AMER. ZOOL., 687-700 (1978).

Correlates and Consequences of Body Size in Nectar-Feeding Birds

JAMES H. BROWN, WILLIAM A. CALDER III, AND ASTRID KODRIC-BROWN

Department of Ecology and Evolutionary Biology, University of Arizona, Tucson, Arizona 85721

SYNOPSIS. Nectar-feeding birds are among the smallest birds and the largest pollinators. Downloaded from https://academic.oup.com/icb/article/18/4/687/2004942 by guest on 23 September 2021 Energetic costs of maintenance, temperature regulation, foraging and reproduction increase in direct proportion to body mass raised to fractional exponents, which may vary from 0.5 to 1.0; overall costs probably vary with an exponent of 0.75. Avian nectarivores acquire most of their energy from flower nectar; in so doing they compete with other nectar feeders and pollinate plants. Larger pollinators are more reliable and move pollen greater distances, but to attract them plants must secrete more nectar and protect it from utilization by smaller animals. Minimum body size of avian nectarivores (2g) appears to reflect both competition with insects and the limited capacity of the smallest birds to acquire and store energy' relative to the demands of fasting, temperature regulation, and reproduction. Hummingbirds have attained significantly smaller size than other nectar feeding birds because lower metabolic rates and use of hypothermic torpor reduce their energy expenditure relative to income. Maximum body size of avian nectarivores (approx- imately 80g) apparently reflects the upper limit of plant energy expenditure for reliable, long distance pollination. Between these limits, size variation reflects divergence to reduce interspecific competition and coevolution with plants to promote specificity.

INTRODUCTION amount of tissue being supported varies with body mass, heat exchange varies with It is appropriate that this symposium body surface area, and capacity to obtain begins with a discussion of body size, for food depends largely on linear dimensions no other characteristic of an organism of locomotor and feeding appendages. influences more aspects of its biology. To Since the bird's mass varies as the cube and appreciate the diverse and pervasive ef- its surface area as the square of its linear fects of body size, consider the energetics dimensions, birds of different body size of a nectar-feeding bird. In order to sur- can exist only if all energetic functions are vive and reproduce, the bird requires adjusted for these relationships. Allomet- energy which it obtains primarily by feed- ric scaling in relation to variation in body ing on sugar solutions secreted by flowers. size is required for all biological functions The bird must ingest sufficient energy to and can be described quantitatively by sustain its tissues, to regulate its body means of simple exponential equations. In temperature and to perform other vital birds, for example, empirically determined functions such as reproduction. The allometric equations describe relationships between body size and such quantitative We gratefully acknowledge the assistance of Renee traits as metabolic rate (Aschoff and Pohl, Vestal in compiling the data in Figure 4. D. Inouye, 1970), length of appendages (Greenewalt, N. Waser, and F. R. Hainsworth kindly supplied 1975), egg size (Rahn et a!., 1975), incuba- unpublished data. Our research on nectar-feeding birds has been funded by NSF Grants DEB 76-09499 tion time (Rahn and Ar, 1974), longevity to J. H. B. and A. K. B., GB-39260 to J. H. B., and (Lindstedt and Calder, 1976), and size of B.M.S. 73-06943, and a National Geographic Society territory (Schoener, 1968). grant to W. A. C. We thank colleagues and students Although correlates of body size varia- too numerous to mention, including many partici- pants in this symposium, for discussions and tion are numerous and well documented, suggestions. F. R. Hainsworth, L. L. Wolf, and an their biological causes and consequences anonymous reviewer made numerous suggestions for often are complex and poorly understood. improving this manuscript. Because virtually all structures and func- 687 688 BROWN ETAL. dons of an organism interact with each source. Body size also influences relation- other and are influenced by body size, it is ships between birds and plants. Relative to difficult to isolate cause and effect. A insect pollinators, birds require large profitable way to investigate the effects of quantities of energy, so we should look for body size is to study intensively organisms special benefits of avian pollinators which which lie at the extremes of established make it advantageous for flowers to secrete allometric relationships. Comparative sufficient nectar to attract birds. studies which include organisms of ex- In the present paper we examine the treme size not only reveal mechanisms influence of body size on the biology of which enable them to achieve physiological these birds. First, we discuss physiological Downloaded from https://academic.oup.com/icb/article/18/4/687/2004942 by guest on 23 September 2021 integration and ecological function, but effects of body size on avian energetics, also indicate compromises and constraints and then we consider their consequences which prevent the evolution of larger or for ecological roles (nectar consumers and smaller forms. pollinators). We also investigate the sig- Nectar-feeding birds provide an excep- nificance of variation in body size among tional opportunity for a comparative avian nectarivores and discuss the com- analysis because they lie at the narrow promises and constraints which determine intersection of two largely nonoverlapping minimum size of birds and maximum size spectra of body sizes. Hummingbirds, sun- of pollinators. birds, honeycreepers, and the other birds which feed extensively on nectar are PHYSIOLOGICAL CORRELATES among the smallest birds and the largest pollinators (Fig.l). Birds vary in size from Nectar-feeding birds obtain most of the 2g bee hummingbird to the 100kg their energy by feeding on floral nectar, a ostrich. Animals that provide pollination plant secretion rich in sugars (Ford and services for plants while foraging for nec- Patton, 1976). Birds metabolize these tar or pollen vary in size from-fig wasps, sugars and use the resulting energy for some of which weigh less than 1 mg, to their own maintenance, temperature reg- some birds and bats, which rarely weigh ulation, foraging, and reproduction. The more than 80g. Although the 40 fold varia- rate at which birds expend energy and the tion in size of nectarivorous birds is minute way they allocate it to different functions compared to the more than 50,000 fold varies with body size. variation in both birds and pollinators, it is Nectar is eaten by many birds. Fisk and accompanied by significant differences in Stein (1976) list 20 species of North the extent to which the birds are American birds in 7 families exclusive of specialized for foraging from flowers and hummingbirds, which have been observed dependent on floral nectar as a food re- feeding on floral nectar. However, the majority of nectar feeders and the most specialized ones are concentrated in four families (Table 1). Note that hum- NECTAR FEEDING BIRDS mingbirds are unique in not belonging to the order of perching birds (Pas- BIRDS seriformes), in being the only group in- habiting the New World, and in attaining a minimum body weight one-third that of POLLINATORS other avian nectarivores. The majority of temperate North American hummingbirds lO"6 lO"4 I0"2 10° I02 I04 I06 weigh approximately 3.5g which is only OOOOOI .0001 .01 1 100 10,000 1,000,000 half the size of the smallest sunbird. BODY MASS (g) Hummingbirds also differ from other nec- FIG. 1. Range of body size in nectar-feeding birds tar feeders in their capacity for sustained (shaded portion) relative to other birds and other hovering flight and for entering hypo- pollinators. Note that nectar feeding birds vary about 40 times in body mass and include both the smallest thermic torpor as a means of emergency birds and some of the largest pollinators. energy conservation (Fig. 2). CORRELATES AND CONSEQUENCES OF BODY SIZE 689

TABLE 1. Sizes and other characteristics of the four primary families of nectar-feeding birds. Range of Minimum Number of body length body mass Family Order species Distribution (mm) (g) Trochilidae Apodiformes 319 North and South 63-216 2.0 America Nectannudae Passe riformes 104 Africa, southern 95-254 6.9 Asia, Malaysia Meliphagidae Passeriformes 167 Australia and 102-432 8.5 Downloaded from https://academic.oup.com/icb/article/18/4/687/2004942 by guest on 23 September 2021 Drepanididae Passeriformes 22 Hawaiian Archipelago 102-203 8.5 6 3 Minimum body mass (mb) was estimated from length (I) using the equation mb = 8 x 10 I . Data are from Van Tyne and Berger (1976).

Since all nectar-feeding birds are rela- low temperatures, all birds maintain rela- tively small and hummingbirds are the tively constant body temperatures of ap- smallest birds of all, it is of interest to proximately 40°C (Caider and King, 1974). consider the physiological correlates of Temperatures well in excess of average small size and to ask what special charac- environmental temperatures are main- teristics of hummingbirds have permitted tained by a balance between heat produc- the evolution of uniquely small size. Such tion and heat loss from the body surface. an analysis should suggest functional pro- Metabolic rate (Hm) ot resting birds in- cesses important in determining the size of creases with body mass (mb) according to a 075 birds. fractional exponent (Hm = kmb ; see Calder, 1974, for a review). The value of k Metabolism, temperature regulation, and torpor differs between passerine and nonpasser- ine birds (Lasiewski and Dawson, 1967; Except when stressed by energy short- Aschoff and Pohl, 1970) so standard (rest- ages resulting from scarce food and/or ing) metabolic rates (SMR) of passerine nectar feeders should be 55 to 65% greater than those of hummingbirds of the same size (Fig. 3). Metabolic rate is elevated above resting levels during activity and for temperature regulation when environmental tempera- ture is below the thermoneutral range (Fig. 3). Since passerines and other birds do not differ significantly in body temperature and insulation (Calder, 1974), differences in metabolic rate are large only in the relatively narrow thermoneutral range (Lasiewski, 1963, 1965; Wolf and Hains- worth, 1972; see Fig. 3). Because insulation increases and relative body surface de- creases with increasing body size, rates of heat loss below the thermoneutral range scale to body mass with smaller fractional exponents (m042 to m0M; Kendeigh, 1969) than SMR. EXPLOITATION OF SMALL FLOWER NECTARS The energetic consequences of these FIG. 2. A hypothetical scheme of some energetic metabolic and thermoregulatory patterns aspects of hummingbird evolution which indicates for birds of varying body size are sum- consequences of small body size (from Calder, 1974). marized in Table 2, which presents calcu- 690 BROWN ETAL.

foraging and reproduction than a small bird whatever the cost per gram (Table 2). M However, it is also true that, in comparison to a larger relative, a small bird: 1) ingests a greater proportion of its body weight per day in food; 2) increases its metabolic rate by a relatively greater factor to ther- moregulate at a given environmental

temperature below the thermoneutral Downloaded from https://academic.oup.com/icb/article/18/4/687/2004942 by guest on 23 September 2021 range; and 3) uses up its fat reserves at a relatively greater rate while fasting. Since food energy and time available to forage AMBIENT TEMPERATURE CC) FIG. 3. Relationship between metabolic rate and are limited, allometric scaling of these ambient temperature predicted for hypothetical 3g energetic costs could present severe prob- passerine and nonpasserine birds from allometric lems for very small birds. equations relating standard metabolic rate and heat Most hummingbirds which have been loss to body mass (from Calder, 1974). Nonpasserities have significantly lower metabolic rates than pas- studied have the capacity to enter into and serines at warm temperatures within the zone of arouse spontaneously from torpor. Origi- thermoneutrality (level portion at lower right), but nally torpor was regarded by physiologists when these birds must expend energy for tempera- as the failure of temperature regulation. ture regulation in cold environments their metabolic rates are very similar. However, demonstrations that hum- mingbirds regularly enter and recover from torpor both in the laboratory lated values derived from allometric rela- (Lasiewski, 1963; Hainsworth and Wolf, tionships for hummingbirds. Note that 1970) and field (Calder and Booser, 1973; standard metabolic rate, rate of heat loss, Carpenter, 1974, 1976a), and that they and daily energy cost decrease with de- regulate body temperature at hypothermic creasing body size. This would appear to levels (Wolf and Hainsworth, 1972; Car- favor birds of small body size because they penter, 1976a) suggest torpor is a normal could subsist on small quantities of nectar metabolic response. Indeed, it is clear that inadequate to support large birds. How- torpor is an adaptive mechanism which ever, because standard metabolic rate and permits energy conservation, but it ap- heat loss rate scale with fractional expo- pears to be used only when food is scarce nents, whereas energy reserves (fat and and birds have difficulty remaining in posi- crop contents) should scale linearly with tive energy balance (Fig. 3; Calder, 1974; respect to body size, small birds must feed Calder and Booser, 1973; Carpenter, more frequently and have less ability to 1976r/; Hainsworth el id., 1977). In such endure periods of food scarcity than large circumstances hummingbirds can enter birds. This disparity increases as environ- torpor at night when they are unable to mental temperature decreases. foruge, temperatures are lowest, and risk Note that we discuss energetics in terms of predation is least (Fig. 2). of metabolic costs to individual birds. Nocturnal torpor has been reported Physiologists often try to make size- only in relatively small birds which have independent comparisons by dividing limited capacity for energy storage but the these rates by body mass and purport to ability to reduce energy expenditure by show an inherent disadvantage of small entering hypothermia. High rates of heat body size because it is cheaper to fuel a loss enable small birds to cool rapidly to gram of a large rather than a small bird. low body temperatures with consequent However, the unit of ecological function is low metabolic rate (Lasiewski and Lasiew- the individual, not the gram (McNab, ski, 1967). Energy conservation associated 1971). A large bird always requires more with torpor is substantial. For example, food energy, has a higher total rate of heat metabolic rate of a 5g hummingbird (Pan- loss, and expends more total energy for tcrpr insignis) torpid at 6° C is only 2O9r that Downloaded from https://academic.oup.com/icb/article/18/4/687/2004942 by guest on 23 September 2021

TABLE 2. Energy and reproductive estimates for hypothetical small hummingbirds.*

Heat loss Daily Crop Fat endur- Pre- Flight rate energy energy energy e+f ance at life- dicted Egg Adult Egg/Ad Egg Egg Body SMR (5x (mW° cost reserves reserves a 5°C span e (% surface surface con tad energy cooling b J gg k 2 Mass Kj/day" SMR) c-'r (Kj)d (Kj)f (Kj)r (days)" (days)* (yr) mass(g) ad 9) (cm2)"1 (cm )" ratio" (%SMRy °C/°C" lg 2.48 12.39 4.72 19.08 1.361 17.39 7.6 1.31 4.4-4.8 .175 17.5 1.47 10 0.29 31.04 13.81 (0.50) (0.60) (0.60) (0.73) (0.75) (0.64) (0.50) (0.84) (0.70) (0.88) (0.59) (1.17) (0.70) (0.63) (1.12) (0.97) (1.24) 4.11 20.54 6.49 25.54 2.137 34.78 9.0 1.88 5.1-5.4 .298 14.9 2.10 15.87 0.26 31.87 11.12 9 2K 73 (0.67) (0.74) (0.74) (0.83) (0.84) (0.77) (0.67) (0.93) (0.81) (0.93) (0.73) (1.10) (0.81) (0.76) (1.04) (0.98) (1.14) 50 »K 5.52 27.32 7.82 30.30 2.77 52.17 10.0 2.32 5.5-5.8 .408 13.6 2.59 20.80 0.25 32.45 9.79 m (0.75) (0.81) (0.81) (0.88) (0.89) (0.83) (0.75) (0.93) (0.86) (0.95) (0.60) (1.07) (0.86) (0.83) (1.04) (0.99) (1.09) 4g 6.82 27.32 8.93 34.20 3.347 69.55 10.7 2.70 5.8-6.1 .509 12.7 3.00 25.20 0.24 32.82 8.95 (0.80) (0.85) (0.85) (0.90) (0.91) (0.86) (0.80) (0.95) (0.89) (0.95) (0.84) (1.05) (0.89) (0.86) (1.04) (0.99) (1.07) 7. 8.02 33.60 9.90 37.57 3.871 86.94 11.3 3.03 6.1-6.4 .604 12.1 3.37 29.26 0.23 33.09 8.35 5g a O * Figures in (parens) signify percentage change going from larger to smaller bird. o " (5) 28.67m0-73 in mW from Calder, 1974. x b Assuming metabolic expansibility for thermoregulation similar to finches (Dawson and Carey, 1977). c From non-passerine heat-transfer coefficient (Calder, 1974). B d From MacMillen and Carpenter (1977). rn * Highest value for each of n=27 birds from Hainsworth and Wolf (19726), least squares log regression, vol. = 0.27Om°. r = 0.789, at 5.04 Kj/ml, obtained from weighted average sucrose-equivalent concentration 0.893 molar from 9 flower spp. in Ecuador, 19 in southwestern U.S. (Hainsworth m and Wolf, 1972r), and 5 spp. in Colorado Rocky Mis. (Waser, personal communication). 'C<* 43.8% of body mass (Odumrtn/., 1961) x 39.7 Kj/g. CO e O Endurance on reserves at SMR. Note that even at summit metabolism, reserves last > 1 day. O " Endurance on reserves at 5°C, (e+f) + C(40° - 5°). 1 (a) 17.6m0-20 for all wild birds and 16.6m018 for wild nonpasserines (Lindstedt and Calder, 1976). h N RahnpJn/. (1975) general equation n\ = 0.277nib°" predicts a value 58% high fora hummingbird (Calypte anna) as estimated, assuming a density of m 1.88g/ml (Romanoff and Romanoff, 1949) from measurements of egg given by Hoyt (1977), and 9 body mass of 3.82 from Pearson (1950). Hence above values are scaled as 0.175mb°". m 0M7 As 4.72 me derived from Hoyt's (1977) egg dimensions and scaled parallel to Meeh equations for surface area (Kleiber, 1961). " Assumed surface area = 10m0-67. ° Contact ratio = (XA egg surface x 2 eggs) -*• 'A adult surface. p Egg energy estimated from 1.05 Kcal/g in altrical eggs (King, 1973, p. 88). " Extrapolated from Fig. 38 of Drent (1975). 692 BROWN ET AL.

of a nontorpid bird resting at the same Foraging environmental temperature (Wolf and Hainsworth, 1972). Small body mass also Nectar feeding birds must expend permits rapid heating and arousal at the energy on foraging to obtain energy. completion of a bout of torpor (Heinrich Foraging takes two forms: perching and Bartholomew, 1971). The capacity of movements and flight. Metabolic costs of small birds for heat production is amazing. hopping about in vegetation and probing Maximum metabolic rates (Hmax) of 14 flowers while perched seldom are mea-

species of small birds vary with respect to sured, but they are generally assumed to Downloaded from https://academic.oup.com/icb/article/18/4/687/2004942 by guest on 23 September 2021 = size according to the relationship Hmax be in the range of 1.5 to 2 times resting 065 kmb (from Calder, 1974). metabolic rate and to scale with respect to Even the largest hummingbirds can body size with the same exponent as SMR utilize the energy saving advantages of (Gill and Wolf, 1975; Carpenter and torpor, as can their even larger relatives, MacMillen, 1976). Flight costs are greater, the swifts and goatsuckers. Passerine nec- and empirical measurements suggest tar feeders, and passerine birds in general, metabolic rate during flight may be as appear to utilize torpor less frequently and much as 5 to 7 times SMR for both hover- effectively (Fig. 4). One small sunbird, iVec- ing in hummingbirds and linear flight in tarinia mediocris (weight approximately other birds (Lasiewski, 1963; Tucker, 7.5g), studied in the laboratory by Cheke 1970; Bernstein et al., 1973; Greenewalt, (1971) entered hypothermia and reached a 1975). Because they usually fly continu- body temperature 17.5°C below active ously while foraging, hummingbirds in body temperature, but we are not aware of particular may expend much energy on evidence that any other nectar feeding flight. Recently several authors have at- birds regularly use torpor under natural tempted to estimate the energy required conditions. (power output) for hovering in hum- mingbirds (e.g., Weis-Fogh, 1972; Hains- worth and Wolf, 1972a, 1975; Epting and Casey, 1973; Greenewalt, 1975; Fein- singer et al., 1978). Most of these esti- mates are based on aerodynamic theory, and use measurements of wing length and 2 Hypothermia body weight rather than direct measure- Heterothermy ments of metabolism of hovering birds. There is some question and disagreement about exactly how energy expenditure for hovering scales with respect to body size; it appears likely that power output varies as •Temperature conformity at least the three-fourths power, and perhaps as the first power, of body mass. Field estimates of energy budgets for ter- ritorial, nectar feeding birds suggest that LOG BODY MASS flight may account for 50% or more of daily energy expenditure (Wolf and FIG. 4. Relationship between metabolic rate for pas- Hainsworth, 1971; Carpenter and MacMil- serine and nonpasserine birds (SMR; from Aschoff and Pohl, 1970) and for Hying insects. Note that with len, 1976). decreasing body size flying nectarivores appear to rely increasingly on energy conservation resulting from reduced body temperature (dashed lines). The exact Reproduction relationship for insects is uncertain because body temperature and metabolic rate of most insects vary with ambient temperature. The line plotted here was Nectar feeding birds expend energy also obtained using Heinrich's (1975«) value for incubat- for the production of offspring. Energetic ing bumblebees and assuming that metabolic rates of costs of reproduction in birds include other inserts varv as nit,0-75. energy invested in eggs and metabolic ex- CORRELATES AND CONSEQUENCES OF BODY SIZE 693

penditures for courtship, mating, nest might be expected, these birds are just building, incubation, brooding and forag- about replacing themselves. Consequences ing for dependent young. Many of these of allometric scaling on survivorship and costs are difficult to measure or estimate, reproduction are extremely important in but in general they probably scale with a the evolution of body size. fractional exponent in relation to body mass. Egg mass (me) varies with body size a 0 77 ECOLOGICAL AND EVOLUTIONARY according to the relationship me mb -

(Rahn ei ai, 1975; see Table 2). Small CONSEQUENCES Downloaded from https://academic.oup.com/icb/article/18/4/687/2004942 by guest on 23 September 2021 females invest less total energy, but a larger proportion of their reserves, in eggs The previous section discussed effects of than larger birds (Skutch, 1962). Small body size on the energetic processes essen- females must also expend relatively more tial for maintenance and reproduction. energy to carry their eggs until they are The present section is concerned with laid. Small females have relatively less ecological consequences of nectarivorous body surface area in relation to relatively birds foraging to obtain energy. By their higher cooling rates of their eggs during feeding activities birds have important ef- incubation. Once the eggs hatch, it should fects on both the plants which produce cost proportionately more to care for the nectar and the other animals which feed relatively larger young. We expect on it. These ecological interactions provide metabolic rates, and hence food require- the selective basis for the coevolution of ments, of dependent young to scale simi- communities of plants and pollinators. Be- 075 cause body size influences energy re- larly to adult SMR, approximately as mb (where m is adult body mass). quirements and foraging behavior, it has b important consequences for ecological and Not all costs can be measured in ener- evolutionary relationships among nectari- getic terms. It is reproductive success or vores and plants. fitness which is the ultimate currency of evolution that determines whether a species persists and how it evolves. Sur- Coevolution of pollinator energetics and plant vivorship and fecundity, the components of fitness, are strongly affected by body nectar secretion 0 20 size. Longevity in birds scales as mb - The basis of relationships between nec- (Lindstedt and Calder, 1976), and small tar feeding birds and plants is coevolved birds have a shorter life in which to repro- mutualism. Floral nectar has evolved spec- duce. Since clutch size tends to decrease ifically to attract animals which move pol- with body size and is typically two in hum- len between plants and promote outbreed- mingbirds (Bent, 1940), nesting success ing. While pollinating, animals obtain and survival, especially of juveniles, should energy from the sugars in nectar. While increase as body size decreases. The con- each partner in such a mutualistic relation- sequences of these relationships for small ship benefits from the activities of the birds are illustrated by data for Broad- other, natural selection acts on each part- tailed Hummingbirds {Selasphorus platycer- ner to obtain the maximum benefit for the cus, body weight 3.5g) at Gothic, Colorado, minimum cost. Most, but unfortunately elevation 2,910 - 3,000m (Waser and In- not all, of the costs and benefits potentially ouye, 1977; Inouye, pers. comm.). Females can be measured in energetic terms. Rates successfully fledge an average of 1.15 of energy intake and expenditure have young per nest, and the season is too short been estimated and net rates of energy to rear more than one brood per year. gain have been calculated for free living, Maximum recorded longevity of six years foraging nectarivores (e.g., Gill and Wolf, compares well with allometrically pre- 1975; Carpenter and MacMillen, 1976; dicted lifespan of 5.8 - 6.1 years (Lindstedt Gill, 1978). Allocation of energy by plants and Calder, 1976), but average lifespan to nectar and other floral attractants pre- appears to be less than three years. As sumably could also be measured, but the 694 BROWN ETAL. benefits of pollen transport and resulting value of nectar secreted per flower per day outbreeding can be evaluated only in terms (Fig. 5). The great variability around this of reproductive success. Lack of a com- relationship can be explained in large part mon, easily measured, currency to quan- by variation in the spatial distribution of tify costs and benefits to each partner flowers or other factors which influence complicates the study of plant-pollinator foraging costs (see Heinrich, 1975; Hein- coevolution, and raises interesting chal- rich and Raven, 1972, for excellent discus- lenges for future investigators. sions). In addition, this pattern is biased by

Nectar-feeding birds share with bats and the difficulty of measuring accurately nec- Downloaded from https://academic.oup.com/icb/article/18/4/687/2004942 by guest on 23 September 2021 a few other mammals the distinction of tar secretion rates for those flowers which being the largest pollinators. They have produce minute amounts and are polli- much higher energy requirements than nated by small insects. Except for this bias', the much smaller insects which pollinate the relationship should have a steeper most plant species. Plants specialized to be slope and nectar secretion might scale with pollinated by birds and mammals must an exponent of 0.5 to 0.75 of pollinator produce sufficient rewards to pay the high body mass as we would predict from energetic costs of their pollinators. Thus, metabolic requirements (Fig. 4). there is a significant positive correlation There appear to be two primary con- between body size of pollinator and caloric sequences of large body size which make it advantageous for some plants to pay the expense of avian pollinators. First, large size confers high mobility which is impor- tant on two scales. In their daily foraging activity large animals are able to move greater distances to satisfy their higher energy demands, and consequently they have larger home ranges and territories.

t In birds, territory area (A,) scales to body mass (mb) according to the allometric equa- 09 tion A, * mb'- (Schoener, 1968). It is potentially beneficial for plants to use large pollinators because they increase the UJ — 0.01 proportion and distance of outcrossing. This should be particularly advantageous uj 0.001 to plants which are sparsely distributed or those in which seed dispersal is inadequate to achieve outbreeding. As might be ex- 0.0001 0.001 0.01 0 1 I 10 100 pected, bird pollinated species tend to be BODY MASS OF POLLINATOR (g) medium- to large-sized, long-lived peren- FIG. 5. Relationship between nectar secretion rate (X) nials which often are sparsely or patchily of (lowers and body mass (mb) of pollinators for a distributed. For example, Grant and Grant large number of flower species pollinated by birds (circles), mammals (triangles), and insects (crosses). (1968) list 127 species of plants from west- 035 The fitted regression equation, X = 0.77mb , prob- ern North America thought to be primar- ably is biased by the difficulties of measuring small ily hummingbird pollinated. Of these, only nectar secretion rates of flowers pollinated by small five are annuals, perhaps an equal number insects. We would expect more points in the lower left hand corner of the graph and a regression equation are biennials, and the remainder are per- with a higher exponent. Data compiled from Car- ennials; a large proportion of the peren- penter (19766), Feinsinger (1976), Gill and Wolf nials are woody shrubs. (1975), Hainsworth and Wolf (19726), Heinrich High mobility conferred by large size (19756), Hocking (1968), Kodric-Brown and Brown (1978), Percival (1965), Stiles (1975). and the authors' also enables avian nectarivores to migrate unpublished data; when necessary body weights were long distances to exploit erratically or sea- estimated from body length using empirically deter- sonally available nectar sources. Thus . mined allometric equations. many nectarivorous birds, including both CORRELATES AND CONSEQUENCES OF BODY SIZE 695 temperate and tropical hummingbirds Community patterns which reflect the (Grant and Grant, 1969; Gass et ai, 1976; effects of body size on competition among Kodric-Brown and Brown, 1978; Car- nectarivores and on plant-pollinator spec- penter, 1976a; Feinsinger, 1976; Fein- ificity are particularly evident in nectar- singer and Colwell. 1978), honeyeaters feeding birds. Many characteristics of (Keast, 1968), and sunbirds (Skead, 1967), bird-pollinated flowers are adaptations migrate in response to local and geo- either to avoid attracting insect com- graphic patterns of flower availability. This petitors (red color, lack of odor) or to ability to migrate makes nectanvorous prevent their taking nectar (long con- Downloaded from https://academic.oup.com/icb/article/18/4/687/2004942 by guest on 23 September 2021 birds reliable pollinators where conditions stricted floral tubes with tough, thickened suitable for plant reproduction occur only walls (see Faegri and van der Pijl, 1971; seasonally or unpredictably. Proctor and Yeo, 1972; Heinrich and Ra- A second consequence of large body size ven, 1972; Heinrich, 1975a). which makes birds particularly effective Where several species of nectar feeding and reliable pollinators in certain habitats birds coexist in the same habitat, they often is thermoregulatory capacity. The ability differ in body size. These species tend to of birds to maintain a constant body temp- subdivide the bird-pollinated plants on the erature makes their activity much more basis of size, thereby minimizing in- independent of environmental tempera- terspecific competition for nectar and ture and insolation than smaller hetero- achieving a degree of plant-pollinator thermic pollinators (Cruden, 1975; Brown specificity. Perhaps the clearest patterns et ai, 1978). This appears to account in are those shown by hummingbirds in the large part for the relatively high abun- West Indies. Lack (1973) noted that al- dance and diversity of nectar feeding though a few small, low elevation islands birds and bird-pollinated plants at inter- have but one hummingbird species, most mediate to high elevations on several con- islands have at least two (Fig. 6). When tinents (North and South America, Africa, only two species are present they usually and southern Asia). differ in wing length by a factor of at least 1.25 (which corresponds to 2-fold differ- ence in weight). Large, mountainous is- Competition among nectar feeders and evolution lands often have more than two species, of plant-pollinator specificity but those of similar size almost invariably Nectar is an energetically rich, easily occur at different elevations or occupy utilized food that is potentially attractive to different habitats. In western Puerto Rico many animals. Competition for nectar two large hummingbirds (Anthracothorax among animals which both do and do not dominicus and A. viridis) replace each other provide effective pollination is severe and altitudinally, and one small species of great importance in the coevolution of (Chlorostilbon maugaeus) occurs at all eleva- plants and nectarivores. If floral nectar is tions (Fig. 6). Hummingbird-pollinated equally available to animals of large and flowers likewise fall into two size small body size, then small nectar feeders categories. One set of flower species has often keep the standing crop of available corolla tubes 17-20mm long, produces nectar so low that large nectarivores can- l-2mg of sucrose per flower per day, and is not forage profitably and are competitively visited almost exclusively by the small excluded (Brown et ai, 1978). Plants are Chlorostilbon. Another group of flower strongly affected by such competition, and species has corollas 30-38mm long, se- those which utilize large pollinators often cretes 4-27mg of sucrose per flower per have evolved mechanisms to prevent con- day, and is pollinated almost exclusively by sumption of their nectar by small nectari- one of the two large Anthracothorax species. vores. As a result, communities of plants This specificity appears to be maintained and nectar feeders are organized in part because the small flowers produce just on the basis of specificity relative to pol- enough nectar to support small, but not linator body size. large, hummingbirds; the large flowers, 696 BROWN ETAL.

GREATER ANTILLES AND BAHAMAS their body sizes are adaptive. The body size of each species should be very close to the Cuba I I [ isle of Pines optimum for its particular ecology; this optimum should reflect the diverse effects | of size which combine to maximize repro- Hispamola 1 I [ Gonave 1 i ductive success. On the other hand, the narrow range of body sizes suggests that Jamaica i 1 11 constraints of being both a bird and a Puerto Rico | | nectar feeder place limits on the range of Downloaded from https://academic.oup.com/icb/article/18/4/687/2004942 by guest on 23 September 2021 Andros A Great flboco sizes which is adaptive. Reexamination of Grand Bahama the patterns discussed in the previous sec- Grand Iguana tions should provide insights into those Caicos and Turks processes which determine the upper and Elegthero lower limits of body size. LESSER ANTILLES

1 I II Minimum size Mart iique "Since smaller birds lose more heat rela- Dominica | II SI Lucto • e n tive to their body size than do larger birds, SI Vincent 8 there is a minimum size below which birds Grenada I • n are incapable of ingesting and metaboliz- SI Croi» Mone Galonte ing sufficient amounts of food to offset SI Kins | this heat loss." (Lanyon, 1963, p. 41) Such Monserrat I sg reasoning is common (e.g., D'Arcy Thomp- 30 10 50 60 70 80 son, 1961; Pearson, 1948, 1953; Greene- WING LENGTH (mm) wait, 1975), but usually it is based on l-'IG. 6. Pattern of body sizes in the hummingbird faunas of representative West Indian islands. Hum- qualitative arguments or untested and mingbird species are indicated by rectangles (shaded tenuous assumptions. It is almost impossi- to indicate altitudinal distribution) on a logarithmic ble to prove that any one factor prevents scale of wing length. Triangles indicate the elevation evolution of smaller size. In fact we suspect (overemphasized) and diameter of each island respec- that minimum size of nectar feeding birds tively. Note that most islands have at least two species of hummingbirds, and that those species which re fleets joint effects of several processes. If coexist at the same elevation usually have wing smaller size were advantageous except for lengths which differ by a factor of at least 1.25. Drawn the influence of one factor, then natural from data in Lack (1973). selection should act strongly to reduce the deleterious effect of that factor and which produce more nectar, have long thereby promote evolution of still smaller corollas which prevent small hirds from size. This indicates a potential hazard in harvesting nectar. In more diverse tropical extrapolating from allometric equations communities where several species of based on a wide range of sizes to make hummingbirds coexist, they often differ inferences about organisms of extreme not only in body size and bill length, but sizes. With this warning to accept our re- also in bill shape, foraging behavior, and sults with care, we shall proceed to make aggressiveness (e.g., Snow and Snow, 1972; such extrapolations. Feinsinger and Colwell, 1978). Allometric scaling of metabolic and other processes creates a number of poten- LIMITS OF SIZE tial problems for very small birds. We have illustrated some of these in Table 2 by The smallest nectar feeding bird weighs using available allometric equations to cal- 2g, the largest may weigh as much as 80g, culate predicted values of some physiologi- and the vast majority of species weigh from cal parameters for hummingbirds of vari- 3 to 30g. Since most appear to be surviving ous size, including the smallest existing and reproducing successfully, we assume species (2g) and a hypothetical bird weigh- CORRELATES AND CONSEQUENCES OF BODY SIZE 697 ing lg. These extrapolations suggest that a occurrence in northern Europe of 2g hummingbird evolving to half its size bumblebee species with substantially would have to contend with: 1) reduced longer proboscides than any North Ameri- endurance to fasting (by 16 and 30% at can species. The inference is that such thermoneutrality and at 5°C respectively); long-tongued bees have been able to evolve 2) a shorter lifespan (by 12%) in which to in Europe because of the absence of com- reproduce; 3) egg mass contributing a rela- peting hummingbirds. Brown, Kodric- tively larger (by 17%) burden for female Brown, Witham, and Bond (unpublished to produce and lift; 4) a relative decrease data) have documented competition be- Downloaded from https://academic.oup.com/icb/article/18/4/687/2004942 by guest on 23 September 2021 in surface of adult which can be applied to tween hummingbirds and insects (particu- heat the egg (37% vs. 30% decrease); and larly bumblebees) and suggested that this 5) a more rapid cooling rate (by 24%) of interaction plays a major role in the evolu- its eggs. tion of plant-pollinator specificity. Brown Small avian nectarivores not only must and Kodric-Brown also have evidence for solve these physiological problems, they similarly important competition between also must be able to obtain sufficient hummingbirds and hawkmoths. energy to meet their metabolic require- These studies suggest that under certain ments. They must evolve mutually advan- conditions either birds or large insects can tageous relationships with plants which competitively exclude the other from a produce nectar, and they must compete flower species, and that the outcome ulti- successfully with other nectar feeders, par- mately is determined largely by the plant, ticularly insects which have smaller body which can evolve characteristics to favor sizes and are subject to different physiolog- the competitor which provides the most ical constraints. We should expect particu- effective pollination. It is difficult to specify larly intense competition between the smal- the extent to which such competition may lest nectar-feed ing birds, hummingbirds, act to prevent evolution of even smaller and the largest insect pollinators, which birds than presently exist. However, since include bumblebees (genus Bombus) and the smallest birds and largest insects are hawkmoths (family Sphyngidae). These similar in size, energy requirements and animals are similar in size: many hum- capacity for pollination, competition may mingbirds weigh 2-4g (Greenewalt, 1962, be as important as physiological constraints cited in Carpenter, 1976a), bumblebee in limiting minimum size in nectar feeding queens often weigh 0.5g or more (B. Hein- birds. This should be a profitable subject rich, personal communication), and some for future research. hawkmoths weigh as much as 6g (Heinrich and Bartholomew, 1971). In addition, all have generally similar metabolic rates and Maximum size hence similar nectar requirements. All are We suggest that maximum body size in endothermic, so that they provide reliable nectar feeding birds (about 80g) is a con- pollination despite low or variable en- sequence of diminishing advantages to vironmental temperatures (Fig. 4; see also plants of producing sufficient nectar to Heinrich, 1975a). Finally, all have elongate attract large avian pollinators. The largest proboscides which they use to extract nec- nectar feeding birds include orioles, dron- tar from tubular flowers specialized to gos, babblers, and some honeyeaters. exclude other nectarivores. These are not only less specialized for Recent studies indicate that competition foraging from flowers than hummingbirds between large nectar feeding insects and and sunbirds, but they are also less depen- hummingbirds can be severe and may play dent on a continuous supply of nectar and a major role in plant-pollinator coevolu- are likely to supplement their diet more tion. Inouye (1976) notes that hum- with insects or fruit. These birds require mingbirds and bumblebees forage on large amounts of energy, and plants must many of the same flowers on his study secrete copious nectar to attract them. For area in Colorado. He calls attention to the only a few plants does the benefit of long 698 BROWN ETAL.

distance pollen movement appear to out- grosse. J. Ornithol. 111:38-47. Baker, H. G. 1961. The adaptation of flowering plants weigh the costs of attracting birds of this to nocturnal and crepuscular pollinators. Quart. size. These plants include some trees and Rev. Biol. 38:64-73. large herbs, particularly species which Bartholomew, G. A., T. R. Howell, and T. J. Cade. occur at low density or reproduce only 1957. Torpidity in the White-throated Swift, Anna once at the end of a long life. Hummingbird, and Poorwill. Condor 59:145-155. Bent, A. C. 1940. Life histories of North American It is of interest that mammalian pol- cuckoos, goatsuckers, hummingbirds, and their al- linators have approximately the same lies. U. S. Natl. Mus., Bull. 176. Downloaded from https://academic.oup.com/icb/article/18/4/687/2004942 by guest on 23 September 2021 upper limit to body size as nectarivorous Bernstein, M. H., S. P. Thomas, and K. Schmidt- birds and show similar patterns of spec- Nielsen. 1973. Power input during flight of the fish ificity. The largest mammalian pollinators, crow, Corvus ossifragus. J. Exp. Biol. 58:401-410. Calder, W. A. 1974. Consequences of body size for most of which are bats, are not highly avian energetics. In R. A. Paynter (ed.), Avian specialized and feed on insects or fruit energetics, pp. 86-151. Publ. Nuttall Ornithological when nectar is not available (Heithaus et Club, No. 15, Cambridge, Mass. ai, 1975). The plants that they pollinate Calder, W. A. and J. Booser. 1973. Nocturnal share many characteristics of those visited hypothermia of broad-tailed hummingbirds during incubation in nature with ecological correlations. by the largest avian nectarivores, including Science 180:751-753. large size, sparse distribution, and copious Calder, W. A. and J. R. King. 1974. Thermal and nectar secretion (Baker, 1961). caloric relations of birds. In D. S. Farner and J. R. King (eds.), Avian biology, Vol. 4, pp. 259-413. CONCLUDING REMARKS Academic Press, New York. Carpenter, F. L. 1974. Torpor in an Andean hum- The influence of body size on nectar- mingbird: Its ecological significance. Science feeding birds is so pervasive and important 183:545-547. Carpenter, F. L. 1976a. Ecology and evolution of an that it affects virtually all aspects of their Andean hummingbird, Oreotrochilvs estella. Univ. biology. We have presented a synthetic Calif. Publ. Zool. 106:1-75. overview of the diverse correlates and con- Carpenter, F. L. 19766. Plant-pollinator interactions sequences of body size from physiological in Hawaii: pollination energetics of Metrosideros and ecological perspectivies which em- collina (Myrataceae). Ecology 57:1125-1144. Carpenter, F. L. and R. E. MacMillen. 1976. Energe- phasize energetics. Body size influences tic cost of feeding territories in an Hawaiian hon- basic processes of metabolism, tempera- eycreeper. Oecologia 26:213-223. ture regulation, locomotion, and repro- Cheke, R. A. 1971. Temperature rhythms in African duction, which determine energy re- montane sunbirds. Ibis 113:500-506. Cruden, R. W. 1972. Pollination in high elevation quirements of free living birds. In foraging ecosystems: Relative effectiveness of birds and bees. to meet these requirements, birds play Science 1976: 1439-1440. important roles in ecological communities Dawson, W. R. and C. Carey. 1976. Seasonal acclima- as consumers of nectar and pollinators of tion to temperature in cardueline finches. J. Comp. plants. In the last decade, great progress Physiol. 112:317-337. Drent, R. H. 1975. Incubation. In D. S. Farner and J. has been made in analyzing these relation- R. King (eds.), Avian biology, Vol. 5, pp. 333-420. ships, but there is much opportunity for Academic Press, New York. further work. The correlates of body size Epting, R. J. and 1. M. Casey. 1973. Power output at all levels of ecological organization and wing disc loading in hovering hummingbirds. suggest that it will be particularly fruitful Amer. Nat. 107:761-765. Faegri, K. and L. van der Pijl. 1971. The principles of to use these patterns and the extensive pollination biology. Pergamon, Oxford. information on energetics of individual Feinsinger, P. 1976. Organization of a tropical guild birds as a basis for investigating aspects of of nectarivorous birds. Ecol. Monogr. 46:257-291. community ecology and plant-pollinator Feinsinger, P., and R. K. Colwell. 1978 Community coevolution which remain poorly docu- organization among Neotropical nectar-feeding birds. Amer. Zool. 18:779-795. mented and understood. Feinsinger, P., R. K. Colwell, J. Terborgh, and S. B. Chaplin. 1978. Elevation and the morphology, REFERENCES flight energetics and foraging ecology of tropical hummingbirds. Amer. Nat. (In press) Aschoff, J. and H. Pohl. 1970. Der Ruheumsatz von Fisk, L. H., and D. A. Stein. 1976. Additional explor- Votjcln als Funktion der Tafjeszeit und der Korper- ers of nectar. Condor 78:269-271. CORRELATES AND CONSFQUENCES OF BODY SIZE 699

Ford, H. A. and D. C. Patton. 1976. The value of Colorado Rocky Mountains. Ph.D. Diss. Univ. insects and nectar to honeyeaters. Emu 76:83-84. North Carolina, Chapel Hill. Gass, C. L., G. Angehr, and J. Centa. 1976. Regula- Keast, A. 1968. Seasonal movements in the Australian tion of food supply by feeding territoriality in the honeyeaters (Meliphagidae) and their ecological Rufous hummingbird. Canad. J. Zool. 54:2046- significance. Emu 68:159-209. 2054. Kendeigh, S. C. 1969. Tolerance of cold and Gill, F. B. 1978. Proximate costs of competition for Bergmann's Rule. Auk 86:13-25. nectar. Amer. Zool. 18:753-763. Kodric-Brown, A. and J. H. Brown. 1978. Influence Gill, F. B. and L. L. Wolf. 1975. Economics of feeding of economics, interspecific competition, and sexual

territoriality in the golden-winged sunbird. Ecology dimorphism on territoriality of migrant Rufous Downloaded from https://academic.oup.com/icb/article/18/4/687/2004942 by guest on 23 September 2021 56:333-345. Hummingbirds. Ecology 59:285-296. Grant, K. A. and V. Grant. 1968. Hummingbirds and King, J. R. 1973. Energetics of reproduction in birds. their flowers.Columbi a Univ. Press, New York. In D. S. Farner (ed.), Breeding biology of birds, pp. Greenewalt, C. H. 1962. Dimensional relationships 78-107. Natl. Acad. Sci., Washington, D.C. for flying animals. Smithsonian Misc. Coll. 144(2): Kleiber, M. 1961. The fire of life. Wiley, New York. 1-46. Lack, D. 1973. The numbers of species of hum- Greenewalt, C. H. 1975. The nignt of birds. Trans. mingbirds in the West Indies. Evolution 27:326- Amer. Philos. Soc. 65(4): 1-67. 337. Hainsworth, F. R., B. G. Collins, and L. L. Wolf. 1977. Lanyon, W. E. 1963. Biology of birds. Natural History The function of torpor in hummingbirds. Physiol. Press, Garden City, New York. Zool. 50:215-222. Lasiewski, R. C. 1962. The energetics of hum- Hainsworth, F. R. and L. L. Wolf. 1970. Regulation of mingbirds. Condor 64:324. oxygen consumption and body temperature during Lasiewski, R. C. 1963. Oxygen consumption of tor- torpor in a hummingbird, Eulampis jugularis. Sci- pid, resting, active, and flying hummingbirds. ence 168:368-369. Physiol. Zool. 36:122-140. Hainsworth, F. R. and L. L. Wolf. 1972a. Power for Lasiewski, R. C. 1964. Body temperature, heart and hovering flight in relation to body size in hum- breathing rate, and evaporative water loss in hum- mingbirds. Amer. Nat. 106:589-596. mingbirds. Physiol. Zool. 37:212-223. Hainsworth, F. R. and L. L. Wolf. 19726. Crop Lasiewski, R. C. and W. R. Dawson. 1967. A re- volume, nectar concentration, and hummingbird examination of the relation between standard energetics. Comp. Biochem. Physiol. 42A:359-366. metabolic rate and body weight in birds. Condor Hainsworth, F. R. and L. L. Wolf. 1972c. Energetics 69:13-23. of nectar extraction in a small, high altitude, tropi- Lasiewski, R. C. and R. J. Lasiewski. 1967. Physiologi- cal hummingbird. Selasphorus flammula.J . Comp. cal responses of Blue-throated and Rivoli's Hum- Physiol. 80:377-387. mingbirds. Auk 84:34-48. Hainsworth, F. R. and L. L. Wolt. 19/5. Wing disc Lindstedt, S. L. and W. A. Calder. 1976. Body size loading: Implications and importance for hum- and longevity in birds. Condor 78:91-94. mingbird energetics. Amer. Nat. 109:229-233. MacMillen, R. E. and F. L. Carpenter. 1977. Daily Heinrich, B. 1975a. Energetics of pollination. Arm. energy costs and body weight in nectarivorous Rev. Ecol. Syst. 6:139-170. birds. Comp. Biochem. Physiol. 56A:439-441. Heinrich, B. 19756. Bee flowers: A hypothesis on McNab, B. K. 1971. On the ecological significance of flower variety and blooming times. Evolution Bergmann's rule. Ecology 52:843-854. 29:325-334. Odum, E. P., C. E. Connell, and H. L. Stoddard. Heinrich, B. and P. H. Raven. 1972. Energetics and 1961. Flight energy and some estimated flight pollination ecology. Science 176:597-602. ranges for some migratory birds. Auk 78:515-527. Heinrich. B. and G. A. Bartholomew. 1971. An Pearson, O. P. 1948. Metabolism of small mammals analysis of pre-flight warm-up in the sphinx moth, with remarks on the lower limit of mammalian size. Manduca sexta. J. Exp. Biol. 55:223-239. Science 108:44-45. Heithaus, E. R., T. H. Fleming, and P. A. Opler. Pearson, O. P. 1950. The metabolism of hum- 1975. Foraging patterns and resource utilization in mingbirds. Condor 52:145-152. seven species of bats in a seasonal tropical forest. Pearson, O. P. 1953. Use of caves by hummingbirds Ecology 56:841-854. and other species at high altitudes in Peru. Condor Hocking, B. 1968. Insect-flower associations in the 55:17-20. high Arctic with special reference to nectar. Oikos Percival, M. S. 1965. Floral biology. Pergamon Press, 19:359-387. Oxford. Hoyt, D. F. 1977. The effect of shape on the surface- Proctor, M. C. F. and P. F. Yeo. 1972. The pollination of volume relationships of birds' eggs. Condor flowers. Collins, London. 78:343-349. Rahn, H., C. V. Paganelli, and A. Ar. 1975. Relation Hull, D. 1973. Thermoregulation in young mammals. of avian egg weighi to body weight. Auk 92:750- In G. C. Whitlow (ed.). Comparative physiology of 765. thermoregulation, Vol. 3, pp. 169-191. Academic Rahn, H. and A. Ar. 1974. The avian egg: Incubation Press, New York. time and water loss. Condor 76:147-152. Inouye, D. W. 1976. Resource partitioning and com- Romanoff, A. L. and A. J. Romanoff. 1949. The avian munity structure: A study of bumblebees in the egg. Wiley, New York. Downloaded from https://academic.oup.com/icb/article/18/4/687/2004942 by guest on 23 September 2021

700 BROWN ETAL.

Schoener, T. W. 1968. Sizes of feeding territories animals. Comp. Biochem. Physiol. 34:841-846. among birds. Ecology 49:123-141. Van Tyne, J. and A. J. Berger. 1976. Fundamentals of ' Skead, C.J.I 967. The sunbirds of southern Africa. A. A. ornithology. Wiley, New York. Balkema, Cape Town. Waser, N. M. and D. Inouye. 1977. Implications of Skutch, A. F. 1962. The constancy of incubation. recaptures of broad-tailed hummingbirds banded Wilson Bull. 74:115-152. in Colorado. Auk 94:393-395. Snow, B. K. and D. W. Snow. 1972. Feeding niches of Weis-Fogh, T. 1972. Energetics of hovering flight in hummingbirds in a Trinidad valley. J. Anim. Ecol. hummingbirds and Drosophila. J. Exp. Biol. 56:79- 41:471-485. 104. Stiles, F. G. 1975. Ecology, flowering phenology, and Wolf, L. L. and F. R. Hainsworth. 1971. Time and hummingbird pollination of some Costa Rican energy budgets of territorial hummingbirds. Ecol- Helkonia species. Ecology 56:285-301. ogy 52:980-988. Thompson, D. W. 1961. On growth and form. Wolf, L. L. and F. R. Hainsworth. 1972. Environmen- Abridged, J. T. Bonner (ed.), Cambridge Univ. tal influence on regulated body temperature in Press, New York. torpid hummingbirds. Comp. Biochem. Physiol. Tucker, V. A. 1970. Energetic cost of locomotion in 41A:167-173.