewhat more robust.
The shaft at its upper end is expanded and transversely flat, gradually becoming narrower, and assuming towards its middle a trihedral shape, after which it flattens again above the fore and aft canal, between the middle and outer metatarsal, near their distal ends.
In Circus the trihedral portion of the bone is much longer, even in comparison to its whole length, than either in the fossil bone or in Aquila.
The form and position of the trochlear condyles agree more closely with Circus, they being broader and with a larger space between them than in Aquila.
The tuberosity for the insertion of tibialis anticus is remarkably developed, another proof of the great power the fossil bird must have possessed.
The ectocondylar concavity is well marked, far more than in the recent species, in both of which the outer side of the proximal surface is almost plane.
The three tendinal grooves between the calcaneal processes and the inner posterior ridge are deeply excavated, much more than in Aquila audax. Half-way down the shaft they unite to form one concave channel, which, above and close to the process for the attachment of the metatarsal of the back toe, runs out to a flat surface.
The two fore and aft foramina in the upper part of the bone, in the grooves near the base of the anterior intercondylar prominences, are well marked.
The surface of the bone running from the outer margin of the ectocondylar ridge down to the outer condyle is very broad and flat, as in Aquila, having its greatest diameter in the middle portion of the bone, thus forming the base of its trihedral form.
In Circus the base of the bone is situated more in its posterior portion, the ridge running towards the centre of the anterior portion of the shaft, giving the latter a triangular form for about two-thirds of its entire length.
|Circumference at its proximal end, ridges included||3.78|
|Circumference at its distal end||4.02|
|Circumference of shaft where thinnest||1.78|
The shaft of this bone, if we compare its total length with that of H. moorei, is generally narrower; this is most conspicuous above the deeply excavated process for the attachment of the back-toe metatarsal, where the shaft is thinnest.
The ectocondylar ridge is also more pronounced, by which the shaft assumes a more triangular form than it possesses in the larger species. The two posterior ridges on both sides of the concave channel are more sharply defined, so that the latter is deeper than in H. moorei, approaching in form more that of the Australian Eagle.
I already observed, in the preliminary remarks, that our search in the turbary deposits of Glenmark after the humerus of this species had been unsuccessful, but that we obtained a fragment of that bone from the postpliocene alluvium on the banks of the Glenmark Creek, about one mile above Glenmark. This fragment consists of the greater portion of the shaft, the proximal and distal extremities being broken off.
The shaft where thinnest has a circumference of 2.20 inches, or 0.15 inch more than the same bone of Harpagornis assimilis, of which we possess a perfect specimen.
It doubtless belonged to an adult bird, and, if restored, would be about an inch longer than the smaller species.
Harpagornis assimilis. Pl. VIII., Figs. 1 and 2.
|Circumference of proximal end||4.49|
|Circumference of distal end||4.10|
|Circumference of shaft where thinnest||2.05|
This important bone, with the exception of a small, portion of the radial crest, is quite perfect. In its general outlines it has, like the other portions of the skeleton, great affinities both to Aquila and Circus.
The shaft is not so straight as that of Aquila, having below the lower termination of the radial crest an outward bend, which is also well exhibited in Circus. At the same time, the proximal extremity is more curved towards the ulnar side in the fossil bone.
The shaft at one-third of its total length above its distal end is nearly round in a transverse section, a feature it has in common with Circus, whereas the shaft of Aquila is more subelliptic.
The articular head forms a more distinct tuberosity than in Aquila, which is also observable in Circus; a broad groove dividing it from the ulnar crest, which advances considerably over the pneumatic foramen.
The radical crest being partly broken off, its whole extent cannot be ascertained. The ridge forming the boundary of the large depression for the insertion of the pectoralis major is well marked.
The articular convexities of the distal extremity are also of considerable size, and well carved out; the pits for the attachment of the muscles are large and deep, all tending to prove that Harpagornis possessed considerable power of flight.
This bone has the following dimensions:—
|Circumference at its proximal end||3.15|
|Circumference at its distal end||2.52|
|Circumference of shaft where thinnest||1.77|
In comparing its total length with the corresponding bone in Aquila audax, as given in the table of measurements, it will be seen that it is only 0.68 inch longer, but that it is distinguished from it by its considerable thickness and the greater expansion of both articular ends. This is most conspicuous when examining the proximal surface, but, considering the great breadth of the distal end of the humerus, quite a natural consequence.
The anconal side of the shaft is rather flatter than in Aquila, so that the bone does not exhibit quite such a great curve as the latter.
The quill knobs are obliterated.
Harpagornis assimilis. Pl. VIII., Figs. 3 and 4.
|Circumference of proximal end||3.00|
|Circumference of distal end||2.32|
|Circumference of shaft where thinnest||1.48|
The pachydermal character of the genus under consideration, when compared with Aquila audax, is well exhibited in this ulna, because, being actually shorter than the corresponding bone in the Australian species, it is much shorter in all its proportions.
The ulna of this species being better preserved than that of Harpagornis moorei has been figured by preference. The two rows of quill knobs, and
principally the one on the ulnar side, are well seen, as well as the intermuscular ridge on the palmar side, and the flat processes for the attachment of the muscles.
Of the left radius of Harpagornis moorei we possess only a fragment of the proximal side. The proximal end is well expanded, and the tubercle for the insertion of the biceps stands considerably forward, the shaft becoming afterwards very flat towards the ulna, not being so much bent as in Aquila.
The radius of Harpagornis assimilis (Pl. VIII., Figs. 5 and 6), which has a total length of 7.62 inches, against 7.90 inches for the corresponding bone in Aquila audax, is, like the ulna, of much stouter proportions. It is more bent towards its distal extremity, so that the same stands at a greater angle to the shaft than any of the recent species.
Metacarpus. Pl. VIII., Fig. 7.
Only one specimen of the metacarpus belonging to Harpagornis assimilis was obtained, being in an excellent state of preservation.
It is not only a little longer than that of Aquila, but also much stouter in its proportions. This is most conspicuous in the medius metacarpal and the proximal end.
The process for the attachment of the index phalanx is broad and heart-shaped, and the two principal intermuscular ridges upon the medius metacarpal enclose a broad and well-defined channel.
Pelvis. Pl. IX., Figs. 1, 2, and 3.
In my introductory remarks I stated that all search after the pelvis of Harpagornis moorei had been unsuccessful, but that we were fortunate enough to obtain this important bone of the smaller species H. assimilis when excavating the other bones belonging to the latter.
Last year, when visiting the Colonial Museum in Wellington, I observed amongst the specimens of our extinct avi-fauna a perfect pelvis, which, on examination, I assigned to Harpagornis. Dr. Hector, at my request, allowed me to take this fine specimen with me for comparison and description. After placing it near the pelvis of H. assimilis, with which it agreed in all main points except its larger size—bearing the same proportion as the bones of H. moorei do to those of the smaller species—I had no hesitation in assigning it at once to the former.
This compound bone, belonging to a fully-grown but still young individual, has all the characteristics which belong to the pelvis of a diurnal raptorial bird, some of the complex features, owing to its enormous size, being developed in a most remarkable degree. It combines great strength with lightness and
elegance of form, of which the drawings attached to this memoir will convey an accurate conception better than words can do.
In the following pages I shall offer a description of the larger and perfect pelvis, which I assigned to Harpagornis moorei, whilst the references to that of the smaller H. assimilis will prove the close generic, if not specific, relations of both.
In comparing the pelvis of H. moorei with those of Aquila audax, the wedge-tailed Eagle of Australia, and of Circus assimilis, the Harrier, and Hieracidea novŒ zealandiŒ, the Sparrow-Hawk of New Zealand, as shown in the following table, the striking difference in size becomes at once manifest.
|Pelvis of||Greatest Length.||Greatest Breadth.|
|Hieracidea novŒ zealandiŒ||2.00||1.13|
When examining this table of measurements another peculiar feature of the fossil bone will present itself to our attention, namely, its great length when compared with its breadth; whilst in the three recent species the double breadth is more than the length, in Harpagornis it is considerably less. This peculiarity is produced principally by the greater steepness of the pelvic roof and by the comparatively greater length of the ilio-ischial plates; moreover, it is also higher in proportion than any of the recent species of Diurnal Raptores with which I could compare it.
When viewed from below the space formed by the hind part of the neurapophysial crest and the two ilia has an oval shape; whereas in the three recent species previously alluded to it is shorter, more open, and semicircular (a).
Beginning with the first sacral vertebra, we observe that the articular surface of its centrum is broader in a transverse than in a vertical direction, 0.69 inch by 0.58 inch. The neural canal has an oval form, its largest diameter, 0.21 inch, being in the vertical line, in this respect resembling Circus; whilst in Aquila, and still more in Hieracidea, the canal approaches the circular form.
The prezygapophyses (pr.) are of middle size and stand forward, their articular surface of a rounded shape, being almost plane. The neural spine is broad and strong at its base, gradually contracting, and forming only near its coalescence a small neurapophysial expansion lying between the iliac plates (n).
A broad and deep ilio-neural opening is formed on each side of the spinal plate, having a greater vertical than lateral extent, and here again differing from the pelvis of the three recent species previously alluded to, the roof
formed by the iliac plates of Harpagornis being consequently considerably steeper.
The surfaces for the head of the two free sacral ribs are strongly developed, the iliac roof extending, however, a little beyond them.
The under surface of the first sacral centrum in its anterior portion is slightly carinate, whilst the centres of the two succeeding ones are rounded, the edges of their articular surfaces being well raised, the posterior one of the third centre the least; after which they flatten and expand to the beginning of the interacetabular region, contracting again to its termination, and possessing a transversely concave, shallow, inferior surface, being broadest near the anterior articular surface of the seventh vertebra.
From the eleventh to the fourteenth they still diminish in breadth, and now exhibit a low but well marked inferior ridge, running out before the last sacral vertebra is reached.
The parapophyses of the third to the sixth sacral vertebra are anchylosed to the lower border of the ilia, forming four interapophysial vacuities on both sides; of these the last parapophysis is the strongest and thickest, standing at right angles to the direction of the axis of the vertebral column.
There is a short parapophysial process starting from the seventh vertebra (the first of the four next vertebræ forming the interacetabular region), which has a downward direction, and is still attached on the left side of the pelvis to the inner edge of the head of the pubic bone (h).
In the pelvis of Harpagornis assimilis this process does not exist, and it resembles in this respect the recent species previously used for comparison. Of the parapophyses of the last four vertebræ, forming the postacetabular region, the first one belonging to the eleventh sacral centrum is a filamentary bone (m) joining the second round and strongest parapophysis, which abuts against the innominate, and with which the posterior ones are also connected by their distal ends.
Of the interapophysial vacuities the first, second, and fourth are elongate, whilst the third and largest is more circular. In the smaller pelvis of Harpagornis assimilis these vacuities are not relatively, but actually, larger than in that of H. moorei.
The coalesced distal portion of these parapophyses runs in an oblique angle from the inner region of the ilia to the abutment of the twelfth sacral centrum, the space between this distal line and the upper side of the ischiadic foramen, below the pelvic disk, being spanned over by a thin deck of bone (d), perforated by a large oval opening 0.48 inch in its largest diameter, which runs parallel to the main axis of the pelvis, and is situated on each side behind the upper and anterior wall of the ischiadic foramen.
The last sacral vertebra of H. moorei is not yet quite anchylosed to the
foregoing vertebra, thus shewing that it belonged to a not quite adult individual; on the other hand, in the pelvis of H. assimilis the articular surfaces of these two last vertebræ are well anchylosed, and the junction of the parapophyses with the lower border of the ilia in its antacetabular part is also well accomplished, which is not quite the case in the pelvis of the larger species under review, so that we may safely assume that the former belonged to a full-grown mature specimen.
The gluteal ridge is decayed in H. moorei, but is well developed and preserved in the smaller species, the gluteal process forming a rounded knob (g), which rises well above the pelvic disk, whilst in Aquila this process has a convex form, directed downwards, and standing well in advance of the ilia. Of the recent species Circus resembles most, in this respect, the extinct gigantic form.
The pre-acetabular iliac plates unite about one-third from their anterior end above the summit of the sacral ridge, diverging again after having been united for 1.70 inch to form a small interposed neural expansion, anteriorly lying scarcely below the upper border of the iliac plates. In this respect it resembles Aquila, whilst in Hieracidea, and still more in Circus, the neural interposition is continuous all the way, but is narrowest in the region where, as observed, the iliac pre-acetabular plates meet in Harpagornis.
The ischium is very strongly developed at the back part of the acetabulum, as might be expected in a bird of such great strength. The tuberosity of the ischium, a roundish flat process, 0.72 inch from its posterior termination, rises conspicuously above its lamelliform surface (k). The posterior termination of the coalesced ischium and ilium is not rounded off, as in Aquila, but has a rather acute form, which, of recent species, Circus, and still more conspicuously Hieracidea, also possess.
The pubic bone, after forming the lower boundary of the obturator notch, gradually loses its trihedral shape and assumes a vertically flattened form, continuing to run for some distance parallel with the ischium; however, as in both specimens its posterior portion is broken off, I cannot say how far it may have extended. In any case it is longer than in Aquila.
A thin plate of bone, closely connected with the lower border of the ischium and gradually thickening, runs to the termination of that latter bone. At its beginning it forms the posterior boundary of the obturator foramen, and fills up the space between the ischium and the pubic bone.
The subacetabular fossæ (f), which are very shallow in Aquila and the Diurnal Raptores now living in New Zealand, are deeply excavated. The pelvic disk is a strong bone separated on each side by a well-marked line from the hind part of the neurapophysial crest, which rises well above it, the latter showing, like all the rest of the bones of which the pelvis is formed, a
remarkable development of all the principal features to be observed in the pelvis of the smaller recent Diurnal Raptores.
Finally, I wish to observe that the pelvis of Harpagornis moorei, from Otago, has still some of its integuments and ligaments attached, of which the lining membrane on the walls of the acetabulum are best preserved; whereas the more fragmentary bone of H. assimilis is in the semi-fossil condition in which all the bones from the remarkable turbary deposits of Glenmark are usually found.
Since my former paper a second ungual phalanx has been obtained, which, applying the same mode of measurement previously used, is 2.75 inches long, and has a circumference of 2.92 inches at its proximal end. It is the third phalanx, and belongs to the second or inner toe of the right foot.
Amongst the smaller bones lately excavated I found also the second phalanx, with which that latter ungual phalanx articulates.
The pachydermal character, even in these toe-bones, is well sustained, and the form and peculiarities of the articular ends, and the large concavity behind and below the trochlear joints of the distal end, are developed in a striking degree.
Of Harpagornis assimilis we possess, as previously observed, several phalanges.
Description of Plates VII.—IX.
Plate VII.—Figs. 1–2. Tibia of Harpagornis moorei.
3-4. Metatarsus " "
VIII.—Figs. 1–2. Humerus of Harpagornis assimilis.
3-4. Ulna " "
5-6. Radius " "
7.Metacarpus " "
IX.—Figs. 1–3. Pelvis of Harpagornis moorei.