A preliminary method for estimating the age of brown kiwi (Apteryx mantelli) embryos

Abstract

Morphological features of a collection of unknown-age wild kiwi (Apteryx mantelli) embryos from early development to point of hatch are described. Using these features, we assign developmental stages to each embryo and compare the progress of development to similar-staged ostrich (Struthio camelus) and chicken (Gallus gallus) embryos. Two ageing schemes for the kiwi embryos are developed by comparing measurements of their hindlimb segments, bills and crown–rump lengths with those of ostrich and chicken embryos at various stages of development. One of the 20 kiwi embryos was of known age. Both the ostrich model and the chicken model gave identical predictions for the marker and four other embryos. Developmental timing of some features differed between all three species, most markedly in the bill, with growth in the kiwi bill being relatively faster to achieve its larger relative and absolute size at hatch.

Keywords:

• embryonic development
• ratite
• precocial bird
• hindlimb
• age estimation

Introduction

Studies of avian embryo growth and development are not common, but exist for poultry (Buckner et al. 1950; Hamburger & Hamilton 1951; Daniel 1957; Fant 1957; Mun & Kosin 1960; Ancel et al. 1995), some waterfowl (Weller 1957; Cooper & Batt 1972; Caldwell & Snart 1974; Montgomery et al. 1978) and some seabirds (Ryder & Somppi 1977; Mahoney & Threlfall 1981). The series by Hamburger & Hamilton (1951) of normal stages of development in the chicken act as a basis for most modern embryological comparisons. They divide the developmental process into 46 stages based on the presence or absence of morphological characteristics independent of incubation length and embryo size. All avian embryos pass through the same series of developmental events (Ricklefs & Starck 1998), but the exact chronology of these events can vary between species even in the earliest stages of development, suggesting that they are not as rigidly conserved as once thought (Richardson 1999; Lilja et al. 2001; Blom & Lilja 2005). Previous studies of morphological development of the embryo of the kiwi (Apteryx spp.) are limited to the observations of Parker (1891, 1892).

Kiwi eggs are unusually large for a bird its size, weighing up to 25% of the female kiwi's body weight, or approximately 400 g (Reid & Williams 1975; Calder 1979). The incubation period is also exceptionally long, typically ranging from 74 to 84 days in the wild (Reid & Williams 1975) but it may extend to 91 days (Colbourne 2002). Artificially incubated eggs typically hatch at 78 days (Bassett 2012).

The kiwi's large egg size and long incubation period likely reflect benefits in retaining an ancestrally large egg in an environment devoid of mammalian predators (Calder 1979; Potter & McLennan 1992). The relatively recent impact of mustelid predation on kiwi chicks has caused a rapid decline in population size, making all kiwi species threatened and requiring active management (Heather & Robertson 2005; Holzapfel et al. 2008). Limited access to kiwi and their eggs and their status as taonga species (i.e. of cultural significance to the native Māori people [Holzapfel et al. 2008]) constrains the availability of specimens and precludes sacrificing healthy embryos for study of normal development. To learn more about the kiwi's embryological development with few available samples, a developmental analogue must be sought for initial comparison.

This study examined the development and growth of the morphological features of brown kiwi (Apteryx mantelli) embryos by comparison with two other precocial species: the chicken (Gallus gallus), which has a comparable adult body size to kiwi; and the ostrich (Struthio camelus), which is closely related phylogenetically to kiwi (Livezey & Zusi 2007). By using detailed descriptions of a collection of kiwi embryos of unknown age, we related developmental stage-defining features of all three species to establish an ageing scheme based on morphological measurements. Based on our analyses, an embryo's developmental stage can be used to estimate the date the egg was laid, the projected date of hatch, or when an embryo died, and to identify critical or high-risk stages of incubation.

Materials and methods

Twenty preserved embryos of wild brown kiwi (Apteryx mantelli) that had died during incubation were examined. Only one of the embryos (designated embryo I*) was of known age (38 days from date of laying with incubation starting immediately).

Developmental stages for each kiwi embryo were determined using Hamburger & Hamilton's (1951) embryo staging methods for domestic chickens. Hamburger and Hamilton stages were also assigned to the ostrich embryo descriptions given in Gefen & Ar (2001). Determination of the stages of development was based on external features only, such as presence of limb buds, limb segments, bill, feathers, nails and scales. The kiwi embryo was presumed to have completed the same portion of incubation as a chicken and as an ostrich at the same stage (which have incubation lengths of 21 days and 42 days, respectively). The age estimations given by these methods are referred to as the chicken and the ostrich models. Various incubation lengths have been reported for brown kiwi: 74–84 days (Reid & Williams 1975), 77–106 days (McLennan 1988), 75 days and 91 days (Colbourne 2002), and 65–75 days (Heather & Robertson 2005). Therefore, the models used both a 78-day and an 85-day incubation length to report an age range for each kiwi embryo. This range of incubation durations was chosen to avoid biasing age estimates by outliers.

Morphological measurements were taken with Vernier calipers (± 0.2 mm) or string and ruler (crown–rump length only). Limb segments were measured on the right hindlimb while it was flexed as much as possible. This was only possible for embryos F–T because of the lack of distinguishable limb segments in earlier embryos. Limb segment lengths included digit III, the longest (i.e. middle) toe (from the proximal phalangeal joint to the distal point of the nail), the tarsometatarsus (from the tarsometatarsal to the proximal phalangeal joint), the tibiotarsus (from the stifle to the tarsometatarsal joint), and the femur (from the greater trochanter of the femur to the stifle joint). Bill length was measured from the distal tip of the bill to the distal end of the cere at the midpoint of the cere's curve. Crown–rump length was measured in all embryos from the cere of the bill to the pygostyle.

Limb segment lengths for embryos F–L were linearly regressed against embryo age as predicted by both the chicken and the ostrich models, assuming both a 78-day (minimum) and an 85-day (maximum) incubation period. For bill-length measurements, a second-degree polynomial regression was fitted. The eight late-stage embryos (M–T) had reached a point in their incubation where they no longer had diagnostic features that could be used to allocate a stage beyond 40. In chickens, Hamburger & Hamilton (1951) determined stages 40–45 by measuring toe and bill lengths. Here, we inserted the toe length of embryos M–T into the toe length regression equation produced using embryos F–L to create age estimations.

To validate our models, a mean bill measurement of 42.8 mm for a point-of-hatch wild kiwi chick (age 0–3 days, lengths 44.5 mm; 42.0 mm; 43.0 mm; 41.6 mm, John McLennan, unpublished data) was used in the bill-length regression line equations to calculate an estimation of incubation length.

Each embryo was photographed with a Fujifilm FinePix S200EXR digital camera using a tripod and light table.

Results

Embryo staging was most effectively determined using limb features, especially the toes. The presence of feathers or degree of joint bending were not useful criteria in this collection as many embryos were preserved in unnatural positions or had suffered varying degrees of damage (Fig. 1).

Figure 1 Kiwi (Apteryx mantelli) embryos are arranged in increasing chronological order. Scale bars appear to the left of each row (20 mm). An asterisk (*) denotes the known-age embryo. See Table 1 for details about each embryo.

Table 1 Estimation of Hamburger and Hamilton (HH) stage for kiwi embryos.

		Predicted age range (days)
Embryo ID	HH stage	Ostrich model	Chicken model	Feature descriptions
A	early 24	14.9–16.2	14.9–16.2	leg bud is approximately twice as long as it is wide, cannot completely distinguish toe plate
B	early 25	18.6–20.2	16.7–18.2	wing straight, legs longer than wings, toe plate present but no demarcation of toes, little embryonic flexure
C	late 28	22.3–24.3	22.3–24.3	wing bent at elbow, middle toe longest, grooves between toes, neck lengthened, bill clearly recognizable and mandible approximately half the length of bill
D	30	26.0–28.3	24.1–26.3	grooves visible between digits, second digit slightly longer, webbing between toes evident despite damage to toes, femur obvious, all limb segments are demarcated, mandible approaching end of maxilla, external auditory openings visible
E	late 31	29.7–32.4	27.9–30.4	wing bent at elbow, toe web contours slightly concave, bill damaged, cannot distinguish feather germs
F	35	33.4–36.4	29.7–32.4	small amount of webbing at base of the toes, beginning to distinguish phalanges, mandible has nearly reached its full length in relation to maxilla, feather germs on dorsal side
G	36	37.1–40.5	37.1–40.5	nails resemble cones, plantar pads just visible in profile, feather germs over the embryo's dorsal surface, thigh and portions of the head
H	late 36	37.1–40.5	39.0–42.5	nearly identical to embryo G, but claws and plantar pads slightly more conspicuous
I*	late 37	40.9–44.5	40.9–44.5	claws and plantar pads are obvious, scales visible on all but toes, feather germs visible on entire surface of head and wings and extend down to the middle of the tibiotarsus, dorsal surface feathers from the back of the neck to the tail are pigmented, feather tracks are apparent, small circular opening between the eyelids
J	early 38	44.6 –48.6	44.6–48.6	feathers germs extend over the majority of the tibiotarsus, scales visible on toes
K	late 38	48.3–52.6	46.4–50.6	nail present on the digit, pigmented feathers cover entire body except top of the head where a portion of skin is missing, nails are semi-transparent, with slight pigmentation at their base, narrow crescent opening between eyelids
L	39	52.0–56.7	48.3–52.6	scales overlap on all but back of the legs, feathers over head and legs, slight pigmentation on back of leg and top of toes, eyelids cover entire eye
M	40 +	60.3–65.7	57.9–63.1	pigmented scales on toes, toenails opaque
				middle toe = 14.8 mm; bill = 29.4 mm
N	40 +	63.0–68.6	60.4–65.8	middle toe = 16.0 mm; bill = 31.8 mm
O	40 +	65.3–71.1	62.5–68.1	middle toe = 17.0 mm; bill = 32.8 mm
P	40 +	67.5–73.6	64.6–70.4	middle toe = 18.0 mm; bill = 32.0 mm
Q	40 +	82.6–90.0	78.6–85.6	claws curve ventrally
				middle toe = 24.6 mm; bill = 41.4 mm
R	40 +	83.9–91.4	79.8–87.0	middle toe = 25.2 mm; bill = 39.0 mm
S	40 +	86.6–94.4	82.4–89.8	middle toe = 26.4 mm; bill = 39.2 Mm
T	40 +	86.6–94.4	82.4–89.8	middle toe = 26.4 mm; bill = 39.8 mm

Embryo age estimates based on ostrich and chicken models and feature descriptions are presented for each kiwi embryo. An (*) indicates the known age embryo (I = 38 days).

Differences in timing of head and bill development between the kiwi and the chickens became apparent around Hamburger & Hamilton's (1951) stage 28 (Table 1, Fig. 1). Embryo C showed a more advanced mandible than the corresponding stage in the chicken. The appearance of feather germs and papillae in the kiwi lagged behind the chicken and did not become apparent until approximately stage 35 in embryo F. As our series is missing stages 32–34, it is possible that feather development occurs earlier. Eyelid development was not apparent to the extent of that in the ostrich and chicken in embryos E and F.

The best measurements for ageing kiwi embryos as determined by the ostrich model were crown–rump length and tarsometatarsus length (Fig. 2, 3). In contrast, bill length and tibiotarsus length gave the best predictions in the chicken model (Fig. 2, 3). The least useful measurement was femoral length (Fig. 2), possibly because of the difficulty in measuring this parameter in these specimens. In embryos A–J, both the ostrich and chicken models gave similar age predictions. An estimation of 40.9–44.5 days for the 38-day-old, known-age embryo (I*) was produced by both the ostrich and chicken models. In embryos K–Q, the ostrich model gave somewhat higher age estimations than the chicken model.

Figure 2 Regressions of selected measurements with predicted age based on ostrich and chicken models. Predictions for both the minimum (78 days) and maximum (85 days) kiwi incubation period are presented. Equations and R² values above the line correspond to the ostrich regression (×, —) and those below correspond to the chicken regression (o, —). A, B, Crown–rump length. C, D, Bill length. E, F, Femur length.

Figure 3 Regressions of selected measurements with predicted age based on ostrich and chicken models. Predictions for both the minimum (78 days) and maximum (85 days) kiwi incubation period are presented. Equations and R² values above the line correspond to the ostrich regression (×, —) and those below correspond to the chicken regression (o, —). A, B, Tibiotarsus length. C, D, Tarsometatarsus length. E, F, Toe length.

Bill growth differed from the other parameters measured and was best described by a second-degree polynomial equation. The regression equation obtained using the ostrich model indicated a greater bill growth rate than that generated using the chicken model. When both models were tested using 42.8 mm as a mean bill length for a wild kiwi at time of hatch, the ostrich model predicted a much longer incubation length of 131.4–148.4 days than the chicken model's prediction of an 89.7–94.5-day incubation length.

Discussion

This study of kiwi embryo development in relation to estimated age is a necessary first step towards a better understanding of kiwi embryology. Like Parker (1891), we found that the length of the embryonic kiwi hindlimb increased rapidly and regularly in a linear fashion. Bill growth was well-described by a second-degree polynomial regression, with the ostrich model predicting a slower rate of growth than the chicken model. The bill growth rate predicted by the ostrich model for the kiwi is also similar to the rate found in ostrich embryos (Gefen & Ar 2001). Despite the ostrich being a much larger bird than the kiwi, the kiwi must attain a longer bill length relative to head size than the ostrich. Our data suggest that kiwi embryos achieve this by growing their bill at a similar rate to ostrich embryos but that they maintain this growth rate throughout their much longer incubation period, which has the potential to be twice as long as the 42-day incubation period for ostrich eggs.

Neither the chicken model nor the ostrich model gave a better estimate of incubation age for the known-age embryo, with both models over-estimating the age of this embryo by 2–7 days. In contrast, middle stages of development had similar age estimations for the kiwi embryos using both models. In later stages, the estimations diverged, with the ostrich model producing older estimated ages than the chicken model. This divergence may indicate that early embryo development follows similar patterns in the chicken and ostrich, but different patterns in late development. For the oldest embryo specimens, which were probably at or close to point of hatch, estimations agree well with the maximum incubation length used in this study. The marked difference between the two estimations of incubation length that was produced by using an average wild hatchling bill size may indicate that the ostrich model does not accurately describe kiwi bill growth throughout the incubation period. At this point, it is unclear whether the chicken, which is more similar to the kiwi in size, or the ostrich, which has a closer phylogenetic relationship, better predicts kiwi embryo age. However, good agreement between ages predicted by the ostrich and the chicken models supports the presented ageing scheme as a useful, but not exact, tool, and highlights the need for validation of this study with more known-age embryos.

The exact mechanism of an embryo's growth rate is not entirely clear, but may result from an antagonistic relationship between ‘supply’ organs such as the intestines and ‘demand’ organs such as the muscle or bone rather than the chick's postnatal developmental mode (Starck 1994; Lilja et al. 2001). The developmental mode is traditionally determined by a chick's physical characteristics, behaviour and level of self-sufficiency at hatch, where the most altricial birds hatch completely dependent upon their parents and the most precocial require little to no parental care (Nice 1962; Starck & Ricklefs 1998). A bird's precociality is not simply a measure of its postnatal growth rate, which is generally, but not exclusively, slower in precocial birds (Ricklefs 1968, 1973, 1979a,b; Starck 1998), but rather a measure of its functional maturity at the time of hatching. In the case of a precocial species such as the kiwi, ostrich and chicken, the limbs must achieve a great deal of functional capability during the incubation period for the chick to walk and feed independently soon after it hatches.

The rate and timing of different embryonic tissues developing functional maturity differs from species to species (Blom & Lilja 2005) and possibly within species (Lilja et al. 2001) and is a consequence of selection for late ontogenetic characters (Richardson 1999). Gefen & Ar (2001) observed temporal differences in several structures between ostrich and chicken embryos, much as we found differences between kiwi embryos and both ostrich and chicken embryos. The relatively rapid growth of the kiwi bill became apparent early in the embryo series, at Embryo C, and persisted throughout development to produce the large characteristic bill that allows for feeding and highly specialized prey detection (Cunningham et al. 2007; Martin et al. 2007). The fast embryonic growth leading to advanced function of the kiwi's limbs and bill during incubation may restrict the ability of those tissues to grow quickly during postnatal development (Kirkwood et al. 1989; Ricklefs et al. 1994; Starck 1994), as is evidenced by kiwi not reaching skeletal maturity until 4–6 years old (Beale 1991; Bourdon et al. 2009).

The obvious limitations of this study are the use of embryos of non-ideal quality that may obscure developmental landmarks, of unknown age and with uncertain cause of death, all of which make it possible that these embryos may not accurately represent normal development. The sample population is also small and skewed towards older embryos. The abundance of late-stage embryos is likely to be a consequence of the BNZ Operation Nest Egg™ programme where kiwi eggs are actively sought and collected from the wild in order to be hatched in captivity. Eggs are less likely to be found until the later stages and the more fragile early-stage embryos are more likely to be damaged or destroyed during collection or to have decayed beyond use by the time the egg is collected. The conservation status of kiwi does not allow for the sacrifice of healthy embryos. Validation of current embryo staging models is necessary, so as more specimens become available, they should be used for this purpose.

In conclusion, by providing detailed descriptions of a collection of kiwi embryos of unknown age, we have directly related the development stages and defining features of the chicken, the ostrich and the kiwi to establish an ageing scheme for brown kiwi embryos based on morphological measurements. This model is of use to conservation management in estimating the date of lay, the date when the embryo would have been expected to hatch, or when the embryo died. Our model may also be of use in identifying critical stages of incubation for kiwi embryo growth and development.

Acknowledgements

We would like to acknowledge the financial assistance provided by the Frank M. Chapman Memorial Foundation. We also thank the Department of Conservation for supplying the eggs used in this study.