Embryonic and Posthatching Development of the Barn Owl (Tyto alba): Reference Data for Age Determination

Transcription

1 DEVELOPMENTAL DYNAMICS 233: , 2005 RESEARCH ARTICLE Embryonic and Posthatching Development of the Barn Owl (Tyto alba): Reference Data for Age Determination Christine Köppl, 1 * Eva Futterer, 1 Bärbel Nieder, 2 Ralf Sistermann, 2 and Hermann Wagner 2 The normal development of the barn owl was documented with the intent of providing a guideline for determining the maturational stage of embryos and posthatching individuals. Embryonic development up to stage 39 could be well described using the well-known developmental atlas for the chicken (Hamburger and Hamilton [1951] J. Morphol. 88:49 92). For later stages, limb size was established as a suitable indicator. In addition, measuring the egg s vascularized area through candling was found to be a useful, noninvasive method for staging very early embryos, up to stage 25. An average relationship between incubation period and embryonic stage was derived, which showed that development in the barn owl initally lags that in the chicken. For posthatching individuals, skeletal measures (tarsal and ulnar length, skull width and length) were the most reliable parameters for judging maturation, up to 1 month. For older individuals, feather development (e.g., length of primary wing feathers) provided the only cue. Developmental Dynamics 233: , Wiley-Liss, Inc. Key words: embryonic stages; bird; avian; altricial; maturation; egg; developmental stage Received 17 December 2004; Revised 25 January 2005; Accepted 26 January 2005 INTRODUCTION The barn owl (Tyto alba) is an effective nocturnal predator with highly specialized visual and auditory systems. The extraordinary precision in sound localization of this bird has interested auditory researchers for several decades, and barn owls are now an established experimental model in sensory research, including developmental studies (e.g., Knudsen, 2002; Kubke et al., 2002; Konishi, 2003). Thus, there is a growing need for basic data on the development of the barn owl, both embryonic and posthatching, which motivated the present study. Researchers need to be able to identify the maturational age of an individual independent of variations in extrinsic factors, e.g., incubation temperature or feeding schedule, and independent of the sensory and neural parameters studied. The reference work for general avian embryonic development is clearly the atlas compiled by Hamburger and Hamilton (1951) for the chicken. To our knowledge, the embryonic development of barn owls has not been investigated at all. Indeed, very few descriptions of the embryonic development of altricial birds are widely available (review in Ricklefs and Starck, 1998). Our aim, therefore, is to provide normative data that can be used for staging barn owl embryos and to emphasize similarities and differences seen between the precocial chicken and the altricial barn owl. The posthatching development of barn owls has already received some attention. Some of these studies provide quantitative growth data that can be used for staging individuals of unknown age (Schönfeld and Girbig, 1975; Bunn et al., 1982; Lenton, 1984; Wilson et al., 1986, 1987; Haresign and Moiseff, 1988; Percival, 1992; Brandt and Seeba, 1994; Taylor, 1994; Durant and Handrich, 1998; Shawyer, 1998). However, different 1 Lehrstuhl für Zoologie, Technische Universität München, Garching, Germany 2 Institut für Biologie II, RWTH Aachen, Aachen, Germany Grant sponsor: Deutsche Forschungsgemeinschaft; Grant number: KO/ *Correspondence to: Christine Köppl, Lehrstuhl für Zoologie, Technische Universität München, Lichtenbergstrasse 4, Garching, Germany. christine.koeppl@wzw.tum.de DOI /dvdy Published online 28 April 2005 in Wiley InterScience ( Wiley-Liss, Inc.

2 BARN OWL EMBRYONIC AND POSTHATCHING DEVELOPMENT 1249 subspecies were studied, which differ substantially in adult size (Taylor, 1994). Our second aim, therefore, was to provide a coherent set of quantitative parameters for aging posthatch barn owls of the subspecies Tyto alba guttata and T.a. pratincola, which are both used in research on sensory development. RESULTS Definition of Embryonic Stages Fig. 1. Correlation between embryonic stage and the diameter of the vascularized area in eggs of T.a. pratincola. Most candled eggs were evaluated several times and thus contributed more than one data point each. The solid line is a linear regression fit to all data shown (y 2.406x ; n 164; r 0.91; P 0.001). Note that the regression line crosses the abscissa significantly above zero, which corresponds to the observation that no blood vessels were discernable in the first few days. Hamburger and Hamilton s (1951) developmental series for the chicken was used as a guideline for staging owl embryos. This system was applicable up to stage 39, where qualitative criteria are used, despite some subtle, characteristic differences between owl and chicken (see list of stages below). However, beyond stage 39, the Hamburger and Hamilton (1951) scheme relies entirely on quantitative measures of beak and toe length in the chicken and, thus, cannot be applied to other species. In a revised, more generally applicable scheme, Ricklefs and Starck (1998) condensed Hamburger and Hamilton s stages into a single stage 40, defined stage 41 as characterized by external pipping and stage 42 as the hatchling. According to this scheme, only 42 (instead of the original 46) stages are thus defined. The first 39 stages are identical to the original description by Hamburger and Hamilton (1951) and cover the formative development of external features. Stages are dominated by species-specific embryonic growth without any qualitative changes in externally visible features. We have adopted this revised 42-stage table for the following description of barn owl embryonic development. In the barn owl, we never observed any development other than the formation of the air bubble in eggs younger than 2 3 days. In individual eggs, this delay could be up to 5 days. Such a delay was also seen independently by a second laboratory (Massoglia, 1997; C.E. Carr, personal communication). Candling the eggs with strong fiber-optic lights proved to be nearly as informative as opening them at early stages and, thus, provides a valuable noninvasive and nondestructive technique for early staging. The diameter of the vascularized area (covered by blood vessels) was most easily identified and correlated significantly with early developmental stage (Fig. 1). Owl embryos judged to be beyond stage 39 (on average 21 days of incubation or older) were staged using limb size, which we defined as the sum of tarsal and ulnar lengths. Those two parameters were selected from a variety of possibilites, because each was well defined and reliable to measure. A linear regression adequately described the limb growth with late incubation age (Fig. 2A), and this regression was then used to define the final stages of embryonic development, assuming an average incubation time of 32 days (Meyer and Wagner, 1995). The mean age for external pipping, given as 2 days before hatching (Rich and Carr, 1999), served as an additional guideline in defining stage 41. Figure 2B then replots limb size as a function of stage. Description of Embryonic Stages in the Barn Owl Due to the restricted availability of barn owls, we did not have as large a Fig. 2. Limb size in embryos of T.a. pratincola. A: Limb size as a function of incubation age. The solid line is a linear regression fit to the data above and including E19 (y 0.916x 1.793; n 20; r 0.832; P 0.001). B: Limb size as a function of embryonic stage. The solid line is an exponential fit, found to represent well the growth up to stage 39 (y e (0.2114x) ). Later stages were not included in the fit, because limb size was the primary criterion used to define those same stages (see text). set of embryos of virtually continuous ages as was used by Hamburger and Hamilton (1951). This limitation explains the irregular age gaps in the following lists. Observations obtained by candling eggs (examples shown in Figs. 3 7). E0: No air bubble. E2 4: Air bubble increases in diameter, yolk darkens. E3 4 (stage 3 6): First blood vessel discernable, first sight of embryo, staging according to the relative

3 1250 KÖPPL ET AL. Fig. 3. Top: A 6-day-old egg of T.a. pratincola, under strong candling lights. A small vascularized patch of 7 mm diameter could be discerned (arrowheads) which is very difficult to see, however, in the photograph. Bottom: The embryo extracted from the same egg, opened immediately after the candling. It was at stage 11. E6, embryonic day 6. length of the primitive streak. E5 6 (stage 9 13): Diameter of vascularized area approximately 10 mm, staging according to the number of somites. E6 8 (stage 14 18): Diameter of vascularized area mm with more branches, embryo appears bent (i.e., flexures discernable), heart beat visible from approximately E7 onward. E8 10 (stage 20 26): Diameter of vascularized area more than 30 mm, blood vessels start to grow over yolk, embryo begins to descend. E11: Eye spot visible. Observations on opened eggs and dissected embryos (Figs. 3 14). Stage 11 (Example shown in Fig. 3): 13 pairs of somites, 3 primary brain vesicles clearly visible. Fig. 4. Top left: Another 6-day-old egg of T.a. pratincola, under strong candling lights. Here, a vascularized patch of 12-mm diameter was discernable (arrowheads). Bottom left and right: The embryo extracted from the same egg, opened immediately after the candling. It was at stage 13.5 (arrowhead points to the otic pit). E6, embryonic day 6. Stage 13.5 (Example shown in Fig. 4): Approximately 22 pairs of somites, cranial and cervical flexures make a broad curve, otic pit clearly visible. Stage 18 (Example shown in Fig. 5): 30 or more somites, leg buds slightly larger than wing buds, trunk rotation approximately up to somites 14 15, otic vesicle and nasal pit clearly identifyable, eyes unpigmented. Stage 18.5 (Example shown in Fig. 6): 38 or more somites, leg buds slightly larger and bulkier than wing buds, body rotation nearly complete (only posterior end of trunk slightly bent), tail bud curved (its tip pointing forward), otic vesicle and nasal pit clearly identifiable, eyes unpigmented. Stage 23.5/24 (Examples shown in Figs. 7 and 8): Both wing and leg buds approximately equally as long as wide, eyes pigmented. Stage 25: Legs and wings bent; digital plate on wings and toe plate on legs present, but no digits discernable; eyes pigmented. Stage 29 (Example shown in Fig. 9): Legs and wings bent; digital plate still round and flat, with faint indications of fingers; toe plate still round, with indications of three toes; visceral arch collar present; neck elongated; external auditory opening present and slit-like; eye lenses facing to the sides; beak clearly visible. Stage 29.5: Small grooves present between fingers, second finger longer than the others; four toes with webbing in between; median limb size (sum of ulnar and tarsal length), 2.25 mm (n 2); visceral arch collar present; mandibular process clearly visible; eye lenses facing to the sides; beak pronounced, but no egg tooth yet. Stage 32 (Example shown in Fig. 10): Bending of wings at elbow joint and legs at knee joint very clear; three fingers with thin webbing in between, second finger being longer than the others; toes with slightly concave webbing in between; median limb size, 5.25 mm (n 2); visceral arch collar

4 BARN OWL EMBRYONIC AND POSTHATCHING DEVELOPMENT 1251 Fig. 5. Top: An 8-day-old egg of T.a. pratincola, under strong candling lights. Note the distinct vascularized area (25-mm diameter) and the embryo situated off-center within it. Bottom: The embryo in situ within the same egg, opened immediately after the candling. It was at stage 18. E8, embryonic day 8. Fig. 7. The same individual egg of T.a. pratincola, illuminated by strong candling lights at different ages, as indicated. Note the clear increase in vascularized area with increasing age (embryonic day [E] 7, 20 mm; E8, 28 mm; E9, 35 mm), as well as the flexed shape of the embryo, which was obvious at E8 and E9. Bottom right: The embryo in situ within the same egg, opened at E9. It was at stage 24. Fig Embryo of T.a. pratincola, at stage Fig. 8. Embryo of T.a. pratincola, at stage has disappeared; anterior part of mandible has reached the beak; both external ear openings clearly visible; eyes starting to orient forward, no scleral papillae like in the chicken; egg tooth discernable, but covered by a membrane and thus not very distinct; faint feather germs on the back, none on the legs; one distinct row of feather germs on each side of the tail. Stage 35 (Example shown in Fig. 11): webs between fingers and toes have nearly disappeared; fingers have lengthened, first and second finger clearly visible; four toes clearly visible, with faint primordia of claws; median limb size, 8.5 mm (n 5); eye lens is larger than in the chicken and clearly oriented forward; possibly two or more faint scleral papillae; eye

5 1252 KÖPPL ET AL. Fig. 9. Embryo of T.a. pratincola at stage 29. To the right of the photograph of a whole embryo, schematic drawings of a wing (top) and a leg (bottom) are shown, illustrating the beginning digit formation, which is typical for this stage. The calibration bar only applies to the photograph. Fig. 10. Embryo of T.a. pratincola at stage 32, shown in the same format as in Figure 9. Note that the digits are not yet separated at this stage. muscles conspicuous; eye lids and nictitating membrane cover the eye s outer half; beak continues to grow longer; faint feather germs on the back (very clear first and second rows); faint third row of feather germs on the tail; also clear scapular feather germs, but still faint on sternum and legs. Stage 35.5: Limbs have grown; claws distinct on fingers faint on toes; median limb size, 10.0 mm (n 7); lids and nictitating membrane approach cornea; lens clearly oriented forward; two or more faint scleral papillae; feather germs like stage 35, only more conspicuous; faint germs of wing feathers. Stage 36.5 (Example shown in Fig. 12): claws on first and second finger and toes clearly present with faint curvature; median limb size, 12.0 mm (n 7); lids and nictitating membrane have reached cornea; feather germs on back, tail, shoulder, breast, and wing conspicuous and conical, smaller on legs, hand, and lower mandible. Stage 37: Median limb size, 14.5 mm (n 7). Stage 37.5: Claws on toes clearly curved; median limb size, mm (n 8); circumference of lids leaves ellipsoid eye opening; feather germs surround the eye (lower lid with one row, none on upper lid); feather germs on back, tail, back of the legs, and on wings have grown to long tapering cones; beak grows conspicuously. Stage 38 (Example shown in Fig. 13): Claws on fingers are curved; both toes and toe claws are curved; median limb size, mm (n 10); circumference of lids leaves narrowed eye opening, the lower lid covering approximately half the cornea; two to three rows of feather germs on lower lid; small feather germs also on toes; large conical feather germs on upper legs, wings, shoulders, and lower mandible; feathers present on back and tail. Stage 39: Median limb size, 19.5 mm (n 4). Stage 40: Toes and toe claws strongly curved; claws now colored white; limb size up to 25.5 mm; eyes completely closed; whole body covered in feathers, including eye. Stage 41 (Example shown in Fig. 14): External pipping and intermittent vocalizing; limb size, mm. Stage 42: Hatching. Limb size, 27 mm. Relationship of Incubation Age and Embryonic Stage in the Barn Owl To provide a description of the average relationship between embryonic stage and incubation age in the barn owl, a third-order polynomial regression was fitted to our data (Fig. 15). The fit was constrained to reach em-

6 BARN OWL EMBRYONIC AND POSTHATCHING DEVELOPMENT 1253 Quantitative Measures of Posthatching Growth Fig. 11. Embryo of T.a. pratincola at stage 35, shown in the same format as in Figure 9. At this stage, the digits are nearly completely separated. Fig. 12. Embryo of T.a. pratincola at stage 36.5, shown in the same format as in Figure 9 (photo and drawings, respectively, are from different individuals of the same stage). Note that feather germs are now conspicuous over most of the body, including the limbs. bryonic stage 42 at incubation day 32, which had been determined previously as the average incubation time for the same breeding group that supplied our embryos (Meyer and Wagner, 1995). The best fit crossed the abscissa at E2.4, reflecting the initial apparently dormant period where no embryonic development could be observed. A range of quantitative measures was taken for posthatching owls, in an effort to provide a combination of parameters that may be used for aging, some of which may be particularly useful in certain age brackets. A traditional, but highly variable measure is body mass. The most rapid increase in body mass occurred between 10 and 35 days posthatching (P10 P35), with a peak reached at P35 P45, and a subsequent phase of reduction to fledgling weight. In both subspecies, a difference was noted between the growth curves of parent-raised and handraised animals, such that the weight of hand-raised owls lagged 3 4 days behind parent-raised ones. Median adult values were 325 g (n 12) for T.a. guttata and 489g (n 42) for T.a. pratincola. Skeletal growth of the limbs was documented in the sum of tarsal and ulnar length, as already used for embryos. The limbs grew rapidly and continuously during the first month and approached adult values after that. The average growth was wellrepresented by sigmoid logistic functions. This finding was true for tarsus and ulna separately (not shown), as well as their sum (Fig. 16). Figure 16 also includes the embryonic data (shown in more detail in Fig. 2A), to illustrate the acceleration of growth after hatching. Adult values were significantly different for T.a. guttata and T.a. pratincola (Mann Whitney U test, P 0.01), with median values of 59.5 mm and 71 mm, respectively, for the tarsus (both n 7); 98 mm (n 4) and 121 mm (n 7), respectively, for the ulna. Head and skull growth followed a similar pattern. Head and skull width were both approximately 15 mm at hatching and grew steadily during the first month posthatching (Fig. 17A,B). Adult values for the subspecies were slightly and significantly different. The size relationships, however, reversed for head and skull width, respectively, with T.a. guttata having a larger head but a smaller skull width than T.a. pratincola. We noted systematic differences in the estimates of head width by different observers and, thus, believe that head width is a less-

7 1254 KÖPPL ET AL. Fig. 13. Embryo of T.a. pratincola at stage 38, shown in the same format as in Figure 9 (photo and drawings, respectively, are from different individuals of the same stage). Feather germs now extend over the toes. Fig. 15. Embryonic stage as a function of incubation age. Filled triangles show stages determined after opening the egg and dissecting the embryo free; the thick solid line shows a third-order polynomial regression to these data, which describes well the average relationship (y x x x 3 ). Open triangles show additional data obtained by candling eggs; thin solid lines join data points obtained by repeatedly candling the same egg. suggesting that most of the difference between the heads of the subspecies was in a more massive beak in T.a. pratincola. For a parameter of feather growth, we followed Shawyer (1998), who documented the development of the seventh primary wing pin and feather lengths for a large sample of T.a. alba. Our data matched those curves well, showing the emergence of the feather at postnatal day (P) 24 P25 and its steady growth until fledging (Fig. 18). The median adult value for the seventh primary wing feather length was 185 mm (n 8) for T.a. guttata, 208 mm (n 5) for T.a. pratincola. Fig. 14. Embryo of T.a. pratincola at stage 41, i.e., pipping, shown in the same format as in Figure 9. reliable measure. Median adult head and skull widths were 48.3 mm (n 6) and 38 mm (n 7) for T.a. guttata and 44.0 mm (n 38) and 40.8 mm (n 9) for T.a. pratincola. Skull length was approximately 20 mm at hatching and grew to a median of 66.3 mm (n 8) in T.a. guttata (Fig. 17C). The median adult value for T.a. pratincola was significantly larger, at 75 mm (n 9), DISCUSSION Embryo Staging in Altricial vs. Precocial Species It may come as a surprise that Hamburger and Hamilton s (1951) definitions of embryonic stages could be readily applied to the barn owl. The chicken is clearly a precocial species, whereas the barn owl is classified as altricial by most authors (reviewed in Starck, 1993). Altricial and precocial

8 BARN OWL EMBRYONIC AND POSTHATCHING DEVELOPMENT 1255 Fig. 16. Limb size as a function of age in days relative to the hatching date, which was defined as zero and is highlighted by the dashed line. Different symbols represent data from different subspecies, as indicated. Each individual contributed only one data point. Adult medians and ranges are shown at the extreme right. The solid line represents a logistic fit through the posthatching data for T.a. guttata (y *[1 e * (x ) ] ), the dashed line is the fit for T.a. pratincola (y * [1 e * (x ) ] ). Note that embryonic data are also plotted to illustrate the change in growth rate after hatching. refer to the developmental stage of the hatchling (review in Starck and Ricklefs, 1998). This designation suggests that prominent differences should also exist between precocial and altricial embryos, at least in the later stages. That this difference is not reflected in the Hamburger Hamilton scheme is a paradox also pointed out by Ricklefs and Starck (1998). The Hamburger Hamilton scheme provides an extremely useful generalization of major, externally visible developmental events; however, it is too crude to take into account more subtle differences of potentially large functional significance. For example, whereas both chicken and owl embryos are fully covered by feathers by stage 39, the feathers of a chicken hatchling clearly have a denser and more insulating appearance than those of an owl hatchling. Also, the eyelids fully enclose the cornea of both chickens and owls in late embryonic development; however, the chicken hatchling will open its eyes immediately after hatching, whereas the owl will not do so for at least another 10 days (e.g., Bunn et al., 1982; Shawyer, 1998; unpublished observations). Similarly, a chicken hatchling will walk soon after hatching, whereas an owl hatchling is nearly immobile. These behavioral differences between altricial and precocial idividuals indicate that the stages of sensory and neural development can be very different, despite similar outward appearance of embryos. For example, Meyer and Wagner (1995) observed that by taking the day of eye opening as a reference, the time course of axon guidance activities in the visual system was comparable between the precocial chicken and the altricial barn owl. Also, with respect to the development of the auditory brainstem, a 2-weekold owlet was found to be equivalent to a hatchling chicken (Kubke and Carr, 2000). The Hamburger Hamilton scheme fails to generalize toward the end of embryonic development. In the absence of external diagnostic features, quantitative measurements of beak and toe sizes were used for late staging above 39 (Hamburger and Hamilton, 1951). Given the different sizes of avian embryos, the final stages, thus, are necessarily species-specific (review in Ricklefs and Starck, 1998). Fig. 17. A C: Head width (A), skull width (B), and skull length, all as a function of posthatching age. Different symbols represent data from different subspecies, as indicated. Adult medians and ranges are shown at the extreme right. The solid and dashed lines represent logistic fits through the posthatching data for T.a. guttata and T.a. pratincola, respectively, which describe well the average growth of each parameter. Each individual contributed only one data point, expect for the T.a. pratincola data shown in A, where 13 individuals were evaluated repeatedly as they grew up. Best fit equations for T.a. guttata wereas follows: for A, y *[1 e * (x ) ] ; for B, y * [1 e * (x ) ] ; for C, y * [1 e * (x ) ] Best fit equations for T.a. pratincola were as follows: for A, y * [1 e * (x ) ] ; for B, y *[1 e * (x ) ] ; for C, y * [1 e * (x ) ] Relationship of Embryonic Stages to Incubation Time Data given for incubation times in barn owls vary between 29 and 34 days (Schönfeld and Girbig, 1975;

9 1256 KÖPPL ET AL. Fig. 18. Length of the unfurled feather of the seventh primary wing feather (the fourth primary feather counting from the outside edge of the wing) as a function of posthatching age. Different symbols represent data from different subspecies, as indicated. Adult medians and ranges are shown at the extreme right. The solid line is plotting data given by Shawyer (1998, Table on p. 21). Fig. 19. Embryonic stage as a function of incubation age, normalized to percentage of total incubation time, for the barn owl and the chicken. Note that the data for the chicken, taken from the stage descriptions in Hamburger and Hamilton (1951) have been slightly modified, according to the 42-stage scheme reviewed in Ricklefs and Starck (1998). Taylor, 1994; Mebs, 2000), with most authors citing an average of days (Smith et al., 1974; Bunn et al., 1982; Shawyer, 1998) or approximately 32 days (Schönfeld and Girbig, 1975; Lenton, 1984; Epple, 1993; Meyer and Wagner, 1995). In relating embryonic stage to incubation age (Fig. 15), we used 32 days, the previously determined value for the breeding group that supplied our embryos (Meyer and Wagner, 1995). Of course, the length of incubation that passes to produce an embryo of equivalent stage is significantly longer in the barn owl than in the chicken (21 days). Therefore, to compare the development of both species, incubation age was normalized as the percentage of total time to hatching (Fig. 19). Even after this normalization, the two curves are clearly not coincident. Development in the owl starts with a delay, however, that is compensated for by a more rapid relative development. For the second half of their incubation time, both species appear to develop similarly. Of interest, a similar developmental delay in the early stages has been seen in a variety of bird species, including two finches (Lonchura), the budgerigar, a duck (Cairina), and a button quail (Turnix; Yamasaki and Tonosaki, 1988; Starck, 1989) and, thus, may be more typical. An immediate start of development at egg laying, as in the chicken, has only been confirmed for two other galliform birds, a turkey and a quail, as well as the pigeon (Mun and Kosin, 1960; Starck, 1989). Qualitative Differences Between Chicken and Owl Embryonic Development Although Hamburger and Hamilton s (1951) description of the chicken embryonic stages was well applicable to the barn owl, subtle differences were noted. Brain growth in the owl appeared to lag in the early stages up to 28, but the forebrain grew disproportionately after that. This finding conforms to the observation that the forebrain of the adult barn owl is substantially larger than in the chicken (compare Knudsen, 1983; Kuenzel and Masson, 1988). A futher prominent feature of the adult owl (which is not found in the chicken) are the frontally directed eyes with largely overlapping visual fields. This difference became obvious from approximately stage 30, when the owl embryos eyes began to rotate forward. The whole eye is relatively smaller in owl embryos than in chicken embryos, where it reaches almost grotesque proportions around stage 30. Another prominent difference was the near-absence of scleral papillae in the owl, which are a very prominent feature in the chicken. Beak growth generally seemed to lead in the owl relative to the chicken and the owl hatchling s beak is considerably larger than the chicken s. However, the egg tooth appeared somewhat later, at stage 32 compared with stage 30 in the chicken. In contrast, most of the feather germ development in the owl lagged that in the chicken by approximately 1 2 stages. Quantitative Measures of Posthatching Growth Various aspects of posthatching growth have been documented previously quantitatively for different subspecies of Tyto alba, including T.a. guttata, the subspecies used to document posthatching growth in this study. The most data by far exist for body mass. All studies show a rapid gain in body mass for the first 40 days or so, with a subsequent slower reduction to fledgling weight (Schönfeld and Girbig, 1975; Lenton, 1984; Wilson et al., 1987; Haresign and Moiseff, 1988; Langford and Taylor, 1992; Durant and Handrich, 1998). Our data for parent-raised T.a. guttata matched previous data from wild T.a. guttata nests very closely (Schönfeld and Girbig, 1975), suggesting that owlets raised by wild and captive parents, respectively, grow nearly identically. However, our data indicate that individuals hand-raised from hatching lag behind owlets raised in their nests. The reasons for this lag are not known, but potentially include differences in the exact composition of the food, the schedule of feeding or the lack of the food-offering call of the mother owl, and the bodily contact with her. All functions of skeletal growth have a similar time course, showing their most rapid growth within the

10 BARN OWL EMBRYONIC AND POSTHATCHING DEVELOPMENT 1257 first 30 days and approximating adult values usually within a further 10 days (tarsus: Wilson et al., 1987; Langford and Taylor, 1992, this study; tarso-metatarsus, tibia, and humerus: Durant and Handrich, 1998; head or skull length: Percival, 1992, this study; head and skull width: this study). Our tarsal data for T.a. guttata matched those of T.a. alba very well (Langford and Taylor, 1992). Head width may appear as a more easily obtained surrogate for the strictly skeletal skull width. However, as an adaptation to precise sound localization, the adult barn owl bears prominent skin flaps extending laterally in front of both ear openings, as well as its characteristic facial disk of feathers (e.g., Payne, 1971). Both features make the head appear considerably wider than the skull and, more importantly, render it somewhat subjective how head width is defined. Haresign and Moiseff (1988) applied callipers to the skin immediately bordering the ear openings. We have adhered to the same method, but still found the measurement subjective, i.e., dependent on the person measuring. Thus, head width is not a measure of equal reliability to skull width. The development of a fully adult plumage takes significantly longer than skeletal growth. Probably the best documented parameter of feather development in the barn owl is the growth of the seventh wing primary in T.a. alba (Shawyer, 1998; Table on p.21). The quill of this feather starts to emerge, on average, at P13, and the feather starts unfurling at P25. Adult feather length is achieved, on average, at P67. The data collected in this study for T.a. guttata fell very close to Shawyer s average values. A similar time course, although not differentiating between quill and feather part, has been documented for the tenth wing primary and the R1 rectrix tail feather (Durant and Handrich, 1998). The quill alone appears to reach a stable size somewhat earlier, around P45, as measured for the eighth wing primary in T.a. affinis (Wilson et al., 1987). The growth of total wing length, a measure that includes both skeletal and plumage parts, has been documented for T.a. alba (Taylor, 1994), T.a. guttata (Schönfeld and Girbig, 1975) and T.a. affinis (Wilson et al., 1987). Wing length increases slowly up to P12. After that, a fast linear growth phase follows, which levels off at adult values between P60 and P65. Quantitative Differences Between Subspecies of Barn Owl The barn owl has a world-wide distribution and several subspecies with restricted geographical distributions are recognized, which differ in plumage coloration and/or body size (review in Taylor, 1994). In this study, data from the two subspecies T.a. guttata and T.a. pratincola are presented and our comparisons of adult data showed significant differences for many parameters. It is important, therefore, to be aware of these differences between subspecies when size criteria are used for determining the age of developing owls. Typical incubation times given for barn owls by different authors vary between 30 and 32 days (see discussion above), but there is no evidence that the subspecies differ in any consistent way. However, egg size and weight appear to correlate with adult size, such that the smaller T.a. alba (Bunn et al., 1982; Taylor, 1994; Durant and Handrich, 1998; Shawyer, 1998), T.a. guttata (Schönfeld and Girbig, 1975), and T.a. affinis (Wilson et al., 1986, 1987) lay smaller and lighter eggs that hatch lighter young than do the larger T.a. javanica (Lenton, 1984) and T.a. pratincola (Howell, 1964; Rich and Carr, 1999). Differences in egg and embryo mass are not expected to affect the staging based on qualitative criteria. However, if the embryos should also differ in skeletal size, it would affect late embryo staging based on these parameters. Therefore, the quantitative data given here for T.a. pratincola embryos should be applied with caution to embryos of smaller subspecies. However, early hatchlings of both subspecies did not differ appreciably from each other. Instead, it appeared that differences only gradually became obvious as the animals grew (Figs. 16, 17). It is also important to note that differences are not equally pronounced for all parameters. The largest differences are seen in body mass, with T.a. pratincola and T.a. javanica being up to 50% heavier than T.a. alba, T.a. guttata, and T.a. affinis. (Schönfeld and Girbig, 1975; Lenton, 1984; Wilson et al., 1987; Haresign and Moiseff, 1988; Taylor, 1994; Durant and Handrich, 1998). This finding does not translate into equally large differences in skeletal size. Whereas T.a. pratincola had significantly longer leg and wing bones and a longer skull (including beak) than T.a. guttata, differences in head and skull width were small in the present study. Thus, barn owls of different subspecies appear to have a similar head size, but the heavier variants have a more stout physique, with longer wings, legs, and talons (own unpublished observations), as well as a more massive beak. Possibly, these differences are related to the prey spectrum. How to Determine the Age of Embryos and Young Barn Owls In this section, we evaluate the suitability of different parameters for aging barn owls of unknown or only vaguely known age and derive recommendations on how to best proceed. For barn owl embryos, we recommend staging them, using the descriptions given in the present study. Because these descriptions are based upon the more extensive chicken material of Hamburger and Hamilton (1951), that study should be familiar to anyone attempting the staging of barn owl embryos as well. At least up to stage 39, staging is definitely preferable even for embryos of known incubation age, because embryonic development can be quite variable, depending on the exact environmental conditions, e.g., natural vs. artificial incubation (Hamburger and Hamilton, 1951; Meyer and Wagner, 1995). If desired, stage may be converted to typical incubation age, according to the polynomial fit shown in Figure 15. In young embryos (up to approximately E10 or stage 25) the stage can be estimated well through candling. Also, measuring the diameter of the vascularized area provides a useful noninvasive parameter for following the progress of development (see Fig. 1). After approxi-

11 1258 KÖPPL ET AL. mately E10, when the embryo has descended into the yolk, it can only be staged reliably by opening the egg. Between about stages 18 and 35 (corresponding to E7 E16), qualitative limb and eye development are the major parameters used to determine stage. After that, qualitative feather development becomes the dominant parameter. Beyond embryonic stage 39 (corresponding to E23 and above), quantitative measures of growth are virtually the only cues. We suggest using the measure of limb size (sum of tarsal and ulnar length) for staging. If available, the incubation age should be given in addition to embryonic stage, because the stages become a much cruder indicator of development for these older embryos. In posthatching barn owls, a rough age bracket can initially be determined by examining the bird s plumage. If there are no primary feather quills present on the wings yet, the owlet is younger than 2 weeks and measures of skeletal size are best used. We again recommend the sum of tarsal and ulnar length, both of which can be obtained from live as well as dead animals. The typical age corresponding to the measured size can then be determined according to the sigmoid fit given in Figure 16. If feasable, skull width and length should also be used in the same way (Fig. 17), and the resulting ages can be averaged. If the owlet shows primary feather quills on the wings already, pin and feather length of the seventh wing primary (after Shawyer, 1998; Fig. 18) are excellent guidelines to aging and can be taken equally well from live and dead individuals. If the bird is still younger than P30, the skeletal measures should still be given equal weight. For older individuals, feather development is the only reliable parameter up to fledging at approximately P65. An alternative to the length of the seventh wing primary is total wing length, using the data of Taylor (1994) for T.a. alba and of Schönfeld and Girbig (1975) for T.a. guttata. Note, however, that most of these skeletal and feather growth data are only applicable to T.a. guttata and T.a. alba, two subspecies of relatively low body mass and small physique. As discussed in the previous section, skull width differed least between the subspecies and would thus be the most reliable indicator of age for other subspecies. EXPERIMENTAL PROCEDURES Determination of Incubation Age Several data sets, collected from captive barn owls over several years, were combined for this report. Barn owl eggs were obtained from a captive breeding group of Tyto alba pratincola (the North and Central American subspecies), held by one of the authors (H. Wagner), at the Max Planck Insitute für Biologische Kybernetik in Tübingen, at the Technische Universität München and at the RWTH Aachen. Regular controls of the nesting boxes ensured that the laying date of each egg was known to within hr. Because owls typically lay eggs in regular 2-day intervals (Taylor, 1994; Shawyer, 1998; unpublished observations), this timing was used as an additional criterion for age determination. Eggs were either left with the parent birds until the desired age or were taken for artificial incubation (KMB-4 incubator, 37 C, 65% humidity). Embryo Staging Hamburger and Hamilton s (1951) developmental series for the chicken was used as a guideline for staging owl embryos. The documentation of early development was initially carried out through candling 40 eggs aged between 2 and 15 days of incubation, using fiber-optic lights. Candling allowed the identification of the main features important for staging, such as the network of blood vessels around the embryo, the primitive streak, the somites, rotation of the embryonic axis, early appearance of the heart, and (later) heart beating. A subset of the candled eggs, as well as all older eggs, were opened at the desired age, the embryos extracted and fixed in 4% paraformaldehyde in 0.1 M phosphate buffer. In total, 81 preserved embryos were evaluated, aged between 4.5 and 30 days of incubation and a further 13 embryos of unknown or only vaguely known incubation age. Most embryos younger than 15 days of incubation were complete, whereas most of the older embryos were either missing the head entirely or the head had had the brain removed, for use in parallel studies of neural development. However, this did not affect the staging critically, because qualitative limb and feather development are the major parameters in the relevant age bracket. Two investigators (H.W., E.F.) first performed the staging independently of each other. Their results varied randomly within 2 stages of each other, i.e., there was no consistent difference. In a second and final round, both persons then re-evaluated the embryos together and agreed on common stages. Determination of Posthatching Age Young barn owls of various ages after hatching were obtained from the same captive breeding group as the embryos, as well as a captive breeding group of Tyto alba guttata (the Central European subspecies), held by one of the authors (C. Köppl) at the Technische Universität München. Direct controls of nesting boxes or video surveillance ensured that the hatching date was known to within 24 hr or, in rare cases, within 2 days. The hatchlings formed two groups. One group remained with the parent birds until the desired age, when they were killed by injection of an overdose of sodium pentobarbital and fixed by transcardial perfusion with 0.9% NaCl, followed by 2% paraformaldehyde and 2.5% glutaraldehyde in 0.1 M phosphate buffer. Most of these individuals were also used in acute physiological experiments on hearing and had been under general anesthesia for several hours before the overdose. A second group of posthatch individuals was hand-raised by the authors, from hatching to fledging. From these individuals, repeated measurements of selected growth parameters, mostly at daily intervals, were obtained. For reference purposes, measurements from adult Tyto alba guttata and Tyto alba pratincola (aged between 4 months and 13 years) were

12 BARN OWL EMBRYONIC AND POSTHATCHING DEVELOPMENT 1259 Fig. 20. Schematic drawings illustrating how growth parameters were defined; the dashed lines indicate the position of the calliper jaws for measurement. The individual drawings show, from top to bottom: a sidewiew of a wing, illustrating ulnar length; a sideview of a leg, illustrating tarsal length; a skull as seen from above, illustrating skull width; and a sideview of a skull, illustrating skull length. also taken. The numbers of adult reference values varied and are given separately for the different parameters (see Results section). Quantitative Measurements of Growth Parameters Body mass was determined with electronic scales accurate to within 0.1 g for specimens below 100 g and accurate to within 5 g for heavier individuals. All linear measurements were performed using callipers with a millimeter scale. Measurements of bilaterally symmetrical structures represent an average of both sides. The following parameters were evaluated. Diameter of vascularized area: measured along the egg surface with a flexible tape measure. Body mass: in cases of repeated measurements of the same hand-raised individuals, weighing was consistently performed in the morning, before feeding. Tarsal length (Fig. 20). Ulnar length (Fig. 20). Limb size: sum of tarsal and ulnar length, used as an indicator of limb growth. Head width: according to Haresign and Moiseff (1988). Skull width (after removal of skin): distance at the widest point, just behind the ear openings (Fig. 20). Skull length: from the back of the skull to the tip of the upper beak (Fig. 20). Statistical Analysis Descriptive statistics, nonparametric testing (Mann Whitney U test) and linear regression fitting were carried out using the software package SPSS 11.0 (SPSS Inc., Chicago, IL). For polynomial and logistic fits, the Solver module of Excel 2002 (Microsoft Corporation) was used. The logistic functions fitted to posthatching growth parameters (Figs. 16, 17) were constrained to asymptote to the (independently derived) adult median value. ACKNOWLEDGMENTS Many thanks to Dr. Catherine E. Carr, who generously shared unpublished data on embryonic development in barn owls. We also thank Ilona Vollpracht, Sandra Brill, and Susanne Scholz for help with embryo staging and maintaining the embryo catalogue and collection. Thanks are also to Monika Franke, who helped digitize literature data. C.K. was funded by a Heisenberg fellowship from the Deutsche Forschungsgemeinschaft. REFERENCES Brandt T, Seeba C Die Schleiereule: Ökologie eines heimlichen Kulturfolgers. Wiesbaden, Germany: AULA- Verlag. 152 p. Bunn DS, Warburton AB, Wilson RDS The barn owl. Calton, England: T. & A.D. Poyser. 264 p. Durant JM, Handrich Y Growth and food requirement flexibility in captive chicks of the european barn owl (Tyto alba). J Zool 245: Epple W Schleiereulen. Karlsruhe, Germany: G. Braun GmbH. 107 p. Hamburger V, Hamilton HL A series of normal stages in the development of the chick embryo. J Morphol 88: Haresign T, Moiseff A Early growth and development of the common barnowl s facial ruff. Auk 105: Howell TR Notes on incubation and nestling temperatures and behavior of captive owls. Wilson Bull 76: Knudsen EI Subdivisions of the inferior colliculus in the barn owl (Tyto alba). J Comp Neurol 218: Knudsen EI Instructed learning in the auditory localization pathway of the barn owl. Nature 417: Konishi M Coding of auditory space. Annu Rev Neurosci 26: Kubke MF, Carr CE Development of the auditory brainstem of birds: comparison between barn owls and chickens. Hear Res 147:1 20. Kubke MF, Massoglia DP, Carr CE Developmental changes underlying the formation of the specialized time coding circuits in barn owls (Tyto alba). J Neurosci 22: Kuenzel WJ, Masson M A stereotaxic atlas of the brain of the chick. London: The Johns Hopkins University Press. Langford IK, Taylor IR Rates of prey delivery to the nest and chick growth patterns of barn owls Tyto alba. In: Galbraith CA, Taylor IR, Percival SM, editors. The ecology and conservation of European owls. Peterborough, UK. p Lenton GM The feeding and breeding ecology of barn owls Tyto alba in peninsular Malaysia. Ibis 126: Massoglia DP Embryonic development of the time coding nuclei in the brainstem of the barn owl (Tyto alba). Master s thesis, Department of Zoology. College Park: University of Maryland. p 75. Mebs Die Eulen Europas. Stuttgart, Germany: Franckh-Kosmos Verlag. 396 p. Meyer SU, Wagner H Guidance of retinal axons by membrane components derived from the barn owl s optic tectum. Zool Anal Complex Syst 99: Mun AM, Kosin IL Developmental stages of the broad breasted bronze turkey embryo. Biol Bull 119: Payne RS Acoustic location of prey by barn owls (Tyto alba). J Exp Biol 54: Percival S Methods of studying the long-term dynamics of owl populations in Britain. In: Galbraith CA, Taylor IR, Percival SM, editors. The ecology and conservation of European owls. Peterborough, UK. p

13 1260 KÖPPL ET AL. Rich V, Carr C Husbandry and captive rearing of barn owls. Poult Avian Biol Rev 10: Ricklefs RE, Starck JM Embryonic growth and development. In: Starck JM, Ricklefs RE, editors. Avian growth and development: evolution withing the altricial-precocial spectrum. Oxford, New York: Oxford University Press. p Schönfeld M, Girbig G Beiträge zur Brutbiologie der Schleiereule, Tyto alba, unter besonderer Berücksichtigung der Abhängigkeit von der Feldmausdichte. Hercynia 12: Shawyer C The barn owl. Chelmsford, Essex: Arelquin Press. 213 p. Smith D, Wilson C, Frost H History and ecology of a colony of barn owls in Utah. Condor 76: Starck JM Zeitmuster der Ontogenesen bei nestflüchtenden und nesthockenden Vögeln. Cour Forsch-Inst Senckenberg 114: Starck JM Evolution of avian ontogenies. Curr Orn 10: Starck JM, Ricklefs RE Patterns of development: the altricial-precocial spectrum. In: Starck JM, Ricklefs RE, editors. Avian growth and development: evolution withing the altricial-precocial spectrum. Oxford, New York: Oxford University Press. p Taylor I Barn owls. Predator-prey relationships and conservation. Cambridge (Great Britain): Cambridge University Press. 304 p. Wilson RT, Wilson MP, Durkin JW Breeding biology of the barn owl (Tyto alba) in central Mali. Ibis 128: Wilson RT, Wilson MP, Durkin JW Growth of nestling barn owls Tyto alba in central Mali. Ibis 129: Yamasaki M, Tonosaki A Developmental stages of the society finch, Lonchura striata var. domestica. Dev Growth Differ 30: