Introduction

Carrion is a highly attractive resource to a diverse array of arthropods (Benbow et al. 2013; Braack 1987; Moura et al. 2005; Parmenter and MacMahon 2009; Schoenly and Reid 1983; Voss et al. 2009). Compared with most other forms of detritus, carrion generates a very intense but brief hotspot of biological activity, including the rapid arrival and exploitation by arthropods. This activity results in the recycling of energy and nutrients through different organisms and trophic pathways, making it a critical part of ecosystem functioning (Barton et al. 2013a; Carter et al. 2007; Doube 1987).

Different arthropod groups have evolved strategies to exploit ephemeral carrion resources (Barton et al. 2013b; Braack 1987; Finn 2001). This includes specialisation on particular stages of decomposition, which partitions the resource among different species (Braack 1987; Schoener 1974). However, competition for such resources is often intense and so rapid dispersal to new resources, fast development times or population growth rates are further strategies used by some arthropod taxa (Barton et al. 2013b; Kouki and Hanski 1995; O’Flynn 1983; Smith and Merrick 2001). Comparing the diversity patterns of taxa with different resource exploitation strategies therefore provides a way to reveal new insight into the potential mechanisms underpinning the spatial and temporal patterns of arthropod biodiversity at carrion.

Two abundant groups of arthropod at carrion are beetles and mites (in addition to flies) (Bornemissza 1957; Braack 1987; Perotti and Braig 2009). Mites are a hyper-abundant component of the arthropod fauna at carcasses, where they may predate on fly larvae or nematodes, and scavenge on animal remains (Perotti and Braig 2009). Many mite species found at carrion are phoretic and use other species, typically flies and beetles, as a method for dispersal between carcasses (Perotti and Braig 2009). In contrast, beetle assemblages found at carrion are less abundant, but can be very species rich, and comprise a mix of resource specialists, generalist predators and opportunist scavengers (Barton et al. 2013b; Braack 1987; Sikes 1994).

In this study, we examined the changes in diversity and composition of mite and beetle assemblages at carrion over time. Detailed studies of arthropods at carrion often document the visitation patterns of species during early stages of decomposition (days to weeks), usually for potential forensic applications (Tomberlin et al. 2011). Fewer studies have examined the diversity dynamics of arthropods at carrion over longer ecological time periods to better understand the significance of carrion as a driver of diversity and heterogeneity in ecosystems (Barton et al. 2013a, b). Further, ecological studies of mite assemblages at carrion are rare (but see Gibbs and Stanton 2001), and the temporal dynamics of mite assemblages at carrion is poorly understood. Here, we compare and contrast the abundance, species richness and composition of mites and beetles collected at kangaroo carcasses over a 6-month period in a grassy eucalypt woodland in south-eastern Australia. Arthropod community dynamics at carrion typically follows a rapid invasion stage, a peak in abundance and richness, and a monotonic decline thereafter (Schoenly and Reid 1987). We therefore expected beetle diversity to follow such a temporal trajectory. By comparison, in was unclear how mite diversity might change after initial colonisation and relative to beetles given that some of the mites were likely to have phoretic associations with the beetles. Previous research has also demonstrated a distinct temporal change in beetle species composition in our study site (Barton et al. 2013b). But compositional changes in mite assemblages at carrion have not previously been documented in temperate eucalypt woodlands. Our study is the first to examine the contrasting dynamics of beetles and mites at vertebrate carrion in southeastern Australia, and contributes new insights into the role of carrion as a driver of diversity and heterogeneity in ecosystems.

Methods

Study area and experimental design

We conducted this experiment at the Mulligans Flat Sanctuary, part of the Mulligans Flat—Goorooyarroo Woodland Experiment (Manning et al. 2011). This reserve is a grassy eucalypt woodland, characterised by two dominant eucalypt species (yellow box Eucalyptus melliodora, and Blakely’s red gum E. blakelyi), interspersed with mixed native and exotic grassland (McIntyre et al. 2010). The dominant native herbivore in much of south-eastern Australia is the eastern grey kangaroo (Macropus giganteus), and is an important source of carrion in ecosystems in much of south-eastern Australia.

In the week beginning 23 November 2010, 18 whole M. giganteus carcasses were sourced from recent road kill in the region, and placed unsecured at the study site within 12 h of death. Carcasses were placed on patches of grass away from any shading effects of nearby large trees. Carcasses were only used where little external damage was observed, thus minimising loss of fluids prior to site placement. The average mass of a carcass was 31 kg (range 17–80 kg), and all were adults without pouch young. No animals were killed directly for the purposes of this study. Further details of site characteristics and temperature during this study are described elsewhere (Barton et al. 2013b; Macdonald et al. 2013).

Arthropod sampling

We used a pair of pitfall traps to sample ground-active arthropods at each carcass. Pitfall traps were placed on opposite sides of and immediately adjacent to each carcass. This allowed them to capture arthropods that actively moved around each carcass. Traps were half filled with propylene glycol as a preservative, and left open for 7 days during week 1, 6, 12 and 26 of the experiment. Each of these four sampling periods was associated with a particular stage of carcass decay. We describe week 1 as the ‘bloat to active decay’ stage, week 6 was during the final stages of ‘advanced decay’, week 12 during ‘dry decay’, and week 26 at a more ‘advanced dry decay’ stage with only few recalcitrant remains left (Carter et al. 2007). All arthropods were extracted from the pitfall traps, stored in 70 % ethanol and sorted to order and family.

Mites attached to beetles were removed from their hosts, and unattached mites that were free in the pitfall trap were also collected. Importantly, our beetle and mite co-occurrence data was not suitable for determining new phoretic associations. We also excluded all mites from the Sarcoptiform infraorder Astigmata, as these mites are typically very small and not able to be sampled effectively with pitfall traps. Thus, we restricted our mite samples to species from the orders Trombidiformes (suborder Prostigmata) and Mesostigmata, both free living and phoretic. We identified all specimens to family and species (where possible) using relevant keys (Halliday 1998, 2000; Lawrence and Britton 1994; Matthews 1980, 1982, 1984, 1987, 1992). Species-level identification was not possible for several beetle specimens, and these were designated a morphospecies code (sensu Oliver and Beattie 1996). Beetle identifications were verified by specialists at the Australian National Insect Collection in Canberra. Mite taxonomy follows that described by the Australian Faunal Directory (http://www.environment.gov.au/biodiversity/abrs/online-resources/fauna/afd/, accessed 05 November 2013). To determine which mite species were phoretic, we used a combination of expert advice from Dr Bruce Halliday of the Australian National Insect Collection, and phoretic associations described in the literature (see Table 1).

Table 1 Summary of the abundant phoretic mites (Acari: Mesostigmata) collected in this study and their known associations with beetles (Coleoptera) described in the literature
Full size table

Statistical analysis

We assessed the thoroughness of our sampling with species accumulation curves (Colwell 2013). For this analysis, we pooled all samples across each of the four sampling times, and used the analytical Mau Tau method to calculate a smoothed curve from 1,000 resamples of the data. We next used analysis of variance (ANOVA) to tested for differences in log10(x + 1) transformed abundance and species richness of mites and beetles between each sampling time. Data were log transformed to improve homogeneity of variances among our data and to meet assumptions of normality. We plotted means of the raw data to visually assess and compare the change in mite and beetle abundance and species richness across the four sampling times.

We used two multivariate methods to examine mite and beetle assemblage differences among decay stages. First, we used a multi-response permutation procedure (MRPP) in PC-Ord 5 software (McCune and Mefford 2006) to test for compositional differences in mite and beetle assemblages between sampling times. This test compares the average within-group distance with the overall average distance between samples (Zimmerman et al. 1985). Greater within-group agreement among samples of a pre-defined group (i.e. sample time) indicate they share greater similarity, and are therefore distinct from the overall set of samples (McCune and Grace 2002; Zimmerman et al. 1985).

Second, we used canonical analysis of principal coordinates (CAP, Anderson and Willis 2003) to examine compositional variation in mite and beetle assemblages. This method is a useful way to analyse patterns in compositional data against a priori hypotheses (Anderson and Willis 2003). In our case, we were interested in describing the variation among a priori groups of samples defined by sampling time (i.e. week 1, 6, 12 or 26). For both multivariate analyses we used the Bray–Curtis dissimilarity index on double square root transformed data to reduce the influence of hyper-abundant species. We plotted the correlations between each species and the two canonical axes in a biplot to identify which mite or beetle species were most strongly associated with each decay stage.

Results

Richness and abundance patterns

We collected 22 mite species from the 18 carcasses over the 6-month duration of the experiment. Eight species of mite (15 individuals) were from the suborder Prostigmata and one species (4 individuals) from the suborder Oribatida. The remaining 13 species of mite were from the order Mesostigmata and were by far the most abundant (8,196 individuals) (“Appendix 1”). We found that the majority of mesostigmatid mites were phoretic, and the most abundant species were Uroseius sp. (Uropodidae), Macrocheles muscaedomesticae, M. peniculatus and M. glaber (Macrochelidae). Two species of Parasitidae were also collected, Parasitus fimetorum and P. coleoptratorum. We collected 88 beetle species from 28 different families (3,614 individuals). The most abundant beetles are listed in “Appendix 2”. Species accumulation curves showed a high level of sampling thoroughness of mite assemblages, but less thorough sampling of the more diverse beetle assemblages, relative to sampling effort (Fig. 1).

Fig. 1
figure 1

Species accumulation curves showing the observed number of species of mites and beetles relative to sampling effort (n = 72)

Full size image

We found a highly significant different in abundance and species richness between the four sampling times for both mites (abundance F 3 = 32.57, P < 0.001; richness F 3 = 27.58, P < 0.001) and beetles (abundance F 3 = 28.43, P < 0.001; richness F 3 = 15.60, P < 0.001). Plots of raw means showed that mites and beetles peaked in abundance and richness at different sampling times. Mites were initially low in abundance and richness at carcasses during week 1 (Fig. 2a), whereas beetles were initially very high in abundance and richness (Fig. 2b). This pattern was reversed by week 12 when mites reached their highest abundance and beetles were lower in abundance and richness.

Fig. 2
figure 2

Changes in mean (±SE) abundance and species richness of a mites and b beetles collected from vertebrate carcasses over a 6-month period. Samples were taken at 1, 6, 12 and 26 weeks after initial placement of carcasses

Full size image

Compositional patterns

The MRPP analysis showed an overall difference among mite assemblages among the four sampling times (T = −16.81, A = 0.23, P < 0.001). Pairwise comparisons indicated a significant difference in mite species composition between adjacent sampling times (P < 0.001), but not between weeks 1 and 26 (P = 0.056). For beetles, MRPP indicated a significant overall difference in assemblage composition among all sampling periods (T = −29.39, A = 0.19, P < 0.001). In contrast to mites, however, pairwise comparisons for beetles revealed that assemblages from each sampling time were always significantly different from every other sampling time (P < 0.001).

Constrained ordinations showed distinct assemblages of both mites (Fig. 3a) and beetles (Fig. 3c) from each sampling time, clearly shown a successional turnover in species composition as decay progressed. For mites, the samples from week 6 and 12 appeared to be the most distinct, whereas there was some overlap among samples from week 1 and 26. The biplot of mite species (Fig. 3b) indicated that M. muscaedomesticae was most strongly associated with week 1, M. glaber and M. peniculatus were associated with week 6, and Uroseius sp. was most strongly associated with week 12. For beetles, there was a tighter grouping of samples from each sample period. The biplot of beetle species (Fig. 3d) indicated that Creophilus erythrocephalus, Ptomaphyla lacrymosa, Aphodius sp. 1 and Saprinus sp. 1 were strongly correlated with week 1, a mix of species contributed to the distinctness of assemblages occurring at week 6, and Aphodius sp. 2 and Ormorgus sp. 1 were correlated with week 12.

Fig. 3
figure 3

Ordinations (left) and biplots (right) of mite a, b and beetle c, d assemblages at carcasses at four sample times. Only mite species with abundance >10, and beetle species with abundance >20 are shown in the biplots

Full size image

Mite–beetle associations

We found direct evidence of an association between Macrocheles sp. and the beetle Saprinus sp. (Fig. 4a), as well as the abundant Uroseius sp. and the beetle Omorgus sp. (Fig. 4b). We summarise in Table 1 the known associations between the phoretic mites we collected and their beetle hosts currently described in the literature. Due to the constraints of our sampling method, we cannot provide definitive evidence of mite and beetle associations. However, the temporal co-occurrence of abundant mites and beetles shown in our biplots (Fig. 3), suggest that the scarabaeines Aphodius and Onthophagus, and the trogid Omorgus, are likely to key hosts for several mite species.

Fig. 4
figure 4

Two examples of a phoretic association between mites and beetles. a A species of Macrocheles (Macrochelidae) attached to the hind legs of the beetle Saprinus sp. (Histeridae), and b a species of Uroseius (Uropodidae) attached to the ventral tergites of the beetle Omorgus sp. (Trogidae). Also note that astigmatid mites (not considered in this study) are present around the prosternum of both beetles, and the terminal abdominal segment of the histerid

Full size image

Discussion

The arrival and departure times of arthropod species at carrion is well documented from many parts of the world (Amendt et al. 2004; Archer 2003; Matuszewski et al. 2010; Matuszewski et al. 2011; Voss et al. 2011), with much of this research aimed at potential forensic applications (Tomberlin et al. 2011). In this study, we took an ecological perspective and examined the temporal dynamics of arthropod diversity at carrion to reveal new insights into the biodiversity associated with the decomposition process. We found that mite assemblages had a delayed peak in richness and abundance relative to beetle assemblages, but that both mites and beetles had clear successional dynamics. This demonstrates for the first time the contrasting diversity patterns of mites and beetles at carrion in temperate eucalypt woodland in southeastern Australia.

The mite assemblages peaked in abundance and diversity at different times to the beetle assemblages. Whereas beetles peaked in richness and abundance during active decay (week 1) and advanced decay (week 6), mites were comparatively much lower in abundance and richness during active decay (week 1), and instead peaked during advanced and dry decay (weeks 6 and 12). Arthropod communities are typically understood to involve a rapid invasion stage, a peak in abundance and richness, and a monotonic decline thereafter (Schoenly and Reid 1987). Our study shows that this general expectation may not occur at the same time for beetles and mites. The reasons for the observed differences in diversity dynamics may be related to their different dispersal and reproduction strategies. The arrival of phoretic mites is dependent on the arrival of their insect hosts, and can only arrive after the primary insect colonisers have arrived. Subsequent to this arrival, the mites proliferate and complete many lifecycles leading to an increase in population abundance and density. In contrast, many individuals of a particular beetle species may colonise a carcass, but then only complete one life cycle. This may lead to a relative lag in peak abundance of each phoretic mite species, even if initial colonisation is synchronous with their insect host. This can be seen for M. muscaedomesticae, which is likely to have arrived with the first calliphorid flies and the beetle P. lacrymosa (see Halliday 1990), but which didn’t reach peak abundance until advanced decay in week 6 of our experiment (“Appendix 1”). Another example is the uropodid mite Uroseius sp. which arrived during week 1 of the experiment, but didn’t reach peak abundance until dry decay in week 12, 6 weeks after beetle host Omorgus sp. (see Philips 2009) peaked in abundance (see “Appendices 1, 2”).

Both mites and beetles displayed successional patterns, reflected in their changes in composition as decomposition progressed over time. Our ordination diagrams showed that key mite and beetle species were driving the compositional distinctness of assemblages at each sample time. The successional dynamics of beetles at carrion is reasonably well documented in Australia (Archer 2003; Barton et al. 2013b; Bornemissza 1957; Voss et al. 2009), but is far less clear for mites. Our study shows that mite assemblages also display successional patterns, despite their more generalised predaceous and scavenging diet, but these patterns were less distinct than beetle assemblages. This could be due to their initial dependence on insect hosts, thus driving distinct arrival times, but delayed and prolonged abundance increases, thus blurring distinct departure times.

Eleven of the 13 mesostigmatid mites we sampled have known phoretic associations with beetles (see Table 1). It is therefore likely that many of the mites in our study had phoretic associations with the beetles from our study area. However, there is likely to be overlap between insect hosts for some mites, perhaps using both flies as well as beetles to move between carcasses. For example, M. muscaedomesticae is a common mite found on flies (Perotti and Braig 2009), but has also been recorded from the beetle P. lacrymosa (Halliday 1990; Philips 2009). Host-specificity between phoretic mites and beetles at carrion is poorly studied in Australia (pers. comm. Bruce Halliday), and there are many gaps in our knowledge of carrion-associated mite and beetle ecology. Further, we cannot make explicit links between mites and beetles in this study as our sampling method allowed only co-occurrence data. Our list of beetles co-occurring with phoretic mite species provides a starting point for further research on host associations and specificity.

Our study showed that mite assemblages at carrion were dominated by phoretic mesostigmatids, although this excludes the numerically dominant Astigmata (O’Connor 2009; Perotti and Braig 2009). We also showed they have a delayed peak in abundance and richness compared with beetle assemblages, which may be due to a lag in their population increase after arrival with their various host flies and beetles. Dispersal method, resource specialisation and life-history strategies are all mechanisms that help explain the contrasting diversity dynamics of different arthropod taxa at carrion (Barton et al. 2013b; Braack 1987). This study provides additional evidence of the important role of these traits, and further highlights the role of carrion as a driver of diversity and heterogeneity in ecosystems.