Dr Robbie McDonald, School of Biological Sciences, University of Bristol, Woodland Road, Bristol BS8 1UG, UK. Tel.: 0117 928 8918, Fax: 0117 925 7374, E-mail: email@example.com
1. Trapping and hunting records are frequently used as an index of animal abundance. This study demonstrates that these records can be misleading if sampling effort is not controlled for.
2. Mean numbers of stoats Mustela erminea and weasels M. nivalis trapped by British gamekeepers have been decreasing since 1975 and 1961 respectively, giving rise to concern that populations of both species may be declining. However, trapping effort has not been quantified over this period.
3. A total of 203 gamekeepers in England were questioned about the trapping effort they expended and the number of stoats and weasels they trapped in 1997. The most significant factor affecting the number of stoats and weasels trapped was trapping effort.
4. Gamekeepers that relied on hand-rearing game birds for shooting regarded stoats and weasels as a less serious problem, and made substantially less trapping effort, than gamekeepers that relied on wild game birds.
5. The national decline in the numbers of stoats and weasels trapped may be the result of a decline in stoat and weasel populations. However, the decline is equally consistent with a reduction in trapping effort, corresponding to a national increase in reliance on hand-rearing game birds for shooting.
6. When the effect of trapping effort was controlled for, the number of weasels trapped by gamekeepers in 1997 was significantly lower in the south-west than in other regions of England and was unusually low in some local areas.
7. Trapping records can be used effectively to monitor populations of stoats and weasels, as long as gamekeepers record the number of traps set in each month and monthly totals of animals killed. Ideally, the sex of each animal and whether it was trapped or shot should also be recorded. Similar modifications should also be made to other wildlife monitoring schemes based on trapping and hunting records.
Hunting and trapping records are frequently used as an index of abundance for mammals and birds (e.g. Strandgaard 1964; Leeuwenberg & Hepburn 1982; Tapper 1992). Such records do not necessarily reflect the abundance of the species and it is important to consider sources of variation that are not related to animal numbers (Leeuwenberg & Hepburn 1982; Myrberget 1988). For example, fur trapping records may be affected by market prices, predator control records by bounty policies, and hunting records by bag limits or even weather (Latham 1952; Brand & Keith 1979; Landry 1982). In most studies of hunting and trapping records, authors recommend that sampling effort should be controlled for when considering the relationship between the record and the actual abundance of the animal. However, few studies have previously been able to quantify sampling effort directly.
The National Game Bag Census reports that the mean bag of weasels has been declining since 1961 (Tapper 1992). The mean bag of stoats increased from 1961 to 1975 but has declined slightly since. This is surprising since bags of stoats are generally closely related to bags of their main prey, rabbits Oryctolagus cuniculus L. (King 1980a; Tapper 1992). Rabbit bags have been increasing dramatically in recent years, following the recovery of rabbit populations from myxomatosis epidemics in the 1950s and 1960s (Tapper 1992). So, stoat bags do not appear to reflect the expected numerical response of stoat populations to increasing abundance of their main prey (King 1983). The decline of weasel bags and the lack of an increase in stoat bags has given rise to concern that stoat and weasel populations may be declining (Harris et al. 1995). However, Tapper (1992) and Macdonald, Mace & Rushton (1998) suggest that one of the reasons why bags may have declined is that the trapping effort made by gamekeepers may also have declined. Clearly, it should be determined whether changes in trapping effort could bring about the apparent decline in bags, in order to judge whether the species are genuinely declining.
In contrast to game animals and fur-bearing species in other countries, there is no statutory regulation of stoat and weasel trapping in Britain and there are no standard management guidelines or bag limits. Therefore, variation in trapping effort is determined by the management policies of the estate and the attitude of the gamekeeper. In this study we quantify the trapping effort made by gamekeepers in England in 1997, relate this to their trapping records of stoats and weasels and assess the degree to which effort would have had to have changed in order to bring about the apparent decline in bags. In order to develop the collation of trapping records into an effective scheme for monitoring national stoat and weasel populations, we highlight additional data requirements and make recommendations for future work.
A questionnaire survey of the complete membership of the National Gamekeepers’ Organization was conducted in April 1998. We requested information about: the area of land they worked on, the main types of game they aimed to promote, the number of stoats and weasels killed by them in 1997, the control methods they employed, the number of traps they set, the months in which traps were set and their predicted response to increased predation by stoats and weasels. We also asked them to rate the seriousness with which they regarded the problems caused by stoats and weasels on a scale of 1 (no problem) to 4 (major problem). Gamekeepers use the same type of trap for both species and set them in locations where all small predators are likely to be caught. There is therefore no difference between the trapping effort made to catch stoats and weasels. The questionnaire was accompanied by a letter from the chairman of the organization recommending support. The form and letter made it clear that the forms should be completed even if the gamekeeper did not attempt to control stoats and weasels or they were rare in their area. A prepaid envelope was provided. Respondents had the choice of providing their name and address or completing the forms anonymously, though all respondents were asked to name the county in which they worked.
The unit of analysis for this survey was the individual gamekeeper. Most gamekeepers have responsibility for the management of an area of land known as their ‘beat’. Many estates employ head gamekeepers who have overall responsibility for the game management of the whole estate, which may comprise several beats, but in most cases head gamekeepers also have their own beat. In other cases, ‘underkeepers’ have joint responsibility for a beat with another individual. Respondents were asked to complete the form with reference to their own beat and their own records only. If it appeared that two gamekeepers had replied in reference to the same beat, only one reply was considered in further analysis. In the case of a response where it was not possible to determine whether it referred to an individual's own beat or to the whole estate, that response was excluded from further analysis.
Gamekeepers were asked to classify themselves into three shoot types, according to the main game bird species they aimed to promote on their beat. Gamekeepers focusing on reared game birds, mainly pheasant Phasianus colchicus L. and red-legged partridge Alectoris rufa L. were described as ‘reared pheasant’, those focusing on wild grey partridge Perdix perdix L. and wild pheasant were described as ‘wild partridge’ and those focusing on wild red grouse Lagopus lagopus scoticus Latham were described as ‘wild grouse’. Where it was not possible to determine the main shoot type – for example, if large numbers of both reared pheasant and wild red grouse were shot – the respondent was excluded from further analysis. Respondents were classified into the seven regions of England used by the National Game Bag Census: south-west, south-east, east, east midlands, west midlands, north-west and north-east (Tapper 1992).
Trapping effort was described by ‘trap months’, i.e. the number of traps set multiplied by the number of months during which they were set. Beat area, bag data and trapping effort data were log10(n + 1) transformed. Data were checked to ensure that they conformed to the assumptions of the statistical tests used (Hair et al. 1995; Zar 1996). Multivariate analysis of covariance (mancova) (Hair et al. 1995) was used to test the variation in both stoat and weasel bags between shoot types, while controlling for variation in trapping effort, which was the same for both species, and beat area. Correlation between these two covariates was not important for the dependent variables because F ratios were calculated using adjusted mean squares. Interaction terms were included in the model initially, but were removed if found to be insignificant and the analysis was rerun. There is extensive colinearity between region and shoot type, since red grouse only occur in upland habitats which are predominantly found in the north of England. Therefore, regional variation in stoat and weasel bags was analysed using a similar model including gamekeepers working on only one shoot type, reared pheasant, which had a countrywide distribution. Unplanned multiple comparisons of group means from anovas were made using Tukey's test with Kramer's Modification for unequal sample sizes (Day & Quinn 1989). Adjusted group means, corrected for covariates, of significant factors from ancovas were compared using the Bryant–Paulson–Tukey test with Kramer's modification for unequal sample sizes (Bryant & Paulson 1976; Day & Quinn 1989). The responses to questions for which a categorical or ordinal response was requested were analysed using non-parametric statistics (Siegel & Castellan 1988; Zar 1996). Following a significant result from a Kruskal–Wallis anova, non-parametric unplanned multiple comparisons were made using Dunn's Q-test (Zar 1996).
To test whether the observed trends in stoat and weasel bags between 1976 and 1996 could be the consequence of changes in trapping effort, we calculated the degree to which trapping effort would have to have changed over this period. We assumed that the observed relationship between stoat and weasel bags and trapping effort in 1997 was applicable to the mean stoat and weasel bags recorded in 1976 and 1996, and solved this relationship for the effort required to produce these bags.
Data were entered into a Microsoft Access database and statistical treatments were applied using Minitab for Windows (Ryan & Joiner 1994) or by customised Microsoft Excel spreadsheets.
In total, 1100 questionnaire forms were sent out and 215 (20%) were returned in time for analysis. Of these, 203 (94%) were in England, could be classified into shoot types and were included in further analyses (Table 1). Of the remaining 12 respondents, six were in Scotland, five were in Wales and one could not be classified into a shoot type. 172 (85%) provided their full names and addresses and 31 (15%) remained anonymous but provided the county in which they worked (Fig. 1). The total area of coverage of this survey was approximately 1760 km2, which constitutes 1·4% of the total land area of England.
Table 1. Shoot types and control methods used for stoats and weasels in 1997 by gamekeepers responding to the questionnaire
Stoats Trapping + shooting
Weasels Trapping + shooting
Full details of their monthly trapping regime, beat area and bags of stoats and weasels in 1997 were provided by 153 respondents (Table 2). Stoat and weasel bags were significantly affected by trapping effort but not by shoot type or beat area (Figs 2 and 3; Table 3). Bags on reared pheasant shoots varied significantly between regions and were significantly affected by trapping effort but not beat area (Table 4) (Fig. 4). Weasel bags in the south-west were significantly lower than in all other regions, except the north-west, where the sample size was insufficient to make statistical comparison (Fig. 4). Stoat bags did not vary significantly between regions.
Table 2. Summary of trapping effort, beat area and stoat and weasel bags in 1997, for gamekeepers supplying the details of all variables
Mean (95% CI)
Mean (95% CI)
Beat area (ha)
Mean (95% CI)
Number of traps set per month Mean (95% CI)
Mean (95% CI)
Table 3. mancova of stoat and weasel bags (n = 153) between shoot types with trapping effort and beat area as covariates. Bag size, beat area and trapping effort data were log10 transformed. Interaction terms were included in the model initially but were removed if not significant
Univariate ancovas Source of variation
Coefficient of covariate (SD)
Shoot type Constant Error Total
2 148 152
Shoot type Constant Error Total
2 148 152
Table 4. mancova of stoat and weasel bags on reared pheasant shoots (n = 121) between regions with trapping effort and beat area as covariates. Bag size, beat area and trapping effort data were log10 transformed. Interaction terms were included in the model initially but were removed if not significant
Univariate ancovas Source of variation
Coefficient of covariate (SD)
Region Constant Error Total
6 112 120
Region Constant Error Total
6 112 120
For a constant area on all shoot types in all regions, stoat bag ∝ effort0·5077 and weasel bag ∝ effort0·5213 (Table 3). Thus, effort = a× stoat bag1·9697 and effort = b× weasel bag1·9183 where a and b are constants. Therefore, in order to bring about a decline in the National Game Bag Census mean stoat bag from 2·40 in 1976 to 1·75 in 1996 (Tapper 1992; S. Tapper, personal communication), trapping effort would have had to have decreased from 5·61a to 3·01a, which is a decline of approximately 46%. In order to bring about a decline in the mean weasel bag from 1·79 in 1976 to 0·86 in 1996, trapping effort would have had to have decreased from 3·06b to 0·75b, which is a decline of approximately 75%.
Trapping effort varied significantly between gamekeepers working on different shoot types (anova, F2,150 = 15·32, P < 0·001). There was no significant difference in the trapping effort made by gamekeepers on wild partridge and wild grouse shoots, but both groups made a significantly greater effort than those working on reared pheasant shoots. The degree to which stoats and weasels were regarded as a problem by gamekeepers varied significantly between shoot types (Kruskal–Wallis anova, stoats, H = 30·1, d_f. = 2, P < 0·001; weasels, H = 31·1, d.f. = 2, P < 0·001). Stoats and weasels were regarded as an equally serious problem by gamekeepers on wild grouse and wild partridge shoots, but were regarded as a more serious problem on both wild grouse and wild partridge shoots than on reared pheasant shoots.
There was no significant difference in the prevalence of trapping in the four seasons between wild partridge and wild grouse shoots (χ2 = 0·32, d.f. = 3, P > 0·1). However, the prevalence of trapping in the four seasons differed significantly between gamekeepers on the wild game bird shoots combined and gamekeepers working on reared pheasant shoots (χ2 = 11·16, d.f. = 3, P < 0·05). Winter trapping was significantly more prevalent than expected on wild game bird shoots than on reared pheasant shoots (χ2-test with partitioned degrees of freedom, χ2 = 8·20, d.f. = 1, P < 0·01) (Fig. 5).
There was no significant difference in gamekeepers’ response to increased predation by stoats and weasels between the three shoot types (χ2 = 0·31, d.f. = 2, P > 0·1). Most gamekeepers (74%) reported that they would increase their trapping effort in response to increased predation by stoats and weasels. Only one gamekeeper (0·5%) reported that he would increase the number of birds reared and released and one gamekeeper reported that he would concentrate trapping effort into a smaller area in response to increased predation. The remaining respondents (25%) would not change their practice at all.
This is the first large-scale survey that quantifies gamekeeper trapping effort and relates this to stoat and weasel trapping records. A total of 203 gamekeepers was surveyed, giving widespread coverage of over 1% of the area of England. In common with other postal questionnaire surveys (Landry 1982; Macdonald & Johnson 1996; Packer & Birks 1999), it was not possible to determine whether this was a representative sample of gamekeepers. It is possible that this survey favoured gamekeepers who have an interest in stoats and weasels because they present a particular problem to their activities or because they appear particularly common in their area. However, the respondents represented a wide range of situations, from those who conducted no trapping at all to those who trapped intensively, and from those in areas where stoats and weasels were considered abundant to those in areas where stoats and weasels were considered absent. Furthermore, it is likely that the respondents to this survey are also contributors to the National Game Bag Census and, since the aim of the current study is to examine the declines exhibited by the National Game Bag Census in greater detail, the two samples may be directly comparable (S. Tapper, personal communication). Therefore, we are confident that, for these purposes, this is a representative sample of gamekeepers in England.
Stoat and weasel bags in 1997 were strongly affected by trapping effort. When the effect of effort is controlled for, apparent group differences, such as those between different shoot types, were not significant. This clearly supports the importance of an explicit consideration of effort in the analyses of large-scale trapping records or similar indices of animal abundance. Earlier studies (Middleton 1934; Jefferies & Pendlebury 1968; Hewson 1972; Tapper 1979, 1982) did not correct stoat and weasel bags for effort but assumed that it was constant between years in the local areas they sampled. In contrast, King (1980b, 1983), Erlinge (1983) and King et al. (1996) did correct records for effort by calculating stoat captures per 100 trap nights but only in small, intensively monitored areas. The National Game Bag Census has not assumed that effort is consistent between years but has not corrected bags for trapping effort (Tapper 1992). The additional collection of trapping effort data as part of this and similar schemes would greatly improve their ability to relate trends in bags to trends in animal populations.
The National Game Bag Census has recorded declines in stoat and weasel bags since 1975 and 1961 respectively (Tapper 1992). These trends have been interpreted as evidence for declines in national stoat and weasel populations (Harris et al. 1995). However, we believe that the decline in national stoat and weasel bags may be due largely to a reduction in trapping effort. Since 1961 the number of game birds reared by gamekeepers has increased substantially and the rate of increase has been especially marked since the mid-1970s (Tapper 1992). In 1997, approximately 250 pheasants were reared and released per km2 on British game estates, whereas in 1960 the average was approximately 50 (Tapper 1992; S. Tapper, personal communication). This suggests that the proportion of gamekeepers relying on reared game birds, and the relative importance of reared birds over wild birds to individual gamekeepers have both increased dramatically during this period. In 1997, more than 80% of gamekeepers relied on reared game birds. These gamekeepers made 60–75% less trapping effort than gamekeepers that relied on wild game birds and, as a result, caught substantially fewer stoats and weasels. Based on stoat and weasel capture rates in 1997, trapping effort would have had to have decreased by 45–75% to bring about the decline in stoat and weasel bags recorded by the National Game Bag Census between 1976 and 1996. A decline in trapping effort of this scale is consistent with a national change from dependence on wild game birds to reared game birds during the same period. Therefore, there is little evidence from game bag data to suggest that national stoat and weasel populations have changed beyond normal annual fluctuations, such as those apparent in long-term local records (Middleton 1934; King 1980a; Tapper 1992). Local and regional changes in stoat and weasel populations cannot be ruled out by this survey. Weasel bags were particularly low in the south-west and several gamekeepers recorded unusually low weasel bags, despite a reasonable level of trapping effort being made. These patterns are hard to interpret, may give cause for concern, and should be investigated further.
In addition to trapping effort, there are other considerations that should be made when using game bag data to monitor stoat and weasel populations; immigration, sex bias, seasonal bias and shot animals. Tapper, Green & Rands (1982) and King (1980b) state that, with intensive trapping effort, resident stoats and weasels are rapidly removed following the start of a trapping campaign and subsequently the majority of the animals trapped are immigrants. Therefore, it is desirable to determine the area over which trapping regimes of varying intensities affect stoat and weasel populations, in order to account for the contribution of immigrants to the total bag. Female stoats and weasels have smaller ranges and lower dispersal distances than males (King 1975a; Erlinge 1977; King & McMillan 1982; Erlinge & Sandell 1986; Robitaille & Raymond 1995). Most intensively studied samples of trapped mustelids exhibit a sex bias in favour of males and have related this to differences in the ranging behaviour of the two sexes (King 1975b; King 1980b; Buskirk & Lindstedt 1989). While the sexes are readily distinguished by experienced observers, gamekeepers are not currently asked to record the numbers of males and females killed. Therefore, no comparison can be made between bags of females, which are more likely to be residents, and bags of males, which are more likely to be immigrants. In addition to variation in trapping effort between seasons, there is a seasonal bias in annual bag size resulting from varying capture rates associated with periods of increased activity, such as juvenile dispersal or mate searching (Erlinge & Sandell 1986), or reduced activity, such as pregnancy. Ideally, monthly variation in capture rates per unit effort should also be quantified. Most gamekeepers in this survey shot stoats and weasels in addition to those they trapped. Tapper (1992) suggests that shot animals account for a significant proportion of the total bag. However, the importance of shot animals has not been quantified either by this survey or the National Game Bag Census.
Gamekeepers are responsible for the management of a large part of the land area of Great Britain (Tapper 1992) and are the only group involved with the management of stoats and weasels. Gamekeepers’ records are a valuable source of data on stoat and weasel populations that are otherwise very hard to monitor (King & Edgar 1977; Macdonald, Mace & Rushton 1998). However, these records can be misleading if information on sampling effort is not available. Game bags are suitable for development into a national monitoring scheme for stoat and weasel populations, but only if gamekeepers record the number of traps set in each month and monthly totals of animals killed. Ideally, the sex of each animal and whether it was trapped or shot should also be recorded.
We are grateful to the Wingate Foundation and the Dulverton Trust for financial support, the National Gamekeepers’ Organization and their chairman David Clark and membership secretary Eddy Graves for supporting this study, Stephen Tapper of the Game Conservancy Trust for unpublished information from the National Game Bag Census and Johnny Birks of the Vincent Wildlife Trust for information about a similar polecat survey. We also thank the many gamekeepers who replied to the survey and provided such detailed information about their work. The questionnaire and this paper were improved by comments from and discussions with Johnny Birks, Jake Fiennes, Carolyn King, Stephen Tapper, Nancy Vaughan, Lindsay Waddell and two anonymous referees. The map was produced using dmap.