Territorial songs indicate male quality in the sac-winged bat Saccopteryx bilineata (Chiroptera, Emballonuridae)

Defense of territories in many animal species involves the advertisementof territory holder quality by acoustic signaling. In the sac-wingedbat Saccopteryx bilineata, males engage in territorial countersingingwhen reoccupying their day-roost territories in the morningand in the evening before abandoning the roost for the night.Females roost mainly in male territories, and territory holdersare reproductively more successful than nonterritorial males.In territorial songs of male S. bilineata, we distinguished6 syllable types and parameterized their acoustic properties.The analysis of 11 microsatellite loci allowed assignments ofjuveniles to their parents. Males had a higher reproductivesuccess both when they uttered more territorial songs per dayand when their long buzz syllables had a lower end frequencyof the fundamental harmonic. Long buzzes had a harsh qualitydue to a pulsation of the fundamental frequency at the syllableonset and also had the highest sound pressure level of all syllabletypes in most territorial songs. Territorial songs and especiallylong buzz syllables are thus likely to advertise territory holderquality and competitive ability.     

Castration Lowers Aggression but not Social Dominance in Male Haplochromis burtoni (Cichlidae)

Adult males of the African cichlid fish Haplochromis burtoni are either territorial or nonterritorial. In nature and aquaria, only territorial males breed and they have markedly more developed testes than do nonterritorial males. Territorial males are clearly dominant over nonterritorial males of the same age and size, and also exhibit higher levels of aggression. In this study, territorial males were gonadectomized in order to assess the effects of lowered androgen levels on their aggression and dominance status. Completely castrated males showed a pronounced reduction in both testosterone and 11-ketotestosterone, as well as significantly lower aggression scores than sham-operated males. Partially castrated males (> 0.005 g gonadal tissue remaining at autopsy) exhibited intermediate levels of both androgen levels and aggression. Surprisingly, in dyadic encounters with normal nonterritorial males of the same size, castrated territorial males retained their dominance despite decreased androgen levels and aggression.     

Mutualism between the territorial intertidal limpet Patella longicosta and the crustose alga Ralfsia verrucosa

Mutualistic relations between plants and animals are well documented on land but have received less attention in marine systems. This study examined the relationship between the territorial intertidal limpet Patella longicosta and the crustose brown alga Ralfsia verrucosa. Adult Patella are found exclusively in association with Ralfsia, on which they feed, while Ralfsia occurs primarily, but not exclusively, in Patella territories. Ralfsia benefits directly from both the presence and the territorial behaviour of Patella. Algal productivity was assessed by measuring oxygen evolution and utilization in situ and deriving photosynthesis/irradiance curves. Productivity was increased by about 30% by the presence of Patella in both summer (P _max of grazed algae 0.0098; ungrazed algae 0.0063 mg C · cm^-2 · h^-1) and winter (P _max grazed algae 0.0081; ungrazed algae 0.0053 mg^-2 · C · h^-1). Algal growth rates were not significantly increased by the application of limpet mucus in the laboratory. We did not examine nutrient regeneration by the limpet, but the increase in photosynthetic rate may depend on the limpet's grazing pattern which creates secondary sites for growth. Ralfsia also benefited from the territorial behaviour of Patella. The effects of different grazing regimes were investigated in different seasons by removing territorial limpets and either excluding all limpets using copper-based antifouling paint, or allowing access to non-territorial limpets (mostly P. oculus) using partial paint barriers. Exclusion of all limpets resulted in rapid overgrowth of Ralfsia plants by the foliose green alga Ulva sp.. Where non-territorial limpets had access to the plants overgrowth was reduced but Ralfsia plants were entirely removed by destructive grazing. Non-territorial grazers removed 90% of Ralfsia plants within 4 weeks in summer and 60% in winter. In control treatments P. longicosta prevented overgrowth by Ulva and actively excluded vagrant grazers, preventing overgrazing. Based on these findings, the association between the limpet and alga can be regarded as a nonobligate mutualism.     

Territories of male and female terrestrial salamanders: costs,benefits, and intersexual spatial associations

Summary I used a mark-recapture study to estimate home areas for 107 red-backed salamanders (Plethodon cinereus) in a natural forest habitat. Both males and females of this species defend feeding territories, but I presume that some individuals in this relatively highdensity population (approximately 2.8 salamanders per m²) are nonterritorial floaters. Although territorial salamanders exhibited greater numbers of tail autotomies, they had significantly longer relative tail lengths. This difference suggests that territorial individuals gain benefits from territorial ownership. From the observation that home area size was inversely correlated with body size, I infer that larger animals gained higher quality foraging areas. Home areas of adults were significantly more segregated intrasexually and more aggregated intersexually than would be expected from a random distribution. Furthermore, intersexual overlap of home areas was significantly greater than intrasexual home area overlap. Territorial defense of feeding areas by male and female red-backed salamanders therefore also may play a role in mating behavior.     

Molluscan grazing and macroalgal zonation on a rocky intertidal platform at Perth,Western Australia

Abstract Intertidal limestone platforms off Perth show a characteristic pattern of algal zonation, with dense macroalgal beds nearshore bounded by a ‘barren zone’ along the seaward edge. Abalone (Haliotis roei) and several species of limpets and chitons are abundant in the barren zone, which is generally devoid of non-coralline macroalgae. The relative importance of abalone versus limpets and chitons in limiting macroalgal abundance in the barren zone was evaluated by manipulating the presence of each group in a factorial experiment. Percentage algal cover was measured photographically in 0.25m² plots at 1–2 month intervals for 9 months. Mean algal cover (mainly the foliose green alga, Ulva rigida) was highest in plots where all grazers were excluded (77–99%), intermediate where only limpets and chitons were excluded (37–85%), and lowest where only abalone were excluded (4–30%) or where no grazers were excluded (2–19%). The effect of limpets and chitons accounted for 55–89% of the variance in total algal cover, whereas the effect of abalone generally accounted for <10% of the variance. Similar results were obtained in terms of the biomass of Ulva rigida at the end of the experiment. Haliotis roei are relatively large and sedentary herbivores, feeding mainly on drift algae. Their effect on algal abundance was mediated both through pre-emption of space, which might otherwise be colonized by algae, and by grazing around their home scar. Limpets and chitons are smaller than abalone, but were much more abundant. Intensive grazing of the reef surface by limpets and chitons precluded the establishment of non-coralline macroalgae, even where abalone were absent.     

Hypertrophy of gonadotropin releasing hormone-containing neurons after castration in the teleost,Haplochromis burtoni

In the African cichlid fish, Haplochromis burtoni, males are either territorial or nonterritorial. Territorial males suppress reproductive function in the nonterritorial males, and have larger gonads and larger gonadotropin-releasing hormone- (GnRH) containing neurons in the preoptic area (POA). We describe an experiment designed to establish the causal relationship between large GnRH neurons and large testes in these males by determining the feedback effects of gonadal sex steroids on the GnRH neurons. Territorial males were either castrated or sham-operated, 4 weeks after which they were sacrificed. Circulating steroid levels were measured, and the GnRH-containing neurons were visualized by staining sagittal sections of the brains with an antibody to salmon GnRH. The soma areas of antibody-stained neurons were measured with a computer-aided imaging system. Completely castrated males had markedly reduced levels of circulating sex steroids [11-ketotestosterone (11 KT) and testosterone (T)], as well as 17β-estradiol (E₂). POA GnRH neurons in castrates showed a significant increase in mean soma size relative to the intact territorial males. Hence, in mature animals, gonadal steroids act as a brake on the growth of GnRH-containing neurons, and gonadal products are not responsible for the large GnRH neurons characteristic of territorial males. © 1992 John Wiley & Sons, Inc.     

Zooplankton-phytoplankton interactions in a eutrophic lake

Enclosure experiments were made in a cyanobacteria dominatedlake (Lake Rotongaio) to assess the impact of zooplankton (>150µm) grazing on algal growth rates and determine the effectof diel and vertical changes in zooplankton grazing intensityand nutrient (NH₄-N) regeneration upon abundance of phytoplankton.The filamentous cyanobacterium Anabaena minutissima var. attenuataand diatom Cyclotella meneghiniana showed a negative linearchange in abundance with a gradient in zooplankton grazing intensity.Phytoflagellates were not grazed and showed a positive linearchange in abundance with increasing zooplankton biomass. Theseeffects, as well as shortening of filament length of Anabaena,were caused by raptorial feeding by the alanoid copepod Boeckellapropinqua which dominated the zooplankton. Phytoplankton growthwas not stimulated by addition of nutrients, suggesting nutrientregeneration was not important. Diel and vertical changes infeeding and NH₄-N regeneration rates were measured in Marchand June 1988. Diel differences were more pronounced in Marchwhen the water column was stratified. Specific feeding rateswere more important than vertical changes in zooplankton biomassin determining community grazing rates in March, but in Junewhen the water column was mixed, vertical distribution of zooplanktonbiomass was important. Zooplankton grazing was an importantloss process for phytoplankton in the lower part of the epilimnionin Lake Rotongaio.     

The effects of grazing on the distribution and composition of low-shore algal communities on the central coast of Portugal and on the southern coast of Britain

A red algal turf is often found just below the barnacle/limpet zone of many European shores, especially on steep shores of moderate exposure. The hypothesis that grazing by limpets determines the upper limit of distribution of this red algal turf was tested on moderately exposed shores in Portugal and Britain. We also aimed to assess whether the grazing effect is modified at various spatial scales. Grazers were excluded by fences, with half-fenced and unfenced controls. Exclusion plots were rapidly colonised by green ephemeral algae in the months immediately after the beginning of the experiment (summer); these algae were later replaced by perennial algae. The percentage cover of turf-forming macroalgae showed a significant increase at both locations. The upper limit of distribution extended more than 50 cm on most of the shores studied. In contrast, control and half-fenced plots remained devoid of algae. After 2 years, ungrazed plots were mainly colonised by a red algal turf (e.g. Caulacanthus ustulatus, Gelidium spp., Laurencia pinnatifida) in Portugal, while canopy cover (Fucus serratus and Himanthalia elongata) dominated in Britain in marked contrast to the grazed plots. Physical factors acting at both local and geographical scales may explain these differences. However, although physical factors probably have an important influence on the identity, size and abundance of sublittoral fringe macroalgae, grazers play a major role in directly setting their upper limits. The effect of grazing by limpets was not consistent for all of the morphological algal groups and spatial scales considered in the present study. The effect of grazing on the cover of turf algae varied between Portugal and Britain (location scale), while effects on ephemeral and canopy algal cover varied at the shore scale within location.     

The influence of crevice size on the protection of epilithic algae from grazers

1. This study investigated how the size of crevices might affect their effectiveness as refuges for diatom‐dominated algal assemblages from the grazing minnow Campostoma anomalum. 2. Crevice size was manipulated by making fired‐clay substrates, using moulds to produce eight substrates with pits from 1.17 to 22.0 mm diameter. Non‐pitted clay‐stones and limestone were also tested. Cages were used to control the access of Campostoma to arrays of the 10 different algal‐colonised substrates. The grazing treatments were: open and grazed, caged and ungrazed, and a grazed cage control. The experiment was replicated in eight large outdoor tanks. After 3 weeks, substrates were brushed and chlorophyll a concentrations of the removed algae and the algae remaining in pits were measured. 3. The experiment was field‐validated by exposing arrays of substrates to grazing Campostoma in five pools of a limestone stream. 4. The clay‐stone and limestone substrates accrued similar algal biomass and assemblages. 5. Smaller crevices provided more protection against grazing than larger crevices. Specifically, pits with openings smaller than 2 mm protected the enclosed algal assemblages in both the tank and field experiments. Larger pits provided less protection and pits over 7 mm in diameter were heavily grazed and may even be preferentially grazed by Campostoma. 6. None of the tested pit sizes were protective against larval chironomid grazers in the tank experiment, demonstrating that differences in the grazer size influence the effectiveness of crevice refuges.     

Top-down impact through a bottom-up mechanism: the effect of limpet grazing on growth,productivity and carbon allocation of Zostera marina L. (eelgrass)

The unusual appearance of a commensal eelgrass limpet [Tectura depicta (Berry)] from southern California at high density (up to 10 shoot^–1) has coincided with the catastrophic decline of a subtidal Zostera marina L. meadow in Monterey Bay, California. Some commensal limpets graze the chloroplast-rich epidermis of eelgrass leaves, but were not known to affect seagrass growth or productivity. We evaluated the effect on eelgrass productivity of grazing by limpets maintained at natural densities (8±2 shoot^–1) in a natural light mesocosm for 45 days. Growth rates, carbon reserves, root proliferation and net photosynthesis of grazed plants were 50–80% below those of ungrazed plants, but biomass-specific respiration was unaffected. The daily period of irradiance-saturated photosynthesis (H _sat) needed to maintain positive carbon balance in grazed plants approached 13.5 h, compared with 5–6 h for ungrazed plants. The amount of carbon allocated to roots of ungrazed plants was 800% higher than for grazed plants. By grazing the chlorophyll-rich epidermis, T. depicta induced carbon limitation in eelgrass growing in an other-wise light-replete environment. Continued northward movement of T. depicta, may have significant impacts on eelgrass production and population dynamics in the northeast Pacific, even thought this limpet consumes very little plant biomass. This interaction is a dramatic example of top-down control (grazing/predation) of eelgrass productivity and survival operating via a bottom-up mechanism (photosynthesis limitation).     

SIZE, SHAPE AND SHELL MORPHOLOGY IN THE ANTARCTIC LIMPET NACELLA CONCINNA AT SIGNY ISLAND, SOUTH ORKNEY ISLANDS

Comparison of the size, shape and shell morphology in littoraland sub-littoral morphs of the Antarctic limpet Nacella concinnareveal differences in shell morphology which are enhanced bystructural anomalies within the shells of the two types. Infestationof sub-littoral shells by the conchocelis phase of an endolithicalga significantly affects shell density and total chlorophylllevels in the two shell morphs. The surface sculpture of sub-littoralshells is characterised by a series of grooves, the configurationof which closely resembles that of the radular teeth in N. concinna.Limpets utilise the available food supply within the shell matrixof other limpets by grazing the shell material. Epibiotic growthof calcareous algae prevent erosion and preserve underlyingshell layers. In severe cases, where protection is lacking,intraspecific shell grazing may remove parts of the shell exposingthe internal tissues. The Dominican Gull, Larus dominicanus, is a major shore predatorof both shell morphs. Gull middens contain both shell typesbut are dominated by the more accessible littoral shells. Comparisonof living populations and midden assemblages indicates thatsize and shape selection of prey occurs, with pear-shaped limpetsbetween 21 mm and 29 mm in length being taken preferentially. Apparent differences in shell form are induced by physical,biological and behavioural influences. Littoral animals arerobust in nature, resist avian pre-dation and are not extensivelygrazed whereas those of the sub-littoral are not subject tothe same degree of predatory attention but suffer a gradualdepletion of their shallower shell form through a combinationof algal infection and intraspecific shell grazing. (Received 21 February 1990; accepted 5 July 1990)     

Zooplankton community size structure and taxonomic composition affects size-selective grazing in natural communities

The body size of an individual zooplankton is well related to its grazing rate and to the range of particle sizes it can ingest, and since cladocerans and copepods feed differently, they follow different relationships. Based on these general patterns in individual organisms, we tested whether the size structure and taxonomic composition of more complex natural zooplankton communities are related to their in situ grazing rate and to the range of algal sizes they graze. We compared community grazing rates on individual algal taxa in two communities dominated by small cladocerans, three communities dominated by large cladocerans and three copepod-dominated communities. Small algae were usually grazed most intensively, but grazing rates were poorly related to algal size alone. The range in size of grazed algae increased with increasing mean zooplankton body size, but differed systematically with their taxonomic composition. Communities dominated by Ceriodaphnia or Holopedium grazed a narrower size range of algae [maximum greatest axial length dimension (GALD)=16–36 μm)] than communities with large biomasses of Bosmina or Daphnia (maximum GALD=28–78 μm). Copepod-dominated communities followed the same general relationship as cladocerans. Daphnia-dominated communities grazed the broadest range of algal sizes, and their total grazing rates were up to 2.4 times their grazing rates on small (<35 μm) “highly edible” algae, a difference of similar magnitude to those found in successful trophic cascade biomanipulations. Received: 31 March 1998 / Accepted: 19 October 1998     

The Role of Encrusting Coralline Algae in the Diets of Selected Intertidal Herbivores

Kalk Bay, South Africa, has a typical south coast zonation pattern with a band of seaweed dominating the mid-eulittoral and between two molluscan-herbivore dominated upper and lower eulittoral zones. Encrusting coralline algae were very obvious features of these zones. The most abundant herbivores in the upper eulittoral were the limpet, Cymbula oculus (10.4 ± 1.6 individuals m⁻²; 201.65 ± 32.68 g.m⁻²) and the false limpet, Siphonaria capensis (97.07± 19.92 individuals m⁻²; 77.93 16.02 g.m⁻²). The territorial gardening limpet, Scutellastra cochlear, dominated the lower eulittoral zone, achieving very high densities (545.27 ± 84.35 m⁻²) and biomass (4630.17 ± 556.13 g.m⁻²), and excluded all other herbivores and most seaweeds, except for its garden alga and the encrusting coralline alga, Spongites yendoi (35.93 ± 2.26% cover). In the upper eulittoral zone, encrusting coralline algae were only present in the guts of the chiton Acanthochiton garnoti (30.5 ± 1.33%) and the limpet C. oculus (2.9 ± 0.34%). The lower eulittoral zone limpet, Scutellastra cochlear also had a large percentage of encrusting coralline algae in its gut with limpets lacking gardens having higher (45.1 ± 1.68%) proportions of coralline algae in their guts than those with gardens (25.6 ± 0.8%). Encrusting coralline algae had high organic contents, similar to those of other encrusting and turf-forming algae, but higher organic contents than foliose algae. Radula structure, grazing frequencies as a percentage of the area grazed (upper eulittoral 73.25 ± 3.60% d⁻¹; lower eulittoral 46.0 ± 3.29% d⁻¹), and algal organic content provided evidence to support the dietary habits of the above herbivores. The data show that many intertidal molluscs are actively consuming encrusting coralline algae and that these seaweeds should be seen as an important food source.     

To be or not to be? Mating success and survival trade‐offs when switching between alternative reproductive tactics

Hormones underlie the decision of assuming a territorial or a nonterritorial role, with territorial individuals usually having higher hormonal levels than nonterritorial individuals. As a territorial status is linked to higher mating opportunities, it is unclear why animals do not keep high hormonal levels and one explanation is that this would imply survival costs. We have tested this using males of the territorial damselfly Argia emma in the field. We increased juvenile hormone (JH) levels using methoprene in both territorial and nonterritorial males and predicted that: (i) males will keep (the case of territorial males) or become (the case of nonterritorial males) territorial after hormonal increase, and (ii) there will be an increase in mating success for nonterritorial males only and an impaired survival for both male tactics. Hormonally treated males remained or became territorial but had their survival impaired compared with control groups. Also, hormonally treated, ex‐nonterritorial males increased their mating success compared with the other control, nonterritorial males. The reduced survival can be explained proximally by the energy devoted either to the enhanced aggression showed during territory defence or immune function (as detected previously in damselflies). Although nonterritorial males may increase their mating success by switching to a territorial tactic, they are possibly unable to do it naturally as JH is dietary dependent and usually nonterritorial animals are in poorer condition than territorial animals.     

The spatial and temporal dynamics of deep chlorophyll layers in high-mountain lakes: effects of nutrients, grazing and herbivore nutrient recycling as growth determinants

Deep chlorophyll layers (DCL) are a common feature of oligotrophiclakes, yet the mechanisms that form and maintain them are notunderstood fully. These phytoplankton populations occur in themetalimnia of lakes where light levels are moderate to low,and where nutrient levels and zooplankton grazing pressure aredifferent than in the epilimnion. To test the importance ofnutrients and grazing pressure for algal growth in differentlake strata, microcosm experiments and monitoring were conductedin two oligotrophic lakes in the Rocky Mountains of North Americathat contain DCL. In situ microcosm experiments with naturalphytoplankton communities from three depth strata were conductedwith macronutrient additions and with and without the naturalzooplankton grazing communities. Alkaline phosphatase assaysand the in situ microcosm experiments indicated less nutrientlimitation in the metalimnia than in the epilimnia of both lakes.Zooplankton grazing in the experiments decreased algal populationgrowth rates by as much as 6% day^–1, with impacts shiftingto progressively deeper strata over the summer. Zooplanktongrazing losses, however, were partially offset by nutrient recyclingthat increased algal growth rates. Depth-differential nutrientdeficiency and zooplankton grazing and recycling interactedto maintain the DCL in these lakes.     

Natural ingestion rates and grazing impact of the marine cladoceran Penilia avirostris Dana in Tolo Harbour, Hong Kong

The gut fluorescence method was used to study algal grazingby Penilia avirostris in a semiestuarine embayment. Grazingimpact estimated from ingestion rate and animal abundance revealedthat only {small tilde}1% of the algal biomass was grazed perday by the cladocerans.     

Indirect effects of a parasite on a benthic community: an experiment with trematodes, snails and periphyton

1. Few studies have directly addressed the role played by parasites in the structure and function of ecosystems. Parasites influence the behaviour, reproduction and overall fitness of their hosts, but have been usually overlooked in community and ecosystem‐level studies. We investigated the effects of trematode parasites on snail–periphyton interactions. 2. Physa  acuta (Gastropoda: Pulmonata) snails infected with the trematode Posthodiplostomum minimum (often >30% of within‐shell biomass) grazed more rapidly than uninfected snails. Trematode effects on snail grazing indirectly affected the standing stock and community structure of periphyton. Populations of snails with 50% infected individuals reduced algal biomass by 20% more than populations with lesser (10% or 0%) infection rates. 3. The alga Cladophora glomerata dominated periphyton communities grazed by snail populations with 50% infection rates, whereas diatoms and blue–green algal taxa dominated when grazed by snail populations with lower infection rates. 4. Thus, trematodes indirectly affected periphyton communities by altering host snail behaviour, a trait‐mediated indirect effect. These results indicate that trematodes can indirectly influence benthic community structure beyond simple population fitness, with possible related effects on ecosystem function.     

Grazing dynamics on a Caribbean reef-building coral

Corals in certain Caribbean coral reef habitats are constantly grazed-on due to the territorial marking behavior of the stoplight parrotfish Sparisoma viride. We studied the grazing dynamics on the Caribbean reef-building coral Montastraea annularis. We transplanted colonies to algae-dominated reefs (Rosario Islands, Cartagena, Colombia), where they encountered higher grazing pressure. We counted grazed polyps every month throughout a year. Over the course of a year, 4,101 different grazed polyps were counted on lobe-like M. annularis transplants ( n =23). Grazing was evaluated on a monthly basis as the probabilities of all the possible transitions among four grazing categories (0%, >0–1%, >1–5%, >5% grazed tissue), uncovering a dynamic process. Higher transition probabilities were always between 0 and 1% (coral tissue grazed) grazing states, indicating that grazing did not usually exceed 1% per coral per month. The probability of remaining uninjured in a month was 0.19, 0.17 of a change from 0–1% grazing state, 0.31 of remaining at 1%, and of full recovery from 1% grazing was 0.16. More than one month was usually required for complete recovery ( P<<1) probably due to both steady grazing pressure and slow regeneration rates. Since the marking behavior of the parrotfish was not as common on other zones of the reef no comparison on the grazing among environments was possible. In spite of this, it is possible to have stable transplanted populations of corals such as M. annularis on algae-dominated Caribbean reef environments due to their tolerance to the natural grazing pressure.Communicated by: K. S. Sealey     

Energetic cost of feeding territories in an Hawaiian honeycreeper

Summary By analysis of time budgets the daily energy expenditure in territorial individuals of a Hawaiian honeycreeper (Vestiaria coccinea, Fam. Drepanididae) were estimated during the nonbreeding season and compared to that of nonterritorial individuals. The mean rise in living costs was 2.3 kcal/24 h or 17% of the nonterritorial energy budget. The most costly territorial behavior was advertisement rather than chasing, and total territorial cost was seen to be little affected by the number of intruders or the size of the territory. These results are compared with data on feeding (nonbreeding) territories of other nectar-feeding birds. The suggestion is made that hummingbirds may be more likely to develop nonbreeding territorial behavior in any set of environmental circumstances than are honeycreepers because of relatively lower total cost of advertisement plus chasing.     

Social class,socially-induced loss,recruitment and breeding of red grouse

Summary Individually marked red grouse (Lagopus lagopus scoticus) could be distinguished as territorial or non-territorial from November till the next spring. Territorial birds survived well and many reared young, but numerous nonterritorial birds died or disappeared, and none paired up or reared young. Differences in spring sex ratio between years had already been set in the previous autumn's territorial stock. Many old birds which had been territorial in earlier seasons lost territories after an annual re-distribution of territory each October–November. Thereafter very few of them were seen alive on the area over the winter, and many were found dead, whereas nearly all old birds that kept their territories were still there in spring. The number of birds in spring and the fate of individuals over winter were predicted reasonably accurately from observations of behaviour in the previous November–December. Thus, differences of social class in autumn were critical to a bird's survival to breed in the next summer and to its chances of contesting for territory and breeding in subsequent years.