The effects of the intensification of agriculture on northern temperate dung beetle communities

Authors


Correspondence: Stephen A. Hutton, Department of Zoology, Ecology and Plant Science, National University of Ireland, Cork, Ireland (fax +353 21 4270562; e-mail stephenhutton@vodafone.ie).

Summary

  • 1There is growing concern that the intensification of agriculture within southern Ireland is having detrimental effects on Irish fauna through habitat loss, reduction in biodiversity and pollution-related events. To assess the impact of farm management on one group of important insects, the dung beetles, their abundance, biomass, diversity and species richness were examined using dung-baited pitfall traps in intensive, organic and rough grazing farms.
  • 2We collected 39 631 dung beetles belonging to 24 different species. Eight species (Aphodius prodromus, A. sphacelatus, A. ater, A. rufipes, A. depressus, Sphaeridium lunatum, S. scarabaeoides and Margarinotus carbonarius) accounted for 94% of the captures, but their relative dominance varied with farm type. 54% of total individuals captured were from organic sites, 30% from intensive sites and 16% from rough grazing sites.
  • 3Organic farms had significantly greater beetle biomass, diversity and species richness compared to intensive and rough grazing farms. Aphodius abundance on organic farms was significantly greater than on intensive and rough grazing farms in spring and autumn. Aphodius biomass on organic farms was significantly greater than on rough grazing farms in spring, late summer and autumn, and greater in autumn compared with intensive farms.
  • 4A colonization experiment demonstrated that the variation in dung beetle abundance among different dung types depended on the farm type. This was not true for beetle biomass colonizing dung pads, where variation among different dung types and different farm types were independent. Multiple comparisons showed that under rough grazing and intensive management there was significantly greater beetle abundance and biomass found in dung from organic and intensive farms than from rough grazing farms. Dung from organic farms held a significantly greater beetle biomass than the intensive and rough grazing dung.
  • 5Synthesis and applications. Intensive agricultural management including the use of chemical fertilisers, veterinary drugs (e.g. ivermectin) and removal of herbaceous field boundaries could be detrimental to dung beetle biodiversity and dung decomposition. Organic farming has beneficial effects on dung beetle communities. Patchy ecosystems characterized by a diversity of ungulate species increase dung beetle diversity and hence improve dung decomposition.

Introduction

Over the past 40 years the more developed countries of the world have experienced a modern agricultural revolution. The impact of this transformation has been such that society's priorities for agriculture and land use have changed substantially (Lafferty, Commins & Walsh 1999). Ireland provides a good example of this agricultural revolution, the essence of which has been higher productivity, but with increasing division of performance between categories of farms and regions of the country. Of the total Irish land area of 7 million hectares, almost 5 million hectares are used for agricultural purposes, including forestry (Lafferty et al. 1999).

Many studies have shown the potentially harmful effect of agricultural landscape change on insect diversity (Morris 1979; Rushton, Luff & Eyre 1989; Morris & Rispin 1993; Di Giulio, Edwards & Meister 2001; Vickery et al. 2001; Kruess & Tscharntke 2002) but there is relatively little information on an important functional group, the north temperate dung beetles (but see Roslin & Koivunen 2001). Dung pats are a major contributor to biodiversity in the agricultural landscape, holding hundreds of species of invertebrate, which include highly specialized species with complex behaviours (Hanski & Cambefort 1991). Dung beetles in the genus Aphodius (Aphodiinae, Scarabaeidae) are the dominant coprophagous beetles found in northern temperate Europe and play an essential role in the breakdown of the dung and the recycling of the organic matter and plant nutrients that it contains (Fry & Lonsdale 1991). Without such recycling, pastures would soon become covered by rank patches of grass with lower digestibility and nutritive value, eventually becoming unsuitable for grazing (Gittings, Giller & Stakelum 1994). There is therefore increasing concern about dung beetle conservation.

Apart from the general decline across virtually all orders of insects due to habitat loss (Lumaret 1994; Samways 1994), dung beetles are threatened by several factors, both at the population and community level. The majority of studies on the impacts of farming practice on dung beetles have mostly investigated the effects of veterinary treatments on beetle communities. In particular, ivermectin, a broad-spectrum drug applied to livestock, appears to affect dung beetle communities causing a reduction in species diversity and an increase in species dominance (Wall & Strong 1987; Madsen et al. 1990; Lumaret et al. 1993; Floate 1998; Dadour, Cook & Neesam 1999), although this may be dependent on the mode of application to the animal (see Errouissi et al. 2001). Little is known about the effect that the change in the agricultural landscape, as a result of intensification, is having on dung beetle community dynamics, although several studies have shown that habitat is an important factor in determining beetle communities (Howden & Nealis 1975; Hanski & Koskela 1977; Koskela & Hanski 1977; Menéndez & Gutiérrez 1996; Barbero, Palestrini & Rolando 1999; Roslin & Koivunen 2001; Roslin 2001; Vessby 2001).

The aims of our study were (i) to evaluate the potential effect of intensification of agriculture in southern Ireland on north temperate dung beetle diversity and population dynamics, and (ii) to assess the relative contributions of habitat change and dung quality to changes in dung beetle assemblages under the different levels of intensification. We investigated the simple null hypothesis that there is no difference in dung beetle abundance, biomass, species richness, diversity, colonization and assemblage structure on cattle dung between different types of farm management.

Materials and methods

study sites

The study sites were situated in Co. Cork and Co. Tipperary in Southern Ireland. The research, carried out between May and October 2000, included three types of farm: intensive (anthelmintic treatment of cattle and high input of synthetic fertilisers/pesticides; soils typically of good quality), rough grazing (some anthelmintic treatment of cattle and low input of synthetic fertilisers/pesticides; soils typically of low quality) and organic (no anthelmintic treatment of cattle and no input of synthetic fertilisers/pesticides; soils typically of good quality). For each farm type, four farms were selected, resulting in a total of 12 study sites. More information on each farm is shown in Table 1.

Table 1.  Farm variables for each site. I = intensive, R = rough grazing, O = organic. Key to cattle breed codes: AA = Aberdeen Angus, BB = Belgian Blue, C = Charolais, F = Friesian, H = Hereford, J = Jersey, L = Limousin, S = Simmental, × denotes a cross breed
Farm typeFertiliserIvermectinFarm size (acres)Habitat compositionSoil typeAltitudeStocking densityBreed(s)Other livestock
TypeAppliedQuantityAppliedFrequencyMethod
11Urea + NPK 27·2·5·5Urea – March NPK – May to SeptemberMarch – 23 kg acre−1May to Sept – 141 kg acre−1Calves – summer Older cattle – May and NovemberCalves – 3 times across summerOral dose18096% grass, 4% woodlandRed sand stone soil (sandy loam)152 m190 cattle (1·1 acre−1)F (90%), BB (10%)2 horses
12Nitrogen + slurrySpring and JuneN – 41 kg acre−1slurry – 2000 gallons acre−1Autumn for cows + adult cattleEvery month for calvesSub injection18994% grass, 6% woodlandClay loam182 m250 cattle (1·3 acre−1)F (90%), H (5%), C (5%)
13Nitrogen P+K as per soil testNovemberN − 121 kg acre−1Young stock during summerTwicePour on dose247100% grassClay loam122–158 m279 cattle (1·1 acre−1)F (100%)
14NPK 27·2·5·5March150 kg acre−110095% grass, 5% woodlandSilt/medium loam152 m56 cattle (0·6 acre−1)F (100%)3 horses
R1NPK 27·2·5·5February and September81 kg acre−1November (cows) June/Sept (cattle)OnceSub injection5080% grass, 10% bogland, 10% woodlandClay loam/peat180 m180 cattle (3·7 acre−1)F (95%), AA (5%)
 Once       
R2NitrogenSpring51 kg acre−15451% grass, 44% shrub, 5% boglandPeat182 m33 cattle (0·6 acre−1)H/F× (60%), F (40%)
R3NPK 27·2·5·5May and August70 kg acre−16050% grass, 50% boglandPeat200 m45 cattle (0·7 acre−1)L (40%), C (30%), AA (15%), S (15%)
R4NPK 18·6·12March and April152 kg acre−1Early spring and autumnTwiceSus-releasebolus8250% grass, 10% crop, 5% woodland, 35% boglandPeat152 m80 cattle (1 acre−1)H (60%), C (40%)30 sheep
O110043% grass, 18% woodland, 39% fenSandy loam50 m30 cattle (0·3 acre−1)AA (100%)14 horses
O252100% grassSilt/medium loam50 m36 cattle (0·7 acre−1)L (60%), AA (40%)6 sheep
O33067% grass, 23% crop, 10% wetlandSilt/medium loam152 m15 cattle (0·5 acre−1)H (100%)3 pigs, 100 hens
O48080% grass, 4% woodland, 16% wetlandPeat229–259 m10 cattle (0·1 acre−1)J (50%), H (50%)

To minimize the effects from variation in other environmental factors, the study sites were selected to be as similar as possible in altitude, geographical location and area. However, due to their economical importance in Ireland, the mean size of intensive farms (179 acres) was significantly greater than rough grazing (61·3 acres) and organic farms (65·5 acres) (F2,9 = 11·12; P < 0·01). For a site to be suitable for inclusion, the farm had to be located inland and to have been under the present form of farming practice for at least 3 years. To reduce the possibility of chemical contamination from adjacent intensive farms, we located all sampling sites on organic farms at least 100 m from their outer boundaries.

experimental methods

Sampling of beetles was carried out using a set of 5 replicate baited pitfall traps, each trap located at 5-m intervals along a transect within the site. The pitfall trap design was based on that of Tyndale-Biscoe, Wallace & Walker (1981). Five percent chloral hydrate was used as the preserving fluid. Each set of traps were located in an area of suitable open pasture, running alongside an existing field perimeter fence, and were baited with 1 L of fresh cow dung (collected from the same farm where the traps were set), wrapped in one layer of muslin. The field perimeter where traps were located was chosen to be typical of that farm to avoid any bias resulting from remnant habitat patches. All farms were rotationally grazed and traps were located in fields in close proximity to cattle. All fields were open pasture although it should be noted that farm O1 was characterized by having well developed hedges around all pastures. Electric fencing was erected to prevent any cattle damage to the traps. The contents of the traps were collected every 10 days at each site and traps were re-baited with fresh dung every 20 days. The trapping commenced on 6 May 2000 and was completed by 4 October 2000, giving a total of eight trapping periods for each site. Due to the distance between sites and the fact that traps had to be baited between 08·00 and 12·00 hours, baiting of the traps was divided equally over two consecutive days (i.e. six sites were baited one day, the remaining six sites the next day). Any possible bias of trapping on the same farm types was avoided by baiting two of each farm type on each of the consecutive days. After separating adult dung beetles from the contents of the traps in the laboratory, specimens were stored in 70% alcohol for later identification and enumeration.

The dung quality (nitrogen levels, organic matter and moisture content) was examined for each farm. A 50–80 g subsample was taken from each dung sample collected and was dried to constant weight at 60 °C to determine moisture content. The subsample was then ground in a mechanical mill. Organic matter content was determined using the loss-on-ignition method. Approximately 1 g of dry subsample was weighed in a pre-cleaned, pre-dried and pre-weighed crucible, and then placed in a muffle furnace at 550 °C for 6 h. To establish nitrogen levels, 0·25 g of dry subsample was digested with 1 Kjeltab (Stieg 1999) and 6 mL of concentrated sulphuric acid at 420 °C in a Tecator 2040 digestor for 1 h. The sample was then filtered through GF/C filter paper and the filtrate was diluted to 250 mL with deionized water. Stock blank was prepared in the same way as samples but without the sample added. A certified reference material Cat. No. CRM129, which is hay powder obtained from the European Commissions Bureau of Common Research, was treated in the same way as the samples. Blank and standards were made up in such a way that their matrix was the same as that for the samples analysed. This was necessary to offset the matrix effect, which is caused by different elements in the sample, and to produce accurate results (Stieg 1999). The chemical composition of the organic matter was calculated on the basis of original dry weight and weight of ash remaining after ignition. A Lachat FIA was used to analyse the prepared samples for nitrogen content (Shaw, Karlsson & Müller 1988).

A colonization experiment was also undertaken to investigate the relative importance of variation in dung type, as against habitat, under the three farm management regimes. Three sites were chosen at random from the above 12 sites, one of each farm practice. At each site, five homogenized dung pads from each farm type (n = 15) were arranged randomly in a grid, each pad located at 5-m intervals from one another to assess relative colonization rates of different dung types. One-and-a-half litres of fresh thoroughly mixed intensive, organic or rough grazing dung was deposited in a plastic former of 22 cm diameter to form each pad. Pitfall traps (three for each dung type, n = 9) baited with the different dung types were run concurrently with the experiment to assess beetle activity. The pitfall traps, and the dung pads with the underlying soil to a depth of c. 5 cm, were removed from the sites two days after the experiment began (cf. Gittings & Giller 1998; Finn & Giller 2000) and were immediately transported to the laboratory and stored briefly at 4 °C before extracting the dung beetles by floatation in a bucket full of water. The underlying soil was sorted by hand for any remaining beetles. The colonization experiment was carried out over a period of 3 days at each site in August 2000. Deposition and collection of the dung pads was conducted between 09·00 and 12·30 hours.

data analysis

To estimate species abundance, richness and diversity of dung beetles, samples were analysed at two scales: (i) separately for each farming regime pooled over its constituent sites and time, and (ii) for individual sites pooled over time. Rank abundance distributions were used to compare the overall beetle assemblages of the three farming regimes, thus taking into account both species number and their abundance. For analysis of overall abundance and biomass, species were divided into Aphodius and non-Aphodius groups and calculations of biomass were based on dry weights of species (Gittings 1994).

Analysis of long-term pitfall trapping data is complicated because of seasonal changes in beetle activity (see Gittings & Giller 1997; Finn, Gittings & Giller 1998). Four time periods were identified by Finn et al. (1998) during which Aphodius species composition is relatively constant, and between which distinct breaks in species composition may be identified (i.e. spring, early and late summer, and autumn). The identification of these periods was generally conservative so as to exclude transitional periods (for further details see Finn, Gittings & Giller 1999). The abundance and biomass data for each of these time periods were obtained by summing the number/biomass of captured beetles per trap across the relevant time period. For the pitfall traps, data were analysed using a single GLM. Both the factors ‘season’ and ‘farm type’ were considered fixed, whilst the individual ‘farm’ factor was random and nested within farm type. Cochran's test (Winer 1971) was used to test for heterogeneity of variance and, where necessary, data were transformed. Multiple comparisons of levels within significant factors were made using Student Newman Keuls (SNK) tests. From the mean square estimates, a quantitative measure of the variation associated with each factor in the analysis can be obtained. Components of variation were calculated for relevant factors using the hierarchical model described by Winer (1971). The percentage of variation for each level of the analysis was calculated as that component of variation divided by the sum of all components of variation multiplied by 100. For the colonization experiment, data were analysed using a 2-factor anova with ‘dung type’ and ‘farm type’ as fixed factors. Transformation of data and comparisons followed the same procedure as above.

To examine the relative assemblage similarity across the different farming practices and seasons, a Detrended Correspondence Analysis (DCA) was performed on log-transformed abundance data, with rare species down-weighted. For the 15 most common species (n ≥ 45 individuals) from the overall data set, analyses of variance were conducted to examine how individual species responded to farm practice.

Results

dung beetle species

Over the entire sampling period a total of 39 631 adult dung beetles belonging to 24 different species were collected. The means (± SE) for each species at each farm type are shown in Table 2. Aphodius was the dominant genus comprising 85·9% (n = 34 050) of all trapped individuals, followed by Sphaeridium 10·3% (n = 4079), Margarinotus 2·7% (n = 1053), Onthophilus 1% (n = 394) and Geotrupes 0·1% (n = 52). The occurrence of species of the genera Onthophagus (n = 1) and Serica (n = 2) was extremely rare. Aphodius prodromus was the most abundant species trapped (n = 10 674), of which 72% were trapped on organic sites. Aphodius rufipes (n = 8988) and A. depressus (n = 8580) were also highly abundant, particularly on organic and intensive farms. Eight species (A. prodromus, A. sphacelatus, A. ater, A. rufipes, A. depressus, S. lunatum, S. scarabaeoides and M. carbonarius) accounted for 94% of the captures, but their relative dominance varied among farm type (Table 2). Overall, 54% (n = 21 316) of beetles were collected from organic sites, 30% (n = 11 953) from intensive sites and 16% (n = 6362) from rough grazing sites. The rank abundance distributions of dung beetle assemblages were similar amongst all farm types over the first 9–10 species, but the intensive and rough grazing assemblages were truncated compared with the organic farms which had a ‘tail’ of rare species (Fig. 1). Indeed, eight of the 24 species in this study were collected solely from organic sites. Although there was significant variation in abundance and biomass between farms within a farm type (see below), the rank abundance distributions were very similar between farms within a farm type (Fig. 1). The distribution pattern observed for the organic assemblages tends towards that of a log normal distribution. This implies that individuals are distributed between species in accordance with the normal distribution, where few species are very abundant or extremely rare and most species are moderately abundant (Magurran 1988). The intensive and rough grazing distributions were closer to MacArthur's broken stick model implying that interactions are not strong among species in the community and that the limited resources are apportioned more evenly and simultaneously rather than sequentially among the species (Magurran 1988).

Table 2.  Mean abundance (± SE) of dung beetle species from intensive, rough grazing and organic farms (n = 4 in each case). The number in parentheses is the percentage of species total across all sites
SpeciesMean abundance
IntensiveRough grazingOrganic
Aphodius prodromus (Brahm)461.0 ± 89·1 (17)284·5 ± 152 (11)1923.0 ± 54 (72)
A. sphacelatus (Panzer) 75·3 ± 32·7 (11)152·5 ± 97·6 (22) 467·8 ± 256·3 (67)
A. ater (Degeer)128.0 ± 8·1 (41) 60·3 ± 14·6 (20) 118·3 ± 30·2 (39)
A. rufipes (L.)850·8 ± 332·9 (38)367·5 ± 136·1 (16)1028·8 ± 496·6 (46)
A. fossor (L.) 52·3 ± 19 (47) 17·8 ± 7·5 (16)  40·3 ± 20·2 (37)
A. depressus (Kugelann)766·5 ± 75 (36)496·5 ± 153·4 (23) 882.0 ± 384·4 (41)
A. fimetarius (L.) 18·8 ± 2·2 (41)  8·5 ± 1·7 (19)  18·3 ± 9·5 (40)
A. rufus (Moll) 10.0 ± 10 (35)     –  18·3 ± 11·7 (65)
A. merdarius (Fabricius)     –     –   0·8 ± 0·8 (100)
A. pusillus (Herbst)  1.0 ± 0·7 (5)     –  18·5 ± 10·2 (95)
A. equestris (Panz.)     –     – 242·5 ± 198 (100)
A. erraticus (L.)  0·3 ± 0·3 (33)     –   0·5 ± 0·5 (67)
A. contaminatus (Herbst)     –     –   2.0 ± 1·2 (100)
A. fasciatus (Olivier)     –     –   0·5 ± 0·5 (100)
Sphaeridium lunatum Fabr.275·3 ± 166·2 (43)108·8 ± 35·2 (17) 253.0 ± 158·9 (40)
S. scarabaeoides (L.)215.0 ± 135 (56) 25·5 ± 9·3 (7) 142.0 ± 87·2 (37)
S. bipustulatum Fabr.     –     –   0·3 ± 0·3 (100)
Geotrupes spiniger (Marsham)  0·5 ± 0·3 (4)  1·5 ± 1·5 (13)   9·5 ± 4·1 (83)
G. stercorosus (Scriba)     –     –   0·3 ± 0·3 (100)
G. stercorarius (L.)     –  0·3 ± 0·3 (20)   1.0 ± 0·7 (80)
Onthophagus similis (Scriba)     –     –   0·3 ± 0·3 (100)
Margarinotus carbonarius (Illiger)133·3 ± 25·5 (51) 43·3 ± 12 (16)  86·8 ± 38·5 (33)
Onthophilus striatus (Forster)  0·5 ± 0·5 (1) 23·8 ± 23·1 (24)  74·3 ± 54 (75)
Serica brunnea (L.)     –     –   0·5 ± 0·5 (100)
Figure 1.

Rank abundance distributions of dung beetle communities for each site. For clarity each farm type is on a separate graph. I = intensive, R = rough grazing, O = organic.

species abundance, richness and diversity

Mean species richness was significantly lower on intensive and rough grazing farms than on organic farms (F2,9 = 13·82; P < 0·01, see Fig. 2a). Dung beetle abundance (Fig. 2b) was also significantly lower on rough grazing and intensive farms than on organic farms (F2,9 = 19·53; P < 0·01). Likewise, mean Aphodius species richness per site (Fig. 2c) was significantly higher on organic than on rough grazing farms (F2,9 = 6·83; P < 0·05; data square-root transformed) but there was no significant difference between organic and intensive farms. The abundance of Aphodius beetles (Fig. 2d) was significantly higher on organic farms than on either of the other farm types (F2,9 = 14·58; P < 0·01). Non-Aphodius species richness (Fig. 2e) was significantly higher on organic than on intensive farms (F2,9 = 4·31; P < 0·05). There were no significant differences in non-Aphodius species abundance (F2,9 = 1·23; P > 0·05) between farm types (Fig. 2f), although the trend was for lower abundance on rough grazing farms. The diversity of dung beetles (alpha diversity index) was significantly higher in organic compared to intensive farms (F2,9 = 6·45; P < 0·05; see Table 3) but there was no significant difference between rough grazing farms and either intensive or organic farms.

Figure 2.

Dung beetle species richness and abundance from pitfall traps in differently managed farms. I = intensive, R = rough grazing, O = organic. (a) Species richness and (b) abundance of all dung beetles per farm type; (c) species richness and (d) abundance of Aphodius species per farm type; (e) species richness and (f) abundance of non-Aphodius species per farm type. Treatments with different letters above standard error bars are significantly different (P < 0·05 from anova).

Table 3.  Characteristics of the dung beetle assemblages for each site where n = the total abundance of beetles trapped over the entire trapping period, SR = species richness, α = alpha diversity index, and TSR = total species richness for each farm type (I = intensive, R = rough grazing and O = organic)
SitenSRαTSR
I12823101·3115
I22525111·48 
I33436131·72 
I43169121·58 
R11911121·7113
R21107101·52 
R3 809111·80 
R42535111·48 
O14350172·2624
O27011172·10 
O34639131·64 
O45316162·04 

The mean abundance and biomass of dung beetles throughout the sampling period varied significantly between farm type and season, with ‘season’ explaining approximately 60% of the variance in five of the six analyses (Fig. 3, Table 4). Despite the fact that there was significant variation amongst sites within a farm type, there was still significant between-farm type variation, hence within-farm type variation was the smaller component and was insufficient to mask the stronger between-farm type variation. A significant interaction (‘season × farm type’) was found for non-Aphodius beetle biomass, indicating that the variation observed between farm types depended on the season (Table 4). The SNK tests examining the differences between farm type within seasons showed that significantly more dung beetles were captured at organic sites than intensive and rough grazing sites in spring and autumn (Fig. 3a). Organic sites also had significantly more beetles than rough grazing sites in early and late summer. Intensive sites had significantly more beetles than rough grazing sites in late summer (Fig. 3a). Significantly more Aphodius beetles were also found at organic sites than intensive and rough grazing sites in spring and autumn, and at organic sites than rough grazing sites in early summer. Also, intensive sites had significantly more Aphodius beetles than rough grazing sites in late summer (Fig. 3c). Although farm type did not significantly affect overall abundance of non-Aphodius beetles, the general trend was for intensive sites to have a higher abundance in spring and early summer due to the presence of large numbers of Sphaeridium species and for organic sites to have a higher abundance of non-Aphodius beetles in late summer and autumn due to a greater number of Geotrupes and, to a lesser extent, Sphaeridium species (Fig. 3e). Total biomass of dung beetles was greater at organic sites than rough grazing sites in spring, late summer and autumn (Fig. 3b). Organic sites also had significantly greater biomass than intensive sites in autumn. Intensive sites had greater beetle biomass than rough grazing sites in late summer (Fig. 3b). The same significant differences between farm types for total biomass were observed for Aphodius biomass (Fig. 3d). Non-Aphodius biomass was characterized by an interaction of the main factors suggesting that the differences in biomass between farm types depended on the season. Thus in general, organic farms held the highest abundance and biomass of beetles and the only obvious exception involved non-Aphodius beetles which were influenced by the high numbers, and hence biomass, of S. lunatum and S. scarabaeoides in intensive farms in spring and early summer. Rough grazing farms were generally the poorest sites for both beetle abundance and biomass (Fig. 3).

Figure 3.

Seasonal dung beetle abundance and biomass in differently managed farms. I = intensive, R = rough grazing, O = organic. The season units are represented by 1 = spring, 2 = early summer, 3 = late summer and 4 = autumn.

Table 4.  GLM analysis of dung beetle abundance and biomass. When no significant interaction between the main factors was present, SNK multiple comparisons are shown. For season, S = spring, ES = early summer, LS = late summer, A = autumn. For farm type, I = intensive, R = rough grazing, O = organic. Significance of the F-values indicated as follows: *P < 0·05, **P < 0·01, ***P < 0·001. Abundance data was square-root transformed. Biomass data was ln(x + 1) transformed
Sourced.f.Total abundanceAphodius abundanceNon-Aphodius abundance
MSFSNKMSFSNKMSFSNK
Season  31150·7425·61***< ES = LS < S1101·2225·62***< ES = LS < S42·8222·08***< ES = LS < S
Farm type  21045·8723·26***R < I < O 962·0515·29**R = I < O24·38 1·42 NSR = O = I
Farm (farm type)  9  44·9711·50***   62·9417·40*** 17·1654·35*** 
Season × farm type  6  69·94 1·56 NS   81·61 1·90 NS  4·19 2·16 NS 
Season × farm (farm type) 27  44·9411·49***   42·9911·88***  1·94 6·14*** 
Residuals192   3·91     3·62   0·32  
Total239         
Sourced.f.Total biomassAphodius biomassNon-Aphodius biomass
MSFSNKMSFSNKMSFSNK
Season  3  77·5240·85***< ES = S < LS  62·2841·22***< ES = S < LS87·2415·01***
Farm type  2  36·8814·51**R < I < O  27·29 9·67**R < O85·42 2·77 NS
Farm (farm type)  9   2·54 8·96***    2·8216·75*** 30·8017·27*** 
Season × farm type  6   1·92 1·01 NS    1·36 0·90 NS 19·92 3·43* 
Season × farm (farm type) 27   1·90 6·69***    1·51 8·97***  5·81 3·26*** 
Residuals192   0·28     0·17   1·78  
Total239         

community ordination

The relative similarity of the assemblages changes across the seasons (Fig. 4). In the spring some separation occurred on axis 2, but most sites tended to cluster along axis 1, with sites O1 and O2 forming outliers, largely due to the occurrence of A. equestris, A. pusillus and A. merdarius in these samples. In the early summer, sites O1, O2, R1 and R4 separated from the main cluster due to the presence of A. equestris and G. stercorosus in O1, the presence of A. pusillus and A. merdarius in O2, and high abundances of A. prodromus, A. sphacelatus and A. rufipes in O2, R1 and R4. In the late summer O1 and O2 again separate from the main site cluster on axis 1 due to the presence of A. equestris and A. pusillus and high abundances of A. rufus and G. spiniger. In the autumn, sites O1, O2, and R1 tended to differentiate from the other sites due to the presence of A. equestris and A. contaminatus in O1 and O2 and G. spiniger in R1. The autumn ordination shows less of an obvious cluster, probably due to the lower abundance of beetles found during this period.

Figure 4.

Detrended correspondence analysis ordination of dung beetle assemblages from differently managed farms. Axis 1 accounted for 34·1% of the species variance, while axis 2 accounted for 21·7%. Key to sample codes: I = intensive, R = rough grazing, O = organic. Key to season codes: ◊ = spring, ▪ = early summer, ▵ = late summer, • = autumn. Key to species codes: ater = Aphodius ater, bip = Sphaeridium bipustulatum, cont = A. contaminatus, dep = A. depressus, eques = A. equestris, errat = A. erraticus, fas = A. fasciatus, fim = A. fimetarius, foss = A. fossor, lun = S. lunatum, merd = A. merdarius, prod = A. prodromus, pus = A. pusillus, ruf = A. rufipes, rufus = A. rufus, scar = S. scarabaeoides, sphac = A. sphacelatus, spin = Geotrupes spiniger, ster = G. stercorarius, sterc = G. stercorosus.

It appears that the organic sites O3 and O4 have a beetle assemblage closely resembling that of intensive and rough grazing farms, although four species of dung beetle captured (A. erraticus, A. fasciatus, S. bipustulatum and Onthophagus similis) at O3 and O4 were not captured (or in the case of A. erraticus, only very rarely captured) at any other site. However, few specimens were caught resulting in little effect on the DCA. Geotrupes stercorarius, while not captured solely at site O4, was captured more frequently at site O4 than any other site.

chemical parameters

Moisture, organic matter and nitrogen content varied significantly within and between farm management (see Table 5 for a summary of the values). Nitrogen levels showed the greatest variability between sites (F11,36 = 281·32; P < 0·001), with intensive and rough grazing sites showing the greatest overlap, and with three of the four intensive farms having a higher nitrogen content than the organic farms. The least variation was observed in the organic matter content (F11,36 = 44·40; P < 0·001), where rough grazing and organic sites generally varied significantly from intensive sites while there was no significant difference within the intensive sites (Table 5). Moisture content was similar amongst most sites apart from O2 and O4 which were significantly different from all other samples (F11,36 = 47·29; P < 0·001). These two sites represented the locations with the highest and lowest dung moisture content, respectively. Intensive, rough grazing and organic sites had a mean moisture content of 89·07%, 88·72% and 88·57%, a mean organic matter content of 81·53%, 86·21% and 83·31%, and a mean nitrogen content of 26·32 mg N g−1, 25·38 mg N g−1 and 22·62 mg N g−1, respectively. Mean values did not vary significantly (moisture: F2,9 = 0·10; P > 0·05; organic matter: F2,9 =  3·02; P > 0·05; nitrogen: F2,9 = 0·84; P > 0·05).

Table 5.  Moisture (Moi), organic matter (Org) and nitrogen (Nit) content of dung samples from the 12 study sites. The mean values (± SD) for each chemical parameter is given for each farm, derived from four subsamples of the farms dung. Means with letters in common are not significantly different across sites for that parameter
 IntensiveRough grazingOrganic
I1I2I3I4R1R2R3R4O1O2O3O4
Moi (%)
88·33 ± 1·2189·07 ± 0·5489·09 ± 0·3789·79 ± 0·0287·52 ± 0·1589·92 ± 0·1488·10 ± 0·4989·32 ± 0·3289.24 ± 0.191·35 ± 0·1688·57 ± 0·1785·13 ± 0·24
aa, ba, bbabaa, ba, bcad
Org (%)81·27 ± 0·3281·54 ± 0·2181·63 ± 0·2181·70 ± 0·3681·83 ± 0·2387·87 ± 0·0688·73 ± 2·7386·40 ± 078.43 ± 0.1584·67 ± 0·483·31 ± 0·286·83 ± 0·35
a
aaaabbb, cec, da, db, c
Nit (mg N/g)19·32 ± 0·226·32 ± 0·1431·07 ± 0·1228·58 ± 0·3527·38 ± 0·7924·67 ± 0·4119·32 ± 0·68 30.15 ± 0.6225·33 ± 0·1823·09 ± 0·4722·61 ± 0·1419·43 ± 0·3
ab, fcdb, deace, fgga

colonization experiment

The abundance of dung beetles colonizing dung pads varied significantly between dung type and farm management (Table 6, Fig. 5a). The variable ‘farm’ explained more of the variance in dung beetle abundance than ‘dung type’ indicating a habitat effect on the colonization of beetles. However, there was also a significant interaction (‘dung type × farm’), indicating that the variation in abundance among different dung types depended on the farm type. The multiple comparisons showed that in rough grazing and intensive management sites there were significantly more beetles colon-izing the organic and intensive dung than the rough grazing dung (Fig. 5a). However, there was no significant difference in beetle abundance between dung types on the organic farm (Fig. 5a). The biomass of dung beetles colonizing dung pads also varied significantly between dung type and farm management (Table 6). ‘Dung type’ (32%) explained more of the variance in biomass than ‘farm’ (26%), indicating a dung type effect on specific species related to site. As there was no significant interaction (‘dung type × farm’) the two factors are independent of one another in relation to biomass. Multiple comparisons showed that in rough grazing and intensive sites there was significantly greater beetle biomass found in the organic and intensive dung than in the rough grazing dung (Fig. 5b). There were also significant differences on the organic site, where the organic dung held a greater beetle biomass than the intensive and rough grazing dung (Fig. 5b).

Table 6.  2-factor anova of abundance and biomass of dung beetles colonizing different dung types under different farm management. Post hoc SNK tests on beetle biomass (pads) were as follows: Dung Type R < I < O and Farm Type O < R = I, where I = intensive, R = rough grazing, O = Organic. Significance of the F-values indicated as follows: *P < 0·05, **P < 0·01, ***P < 0·001. Data for the pitfall traps were square-root transformed
Sourced.f.Abundance – padsBiomass – padsAbundance – pitfallsBiomass – pitfalls
MSFMSFMSFMSF
Dung type 23665·2717·6***1708201·8420·6*** 18·89 37·3*** 560·31 35·6***
Farm type 24734·2022·7***1436094·1117·3***124·16245·1***4144·88263·2***
Dung type × farm type 4 575·87 2·8* 186174·43 2·2 ns  7·33 14·5*** 219·20 13·9***
Residuals36 208·19   83130·85   0·51   15·75 
Total44        
Figure 5.

Mean (± SE) abundance and biomass (mg dry weight) of dung beetles captured in dung pads and dung-baited pitfall traps for the three dung types sampled at each of the different farming regimes. I = intensive, R = rough grazing, O = organic. Treatments with different letters above standard error bars are significantly different (P < 0·05 from SNK).

pitfall traps

The abundance of dung beetles attracted to dung baited pitfall traps varied significantly between dung type and farm management (Table 6, Fig. 5c). As for the dung pads, the variable ‘farm’ explained more of the variance in abundance than ‘dung type’. However, there was also a significant interaction (‘dung type × farm’), indicating that the variation in abundance among different dung types depended on the farm type. Multiple comparisons showed that on the intensive site significantly more beetles were found in the organic dung traps than the intensive and rough grazing traps (Fig. 5c). There were no significant differences in abundance between dung types in the rough grazing or organic sites (Fig. 5c). The biomass of dung beetles attracted to dung baited pitfall traps also varied significantly between dung type and farm management (Table 6). Again, the variable ‘farm’ explained more of the variance than ‘dung type’. However, there was a significant interaction (‘dung type × farm’) indicating that the variation in biomass among traps baited with different dung types depended on the farm type. Multiple comparisons showed that significantly greater beetle biomass was found in the organic dung traps on the intensive site and in the organic dung traps compared with the intensive dung traps on the rough grazing site (Fig. 5d). There was no significant difference in beetle biomass between dung types on the organic site.

Discussion

Our results have shown that the abundance, biomass, richness and diversity of dung beetle species was higher on organic than on intensive and rough grazing farms. The dung beetle community structure across all three farm types was typical of north European temperate communities in that it was dominated by Aphodiidae (Hanski & Cambefort 1991). Whilst there appears to be some variation between farms within a farm type in abundance and biomass (see also below), the overall community structure within each farm type appears more stable, as shown by the rank-abundance distributions and the community ordination data. The within-farm type variation is the smaller component and is insufficient to mask the larger between-farm type variation.

The colonization experiment illustrated the importance of both the habitat differences in the farm or surrounding habitat (i.e. variation in the potential species and population pools available), as well as in the different dung types. Dung beetles are known to prefer certain dung types to others (cf. Gittings & Giller 1998; Finn & Giller 2002), even though they were found in all the dung types in the present study. Generally, significantly greater abundance and biomass of dung beetles were found in dung from organic and intensive farms than from rough grazing farms at each of the farm types. The type of dung was more important when considering beetle biomass, as this explained more of the variation than the type of farm sampled.

One area of debate that needs further investigation is the use of veterinary medicinal products to control endoparasites in farm livestock (e.g. avermectins). These products may be excreted in the dung for a period of time after treatment, which in turn leads to dung-inhabiting invertebrates being exposed with the potential for toxic effects. Errouissi et al. (2001) showed that ivermectin, in the form of a sustained-release bolus, was highly effective in killing Aphodius constans larvae for approximately 143 days after treatment. Wardhaugh, Longstaff & Morton (2001) compared pour-on formulations of eprinomectin and moxidectin on the development and survival of the dung beetle Onthophagus taurus. While moxidectin had no detectable effects, eprinomectin caused high juvenile mortality during the first 1–2 weeks after treatment. Through modelling, the authors predicted that eprinomectin was capable of reducing beetle activity in the next generation by 25–35%. In relation to the present study, this could partially explain the low numbers of second generation A. prodromus and A. sphacelatus (that appear in the autumn) on sites (i.e. I1 and R4) that applied ivermectin in spring (when the first generation of A. prodromus and A. sphacelatus appear). This pattern was observed in the community ordination where in late summer and autumn, clear separation occurred between sites that administered ivermectin and those that did not, particularly on intensive farms.

When comparing the community structure across the three farm types, eight species (including one chafer) were found solely on organic farms. The community ordination shows clear separation of sites O1 and O2 from the majority of other sites. While it would be erroneous to suggest that these species are only associated with organic farms, it strongly suggests that certain facets of organic farming are beneficial to the biodiversity of north temperate dung beetles (be it farm management, dung type or local or surrounding habitat type). Other studies that have compared organic/low intensive farming to conventional/high intensive farming have found similar results for a variety of other organisms. Blackburn & Arthur (2001) found a significantly greater density of centipedes in organic farm field margins compared with conventional farms, most probably as a result of pesticide use. The results obtained by Freemark & Kirk (2001) revealed that species richness and total abundance of birds was significantly greater on organic than conventional farms. The authors emphasized the importance of non-crop habitats and less intensive management practices to the conservation of avian biodiversity on farmland. Peña et al. (2003) demonstrated that agricultural landscapes in western France transformed by intensification, field enlargement and removal of hedges, have affected species composition and reduced abundance of carabid beetles.

In Ireland, intensive farming has been heavily criticised because of its tendency to reduce the diversity of the landscape by removing valuable habitats (e.g. hedges, ditches and woodland). It has resulted in a loss of species, for example the corncrake Crex crex Linne (Green 1999). Based on the results of the present study, it appears that such changes may also have reduced species richness of dung beetles. Rushton et al. (1989) demonstrated that species composition and diversity of ground beetles and spiders decreased when upland pastures were agriculturally improved. The resulting pasture was more similar to that of intensively managed pastures at lower altitudes. Intensive farms tend to replace field boundaries with electric fences that greatly reduce, or completely eradicate, the shade capacity of field boundaries, such that species that are shade specialists might become excluded from large areas of pasture (Horgan 2002). This could explain the absence of species from certain sites in this study. Field boundaries of organic farms were more diverse, consisting of a mixture of shrubs and trees providing cover for shade specialist species such as A. equestris and A. fasciatus. A total of 970 specimens of A. equestris were captured from two organic farms, which had the highest quantity of vegetative cover compared to that of any other site (S.A. Hutton, pers. obs.). Sampling effort for each site was identical, supporting the notion that features of farm habitat are playing an important role in the distribution of specialist dung beetles and hence the diversity and richness of the dung beetle community as a whole.

One further explanation may lie in the nature of the soils. Rough grazing sites are generally located on low quality peat soils, which are prone to flooding. Regular flooding of soils could cause serious implications for dung beetle populations. Most Aphodius species lay their eggs in the underlying soil beneath the dung pad and Geotrupes bury dung in brood masses in the soil underneath the dung pad, where their offspring develop. Typhaeus typhoeus (a member of the family Geotrupidae) larvae develop in brood masses up to 40 cm below the soil surface, provided the water table was permanently below 80 cm (Brussaard & Slager 1986). In this present study, 83% of the total abundance of Geotrupes species were found on organic sites, 13% on rough grazing sites and 4% on intensive sites. These species, being much larger than Aphodius, can remove substantial proportions of single dung pads, so these results may be of importance in terms of dung breakdown potential.

Species abundance values for rough grazing sites were generally low. In fact, rough grazing sites had approximately 50% fewer beetles than intensive sites. In addition, organic farms had approximately 50% more beetles than intensive sites. Further, given the fact that the differences were found in both pads and pitfall traps (with and without beetle–dung interactions, respectively) it is suggested that the chemical differences between the dung types are sufficient to allow ready detection by the beetles and thus explaining the variable attractiveness.

Previous studies in southern Ireland have shown dung type preferences by colonizing dung beetles. Gittings & Giller (1998) investigated dung beetle colonization of dung from cows compared to a range of exotic animals and found that dung beetle species usually displayed distinct and consistent preferences in their colonization of particular types of dung. Finn & Giller (2002) examined dung beetle colonization of native herbivore dung and discovered that sheep dung pats generally had the highest values of total beetle biomass, compared to that from cows and horses. While this area of research was not explored in the present study, sites that contained a mixture of livestock (including horses, sheep and pigs) tended to have greater beetle abundance, biomass, diversity and species richness, supporting previous work (Barbero et al. 1999).

conservation

Conservation biologists are concerned with the factors necessary for population persistence. Adult dung beetles use dung both for feeding and laying eggs. Accordingly, the first step in the conservation of dung beetles is to facilitate the choice of dung available to adult beetless. One would expect, from the preferences shown by the adults, that larval abundance and survival would be greatest in organic dung and that features of organic farms could benefit their survival. For example, dung in the shade from field boundaries dries at a slower rate than exposed pats, potentially increasing the larval feeding time available. It follows that patchy ecosystems characterized by open pasture and dense surrounding vegetation and inhabited by several ungulate species (i.e. cattle, horses and sheep) can support the highest levels of dung beetle diversity and hence improve dung decomposition.

Our results emphasize the importance of vegetative field boundaries and less intensive management practices for the conservation of dung beetle biodiversity. The fact that many beetle species were more abundant on organic than intensive/rough grazing sites, suggests that increasing the area of land under an organic or ecological farming regime (low input system encouraging diverse ecosystems) might increase regional beetle populations. Currently, as in other countries, the acreage of land managed organically in Ireland is relatively small and is therefore unlikely to have much effect on dung beetle populations except locally.

Intensive farms will always be a feature of the Irish landscape because of their economic importance; however, from a biodiversity perspective, it would appear essential that the rate at which intensive farms are expanding and engulfing valuable habitats should be curtailed. This study has shown that intensive farms support 38% less dung beetle species than organic farms, giving an indication of the potential biodiversity loss (e.g. decline of Aphodius erraticus, an important species in the nutrient cycle of pasture ecosystems). Farm inputs also need to be controlled at levels which allow for high levels of diversity of dung dwelling species, and their associated species.

Acknowledgements

The authors would like to thank Áine Healy, Daithí McSweeney, Sofia Gripenberg and all the farmers for their cooperation and assistance with this study. The authors also thank three anonymous referees for their valuable comments. This work was funded by the Irish Heritage Council.

Ancillary