Left: Artificial solitary bee nest structures   Right: Investigating invertebrates on nettles using a "pooter"
 
 
The results confirmed that breeding populations of solitary bees and wasps might be increased by provision of simple, inexpensive nests. However, nests for bumblebees were conspicuously unsuccessful, despite trying various different recommended designs. The results also suggested that dead wood would be valuable if left undisturbed for a sufficient duration. Even within the short period of the study, the fruits of four fungi and a slime mould appeared on the wood, and it provided shelter for invertebrates and amphibians.
 
Ponds, though in some cases used by frogs, were very slow to build up invertebrate communities of any diversity by natural colonization. However, if established artificially, even small ponds could maintain their populations over several seasons. Planting nettles for larvae of nettle-feeding butterflies appeared to be a waste of time, perhaps because of the occurrence elsewhere in the wider urban landscape of larger patches than most garden owners would be prepared to tolerate.
 
The broad conclusion is that while some widely-recommended kinds of ‘wildlife gardening’ may be very effective, others have a low probability of success on the time scales and spatial scales likely to be acceptable to many garden owners.
 
BUGS 1 conclusions
 
The results of the BUGS project confirm that gardens are a unique and important feature of urban systems, probably harbouring the bulk of biodiversity in such areas. There are a variety of features of gardens that seem to have positive effects on biodiversity, many of them under the control of garden owners; if even a modest proportion of the huge effort invested in garden management were directed at enhancing these features of gardens then the effects could be considerable. However, the greatest value will come when the quality and quantity of garden space is also maximized at the neighbourhood scale, which requires both planners and garden owners to take the biodiversity value of gardens to heart.
 
BUGS 2 Project
 
The first Sheffield BUGS project had three key elements. It provided information about the scale and nature of the garden resource in that city. It also provided detailed information about the plant and animal biodiversity in 61 Sheffield gardens, and whether some frequently recommended techniques for increasing biodiversity in gardens actually work.
 
Much of that work, specifically that concerned with trapping and identifying invertebrates, was both expensive and extremely labour-intensive, and is unlikely to be repeated. But it was clear that expanding other parts of the project to a larger (national) scale would be very valuable. One obvious outstanding question is whether the gardens of other UK cities, in terms of their quantity, nature and floristic composition, are similar to those of Sheffield.
 
The Biodiversity in Urban Gardens 2 project was carried out over three years (2004-2007) in five cities: Leicester, Oxford, Cardiff and Belfast and Edinburgh. It had four main components:
 
1.  To characterise the size distributions of gardens, and patterns of spatial variation in size, shape and connectivity, and the relationship of these to housing type, using aerial photography and Ordnance Survey mapping.
 
2.  To characterise the features of about 50 sample domestic gardens from each city, including trees and mature shrubs, lawn areas, ponds and compost heaps.
 
3.  To determine the floristic composition and diversity of the same sample gardens.
 
4.  Using the same sample gardens, householders were asked to complete a brief questionnaire about their gardening activities, including for example weeding, lawn mowing, dead-heading, watering and use of fertilisers and pesticides. Householders were also asked about attitudes to wildlife, and which specific (vertebrate) species they had observed in their gardens (from a list of 28 species).
 
The scale of the resource
 
The proportion of the urban area of each city covered by domestic gardens ranged from 21.8 % to 26.8 % (so Sheffield, at 23 %, is typical). In a random sample of at least 500 houses in each city, 99 % had gardens, the mean areas of which ranged from 155 m2 to 253 m2 (so once again Sheffield, at 173 m2, is not unusual). Not surprisingly, garden area was closely associated with housing type; taking median values across all five cities, detached houses had approximately twice the garden area of semi-detached houses, which in turn had approximately twice the garden area of terraced houses. As in Sheffield, relatively small gardens contributed disproportionately to the total garden area of each city, being more numerous than larger gardens. Perhaps surprisingly, there was no obvious relationship between garden size and distance to the edge of any of the cities.
 
Garden features
 
As in Sheffield, garden size played an overwhelming role in determining garden composition. Larger gardens contained more different kinds of land-use, and several land-uses (including cultivated borders, mown and unmown grass, trees >3 m, uncultivated areas, vegetable patches, and, to a lesser extent, ponds and compost sites) were more likely to be found in large gardens. Larger gardens also had proportionately larger areas of cultivated border, unmown grass, uncultivated land and vegetable patches. Larger gardens are also more likely to support a greater number of taller trees and large shrubs, and therefore to possess a disproportionately greater extent of vegetation cover >2 m in height.
 
On the other hand, buildings and non-vegetated features (e.g. patios, decking, areas of gravel, ponds, sheds, greenhouses and garages) were more likely to maintain a relatively constant absolute size, and therefore proportionally decreased with increasing garden area. In its effect on garden features, house age was much less important than garden size.
 
All these results are broadly consistent with those from Sheffield. Therefore, given the positive relationship between the extent of taller vegetation canopies and invertebrate species richness and abundance found in Sheffield, larger gardens should generally provide more potential habitat and thus greater opportunities for wildlife. But note that this is not an effect of garden size per se, but an indirect one of larger gardens being more likely to contain some (but not all, e.g. ponds) wildlife-friendly features.
 
Garden floras
 
The entire garden flora consisted of 1056 species. Numbers of plant species recorded in individual gardens ranged from 7 to 157 with a mean of 58 across all five cities. Of the total flora, 30% of species were native and 70% alien. These proportions are similar to those found in Sheffield, but the total number of species, and species per garden, is lower. A major reason for this is that the Sheffield survey also included lawns, which contain several species (many of them native) that tend not to be found elsewhere in the garden. For this reason, the Sheffield BUGS and BUGS 2 floristic surveys are not strictly comparable. Thus, for example,  although a high proportion of the plant species recorded most frequently in BUGS 2 were native, this tendency was not as marked as in Sheffield.
 
Floras of gardens in the five cities were similar in most respects, but a conspicuous feature of the results was that Belfast gardens were less species-rich. The most diverse were in Leicester, followed by Cardiff, Oxford, Edinburgh and then Belfast.  The graph below (adapted from the BUGS XII paper listed below) shows species accumulation curves, in which the total number of plant species found is plotted agains the sample number (quadrats). The red lines show the upper and lower bounds for gardens. For the same number of quadrats, Leicester gardens had roughly 50% more species than Belfast gardens. This may in part be because Belfast had many small paved yards rather than cultivated borders, as to some extent did Edinburgh.
 
 
 
BUGS Publications 
 
BUGS 1 Publications
Thompson, K., Austin, K.C., Smith, R.H., Warren, P.H., Angold, P.G. & Gaston, K.J. 2003. Urban domestic gardens (I): Putting small-scale plant diversity in context. Journal of Vegetation Science 14, 71-78.  Read here
 
Gaston, K.J., Smith, R.M., Thompson, K. & Warren, P.H. 2005. Urban domestic gardens (II): experimental tests of methods for increasing biodiversity. Biodiversity and Conservation 14, 395-413.    Read here
 
Thompson, K., Hodgson, J.G., Smith, R.M., Warren, P.H. & Gaston, K.J. 2004. Urban domestic gardens (III): Composition and diversity of lawn floras. Journal of Vegetation Science 15, 371-376.   Read here
 
Gaston, K.J., Warren, P.H., Thompson, K. & Smith, R.M. 2005. Urban domestic gardens (IV): the extent of the resource and its associated features. Biodiversity and Conservation 14, 3327-3349.   Read here
 
Smith, R.M., Gaston, K.J., Warren, P.H. & Thompson, K. 2005. Urban domestic gardens (V): relationships between landcover composition, housing and landscape. Landscape Ecology 20, 235-253.   Read here
 
Smith, R.M., Warren, P.H., Thompson, K. & Gaston, K.J. 2006. Urban domestic gardens (VI): environmental correlates of invertebrate species richness. Biodiversity and Conservation 15, 2415-2438.   Read here
 
Thompson, K., Colsell, S., Carpenter, J., Smith, R.M., Warren, P.H. & Gaston, K.J. 2005. Urban domestic gardens (VII): a preliminary survey of soil seed banks. Seed Science Research 15, 133-141.   Read here
 
Smith R.M., Gaston K.J., Warren P.H. & Thompson, K. 2006. Urban domestic gardens (VIII): environmental correlates of invertebrate abundance. Biodiversity and Conservation 15, 2515-2545.   Read here
 
Smith, R.M., Thompson, K., Hodgson, J.G., Warren, P.H. & Gaston, K.J. 2006. Urban domestic gardens (IX): Composition and richness of the vascular plant flora, and implications for native biodiversity. Biological Conservation 129, 312-322.   Read here
 
Gaston, K.J., Smith, R.M., Thompson, K. & Warren, P.H. 2004. Gardens and wildlife – the BUGS project. British Wildlife 16, 1-9.
 
BUGS 2 Publications
 
Loram, A., Tratalos, J., Warren, P.H. & Gaston, K.J. 2007. Urban domestic gardens (X): the extent & structure of the resource in five cities. Landscape Ecology 22, 601-615  Abstract here
 
Gaston, K.J., Fuller, R.A., Loram, A., MacDonald, C., Power, S and Dempsey, N (2007). Urban domestic gardens (XI): variation in urban wildlife gardening in the UK. Biodiversity and Conservation 16, p 3227-3238 Abstract here
 
Loram, A., Thompson, K., Warren, P.H. & Gaston, K.J. (2008) Urban domestic gardens (XII): the richness and composition of the flora in five cities. Journal of Vegetation Science 19:321–330 Abstract here
 
Loram, A., Warren, P.H. & Gaston, K.J. (2008) Urban domestic gardens (XIV):
The characteristics of gardens in five cities. Environmental Management. 2008 Sep;42(3):361-76. doi: 10.1007/s00267-008-9097-3. Epub 2008 Apr 25.  Abstract here
The Biodiversity in Urban Gardens in Sheffield (BUGS) project
 
Preamble
The two Bugs projects, from 2001 to 2008, still constitute the most complete and scientifically sound study of the garden resource, biodiversity and other attributes of wildlife in domestic gardens anywhere in the world. The papers resulting (listed at the end of this page) are fundamental starting points for any subsequent study of garden ecology.
 
 
Introduction
There are around 25 million dwellings in the UK, many with private/domestic gardens (no-one seems to know exactly how many), but surprisingly little is known about the ecology of almost all of them. For a long time, they were considered unworthy of study by ‘serious’ ecologists, and they’re also  surprisingly hard to study. Their highly fragmented ownership makes access difficult, and what can be done in the way of surveys or experiments can’t be too intrusive or introduce any hazards.
 
Most of what we know about garden biodiversity was based for a long time on a single garden in Leicester, intensively studied by Jennifer Owen for 30 years. You can read more about her seminal work in our paper on Jennifer Owen's Studies and another on the species she recorded.
 
BUGS 1 project
 
To address this lack of information, the first Biodiversity of Urban Gardens in Sheffield (BUGS) project was developed using the city of Sheffield, UK. It asked three main questions: how much of the urban environment is gardens; how much biodiversity is present and what determines its size and richness; and do some popular and frequently recommended techniques for increasing biodiversity actually work?
Representative Sheffield gardens from the study
Scale of the resource
 
With a mean individual area of 173 m2, domestic gardens cover approximately 33 km2 or 23% of the predominantly urban area of Sheffield, figures broadly consistent with previous estimates for other UK cities. In other words, any consideration of the quantity, quality, and connectivity of green space in cities which does not include domestic gardens is missing much, and perhaps most, of the story. Smaller gardens contribute disproportionately to the total area of gardens, because of their much larger numbers.
 
A telephone-based survey revealed that across Sheffield 14.4% of dwellings with gardens had ponds, 26% had nest-boxes, 29% had compost heaps, and 48% held trees (more than 3m tall). These values, scaled up, yield total estimates for domestic gardens in Sheffield of 25,200 ponds, 45,500 nest boxes, 50,750 compost heaps, and 360,000 trees. These figures are quite impressive, and in some cases, e.g. ponds, exceed densities of the same features in the wider countryside.
 
A detailed study was conducted of 61 gardens, chosen to vary as much as possible in key attributes, such as age, size and location in the city. Garden size played an overwhelming role in determining garden composition: larger gardens supported both a greater diversity and greater actual areas of most types of garden land uses, and were more likely to contain trees, vegetable patches and composting sites.
 
Garden biodiversity; plants
 
Standard survey methods were used to examine the plant and animal (invertebrate) diversity of the 61 study gardens. They were found to contain 1166 vascular plant species in all; 31 % of these were natives, but native species were generally more widely distributed than aliens, so individual gardens averaged 42 % native species.
 
The average garden contained 119 different plant species, but the range was very broad (48-268), and the number was quite strongly related to garden area: big gardens had more species. A conspicuous feature of garden floras was that most species are very rare; 490 species occurred only once, while only 35 species (3 % of the total) were found in at least half the gardens. A quarter of non-native plants originated from temperate Europe, nearly one fifth from Asia (excluding China), and more than a tenth from North America.
 
 
Surveying a Sheffield lawn with a quadrat.  A Malaise trap is being set up in the background
A separate survey of the floras of lawns (in those gardens that had them) revealed a surprising amount of floristic diversity. In 52 lawns there were 159 species of vascular plants, while individual lawns contained anything from 12 to 42 species, with an average of 24 species per lawn. Only one species, rough meadow grass Poa trivialis, was found in every lawn. The overwhelming majority of lawn species (94 % in the average garden) were natives. The composition of lawns showed a strong influence of local climate; even though common bent grass Agrostis capillaris is one of the most commonly sown lawn grasses, it was found in substantial amounts only in lawns in the higher, cooler and damper west of the city. In the lower and warmer east, it was largely replaced by creeping bent Agrostis stolonifera.
 
A study of soil seed banks of the study gardens revealed 2759 seeds in total from 366 soil samples (35 litres of soil in all). Garden seed banks were dominated by common weeds, with pearlwort Sagina procumbens and willowherbs Epilobium spp. topping the list. However, garden seed banks were surprisingly diverse, with 120 species in total, including seeds of many garden plants. Buddleia Buddleja davidii was easily the commonest garden plant in the seed bank, while tutsan Hypericum androsaemum, a native shrub commonly grown in gardens, was a close second. Garden pansy Viola x wittrockiana, columbine Aquilegia vulgaris, Atlas poppy Papaver atlanticum and large-flowered evening primrose Oenothera glazioviana were also frequent.
Tatty old wooden garden structures revealed a remarkable number of lichen species.
 
A detailed inventory of the lichens in each survey garden revealed nearly 80 species, ranging from 2 to 30 per garden, with an average of about 15. Moss on one tarmac path supported what might be a new, minuscule species of Macentina. The presence or absence of a suite of species depended on whether a garden contained acid or alkali stone surfaces (including concrete), weathered timber, or trees and shrubs with branches exposed to the light. The largest garden would have lost half of its lichen flora by the simple removal of an old wooden bench, which was also home to two species found nowhere else in the study.
 
Sampling, rather than a comprehensive inventory, also revealed 68 species of bryophytes, ranging from 3 to 25 per garden, with an average of 11 species. Some of the more unexpected finds were mosses considered characteristic of exposed ground in the region’s woodland or farmland, such as Dicranella heteromalla and Barbula hornschuchiana. While larger gardens tended to support more species of both bryophytes and lichens, the richness of bryophytes was also associated with that of the higher plants. Thus some gardens, irrespective of their size, were ‘good’ for higher plants and bryophytes, but not necessarily lichens. Gardens at higher altitude also tended to support more lichen and bryophyte species.
 
Garden biodiversity; invertebrates
 
Building on the methods employed in Jennifer Owen’s long-term study, invertebrates were sampled by pitfall trapping, Malaise trapping, and leaf-litter sampling. In addition, the garden flora was searched for leaf-mining insects (which were subsequently reared so they could be identified). In the end some 40,000 specimens were collected and sorted, and in excess of 700 species were positively identified (although many groups could not be tackled with the resources available).
 
Specimens were sorted to species in 13 major taxonomic groups (molluscs, millipedes, centipedes, woodlice, hoverflies, bumblebees, craneflies, sawflies, true bugs, solitary bees and wasps, beetles and arachnids). In the more diverse groups there was high turnover between gardens; in the last seven groups above, around half of their constituent species occurred in only a single garden. In contrast, many of the members of several species-poor groups of invertebrates (e.g. molluscs, woodlice and bumblebees) were virtually ubiquitous.
 
Overall, the most important predictors of both richness and abundance were related to a garden’s position in the urban area (especially location on an East-West or altitudinal gradient; a somewhat Sheffield-specific result, since few other UK cities extend over such a large altitudinal range) and the provision of garden vegetation (especially the number of trees and the area of canopy vegetation more than 2m above the ground). This result shows that efforts to enhance gardens for biodiversity seem to benefit the richness and sizes of wildlife populations simultaneously.
 
There was some indication that the more ground-based species (with typically lower dispersal abilities) were more influenced by factors associated with an individual garden,  while the more aerial species (with typically higher dispersal abilities) were more influenced by broader-scale factors, reflecting the landscape in which the garden was set (e.g. altitude).
 
"Myths" tested
 
Some popular ideas about biodiversity in gardens were not supported by the results. Aside from the fact that larger gardens tended to contain more trees, big gardens were not superior to small ones, there was little difference between suburban and urban gardens, and the richness and abundance of invertebrate wildlife were rarely related to the number of native plant species in individual gardens. One exception to the latter was the richness of leaf miners (see our pages on leaf mining moths and flies) : fifty-four leaf-mining insects were identified from more than one hundred plant species; of these hosts, only 15 out of 89 genera were non-native, and only one of these - nasturtiums Tropaeolum - belonged to a non-native family, the Tropaeolaceae.
 
The results suggest that garden biodiversity is responding to two sets of processes. First, the resources provided by individual gardens, which can readily be manipulated by individual garden owners, but tend to be partially limited by garden size. Second, the broader context, which is not something that an individual garden owner can do much about, short of moving house, but is something over which planning processes and government policies could exert some level of control.
 
Macroinvertebrates were also sampled in the thirty-seven garden ponds found in the survey gardens. The numbers of macroinvertebrate taxa found in ponds varied from five to fifteen; larger ponds and those with clearer water, or more submerged vegetation, tended to have more invertebrate species, though heavily shaded ponds had fewer. The presence of fish did not have a strong effect on invertebrate diversity, though it did affect the abundances of particular species; whether due to the direct effects of fish predation or due to the environmental conditions in ponds with fish in is not clear, since some taxa were more abundant in ponds with fish.
 
Wildlife gardening
 
The final part of the BUGS 1 project comprised a series of replicated experimental tests of five common wildlife gardening recommendations, involving the introduction to each of 20 gardens of (i) artificial nest sites for solitary bees and wasps; (ii) artificial nest sites for bumblebees; (iii) small ponds; (iv) dead wood for fungi & other saproxylic organisms; and (v) patches of stinging nettle.
The green lines show similar curves for the highest and lowest diversity urban and semi-natural habitats, which were from the highest: limestone grassland, then urban derelict land, scrub, urban brownfield grassland, acidic woodland and finally acidic grassland.  The difference in shape between these  and the garden curves is significant.  While limestone grassland shows great diversity at small sample size, the curve quickly flattens, with few more species found as the sample area increases.  With the garden plots, there is a lower initial diversity, but much less tendency to decline with increasing sample area, indeed the larger the area sampled the greater the number of species found.
 
This suggests that while the urban and semi-natural habitats are are relatively homogeneous, garden habitats are much more heterogeneous. Even small gardens vary greatly, as might different areas within larger gardens. Gardens as contrived habitats have a much greater pool of species. It is estimated that there are around 15,000 plant species commercially available to gardeners in the UK (Brickell 1996), compared with only 1,600 native, and about 1,800 naturalised species in the countryside.
 
While the higher plant biodiversity of the gardens was down to the number of alien species, the shape of the accumulation curves in when natives alone were plotted still showed little sign of flattening, so even for native species, the gardens were surprisingly heterogeneous.
 
Garden management and wildlife
 
The questionnaires revealed that the majority of householders used their gardens chiefly for relaxation, recreation, and eating. Fewer than one fifth included ‘gardening’ amongst their garden uses, although all performed some garden management, so clearly not all ‘gardening’ reflects an interest in gardening.
 
Watering and lawn-mowing were the most common activities and were predictors of other types of management including weeding, vegetation-cutting, leaf-collection, and dead-heading. The number of land uses (e.g., cultivated borders, vegetable patches, mown and unmown grass) tended to increase with the extent or frequency of garden-management activities. In addition, increasing levels of all forms of garden management were positively related to the total number of plant species recorded in gardens; in other words, and hardly surprisingly, those who did the most ‘gardening’ also tended to have the most varied and interesting gardens.
 
Sixty percent of respondents said they made some effort to attract wildlife, e.g., putting out food for birds, providing nest boxes, and growing so-called ‘wildlife-friendly’ plant species, although only 9% said that most of their garden management was aimed at attracting wildlife.
 
On average, a significantly greater number of wildlife species (from the list of 28 selected vertebrates) was observed in gardens where efforts were made to attract wildlife compared with those in which no effort was made, with the number of species increasing slightly with extent of the effort reportedly made. Householders with ponds in their gardens also tended to make more effort to attract wildlife, provide water for wildlife, and observe more wildlife (from the list of 28 selected vertebrates) than householders who did not have ponds. But note that, as in all such relationships, it’s hard to separate cause and effect; any correspondence between increased observations of wildlife species and effort aimed at attracting wildlife is dependent on the level of interest and activity of respondents, their interest in wildlife, and ability to recognise different species. Someone who is indifferent to their garden and/or wildlife is unlikely to be aware of the species that frequent that garden, while someone who spends time in their garden and is interested in wildlife is more likely to observe the various species and be curious as to their identity.
 
Given that the 28 vertebrates about which householders were questioned included very common species such as blackbird, blue tit, fox, frog, great tit, hedgehog, house sparrow, magpie, newt, robin, squirrel, starling and wren, observations of (vertebrate) wildlife were surprisingly low, with a mean of 7.8 species observed, and a modal value (i.e. the number observed by the largest number of respondents) of zero.
 
Acknowledgements
 
BUGS 1 was supported by funding from the URGENT programme of the Natural Environment Research Council (grant GST/02/2592/), and from the Department of the Environment, Transport & the Regions. BUGS 2  was supported by funding from the Countryside Council for Wales, the Department for Environment, Food and Rural Affairs, English Nature, the Environment and Heritage Service and the Scottish and Northern Ireland Forum for Environmental Research (SNIFFER).
 
Thanks are due to the householders who made their gardens available, to numerous experts for their immense assistance in identifying specimens, and for the assistance of many helpers during the project.
 
 
Page written by Dr Ken Thompson edited by Steve Head
 
The Biodiversity in Urban Gardens in Sheffield (BUGS) project
 
Preamble
The two Bugs projects, from 2001 to 2008, still constitute the most complete and scientifically sound study of the garden resource, biodiversity and other attributes of wildlife in domestic gardens anywhere in the world. The papers resulting (listed at the end of this page) are fundamental starting points for any subsequent study of garden ecology.
 
 
Introduction
There are around 25 million dwellings in the UK, many with private/domestic gardens (no-one seems to know exactly how many), but surprisingly little is known about the ecology of almost all of them. For a long time, they were considered unworthy of study by ‘serious’ ecologists, and they’re also  surprisingly hard to study. Their highly fragmented ownership makes access difficult, and what can be done in the way of surveys or experiments can’t be too intrusive or introduce any hazards.
 
Most of what we know about garden biodiversity was based for a long time on a single garden in Leicester, intensively studied by Jennifer Owen for 30 years. You can read more about her seminal work in our paper on Jennifer Owen's Studies and another on the species she recorded.
 
BUGS 1 project
 
To address this lack of information, the first Biodiversity of Urban Gardens in Sheffield (BUGS) project was developed using the city of Sheffield, UK. It asked three main questions: how much of the urban environment is gardens; how much biodiversity is present and what determines its size and richness; and do some popular and frequently recommended techniques for increasing biodiversity actually work?
Representative Sheffield gardens from the study
Scale of the resource
 
With a mean individual area of 173 m2, domestic gardens cover approximately 33 km2 or 23% of the predominantly urban area of Sheffield, figures broadly consistent with previous estimates for other UK cities. In other words, any consideration of the quantity, quality, and connectivity of green space in cities which does not include domestic gardens is missing much, and perhaps most, of the story. Smaller gardens contribute disproportionately to the total area of gardens, because of their much larger numbers.
 
A telephone-based survey revealed that across Sheffield 14.4% of dwellings with gardens had ponds, 26% had nest-boxes, 29% had compost heaps, and 48% held trees (more than 3m tall). These values, scaled up, yield total estimates for domestic gardens in Sheffield of 25,200 ponds, 45,500 nest boxes, 50,750 compost heaps, and 360,000 trees. These figures are quite impressive, and in some cases, e.g. ponds, exceed densities of the same features in the wider countryside.
 
A detailed study was conducted of 61 gardens, chosen to vary as much as possible in key attributes, such as age, size and location in the city. Garden size played an overwhelming role in determining garden composition: larger gardens supported both a greater diversity and greater actual areas of most types of garden land uses, and were more likely to contain trees, vegetable patches and composting sites.
 
Garden biodiversity; plants
 
Standard survey methods were used to examine the plant and animal (invertebrate) diversity of the 61 study gardens. They were found to contain 1166 vascular plant species in all; 31 % of these were natives, but native species were generally more widely distributed than aliens, so individual gardens averaged 42 % native species.
 
The average garden contained 119 different plant species, but the range was very broad (48-268), and the number was quite strongly related to garden area: big gardens had more species. A conspicuous feature of garden floras was that most species are very rare; 490 species occurred only once, while only 35 species (3 % of the total) were found in at least half the gardens. A quarter of non-native plants originated from temperate Europe, nearly one fifth from Asia (excluding China), and more than a tenth from North America.
 
 
Surveying a Sheffield lawn with a quadrat.  A Malaise trap is being set up in the background
A separate survey of the floras of lawns (in those gardens that had them) revealed a surprising amount of floristic diversity. In 52 lawns there were 159 species of vascular plants, while individual lawns contained anything from 12 to 42 species, with an average of 24 species per lawn. Only one species, rough meadow grass Poa trivialis, was found in every lawn. The overwhelming majority of lawn species (94 % in the average garden) were natives. The composition of lawns showed a strong influence of local climate; even though common bent grass Agrostis capillaris is one of the most commonly sown lawn grasses, it was found in substantial amounts only in lawns in the higher, cooler and damper west of the city. In the lower and warmer east, it was largely replaced by creeping bent Agrostis stolonifera.
 
A study of soil seed banks of the study gardens revealed 2759 seeds in total from 366 soil samples (35 litres of soil in all). Garden seed banks were dominated by common weeds, with pearlwort Sagina procumbens and willowherbs Epilobium spp. topping the list. However, garden seed banks were surprisingly diverse, with 120 species in total, including seeds of many garden plants. Buddleia Buddleja davidii was easily the commonest garden plant in the seed bank, while tutsan Hypericum androsaemum, a native shrub commonly grown in gardens, was a close second. Garden pansy Viola x wittrockiana, columbine Aquilegia vulgaris, Atlas poppy Papaver atlanticum and large-flowered evening primrose Oenothera glazioviana were also frequent.
Tatty old wooden garden structures revealed a remarkable number of lichen species.
A detailed inventory of the lichens in each survey garden revealed nearly 80 species, ranging from 2 to 30 per garden, with an average of about 15. Moss on one tarmac path supported what might be a new, minuscule species of Macentina. The presence or absence of a suite of species depended on whether a garden contained acid or alkali stone surfaces (including concrete), weathered timber, or trees and shrubs with branches exposed to the light. The largest garden would have lost half of its lichen flora by the simple removal of an old wooden bench, which was also home to two species found nowhere else in the study.
 
Sampling, rather than a comprehensive inventory, also revealed 68 species of bryophytes, ranging from 3 to 25 per garden, with an average of 11 species. Some of the more unexpected finds were mosses considered characteristic of exposed ground in the region’s woodland or farmland, such as Dicranella heteromalla and Barbula hornschuchiana. While larger gardens tended to support more species of both bryophytes and lichens, the richness of bryophytes was also associated with that of the higher plants. Thus some gardens, irrespective of their size, were ‘good’ for higher plants and bryophytes, but not necessarily lichens. Gardens at higher altitude also tended to support more lichen and bryophyte species.
 
Garden biodiversity; invertebrates
 
Building on the methods employed in Jennifer Owen’s long-term study, invertebrates were sampled by pitfall trapping, Malaise trapping, and leaf-litter sampling. In addition, the garden flora was searched for leaf-mining insects (which were subsequently reared so they could be identified). In the end some 40,000 specimens were collected and sorted, and in excess of 700 species were positively identified (although many groups could not be tackled with the resources available).
 
Specimens were sorted to species in 13 major taxonomic groups (molluscs, millipedes, centipedes, woodlice, hoverflies, bumblebees, craneflies, sawflies, true bugs, solitary bees and wasps, beetles and arachnids). In the more diverse groups there was high turnover between gardens; in the last seven groups above, around half of their constituent species occurred in only a single garden. In contrast, many of the members of several species-poor groups of invertebrates (e.g. molluscs, woodlice and bumblebees) were virtually ubiquitous.
 
Overall, the most important predictors of both richness and abundance were related to a garden’s position in the urban area (especially location on an East-West or altitudinal gradient; a somewhat Sheffield-specific result, since few other UK cities extend over such a large altitudinal range) and the provision of garden vegetation (especially the number of trees and the area of canopy vegetation more than 2m above the ground). This result shows that efforts to enhance gardens for biodiversity seem to benefit the richness and sizes of wildlife populations simultaneously.
 
There was some indication that the more ground-based species (with typically lower dispersal abilities) were more influenced by factors associated with an individual garden,  while the more aerial species (with typically higher dispersal abilities) were more influenced by broader-scale factors, reflecting the landscape in which the garden was set (e.g. altitude).
 
"Myths" tested
 
Some popular ideas about biodiversity in gardens were not supported by the results. Aside from the fact that larger gardens tended to contain more trees, big gardens were not superior to small ones, there was little difference between suburban and urban gardens, and the richness and abundance of invertebrate wildlife were rarely related to the number of native plant species in individual gardens. One exception to the latter was the richness of leaf miners (see our pages on leaf mining moths and flies) : fifty-four leaf-mining insects were identified from more than one hundred plant species; of these hosts, only 15 out of 89 genera were non-native, and only one of these - nasturtiums Tropaeolum - belonged to a non-native family, the Tropaeolaceae.
 
The results suggest that garden biodiversity is responding to two sets of processes. First, the resources provided by individual gardens, which can readily be manipulated by individual garden owners, but tend to be partially limited by garden size. Second, the broader context, which is not something that an individual garden owner can do much about, short of moving house, but is something over which planning processes and government policies could exert some level of control.
 
Macroinvertebrates were also sampled in the thirty-seven garden ponds found in the survey gardens. The numbers of macroinvertebrate taxa found in ponds varied from five to fifteen; larger ponds and those with clearer water, or more submerged vegetation, tended to have more invertebrate species, though heavily shaded ponds had fewer. The presence of fish did not have a strong effect on invertebrate diversity, though it did affect the abundances of particular species; whether due to the direct effects of fish predation or due to the environmental conditions in ponds with fish in is not clear, since some taxa were more abundant in ponds with fish.
 
Wildlife gardening
 
The final part of the BUGS 1 project comprised a series of replicated experimental tests of five common wildlife gardening recommendations, involving the introduction to each of 20 gardens of (i) artificial nest sites for solitary bees and wasps; (ii) artificial nest sites for bumblebees; (iii) small ponds; (iv) dead wood for fungi & other saproxylic organisms; and (v) patches of stinging nettle.
Left: Artificial solitary bee nest structures   Right: Investigating invertebrates on nettles using a "pooter"
 
 
The results confirmed that breeding populations of solitary bees and wasps might be increased by provision of simple, inexpensive nests. However, nests for bumblebees were conspicuously unsuccessful, despite trying various different recommended designs. The results also suggested that dead wood would be valuable if left undisturbed for a sufficient duration. Even within the short period of the study, the fruits of four fungi and a slime mould appeared on the wood, and it provided shelter for invertebrates and amphibians.
 
Ponds, though in some cases used by frogs, were very slow to build up invertebrate communities of any diversity by natural colonization. However, if established artificially, even small ponds could maintain their populations over several seasons. Planting nettles for larvae of nettle-feeding butterflies appeared to be a waste of time, perhaps because of the occurrence elsewhere in the wider urban landscape of larger patches than most garden owners would be prepared to tolerate.
 
The broad conclusion is that while some widely-recommended kinds of ‘wildlife gardening’ may be very effective, others have a low probability of success on the time scales and spatial scales likely to be acceptable to many garden owners.
 
BUGS 1 conclusions
 
The results of the BUGS project confirm that gardens are a unique and important feature of urban systems, probably harbouring the bulk of biodiversity in such areas. There are a variety of features of gardens that seem to have positive effects on biodiversity, many of them under the control of garden owners; if even a modest proportion of the huge effort invested in garden management were directed at enhancing these features of gardens then the effects could be considerable. However, the greatest value will come when the quality and quantity of garden space is also maximized at the neighbourhood scale, which requires both planners and garden owners to take the biodiversity value of gardens to heart.
 
BUGS 2 Project
 
The first Sheffield BUGS project had three key elements. It provided information about the scale and nature of the garden resource in that city. It also provided detailed information about the plant and animal biodiversity in 61 Sheffield gardens, and whether some frequently recommended techniques for increasing biodiversity in gardens actually work.
 
Much of that work, specifically that concerned with trapping and identifying invertebrates, was both expensive and extremely labour-intensive, and is unlikely to be repeated. But it was clear that expanding other parts of the project to a larger (national) scale would be very valuable. One obvious outstanding question is whether the gardens of other UK cities, in terms of their quantity, nature and floristic composition, are similar to those of Sheffield.
 
The Biodiversity in Urban Gardens 2 project was carried out over three years (2004-2007) in five cities: Leicester, Oxford, Cardiff and Belfast and Edinburgh. It had four main components:
 
1.  To characterise the size distributions of gardens, and patterns of spatial variation in size, shape and connectivity, and the relationship of these to housing type, using aerial photography and Ordnance Survey mapping.
 
2.  To characterise the features of about 50 sample domestic gardens from each city, including trees and mature shrubs, lawn areas, ponds and compost heaps.
 
3.  To determine the floristic composition and diversity of the same sample gardens.
 
4.  Using the same sample gardens, householders were asked to complete a brief questionnaire about their gardening activities, including for example weeding, lawn mowing, dead-heading, watering and use of fertilisers and pesticides. Householders were also asked about attitudes to wildlife, and which specific (vertebrate) species they had observed in their gardens (from a list of 28 species).
 
The scale of the resource
 
The proportion of the urban area of each city covered by domestic gardens ranged from 21.8 % to 26.8 % (so Sheffield, at 23 %, is typical). In a random sample of at least 500 houses in each city, 99 % had gardens, the mean areas of which ranged from 155 m2 to 253 m2 (so once again Sheffield, at 173 m2, is not unusual). Not surprisingly, garden area was closely associated with housing type; taking median values across all five cities, detached houses had approximately twice the garden area of semi-detached houses, which in turn had approximately twice the garden area of terraced houses. As in Sheffield, relatively small gardens contributed disproportionately to the total garden area of each city, being more numerous than larger gardens. Perhaps surprisingly, there was no obvious relationship between garden size and distance to the edge of any of the cities.
 
Garden features
 
As in Sheffield, garden size played an overwhelming role in determining garden composition. Larger gardens contained more different kinds of land-use, and several land-uses (including cultivated borders, mown and unmown grass, trees >3 m, uncultivated areas, vegetable patches, and, to a lesser extent, ponds and compost sites) were more likely to be found in large gardens. Larger gardens also had proportionately larger areas of cultivated border, unmown grass, uncultivated land and vegetable patches. Larger gardens are also more likely to support a greater number of taller trees and large shrubs, and therefore to possess a disproportionately greater extent of vegetation cover >2 m in height.
 
On the other hand, buildings and non-vegetated features (e.g. patios, decking, areas of gravel, ponds, sheds, greenhouses and garages) were more likely to maintain a relatively constant absolute size, and therefore proportionally decreased with increasing garden area. In its effect on garden features, house age was much less important than garden size.
 
All these results are broadly consistent with those from Sheffield. Therefore, given the positive relationship between the extent of taller vegetation canopies and invertebrate species richness and abundance found in Sheffield, larger gardens should generally provide more potential habitat and thus greater opportunities for wildlife. But note that this is not an effect of garden size per se, but an indirect one of larger gardens being more likely to contain some (but not all, e.g. ponds) wildlife-friendly features.
 
Garden floras
 
The entire garden flora consisted of 1056 species. Numbers of plant species recorded in individual gardens ranged from 7 to 157 with a mean of 58 across all five cities. Of the total flora, 30% of species were native and 70% alien. These proportions are similar to those found in Sheffield, but the total number of species, and species per garden, is lower. A major reason for this is that the Sheffield survey also included lawns, which contain several species (many of them native) that tend not to be found elsewhere in the garden. For this reason, the Sheffield BUGS and BUGS 2 floristic surveys are not strictly comparable. Thus, for example,  although a high proportion of the plant species recorded most frequently in BUGS 2 were native, this tendency was not as marked as in Sheffield.
 
Floras of gardens in the five cities were similar in most respects, but a conspicuous feature of the results was that Belfast gardens were less species-rich. The most diverse were in Leicester, followed by Cardiff, Oxford, Edinburgh and then Belfast.  The graph below (adapted from the BUGS XII paper listed below) shows species accumulation curves, in which the total number of plant species found is plotted agains the sample number (quadrats). The red lines show the upper and lower bounds for gardens. For the same number of quadrats, Leicester gardens had roughly 50% more species than Belfast gardens. This may in part be because Belfast had many small paved yards rather than cultivated borders, as to some extent did Edinburgh.
 
 
 
The green lines show similar curves for the highest and lowest diversity urban and semi-natural habitats, which were from the highest: limestone grassland, then urban derelict land, scrub, urban brownfield grassland, acidic woodland and finally acidic grassland.  The difference in shape between these  and the garden curves is significant.  While limestone grassland shows great diversity at small sample size, the curve quickly flattens, with few more species found as the sample area increases.  With the garden plots, there is a lower initial diversity, but much less tendency to decline with increasing sample area, indeed the larger the area sampled, the greater the number of species found.
 
This suggests that while the urban and semi-natural habitats are are relatively homogeneous, garden habitats are much more heterogeneous. Even small gardens vary greatly, as might different areas within larger gardens. Gardens as contrived habitats have a much greater pool of species. It is estimated that there are around 15,000 plant species commercially available to gardeners in the UK (Brickell 1996), compared with only 1,600 native, and about 1,800 naturalised species in the countryside.
 
While the higher plant biodiversity of the gardens was down to the number of alien species, the shape of the accumulation curves in when natives alone were plotted still showed little sign of flattening, so even for native species, the gardens were surprisingly heterogeneous.
 
Garden management and wildlife
 
The questionnaires revealed that the majority of householders used their gardens chiefly for relaxation, recreation, and eating. Fewer than one fifth included ‘gardening’ amongst their garden uses, although all performed some garden management, so clearly not all ‘gardening’ reflects an interest in gardening.
 
Watering and lawn-mowing were the most common activities and were predictors of other types of management including weeding, vegetation-cutting, leaf-collection, and dead-heading. The number of land uses (e.g., cultivated borders, vegetable patches, mown and unmown grass) tended to increase with the extent or frequency of garden-management activities. In addition, increasing levels of all forms of garden management were positively related to the total number of plant species recorded in gardens; in other words, and hardly surprisingly, those who did the most ‘gardening’ also tended to have the most varied and interesting gardens.
 
Sixty percent of respondents said they made some effort to attract wildlife, e.g., putting out food for birds, providing nest boxes, and growing so-called ‘wildlife-friendly’ plant species, although only 9% said that most of their garden management was aimed at attracting wildlife.
 
On average, a significantly greater number of wildlife species (from the list of 28 selected vertebrates) was observed in gardens where efforts were made to attract wildlife compared with those in which no effort was made, with the number of species increasing slightly with extent of the effort reportedly made. Householders with ponds in their gardens also tended to make more effort to attract wildlife, provide water for wildlife, and observe more wildlife (from the list of 28 selected vertebrates) than householders who did not have ponds. But note that, as in all such relationships, it’s hard to separate cause and effect; any correspondence between increased observations of wildlife species and effort aimed at attracting wildlife is dependent on the level of interest and activity of respondents, their interest in wildlife, and ability to recognise different species. Someone who is indifferent to their garden and/or wildlife is unlikely to be aware of the species that frequent that garden, while someone who spends time in their garden and is interested in wildlife is more likely to observe the various species and be curious as to their identity.
 
Given that the 28 vertebrates about which householders were questioned included very common species such as blackbird, blue tit, fox, frog, great tit, hedgehog, house sparrow, magpie, newt, robin, squirrel, starling and wren, observations of (vertebrate) wildlife were surprisingly low, with a mean of 7.8 species observed, and a modal value (i.e. the number observed by the largest number of respondents) of zero.
 
Acknowledgements
 
BUGS 1 was supported by funding from the URGENT programme of the Natural Environment Research Council (grant GST/02/2592/), and from the Department of the Environment, Transport & the Regions. BUGS 2  was supported by funding from the Countryside Council for Wales, the Department for Environment, Food and Rural Affairs, English Nature, the Environment and Heritage Service and the Scottish and Northern Ireland Forum for Environmental Research (SNIFFER).
 
Thanks are due to the householders who made their gardens available, to numerous experts for their immense assistance in identifying specimens, and for the assistance of many helpers during the project.
 
 
Page written by Dr Ken Thompson edited by Steve Head
 
BUGS Publications 
 
BUGS 1 Publications
Thompson, K., Austin, K.C., Smith, R.H., Warren, P.H., Angold, P.G. & Gaston, K.J. 2003. Urban domestic gardens (I): Putting small-scale plant diversity in context. Journal of Vegetation Science 14, 71-78.  Read here
 
Gaston, K.J., Smith, R.M., Thompson, K. & Warren, P.H. 2005. Urban domestic gardens (II): experimental tests of methods for increasing biodiversity. Biodiversity and Conservation 14, 395-413.    Read here
 
Thompson, K., Hodgson, J.G., Smith, R.M., Warren, P.H. & Gaston, K.J. 2004. Urban domestic gardens (III): Composition and diversity of lawn floras. Journal of Vegetation Science 15, 371-376.   Read here
 
Gaston, K.J., Warren, P.H., Thompson, K. & Smith, R.M. 2005. Urban domestic gardens (IV): the extent of the resource and its associated features. Biodiversity and Conservation 14, 3327-3349.   Read here
 
Smith, R.M., Gaston, K.J., Warren, P.H. & Thompson, K. 2005. Urban domestic gardens (V): relationships between landcover composition, housing and landscape. Landscape Ecology 20, 235-253.   Read here
 
Smith, R.M., Warren, P.H., Thompson, K. & Gaston, K.J. 2006. Urban domestic gardens (VI): environmental correlates of invertebrate species richness. Biodiversity and Conservation 15, 2415-2438.   Read here
 
Thompson, K., Colsell, S., Carpenter, J., Smith, R.M., Warren, P.H. & Gaston, K.J. 2005. Urban domestic gardens (VII): a preliminary survey of soil seed banks. Seed Science Research 15, 133-141.   Read here
 
Smith R.M., Gaston K.J., Warren P.H. & Thompson, K. 2006. Urban domestic gardens (VIII): environmental correlates of invertebrate abundance. Biodiversity and Conservation 15, 2515-2545.   Read here
 
Smith, R.M., Thompson, K., Hodgson, J.G., Warren, P.H. & Gaston, K.J. 2006. Urban domestic gardens (IX): Composition and richness of the vascular plant flora, and implications for native biodiversity. Biological Conservation 129, 312-322.   Read here
 
Gaston, K.J., Smith, R.M., Thompson, K. & Warren, P.H. 2004. Gardens and wildlife – the BUGS project. British Wildlife 16, 1-9.
 
BUGS 2 Publications
 
Loram, A., Tratalos, J., Warren, P.H. & Gaston, K.J. 2007. Urban domestic gardens (X): the extent & structure of the resource in five cities. Landscape Ecology 22, 601-615  Abstract here
 
Gaston, K.J., Fuller, R.A., Loram, A., MacDonald, C., Power, S and Dempsey, N (2007). Urban domestic gardens (XI): variation in urban wildlife gardening in the UK. Biodiversity and Conservation 16, p 3227-3238  Abstract here
 
Loram, A., Thompson, K., Warren, P.H. & Gaston, K.J. (2008) Urban domestic gardens (XII): the richness and composition of the flora in five cities. Journal of Vegetation Science 19:321–330 Abstract here
 
Loram, A., Warren, P.H. & Gaston, K.J. (2008) Urban domestic gardens (XIV):
The characteristics of gardens in five cities. Environmental Management. 2008 Sep;42(3):361-76. doi: 10.1007/s00267-008-9097-3. Epub 2008 Apr 25. Abstract here