Patterns of Growth in Birds. v. a Comparative Study of Development in the Starling, Common Tern, and Japanese Quail

--This study compares developmental changes in the behavior, morphology, and body composition of three species of birds with different types of development: Starling (Sturnus vulgaris) (altricial), Common Tern (Sterna hirundo) (semi-precocial), and Japanese Quail (Coturnix coturnix japonica) (precocial). The Starling grows four times and the tern, two and one-half times more rapidly than the quail. Although the adult weights of the three species are similar, the tern neonate is about twice as large (12.8% adult weight) as Starling (6.7%) and quail (5.8%) hatchlings. Both the tern and quail move about and regulate their body temperatures at hatching. The quail chick begins to fly when it is half-grown, the Starling and tern when they are nearly fullgrown. At hatching, the tern and quail have larger proportions of nonlipid dry matter and less water in their bodies than does the Starling, indicating greater maturity. The body proportions of neonates are similar, except that the integument and legs of the Starling are smaller than those of the tern and quail. During development, proportion of pectoral muscles and, to a lesser extent, wings, increases in all three species, proportion of legs in the tern decreases and, proportion of head in the quail and, to a lesser extent, the Starling and the tern decreases. Development of homeothermy in the Starling is accompanied by the development of plumage and maturation of the leg muscles to provide a source of heat. In most tissues, increasing function is accompanied by decreasing water content and decreasing growth rate. The functional maturity of the leg appears to set the overall pace of development of the chick. In the tern, rapid body growth is consistent with slow leg growth because the relative size of the legs decreases continually during development. Adult terns have small legs compared both to other adult birds and to newly hatched chicks. Large egg size in the tern results in large hatching size and further reduces the postnatal growth increment of the legs, thereby reducing the postnatal development period. The results of this study support the hypothesis that rate of increase in body weight is inversely related to degree of functional maturity. Several refinements of this general pattern are also evident: (1) the growth rate of the organism is limited by the most slowly growing component, (2) the relative proportions of body components must remain within limits determined by their function during the development period, and (3) post-hatching growth rate may be increased under certain circumstances by increasing embryonic growth of organs that are functionally mature at hatching. Received 17 May 1976, accepted 5 September 1978. ORNITHOLOGISTS have paid little attention to growth rate (here meaning the rate of increase in body weight during the development period) except where variation in growth rate between species either is associated with differences in clutch size or reflects differences in the nutrition of the young (Lack 1968). The adaptive basis of growth rate traditionally has been viewed as a balance between selection of faster growing individuals, because they are vulnerable to predators for shorter periods, and selection of slower growing individuals, because they require less energy per unit time and thereby permit larger family size. According to this view, optimum growth rate is determined by selection directly upon growth rate and is independent of other aspects of development. The hypothesis that growth rate represents an optimum balance between selection for low rates of energy requirement and short development time would predict that variation in growth rate between species should be related to predation rate and pattern of energy utilization. Lack (1968) found a direct relationship between nestling 10 The Auk 96: 10-30. January 1979 January 1979] Avian Growth Patterns 1 1 period, which he used as an index of growth rate, and clutch size in passerines and other orders of birds having altricial development. A limitation to Lack's analysis is that nestling period need not correspond to growth rate; in passerines and, to some extent, in larger altricial species, age at fledging varies with respect to the body weight growth curve (Ricklefs and Hainsworth 1968, Ricklefs 1973). Furthermore, defining the age at fledging for nidifugous species presented such problems that Lack could not include their growth rates in his discussion of adaptations for reproduction. The application of curve-fitting techniques to growth curves (Laird 1965, Ricklefs 1967, Taylor 1968) has permitted more direct analysis of variation in growth rate than afforded by comparison of nestling periods. I have found that the characteristic growth rates of species are related to adult body weight, nestling period, food availability (particularly in seabirds with broods of one chick and possibly in tropical birds, especially those that feed fruit to their young), and type of development pattern (altricial versus preco1⁄2ial and intermediate types) (Ricklefs 1968, 1973). But I failed to find any relationship between growth rate and nesting mortality rate, which would be expected in accordance with the traditional view of growth rate (Ricklefs 1969a). Tropical passerines grow more slowly than temperate-zone passerines, even though they suffer twice the nest mortality on average (Ricklefs 1968, 1969b, 1976). Among temperate-zone passerines, growth rate and nest mortality rate are positively, but not significantly, correlated (Ricklefs 1969a), although growth rate and length of the nestling period are strongly related (Ricklefs 1968). Raptors, which as a group have low nest mortality, grow several times more rapidly than galliforms of similar body weight, which suffer relatively high mortality during the early development period. Additional examples further emphasize the lack of empirical relationship between growth rate and mortality. According to a mathematical model of factors influencing growth rate in altricial species (Ricklefs 1969a), the optimum growth rate is the most rapid that can be supported by the parents' ability to feed their young. The model, based solely on the relationship of growth rate to energy requirement and nestling mortality, predicts that birds will always sacrifice family size to reduce development period and therefore will rear a single offspring at a rate limited only by the parents' capacity to satisfy the energy and nutrient requirements of the growing chick. The inconsistency of this theoretical result with the breeding behavior of birds, particularly the prevalent characteristic of laying more than one egg per clutch, illustrates the inadequacy of the model, particularly the assumption that growth rate evolves without internal constraints. I have proposed two types of internal constraints on growth rate based on the requirement of growth processes for energy and nutrients, on one hand, and for tissue capable of embryonic activity, on the other (Ricklefs 1969a, 1973). First, the allocation of energy by precocial birds to such mature functions as activity and temperature regulation could limit the energy available for growth. Total energy intake may also be limited by the development of feeding behavior. Second, the acquisition of functional maturity through cellular differentiation could limit the proportion of a tissue that is capable of cell proliferation and growth. If these processes were themselves limited by cellular (biochemical and molecular) constraints whose nature is determined without regard to selection on the growth rate of the organism, the proportion of embryonic tissue in an individual could determine its rate of growth. 12 ROBERT E. RICKLE3⁄4S [Auk, Vol. 96 Hypotheses based on internal constraints do not exhaust the potential factors underlying variation in growth rate among birds (see Dunn 1973, in press; O'Connor 1975, 1977 for further discussion of growth rates). Nor are the two hypotheses described above necessarily mutually exclusive or even distinguishable by experimental analysis. At this point, our understanding of the factors that determine growth rate is limited by our understanding of growth and development patterns themselves, particularly the internal changes in body proportions and constituents that accompany the transition from egg to adult (Ricklefs 1967, 1975; O'Connor 1977). This paper examines developmental changes in the behavior, morphology, and anatomy of the Starling (Sturnus vulgaris), Common Tern (Sterna hirundo), and Japanese Quail (Coturnix coturnix japonica). The species were chosen for their varied types of development. The Starling is altricial, lacking mature function and self-sufficiency at hatching (Nice 1962), and grows rapidly. The Japanese Quail is precocial, that is, highly self-sufficient in feeding and temperature regulation at hatching (Wetherbee 1961, Spiers et al. 1974), and grows slowly. The Common Tern mixes altricial and precocial traits. Tern chicks are nearly self-sufficient in temperature regulation at hatching (LeCroy and Collins 1972) but are fed by their parents. Like the quail chick, the newly hatched tern has open eyes, well-developed down, and runs about (though neither can fly at hatching), but it grows almost as rapidly as the Starling (Ricklefs 1968). The tern's rapid growth is inconsistent with the idea that growth rate is constrained by maturity of function. One purpose of comparing tern chicks with those of species having different types of development is either to resolve this inconsistency or to reject the hypothesis about growth rate variation that is based on a precocity-growth rate trade off. MATERIALS AND METHODS Starlings were studied between 1970 and 1974 at a colony of free-living birds attracted to nesting boxes at the Stroud Water Research Laboratory of the Academy of Natural Sciences of Philadelphia near Kennett Square, southeastern Pennsylvania. Japanese Quail eggs obtained from Truslow Farms, Chesterton, Maryland, were incubated, and the chicks raised, in my laboratory. Common Terns were studied in 1972 at the Great Gull Island, New York, field station of the Linnean Society of New York City. Capacity for temperature regulation in the Starling and quail was determined from the body temperatures of chicks exposed to ambient temperatures of 17-23øC for 30-45 min. Temperatures were obtained with a telethermometer probe (Yellow Springs Instruments) inserted through the mouth into the proventriculus. During the test, the young were suspended in small cheese-cloth bags, within which their motion was restricted. Measurements were made during the middle of the day; none of the subjects had been deprived of food. LeCroy and Collins (1972) used a similar technique to measure temperature regulation in 1 to 4 day-old Common Terns. A coefficient of temperature regulation was calculated by the expression (Tyoung Tamt•lent)/(Tadult Tarns)lent) where T is temperature; the body temperature of adults was designated as 42øC. The coefficient of temperature regulation thus measures the fraction of the temperature gradient between an adult and its environment that can be maintained by a young bird. Pedal activity was determined for the Starling and quail in a circular pen 1 m in diameter with sides 20 cm in height. The floor of the pen was marked in squares 10 cm on a side. Young were placed in the pen and the number of squares entered in a 30-s trial period was counted. Activity was expressed as number of squares per minute. Flight capability in the Starling and quail was determined by dropping young from a height of 5 ft (1.52 m) onto a soft mat. The horizontal distance covered by the bird before landing was measured, and the angle of descent, measured from the horizontal, was calculated as arctan (h/d) where h is the height from which the chick was dropped and d the horizontal distance. Birds less than 10 days old were not tested. The length of the wing, tarsus, and a primary feather (outer for the tern, and fifth for the Starling and quail), and the area of the wings were measured on all the birds collected for anatomical analysis. Wing January 1979] Avian Growth Patterns 13 area was measured by tracing the outline of the outspread wings onto paper and comparing the weight of the outlined area to a piece of paper of known area. Series of chicks of each species were dissected into 10 components, each of which was analyzed separately to determine amounts of water, lipid, and nonlipid dry constituents. Although all the specimens were frozen and stored for variable periods before analysis, the difference between the fresh weight and the summed weights of the constituents were generally less than 5% of the fresh weight for the smallest young and less than 2% for adults. The components analyzed were integument (skin and feathers together), leg, wing, pectoral muscle, heart, liver, stomach, intestine, head, and remainder of body (primarily the back, neck, rib cage, pelvic girdle, lungs and kidneys). Components were dried to constant weight in a vacuum oven at 50-60øC. Lipids were extracted with a 5:1 mixture of petroleum ether (30-60øC b.p.) and chloroform (see Ricklefs 1975). Components were placed in pre-weighed aluminum pans and weighed on a Sartorius analytical balance. Although the ages of all young used in the analysis were known to within 1 day, a developmental chronology was constructed from wing length and, for younger birds, partly from body weight and tarsus length. Because this study examined differences between species rather than variation within a species, developmental age was used so as to minimize the deviations of individuals from the average development pattern of the population. The growth rate of each species was determined by fitting the growth curve by a logistic equation, having the form W(t) = A/(1 + exp[-K(t t0]) where W(t) is the weight at age t, A is the asymptote, or weight plateau, of the growth curve, K is the growth rate constant, and ti is the age at the inflection point of the growth curve•the point of maximum growth rat•which occurs at one-half the asympotic weight on a logistic curve. The logistic equation and methods used to fit the equation to growth data are described by Ricklefs (1967, 1968). A growth index was calculated for each species from the age at inflection and the growth rate constant. The growth index is a time scale adjusted by the growth parameters so as to make the growth curves of all species coincide when plotted as a function of their particular growth indices. The growth index is 0 at the age of inflection; its absolute value increases in both directions along the time axis by units equal to the time required for growth between 10 and 50% of the asymptote. This growth unit may be calculated directly from the growth rate constant by the expression, growth unit (days) = 2.20/K (Ricklefs 1967). The growth index allows one to portray the growth of body components, the acquisition of mature function, and the timing of such discrete events as hatching and fledging with respect o growth in body weight, yet retain their proper temporal relationships. The growth index thus reveals differences in the development patterns of species independently of differences in rate of increase of body weight. Growth indices used in this analysis are based on logistic equations fitted to data of Kessel (1957) for Starlings obtained from early broods of one season near Ithaca, New York, data of LeCroy and Collins (1972) for Common Terns obtained on Great Gull Island in 1968, and unpublished data for Japanese Quail raised in the laboratory. The first day of post-hatching life was designated as day 1; young aged 1 day were thus between 0 and 24 h old.