“ ‘, .Is 3»! iit!rlllE wagons 1J‘5’ahb3‘x. gs-r..~'~1‘::'.u.'.c Va.- ".77.:"(afLZTDM | q I V. 9 5-1. .0“ '7”. M '6 {,3 ‘-. T5 9 ‘3? 61:. .. ' _ . . r ‘ . V \ U 4 f. 9 J ‘ This is to certify that the thesis entitled THE EFFECT OF AGGRESSION 0N BREEDING TERRITORY SIZE IN THE SAVANNAH SPARROW PASSERCULUS SANDWICHENSIS presentedby CHRISTOPHER MARC ROGERS has been accepted towards fulfillment of the requirements for M . S . degree in ZOOLOGY (gm/{£32m Major professor Date_§.?_(D W’VZ //°L2_— 0-7639 MS U is an Affirmative Action/Equal Opportunity Institution «III ll!llllllllllllllllllllllll L 1293 00658 9687 MSU LIBRARIES -—_. RETURNING MATERIAL§: Place in book drop to remove this checkout from your record. FINES will be charged if book is returned after the date stamped below. '— THE EFFECT OF AGGRESSION 0N BREEDING TERRITORY SIZE IN THE SAVANNAH SPARROW PASSERCULUS SANDWICHENSIS By Christopher Marc Rogers A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE Department of Zoology 1982 L? I! r 4 ABSTRACT THE EFFECT OF AGGRESSION ON BREEDING TERRITORY SIZE IN THE SAVANNAH SPARROW PASSERCULUS SANDWICHENSIS BY Christopher Marc Rogers Following a review of the literature on avian territoriality, two important correlates of breeding territory size are recognized: prey density and intrusion pressure. It is concluded that a third correlate is possible in passerine breeding systems in which a relatively low pop- ulation density may preclude significant rates of intrusion: male aggre- sion. In this study, the hypothesis that aggression and breeding terri- tory size are positively related in Savannah Sparrows is tested by grading the aggressive response of males to a novel stimulus. Breeding territory size is statistically significantly but moderately correlated with aggression scores at the population level. However, at the level of contiguous neighbors, males with high aggression scores defended star tistically significantly larger territories than their less aggressive neighbors. These patterns parallel those observed in two other terri- torial species. In the study population, competition between males for breeding space is not population-wide but instead occurs between neigh- boring territorial male sparrows. TABLE OF CONTENTS LIST OF TABLES . . . . . . . . . . . . . . . . . . . . . . . LIST OF FIGURES . . . . . . . . . . . . . . . . . . . INTRODIICTION O O O O O O O O O O 0 O O O O O O O O O O O O O The evolution of behavior . . . . . . . . . . . . . . . Historical perspective of territoriality . . . . . . . Quantification of the adaptive functions of territory Recent hypotheses . . . . . . . . . . . . . . . . . . . . MATERIALS AND METHODS . . . . . RESULTS 0 O O O O O O O O O O O O O O O O O O O O O O O O The sparrow population . . . . . . ... . . . . . . Nature of territories . . . . . . . . Field measurement of aggression . . . . . . . . . . . . . Behavioral assessment of the aggression test . Relationship of territory size to aggression . Morphological measurement . . . . . . . . . . . . . DISCUSSION . . . . . . . . . . . . . . . . . . . . . . . . . . Nature of territories . . . . . . . . . . . . . . . Field measurement of aggression . . . . . . . . . . Behavioral assessment of the aggression test . . . . . . Relationship of aggression and territory size at the popu- lation level . . . . . . . . . . . . . . . . . . Relationship of aggression and territory size at the level of neighbors . . . . . . . . . . . . . . . . . . . . . . . 'Predictions outlining future work . . . . . . . . . . . . LIST OF REFERENCES . . . . . . . . . . . . . . . . . ii iii iv O‘UINH l6 16 20 20 25 26 31 36 36 37 39 41 44 46 48 LIST OF TABLES Table 1 Rating scale used to quantify aggressive responses of Savannah Sparrows to the aggression test . . . . . . . 2 Breeding chronology of the Savannah Sparrow study population in 1981 O O O O O O O O O O O I O O O O O O I O O I O I O O O 3 Initial capture dates and AT score on two ATS given to male sparrows during the first field exam . . . . . . . . . . . . 4 AT scores for 9 territorial males captured twice . 5 Statistical analysis of variation in AT scores within and between field exams of male sparrows . . . . . . . . . 6 Dominance hierarchy of 7 territorial male sparrows in the aViary I O O O O 0 O O O O O O 0 O- O O O O O O 0 O O 0 O O 7 Values of TA, AT, and SA for each territory (n = 18) . . . . 8 One-way analysis of covariance of TA (X; n = 18) for 3 levels of AT (SA is a linear covariate = Y) . . . . . . . . . . . . 9 Duncan's New Multiple Range Test of TA for the 3 levels of AT 10 Mean : SE and statistical analysis of morphological variables of 18 territorial male and 13 female sparrows . . . . . . . . ll Differences in territory size (m2) between dominant-subordin- ate pairs at the population and contiguous-territory levels of competition . . . . . . . . . . . iii 14 18 19 23 24 27 28 32 33 34 35 LIST OF FIGURES Figure 1 Map of the study site occupied by the breeding population . 2 Map of the 18 territories comprising the study population . 3 The relationship between territory area (TA) and aggression (AT) . . . . . . . 4 The relationship between territory area (TA) and summed ag- gression (SA) . . . . . iv 11 22 29 3O INTRODUCTION The evolution of behavior Like morphologists and physiologists, behavioral ecologists have chosen the problem of unambiguously accounting for specific attributes in animal populations. Beginning with Darwin (1859) an effective analytical approach has been to hypothesize about the adaptive value of species-specific behavior relative to environmental factors (e.g. temperature, competition) with which the behaviors are thought to interact. For example, behavioral thermoregulation in lizards is "adaptive" behavior in that it assures general physiological homeostasis in a near-optimal manner and therefore increases the probability of survival. Assuming an amount of variation in a critical behavioral trait of a species, a result of species-wide interaction with the environment may be increased survival and subsequent reproductive success of particular phenotypes. To continue with the above example, lizards of a certain phenotype in a given population comprised of varying phenotypes may tend to remain in direct insolation too long and thereby become incapable of reproduction. Such nonrandom directional change produced in the genetic constitution of a species is termed evolution. This dynamic process is effected by the elimination from breeding of individuals that interact maladaptively with critical environmental factors (natural selection) and the resultant reproductive success of survivors. The message is that adaptive behavior can evolve within a species in response to a selective environment (Brown 1975, Alcock 1979). Just as morphological adaptation reflects natural selection (Darwin 1859), so does behavioral adaptation, though in an anthrOpomorphic sense, less explicitly. Examples of evolved, adaptive behavior in animals include habitat selection (Fretwell 1969, 1974) foraging strategies (Orians 1971, Holmes 25 El. 1977), time-energy utilization strategies (Schoener 1971, King 1979), and mating systems (Orians 1969, Wolf and Stiles 1970). Territoriality has long been considered adaptive behavior by ecologists (Howard 1920, FOX.EEH§£' 1981) since it encompasses all vertebrate classes and many invertebrate groups (Carpenter 1958). Territoriality is conspicuously developed in many forms ranging from large exclusive all-purpose to diminutive nesting territories (Nice 1941a). It is certain that territorial behavior, present in many divergent taxa, has evolved in numerous ecological situations (Brown 1964). Yet, the overall "currency" of evolutionary fitness gained directly from territorial defense (i.e. quantifiable adaptive value in units) is unclear. This is understandable when one considers the arbitrary nature of most proposed measures of fitness, e.g. calories gained or young produced per unit time. Thus, the adaptive function of territorial behavior, given it exists, has long been the subject of sometimes acrimonious debate among evolutionary biologists. This debate has its beginnings in early concepts of territoriality and continues to the present time. Historical perspective of territoriality An understanding of territorial behavior in animals began with simple demographic observations made by Aristotle in 350 B.C. (Welty 1974). The idea of exclusive breeding areas defended against conspecifics later appeared in European scientific literature through Olvina in 1622 (Nice 1941a). These first concepts of avian territoriality were derived from conspicuous avian behavior, and later gave rise to more comprehensive theories of animal territoriality. In his book, Der Vogel und Sein Leben (1868) Bernard Altum originally and clearly presented a unified theory of avian territoriality (Stokes 1974). Altum's work soon became widely accepted though restricted by geographical isolation to EurOpe. His notable ideas were 1) the importance of intermale fighting facilitating spacing of nesting pairs 2) avoidance of intraspecific competition and subsequent food shortage 3) the fixation of territory ownership and boundaries by male song. Most important to future thinking, Altum stressed that the pattern of resource dispersion necessarily determines whether or not resources will be defended, and that the most aggressive individuals will do so when feasible. In 1903, the Irish naturalist C.B. Moffat published an independently derived scheme of avian territoriality that closely resemble Altum's synthesis. Keenly aware of Darwin's work, his major contribution was placing territorial behavior in an evolutionary context, accountable for by natural selection. To Moffat, evidence of excess nonbreeding birds in actively breeding populations suggested that territorial behavior sets an upper limit on birth rate producing fitness differences between individuals. Howard (1920), often called the "father" of the territory concept, presented a final definitive theory before rigorous description and classification of avian territories began. The accessibility of his work to an international scientific audience, and not its originality, made Howard the paramount figure in early territory theory. Howard defended the hypothesis that territoriality spaces individual breeding pairs thereby facilitating efficient resource utilization. He also resolved that space, not a mate, is defended, insuring "successful discharge of the sexual function”, an hypothesis later supported by Nice (1937, 1941a). By this reasoning, territoriality guarantees the close proximity of the sexes, increasing the probability of reproduction. Such a mechanism may be important in highly vagile organisms such as birds. To the contrary, territoriality is highly developed in far less vagile groups, e.g. fish and reptiles, where the probability of sexual encounters is high in the absence of territorial defense. Definition and classification of avian territory followed development of early theories. Noble (1939) provided the most ubiquitous definition of territory to date, calling it "any defended area". Relying on this definition, Nice (1941a) placed avian territories into seven distinct categories (Types A - H) ranging from all-purpose (Type A - mating, nesting, and feeding) to winter territories (Type H). Defense of space via aggression or threat display was determined a factor common to all categories. Systematic classification of avian territories revealed the correlation between diversity of adaptive function and diversity of type. In his extensive review, Hinde (1956) stressed this point by citing an extensive array of adaptive functions associated with many different types of territories. Site familiarity, disease buffer, buffer against predation (plus others) were recognized, only to be later rejected as unimportant (Verner 1977). This pre-quantitative period in the ontogeny of territorial theory ended when Carpenter (1958) concluded existing territorial theory to be "limited, unsystematic, and qualitative ... in the experimental stage". Quantification of adaptive functions of territory Statistical correlation of ecological and behavioral factors with territory size has proven fruitful in clarifying adaptive functions of territorial behavior. Stenger (1956) showed that forest-floor invertebrate density and territory size in the Ovenbird (Seirus aurocapillus) were inversely related, supporting the food hypothesis of Howard (1920). Further studies revealed a similar relationship for many unrelated species in unalike environments during both breeding and nonbreeding periods (Pitelka st 31. 1955, Orians 1961, Wolf 1969, Holmes 1970, Cody and Cody 1972, Gill 1975, Gass 1976, Lyon 1976, Kodric-Brown and Brown 1978, Gass SE 31. 1979). Animal populations adjust their density according to limiting environmental conditions such as food availability (Lack 1954, Wynne-Edwards 1962). Obvious adjustments of mean breeding territory size within and between years (Holmes 1970) coupled with existence of "floating" populations comprised of socially subordinate individuals (Garrick 1963, Smith 1979) indicate that territorial behavior functions as a population regulator (Huxley 1934, Stewart and Aldrich 1951, Tinbergen and Kluyver 1953, Tinbergen 1957, Glas 1960, Orians 1961, Tompa 1962, Garrick 1963, Brown 1964, 1969, Watson and Miller 1971, Verner 1977). Unfortunately, only a few studies (Wolf 1969, Holmes 35 '31. 1970) have examined resource and population dynamics simulta- neously. Therefore the conclusion that population regulation is achieved by adjusting breeding territory size relative to proximate food supply is tenuous at best. A more general approach to addressing the evolutionary role of territorial behavior is to consider territoriality as a general spacing system influencing distribution, abundance, and hence evolution of organisms (Brown and Orians 1970). To this end recent studies have challenged the specific food hypothesis of Howard in favor of more encompassing behavioral hypotheses. Recent hypotheses Currently, (Myers EEHEl' 1979, Ewaldlsgflal. 1980) avian territory size is believed to be proximally controlled by intrusion rate of conspecifics attempting to initiate or expand an exclusive territory at the expense of neighboring males. To illustrate, in their study of the Sanderling (Calidris alba) Myers E£“3£' obtained a statistically significant positive simple correlation between density of intertidal invertebrates (Excirolana) and size of feeding territory. However, using partial correlation analysis, intrusion “pressure" (rate of intrusion of conspecifics per meter of territorial perimeter) was found to account for nearly all variation in territory size. Prey density thus became independent of territory size, originally correlating with territory size only through spurious correlation with intruder pressure. This spurious correlation resulted because areas of high prey density attracted many foraging Sanderlings. Simlarly, Ewald e£_ ‘31. (1980) showed convincingly that nesting territory size in colonial western Gulls (Larus occidentalis) is causally related to intrusion pressure. Two major criticisms of the above studies of intrusion pressure exist. Firstly, no consideration is made of the effect of individual variation in aggressiveness on territory size. Given the well documented effects of age (Watson 1971, Ketterson 1979) and general health (Watson and Moss 1972) on defensive ability in birds, it is highly unlikely that all members of a territorial population will respond homogeneously to any level of conspecific intrusion pressure. Indeed, most breeding populations do consist of territorial individuals of different ages (Nice 1937, Robel 1966, Watson 1971, Ballard and Robel 1974, Greenwood 35 El. 1979). Nice (1937) qualitatively noted that older more aggressive Song Sparrows (Melospiza melodia) defended larger territories than younger, less aggressive birds in the same breeding population. Watson and Miller (1971) showed that variation in male aggression during territorial defense is positively correlated with territory size in breeding Red Grouse (Lagopus lagopus). A second criticism of the studies of intrusion pressure concerns scope of conclusions. It is conceded (Myers £5.3if 1979) that in territorial systems illustrating inflexible spacing behavior (e.g. exclusive breeding territories) when tenure of territorial individuals is long, intrusion pressure is not likely to have a dramatic effect on territory size compared to systems exhibiting flexible spacing behavior (e.g. Sanderling winter territories). The rigorous hyperdispersion (Brown and Orians 1970) of territorial males in breeding systems precludes convergence of foraging birds on areas of high prey density, a phenomenon observed in wintering Sanderlings. Hyperdispersion, a function of aggression, largely removes the social potential for intrusion pressure to determine territory size. Breeding territories show extensive variation in size (Nice 1937, Hinde 1956, Tompa 1962, Weeden 1965, Stefonski 1967, Ewald Efinii‘ 1980 plus many others). Proximate determinants of territory size are largely unknown (Myers 35 El. 1979) but apparently act as determinants of reproductive fitness (Kluyver and Tinbergen 1953, Glas 1960, Robel 1966, Orians 1966, Zimmerman 1966) since owners of small territories in breeding populations often experience low reproductive output relative to owners of large territories (e.g. Kluyver and Tingergen 1953, Zimmerman 1966). That hyperdispersion of territorial males is causally related to male aggression seems indisputable at this point (Howard 1920, Brown and Orians 1970). It is therefore possible that variation in territory size may be causally related to variation in inherent aggressiveness of territorial males. This hypothesis was tested with a passerine breeding system in the present study. The effect of individual variation in aggression on avian breeding territory size was examined in a population of Savannah Sparrows (Passercerlus sandwichensis: Fringillidae), a locally common species in southern Michigan. Comprised of at least sixteen subspecies,;3; sandwichensis ranges over the entire North American continent excluding the southeastern United States. It nests in fallow fields, sedge bogs, short-grass prairies, and coastal salt marshes (Bent 1968). The Eastern Savannah Sparrow (323' savannah) is the subject of the present study. It ranges from southern Ontario, east to Nova Scotia, south to central Ohio, and west to northern New Jersey (Bent 223 ££E.). Following spring time arrival of single individuals in central Michigan (average arrival: third week of April, McWhirter and Beaver 1977) each usually monogamous pair feeds and rests in an exclusive Type A territory defended by the male bird. Average fall departure from central Michigan is the second week of October. During the spring and summer periods the diet consists of 63 and 26% plant matter (grass seeds), respectively (Bent 223 gig.). Insects form an important dietary component during the nestling and fledgling phases of the breeding cycle. The specific features of homogeneous breeding habitat, quantifiable resource base, and long-term rigorous defense of an exclusive all-purpose breeding territory make 2, sandwichensis an ideal species for testing hypotheses regarding adaptive functions of territorial behavior. MATERIALS AND METHODS The study population occupies a fenced short-grass area in the Water Quality Management Aquatic Site (WQMAS) on the south campus of Michigan State University (MSU). The southern portion of the MSU campus is an extensive agricultural area where Savannah Sparrows commonly breed in suitable habitats (fallow fields). The dominant grass on the study site is Digitari 32. growing to a seasonal maximum of 45-60 cm in August. Red clover (Trifolium pratense) grows to the same general height in dense numerous patches scattered throughout the area. Two species of wild mustard (Cruciferae) generally occur sparsely but densely along the shoreline of artificial ponds created within the study area (Figure l). The sparrow population has been the subject of previous study of avian territoriality conducted by Donald Figure 1. Map of the study site occupied by the breeding population. The surrounding line is a wire fence. Numbered areas are artificial ponds; a road enters the site from its east side. 10 -\F_ Figure l. L. Beaver (DLB), Dept. of Zoology, Michigan State University. Consequently part of the population was color-banded in 1980 providing one year of data on territorial behavior. No numerical manipulation of the papulation (e.g. removal or addition of individuals) has occurred during past research. The study population is considered to be naturally fluctuating in a preferred environment, utilizing a naturally varying resource base. Field observations of territoriality began on 30 March and continued through 19 June 1981. During the period of intense territorial behavior (before the fledgling of young, 1 April to mid June) male sparrows were captured on their territories by luring them into a mist net erected near a mounted male Savannah Sparrow model. Past experience has shown that most territorial males will attack similar models and are easily captured. A tape recorder hidden near the model played conspecific songs, providing a second attractant for territorial males. Habituation was avoided by curtailing model presentation after 2-3 minutes of exposure to an unresponsive bird. Rarely females were captured by this method. Several males that consistently detected the mist net were captured with nets set in the dim light of dusk or early morning. Most captured females were caught with this method. To allow individual recognition all banded birds received a unique combination of colored plastic leg bands. A11 color-banded birds were given a two or three digit color code based on the following scheme: A-aluminum U.S. Fish and Wildlife Serivce band, B-black, G-dark green, L-light blue, LG-light green, M-mauve, O—orange, P-purple, Rrred, W-white, Y-yellow. From here on, territories are labelled according to the color code of the male owner. Inherent aggression was estimated in each male and female sparrow. Immediately upon capture, the aggression test (AT) developed by Burt and Giltz (1969) was administered twice to each bird, once before and after measurement of morphological variables. A11 examinations took place off the territory in an automobile. Morphological measurement typically required between S and 10 minutes. Within a single examination two ATS were administered to each bird. The mean of these two tests was used in all subsequent analyses of AT relative to territory area (TA). This test provides quantification of inherent aggression that is site-independent, i.e. distinct from defense of a familiar area (territory). At time zero, a bird restrained by the closed left hand was freed from restraint and held only by the right hand, dorsal surface toward the palm, with the thumb and first finger encircling the folded wings. The upper one-third of the body including the head, neck, and shoulder region was exposed. The bird was observed unthreatened in this position for 10 seconds. For three subsequent lO-second intervals threatening was accomplished by lightly tapping (appprox. 2 seconds/tap) the bill and chest region using the left middle finger, index finger, and thumb in that order. Total AT time was 40 seconds. The rating scale used to quantify aggressive responses to threat ranges from one to 10 and is presented in Table 1. Male birds trapped a second time up to 2 months into the breeding season (n=9) were retested without knowledge of the first AT score. The avian behavioral parameter quantified by the AT score is probably an inherent tendency to aggress, at least in response to a novel stimulus (Burtt and Giltz 1969). Verification of whether or not 13 Table 1. Rating scale used to quantify aggressive responses of Savannah Sparrows to the agression test (from Burtt and Giltz 1969). Response to stimulus AT Score Spontaneously begins biting rapidly and vigorously the moment the hand is Opened and continues for 40 seconds. 10 Spontaneously bites rapidly and vigorously, but does not begin immediately when the hand is opened. _ 9 Spontaneous, but less vigorous biting continuously through most of the initial period. 8 A few spontaneous bites followed by rapid and vigorous biting when threatened. 7 Quiet during the initial 10 seconds, but rapid and continuous biting when threatened. 6 Several vigorous bites at each threat, but not a continuous attack. 5 Some biting at each threat, but not vigorous. 4 No biting until the last threat, but several bites there- after. 3 A single bite on the last threat. 2 No biting whatever; completely passive. 1 14 the AT score measures overall inherent aggression in the sparrows was attempted by placing I previously tested territorial sparrows representing a broad range of TA (1357-3391 m2) into a 1.8 x 2.7 x 27m aviary and quantitfying the resultant dominance hierarchy. Three males trapped in modelless mist nets on 15 June and four on 16 June were placed in the aviary at 1300 on 18 June following a day in a second aviary with food and water provided ad libitum. Close-range observations from behind a screen took place from 1630 to 2030 on 18 June and 0950 to 1050, 1130-1230 on 19 June for a total of 6.0 hours. Pairwise aggressive interactions were induced via competition for food at a Spatially limited food source (a 45x4 cm tray of continuously available seed mix). Water was provided ad_libitum in a large bowl immediately adjacent to the food tray. Interactions were recorded with respect to identity of the interacting pair and type of interaction, i.e. displacement from the food tray with or without direct attack. Retreating birds were judged subordinate to the displacing member of the interacting pair. The percentage of total interactions a bird won was correlated with its mean of two AT scores obtained 1) during the first field exam (correlation l) 2) immediately before placement into the aviary (correlation 2). All captive birds were released directly back onto their territories on 19 June. Morphological data recorded were 1) subcutaneous furcular and abdominal fat using a semi-quantitative index (see WOlfson 1945 and weise 1956b), 2) total body weight (nearest 0.1 g) on a Pesola spring scale, and 3) lengths of protruding cloaca (males only; nearest mm), bill (nearest 0.1 mm), unflattened wing chord and tail (nearest mm), 15 and tarsus (nearest 0.01 mm on a micrometer caliper). Qualitative data on contour and flight feather molt and development of the incubation patch in females were recorded along with capture date, time, method, and time of release. Territorial behavior in Savannah Sparrows is well-documented and considered stereotypic (Potter 1972, Welsh 1975). Particular displays serve to mark permanent, season-long territorial boundaries. Between 15 April and 6 June boundaries of 20 breeding territories were determined by observing locations of "parallel walk" displays. Using this behavior, male sparrows determine exact boundaries by walking slowly side-by-side along a mutually decided line (Potter 1972). Overt fighting between two territorial males at a common border gave a second specific determinant of territorial limits aside from parallel walking. Locations of song perches lent a general indication of territory ownership, especially when a male sang continuously from an area of nebulous ownership. Singing rarely occurred directly on a boundary, presumably due to the paucity of adequate song perches. All territory boundaries were marked with upright blue and/or yellow flags 24" high. Flags were occasionally used as border song perches, furnishing a "check" on territory ownership. Following final boundary establishment (6 June) each territory was divided into a series of triangles whose perimeter and relative location were measured with a transit and compass. Geometric data were transferred to graph paper in the field and territory area in or2 calculated by summing the composite triangular areas. RESULTS The sparrow population 16 A general breeding chronology of the sparrow population is presented in Table 2. Inspection of Table 2 shows that two simple but critical assumptions were met: the population was breeding, and individual males defended exclusive breeding territories. The first migrant males arrived on the site around 30 March. Arrival of males from wintering grounds (southern Indiana south to Florida and east to the U.S. coastline) continued through the third week of April. All males to subsequently occupy territories were present but not necessarily exclusively territorial by 15 April; 7 males banded in 1980 returned to defend territories in 1981 (B,R/Y,Y/O,R,O,R/O,R/W). Females arrived much later than males, arriving on the site between 10 and 21 April. Territorial skirmishes between males began the first week of April as population density increased. Territorial skirmishes gradually became stereotyped defense of particular territorial borders at equilibrium population density (5.39 breeding pairs/lOOha). Between 4 April and 10 June, 40 sparrows were captured and examined (27 males, 13 females). Birds retrapped during this period totalled 22 (17 males, 5 females); a total of 62 captures was made. Retrapped birds were not always tested for aggression since field conditions often precluded efficient testing. Initial capture dates and times of males are presented in Table 3. Of the 27 captured males, 20 were caught between 0630 and 1030, 1 between 1030 and 1400 (1130), and 6 between 1430 and 2000 (Table 3). Only eighteen of these 27 birds were used in analysis of aggression and territory size, since the remainder disappeared from the study site, presumably occupying surrounding breeding areas. The exact status of these latter birds is 17 Table 2. Breeding chronology of the Savannah Sparrow study population in 1981. Breeding behavior Period (1981 dates) Spring arrival males females Territorial defense Nesting nest building nestlings present last week of March - 21 April last week of March - 15 April 10 - 21 April April through July late April - early May first detected on 12 May Table 3. Initial capture dates and AT score on two ATs given to male sparrows during the first field exam (n=27). Bird Capture First Second Change Average date (time) Score(l) Score(2) [(2) - (1)] [(1) + (2)/2] Y/P 4-15 (0915) 6.0 5.0 -l.0 5.5 R/O 4-16 (0730) 1.0 1.0 0.0 1.0 LG/R 4-16 (0820) 5.0 5.0 0.0 5.0 L/G 4-16 (0830) 2.0 1.0 -l.0 1.5 0/G 4-16 (0845) 1.0 1.0 0.0 1.0 R/Y 4-18 (0730) 6.0 4.0 -2.0 5.0 W/B 4-18 (0815) 5.0 4.0 -1.0 4.5 M/A 4-18 (1000) 4.0 5.0 1.0 4.5 Y/A 4-18 (1015) 6.0 6.0 0.0 6.0 B 4-18 (1130) 2.0 4.0 2.0 3.0 P/O 4-21 (0820) 4.0 1.0 -3.0 2.5 R/A 4-21 (1800) 3.0 2.0 -1.0 2.5 A/B 4-21 (1830) 2.0 2.0 0.0 2.0 P/G 4-25 (1650) 6.0 5.0 -1.0 5.5 A/W 4-25 (1915) 1.0 3.0 2.0 2.0 M 4-25 (1930) 4.0 1.0 -3.0 2.5 Y/G 4-25 (1945) 1.0 1.0 0.0 1.0 R 4-26 (0900) 1.0 1.0 0.0 1.0 A/0 4-30 (1030) 1.0 1.0 0.0 1.0 Y/O 5-02 (0920) 5.0 4.0 -1.0 4.5 0 5-03 (0730) 3.0 1.0 -2.0 2.0 R/P 5-03 (0940) 4.0 2.0 -2.0 3.0 R/W 5-04 (0815) 2.0 1.0 -l.0 1.5 P/R 5-04 (0900) 4.0 3.0 -1.0 3.5 Y/L 5-04 (0900) 5.0 3.0 -2.0 4.0 P/A 5-04 (0935) 6.0 5.0 -1.0 5.5 W/W 5-19 (1430) 5.0 4.0 -1.0 4.5 unclear; further migration as a cause of their disappearance is unlikey given the early spring passage of the species through central Michigan. It is possible that the missing birds constituted a floating population of nonterritorial, non-breeding males, a relatively common phenomenon in passerine breeding populations (reviewed by Brown 1969). Nature of territories Territories varied in shape from roughly circular (e.g. M) to rectangular (e.g. R/A) and appeared evenly distributed throughout the study area (Figure 2). Some males had uniquely shaped territories; for example W/B had a long, hour glass-shaped territory (Figure 2). Five birds (R/W,R/Y,P/G,B,P/A) had linear boundaries along the fenceline surrounding the study area. Social determinants of shape are not clear from the present analysis. Territory size averaged 1829.4 :_205.6 m2 (E :_SE, nal8). No extensive interstices existed between territories (Figure 2) as have been documented in one other Savannah Sparrow population in Michigan (Potter 1972). Nearly the entire study area was partitioned by aggressive behavior into exclusive breeding territories. Field measurement of aggression Tables 3 and 4 list AT scores on first and second field exams of males, where two ATs were administered during each field exam. The mean of these two exams is used for between-exam comparisions as well as in analysis of AT relative to TA. Statistical analysis of variation in AT scores within and between field exams revealed that the AT accurately and precisely measured a behavioral parameter, most likely inherent aggression (Table 5). Spearman rank correlation (rs) and Figure 2. Map of the 18 territories comprising the study population. Two- letter code is the color code of the defending territorial sparrow; symbol- ogy as in Methods section (UB 8 unbanded male). Dashed lines indicate territorial boundaries. 3c 2 .r“\ , \{in us / ""':’/r If \\ ' \/ ‘7 ‘4 5 a ' , .25), AT score immediately before placement into the aviary (1-tailed rs = -.20, P > .25), or TA (1-tailed rs = -.243, P > .25). Relationship of territory size to aggression Table 7 summarizes AT, TA, and summed aggression (SA) values for 18 territories. For each territory, SA is the sum of AT scores of territorial male sparrows on all adjacent territories. Summed aggression includes complete information on number of territorial neighbors as well as their relative aggression. Relationships between TA and AT, TA and SA are illustrated graphically in Figures 3 (AT) and 4 (SA), which reveal no precise trends. A possible exception is the increase of TA at high levels of AT (Figure 3). A one-way analysis of covariance (ACOVA) was performed on TA (m2) with 3 levels of AT as 3 treatments using SA as a linear covariate. The 3 levels of AT became obvious by examining the frequency of birds exhibiting each of 11 possible AT scores. These .mwuua woooam comuummaoo uooufiv oumuaaaomm cu “mononucoumq ca oumcawuon=m\maomuo>ou moan some now monocuvuonom mo gonad: ago an wooa>wv mean some an voocoaumaxo mammuo>ou mo gonads onu mamavo moam> «gnu a 7 ~c awn com «on Hooch 9. o.m moa. o mma mHH «a m\oA o.a 5mm. Ao.ava Had no em o\< m.H Nae. Ao.ova coa mm cc 3\M m.~ Ame. Am.~v¢ ewd «a mm 2 o.~ one. Ao.mvw mm wm he 3\< m.H ecu. A~.mvo~ omd mm mm U\A m.q can. Am.~vm~ am ma «a m\3 Amzv :63 vooooauomxo ouoom H< acofiuomuouoa a meowuoououofi mommoa was: roam amxo umuam mo mooauuoaoum mammuo>ou a Hooch a a .humfi>m man :u maouummm mama HmHHOuwuuou m we mnoumuow: mocmoaaon .o canny Table 7. Values of TA, AT, and SA for each territory (n - 18). Territory TA, m2 AT score1 SA score P/O 897 2.5 10.0 0 900 2.0 10.0 Y/O 900 4.5 5.5 R/0 1119 1.0 13.5 R/A 1151 2.5 5.0 R 1306 1.0 10.0 L/G 1357 1.5 13.5 R/W 1467 1.5 7.5 M 1535 2.5 18.0 A/W 1613 2.0 11.0 R/Y 1629 5.0 7.0 A/B 2139 2.0 6.5 B 2167 3.0 10.5 A/O 2262 1.0 13.0 P/G 2418 5.5 3.5 LG/R 2693 5.0 12.0 W/B 3391 4.5 4.0 P/A 3985 5.5 6.0 Mean 1: SE 1829.41205.6 2.92:0.38 9.25:0.92 1mean of 2 ATs, first exam 28 40l- . o 301- " O E - o" . 2 x . . < 20l— . p- 10- ' . I'. 1 ’1 AT Figure 3. The relationship between territory area (TA) and aggression (AT). 40' " ‘ P o 30. .. N o 5.. O .. O O / . ' .5 20 - ° . 0 10 _ ° ' ° : 1 I l 11 l I l L 1 2 4 6 8101214161820 SA Figure 4. The relationship between territory area (TA) and summed aggres- sion (SA). levels are 1.0-1.5 (n a 6), 2.0-3.0 (n = 6), and 4.5-5.5 (n a 6). By examining only the frequency of TA values for each possible AT score, AT treatment levels were derived independently of knowledge of actual values of TA. The ACOVA showed no significant effect of AT on TA (F a 2.04, n.s at P < .05) and SA was not a significant linear covariate of TA (F s .0014, n.s., beta regression coefficient - .0000452, Table 8). Table 9 summarizes the results at Duncan's multiple range test (Steel and Torrie 1960) applied to the three levels of AT (three treatment levels), each based on 6 values of TA. Birds of high aggression (level 3 of AT) defended significantly larger territories than birds of moderate aggression (P < .05, see Table 9 for x':;SE). Level l-level 3 and level l-level 2 comparisons were statistically nonsignificant. Morphological measurement Morphological comparision of 18 territorial males with 13 females revealed a marked degree of sexual dimorphism based on overall size (Table 10). Compared to females, males weighed more (19.41 vs. 16.78 g, P < .001), had longer wing chords (70.22 vs. 65.54 mm, P < .001), bills (10.25 vs. 10.17 mm, P > .50) but were leaner (1.42 vs. 2.31 fat index, P < .01) (see Table 10 for x :_SE of morphological variables). Cloacal length in males averaged 6.5 i_.44 mm (i t SE), indicating physiological reproductive capacity in the population by as early as 4 April. All 18 territorial males developed extended cloacas. Reproductive capacity in females (incubation patch development, 31 .3. .588. ... among 38 Banana... 3m 8 a. mo Shams ...om Banana N a o .9: $0.? 858:8 a ”86.2 N Go Vs a 8.8 N 3.3 .2 Banana 3 Ram 8&3 3.5:- 8.8882 2 Soy 8.2 S «on: 3.81 8.8m 86531 2 08m 8.8 N 3.? 8.8 8.5:- $683.0 N .2 m: a. £59.“ GEN 80% 98% a. 8&8 A» n 83.8160 ummcfi m 3 <3 be no 393 m now 8T: 33 <9 no a «933960 «6 38358 hos-mad .m 03mm. monogamoumm ma ao>oa comm qu H< mo owcmu N $.8NIN 3.5 TN 0 as: H TNRN 36-0.: m 3...: TN 0 N5: H NNNS 8.233 N 3.3 N1 0 E: H 08.: 3.79: N Ho>og Amocmoamucwwmv comaumaeoo : ANEVMm.H x B< .H< wo mHo>oH m on» new <9 «6 amok omcmm onHuHoz Boz n.smosnn .a manna 33 8:39 Econ 53 835mm Bugs mo umou I m. «03503..» «0 H93 Danmneom 30308 2303 N03 5 00032: 08 3838 82088801 SE 8093 N0 8.0x 0000 N000 H000 N00 N80 N0 2 a 2 a 2 0 0 2.0 H 2.01 20 H 2.8 3.0 H 00.00 :0 H 3.00 .000 H 5N N00 H 0%.: 8180 01 01 01 01 01 01 0a a 00.0 H 3.2 2.0 H SEN 00.0 H 00.8 9.0 H NN.0N 3. H 0.0 80 H N.) NN.0 H 3.3 831 2308.3 “6050.3 33% SEER saying an: 2083 a 30 9800. :80. was 888 000 >000 0.500000 38.8 2 0:0 BE 3039.8 04 No 031309 1003238 No 39:20 830380 08 mm H 082 .S 300.... mo. v m nu o.mmH.HVo.mom mmHHOufiuumu macawaucou mg. v m mm m.om~ an.wmm coaumasaoa I. aowuwquaoo mocmofimacmam : Away mm H + cam: mo Ho>mq .cofiuaumasoo mo mam>ma mucuauHMu msosmaucoo vcm acaumasmom 05» um mufimm mumcawuonsmlucmcfiaow cmoauwn ANEV muam >u0uauumu ca mwocmuwHMHo .- manna 35 oviducal expansion) was first detected on 19 May, three weeks after the earliest arrivals. DISCUSSION Nature of territories Breeding territory size averaged 1829.4 : 205.6 m2 and is obviously quite variable (Table 7). This average figure is slightly larger than the average (1740 m2, n - 13) reported by Welsh (1975) in Nova Scotia beach-dune habitat and much larger than that reported by Potter (1972) (1068 m2, n - 68 over 3 years) in fallow field habitat in south-central Michigan. However, direct comparision of mean territory area is best restricted to Michigan studies due to the unique habitat of the Nova Scotia pOpulation. There is no apparent explanation for the difference in average territory sizes obtained from studies in Michigan (this study and Potter 1972). No physiognomic or ecological characterstics of the study site appeared to influence the distribution, shape, or size of territories (Figure 2). For example, territories bordered by fenceline (e.g. B) or shoreline (e.g. LG/R) were not recognizably different from territories not so bordered. Ecological (i.e. habitat) bias of size and/or shape was virtually absent since all territories were situated in homogeneous short-grass habitat, apparently the preferred habitat. Territories did differ on a qualitative basis with respect to vegetative composition (proportion of wild clover and wild mustard) but this aspect was not quantified. Marginal areas of tall weedy growth, mostly goldenrod (Solidago 32.) were exclusively occupied by breeding Tree Swallows 36 (lridoprocne bicolor), Red-winged Blackbirds (Agelaius phoeniceus), Bobolinks (Dolichonyx oryzivorus), Eastern Medowlarks (Sturnella neglecta), and Song Sparrows (Melospiza melodia), but not by Savannah Sparrows of this study. Interspecific competition for territory (Orians and Willson 1974) had no apparent effect on territory size in the study population. Territory boundaries were generally stable before measurement of TA, by the last week of April. The only major changes in territorial boundaries (expansions after the disappearance of two males, B and R/O; see below) occurred following observer marking of pre-expansion boundaries. The social behavior determining boundaries in this species is not relentless, season-long interference competition; rather, boundaries are "agreed upon" by neighboring males and occasionally disputed following agreement. This pattern of territorial boundary determination has been observed in one other breeding population of Savannah Sparrows (Welsh 1975). Field measurement of aggression Consistency of AT scores of male sparrows within and between field exams (Table 5) clearly indicates that the aggression test measures a specific behavioral response to a novel stimulus. These results exactly parallel those of Burtt and Giltz (1969). Aggression tests administered to individual birds of four species consistently predicted a second score (Pearson r - .82, P < .001). Between-individual variation in aggressiveness and seasonal consistency of scores were also shown for two species (Brown-headed Cowbird, Molothrus ater; Common Crackle, Quiscalus quiscala). 37 Given the accuracy and precision of the aggression test, the question of what actually is measured remains. Originally, the aim was to quantify inherent aggression independent from site-related aggression. This approach was chosen over model presentation to territorial males (Brown 1975) because aggressive responses to such tests (Coulson 1968, Watson and Miller 1971) can vary over short-term time and Space, depending on the extent of previous habituation or model placement within the territory. Nevertheless, three territorial males were presented with realistic "singing" models of intruding male conspecifiics for a period of one to four days each. Responses to the model varied between males suggesting real differences in aggression but this variation was equalled by variation within males. This problem limits interpretation of male responses to the singing model. However, the statistically significant decrease of AT scores of territorial males over the breeding season (Table 5) was coincident with an observed decrease in the tendency to expand territory area through aggressive behavior. Whether the AT directly measures this tendency or a correlate thereof (e.g. general attentiveness associated with territorial vigilance) is irrelevant to statistical analysis of the relationships between AT and TA. Further, interpretation of the correlation between AT and TA does not depend upon whether the AT directly or indirectly measures the inherent tendency to defend a territory, but does depend on the accuracy with which the AT measures such a tendency or a correlate thereof. It is not possible to demonstrate whether the AT is a direct or indirect index of territorial behavior, and so the relationship between seasonally decreasing AT scores and territorial defense must be relied upon for assurance that the AT indeed quantifies territorial behavior along a continuous scale. Unfortunately, the link between seasonally decreasing AT scores and territorial defense, though validating comparisions of AT and TA, is merely qualitative. Behavioral assessment of the aggression test Proportion of interactions won in the aviary (winning percentage, WP, Table 6) by captive male sparrows was not significantly correlated with AT (l-tailed rS - .04, P < .25, n . 7 males). These results can be interpreted in three possible ways: 1) some unknown factor, uncorrelated with true inherent aggression, influenced competitive ability among captive birds unequally 2) the AT precisely quantifies a behavioral parameter other than inherent aggression 3) site-dependent dominance occurred in the aviary, a fact suggested by the frequent dominance reversals detected in the captive group of territorial males (Table 6). The first possibility, that of an uncontrolled variable, is supported by the significant linear regression of WP and fat index measured immediately upon capture, before placement into the aviary (linear regression equation: WP - .222 + .210 fat index, P of regression coefficient < .01). This analysis was carried out after the relationship between WP and AT proved statistically nonsignificant, and other meaningful considerations were made (see below) concerning possible alternate interpretations of the nonsignificant correlation. The second possible interpretation of the nonsignificant correlation of WP and AT is that each variable is a measure of a different behavioral parameter. This possibility came to light following observation of the captive group of territorial males. To illustrate, limited access measures such as that used in the present study may quantify skill at the competitive task rather than inherent aggression or competitive ability (Syme e£_gl, 1974). Further, measures of social dominance in one situation are often uncorrelated with measurements of aggression in a second situation (reviewed by Syme ‘gt‘al. 1974; e.g. Bennstein 1969). These considerations have led to a rejection of a unidimensional view of social dominance in favor of a multidimensional approach to understanding the allocation of group resources among individuals. Future assessments of aggression in more than one situation should take into account these pivotal findings. The third and final possible intrepretation concerns the frequent dominance reversals detected in the captive group of territorial males. Dominance hierarchies exhibiting reversals are functionally dependent on site-dependent dominance and aggression (Brown 1975), where individual members exhibit space-related changes in dominance. It is not surprising that territorial males should display site-dependent behavior, even under captive conditions that should not a priori reduce fundamentally the expression of territorial behavior. Site-dependent dominance has been demonstrated in natural populations of Steller's Jay (Cyanocitta stelleri, Brown 1963) and two species of parids (Odum 1942, Colquhoun 1942), and has been qualitatively noted in many studies of avian territorial systems. To summarize, three possible interpretations of the statistically nonsignificant correlation of WP and AT are not readily distinguishable. The alternate interpretations came to light following the completion of the behavioral assessment of the AT, and best serve not as validation 40 of what the AT directly measures but instead to point out potential problems of future similar assessments. In the following discussion of the relationship between aggression and territory size, AT scores of teritorial males will be used as a measure of inherent aggression (aggression that results in the retention or expansion of a previously claimed territory) due to the relationship between AT and territorial behavior observed in the field. Both variables showed a consistent decrease over the breeding season. Relationship of aggression and territory size at the population level Spatial competition is a sine qua non for aggression to operate as a determinant of territory size. Within the preferred habitats, all available space on the study site was divided into breeding territories via aggressive male behavior (Figure 2), supporting the notion that space was a contested resource in the sparrow population. In fact, intermale competition for Space may have eliminated a small number of males from breeding. For example, on May 8 a male banded in 1980 (Y) was observed attempting to copulate with a female present on territory A/B. This male, perhaps a late migrant, did not defend a territory in 1981 (but did in 1980) and behaved much like a nonbreeding "floater" (Brown 1969). Nonbreeding floaters without territories were detected in another Savannah Sparrow population (Welsh 1975). Another line of evidence illustrating a high level of spatial competition concerns the rapid reoccupation of newly vacated territories. In both cases where territorial males were suddenly 41 removed from their territories (B disappeared unexplainably by mid-May, R/O was killed in a mist net on May 16), neighboring males expanded their territories in a matter of days to include much of the defended area. These observations suggest that space for breeding was limited but do not allow further inferences about the existence of a floating population. Thus far it has been shown that space available for breeding was limited in the study area. Extensive variation in breeding territory size (TA) occurred in the study population (range of TA 8 897-3985 m2, x - 1829 m2). Further, aggressiveness varied significantly between but not within territorial males (Table 5). These features of the study population of territorial male Savannah Sparrows allow evaluation of the relationship between aggression and territory size. Variation in territory size at the population level cannot be explained on the basis of varying individual aggression (Table 8). Aggression appears to be important at high but not intermediate or low levels (Table 9). The dispersion of the 14 lowest TA values in Figure 1 appears random relative to the AT axis; in fact, the significant positive correlation (l-tailed rs - .48, P < .025, n . 18 territories) of AT and TA is eliminated when the four greatest values of AT are excluded (l-tailed r8 2 -.02, P > .50). The same qualitative results can be expected if the four values were to be excluded from the range test of Table 9. Excluded males (AT) are LG/R (4.5), W/B (4.5), D/A (5.5) and P/G (5.5). Exclusion of these males from analysis of TA is not warranted because of their great impact on 42 Spacing in the population; they occupied 40% of the total defended area. These findings are in disagreement with those of other studies of aggression relative to territory size at the population level in vertebrates. Territory area and an index of aggression were strongly correlated within years in breeding Red Grouse in Scotland (Watson and Miller 1971, Watson and Moss 1971). Aggression was shown to be causally related to territory size by testosterone implantation of subordinate male grouse which subsequently increased territory size and mating success. Territory size and aggression are clearly positively related in several species of lizards (Rand 1967, Brattstrom 1974, Fox 35. 31. 1981). Aggression is strongly selected for in natural populations of birds via differential reproduction of aggressive males, in lekking species (Robel 1972, Ballard and Robel 1974) colonial species (Coulson 1968, Spurr 1974) and species defending exclusive breeding territories, including Fringillids (Brown 1964, 1969, Watson and Miller 1971). Craig 32 31. (1965) artifically increased aggressiveness in strains of chickens in five generations by selecting only the most aggressive males for breeding. The mating system of the study population differed greatly from these highly selective, usually polygamous mating systems in which a very small proportion of territorial males are responsible for nearly all inseminations of females. To illustrate, sparrows of this study were monogamous; nearly all territorial males succeeded in obtaining females with whom they presumably mated (DLB, pers. comm.). This observation is supported by data on nesting: at least five nests were found containing either eggs or nestlings, each nest located in a 43 territory defended by a single male associated with a single female. These are territories (AT of defending male) L/G (2.5), LG/R (4.5), R/A (2.5), R/O (1.0), and W/B (4.5). If these data represent the entire breeding p0pulation, breeding opportunities were not limited to highly aggressive male sparrows. Thus it is not unusual that only a few males in the study population were highly aggressive (Figure 3), since strong selection for aggression in avian populations seems functionally dependent on nonrandom mating of only highly aggressive males (Craig 35 ‘21. 1975, Robel 1972). The nature of selection for male aggressiveness in the study population is made clear here in order to unambiguously discuss possible selection for male experience (below). Relationship of aggression and territory size at the level of neighbors Aggression may be a significant determinant of territory size at the level of contiguous pairs of territorial males. Territorial male Savannah Sparrows defended the same general area between years in the study area (DLB, pers. comm.), a behavioral trait tempering a population-wide effect of aggression on territory size. This effect is manifest in the selection of territory size by returning males. In early spring, areas defended the previous year or years are chosen instead of presumably available larger sites. For example, males B and R/Y occupied the same territories in 1980 and 1981; initial defense in the latter year occurred while other larger areas in similar habitat were undefended and presumably available. While they occupy the same general area between years, territory size of site-faithful males varies considerably, suggesting competition 44 with neighboring males might affect territory size. This phenomenon has been demonstrated in a lizard population (Uta stansburiana) in Oklahoma (Fox 35.2i‘ 1981). Dominant lizards in dominant-subordinate pairs defended larger and more vegetatively diverse home ranges than subordinates. Home ranges of dominant lizards resembled survivor rather than non-survivor home ranges in a second population. An important conclusion was that during formation of individual home ranges, aggressive behavior likely defines a matrix of dominant- subordinate dyads among neighbors that influence territory acquisition (F°X.EE“3l' 1981), and, importantly, the genetic nature of future populations. The hypothesis that spatial competition between territorial males occurs at the dyad and not the population level was tested in the present study (Table 11). 0f 27 pairs of contiguous territorial males known to have aggressively interacted, birds with greater AT scores defended significantly larger territories than birds with lower AT scores (l-tailed paired E-test, {d _+_- SE of difference - 365.6 1 199.6, t - 1.83, P < .05). A similar analysis performed on 28 of 125 possible non-contiguous pairs (28 pairs were randomly selected to eliminate a sample size effect) revealed an expected positive but statistically nonsignificant difference in territory size (Edi: SE = 288.1 1: 230.9, t - 1.25, P < .15) This analysis shows that competition for territory area occurs primarily between neighboring male sparrows, explaining the nonsignificant relationship between TA and AT at the population level. The historical effect of male site fidelity is taken into account. 45 Aggressive males defend larger territories than less aggressive neighboring males, although all birds do not attain territory area in direct proportion to aggressiveness due to the presence of site-faithful males. Site fidelity merely dampens the overall positive effect of aggression on territory size. Predictions outlining future work What is the motivation for defense of larger areas by male Savannah Sparrows? A hypothesis is that defense of a large territory may increase the probability of mating success and related increases in fitness. By this logic, females improve their fitness by selecting aggressive males that defend large, high quality territories (Searcy 1979). Such males experience enhanced reproductive success if female intersexual selection is biased toward aggressive males. To the contrary, female choice may not always favor aggressive males (Searcy 1979). As argued above, selection for very aggressive males appears weak in the sparrow population. Aggression and territory area are related but more aggressive males may not experience optimal reproductive output, a variable not measured in 1980 or 1981. Experience of defense against conspecifics, and not aggression, was positively correlated with male reproductive output in Red-winged Blackbirds (Yasukawa 1979). Late season territoriality by young males in a Washington population provided opportunity for gaining experience defending area. Similar Opportunities were available in the study population. The hypothesis that territory size and aggression are causally related does not rule out the hypothesis that experienced (older) males If generate greater reproductive output than inexperienced males. Inexperienced males defending territories for the first time should be aggressive to gain experience in holding a territory, establishing the correlation between TA and AT; inexperienced males, however, may not reproduce as effectively as experienced males. Assuming females choose experienced males for breeding, male qualities directly associated with experience and not aggression will be strongly selected for. Several testable predictions can be made based on the above line of reasoning. Aggressiveness and age should be inversely related. If female choice is near optimal, experienced males not spending a great deal of time and energy in territorial defense can allocate a maximum of resources for reproduction, thus generating a greater number of progeny or more physiologically fit progeny than inexperienced males. In either case experienced males and females will contribute maximally to future generations. Territory size between years should shrink consequently with increasing breeding success since experience and not large territory area is thought to augment breeding success of males. Reductions in territory size and attendant reproductive experience predict that territory size is not related to resource density. 47 LIST OF REFERENCES LIST OF REFERENCES Alcock, J. 1979. Animal behavior: an evolutionary approach. 2nd ed. Sinauer, Sunderland, Massachusetts. Ballard, W.B. and R.J. Robel. 1974. Reproductive performance of dominant male Greater Prairie Chickens. Auk 91: 75-85. Bennstein, I.S. 1969. Stability of the status hierarchy in a pigtail monkey group (Macaca nemistrina). Anim Behav. 17: 452-458. Bent, A.C. 1968. Life histories of North American cardinals, grosbeaks, buntings, towhees, finches, sparrows and their allies. Dover, New York. Brattstrom, B.H. 1974. The evolution of reptilian social organization. Am. 2001. 14: 35-49. Brown, J.L. 1963b. Aggressiveness, dominance, and social organization in the Steller Jay. Condor 65: 460-484. 1964. The evolution of diversity in avian territorial sys- tems. Wil. Bul. 76: 160-169. 1969. Territorial behavior and population regulation in birds: a review and re-evaluation. Wil. Bul. 81: 293-329. 1975. The evolution of behavior. Norton, New York. and G.H. Orians. 1970. Spacing patterns in mobile animals. Ann. Rev. Ecol. Syst. 1: 239-262. Burtt, H.E. and M.L. Giltz. 1969. A statistical analysis of black- bird aggressiveness. Ohio J. of Sci. 69: 58-62. Carpenter, C.R. 1958. Territory: a review of concepts and problems. In_ Behavior and evolution, A. Roe and G.G. Simpson eds. Yale Univ.' Press, New Haven. Carrick, R. 1963. Ecological significance of territory in the Austra— lian Magpie, Gymnorhina tibican. Proc. 13th Int. Ornith. Congr.: 740-753. Cody, M.L. and C.B.J. Cody. 1972. Territory size, clutch size, and food in populations of wrens. Condor 74: 473-477. [.0 Colquhoun, M.K. 1942. Notes on the social behavior of blue tits. Brit. Birds 35: 234-240. Coulson, J.C. 1968. Differences in the quality of birds nesting in the centre and on the edges of a colony. Nature 217: 478-479. Craig, J.V., Ortman, L.L. and A.M. Guhl. 1965. Genetic selection for social dominance ability in chickens. Anim. Behav. 13: 114-131. Darwin, C. 1859. On the origin of species by means of natural selection. Murray, London. Ettinger, A.O. and J.R. King. 1979. Time and energy budgets of the Wil- low Flycatcher (Empidonax trailli) during the breeding season. Auk 97: 533-546. Ewald, P.W., Hunt, G.L. and M. Warner. 1980. Territory size in Western Gulls: importance of intrusion pressure, defense investments, and vegetation structure. Ecology 61: 80-87. Fox, S.F., Rose, E. and R. Myers. 1981. Dominance and the aquisition of superior home ranges in the lizard Uta stansburiana. Ecology 52: 883-893. Fretwell, S.D. 1969. Dominance behavior and winter habitat distribution in juncos (Junco hyemalis). Bird-banding 40: 1-25. 1972. Populations in a seasonal environment. Princeton Monogr. in Pop. Biol. 5. Princeton Univ. Press, Princeton. Gass, C.L. 1979. Territory regulation, tenure, and migration in Rufous Hummingbirds. Can. J. Zool. 57: 914-923. Gass, C.L., Angehr, G. and J Centa. 1976. Regulation of food supply by feeding territoriality in the rufous hummingbird. Can J. Zool. 54: 2046-2054. ‘ Gill, F.B. and L.L. Wolf. 1975. Economics of feeding territoriality in the Golden-winged Sunbird. Ecology 56: 333-345. Glas, P. 1960. Factors governing density in the Chaffinch (Fringilla coelebs) in different types of wood. Arch. Neerl. Zool. 13: 466-472. Greenwood, P.J., Harvey, P.H. and C.M. Perrins. 1979. The role of dis- persal in the Great Tit (Parus major): the causes, consequences, and heritability of natal dispersal. J. Anim. Ecol. 48: 123-142. Hinde, R.A. 1956. The biological significance of territories of birds. Ibis 98: 340-369. Holmes, R.T. 1970. Differences in population density, territoriality, and food supply of Dunlin in arctic and subarctic tundra. 12_Anima1 populations in relation to their food resources, A. Watson ed. Black- well, Oxford. 49 Howard, E. 1920. Territory in bird life. Collins, London. Huxley, J.S. 1934. A natural experiment on the territorial instinct. Brit. Birds 27: 270-277. Ketterson, E.D. 1979. Aggressive behavior in wintering Dark-eyed Jun- cos: determinants of dominance and their possible relation to geo- graphic variation in sex ratio. Wil. Bul. 91: 371-383. Kluyver, H.N. and N. Tinbergen. 1953. Territory and the regulation of density in titmice. Arch. Neerl. Zool. 10: 265-287. Kodric-Brown, A. and J.H. Brown. 1978. Influence of economics, inter- specific competition, and sexual dimorphism on territoriality in migrant rufous hummingbirds. Ecology 59: 285-296. Lack, D. 1954. The natural regulation of animal numbers. Clarendon Press, Oxford. Lyon, D.L. 1976. A montane hummingbird system in Oaxaca, Mexico. Wil. Bul. 88: 280-299. McWhirter D. and D.L. Beaver. 1977. Birds of the Capital Count Area of Michigan. Publ. Mus. Biol Ser., Michigan State University 5: 353- 442. Moffat, C.B. 1903. The spring rivalry of birds. Some views on the limit to multiplication. Irish Naturalist 12: 152-166. Myers, J.P., Connors, P.G. and F.A. Pitelka. 1979. Territory size in wintering sanderlings: the effects of prey abundance and intruder density. Auk 96: 551-561. Nice, M.M. 1937. Studies in the life history of the Song Sparrow I. Trans. Linnean Soc. N.Y. 4: 57-83. 1941a. The role of territory in bird life. Amer. Midl. Nat. 26: 441-487. Noble, G.K. 1939. The role of dominance in the social life of birds. Auk 56: 263-273. Odum, E.P. 1942. Annual cycle of the black-capped chickadee. Auk 59: 499-531. Orians, G.H. 1961. The ecology of blackbird (Aegelaius) social systems. Ecol. Monogr. 31: 285-312. 1971. Ecological aspects of behavior. In_ D.S. Farner v and J.R. King, eds. Avian Biology, V. 1. Acadenic Press, New York. Pitelka, F.A., Tomjich, P.Q. and T.W. Treichel. 1955. Ecological re- lations of jaegers and owls as lemming predators near Barrow, Alaska. Ecol. Monogr. 25: 85-117. 50 Potter. P.E. 1972. Territorial behavior in Savannah Sparrows in south- eastern Michigan. Wil. Bul. 84: 48-59. Rand, A.S. 1967. Ecology and social behavior in the iguanid lizard, Anolis lineatopus. Proc. U.S. Natl. Mus. 122: 1-79. Robel, R.J. 1972. Possible function of the lek in regulating tetraonid populations. Proc. 15th Int. Ornith. Congr. (1970): 121-133. Schoener, T.W. 1971. Theory of feeding strategies. Ann. Rev. Ecol. Syst. 2: 369-404. Searcy, W.A. 1979. Male characteristics and pairing success in male Red- winged Blackbirds. Auk 96: 353-363. Smith, S.M. 1978. The "underworld" in a territorial sparrow: adaptive strategy for floaters. Am. Nat. 112: 571-582. Spurr, E.B. 1974. Individual differences in the aggressiveness of Adelie Penguins. Anim. Behav. 22: 931-940. Stefanski, R.A. 1967. Utilization of the breeding territory in the Black- capped Chickadee. Condor 69: 259-267. Stenger, J. 1958. Food habits and available food of ovenbirds in relation to territory size. Auk 75: 335-346. Stewart, R.E. and J.W. Aldrich. 1951. Removal and repopulation of breeding birds in a spruce-fir forest community. Auk 68: 471-482. Stiles, F.G. and L.L. Wolf. 1970. Hummingbird territoriality at a trop- ical flowering tree. Auk 87: 467-498. Stokes, A.W. ed. 1974. Territory. Dowden, Hutchinson, and Ross, Strouds- burg, Pennsylvania. Syme, G.J. 1974. Competetive orders as measures of social dominance. Anim. Behav. 22: 931-940. Syme, G.J., Pollard, J.S., Syme, L.A. and R.M. Reid. 1974. An analysis of the limited access measure of social dominance. Anim. Behav. 22: 486- 500. Tinbergen, N. 1957. The functions of territory. Brid Study 4: 14-29. Tompa, F.S. 1962. Territorial behavior: the main controlling factor of a local Song Sparrow population. Auk 79: 687-697. Verner, J. 1977. On the adaptive significance of territoriality. Am. Nat. 111: 769-775. Watson, A. and R.L. Miller. 1971. Territory size and aggression in a fluctuating Red Grouse population. J. Anim. Ecol. 40: 367-383. Watson, A. and R. Moss. 1971. Spacing as affected by territorial beha- vir, habitat, and nutrition in Red Grouse (Lagopus 1;_scoticus). .In' A.H. Esser, ed. Behavior and envorironment: the use of space by animals and man. Plenum Press, New York. 1972. A current model of population dynamics in Red Grouse. Proc. 15th Ornith. Congr. 134-149. Weeden, J.S. 1965. Territorial behavior of the Tree Sparrow. Condor 67: 193-209. WeiSE, C.M. 1956b. .Nightly unrest in caged migratory sparrows under out- door conditions. Ecology 47: 274-287. Welsh, D.A. 1975. Savannah sparrow breeding and territoriality on a Nova Scotia dune beach. Auk 92: 235-251. Welty, J.C. 1974. The life of birds. 2nd ed. Saunders, Philadelphia, London, Toronto. Wolf, L.L. 1969. Female territoriality in a tropical hummingbird. Auk 96: 490-504. Wolfson, A. 1945. The role of the pituitary, fat deposition, and body weight in bird migration. Condor 47: 95-127. Wynne-Edwards, V.C. 1962. Aminal dispersion in relation to social be- havior. Hafner, New York. Yasukawa, K. 1979. Territory eStablishment in Red-winged Blackbirds: importance of aggressive behavior and experience. Condor 81: 258-264. Zimmerman. J.L. 1966. Polygamy in the dickcissel. Auk 83: 534-546. 52