Introduction
Dung is a rich resource. It can be highly abundant and relatively predictable in its occurrence compared with other patchily distributed resources, such as carrion, particularly where large populations of grazing vertebrates are present. Herbivore dung is very similar to leaf litter in composition, consisting mainly of water and undigested plant material along with the products of metabolism, gut epithelial cells and a distinct array of micro-organisms (Marsh & Campling, Reference Marsh and Campling1970; Greenham, Reference Greenham1972; Stevenson & Dindal, Reference Stevenson and Dindal1987; Aschenborn et al., Reference Aschenborn, Loughnan and Edwards1989). As a result, the insect community that inhabits this environment is particularly diverse, consisting of as many as 400 species (Hanski, Reference Hanski, Hanski and Cambefort1991; Skidmore, Reference Skidmore1991). As well as high species diversity, the dung environment supports a large number of individuals. It has been estimated that a single cow can support an insect community at least one fifth its own weight purely on the food left in its faeces (Laurence, Reference Laurence1954) and an individual dung pat may contain 1000 or more insect inhabitants (Laurence, Reference Laurence1954; Hanski, Reference Hanski, Hanski and Cambefort1991).
The dung-colonizing insects play a particularly important role in ecosystem function; the timely decay of plant and animal remains is essential to the carbon and nitrogen cycles, soil fertility and the population dynamics of a wide diversity of species at a range of trophic levels. The importance of these invertebrate colonizers in pat decomposition has been clearly demonstrated by exclusion experiments, where mesh covers have been used to prevent colonization. Lumaret & Kadiri (Reference Lumaret and Kadiri1995) found that pats from which all insects were excluded for one month, took 1.7 to 2.2 times longer to completely disintegrate than uncovered pats. Holter (Reference Holter1979) also showed that selective exclusion of nocturnal colonizers, particularly Aphodius rufipes L., by covering pats overnight during the first week, resulted in only half of the disappearance found in freely exposed pats. Exclusion of insects from cow pats for only two days following deposition resulted in a significant increase in the amount of dung that remained after 35 days (Lee & Wall, Reference Lee and Wall2006a). Furthermore, several authors have shown that a range of insecticides and anthelmintics administered to livestock and excreted in faeces may, under some circumstances, kill dung-colonizing insects and retard decomposition (Wall & Strong, Reference Wall and Strong1987; Madsen et al., Reference Madsen, Overgaard Nielsen, Holter, Pederson, Brochner Jespersen, Vagn Jensen, Nansen and Gronvold1990; Sommer et al., Reference Sommer, Steffansen, Nielsen, Gronveld, Vagn Jensen, Jespersen, Springborg and Nansen1992; Strong et al., Reference Strong, Wall, Woolford and Djeddour1996; Floate, Reference Floate1998; Floate et al., Reference Floate, Wardhaugh, Boxall and Sherratt2005).
While tropical dung communities dominated by dung-burying scarabaeid beetles can completely remove a dung pat in a few hours (Anderson & Coe, Reference Anderson and Coe1974; Hanski, Reference Hanski, Hanski and Cambefort1991), dung disappears much more slowly in temperate regions, over periods of weeks or months, or even sometimes years given the occurrence of unfavourable winter or drought conditions. This slower decomposition allows a complex community of insects to develop in and around the dropping during its degradation. Even in temperate regions, however, the successional changes in the dung usually make it suitable for only one generation of most insects, so individuals must be able to disperse and colonize new pats during their life. Hence, selection pressures acting on adults and larvae are very different. Adults must be highly mobile and able to find new resource (dung) patches in which to lay eggs, while larvae do not disperse far, but must complete their growth before the dung is exhausted or no longer suitable for their development. Clearly, these development rates and selective pressure also vary with the colonising species; egg-to-adult development for dung beetles may require several months, whereas the development of larval flies may require only one to two weeks. During this period, larvae must also be able to survive high levels of predation within the pat.
The dung-breeding flies Neomyia cornicina (Fabricius) and N. viridescens (Robineau-Desvoidy) (Diptera: Muscidae) are widespread and abundant in cattle pastures. Neomyia cornicina is widely distributed throughout the Holarctic, Neotropical and Oriental regions, while N. viridescens is thought to be restricted to the Palearctic. They can be seen alighting on fresh droppings and are conspicuous iridescent green-coloured muscids, with a characteristic iridescent frons and face. Adult females oviposit batches of eggs in cavities near the surface of the pat and the egg batches of several females may be aggregated in the same cavity (personal observation). Neomyia spp. are important components of the dung community but have also been used as indicator species in studies of the cow-dung invertebrate community, since they are highly sensitive to the presence of insecticide residues (Gover & Strong, Reference Gover and Strong1995, Reference Gover and Strong1996; Sommer et al., Reference Sommer, Vagn Jensen and Jespersen2001; Iwasa et al., Reference Iwasa, Nakamura, Fukaki and Yamashita2005; Lumaret et al., Reference Lumaret, Errouissi, Galtier and Alvinerie2005). However, despite being abundant in the field and widely used in the laboratory, the basic ecology of these two species is not well understood; N. viridescens in particular has been little studied. As a complicating factor, the nomenclature and taxonomy of these species has been confused; the classic study by Hammer (Reference Hammer1941) in Denmark uses the name Cryptolucillia caesarion. The name Orthellia caesarion has been particularly widely employed (e.g. Stoffolano & Streams, Reference Stoffolano and Streams1971) as has Orthellia cornicina (Wardhaugh & Rodriguez-Mendez, Reference Wardhaugh and Rodriguez-Mendez1988). There is one report which uses the term Orthellia viridis for flies collected in the south of France (Kirk, Reference Kirk1992). Whether the various Palearctic studies examined only N. cornicina or did not differentiate N. cornicina from N. viridescens is unknown, but the latter is probable. The aim of the present work, therefore, was to examine the seasonal abundance and reproductive output of both N. cornicina and N. viridescens in cow dung in pastures in southwest England to attempt to identify possible ecological differences between these two species.
Methods and materials
Seasonal abundance
The pattern of seasonal abundance was examined in an area of permanent grassland pasture on a farm located approximately 20 km southwest of Bristol, UK. The pasture used was grazed by a dairy herd of about 250 Holstein-Friesian cows and a small number of sheep. The majority of the cattle were let out to pasture in mid-May and were rotated between contiguous fields until mid-October, when they were brought back in for the winter. The cattle were not treated with anthelmintics while maintained in the milking herd.
Batches of ten artificially-constructed cow pats, formed from fresh dung, were placed out each week between the 21st May and 29th October 2001 to allow for insect colonization. When pats were required, fresh dung from several cows was collected from the milking parlour during afternoon milking. This was thoroughly mixed to ensure uniform constituency and texture and used immediately. Using a hand-held spring balance, 1.5 kg of fresh dung was weighed out and circular pats, 4–5 cm deep with a diameter of approximately 19 cm, were produced using a polythene former. The former was removed once the pat had been created. Plastic netting (2 cm mesh) was placed under the pats to assist with their recovery; the netting was not considered likely to have affected invertebrate movement, given the size of the dung-colonising species relative to the mesh width. Samples of the dung used were brought back to the laboratory for analysis of their water content (Lee & Wall, Reference Lee and Wall2006b). Throughout the study, daily temperature and humidity were recorded at the study site using an automated weather station (DataHog, Skye Instruments Ltd, Llandrindod Wells, UK).
Artificial pats were left exposed in the field for one week to allow colonization, following which they were retrieved. After collection, each pat was returned to the laboratory, placed onto a thin layer of sawdust on a polythene sheet and put in an individual fine-mesh bag, at approximately 20–25°C, to await the emergence of invertebrate colonizers. The mouth of the mesh bag was attached to a plastic collecting beaker and, as they emerged, insects were funnelled from the mesh bag into the beaker, where they quickly died. This funnel system helped to ensure that there was no recolonization of dung by newly-emerged insects in the laboratory. All insects to emerge were collected and counted; however, only the data for the adult Neomyia are presented here. Details of the other insects recovered from these pats are given elsewhere (Lee & Wall, Reference Lee and Wall2006b).
All adult Neomyia to emerge were identified to species level. The two Neomyia species were differentiated using the chaetotaxy of the thorax; N. cornicina has two pairs of dorso-central and two single acrostichal bristles on the prescutum, whereas N. viridescens has the pairs of dorso-central bristles only.
Comparison between farm types
In 2004, Neomyia abundance was studied in two areas of permanent grassland pasture. The first site was the same as was used in 2001 (farm 1), while the second was an organic beef farm located approximately 5 km north of Bristol (farm 2), and grazed by approximately 130 South Devon beef cattle over an area of 210 acres. In this year, batches of 1 kg cow pats (approximately 15 cm in diameter) were formed each week from fresh dung as described above. Five pats were placed out each week at each farm between the 28th of June and the 3rd of September. The artificial pats were left exposed in the field for 3–5 days to allow colonization, following which they were retrieved, returned to the laboratory and treated as described previously.
Reproductive output in the field
In 2004, at both farms, flies were collected at 3–5 day intervals using a hand-held net as they alighted on cow pats. In addition, at farm 1, flies were trapped using 40-cm-sided square white targets (Wall et al., Reference Wall, Green, French and Morgan1992) covered by a polybutene-based non-setting adhesive (Oecotak, Oecos Ltd, Kimpton UK). Each target was baited with 500 ml of fresh cow dung which was replaced weekly. Five targets were placed vertically around the edges of the cattle pastures at least 100 m apart with the support-pole pushed into the ground, so that the target base was approximately 30 cm above ground level. Targets were inspected every 3–4 days from July to September 2004. At each inspection, all green-coloured Diptera were removed and returned to the laboratory where Neomyia were identified, sexed and counted under a binocular microscope.
For each female collected in a hand-net or on the sticky-target, its left wing was removed and mounted on a sheet of paper using sticky-tape. The wing was then examined under a binocular microscope and the length of the posterior cross vein between the fourth and fifth longitudinal veins (dm-cu between veins CuA1 and M; McAlpine, Reference McAlpine, McAlpine, Peterson, Sherwell, Teskey, Vockeroth and Wood1981) was measured to give an index of size. Subsequently, the abdomen of each female was detached from the thorax using a pair of dissecting scissors and placed in Ringer's solution (0.9% saline) on a glass microscope slide. Using mounted needles and under a dissecting microscope, the ovaries were gently removed from the abdomen and teased apart. Egg follicles within the ovaries were measured to the nearest 0.025 mm using an eye-piece graticule. The number of mature eggs was counted.
Results
Seasonal abundance and distribution
A total of 1681 N. cornicina and 123 N. viridescens emerged successfully from the 240 artificial cow pats placed out during the summer of 2001. Individuals of both Neomyia species first appeared in the cow pats placed out into the field on day 56 (25th June), with low numbers (15 or less) in each batch until those placed out between days 119 and 133 (27th August–10th September), when they showed a sharp peak followed by a rapid decline (fig. 1). Multiple regression analysis found no significant effects of mean temperature, mean relative humidity or dung water content on either the log10 transformed abundance or the abundance data after it had been de-trended to remove the seasonal pattern. For the later analysis, a third order polynomial regression was plotted through the seasonal abundance data, and the residuals were then treated as the independent variable in the multiple regression analysis.
The numbers of Neomyia cornicina (●) and N. viridescens (○), recovered from batches of ten artificial 1.5 kg cow pats placed out at weekly intervals at farm 1 in southwest England between May and November (day 1: 1st of May 2001). Zero counts not shown.
Neomyia cornicina was present in 31 pats (13%); and, where it was present (excluding zero counts), a median of 19 adults (interquartile range 54) emerged from each pat. Although numbers were generally relatively low, for one pat placed out on day 126 (3rd September), 266 N. cornicina were recovered from this pat alone (fig. 2). Neomyia viridescens was present in 18 pats (8%); and, where it was present (excluding zero counts), a median of three adults (interquartile range=6) emerged from these pats (fig. 2). Both species were found together in only nine pats. Analysis of variance of the log10 (+1) transformed number of each species showed that N. cornicina were significantly more abundant than the N. viridescens (F=15.2, P<0.001).
The detransformed mean log10 number (+1) of Neomyia cornicina (●) and N. viridescens (○) (±detransformed 95% confidence intervals) recovered from batches of five artificial 1 kg cow pats placed out at weekly intervals at farms 1 and 2 in southwest England between June and September (day 1: 1st of May 2004).
Of the 94 pats put out in 2004, N. cornicina emerged from 43 (46%). A median of three adults emerged from each pat (interquartile range=5) and only six pats gave rise to more than ten N. cornicina (fig. 2). Neomyia viridescens emerged from only 11 of the 94 pats (12%) and all of these gave rise to less than five adults, at a median of one adult per pat (interquartile range=1) (fig. 2). Both species were found together in only five pats. Again, analysis of variance of the log10 transformed number of each species showed that there were significantly fewer N. viridescens than N. cornicina at both farms (F=34.03, P<0.001) but that the numbers of both species were higher at the dairy farm than the organic beef farm (F=5.42, P=0.02; fig. 3). There was no significant interaction between farm type and Neomyia species (F=2.87, P=0.09).
The number of pats from which adult Neomyia cornicina (solid bars) and Neomyia viridescens (open bars) emerged in frequency classes of five, in 2001 (upper) and 2004 (lower).
If oviposition by the two species had occurred at random in the available pats (240 in 2001 and 94 in 2004), the probability of co-occurrence in each year can be calculated as the product of the proportionate occurrence of each species multiplied by the number of pats available, which gives an expected co-occurrence in two pats in 2001 and five pats in 2004. This can be compared to the nine and five pats where co-occurrence was actually recorded in 2001 and 2004, respectively. Chi-squared analysis (with Yeats' correction for small sample size) suggests that the incidence of co-occurrence was significantly greater than expected in 2001 (χ2=21.1, P<0.001) but not in 2004.
Reproductive output
One hundred and thirty nine female N. cornicina were caught by the traps and 23 by the hand-nets; 39 (28.1%) and 12 (52.2%) of these females contained mature eggs, respectively. There were no differences in the size of the females or numbers of eggs they were maturing for females caught by the two capture methods, which were, therefore, pooled for subsequent analysis. Overall, a mean of 28.8 (±13.2) eggs were matured by the gravid female N. cornicina and larger females matured significantly larger numbers of eggs (F=17.25, P<0.001; fig. 4). This relationship persists even if the three largest females are removed from the analysis.
The numbers of eggs matured by adult female Neomyia cornicina of differing body size, where an index of size was measured as the length of the posterior cross vein between the fourth and fifth longitudinal veins (dm-cu between veins CuA1).
A total of 27 N. viridescens were caught by the traps and 17 in the hand net. Of these, 11 in the trap were gravid (40.7%) as were 13 of those caught by the hand net (76.5%). There were no differences in the size of the females or numbers of eggs they were maturing for females caught by the two methods, which were, therefore, pooled for subsequent analysis. Overall, a mean of 37.1 (±16.9) eggs were matured by gravid N. viridescens females, but there was no significant relationship between female size and the number of eggs matured (F 23=0.83, P<0.37).
The female N. viridescens were significantly larger than the N. cornicina (F=29.1, P<0.001) and matured significantly higher numbers of eggs (F=5.33, P=0.02; fig. 5).
Distribution of egg batch sizes for Neomyia cornicina (open bars) and N. viridescens (solid bars).
Discussion
Despite the large number of studies that have been undertaken on the biology of components of the cow dung colonizing invertebrate community, for many species little is still known. Adults of N. cornicina and N. viridescens are good examples of such common dung inhabitants, where information on their basic ecology and functional contribution to the decomposition process is lacking; indeed, the literature would suggest that in many studies the two species are seldom differentiated. The present study, however, has shown that there are a number of fundamental life-history differences, despite their superficial morphological similarity. Adult females of N. viridescens are significantly larger than N. cornicina and mature significantly larger egg batches. A significant relationship between individual female size and the numbers of eggs matured was found for N. cornicina but not N. viridescens, although the latter may simply reflect the relatively low sample size obtained for this species in 2004. It is also worth noting that eggs batches were measured from adult females of unknown age and oviposition history in the field and that this is likely to have accounted for a proportion of the variability seen in these data.
Most dung-colonizing species are found in highly aggregated distributions, but the degree of aggregation may vary widely. In the present study, N. cornicina emerged from 13% and 46% of the pats, while N. viridescens emerged from only 8% and 12% of the pats deployed in 2001 and 2004, respectively. Since most insect larvae cannot disperse between pats, their distribution and abundance is primarily determined by the oviposition behaviour of adults and the levels of mortality within the pat (Sowig, Reference Sowig1996). Females of most insects deposit their eggs in clutches, and clutch size is a fundamental aspect of the life-history of each species (Godfray & Parker, Reference Godfray and Parker1992; Withers et al., Reference Withers, Madie and Harris1997). In patchy ephemeral habitats such as dung, the average patch quality, the probability of finding a suitable patch and the distance between patches influences the range of strategies that are successful (Beaver, Reference Beaver1977; Hanski, Reference Hanski, Gray, Crawley and Edwards1987; Withers et al., Reference Withers, Madie and Harris1997; Thiel & Hoffmeister, Reference Thiel and Hoffmeister2004). Species breeding in rarer, less predictable but high quality habitat patches should typically deposit their eggs in large clutches; while smaller clutch sizes are expected where patches are more frequent and predictable, or of lower quality (Godfray et al., Reference Godfray, Partridge and Harvey1991; Heard, Reference Heard1998). Carrion breeding blowflies produce relatively large clutches, of between 200 or more eggs (Ives, Reference Ives1991; Wall, Reference Wall1993), resulting in high levels of aggregation and severe competition (Holter, Reference Holter1979; Ives, Reference Ives1991). In contrast, dung-breeding Diptera usually oviposit in much smaller clutches than carrion-breeding species, because dung is a more predictable and abundant resource. The mean egg batch sizes for N. cornicina and N. viridescens were 29 and 37, respectively, which is similar to the egg batch sizes found in other dung-breeding flies, such as Scathophaga stercoraria L. (Hammer, Reference Hammer1941).
The number of pats in which the two species co-occurred was higher than would have been expected by chance in 2001 but not in 2004. However, from these data, it is difficult to draw any clear conclusions about whether there was any interaction between the two species in their distribution between pats. It may simply have been that in 2001 some pats were more suitable than others for oviposition, perhaps as a result of their location in the field. This issue requires further, more detailed study.
Given the numbers of eggs matured by adult female Neomyia, the small numbers of each species generally recovered from the pats is of considerable interest. In particular, for N. viridescens, numbers were significantly lower than those of N. cornicina, despite the larger clutch sizes of the former. The data suggest that high levels of mortality may occur in the pats and that this mortality may be relatively higher for N. viridescens than for N. cornicina. Why this should be the case remains to be determined. Mortality, resulting directly from predation or parasitism or indirectly from the presence and activity of other dung colonizers, has previously been implicated as major determinants of fly abundance in pats (Kirk, Reference Kirk1992).
Dung-colonizing Diptera generally reach their peak abundance in late summer in southwest English pasture habitats (Lee & Wall, Reference Lee and Wall2006b). In the present study in 2001, both N. cornicina and N. viridescens were only first detected in July and then showed a very short, pronounced seasonal peak in larval abundance of only 1–2 weeks centred around pats placed out on September 3rd. It is possible, however, given the highly aggregated distribution of eggs and larvae within pats discussed above, that the detection method used in the present study, with batches of ten artificial pats deployed at weekly intervals in 2001, may not have been sufficiently sensitive to detect very low populations earlier in the year and that considerably greater sample sizes might have been needed. In one of the only other similar studies that has been carried out previously, Hammer (Reference Hammer1941) suggested that Neomyia overwinter as adults, become active in Spring and then pass through four or five generations, increasing exponentially in abundance in the field over summer. The rate of oviposition and larval ectivity would be expected to be highly dependant on temperature, so the colonization and subsequent decomposition of a dropping will depend on the weather immediately following deposition. While low temperatures appear to lengthen the time during which dung remains attractive to some invertebrates (Vessby, Reference Vessby2001), this period still only lasts for a few days and in bad weather a dropping may remain completely uncolonized by insects (White, Reference White1960). The suitability of a dropping as a resource, therefore, relies not only on its physical properties but also on its availability to colonizers at a certain time. However, here no relationship between abundance and any of the weather parameters in the week following deposition was observed. The reasons for this are unknown and few data exist on the relationships between weather conditions and oviposition or larval development in these species. Differences between the dairy and organic beef farms in Neomyia abundance were minor and, where differences were present, abundance was higher on the dairy than the beef farm.
Precisely how the ecological differences between Neomyia cornicina and N. viridescens translate into functional specialisation or how niche separation is maintained is unknown. Clearly, further work is required to identify the nature and cause of the mortality experienced by N. viridescens and N. cornicina.
Acknowledgements
We are grateful to the University of Bristol for its financial support of C.L. We would like to thank Andrew Forbes for his personal interest and helpful comments on this manuscript and Merial Animal Health for financial support of EA. We are grateful to John Keedwell and Mike Amos for permission to collect dung and use their farms as field sites.
