SEARCH

SEARCH BY CITATION

Keywords:

  • experiment;
  • testosterone;
  • territorial behaviour;
  • unstable population dynamics

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Appendix
  • 1
    According to the ‘territorial behaviour’ hypothesis, population cycles of red grouse are caused by delayed density-dependent changes in the aggressiveness of territorial cocks. We report here on a replicated population experiment testing assumptions of this hypothesis.
  • 2
    We used testosterone implants to increase aggressiveness of cocks for 3 months during autumn, when recruitment and territory establishment take place. On two moors located in northern England, and on two 1-km2 areas within each moor, we implanted adult cocks with testosterone on an experimental area and with sham implants on a control area.
  • 3
    During the first autumn, the testosterone treatment prevented recruitment of young cocks into the territorial populations. This reduced breeding density and altered the age ratio among territorial cocks, and possibly levels of kinship. If so, the ‘kinship’ hypothesis predicted that density and recruitment should continue to differ between testosterone-treated and control areas.
  • 4
    Grouse density remained significantly lower on the experimental than on the control areas for two consecutive breeding seasons. This confirmed a strong spatial structuring within grouse populations, which prevented immigration from neighbouring higher-density areas. In the second autumn, testosterone was not implanted but the recruitment rate remained significantly lower and cock density continued to decline more on the experimental than on the control areas.
  • 5
    The results suggest that cocks continued to be aggressive and to maintain large territories for at least a year after aggressiveness was increased experimentally, and therefore that autumn aggressiveness is influenced by previous territorial contests.
  • 6
    The experiment validates key assumptions of the ‘territorial behaviour’ hypothesis for red grouse cycles. Population models in a subsequent paper demonstrate how changes in aggressiveness can cause population cycles.

Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Appendix

The mechanisms driving cyclic fluctuations in the size of populations of some species have fascinated ecologists for a long time (Stenseth 1999; Bjornstad & Grenfell 2001; Moss & Watson 2001; Berryman 2002). Density-dependent processes acting with a time delay are probable mechanisms for cycles, although there has been disagreement about the relative importance of intrinsic and extrinsic mechanisms (Lance & Lawton 1990; Stenseth, Bjornstad & Falck 1996). Extrinsic mechanisms include interactions with resources, parasites or predators (e.g. Krebs et al. 1995; Hudson et al. 1998; Turchin, Taylor & Reeves 1999; Turchin et al. 2000; Hudson et al. 2002; Korpimaki et al. 2002) and intrinsic mechanisms involve genotype selection, changes in aggressiveness or kin-facilitation (e.g. Chitty 1967; Charnov & Finnerty 1980; Krebs 1985; Lambin & Krebs 1993; Lambin & Yoccoz 1998; Moss & Watson 2001).

The debate over the importance of extrinsic and intrinsic factors is well illustrated by studies on red grouse Lagopus lagopus scoticus (Lath.) population dynamics (Lance & Lawton 1990). Red grouse populations commonly show cycles with periods of 4–10 years (Moss & Watson 2001). Studies conducted in northern England tested the hypothesis that these cycles are caused by a parasitic nematode (Trichostrongylus tenuis Eberth.). The ‘parasite hypothesis’ states that parasite-induced reduction in grouse fecundity, coupled with low parasite aggregation in the host population, would generate population cycles (Hudson 1986; Dobson & Hudson 1992; Hudson et al. 1992; Hudson et al. 1998; Hudson et al. 2002). Meanwhile, studies in north-east Scotland focused on spacing behaviour, specifically changes in the aggressiveness of territorial cocks, as a primary cause of cycles (‘territorial behaviour’ hypothesis: Watson 1985; Mountford et al. 1990; Moss, Parr & Lambin 1994; Watson et al. 1994; Moss, Watson & Parr 1996; MacColl et al. 2000; Moss & Watson 2001). Variants of the ‘territorial behaviour’ hypothesis share the premise that at high densities individuals become more aggressive, hence reducing the recruitment of young animals into the territorial population. This is thought to precipitate a population decline, reversible only when density drops low enough for aggressiveness and territorial requirements to lessen.

Variation in recruitment, the main demographic cause of population change in red grouse (Moss & Watson 1991), occurs in two stages: the number of young reared (breeding production) and the proportion of these young that become territorial (Moss et al. 1996; Moss & Watson 2001). Cocks establish territories in autumn and, despite over-winter mortality, recruitment in autumn determines subsequent spring density (Watson 1985; Mougeot et al. 2003). Aggressiveness varies during a cycle, following changes in density with a one-year time lag (Watson et al. 1994; Moss et al. 1996). A possible mechanism for these changes in behaviour is based on differential aggressiveness towards kin and non-kin and changes in the kin structure of male populations (Mountford et al. 1990; Watson et al. 1994; Hendry et al. 1997; Matthiopoulos, Moss & Lambin 1998, 2000, 2002). Territorial cocks are less aggressive towards close kin than to non-kin and favour their recruitment, so forming territory clusters of related cocks (Watson et al. 1994; MacColl et al. 2000). The ‘kinship’ hypothesis (a specific formulation of the ‘territorial behaviour’ hypothesis) states that, as population density increases, tolerance among related cocks decreases and recruitment is curtailed. Without recruitment sufficient to compensate for mortality, kin clusters disintegrate and consequent restriction of recruitment drives density into a trough.

Testosterone implants increase the aggressiveness of individual grouse (Moss et al. 1979; Watson & Parr 1981), so enabling workers to investigate how territorial behaviour affects demographic processes in natural populations (e.g. Moss et al. 1994). We increased the aggressiveness of old cocks during autumn and investigated subsequent effects on grouse density over two breeding seasons. On two moors, and on two 1-km2 study areas within each moor, we implanted the old (territorial) cocks with testosterone in one area and with sham implants in the other. The short-term (2–3 months) increase in aggressiveness induced by the testosterone in autumn t reduced grouse density in spring and summer t+ 1, demonstrating that autumn territorial behaviour regulates subsequent breeding density (Mougeot et al. 2003). A riskier prediction from the ‘kinship’ hypothesis was that the treatment in autumn t would influence territorial behaviour in autumn t+ 1 and hence breeding densities in spring t+ 2. This should work by preventing young cocks from recruiting in autumn t, to the population as a whole and to their natal kin clusters. In turn, fewer or smaller kin clusters should increase autumn aggressiveness and reduce recruitment (Watson et al. 1994; MacColl et al. 2000). Hence, the reduced recruitment of young cocks in autumn t should influence recruitment of the next cohort, in autumn t+ 1. Here, we report on the long-term effects of the testosterone treatment on density and recruitment. We did not measure kinship or kin clusters, but looked for a lagged effect of the treatment on recruitment and population change.

Materials and methods

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Appendix

experimental protocol

We conducted this experiment on two grouse moors in northern England (Catterick Moor, North Yorkshire and Moorhouse, Cumbria, hereafter referred as to moors 1 and 2, respectively). Populations on both moors were cyclic, increasing and close to their expected peaks. On each moor, two 1-km2 areas separated by 0·5 km (buffer area) were assigned a treatment randomly (C- or control area, or T- or testosterone area). Between mid-September and mid-November 2000, we caught 338 grouse (163 cocks) by lamping and netting them at night. Each bird was sexed and aged (young vs. > 1 year old) from plumage and morphology (Cramp & Simmons 1980). We refer to birds as ‘reared young’ in their summer of hatching, as ‘young adults’ until their second autumn (young recruited the previous autumn) and as ‘old adults’ thereafter. Grouse referred to as ‘adults’ are ‘young adults’ and ‘old adults’ combined. All cocks were ringed and colour-marked with patagial wing-tags, using different colours for reared young and adult cocks on each area on a moor. We implanted adult cocks, but not reared young. The implants comprised two silastic tubes (each 20 mm long, 0·62 mm inner and 0·95 mm outer diameter) filled with crystalline testosterone proprionate (T-areas) or empty (C-areas). This length of implants, determined during laboratory trials, increased testosterone plasma levels 7–10-fold for 3 months, and increased comb size close to that usual for spring (authors’ unpublished data). Implants were inserted in the chest, between skin and breast muscles, under local anaesthesia. We implanted 71 old cocks (17–18 cocks per area). Median implanting dates did not differ significantly between moors (Kruskall–Wallis test; inline image = 0·63, NS) or between areas within moors (inline image = 0·63 for moor 1 and inline image = 1·98 for moor 2, both NS). From the number of cocks caught in autumn and counted in August 2000 (see below), we estimated that 81% and 83% of cocks were caught on the T- and C-areas, respectively, and that 86% and 87% of old cocks (46% and 41% of all cocks) were implanted. All procedures were carried out under Home Office licence (PPL 80/1437).

effect of treatment on changes over time in grouse density

We counted grouse with dogs (Jenkins, Watson & Miller 1963). Each count consisted of six 1-km transects spaced regularly, such that the dog covered the whole area. Grouse were flushed, counted and sexed from their call and plumage. In summer, reared young are in broods and distinguished easily from adults by their smaller size and fresh plumage. From summer counts, we estimated adult grouse density, breeding success (number of young per hen) and breeding production (the total number of reared young). We counted grouse on each area in summer 2000 (early August, prior to the manipulation of aggressiveness), autumn 2000 (early December, about 6 weeks after treatment), summer 2001 (early August), autumn 2001 (early December) and spring 2002 (early April). All study areas were of 1-km2, so counted numbers equalled density (birds per km2). Variation in recruitment to the territorial population depends on both previous breeding production and on the proportion of reared young cocks establishing territories (Moss & Watson 1991, 2001). From the summer counts, we estimated the number of cocks available for recruitment (prior to the experiment) as the number of adult cocks plus half the number of reared young counted in a given area (sex ratio at fledging averages 1 : 1; R. Moss, unpublished data). From the December counts, we estimated numbers of cocks present after the manipulation of aggressiveness (first year) and after the autumn territorial contests (second year). Restrictions associated with foot and mouth disease (Robertson, Crowle & Hinton 2001) prevented counts in spring 2001. From previous work, we established that the densities of adult cocks and hens in summer (late July, early August) are correlated strongly with those of the previous spring (SAS GLM procedure: F1,31 = 151·9 and 138·8, R2 = 0·82 and 0·83, respectively, both P < 0·0001, n = 32 areas; authors’ unpublished data). We thus used the number of adult cocks and hens counted in August 2001 (i.e. excluding the reared young) as an estimate of relative spring cock and hen density, respectively (see also Thirgood et al. 2000).

effect of the treatment on the recruitment of young cocks

Having manipulated aggressiveness in adult cocks only, we expected to affect the recruitment of reared young cocks and hence the age structure in the subsequent territorial male populations. We therefore tested for differences between treatment and control areas in (1) changes over time in the young to old ratio of adult territorial cocks, (2) per capita recruitment rate and (3) recruitment success of a sample of radio-tagged reared young cocks.

Changes over time in young to old ratios

From September onwards, young and old adults cannot be told apart by field observation. However, we used different colour tags for reared young and adult cocks and were thus able to age samples of cocks during counts. The young to old ratio before the manipulation of aggressiveness was estimated from the number of reared young and adult cocks caught and tagged within each area, and subsequently from tagged cocks observed during counts in December 2000 and August 2001 (10–19 cocks/area/count).

Per capita recruitment rate

Recruitment rate (number of young adult cocks recruited in year t divided by the number of adult cocks in year t − 1) in each area was estimated, in the first year, from the number of young adult cocks in August 2001 (estimated from wing-tagged birds) divided by the number of adult cocks in August 2000. In spring 2002, the proportion of young adult cocks within the territorial population on each study area was estimated from a sample of 15–25 captured cocks. The per capita recruitment rate was then calculated as the number of young adult cocks in spring 2002 divided by the number of adult cocks in August 2001.

Recruitment success of reared young cocks

On moor 2, we fitted a sample of reared young cocks on the control (n = 7) and testosterone areas (n = 8) with radio-tags (TW3-necklace radio-tags, Biotrack, Wareham, Dorset, UK) before the manipulation of aggressiveness. Cocks were located regularly during the autumn, by triangulation or by flushing. They were located again in early May 2001 and in early August 2001, when the area where they were flushed was searched for chicks (proof of breeding) with a dog.

statistical analyses

We performed all statistical analyses with sas 8·01 (SAS 2001), using generalized linear models to analyse the count data (sas genmod procedure). Models for the numbers of cocks or hens per km2, and for the number of young cocks per old cock, were fitted to the data using a Poisson error distribution corrected for over-dispersion and a log-link function. Two analyses contrasted changes in grouse densities between T- and C-areas, the first in the year that aggressiveness was manipulated (August 2000–August 2001) and the second in the following year (August 2001–April 2002). In each analysis, the categorical effect of time was decomposed into two periods. The first (period 1; August–December) contrasted changes in density before vs. after autumn recruitment, while the second (period 2; December to the following spring or summer) contrasted changes in density after autumn recruitment. Explanatory factors included moor (moor 1 vs. moor 2), treatment (T- vs. C-areas), period 1, period 2 and all the interactions between these factors. We particularly tested whether changes in numbers over time differed between treatment areas (testing for period1 × treatment and period2 × treatment interactions) and whether these changes differed between replicate moors (testing for period × treatment × moor interactions) after controlling for other variables. We also tested for differences between T- and C-areas in per capita recruitment rate for each year separately: a model of the number of young adult cocks in year t was fitted using a Poisson error distribution and a log-link function, with the number of adult cocks in year t − 1 included as an offset. Models included T- or C-area, moor and their interaction between these factors as explanatory variables. We tested for differences in breeding success between T- and C-areas for each moor and year separately: models for the numbers of young per hen were fitted using a Poisson error distribution and a log-link function.

Results

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Appendix

effects of treatment on recruitment and density in the first year

In summer 2000, before implanting, breeding success did not differ significantly between the T- and C-areas on either moor (Table 1; inline image = 0·01 for moor 1 and inline image = 0·82 for moor 2, both NS). We tested for an effect of treatment on changes in cock density between August 2000, December 2000 and August 2001. The August 2000 counts included the numbers of adults and reared young (birds available for recruitment, prior to experiment) and the August 2001 counts included adults only, as a surrogate for previous spring density. Cock density decreased proportionately more on the T- than on the C-areas between August and December 2000 (Fig. 1a; Table 2; significant period1 × treatment interaction) on both moors (Table 2; non-significant period1 × treatment × moor interaction). In contrast, changes in cock density between December 2000 and August 2001 did not differ between T- and C-areas (Fig. 1a; Table 2; non-significant period2 × treatment interaction) on either moor. Increased aggressiveness thus resulted in fewer cocks by December, and relative differences in cock density between the T- and C-areas were maintained subsequently, until the following August. Changes in hen density also differed between treatment areas but apparently lagged behind changes in cock density in that proportionately more of the decline occurred in period 2 than in period 1 (Fig. 1c, Table 2). However, an analysis similar to that shown in Table 2, but including the count data for both sexes and adding the extra categorical variable sex, revealed that the effect of treatment on changes in numbers did not differ significantly between sexes (F1,11 = 0·99 and 1·05 for the sex × period1 × treat and sex × period2 × treatment interactions, respectively, both NS).

Table 1.  Breeding success (mean number of young per hen and 95% confidence interval) on the testosterone and control areas in the summers before (2000) and after the experiment (2001). Sample sizes, in brackets, refer to number of hens
YearMoor 1Moor 2
TestosteroneControlTestosteroneControl
20003·4 [2·6–4·4] (17)3·4 [2·6–4·4] (16)3·6 [2·7–4·7] (14)3·0 [2·3–3·9] (20)
20014·6 [3·6–5·8] (15)3·8 [3·1–4·6] (26)4·0 [3·2–5·1] (17)4·4 [3·8–5·1] (40)
image

Figure 1. Changes over time in the number of (a) cocks, (b) young relative to old cocks and (c) hens counted on control (○) and testosterone (•) areas on each moor. For August counts, we plotted separately the number of adult grouse (as an index of previous breeding density) and the number of adults and reared young (birds available for recruitment in autumn 2000 and 2001), the dashed vertical bars representing breeding production. Vertical grey bars indicate when old cocks were implanted.

Download figure to PowerPoint

Table 2.  Generalized linear models (Genmod procedure, SAS 2001) testing for an effect of treatment on changes in grouse numbers (see Fig. 1). Analyses were conducted separately for (1) the year aggressiveness was manipulated (from August 2000 to August 2001) and (2) the year after (from August 2001 to April 2002)
Dependent* d.f.Cocks km−2Young/old§Hens km−2 
χ2Pχ2Pχ2P
  • *

    Models were fitted to the data (numbers of cocks km−2, number of young relative to old cocks, and number of hens km−2) using a Poisson error distribution corrected for over-dispersion and a log-link function.

  • Treatment area (control vs. testosterone).

  • The categorical effect of time was decomposed into two time periods that contrasted differences before and after autumn recruitment (period 1, differences between August and December) and after autumn recruitment (period 2, differences between December and the next spring or summer). This was done by coding Period1 as 2, −1, −1- and Period2 as 0, 1, −1 for the August, December and spring/summer counts, respectively.

  • §

    The dependent variable is the number of young cocks, with number of old cocks included as an offset in the model.

  • P-values: NS: P > 0·05;

  • *

    P < 0·05;

  • **

    P < 0·01;

  • ***

    P < 0·001.

(1) August 2000–December 2000–August 2001
Moor1  0·02NS0·43NS 0·73NS
Treat1 19·32***7·68** 9·26**
Period11 26·43***5·48*20·56***
Period21  6·48**0·11NS 7·09**
Period1 × moor1  3·46NS0·05NS 5·23*
Period2 × moor1  0·00NS0·03NS 0·00NS
Treat × moor1  2·32NS1·23NS 3·58NS
Period1 × treat1  9·38**7·57** 1·82NS
Period2 × treat1  0·42NS0·00NS 4·70*
Period1 × treat × moor1  0·06NS0·14NS 0·27NS
Period2 × treat × moor1  0·12NS0·02NS 0·14NS
(2) August 2001–December 2001–April 2002
Moor1 37·07***  49·52***
Treat1151·12***  96·99***
Period11 12·18***  32·61***
Period21  6·28*   0·93NS
Period1 × moor1  3·48NS   5·51*
Period2 × moor1  0·07NS   0·46NS
Treat × moor1  3·06NS   1·81NS
Period1 × treat1  5·95*   0·79NS
Period2 × treat1  1·11NS   5·52*
Period1 × treat ×  moor1  2·84NS   0·27NS
Period2 × treat × moor1  0·10NS   4·60*

Between August and December 2000, the ratio of young to old cocks decreased threefold on the T-areas (from 1 : 1 to 1 : 3) but changed little on the C-areas, where it remained close to 1 : 1 (Fig. 1b; Table 2; significant period1 × treatment interaction). These differences in young to old ratios were maintained from December 2000 to August 2001 (Fig. 1b; Table 2; non-significant period2 × treatment interaction). On both moors, implanting adult cocks with testosterone resulted in fewer reared young cocks recruiting into the territorial populations during the autumn of implantation, and in smaller proportions of young adult cocks in the subsequent breeding populations. This effect of treatment on the recruitment of young cocks was also reflected by a significantly lower per capita recruitment rate (number of young adult cocks in summer 2001/number of adult cocks in summer 2000) on the T- than on the C-areas, on both moors (Fig. 2; moor: inline image = 1·29, NS; treatment area: inline image = 16·15, P < 0·001; interaction: inline image = 0·07, NS).

image

Figure 2. Per capita recruitment rate (young adult cocks year t/adult cocks year t − 1) according to treatment (white bars: control areas; black bars: testosterone areas) in the year aggressiveness was manipulated (2000) and in the following year (2001).

Download figure to PowerPoint

On moor 2, we followed a sample of radio-tagged reared young cocks from the C- and T-areas. The proportions of these that had died, remained alive within the area where they were marked or survived elsewhere (Table 3) did not differ significantly between the C- and T-areas in November (G-test: G= 4·27, NS), but did in the following spring (May; G= 11·17, P < 0·05) and summer (August; G= 9·78, P < 0·05). Overall, fewer reared young cocks from the T-area survived over winter and recruited within the area they originated from. All the radio-tagged cocks alive in August 2001 had chicks, and thus had successfully established a territory and bred.

Table 3.  Survival and recruitment success of radio-tagged reared young cocks from the control and testosterone areas on moor 2
Treatment area*September 2000November 2000May 2001August 2001
CTCTCTCT
  • *

    C: control area; T: testosterone area.

  • Either the bird was dead or the radio failed.

Alive within study area78636151
Alive outside study area00141222
Dead00000204
Lost00010301

effects of treatment on recruitment and density one year after the manipulation

In summer 2001, breeding success did not differ significantly between the T- and C-areas on either moor (Table 1; inline image = 1·43 for moor 1 and inline image = 0·40 for moor 2, both NS). We tested whether changes in grouse numbers continued to differ between areas a year after treatment by considering the numbers of adult and reared young grouse counted in August 2001 (birds available for recruitment), and numbers in December 2001 and April 2002. As in the previous year, cock density decreased in greater proportion on the T- than on the C-areas between August and December 2001 (Fig. 1a; Table 2: significant period1 × treatment interaction). This occurred on both moors (non-significant period1 × treatment × site interaction), although it was more pronounced on moor 1 than on moor 2 (Fig. 1a). Per capita recruitment rate of cocks was higher on moor 2 than on moor 1 and was again significantly lower on the T- than on the C-areas on both moors (Fig. 2; moor: inline image = 12·07, P < 0·001; treatment area: inline image = 6·86, P < 0·01; interaction: inline image = 2·18, NS).

As in the first year, changes in hen density apparently lagged behind changes in cock density, with a greater decline on the T- than on the C-areas between December 2001 and April 2002 (Fig. 1c; Table 2; significant period2 × treatment interaction). However, again as in the first year, this difference between sexes was not significant (F1,11 = 1·09 and 0·34 for the sex × period1 × treat and sex × period2 × treatment interactions, respectively, both NS). One-and-a-half years after treatment, both cock and hen densities were significantly lower on the T- than on the C-areas (Fig. 1a,b; Table 2).

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Appendix

The short-term increase in cock aggressiveness in autumn, induced by testosterone, reduced grouse density and recruitment for two consecutive breeding seasons. These long-lasting effects of the treatment on density and recruitment suggest that the grouse populations had, in effect, a memory of the manipulation, and that autumn aggressiveness was influenced by previous territorial contests. Below we discuss these findings and their implications for our understanding of the processes regulating grouse densities.

effects of the testosterone treatment on recruitment and subsequent density

Male red grouse implanted with testosterone exhibit typically bigger combs, are more aggressive and expand their territories to the detriment of other cocks (Watson & Parr 1981; Moss et al. 1994). We implanted old cocks in the autumn, when they attempt to maintain their territory while young cocks try to establish a territory, often near to their natal site. Increased aggressiveness of old cocks in autumn significantly reduced cock density (Table 2), presumably by evicting previously established cocks or by preventing others, particularly young cocks, from settling nearby (Watson & Parr 1981; Moss et al. 1994). Elevated testosterone levels might also reduce the cocks’ condition, increase their susceptibility to parasites or predators and might therefore increase mortality (e.g. Duckworth, Mendonca & Hill 2001; Hughes & Randolph 2001; Lindstrom et al. 2001). However, the testosterone manipulation reduced recruitment and hence the proportion of young adult cocks in the territorial population. The treatment thus worked mainly by preventing young cocks from obtaining territories, while most implanted old cocks retained their territories. The fate of the radio-tagged young cocks on moor 2 illustrated this: most young cocks from the treatment area died in winter or early spring, whereas all young cocks from the control area survived over winter, and most of them recruited and bred locally. Thus, it seems that most of the young cocks that failed to establish a territory died over winter, as found in another study (Watson 1985). Increased aggressiveness also resulted in fewer breeding grouse (both cocks and hens) in the following summer. We manipulated only cock behaviour, but found that the treatment affected breeding hen numbers similarly. This is also consistent with previous work showing that changes in breeding hen numbers are driven by changes in the number of territorial cocks (Moss et al. 1994, 1996).

longer-term differences in grouse density between treated and control areas

The most interesting result was that the manipulation of cocks’ aggressiveness in autumn reduced grouse density over two subsequent breeding seasons. In spring 2002, 1·5 years after the manipulation, overall grouse density was 66% and 67% lower on the T- than on the C-areas on moors 1 and 2, respectively. Given that the T- and C-areas were only 0·5 km apart, the findings suggest a strong structuring of grouse populations that maintained such differences in density, and prevented immigration and compensatory recruitment from neighbouring higher density areas. This is consistent with other population experiments, in which the demographic effects of manipulations conducted on small areas have been confined largely to these areas (e.g. Watson, Moss & Parr 1984; Moss et al. 1994, 1996).

In another similar population experiment, Moss et al. (1994) implanted cocks with testosterone in spring, when the territorial structure had been established. This led to a short-term decrease in population density in both sexes, which was not maintained until the following spring. This contrasts with our finding of a long-lasting effect of the testosterone treatment on grouse density. Our results suggest that the grouse populations had, in effect, a memory of the manipulation, and that cocks that were aggressive and held large territories, either because they were given testosterone implants or because they were recruited when aggression levels were high, continued to do so in the following autumn. The difference between our results and those of Moss et al. (1994) might have occurred because we manipulated aggressiveness when recruitment takes place in autumn, and changed the age-structure of subsequent territorial populations. The effective memory of the manipulation might thus be related to recruitment and to the establishment of territory boundaries in autumn.

delayed effects of the treatment on recruitment

Male recruitment continued to differ between the treatment and control areas in the autumn following the manipulation of aggressiveness. Changes in cock numbers between August 2001 and April 2002 differed significantly between the treatment and control areas, with a proportionately greater decline on the treatment areas from August to December. Per capita recruitment rate for 2002 was also significantly lower on the treatment than on the control areas.

Aggressive behaviour can be affected by factors other than hormones. Another natural determinant of aggressiveness and recruitment in grouse is kinship: territorial cocks are less aggressive towards close kin than to unrelated neighbours, and kin clusters of related cocks facilitate the recruitment of their joint offspring (Watson et al. 1994; MacColl et al. 2000). One prediction of the ‘kinship’ hypothesis, supported by empirical data (Watson et al. 1994; MacColl et al. 2000), is that a reduction in recruitment should lead to a decrease in levels of relatedness among territorial cocks and in the size of their kin clusters. This, in turn, should increase autumn aggressiveness and reduce recruitment rate. By increasing aggressiveness in adult cocks and curtailing the recruitment of young cocks in autumn 2000, we might thus have reduced levels of relatedness among territorial cocks, with fewer or smaller kin clusters on the treatment than on the control areas. If so, the kinship hypothesis predicts relatively higher levels of aggressiveness and a reduced recruitment rate on the treatment areas in the following autumn. The kin structuring of male populations might therefore constitute their ‘memory’ of the manipulation, and can explain the reduction in recruitment rate observed on the treatment areas as well as the lack of compensatory recruitment from neighbouring higher density areas. The hypothesis that increasing the aggressiveness of old cocks in autumn reduces subsequent levels of kinship, however, remains to be tested.

implications for red grouse cycles

The results support key assumptions of the ‘territorial behaviour’ hypothesis and are also consistent with predictions of the ‘kinship’ hypothesis. Population models presented in the following paper (Matthiopoulos et al. 2003), based on assumptions now verified by field observations and experiments, demonstrate how changes in aggressiveness can cause population cycles. Our work was conducted in northern England, where cyclic population declines have been associated with heavy parasite burdens (Hudson 1986; Dobson & Hudson 1992; Hudson et al. 1992; Hudson, Dobson & Newborn 1998; Hudson et al. 2002) but where the role of territorial behaviour has been little studied. Clearly, parasites and territorial behaviour can act together, with parasites reducing breeding success and the number of grouse available for recruitment, and territorial behaviour regulating the recruitment of reared young. In this experiment we did not manipulate parasite burdens directly, and parasites could well have contributed to the observed changes in numbers. However, we found no significant difference in breeding success between the treatment and control areas and so differences between them were unlikely to be due to parasites affecting breeding production. Parasites can limit aggressiveness (Fox & Hudson 2001) and androgens, by impairing immune function, might increase parasite infections (Folstad & Karter 1992). In our study, however, any effects of testosterone implants upon parasite establishment should have been restricted to the animals implanted and to the duration of the implants. Future studies could investigate how parasites and aggressiveness interact in red grouse, in order to better understand the relative roles of territorial behaviour and parasites in regulating grouse populations.

Acknowledgements

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Appendix

We are grateful to Major T.P.J. Helps from the Catterick Garrison army training facility for allowing us to conduct the work on Catterick Moor, and to the personnel of Wathgill Camp for logistical support. We are also grateful to English Nature for access to Moorhouse Nature Reserve and to C. McCarthy and J. Adamson for their help with organizing the work there. We thank R. Cox, N. Green, F. Leckie and D. Luccini for their help with the fieldwork and D. Elston for his help with statistics. We also thank S. Albon for helpful comments on previous drafts. This work was funded by a NERC grant (NER/A/S/1999/00074).

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Appendix
  • Berryman, A.A. (2002) Population cycles: causes and analysis. Population Cycles: the Case for Trophic Interactions (ed. A. A.Berryman), pp. 328. Oxford University Press, Oxford.
  • Bjornstad, O.N. & Grenfell, B.T. (2001) Noisy clockwork: time series analysis of population fluctuations in animals. Science, 293, 638643.
  • Charnov, E.L. & Finnerty, J.P. (1980) Vole population cycles: a case for kin selection. Oecologia, 45, 12.
  • Chitty, D. (1967) The natural selection of self-regulatory behaviour in animal populations. Proceedings of the Ecological Society of Australia, 2, 5178.
  • Cramp, S. & Simmons, K.E.L. (1980) The Birds of the Western Palearctic, Vol. 2. Oxford University Press, Oxford.
  • Dobson, A.P. & Hudson, P.J. (1992) Regulation and stability of a free-living host–parasite system −Trichostrongylus tenuis in red grouse. II. Population models. Journal of Animal Ecology, 61, 487498.
  • Duckworth, R.A., Mendonca, M.T. & Hill, G.E. (2001) A condition dependent link between testosterone and disease resistance in the house finch. Proceedings of the Royal Society of London Series B, Biology Sciences, 268, 24672472.
  • Folstad, I. & Karter, A.J. (1992) Parasites, bright males, and the immunocompetence handicap. American Naturalist, 139, 603622.
  • Fox, A. & Hudson, P.J. (2001) Parasites reduce territorial behaviour in red grouse (Lagopus lagopus scoticus). Ecology Letters, 4, 139143.
  • Hendry, R., Bacon, P.J., Moss, R., Palmer, S.C.F. & McGlade, J. (1997) A two-dimensional individual-based model of territorial behaviour: possible population consequences of kinship in red grouse. Ecological Modelling, 105, 2339.
  • Hudson, P.J. (1986) The effect of a parasitic nematode on the breeding production of red grouse. Journal of Animal Ecology, 55, 8592.
  • Hudson, P.J., Dobson, A.P., Cattadori, I.M., Newborn, D., Haydon, D.T., Shaw, D.J., Benton, T. & Grenfell, B.T. (2002) Trophic interactions and population growth rates: describing patterns and identifying mechanisms. Philosophical Transactions of the Royal Society of London Series B, Biology Sciences, 357, 12591271.
  • Hudson, P.J., Dobson, A.P. & Newborn, D. (1998) Prevention of population cycles by parasite removal. Science, 282, 22562258.
  • Hudson, P.J., Newborn, D. & Dobson, A.P. (1992) Regulation and stability of a free-living host-parasite system −Trichostrongylus tenuis in red grouse. I. Monitoring and parasite reduction experiments. Journal of Animal Ecology, 61, 477486.
  • Hughes, V.L. & Randolph, S.E. (2001) Testosterone increases the transmission potential of tick-borne parasites. Parasitology, 123, 365371.
  • Jenkins, D., Watson, A. & Miller, G.R. (1963) Population studies on red grouse, Lagopus lagopus scoticus (Lath.), in north-east Scotland. Journal of Animal Ecology, 32, 317376.
  • Korpimaki, E., Norrdahl, K., Klemola, T., Pettersen, T. & Stenseth, N.C. (2002) Dynamic effects of predators on cyclic voles: field experimentation and model extrapolation. Proceedings of the Royal Society of London Series B, Biology Sciences, 269, 991997.
  • Krebs, C.J. (1985) Do changes in spacing behaviour drive population cycles in small mammals? Behavioural Ecology (eds M.R.Sibly & R.H.Smith), pp. 295312. Blackwell Scientific Publications, Oxford.
  • Krebs, C.J., Boutin, S., Boonstra, R., Sinclair, A.R.E., Smith, J.N.M., Dale, M.R.T., Martin, K. & Turkington, R. (1995) Impact of food and predation on the snowshoe hare cycle. Science, 269, 11121115.
  • Lambin, X. & Krebs, C.J. (1993) Influence of female relatedness on the demography of Townsend's vole populations in spring. Journal of Animal Ecology, 62, 536550.
  • Lambin, X. & Yoccoz, N.G. (1998) The impact of population kin-structure on nestling survival in Townsend's voles, Microtus townsendii. Journal of Animal Ecology, 67, 116.
  • Lance, A.N. & Lawton, J. (1990) Red Grouse Population Processes. Royal Society for the Protection of Birds, Sandy, UK.
  • Lindstrom, K.M., Krakower, D., Lundstrom, J.O. & Silverin, B. (2001) The effects of testosterone on a viral infection in greenfinches (Carduelis chloris): an experimental test of the immunocompetence–handicap hypothesis. Proceedings of the Royal Society of London Series B, Biology Sciences, 268, 207211.
  • MacColl, A.D.C., Piertney, S.B., Moss, R. & Lambin, X. (2000) Spatial arrangement of kin affects recruitment success in young male red grouse. Oikos, 90, 261270.
  • Matthiopoulos, J., Moss, R. & Lambin, X. (1998) Models of red grouse cycles. A family affair? Oikos, 82, 574590.
  • Matthiopoulos, J., Moss, R. & Lambin, X. (2000) The kin-facilitation hypothesis for red grouse population cycles: territory sharing between relatives. Ecological Modelling, 127, 5363.
  • Matthiopoulos, J., Moss, R. & Lambin, X. (2002) The kin facilitation hypothesis for red grouse population cycles: territorial dynamics of the family cluster. Ecological Modelling, 147, 291307.
  • Matthiopoulos, J., Moss, R., Mougeot, F., Lambin, X. & Redpath, S.M. (2003) Territorial behaviour and population dynamics in red grouse Lagopus lagopus scoticus. II. Population models. Journal of Animal Ecology, 72, 10831096.
  • Moss, R., Kolb, H.H., Marquiss, M., Watson, A., Treca, B., Watt, D. & Glennie, W. (1979) Aggressiveness and dominance in captive cock red grouse. Aggressive Behavior, 5, 5884.
  • Moss, R., Parr, R. & Lambin, X. (1994) Effects of testosterone on breeding density, breeding success and survival of red grouse. Proceedings of the Royal Society of London Series B, Biology Sciences, 258, 175180.
  • Moss, R. & Watson, A. (1991) Population cycles and kin selection in red grouse Lagopus lagopus scoticus. Ibis, 133, 113120.
  • Moss, R. & Watson, A. (2001) Population cycles in birds of the grouse family (Tetraonidae). Advances in Ecological Research, 32, 53111.
  • Moss, R., Watson, A. & Parr, R. (1996) Experimental prevention of a population cycle in red grouse. Ecology, 77, 15121530.
  • Mougeot, F., Redpath, S.M., Leckie, F. & Hudson, P.J. (2003) The effect of aggressiveness on the population dynamics of a territorial bird. Nature, 421, 737739.
  • Mountford, M.D., Watson, A., Moss, R., Parr, R. & Rothery, P. (1990) Land inheritance and population cycles of red grouse. Red Grouse Population Processes (eds A. N.Lance & J. H.Lawton), pp. 7883. Royal Society for the Protection of Birds, Sandy, UK.
  • Robertson, H.J., Crowle, A. & Hinton, G. (2001) Interim Assessment of the Effects of the Foot and Mouth Disease Outbreak on England's Biodiversity. English Nature, Peterborough.
  • SAS (2001) SAS / STAT User's Guide, Version 8.01. SAS Insitute Inc., Cary, NC.
  • Stenseth, N.C. (1999) Population cycles in voles and lemmings: density dependence and phase dependence in a stochastic world. Oikos, 87, 427461.
  • Stenseth, N.C., Bjornstad, O.N. & Falck, W. (1996) Is spacing behaviour coupled with predation causing the microtine density cycle? A synthesis of current process-oriented and pattern-oriented studies. Proceedings of the Royal Society of London Series B, Biology Sciences, 263, 14231435.
  • Thirgood, S.J., Redpath, S.M., Rothery, P. & Aebischer, N.J. (2000) Raptor predation and population limitation in red grouse. Journal of Animal Ecology, 69, 504516.
  • Turchin, P., Oksanen, L., Ekerholm, P., Oksanen, T. & Henttonen, H. (2000) Are lemmings prey or predators? Nature, 405, 562565.
  • Turchin, P., Taylor, A.D. & Reeve, J.D. (1999) Dynamical role of predators in population cycles of a forest insect: an experimental test. Science, 285, 10681071.
  • Watson, A. (1985) Social-class, socially-induced loss, recruitment and breeding of red grouse. Oecologia, 67, 493498.
  • Watson, A., Moss, R. & Parr, R. (1984) Effects of food enrichment on numbers and spacing behavior of red grouse. Journal of Animal Ecology, 53, 663678.
  • Watson, A., Moss, R., Parr, R., Mountford, M.D. & Rothery, P. (1994) Kin landownership, differential aggression between kin and non-kin, and population fluctuations in red grouse. Journal of Animal Ecology, 63, 3950.
  • Watson, A. & Parr, R. (1981) Hormone implants affecting territory size and aggressive and sexual behavior in red grouse. Ornis Scandinavica, 12, 5561.

Appendix

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Appendix

Appendix I

cock numbers observed and postdicted by the kinship model in the second year

Matthiopoulos et al. (1998) developed a simple age-structured model for red grouse cycles which focuses on changes in cock numbers and has space as a limiting resource. We investigate whether this model is consistent with the present results. This model can be used to predict cock density in year t+ 1 given the following for the year t:

  • Ot: number of old adult cocks

  • Yt: number of young adult cocks (recruited in the previous year)

  • Rt: number of reared young cocks

  • s: yearly, per capita survival rate of cocks

  • b: breeding success (male offspring only)

  • A: available area in study site

  • at: minimum territory size required for the establishment of cocks.

Territory size is related closely to aggressiveness and so it is also reasonable to call at‘aggressiveness’, as explained fully in the accompanying paper (Matthiopoulos et al. 2003). The question we asked of the model was: what changes in aggressiveness (at) are needed to explain the changes in cock density observed on the testosterone and control areas 1 year after the manipulation (from August 2001 to April 2002)? The available area (A) was constant (1-km2 study areas) and Ot, Yt and Rt were all measured for each area and each population during the August 2001 counts. However, survival (s) was not known. We used average s-values of 0·5 and allowed for it to vary between 0·3 and 0·7 (the range of survival rates observed in natural populations). We investigated which value of at would best postdict the number of cocks counted on each area in April 2002 (Table A1). Varying survival had relatively little effect on postdicted cock densities, which depended mainly on at. The results show that for the model to postdict cock density correctly, the aggressiveness of cocks (at) needed to be higher on the T- than on the C-areas on both moor (Table A1). We did not verify this conclusion quantitatively by measuring aggressiveness and so conclude only that the model is qualitatively consistent with our results.

Table A1.  Numbers of cocks observed and postdicted by the kinship model (see Matthiopoulos et al. 1998) on the control and testosterone areas according to varying levels of territory requirements or aggressiveness (at)
Moor Treatment area*12
CTCT
  • C: control area; T: testosterone area.

  • Numbers of cocks counted in April 2002.

  • Postdicted numbers are those obtained for a survival of 0·5. Those highlighted in bold are closest to observed numbers. Numbers in bracket refers to the range of postdicted numbers obtained for survival values ranging from 0·3 to 0·7.

Observed581710532
Postdicted    
at =0·01  106 [98–112]41 [39–43]
0·05  102 [96–106]38 [38–39]
0·161 [60–63]  98 [95–99]35 [34–36]
0·1560 [57–61]  33 [29–35]
0·258 [57–58]  30 [27–33]
0·355 [54–55]  28 [25–31]
0·5 26 [24–28]  
0·75 22 [21–25]  
1 20 [18–21]  
1·5 17 [16–18]  
1·75 16 [16–17]  
2 15 [15–16]