An appraisal of: Harris (2021) A review of the animal welfare, public health, and environmental, ecological and conservation implications of rearing, releasing and shooting non-native gamebirds in Britain. Report to the Labour Animal Welfare Society.
Last year, GWCT research were commissioned by Natural England to review the ecological effects of releasing gamebirds and have published on this subject area widely in the scientific literature in recent decades (see Madden & Sage 2021; Sage et al. 2020). As such, we were approached by BASC to consider whether this report produced by Stephen Harris provides a fair and balanced assessment of the relevant scientific evidence.
This appraisal relates to the ecological, environmental and conservation considerations in Harris (2021) only – we (the authors) are not sufficiently familiar with the literature on welfare or public health to investigate those aspects of this report in a reasonable timescale.
A scientific literature review is required to provide the audience with ‘a balanced overview of the existing knowledge on the topic under question’ or ‘a critical account of what has been published’. A systematic review focuses on a specific research question, and is required to ‘identify, appraise and synthesize all good relevant research evidence’. Systematic reviews are the best way to avoid bias and this approach is now usually required for publication in peer-reviewed journals.
Harris (2021) does not pretend to be a systematic review, quite the opposite. The last key point on the first page (Instructions) states: ‘It is my decision as to what evidence I wish to include in my review.’ At the same time, elsewhere the report does claim to be a scientific review. On the first page it says: ‘I was asked to review the scientific evidence on…’ and so on.
Our task is to consider whether the report meets the most basic requirements of a scientific review, i.e. that it is not biased by a process of selecting and ignoring references to suit a particular argument (cherry picking), and/or by misinterpreting selected references.
In this appraisal, we give an account of the report through three examples. In each example, we examine:
- the way references are cited in the text,
- the selection and exclusion of references, and
- the interpretation of what the references say and whether this is accurate.
The overall conclusion of this process is that the report is not a scientific review but projects the image of being so. We have found that Harris (2021) both cherry picks the evidence and misrepresents those that he cites in the three examples considered here. The result is that the headline statements in the report look like they are scientifically supported when, in fact, they are not.
It is reasonable to assume that the same approach is taken in other sections of the report, rendering it an opinion piece, in which many of the ‘findings’ are influenced by the views of the author, while others do have some evidence around them. However, there are at least several occasions where the science has been misinterpreted and it appears that efforts are made to cover this by citing difficult-to-access references and by taking speculative ideas and reporting them as actual findings.
Example 1: Releasing and foxes
A key part of the report and Harris’ overall argument is about the scale of gamebird releasing, the effect this has on foxes and on predation. Initially, he makes a reasonable assessment of gamebird numbers and biomass during the year (pages 17 -22). To provide context, comparisons are made with biomass and numbers of other birds and not with, for example, free-range livestock.
However, the effect of these releases on foxes and predation is not reasonably assessed in the following sections. Page 23 refers to a series of radio-tracking studies undertaken on large release-based shoots where the fate of a significant number of those released birds is to be predated by foxes.
According to Harris these studies give an ‘estimate of the importance of pheasants in the diet of British foxes’, but they actually tell us nothing about that. Foxes do eat pheasants and given the evidence base cited in this section it would be reasonable to say something like: ‘released gamebirds can form part of the diet of foxes on releasing shoots at certain times of the year.’ But no more than that.
Harris then cites another reference of his own (Baker and Harris 2003, ref 80) and lists on page 23 the number of pheasants a family of 6.5 foxes consumes in a year, as apparently stated by that paper. It is not an easy reference to access. In it the authors review and categorize the diet of foxes from 28 sources into small, medium and large mammals, birds, other vertebrates and a couple of other groups for analysis.
There are no data on pheasants or on gamebirds as a group. The last table in the discussion of that paper gives a hypothetical example indicating that if all the different bird species eaten by foxes in all the studies reviewed had been pheasants, this is how many it would have been. Harris however takes these theoretical figures from the paper and on page 23 of his 2021 report, he presents them as actual data on the number of pheasants a fox family eats while in reality no data on this are presented in the paper.
Foxes do predate pheasants and there are GWCT radio tracking data on this that Harris summarizes. However, on page 24 in the first paragraph, two references are then cited in a way that suggests they support the statement ‘there has been a marked increase in the contribution of predation on released gamebirds to supporting the British fox population over the last 20 years’. This is not the same thing and no other references are provided. But these two papers (numbers 33, 120) both by Aebischer (2019; 2013) do not show any kind of link between fox populations and gamebirds. One did not even mention foxes.
The other provides NGC data on fox culling trends over time which do not correlate with gamebird trends. In the GWCT Annual Review of 2013 Aebischer (2014) gives NGC data for foxes. But the graph shows no increase in foxes over 20 years. Similarly, Harris could have referenced the BTO BBS fox trend data which is freely available on the BTO website.
These data show a significant decline in foxes over the last 20 years or so. The references he cites, and the ones he avoids, all do not support his assertion. Harris covers up the fact that he has no supporting information on his central point and presumably hopes the reader does not check those references he cites (or the ones he doesn’t).
In the last part of this section on page 24 Harris goes on to cite an article in the GWCT Annual Review of 2003 (ref 132) to support the suggestion that the increase in releasing and hence food supply is linked specifically to an increase in fox culling, still clearly talking about the 20-year time frame.
But, obviously, virtually no data are presented on culling for the last 20 years in an 18-year-old reference. In fact, the review article shows that culling of foxes had decreased slightly in the 10 years between 1993 and 2003 (after increasing before that).
Harris could have used the Annual Review article of 2013 (Aebischer 2014) or the peer reviewed paper (Aebischer 2019) but these would have shown, instead, that fox culling has variously decreased, flat-lined or increased very slightly over the last 20 or 30 years, not following the substantial increase in releasing gamebirds over that period at all (pheasants doubled and partridge increased five-fold). Harris appears to obscure this false interpretation by citing a document that most people cannot access (Aebischer 2003) rather than citing one they can (Aebischer 2019).
Towards the end of the next section on carrion (second to last paragraph, page 26) there are a series of false assumptions for which there are no references. These detail how many foxes are supported per year by gamebird carrion, and the point finishes by saying ‘in the last 20 years there has clearly been a very substantial increase in the impact of the gamebird-shooting industry on the numbers of foxes in Britain’.
The numbers in that paragraph have no supporting evidence and as indicated in the previous paragraph while gamebird releasing has increased substantially in recent decades the fox population hasn’t. It is also noteworthy that the total contribution of all birds to the diet of British foxes according to Baker & Harris from 2003 (the paper we discussed the misrepresentation of earlier) was 19.5%.
But Harris doesn’t mention that number here even though he is an author on that paper. Even if he did another straight swap of the word ‘bird’ in Baker & Harris to ‘pheasant’ here, the numbers in this paragraph would be completely different.
The idea that gamebird release may be one of various contributing factors to fox numbers via predation or carrion provision either locally or at a wider scale is very plausible, as suggested in the discussion in Aebischer (2019). However, these two sections on predation and carrion present no evidence to support their key statements.
This process leads to a false picture in the summary parts of the report, which most people will read. In this case, in the conclusions on page 44, the report firmly states: ‘In Britain, industry data show that between 80,000 and 200,000 foxes a year [ie most foxes in Britain] are supported solely by predating or scavenging non-native gamebirds.’
Then in the summary there are several similar statements ‘’…there would be significantly fewer foxes in Britain if the gamebird shooting industry stopped providing the supplementary food [gamebirds]…’, ‘The gamebird shooting industry has played a major role in supporting the British fox population…’.
There are no papers or other sources referred to in this report or elsewhere that support such statements. This is not simply because the work has not been done and there is a lack of evidence: all these statements and figures are unlikely to be true.
Foxes are common throughout much of the Northern hemisphere and most scientists and well informed commentators would agree with the following: ‘Known or potential drivers of high mesopredator abundance such as foxes in Europe include forestry and agricultural influences on landscape fragmentation and configuration, aspects of livestock farming, high abundance of deer, lagomorphs and other food resources (such as gamebirds), the extirpation of most apex predators and suppression of those that remain, and urbanisation as a key source of anthropogenic food resource.’ There is no discussion of these things in this report.
Example 2: Releasing, dispersal and impacts on food resources
The section on dispersal and food requirements of released gamebirds (page 37-38) begins by discussing long-term dispersal of released gamebirds. This discussion ignores the process by which numbers of birds reduce significantly by the end of shooting and beyond, and that the density of any dispersing birds in the landscape rapidly becomes negligible in terms of any direct ecological impacts simply because the area of land involved is so large.
The next target is game crops which according to Harris is another form of supplementary feeding that ‘does not conform to conservation principles’ and ‘has [unwanted] ecological effects beyond the Defra 500 m buffer zone’. No reference is offered to support this. While most game crops on shoots are within 500 m or so of release areas there are also many papers in good peer-reviewed journals which show game crops support high densities of wintering birds on farmland (see Sage et al. 2020 and Mason et al. 2020 reviews).
They are widely accepted amongst conservation NGOs as beneficial. In spite of this Harris’s report implies that there is no scientific evidence for the benefits of game crops and states that ‘the gamebird industry claims that [game crops] are an important resource for wild bird populations’. To support that idea he chooses some industry (GWCT) guidelines to cite (ref 220) and discards all the actual evidence.
Harris then cites a PhD by Pressland (ref 229), which was supervised by himself but from which there are not yet any published papers, to state that pheasants eat an estimated 12,000 kg of invertebrates each day in early summer and 150,000 kg in September. He goes on to say that pheasants cause substantial pressure on rural invertebrate populations.
There are several papers that show that pheasants can affect some invertebrate groups inside release pens (see the reviews), but there is no evidence (and no real mechanism) by which pheasants will be significantly affecting invertebrate populations at distance from release sites.
These are the key points from the unpublished PhD thesis, which Harris cites:
- Pressland (2009) only looked at woods and wood edges near to pens (in the pheasant release sample).
- There was no difference in insect numbers caught in pitfalls in wood edge plots with or without releasing before or after releasing, and in any plot type after release.
- There were fewer insects overall caught in pitfalls in field edges alongside releasing woods just before releasing (which is difficult to explain).
- Some insect groups were caught more frequently from sites with releasing and some without in late spring and these variations were also not easily explained.
- Using faecal analysis, the only finding was that the proportion in pheasant diet increased in spring (when most insects are out).
- In relation to caterpillars and butterflies there were no significant negative effects.
Overall, there is no clear picture of substantial effects of released pheasants on invertebrates close to the release sites in Pressland’s thesis. There is no investigation at all of effects further away from release sites. There are no results that refer to 12,000 kg of invertebrates eaten per day in summer and 150,000 kg in winter in the thesis and there is no way by which these numbers could have been calculated from the thesis.
The figures appear in the Introduction of chapter 2 only, in relation to the crop size of a pheasant, in a discussion around how much of anything a pheasant can theoretically eat. This technique, of taking hypothetical numbers from an obscure reference and reporting them as findings of that study, was used in the previous example on fox diet (Baker & Harris 2003). The statement by Harris in this example, that pheasants cause a substantial pressure on rural invertebrate populations, is not supported by this or any other reference.
Example 3: Ecological reviews on the effects of releasing
In Harris’ conclusion (page 44, column 1), when referring to the two ecological and environmental reviews of gamebird releasing (Madden and Sage 2021; Sage et al. 2020) he makes the statement: ‘Had a structured and rigorous and quality assessment of the evidence assessed all of the available literature, it would have shown…’ He also dismisses those reviews (and others) in the last paragraph of the Introduction.
However, both of these publications were systematic reviews, following a standard procedure (Collins et al. 2015), with a clear account of the methods used to systematically identify all relevant references and to interpret their findings correctly. The first review was approved by Natural England and the second was peer reviewed and published in a good independent scientific journal.
In spite of this, these comprehensive and unbiased reviews are dismissed. They identified 100 plus references relevant to ecological/ environmental aspects, all of which are reported without selection or bias. Yet around half of these papers are not reported in Harris (2021) even though they are all relevant to the stated aims of a large part of it. As he says: ‘It is my decision as to what evidence I wish to include in my review’.
References
Aebischer NJ 2004. Recent trends in predator culls. Game Conservancy Review, 35, 92-95.
Aebischer NJ 2013. National gamebag census: released game species. Game and Wildlife Conservation Trust Review, 44, 34-37.
Aebischer NJ 2014. National Gamebag Census: Rabbits, foxes and mustelids. Game and Wildlife Conservation Trust Review 45, 30-33.
Aebischer NJ 2019. Fifty-year trends in hunting bags of birds and mammals and calibrated estimation of national bag size using GWCT’s National Gamebag Census. European Journal of Wildlife Research, 65, 64-76.
Baker PJ, Harris S 2003. A review of the diet of foxes in rural Britain and a preliminary assessment of their impact as a predator. In: Tattersall F, Manley W (eds) Conservation and conflict – mammals and farming in Britain, pp. 120-140. Westbury Publishing, Otley
Collins A, Coughlin D, Miller J, and Kirk S 2015. The production of quick scoping reviews and rapid evidence assessments: a how to guide. NERC, UK
Harris S 2021. A review of the animal welfare, public health, and environmental, ecological and conservation implications of rearing, releasing and shooting non-native gamebirds in Britain. Report to the Labour Animal Welfare Society.
Madden, JR and Sage, RB 2020. Ecological Consequences of Gamebird Releasing and management on Lowland Shoots in England. Natural England report.
Mason LR, Bicknell JE, Smart J et al. 2020. The impacts of non-native gamebird release in the UK: an updates evidence review. RSPB Research Report No. 66. RSPB, Sandy, UK.
Pressland C 2009. The impact of releasing pheasants for shooting on invertebrates in British woodlands. PhD thesis, University of Bristol.
Sage R.B., Hoodless A.N., Woodburn M.I.A et al. 2020. Summary review and synthesis: effects on habitats and wildlife of the release and management of pheasants and red-legged partridges on UK lowland shoots. Wildlife Biology 2020. wlb.00766.