re 6” 2 TC eae
4 vt

Pieenat sreteter
et ES pari

ee Saree :

ar

ee?

—
—)

HARVARD UNIVERSITY

a

LIBRARY

OF THE

Museum of Comparative Zoology

MUS. COMP, ZU0OL
LiBRARY,

sep 8 1964

HARVARD
UNIVERSITY

pela

YALE PEABODY MUSEUM

oF NatTurAL History

Number 78 April 15, 1964 New Haven, Conn.

ONTOGENY AND EVOLUTION IN THE MEGAPODES
(AVES: GALLIFORMES)*

Grorce A. Ciark, JR.

DEPARTMENT OF BroLocy, YALE UNIveERsITY2

INTRODUCTION

Unlike all other birds, megapodes of Australia and the
Pacific Islands incubate their eggs in mounds or holes by heat
from fermentation, sun, or volcanic activity. Furthermore, meg-
apodes are unique among birds in being able to fly weakly
on the day of hatching and in having no parental care for
young.

These and other reptile-like aspects of megapode reproduc-
tion have been interpreted in two contradictory ways. Some
authorities (e. g. Portmann, 1938, 1950, 1955) have main-
tained that megapodes are the most primitive of living birds,
while others (e. g. Pycraft, 1910) have stated that the similari-
ties of megapodes and certain reptiles are due to convergent
evolution. A related and also unresolved problem has been the
primitiveness of mound-building megapodes relative to those
laying their eggs in holes (cf. Frith, 1962).

1 This study is based on a dissertation presented for the degree of Doctor
of Philosophy at Yale University.

2 Present address: Department of Zoology, University of Washington,
Seattle, Washington.

2 Postilla Yale Peabody Museum No. 78

Despite their anomalous nidification and precocity of young,
megapodes as adults are structurally similar to other members
of the Order Galliformes (e. g. pheasants) as exemplified by the
overlap in adult osteological proportions (cf. data of Verheyen,
1956). Morever, the family Megapodiidae and the New World
gallinaceous family Cracidae (chachalacas, guans, curassows)
are particularly difficult to separate at the family level on a
morphological basis (cf. Miller, 1924). Megapodes and cracids
have been classified as the two most primitive gallinaceous fami-
lies (Huxley, 1868; Peters, 1934).

Unlike the megapodes, most other species of Galliformes have
a simple nest on the ground, but some pheasants, cracids, and
the highly aberrant hoatzin (Opisthocomus) nest in trees.
Since avian development often varies in accord with nidification,
it was anticipated that the study of megapode embryos and
juveniles would reveal clues pertinent to the analysis of mega-
pode phylogeny.

Prior to this study, the only detailed accounts of structure
of embryonic or juvenile megapodes were based on the genus
Megapodius (cf. Pycraft, 1900; Friedmann, 1931; Becker,
1959). These previous investigations had led to contradictory
conclusions on the homologies of the early plumages (cf. Nice,
1962) and on the phylogenetic origins of the family (cf.
Frith, 1962). The object of the present study was to attempt
to resolve the controversy over the phylogeny of megapodes
through examining the morphology of embryos and juveniles
representing several genera of megapodes.

MATERIALS AND METHODS

Specimens. Thirty embryos of the megapodes Talegalla
jobiensis and Leipoa ocellata were studied (Tables 2, 3).
The 11 Talegalla embryos were collected for this investigation
in New Guinea during 1959-60 by E. T. Gilliard and 8. D.
Ripley in separate expeditions. The 19 Leipoa embryos were
collected by me during 1960 in the mallee about 25 miles north
of Griffith, New South Wales, Australia; the collecting area
was favorable in having an unusually high density (Frith,
1959) of active Letpoa mounds which were as frequent as one
per 50 acres in the limited suitable terrain. The eggs of Leipoa

Bree rvrars f

SEP 8 1964
See As ARVARD
April 15, 1964 Ontogeny and Evolution i Sv 3

were marked as found and allowed to incubate in the mounds.
Subsequent collections provided accurate ages for three embryos
and minimal ages for certain others (Table 2). The tempera-
ture is ordinarily relatively uniform for Leipoa eggs together
in a mound (Frith, 1959), and the time between egg layings
by a hen is usually four or more days (Frith, 1959, and a
few cases in this study). Thus when actual or minimal age of
one embryo was known, minimal ages of progressively larger
embryos in that mound were estimated by adding four more
days for each. Since the first eggs were probably laid on
September 4 or later, as judged from previous years (Frith,
1959), some specimens (Nos. 12, 16, 19, 1195, of Table 2)
could be assigned presumed maximal ages; smaller embryos
from the same mounds could also be assigned maximal ages,
again using the hypothesis of four or more days between egg
layings in a mound.

Kighty-two juvenile specimens (including 79 study skins)
of megapodes were examined at the American Museum of Nat-
ural History and Yale Peabody Museum (YPM). Among these
were the following species (with numbers of each) : Megapodius
freycinet (59), M. laperouse (5), M. pritchardii (1), Macro-
cephalon maleo (1), Aepypodius arfakianus (2), Talegalla
cuviert (2), T. fuscirostris (4), T. jobiensis (4), Alectura
lathami (3), and Leipoa ocellata (1). More than 140 embry-
onic and juvenile specimens representing 22 genera of non-
megapode Galliformes were used for comparison.

Methods. Characters were chosen for interspecific morpho-
logical comparisons according to 1) potential accuracy of
description or measurement, as determined by reproducibility
in repeated examinations, and 2) potential phylogenetic signifi-
cance demonstrated by the extent of intergeneric variation and
its possible phylogenetic interpretations.

Measurements. Measurements, selected for their applica-
bility over a wide range of sizes, were:

Wine: folded and flattened, with a rule from the anterior
edge of the wrist to the end of the manus, or, in feathered speci-

4 Postilla Yale Peabody Museum No. 78

mens, to the most distant tip of a remex. Due to the distal
shriveling of the ensheathed remiges of embryonic megapodes,
wing lengths over 20 mm (Tables 2, 3, 5) were rounded to the
nearest 5 mm.

Tarsus: with Vernier calipers from the posterodorsal sur-
face of the ankle along the tarsometatarsus to the level of the
proximal surface of the base of the hallux.

CuLMEN: with calipers from the tip to the most posterior
unfeathered point on the dorsal midline.

HuMERvUs; RADIUS: respective maximal lengths with calipers.

Tuirp (MIDDLE) DicIT: straightened, with a rule from the
tip to the most distal point of webbing connecting with an
adjacent toe.

Megapode embryos Nos. 1, 20, and 21 (Tables 2, 3) were
too immature to measure by these criteria.

Values in the Tables (2, 3, 5) are means of two measure-
ments, each of which, unless otherwise noted, was rounded to
the nearest millimeter. Estimated maximal ranges of variation
in measuring were + 1 mm for dimensions of 2 to 10 mm and
up to + 3 mm for dimensions of 150 mm; these maximal esti-
mates were derived from the ranges in duplications of more than
500 measurements. Among the factors possibly affecting the
accuracy In measuring were 1) unavoidable errors in aligning
and reading calipers and rule, 2) structures changing in shape
as well as length, 3) variations in the positions of parts of
specimens at fixation, and 4) (for anatomical specimens) rate
of fixation with 10 per cent formalin.

Weights (Tables 3, 5), recorded by collectors in the field,
are given only for fresh specimens, as weights of preserved
specimens would be unreliable. The weights and their cube
roots were plotted on arithmetic and double logarithmic graphs
against the various linear dimensions; if any one of the weights
for Talegalla were grossly in error, this would have been seen
as a point lying relatively far from the plot for the other
points. Factors possibly influencing accuracy in weighing in-
clude uneven removal of the yolk sac of embryos before weigh-

April 15, 1964 | Ontogeny and Evolution in Megapodes 5

ing, uneven drying of surface moisture on the feathers of
embryos, and variations in the contents of the digestive tract
of juveniles.

MORPHOLOGY OF EMBRYOS AND JUVENILES

Time in embryonic development. Young embryos of the
megapode Leipoa developed slowly compared with embryos of
phasianids (e. g. Gallus, Phasianus, Coturnix), as shown by
the much later occurrence of the first gross appearance of
egg tooth, feathers, labial groove, etc., in Leipoa (Table 1).
Through the first 20 days, these Leipoa embryos attained a
much smaller absolute size than did embryonic chickens (domes-
tic G. gallus) as illustrated by comparing linear dimensions of
Leipoa and chickens (Fig. 1; Tables 2 and 3). As an example,
after 20 days of incubation the wing of an embryonic Letpoa
was less than 50 per cent as long as that of a chicken (Fig. 1).

The normal incubation period of Leipoa is generally at least
twice as long as that of known phasianids or turkeys (cf.
Table 1; see also Frith, 1959, on Leipoa, and Romanoff, 1960,
on phasianids). This lengthy incubation period of Leipoa ts

Tasie 1. Time of certain gross morphological changes in embryos of the
megapode Leipoa and of phasianids. Age in days after laying of the

egg.
Leipoa Phasianus Coturnix Gallus
age age age age
Egg tooth formed ..... 21-22 9 5-6 61-7
Labial groove formed .. 21-22 ? 2 10
Heathers) appear)... .-1-- 11-21 9 5-6 612-7
Toes are first separated 11-21 10-12 7-8 8-9
Seales appear on legs .. 29-54 13 8?-9 11-12
Eyelids come together .. 29-61 15 10-11 13
Labial groove lost ..... 29-61 ? ? 19
latching 72 o-.6 25 60-73 23-24 16 20-21

Sources of data: Leipoa ocellata, specimens of this study; Phasianus
colchicus, Fant, 1957, and Westerskov, 1957; C. coturnix japonica,
Padgett and Ivey, 1960; domestic G. gallus, Hamilton, 1952.

6 Postilla Yale Peabody Museum No. 78

related to both the slow early development and the large size
at hatching (see p. 27 for discussion of the effects of incuba-
tion temperatures).

: /

Gallus

Wing (mm)

10 30 50 70

Days

Figure 1. Chronological growth of the wing in Leipoa ocellata and do-
mestic G. gallus (data from Tables 2 and 3). Curves showing length against
time were fitted by inspection and should not be considered as quantita-
tively accurate.

Relative proportions and growth. At hatching Talegalla
and Leipoa are about two to 15 tinfes heavier than other newly
hatched Galliformes of the genera Coturnia, Colinus, Phasianus,
Gallus, and Meleagris (Lyon, 1962; Westerskov, 1957 ; Roman-
off, 1960; see also Table 6). It is of interest that Letpoa and
Talegalla at hatching have proportions and size like those of
adult C. coturnix japonica (Table 5). The genus Megapodius
is intermediate in hatching weight (Table 5) between Talegalla
and phasianids or turkeys.

As a means of comparing changes in proportions during the
growth of different species of the Order Galliformes, arithmetic
and double logarithmic plots (e. g. Figs. 3, 4, 5) were prepared
using the linear measurements of embryos and juveniles (data
of Tables 2, 8, and 5). Such proportional growth was described

April 15, 1964 | Ontogeny and Evolution in Megapodes 7

approximately in certain cases by using the conventional allo-
metric equation, Y = AX®, or the equivalent form, log Y = log
A+B log X, where X and Y are the values of two dimensions.
A and B (Table 4) were calculated using Bartlett’s method as
described by Simpson, Roe, and Lewontin (1960). B values for
different species were compared using a modified t-test (Simp-
son et al., 1960). Correlation coefficients for the sets of data
expressed as B values in Table 4 were all significant at the
0.001 level.

To compare growth of linear dimensions relative to total
body size in different species, the cube root of weight was used
as one criterion for body size (see Amadon, 1943, for the

Taste 2. Data for specimens of Leipoa ocellata. All are embryos except
1195. For procedures of measuring, see text. All lengths in milli-
meters. Estimated ages in days. Symbols: S, specimen number; W,
wing length; T, tarsal length; C, culmen length; H, humerus length;
R, radius length; Td, length of third digit; m, male; f, female;—,
observation could not be made.

S) WwW Ab Cc H R Td Sex Age
1 = = — —- 11-?
2 7 3 3 5 3 3 = 21-22
3 9 6 5 6 6 4 — 22
4 12 8 7 10 10 6 29
5 13 8 8 10 9 a — 2-54
6 20 12 9 13 13 11 m 19-58
uf 20 13 9 13 12 10 m 2-55
8 25 14 10 16 14 12 f 2-59
4) 25 15 10 17 18 13 m ?
10 40 19 12 20 20 15 m ?
11 45 20 11 23 20 15 m ?-62
12 45 21 12 22 21 15 m 45-70
13 55 22 12 22 23 16 fi 48-61
14 70 24 14 28 26 19 m 2-63
15 70 23 13 27 25 17 — 52-64
16 75 25 14 28 27 18 m 49-74
17 80 26 13 28 29 20 if; 56-69
18 80 24 14 31 28 20 f 2-67
19 85 26 14 29 30 20 f 60-73
1195* 115 28 — 38 37 24 _— 2-73

* This specimen, found dead in the field, was lacking its head.

8 Postilla Yale Peabody Museum No. 78

LEGEND
8 a eae
B Talegalia °
Us @ Megapodius > 2 a
O Aepypodius
6 ; Vv
V Alectoris v
© Phasianus e a
5 ® ¥ 4

Cube Root of Weight

Culmen (mm)

Figure 2. Relationships of the cube root of body weight in grams to the
culmen length in the megapodes Talegalla jobiensis, Aepypodius arfa-
kianus, Megapodius freycinet and the phasianids Alectoris chukar and
Phasianus colchicus. All data from this study except that for Phasianus,
for which mean values for males were taken from Westerskov, 1957. See
text for discussion,

explanation of this procedure). Since weights were unknown
for most specimens, a linear criterion for body size was also
chosen. As the culmen length had a relatively direct relation-
ship to the cube root of body weight over a fifty fold range of
weights for eight specimens of embryonic and juvenile T'alegalla
jobiensis and for six juvenile specimens of the phasianid Alec-
toris chukar (Fig. 2), culmen was selected as a convenient
linear measure for body size in these specimens. Moreover, simi-
lar analyses revealed that culmen is a relatively good measure
for body size in embryonic chickens (10-21 days; matched
lengths from figures of Hamilton, 1952, with weights from
Romanoff, 1960) and in juvenile Phasianus colchicus from
zero to nine weeks posthatching (Westerskov, 1957; see also
Fig. 2 of this study). Since the culmen is a less sensitive and
less accurate indicator of body size than is the cube root of
body weight, certain interspecific differences have possibly gone
undetected due to the use of culmen as a major standard for
body size.

April 15, 1964 | Ontogeny and Evolution in Megapodes 9

The scales for the cube root of weight in Figs. 3, 4, 5, are
calculated from a mean value of 3.24 for the ratio of culmen
length to the cube root of body weight in grams for the eight
weighed specimens of T'alegalla jobiensis. Due to the relative
imprecision of culmen measurements (compared with weights)

Taste 3, Data on specimens of Talegalla jobiensis and domestic G. gallus.
Nos. 1196, A, B, C, S, T, U, V, W are posthatching specimens. For
procedures of measuring, see text. All lengths are in millimeters.
Weights in grams. Ages in parentheses are estimated from stages in
Hamilton (1952). Symbols: S, specimen designation; W, wing length;
T, tarsal length; C, culmen length; H, humerus length; R, radius
length; Td, length of third digit; m, male; f, female;—, observation
could not be made.

Talegalla:

SS) WwW as Cc H R Td Sex Weight
21 — — = —_ —- _ —_— 3.5
22 9 6 5 6 6 4 — 4.7
23 11 6 6 8 7 6 = 5
24 12 9 6 9 8 7 = =
25 18 13 8 11 12 9 m 14.3
26 20 15 g) 14 12 11 m 22
27 45 23 12 20 19 15 = =
28 80 30 16 29 28 24 — 101
29 100 30 15 31 29 24 f 108
30 100 33 17 34 33 27 =
1196 115 38 16 38 35 28 — —

A 115 35 16 — — — t 125

B 160 47 21 —— — = m 292

C 164 49 18 — — — — =

Gallus: Age (days)

N 8 5 5 5 4 4 (10-11)

O 10 8 uf 7 6 6 (11-12) .
P. 11 9 6 8 7 t (12)

Q 25 17 10 12 10 16 (19)

R 27 19 10 13 11 17 (19-20)

S) 85 26 15 29 25 24 —

T 110 36 18 35 33 30 —

U 135 44. 21 40 36 35 53

Vv 170 51 24 AT 44, 42 63

W 160 54 24 50 46 41 =

Note: Specimens A, B, C are study skins.

10 Postilla Yale Peabody Museum No. 78

50] 50]
LEGEND LEGEND
Q& Leipoa 1°) Q Leipoa
B Jolegalla ie) B Jolegalla
O Ggallus O Ggoallus 8
404 404
a ge oO
€ 9) Pe
— | S
= me — 301 fe)
3 a : e
rC)
& 2 3 on
a A =
| te
20 . E 20} as
aN Q me
mA
A
‘ 2
10 A 8 10 ms
ad
OA
a =
A O A 3)
a SS ee
5 10 15 20 25 5 10 15 20 25
Culmen (mm) Culmen (mm)
————_
SiS ia entG. 2 Mine GS
AY Weight AY Weight

Figure 3A, (Left) Growth of the radius relative to the culmen in
Leipoa ocellata, Talegalla jobiensis and domestic G. gallus (see alsa
Fig. 4). B. (Right) Growth of the third digit relative to the culmen in
these three species (see also Fig. 4).

Cube roots of weights in grams calculated by the method indicated in
text(ps 9):

and probable interspecific variations in the mean ratio of cul-
men to the cube root of body weight, the cube root values in
Figs. 3, 4, and 5, are probably not precise for individual speci-
mens shown on the graphs? nevertheless, these cube roots of
weights help to indicate, in an approximate way, the relative
growth of the different species.

As shown by either arithmetic (e. g. Figs. 3, 5A) or loga-
rithmic plots (Fig. 4), growth of linear dimensions relative to
culmen in the two species of megapodes is generally similar to
that of Gallus (see also Table 4). It should be emphasized,
however, in view of the necessarily small sample sizes and
inherent limits of accuracy in measurement, that these analy-
ses tend to mask certain differences in relative growth. For
example, in embryonic chickens the radius (Fig. 4A) and
humerus temporarily have lower rates of relative growth fol-
lowed again by higher rates (this study) ; the data of Roman-

April 15, 1964 Ontogeny and Evolution in Megapodes 11

off (1960: 1146) show that the slow growth of these structures
in chickens occurs about 14-17 days of incubation. As a con-
sequence, the radius and humerus of chickens near hatching
are a few millimeters shorter than those of similar-sized embryos
of Talegalla or Letpoa (data in Tables 2, 3). In addition,
measurements of three juveniles of the phasianid C. coturni«
japonica revealed for this form also a slow mean rate of embry-
onic growth of radius and humerus relative to other dimensions
followed by increased relative rates after hatching. The rela-
tively short radius and humerus of Gallus and Coturnix in
older embryos and at hatching are possibly adaptive in prevent-
ing premature flying of the young birds; such an adaptation
would be analogous to the retarded development of remiges in
juveniles of forms such as petrels and hawks. No trace of a
relatively slow embryonic growth of radius and humerus was
found in the megapodes.

Culmen measurement in the utilized samples covers a rela-
tively small range (less than 20 mm), but this handicap is offset
somewhat by the utility of this measurement for study skins.
The culmen is measured linearly over a curved surface but
nevertheless is empirically useful. In measuring the culmen of

Taste 4. Interspecific comparison of allometric growth of dimensions rela-
tive to culmen. None of the interspecific differences in exponent is sta-
tistically significant. See text for details.

Exponent (B) with Size
95 per cent Coefficient of

Dimension Species _ confidence interval (A) sample
Tarsus Leipoa 1.6 + 0.2 0.24 18
wv Talegalla 1.5 + 0.3 0.20 - 13
2. Gallus 1.4 + 0.2 0.18 10
Humerus Leipoa as) te Oe! 0.19 18
2 Talegalla 1.4 + 0.2 0.15 10
2 Gallus 1.4 + 0.2 0.19 10
Radius Leipoa 1.6 + 0.2 0.26 18
2 Talegalla 1.4 + 0.2 0.17 10
2 Gallus 1.5 + 0.2 0.27 10
Third digit Leipoa 1.5 + 0.3 0.24 18
ae ed Talegalla 1.5 + 0.2 0.23 10

ee tae Gallus 1.5 + 0.2 0.20 10

12 Postilla Yale Peabody Museum No. 78

late embryonic and juvenile chickens (Gallus; Table 3), the
presence of the comb necessitated estimating culmen lengths
in eight specimens through the projection of lines from the
postero-lateral margins of the horny bill dorsally to the mid-
line; however, this approximation did not alter the interpreta-
tions as shown by using other combinations of dimensions. At
hatching in Gallus, Leipoa, and T'alegalla, the culmen may lose
up to 1 mm in length through loss of periderm, but this small
change does not affect the interpretations of relative growth.

Analogous to the shorter culmen after hatching are reduc-
tions (about 5 mm) in wing length of juveniles of these species
through loss of natal downs and also the decrease (less than
1 mm) in length of the third digit through loss of the claw
pad at hatching. Here again the interpretations of relative
growth were not affected.

Relative and proportional growth of gallinaceous wings was
too complex to permit adequate representation in a simple
equation, but, as shown by graphs (e. g. Fig. 5), relative
growth of the wing in T'alegalla, Leipoa, and other Galli-
formes was similar within the size range considered. The propor-
tional growth illustrated in Fig. 5B suggests possible interspe-
cific differences which, however, are not especially striking. Data
for the Jungle Fowl (G. gallus) were used in Fig. 5 to provide
a larger sample, but data for chickens (domestic G. gallus;
Table 3) gave similar results.

Juvenile Megapodius have an unusually short culmen con-
trasted with those of juveniles of other megapodes or other
Galliformes ; the mean ratio of culmen length to the cube root
of body weight for three Megapodius freycinet (Fig. 2; Table
5) was 2.1, compared with 3.24 for eight T'alegalla jobiensis.
Young juvenile Megapodius (Table 5) also differ from young
juveniles of Talegalla in having a longer wing relative to the
cube root of body weight.

Measurements of wing, tarsus and culmen of more than 110
other juvenile specimens representing 22 genera of non-mega-
pode Galliformes (cf. Table 5) were plotted on graphs and
compared. These species generally appear to have proportional
growths similar to those of T'alegalla, Leipoa, and Gallus.
However, a juvenile Craa rubra of the cracids (Table 5) was

Radius (mm)

April 15, 1964 | Ontogeny and Evolution in Megapodes_ 13

exceptional in having a relatively short wing (shown also by
figures of young Crawx globicera (= rubra) in Heinroth, 1931).
The shorter wing at hatching in Craw is apparently associated
with the generally less well-developed feathers (p. 24). Forms
such as ducks (e. g. Anas) which have delayed formation of
juvenal remiges show plots of alar growth quite unlike those of
Galliformes.

These analyses, although necessarily based on small samples,
indicate that embryonic megapodes undergo proportional and
relative growth analogous to that occurring up to several weeks
posthatching in phasianids. Certain forms such as Megapodius
and Crax show interesting deviations from the general gal-
linaceous conditions. Larger samples might reveal additional
interspecific differences and possibly intraspecific variations
according to individuals, sex or locality.

Some qualitative comparisons of embryos and juveniles.
Embryos of Leipoa (e. g. Nos. 2 and 19) and of chickens
shortly prehatching behaved similarly when taken from the
shell, i. e. the embryos gaped and kicked. Even Leipoa embryos

5 LEGEND 5
aR 8 LEGEND
- oe poa 40 6 Lelpoa 8
alegalla ho ° = Talegalia °
3 ©G. gallus — 30 °G. gallus °
E %
E
2 -
ca
a
=
=
=

3 5 7 10 20 30
Culmen (mm)

3 5 if 10 20 30
Culmen (mm)

—
. I. z 3 y

5 22 3.1 6.2 ii zZ A. oe
V Weight V Weight

Figure 4A. (Left) Double logarithmic plot of growth of the radius rela-
tive to the culmen in Leipoa ocellata, Talegalla jobiensis and domestic
G. gallus. B, (Right) Double logarithmic plot of growth of the third digit
relative to the culmen in these three species.

Cube roots of weights in grams calculated by the method indicated in
Texte):

14 Postilla Yale Peabody Museum No. 78

Taste 5. Comparison of dimensions of some juvenile Galliformes. Speci-

mens arranged by increasing tarsal length. For procedures of measuring,

see text. Lengths in millimeters. Symbols: W, wing; T, tarsus; C, culmen;
m, male; f, female; g, grams.

We ae We Dae

Numida meleagris (f) 20 17 10 Penelope purpurascens 105 27 12
Chrysolophus pictus 20 18 7 Megapodius pritcharditi 85 28 6
Phasianus colchicus 30 19 8 Gennaeus leucomelanos 105 28 15
Opisthocomus hoazin 40 19 12 Chrysolophus pictus 115 30 14
Gennaeus leucomelanos 50 19 10 Alectoris chukar

Syrmaticus mikado 30-20 8 (f; 154g) 120 30 16
Chrysolophus pictus 26 20 7 ‘Tragopan temmincki 130 30 14
Phasianus colchicus 35 21 10 Ortalis wagleri 110 31 15
C. coturnix japonica 90 22 12 Alectoris chukar
Alectoris chukar (£5198 g) 180 31 16
(m; 73 g) 95 22 14 £.,Mitu tomentosa 85 32 14
Phasianus colchicus 85) 2313 Talegalla fuscirostris 120 32 15
Meleagris gallopavo 45 24 10 Alectura lathami 105 33 15
Ortalis wagleri 45 24 10 Phasianus colchicus (f) 110 33 18
Ortalis vetula 70 24 12 Francolinus gularis 120 33 15
Megapodius freycinet Megapodius freycinet
(f; 63.6 g) 100 24 8 (f; 117 g) 125 33 10
Numida meleagris (m) 110 24 14 Megapodius freycinet
Megapodius laperouse 95° 25 8 (123.5 g) 130 33 11
Alectoris chukar Opisthocomus hoazin 165 34 19
(m; 121 g) 115 26 15 Macrocephalon maleo 140 35 15
Meleagris gallopavo 55 27 9 Crax rubra (m) 75 36 15
C. coturnix japonica Dendragapus obscurus
(adult) 100+ 27 13 (f) 175 36 18

considerably larger than chickens at hatching show this charac-
teristic embryonic behavior.

Meyer (in Meyer and Stresemann, 1928) noted the large fat
deposits in late embryonic Megapodius; both Talegalla and
Leipoa embryos (this study) also have subcutaneous fat bodies
distributed similarly to those of chicken embryos but covering
a wider area in embryos near hatching. These deposits in older
Talegalla and Leipoa embryos are especially well developed
laterally along the neck and beneath portions of the ventral
feather tract.

The genus Megapodius (Miller, 1924; confirmed in this
study) is unusual among Galliformes in having a small web

April 15, 1964

Ontogeny and Evolution in Megapodes

180 1@)
LEGEND
170 Leipoo &
s Tolegal/ia @
G. gallus O
160 a
(e)
150
1e)
140 fe)
130
120
a
— 10 12)
=
i= Oo
100 ae
oO
2 oO
= 90
Ao
804 as
w
“ &
60 1@)
@
50
404 a
()
304
(e.ay
204
&
104 F
a
10 20 30
Culmen (mm)
SS
3.1 6.2
Ay Weight

Figure 5A. (Left) Growth of the wing relative to the culmen in Leipoa
ocellata, Talegalla jobiensis and G. gallus. Cube roots of weights in grams
calculated by the method indicated in the text (p. 9). B. (Right) Propor-
tional growth of the wing versus the tarsus in these three species.

(mm)

WING

160 4

150 4

140

130 4

120 4

90 4

80

704

60 4

50 4

404

30 4

LEGEND
Q Le/poa
® Zolegal/a
O Ggallus

(e)

—

oO

10 20 30 40

Tarsus (mm)

15

between the second and third toes but, unlike forms such as
Leipoa, Talegalla and Gallus, none between the third and

fourth toes.

A few qualitative gross morphological changes appear at
a greater absolute weight, and, for larger embryos, at a detec-

16 Postilla Yale Peabody Museum No. 78

tably greater linear size. in the megapodes (Letpoa and Tale-
galla; this study) than in Phasianus (Westerskov, 1957) or
Gallus (structures from Hamilton, 1952, matched with weights
from Romanoff, 1960). Examples of these phenomena in T'ale-
galla versus phasianids (Table 6) include first appearance of
feathers, egg tooth, labial groove, and coming together of the
evelids.

Taste 6. Comparison of weights at times of certain qualitative morpholog-
ical changes in Talegalla jobiensis, Phasianus colchicus, and domestic
G. gallus. Weights in grams. Talegalla weights in parentheses were esti-
mated from culmen lengths using the relationship reported in the text
(ips Qe

Macroscopic Talegalla Phasianus Gallus
character weight weight weight
Hirst: appearance,» feathers, 322... - 3.5 - 4.7 0.7-1.7 0.4- 1.2
First appearance, egg tooth ........ 4.7 - 5.0 0.7-1.7 0.4- 1.2
Formation of separate toes ......... 3.5 - 4.7 1.4-4.8 0.7- 2.3
Formation of scales on legs ........ (5.5)-14.3 3.2-5.8 2.3- 7.3
Eyelids coming together ........... 22- (40) 4.7-8.5 §.2-11.0
Pbatichin gyre a cine cine wislovnieteye, Saieusieysteiss 110+ 23 33

Sources of data: Talegalla from this study; Phasianus from Westerskov
(1957); Gallus morphology from Hamilton (1952) combined with Gallus
weights from Romanoff (1960: 1147).

Tarsal seutellation. My observations on the tarsal scutella-
tion of megapodes support the findings of Ogilvie-Grant (1893).
Megapodius, Aepypodius, and Talegalla are alike in having a
single row of large scutes down most of the foresurface of the
tarsus (tarsometatarsus), but Aepypodius has two rows dis-
tally. Alectura and Leipoa have two rows of large scutes down
the foresurface, while Macrocephalon has many small scutes.
Tarsal scutellation is similar in juveniles and adults within a
species of megapode.

Turkeys, many phasianids and some cracids have two rows
of large scutes on the foresurface, while many cracids possess
only one row; Opisthocomus has many small scutes.

Feathering of the oil gland. Talegalla jobiensis has a naked
oil gland (no feathers on the tip; Fig. 6, this study) and thus

April 15, 1964 | Ontogeny and Evolution in Megapodes 17

y My
NY SZ

Figure 6. Oil glands of domestic G. gallus (19 day embryo; ca. 5.5 X ),
Magapodius laperouse (YPM 89; juvenile; ca. 3X), and Talegalla
jobiensis (No. 29, embryo; ca. 1.5), from left to right. Dorsal view.

is like Alectura and Leipoa (Miller, 1924; confirmed in this
study). In contrast, Megapodius laperouse (Fig. 6, this study)
has a tufted oil gland as was reported by Miller (1924) for
other species of Megapodius and for Macrocephalon. Most
Galliformes, excluding megapodes, have tufted oil glands (Fig.
6 of this study; Miller, 1924; see also Table 7 for a summary
of this character in other birds).

Eutaxy. Unlike other gallinaceous families, megapodes have
variation in eutaxy (presence of the fifth secondary; Steiner,
1918; Miller, 1924). As anticipated from reports on allied
species (i. e. Alectura and Leipoa; Miller, 1924), Talegalla
jobiensis is eutaxic (this study). Both Talegalla and Leipoa
are eutaxic at the first embryonic appearance of the second-
aries. Megapodius laperouse (YPM 89) is also eutaxic, but
M. pritchardii (Pycraft, 1900) and some (but not all) mem-
bers of M. freycinet (Steiner, 1918; Miller, 1924) are diasta-
taxic (lacking the fifth secondary). Macrocephalon is also
diastataxic (Miller, 1924). In contrast, all other Galliformes,
including chickens, are eutaxic (Miller, 1924; see also Table 7
for a summary of diastataxy and eutaxy in other birds).

Carotid arteries. In agreement with the data reviewed by
Glenny (1955) for Megapodius freycinet, M. pritchardu,
Macrocephalon, and Alectura, the megapodes dissected in this
study (e. g. Leipoa No. 17, Talegalla No. 29, Megapodius
laperouse YPM 89) had a left dorsal carotid artery but none

18 Postilla Yale Peabody Museum No. 78

on the right side; in contrast, chicken embryos possessed both
right and left dorsal carotids. Gleniry (1955) has reported that
all Galliformes except megapodes are bicarotid (see Table 7
for a summary of this feature in other birds).

Early plumages. Studer (1878) and Pycraft (1900) be-
lieved that megapodes molt natal downs before hatching,
but Portmann (1955) and Becker (1959) have contended

Taste 7. Status of dorsal carotid arteries, disastataxy versus eutaxy, and
oil gland feathering in nongallinaceous birds. Symbols: 2, bicarotid; 1, uni-
carotid; E, eutaxy; D, diastataxy; T, tufted oil gland; N, naked oil gland;
O, no oil gland.

Taxonomic Carotid Fifth Oil
group arteries secondary gland

Minami Gd Aes (sieve nis ore caterers 2 E Av
EVA UIUCS ase cea io. catieeesusks oki tesnrake 12 D,E O,?
GaviiGae acres sersicleerna sere os 2 D aly
HOGICIPEGILONMES we arreeternerle ret 1 D T
rocellariiformesin- terriers 2 (1) D T
Spheniscidaew ceil 2 D dt
Pelecaniformes a eeeeneceee ec 3 1,2 D, E ty
Ciconiiformesme eee eee eee 12, D TN
AMATI ae erscecxcmuscsee cecmersncie serene 2 D T
Amatidlalet Aly Ait eacemred ote eee cts 2 D At
Ral conitornmes eee eee 2 D RUN
Gruiformess (aoe oat eres 1,2 D, E T,N,O
Charadriiionmesee renee ree 2 (1) D, E T
Columbifornmes = a-aree see 2 D, E N,O
JERMHEKOUOISONOS Gancccsngceccancs 1,2 D sre)
Musophagidaemrrseeri riser 2 E T
Cuculidae (yee eo toe eee 2 E N
DUA MOMS soogocdbucoseuabes 2 D aN
Caprimulcitormes esse eee eee 5% D N,O
APodiiOrmeseee rece eee 1,2 D,E N
Coltiionmess sane eee ene 1 K N
AMOYEOVUBKOIOTNES Sesonacconsdavae 1 EK N
Coraciftormesmnoe enone 1,2 D, E TN
Riciiormes) erro eC 2 E T,N;O
asserifonmeswce a ceaiceeieeerioe 1 E N

Sources of data: arteries, Glenny, 1955; eutaxy and diastataxy, Steiner,
1956; oil gland, Beddard, 1898, and Miller, 1924.

April 15, 1964 | Ontogeny and Evolution in Megapodes 19

that megapodes lack natal downs and that their first feathers
represent the phylogenetic precursors of natal downs. In con-
trast, Friedmann (1931) stated that megapodes at hatching
bear juvenal feathers in opposition to several authors (e. g.
Ogilvie-Grant, 1893), who referred to the downy young. In
order to determine which, if any, of these conflicting views is
correct, it was necessary to analyze many features of pterylo-
sis, feather growth, and molt.

In the embryonic early growth of the megapode feathers,
those of the tail are longest. For example, on one Leipoa
(No. 5) the caudal sheaths (10 mm long) were 5 mm longer
than the next longest ones on the cervical region and femoral
tract. Similarly, a T'alegalla embryo (No. 24) with tail feath-
ers of 10 mm had the next longest sheaths (38 mm) on the
cervical region. Precocious embryonic early growth of caudal
natal downs occurs in chickens (Hamilton, 1952) and Coturnix
Quail (Padgett and Ivey, 1960) and is apparently a gallina-
ceous trait.

Although a row of 9 or 10 relatively large papillae initially
were formed on the posterior surface of the manus (e. g. on
Nos. 3, 22, 23), of these only primaries 1 through 8 were large
on older embryos and newly hatched Talegalla and Letpoa
(see also Pycraft, 1900, for Megapodius). Such embryonic
repression of the juvenal outer primaries (9 and 10) is charac-
teristic for many Galliformes.

Embryonic megapodes do not molt, contrary to the report
of Studer (1878), who was misled partly by the ease with which
immature sheaths are dislodged from the skin. Indeed, feather
maturation, manifested by hardening, does not occur on the
body in T'alegalla and Leipoa until the last quarter of incuba-
tion as determined by dissection of sheaths from eight tracts.
At hatching, as in other Galliformes, the feathers on the body
are fully grown or nearly so, but the vanes of the remiges
continue growing.

Feather sheaths at hatching are longer on Talegalla and
Leipoa than on chickens. To illustrate this condition, the mean
lengths (M) and coefficients of variation (CV) were calculated
for six sheaths from each of three embryos near hatching. The
six sheaths were taken from corresponding positions on six

No. 78

TH ZA
sp | ) } GZ WE ye

Postilla Yale Peabody Museum

a2

ing

3 ca. 3X).

ison of the tip of secondary No. 9 of the right w

Figure 7. Compar
(top; ca. 4X) with a natal down from the body (bottom

Leipoa ocellata No. 19; 60-73 days of incubation.

April 15, 1964 Ontogeny and Evolution in Megapodes_ 21

tracts on the body of each of the embryos. The values were:
Gallus (19 day) M 13.8 mm (CV 37.6); Letpoa (No. 19) M
28.6 (CV 39.9); and Talegalla (No. 30) M 36.5 mm (CV
38.8). In view of the great variation in lengths of sheaths
within a tract, these values are useful only to indicate the
great difference between megapodes and chickens.

Sheaths on the body of Talegalla and Leipoa embryos
appeared conventional, having opaque and unshriveled tips,
but sheaths of remiges, alula quills, and certain alar upper
coverts of the older T'alegalla and Leipoa embryos had unusual
translucent and shriveled tips as noted by Pycraft (1900) for
remiges of embryonic Megapodius. Pycraft (1900) figured a
constriction of the sheath of the Megapodius remex in the
region of transition from opaque to translucent portions. This
constriction does not occur in Leipoa and Talegalla (this
study); due to lack of a suitable specimen of Megapodius,
it was not possible to check Pycraft’s report of a constriction
in that genus.

Within the translucent tips of the sheaths of remiges on
older Talegalla and Leipoa embryos are weak filaments which
are distal portions of the central barbs of the tip of the remex
(Fig. 7). These distal filaments are easily dislodged in removing
remiges from the sheaths so that some or all filaments are
missing from the expanded remiges of embryos (as in Fig. 7)
and juveniles. Unlike the correspondingly placed natal downs
on the tips of juvenal remiges of phasianids or cracids, these
filaments on the tips of remiges of embryonic megapodes are

weakly developed and lack barbules.

On juveniles of six megapode genera (this study), the feath-
ers at hatching have 1) barbule-free distal ends of central barbs
of body feathers (Fig. 7); 2) a central rhachis; 3) a large
aftershaft on the body feathers (Fig. 7) ; 4) a well-formed vane
in the remiges; these features in common demonstrate that
megapodes had common ancestors possessing such features at
hatching. In contrast, the feathers of chickens at hatching have
1) barbule-free distal ends of central barbs; 2) a distinct
rhachis only in the short and growing juvenal remiges; 3) no
aftershaft ; 4) a well formed vane only in the growing remiges.

22 Postilla Yale Peabody Museum No. 78

Hall (1901), Blasyzk (1935), and Frith (1962) have
reported for juvenile Leipoa and Alectura that the feathers
on the body at hatching are later carried out on the tips of
the growing second feathers. The finding of these connections
(this study) on Letpoa ocellata (Fig. 8), Alectura lathami,
Talegalla jobiensis, and Megapodius freycinet, demonstrates
that this is another general feature of megapodes. As the first
feathers are easily dislodged from the tips of the second ones,
the rarity of observations of these junctions on preserved speci-
mens is to be expected. These connections resemble those be-
tween natal downs and juvenal feathers in other Galliformes.

en
+ St
Ws x

Za

Figure 8. A natal down attached to the tip of a juvenal rectrix from
juvenile Leipoa ocellata. (YPM 1195) ca. 3X.

April 15, 1964 | Ontogeny and Evolution in Megapodes 23

However, since similar connections occur between other genera-
tions of feathers in Galliformes (Watson, 1963), these attach-
ments, considered alone, do not demonstrate conclusively that
the first feathers on the body of megapodes are natal downs.

Nevertheless, the homology of megapode feathers on the
body at hatching with the natal downs of other Galliformes
is shown by the following features in common: 1) the preco-
cious early growth of embryonic tail feathers; 2) the plumula-
ceous structure of the feathers on the body at hatching relative
to the more pennaceous structure of later generations of feath-
ers and of the first remiges; 3) attachment of the first feathers
to the tips of growing feathers of the second generation; 4)
barbule-free distal ends of central barbs; 5) start of the first
body molt within two weeks posthatching (data on Leitpoa
timing from Hall, 1901, and Frith, cited in Nice, 1962).

The following group of characters demonstrates that the
inegapode first remiges are juvenal like those of other Galli-
formes: 1) only eight primaries at hatching but ten on older
juveniles and adults; 2) similar lengths of growing primaries
Nos. 1( first basic = postjuvenal) and 10 (juvenal) on juve-
nile Megapodius (YPM 89) as in certain juvenile phasianids
(cf. Heinroths, 1928) ; 8) remiges more pennaceous than other
feathers at hatching; 4) similar location of the distal filaments
on the embryonic remiges of megapodes and of the correspond-
ing natal downs on other Galliformes ; 5) time of initial loss of a
first remex (two weeks posthatching in Leipoa; Hall, 1901) ;
synchrony of molt of natal downs on the body and juvenal
remiges is characteristic for Galliformes.

The lengths of rhachises in the natal downs of Galliformes
can be partly correlated with the size of the newly hatched
birds. For example, the young of small phasianids, e. g. Cotur-
nix, lack rhachises in their natal downs, while turkeys (Melea-
gris; Pycraft, 1900, and confirmed in this study; and Agrio-
charis; this study) and tragopan pheasants (this study), both
of which are larger at hatching than are the small phasianids,
have short rhachises in their natal downs. Megapodes, still
larger at hatching, have longer rhachises (Fig.7). Certain cra-
cids, e. g. Crax, are exceptional in being large at hatching

24 Postilla Yale Peabody Museum No. 78

(over 100 grams; Heinroth, 1931), while lacking or having
only short rhachises in their natal downs (this study).

As might be expected from the data thus far presented,
many phasianids molt the last of their natal downs at a body
size smaller than that of juvenile megapodes at the time of loss
of the last natal downs. For example, Phasianus colchicus at
160 grams has lost nearly all the natal downs (Westerskov,
1957), while Talegalla (e. g. B of Table 3) at this weight
retains many natal downs on the breast, back and head.

Thus the hatching plumages of megapodes and other Galli-
formes are homologous but differ structurally.

Structures associated with hatching. Several authors (e. g.
Frith, 1959) have reported megapodes at hatching kicking
their way out of the shell, and some observers (e. g. Elvery in
Campbell, 1901) have emphasized the difference from hatching
m chickens. A relatively detailed description of megapodes at
hatching is that of Bergmann (1961), who observed that, in
Talegalla cuvieri, at the time of breaking open of the shell, the
only parts of the body to break through the shell membrane
were the legs and feet. Thus T'alegalla is unlike both chickens
(Hamilton, 1952) and Coturnix Quail (Clark, 1960) which
use the egg tooth of the beak conspicuously in breaking open
the shell.

Although Friedmann (1981) could not find an egg tooth on
one Megapodius pritchardii embryo, and Bergmann (1961)
could not find an egg tooth on T'alegalla cuvieri at hatching, I
(1960, 1961) have found egg teeth on both T'alegalla jobiensis
and Leipoa ocellata embryos (latter observation made indepen-
dently by Frith, 1962). Frith has kindly shown me one speci-
men of prematurely hatched Leipoa bearing an egg tooth,
which, together with my finding that many other specimens of
newly hatched megapodes lack egg teeth, suggests that egg
teeth are usually lost about the time of hatching in megapodes.
The egg teeth of chickens near hatching are approximately
two times larger in linear dimensions than the fully grown egg
tooth of Letpoa (Fig. 9) or Talegalla. Especially when con-
sidered relative to body size at hatching, the megapode egg
tooth is quite small. I (1961) have reviewed the occurrence

April 15, 1964 Ontogeny and Evolution in Megapodes 25

of egg teeth in birds as a whole; egg teeth probably occur on
most, if not all, birds. Megapodes are the only birds for which
egg teeth are thought to be nonfunctional at hatching.

In Talegalla and Leipoa the Musculus complexus or “hatch-
ing muscle” is located dorsally on the neck immediately under
the skin (and under fat deposits in larger embryos), attached
anteriorly to the parietal of the skull, and posteriorly con-
nected to the third, fourth, and fifth cervical vertebrae and the
muscular complex overlying these vertebrae. The two complexus
muscles were separated in the dorsal midline in the 20 examined
anatomical specimens of megapodes: in Leipoa by minimal

Figure 9. Egg tooth of an embryonic Léipoa ocellata. (No. 9) Overlying

oa

periderm removed. Ca. 7X.

distances of 1.5 (No. 4) to 38 mm (No. 19) and in Talegalla
by 2.5 (No. 26) to 5 mm (No. 30). In contrast, in chicken
embryos near hatching, the two complexus muscles met in the
dorsal midline (Fig. 10). The anterior insertions meet in the
dorsal midline long before hatching and after hatching move
laterally, separating in the dorsal midline (Fisher, 1958; this
study). The M. complexus of megapodes and chickens also
differed in the apparent lack of a temporary enlargement about
the time of hatching in megapodes. In chickens near hatching
this muscle appears swollen, protruding above the level of
adjacent cervical muscles and reaching a thickness of at least
2.5 mm, whereas in megapodes no swelling was observed and
maximal thickness was always less than 1 mm. Similarly,
although maximal width of the complexus muscle in each of
four chickens near hatching was 7 mm, in none of the mega-
podes did this width exceed 5-7 mm, which was reached only
in the largest specimens (e. g. Nos. 19, 30).

26 Postilla Yale Peabody Museum No. 78

Length measurements of the M. complexus were unreliable
due to the lack of a clear posterior boundary of the muscle.
When measurements of width and midline separation were ana-
lyzed relative to body size by plotting on arithmetic and double
logarithmic graphs, no indications of prehatching variations
other than growth and individual variations were detected for
the megapodes, but the precision of these measurements (about
+ 0.5 mm) is not very great relative to the dimensions meas-
ured. These observations do not eliminate the possibility of a

Figure 10. The Musculus complexus of domestic G@. gallus (19 day
embryo; ca. 1.2) and of Talegalla jobiensis (No. 30; ca. 14%).
Talegalla on the right.

transient enlargement of the M. complexus at hatching in meg-
apodes, but they provide no support for such a view. The
separation in the dorsal midline and apparent lack of special
enlargement of the complexus muscle at hatching in megapodes
are very likely correlated with the larger size of megapodes
at hatching.

The small egg tooth and unusual features of development
of the M. complexus of megapodes appear to be associated with
the different methods of hatching in megapodes and phasianids.

DISCUSSION AND CONCLUSIONS

Gallinaceous growth and maturation. The embryonic mega-
podes Letpoa after the first 20 days were relatively immature
compared with chickens of similar age. Although slow early
embryonic development is a reptile-like character, not too
much phylogenetic significance can be attributed to this con-

April 15, 1964 | Ontogeny and Evolution in Megapodes 27

dition in Leipoa, since the slow developmental rate is asso-
ciated with the methods of incubation including relatively
low incubating temperatures. It is possibly phylogenetically
significant that Leipoa can hatch successfully (Frith, 1959)
at incubating temperatures so low (below 95°F) as to be lethal
for chicken embryos (Romanoff, 1960) ; however, data on the
normal range of egg temperatures of wild birds in general
(Huggins, 1941) ridiiate that megapodes are perhaps not
unusual among birds with respect to tolerated incubating tem-
peratures.

Interpretation of the chronology of embryonic megapodes
is complicated by great individual variation. For example, nor-
mal prehatching periods in Leipoa from different mounds range
from 50 to 90 days in association with intermound variations
from 96° down to 80°F in incubating temperatures (Frith,
1959). Since incubating temperatures of the megapode T'ale-
galla jobiensis (Ripley, 1964) are within the range for
Leipoa (Frith, 1959), it is possible, though unproven, that
Talegalla has an embryonic chronology similar to that of
Leipoa. Analysis of differences in embryonic chronology be-
tween megapodes and phasianids is further complicated by the
great interspecific variation among phasianids incubated at
100°F. For example, Colinus weighing 6 grams (egg weight,
9 g) and Phasianus weighing 18 grams (egg weight, 32 g)
are both hatched in 24 days, while chickens of 31 grams (egg
weight, 60 g) are hatched in only 21 days (Romanoff, 1960:
1143). Data are not available for a quantitative comparison
of the effects of varied incubation temperatures on the devel-
opment of chickens versus megapodes.

Both the phasianid Phasianus colchicus (Westerskoy, 1957)
with an adult (male) weight of 1400 grams and the megapode
Alectura lathami (Coles, 1937) with a slightly higher adult
weight (Heinroth, 1922) reach full size about 25-30 weeks
after laying of the egg, indicating that the posthatching growth
of Alectwra is neither unusually fast nor slow compared with
that of phasianids.

The data of this study show that Leipoa and Talegalla
before hatching undergo proportional and relative growth
analogous to that occurring up to several weeks posthatch-

28 Postilla Yale Peabody Museum No. 78

ing in other Galliformes. The similarity of relative growth in
young Galliformes is in agreement with the morphological
homogeneity of adults (cf. data of Verheyen, 1956). The rel-
ative growth appears, in this case, to be phylogenetically
generally more conservative than chronological growth. The
differences in relative growth of radius and humerus between
megapodes and phasianids do not indicate that either group
is more primitive than the other.

The noted interspecific variations in the size of embryos at
the first macroscopic appearance of certain structures may
represent interspecific differences in the growth of anlage of
these structures, for, as Schmalhausen (1926) and others have
pointed out, relative growth itself can produce qualitative
changes in form.

Although the weight of a bird at hatching is relatively
directly correlated with the weight of the egg (Heinroth,
1922), the ratio of the size of the egg relative to that of
adults often shows considerable intergeneric variation (Hein-
roth, 1922). Megapodes and certain small phasianids (e. g.
Coturnix) have eggs generally in the range from 8 to 18 per
cent of adult body weight in contrast to other phasianids and
turkeys with eggs weighing less than 5 per cent of adult body
weight (Heinroth, 1922).

The precocity of megapodes at hatching is associated with 1)
the large absolute egg size and correspondingly large size of
young at hatching together with 2) an embryonic relative
growth of the wing analogous to that occurring up to several
weeks posthatching in phasianids. No birds other than mega-
podes have large eggs plus extensive embryonic growth of the
wings.

Megapodes and reptiles. Portmann (1938) listed the fol-
lowing as primitive (reptile-like) traits of megapodes: lack of
natal downs, possible lack of an egg tooth at hatching, absence
of parental care for young, eggs incubated in sand by solar
heat, long incubation period, large clutch size, slow growth to
adult size, and precocity of young at hatching. However, as
shown by my study, megapodes do have natal downs, and at
least some species have egg teeth. Furthermore, there is no

April 15, 1964 Ontogeny and Evolution in Megapodes 29

good evidence for an especially slow posthatching growth of
megapodes.

Moreover, the many adaptive interrelationships (coadapta-
tions) of the reptile-like characters of megapodes should be
considered. For example, the long incubation period is cor-
related with the methods of incubation and the large size and
precocity of young at hatching. The precocity of young is also
correlated with the lack of parental care which in turn is asso-
ciated with the incubating methods and clutch size. The reptile-
like traits of megapodes all belong to one, or perhaps two,
group(s) of coadapted characters. Considered in this way, the
evidence for special affinities of megapodes and reptiles is uncon-
vincing, since the points of similarity are all related to com-
mon reproductive adaptations.

The case for special reptilian affinities of megapodes would
be greatly strengthened if there were reptile-like characters
relatively independent of the central adaptation in megapodes :
however, no such characters have yet been found. As one exam-
ple, there is reported to be a significant difference in the caloric
values of reptilian and avian egg yolks (Slobodkin, 1962), yet
samples of yolk collected during this study from relatively
fresh eggs of Leipoa and Gallus had values agreeing with
those of other avian species (Slobodkin, 1962).

Furthermore, advocates of the primitiveness of the mega-
podes among birds as a whole have generally failed to analyze
the possibility of convergent evolution. In short, evidence for
the primitiveness of megapodes among birds as a whole is
unacceptable.

Evolution of the megapode family. Megapodes are basi-
cally similar in morphological development to phasianids. Dif-
ferences in the structure of natal downs, in absolute and rela-
tive sizes of eggs, in sizes of subcutaneous fat bodies, in develop-
ment of the hatching apparatus, ete., are all directly or
indirectly correlated with the sizes of the young at hatching.

Huxley (1868) emphasized that, in contrast to other Galli-
formes, megapodes and cracids are alike in depth of the sternal
notches and in position of the hallux. From this anatomical
basis, he postulated that these forms, isolated respectively in

30 Postilla Yale Peabody Museum No. 78

the Australian and Neotropical regions, are remnants of an
ancestral gallinaceous stock which has been replaced through
most of the Old World and Nearctic region by more modern
Galliformes.

However, the differences at hatching in feather structure be-
tween cracids and megapodes support the generalization that
megapodes and cracids are not especially closely related in
evolution, contrary to some current classifications (e. g. Peters,
1934).

The contemporary megapodes are characterized by 1) rha-
chidial natal downs on the body, 2) long juvenal remiges and
large body size at hatching, 3) a relatively high ratio of egg
to adult weights compared with other Galliformes, and 4) the
unicarotid condition; it is likely that these distinctive traits
were present in a population ancestral to all living megapodes.
Megapodes are apparently unique among birds in having such
long and weak natal downs preceding the embryonic juvenal
remiges. These weak natal downs are clearly vestiges rather
than preadaptations and indicate the evolution of megapodes
from unknown gallinaceous ancestors possessing a natal plum-
age and less precocious chicks resembling those of extant phasi-
anids.

This phylogenetic interpretation is also supported by the
finding of a vestigial egg tooth and the apparent lack of special
enlargement of the complexus muscle at hatching; these fea-
tures strongly indicate an evolutionary origin of megapodes
from forms less precocious at hatching. One aspect of the evolu-
tion of megapodes has been the transition from the use of the
egg tooth in hatching to kicking open the shell.

The variation in the number of carotid arteries in birds as
a whole (Table 7) appears to be due to much convergent evolu-
tion. The most readily conceived sequence is a loss of one
carotid artery (Glenny, 1955), but a possible evolutionary
increase cannot be excluded. The occurrence of only one carotid
in megapodes in contrast to two in all other known Galliformes
suggests that megapodes are specialized in this respect.

My conclusions, based on morphology, are compatible with
the concept of Mainardi and Taibel (1962: Fig. 4), based

largely on erythrocyte antigens, that megapodes, cracids, and

April 15, 1964 | Ontogeny and Evolution in Megapodes_ 31

phasianids have evolved as three separate lines from unknown
gallinaceous ancestors.

It is pertinent that there are living forms intermediate in
structure of feathers at hatching and in precocity of young
between megapodes and phasianids such as Phasianus or Gallus.
For example, the phasianid genus T'ragopan has natal downs
with short rhachises (this study), relatively long juvenal remi-
ges at hatching (Beebe, 1918), and initial flight on the third
day posthatching (Nice, 1962; after the Heinroths). Although
Tragopan probably does not represent the phylogenetic ances-
tors of megapodes, certain aspects of its structure and behavior
of young aid in visualizing the evolutionary origin of the mega-

podes.

Evolution within the megapodes. Megapodius and Macro-
cephalon lay their eggs in holes (Megapodius also uses mounds )
and are known to lay their eggs communally, while the four
other genera use mounds exclusively as far as known. (In
accord with the study of Ripley (1964) the form Eulipoa
wallacei is here included in the genus Megapodius. )

The specialized Macrocephalon is somewhat intermediate in
adult proportions of wing, tarsus, and tail between other large
megapodes (4 genera) and the smaller Megapodius (data in
Ogilvie-Grant, 1893). The relatively uniform color of Megapo-
dius and its relative simplicity of nesting habits have led some
authors (e. g. Becker, 1959) to consider Megapodius primitive
among the megapodes. The uniform color pattern of Mega-
podius resembles that of Aepypodius or Tialegalla and may
indeed be a primitive trait among living megapodes. But sim-
plicity of nesting site (e. g. the incubation of eggs in holes in
the ground) does not necessarily imply primitiveness as illus-
trated by the specialized brood-parasitic avian species which
also build no nests.

Since one trait of the megapodes is the relatively high ratio
of egg weight to adult weight, and since megapodes have evolved
from apparently more conventional gallinaceous ancestors, it is
likely that, during megapode evolution, sizes of eggs increased
relative to adult size. Although megapode evolution has very
likely also involved an increase in the absolute size of eggs and

32 Postilla Yale Peabody Museum No. 78

chicks at hatching, the absolute sizes of newly hatched young
do not necessarily indicate the relative primitiveness of the
contemporary megapodes. Indeed, if, as seems likely, the evolu-
tion of megapodes has involved an increase in the absolute size
of eggs and hence of young at hatching, then a large ancestral
adult would have been better preadapted, in terms of size, than
a small ancestral adult for the evolution of larger absolute sizes
of eggs.

More critical features suggesting the direction of evolution
within the megapodes are the proportions at hatching. In this
respect Megapodius is more remote than T'alegalla or Leipoa
from the conditions in non-megapode Galliformes. In view of
the relatively shorter bill and longer wing at hatching and the
unusual webbing of the toes in Megapodius, the simplest hypo-
thesis is that Megapodius has secondarily evolved from a form
hke T'alegalla or Aepypodius. Thus Megapodius, perhaps most
reptile-like of the megapodes in certain respects, is structurally
specialized.

The small size (and relatively short culmen) of adult Mega-
podius appear to be adaptive in reducing potential ecological
competition where Megapodius and other megapode genera
occur sympatrically (Ripley, 1960). From the present study
it is apparent that a shorter culmen and smaller body size at
hatching also characterize Megapodius when compared with
other megapodes.

Megapodius and Macrocephalon have possibly primitive char-
acters in the occurrence of diastataxy (variable in Megapo-
dius) and the tufted oil gland. Distribution of these characters
in birds as a whole (Table 7) indicates that there is no neces-
sary correlation in the presence of these features and that they
have been subject to considerable convergent evolution. Despite
the contention of Steiner (1918, 1956) that diastataxy is prim-
itive because it occurs in “primitive” birds, there is no con-
vineing evidence against the possibility that diastataxy might
evolve from eutaxy (see Humphrey and Clark, 1961, for a
review of the various hypotheses on the origin of diastataxy).
Similarly, there is no reason to assume that a tufted oil gland
is necessarily primitive.

April 15, 1964 Ontogeny and Evolution in Megapodes 33

In view of the intraspecific constancy of tarsal scutellation
and its intergeneric variation in the megapodes, it appears use-
ful in dividing the megapodes into subgroups; however, in view
of the range of variation within the megapode family, it would
probably be unwise to emphasize this feature in attempting to
determine the affinity of megapodes with other gallinaceous
families.

\lll

Megapodius Leipoa

Alectura

-—_
Talegalla =

Aepypodius =

Macrocephalon

Stem megapode
population

Pheasant-like
gallinaceous
ancestors
Figure 11. Provisional phylogeny of the family Megapodiidae. ‘The
smaller branches leading from the genera represent speciation.

From these considerations, the first phylogeny to cover inter-
generic relationships within the megapodes has been developed
(Fig. 11). The ancestral stem population (Fig. 11) would
have possessed large adult and chick sizes, like Talegalla,
rhachidial natal downs, a relatively long culmen at hatching,
and egg laying in mounds. If this phylogeny is correct, then
current classifications (e. g. Peters, 1934) are misleading in
placing Megapodius first in the sequence of megapode genera.

In examining megapode development, I have found no charac-
ters indicating that megapodes are especially primitive birds;
indeed, the evidence demonstrates the specialized nature of
megapode ontogeny which has probably evolved from a phasi-
anid-like condition.

34 Postilla Yale Peabody Museum No. 78

ACKNOWLEDGMENTS

Valuable suggestions and constructive criticism were given
by Professors S. D. Ripley, G. E. Hutchinson, J. L. Brooks,
and J. P. Trinkaus. Dr. E. J. Boell gave much excellent coun-
sel. Dr. P. S. Humphrey initially suggested the topic and
provided many helpful suggestions. Dr. H. J. Frith, of the
Wildlife Survey Section, Australian Commonwealth Scientific
and Industrial Research Organization, made arrangements
which greatly facilitated my field collecting. I am also much
indebted to Dr. D. Amadon, the American Museum of Nat-
ural History, Mr. B. K. Brown, Dr. E. T. Gilliard, Mr. D.
Heath, Mr. E. A. Heath, Dr. H. Levene, and many others
who have helped in a variety of ways. Mrs. Shirley Hartman,
Mr. G. di Palma, and my wife prepared the figures. I am
especially grateful to my wife for her encouragement and aid.

Financial support was provided by the National Science

Foundation Grant G-10735, awarded to Dr. S. D. Ripley.

SUMMARY

Many differences found in development between megapodes
and phasianids are associated with megapodes having before
hatching proportional and relative growth equivalent to that
occurring up to several weeks posthatching in phasianids.

Contrary to published reports, megapodes at hatching bear
juvenal remiges and natal downs on the body and are thus like
other Galliformes, although there are structural differences in
the natal downs. Vestigial natal downs preceding the embry-
onic juvenal remiges indicate that megapodes evolved from
forms with more conventional gallinaceous feathering at hatch-
ing and less precocious young.

This interpretation of megapodes as evolutionarily special-
ized is also upheld by their vestigial egg teeth and apparent
lack of a special enlargement of the complexus muscle which *
aids in the hatching of other Galliformes.

Compared with other megapode genera and other Gallifor-
mes, young juvenile Megapodius have a long wing and unus-
ually short bill. It is therefore concluded, contrary to published

April 15, 1964 | Ontogeny and Evolution in Megapodes 35

reports, that, despite its apparent simplicity in color pattern
and egg laying habits, Megapodius is specialized among mega-
podes.

A phylogeny of the megapode genera is proposed on the
basis of proportions at hatching, tarsal scutellation, foot web-
bing, eutaxy, oil gland feathering, and other characters.

REFERENCES CITED

Amadon, D., 1943. Bird weights as an aid in taxonomy. Wilson Bull. 55:
164-177.

Becker, R., 1959. Die Strukturanalyse der Gefiederfolgen von Megapodius
freyc. reinw. und ihre Beziehung zu der Nestlingdune der Hiihnervégel.
Rev. suisse Zool. 66: 411-527.

Beddard, F. E., 1898. The structure and classification of birds. Longmans,
Green, London. p. xx + 1-548.

Beebe, W., 1918. A monograph of the pheasants. Vol. 1. Witherby, London.
Dy 2lbee ae ileiesy

Bergmann, S., 1961. My father is a cannibal. Hale, London. 192 p.

Blaszyk, P., 1935. Untersuchungen iiber die Stammesgeschichte der
Vogelschuppen und Federn .... Morph. Jahrb. 75: 483-521.

Campbell, A. J., 1901. Nests and eggs of Australian birds. Pt. 2: 525-1102.
Sheffield, England.

Clark, G. A., Jr., 1960. Notes on the embryology and evolution of the
megapodes (Aves: Galliformes). Yale Peabody Mus. Postilla no. 45:
1-7.

Clark, G. A., Jr., 1961. Occurrence and timing of egg teeth in birds. Wilson
Bull. 73: 268-278.

Coles, C., 1937. Some observations on the habits of the Brush Turkey
(dlectura lathami). Proc. Zool. Soc. Lond, 107A: 261-273.

Fant, R. J., 1957. Criteria for aging pheasant embryos. J. Wildl. Mgt. 21:
324-328.

Fisher, H. I., 1958. The “hatching muscle” in the chick. Auk 75: 391-399.

Friedmann, H., 1931. Observations on the growth rate of the foot in the
mound birds of the genus Megapodius. Proc. U. S. Nat. Mus. 80 (Art.
1): 1-4.

Frith, H, J., 1959. Breeding of the Mallee Fowl, Leipoa ocellata Gould
(Megapodiidae). CSIRO Wildl. Res. 4: 31-60.

Frith, H. J., 1962. The Mallee-fowl. Angus & Robertson, Sydney. p. xii +
1-136.

Glenny, F. H., 1955. Modifications of pattern in the aortic arch system of
birds and their phylogenetic significance. Proc. U. S. Nat. Mus. 104:
525-621.

Hall, R., 1901. Descriptions of the young of some Australian birds. Vict.
Nat. 18: 19-25.

Hamilton, H. L., 1952. Lillie’s development of the chick. 3rd ed. Holt,
New York. p. xv + 1-624.

Heinroth, O., 1922. Die Beziehungen zwischen Vogelgewicht, Eigewicht,
Gelegegewicht und Brutdauer. J. f. Orn. 70: 172-285.

36 Postilla Yale Peabody Museum No. 78

Heinroth, O., 1931. Beobachtungen bei der Aufzucht eines Knopfschnable-
Hokko’s (Crax globicera) und eines Mitu’s (Mitua mitu). J. f. Orn.
79: 278-283.

Heinroth, O. and M. Heinroth, 1928. Die Vogel Mitteleuropas. Bd. 3.
Bermihler, Berlin. p. x + 1-286.

Huggins, R. A., 1941. Egg temperatures of wild birds under natural condi-
tions. Ecology 22: 148-157.

Humphrey, P. S., and G. A. Clark, Jr., 1961. Pterylosis of the Mallard
duck. Condor 68: 365-385.

Huxley, T, H., 1868. On the classification and distribution of the Alectoro-
morphae and Heteromorphae. Proc. Zool. Soc. Lond., 294-319.

Lyon, D. L., 1962. Comparative growth and plumage development in
Coturnix and Bobwhite. Wilson Bull. 74: 5-27.

Mainardi, D., and A. M. Taibel, 1962. Studio immunogenetico sulle
parentele filogenetiche nell’ordine dei Galliformi. Inst. Lombardo, Rend.
Se., B 96: 131-140.

Meyer, O., and E. Stresemann, 1928. Zur Kenntnis der Entwicklung von
Megapodius und Oxyura im Ki. Orn. Monatsber. 36: 65-71.

Miller, W. DeW., 1924. Further notes on ptilosis. Bull. Am. Mus. Nat.
Hist. 50: 305-331.

Nice, M. M., 1962. Development of behavior in precocial birds. Trans. Linn.
Soc; N. Y. 8: xii —-— 1-211.

Ogilvie-Grant, W. R., 1893. Catalogue of the game birds .. . of the
British Museum. p. 445-472.

Padgett, C. S., and W. D. Ivey, 1960. The normal embryology of the
Coturnix Quail. Anat. Record 137: 1-11.

Peters, J. L., 1934. Check-list of the birds of world. Vol. 2. Harvard Univ.
Press, Cambridge, Mass. p. xvii + 1-401.

Portmann, A., 1938, Beitriige zur Kenntnis der postembryonalen Entwick-
lung der Vogel. Rev. suisse Zool. 45: 273-348.

Portmann, A., 1950. Le développement postembryonaire. p. 521-535; in:
Grassé, P.-P. (ed.) Traité de Zoologie, Tome XV, Oiseaux.

Portmann, A., 1955. Die postembryonale Entwicklung der Vogel als
Evolutionsproblem. Acta XI Congr. Internat. Orn., Basel 1954: 138-151.

Pycraft, W. P., 1900. A contribution towards our knowledge of the
pterylography of the megapodii. p. 483-492; in: A. Willey (ed.)

Zoological results ... New Britain... Part IV. Cambridge at Univ.
Press.

Pycraft, W. P., 1910. A history of birds. Methuen, London. jab zeal Se
1-458.

Ripley, S. D., 1960. Distribution and niche differentiation in species of
megapodes in the Moluccas and Western Papuan area. Proc. 12th
Internatl. Orn. Congr., Helsinki 1958: 631-640.

Ripley, S. D., 1964. A systematic and ecological study of birds of New
Guinea. Yale Peabody Mus. Bull. no. 19: 87 p.

Romanoff, A. L., 1966. The avian embryo. Macmillan, New York. p. xvi
+ 1-1305.

Schmalhausen, I., 1926. Studien iiber Wachstum und Differenzierung. III.

Die embryonale Wachstumskurve des Hiihnchens. Roux’ Archiv 108:
322-387.

April 15, 1964 | Ontogeny and Evolution in Megapodes 37

Simpson, G. G., A. Roe, and R. C. Lewontin, 1960. Quantitative zoology.
Rev. ed. Harcourt, Brace, New York. p. vii + 1-440.

Slobodkin, L. B., 1962. Energy in animal ecology. p. 69-101; in: J. B. Cragg
(ed.) Advances in ecological research. Vol. 1.

Steiner, H., 1918. Das Problem der Diastataxie des Vogelfliigels. Jena.
Z., N. F. 48: 221-496.

Steiner, H., 1956. Die taxonomische und phylogenetische Bedeutung der
Diastataxie des Vogelfliigels. J. f. Orn. 97: 1-20.

Studer, T., 1878, Beitrige zur Entwicklungsgeschichte der Feder. Z. f.
wiss. Zool. 31: 421-436.

Verheyen, R., 1956. Contribution a l’anatomie et a la systematique des Gal-
liformes. Inst. Roy. Sci. Nat. Belgique 32 (42): 1-24.

Watson, G. E., 1963. Feather replacement in birds. Science 189: 50-51.

Westerskoy, K., 1957. Growth and moult of pheasant chicks. New Zealand
Dept. Int. Aff. Wildl. Publ. 47: 1-64.

| Harvard MCZ Libra’
IUSLO NN
3 2044 066 305 236

Date Due

MAR ——1976

AUS HB

aes gael
RAPT

i a ain rnrna tees tart
eno: it OR ON areal =>
at ee ae ee eae pate eee

pa PE GOEL

fa art eae