Two dung beetle species that disperse mimetic seeds both feed on eland dung

HOW TO CITE: Midgley JJ, White JDM. Two dung beetle species that disperse mimetic seeds both feed on eland dung. S Afr J Sci. 2016;112(7/8), Art. #2016-0114, 3 pages. http://dx.doi.org/10.17159/ sajs.2016/20160114 Scarabaeus spretus zur Strassen was observed to roll and bury Ceratocaryum argenteum (Restionaceae) seeds in the sandplain fynbos of the Potberg area of the De Hoop Nature Reserve, South Africa. This species is the second dung beetle species found to be deceived by the faecal mimicry of C. argenteum seeds – the first species being Epirinus flagellatus. An isotopic analysis suggests that both these dung beetle species most likely feed on eland (Taurotragus oryx), not bontebok (Damaliscus pygargus pygargus), dung. Thus the model in this mimicry is eland dung; this interaction suggests large herbivores are an integral part of this fynbos.


Introduction
The dung beetle Epirinus flagellatus was observed to roll and bury seeds of the Cape plant Ceratocaryum argenteum (Restionaceae) at a site in the Potberg part of the De Hoop Nature Reserve in South Africa. 1 This primary dispersal of seeds involves chemical and visual mimicry because neither the dung beetle nor its larvae can eat these hard seeds.Chemically the seeds have characteristics of the dung of both of the most common large herbivores in the reserve: the eland (Taurotragus oryx) and the bontebok (Damaliscus pygargus pygargus). 1However, the seeds are more similar in shape and size to the smaller faeces of the bontebok, which is then the possible visual model that C. argenteum mimics.At the same Potberg site, during February 2016, we observed similar seed dispersal of C. argenteum seeds by another dung beetle, Scarabaeus spretus zur Strassen.The aim of this paper is to document this new burial behaviour and to investigate both dung beetle species to determine whether the faeces of the bontebok or the eland is the likely model of the mimic.
The bontebok is a short grass grazer whereas the eland is a mixed feeder. 2,3The two dominant grassland/ renosterveld grass species at Potberg are Cymbopogon popschilli (Andropogoneae) and Cynodon dactylon (Chloridoideae), 3 to which can be added the relatively widespread Themeda triandra (Andropogoneae).All three species utilise the C4 photosynthetic pathway rather than the C3 pathway. 4This pathway is common in tropical grasses whereas the C3 system is more common in woody plants and temperate grasses.The enzymes of these two different photosynthetic pathways produce different carbon δ 13 C signatures in their photosynthetic products.The relatively rare stable isotope of carbon, 13 C, is slightly heavier than the more common 12 C, which affects the ratios of these isotopes in different plants depending, for example, on enzyme preferences for the lighter isotopes.Fractionation is the process which reflects changes in relative proportions of isotopes, such as 13 C: 12 C during C3 photosynthesis.Fractionation can also occur in 15 N because, as it is heavier than 14 N, it may increase in tissues depending on factors such as levels of metabolism, catabolism and excretion.Thus animals are typically enriched by +3-5‰ in δ 15 N compared to their diet, although typically they are less than +1‰ enriched in δ 13 C. 5 During metamorphosis, larval tissue is broken down and then used to form new adult tissue and thus metamorphosis is also known to increase both δ 13 C and δ 15 N in the adult tissue in much the same way as would happen in adult tissue with an increase in trophic level. 6Thus the hypothesis that bontebok dung is the likely model C. argenteum mimics can be tested using the isotopic method for diagnosing animal diets, including those of dung beetles. 7

Methods
The study took place in the Potberg area of the De Hoop Nature Reserve (34.374420 S, 20.533060 E) in the sand plain vegetation type in which C. argenteum grows.During 3 days in early February 2016, we placed out 5 to 10 piles of C. argenteum seeds, with each pile comprising 10-20 seeds.Piles of seeds were 10 m apart on the edge of a 100-m stretch of a sand road through natural vegetation.We monitored the seed piles in the early morning for approximately 2 h (starting at about 08:00).This experiment took place after a 24-h rain event.
Dung samples were taken in various vegetation types in the Potberg reserve.These types were renosterveld (dominated by the shrub Elytopappus rhinocerotis (Asteraceae)), grassland (dominated by Cynodon dactylon), salt marshes (dominated by Chenopodiaceae), valley bottom fynbos (dominated by the Proteaceae shrubs Leucadendron linifolium/ L. coniferum), sand plain fynbos (dominated by Leucadendron laureolum, where C. argenteum occurs) and limestone fynbos (dominated by Leucadendron meridianum).Eland dung was found at all six sites whereas bontebok dung was found at all but the last two fynbos sites (sand plain fynbos and limestone fynbos).Previously, Radloff et al. 3 noted that bontebok avoid fynbos whereas eland are found throughout fynbos, including limestone fynbos.To reduce chances of pseudoreplication, we sampled only a single pellet of bontebok or eland dung from a dung pile; only dung piles greater than 5 m apart were sampled and, as judged by colour, only relatively fresh samples were collected until a total of 10 pellets had been sampled from within each vegetation type.Dung pellets and dung beetle exoskeletons were dried and analysed for δ 13 C and δ 15 N, in ‰, using standard techniques at the Archaeometry Lab at the University of Cape Town.Dung beetle larvae have chewing teeth and are vigorous detritivores that depend on the plant remains that constitute the dung ball, rather than being dependent on microbiota associated with the ball. 8Although female beetles do select small fragments of plant remains from dung to constitute brood balls, 8 this is not likely to significantly affect the isotopic signature of these balls nor the signature of the exoskeletons of adults that emerge from these balls.

Results
Dung beetles arrived at seed stations within a few minutes of placing seeds out; thus within 2 h each day, more than 10 beetles had arrived at seed piles along our short 100-m transect and had started burying seeds (Table 1 and Figure 1a-c).1d), indicating that they too are deceived by the scent of the seeds.These 'flesh flies' are typically attracted to dung, carrion or rotting vegetation. 9Lesser dung flies (Sphaeroceridae) were observed on S. spretus (Figure 1b).

Discussion
The increase in δ 15 N as a result of metamorphosis is in the range of +3‰ and +5‰ for a selection of insects ranging from Diptera to Coleoptera to Lepidoptera 6 and the increase in δ 13 C is about +1‰ 5 .Dung beetle adults whose larvae fed on bontebok dung should thus have values of slightly more than -20‰ δ 13 C but up to 7.6‰ δ 15 N.The mean δ 13 C values of both E. flagellatus and S. spretus clearly indicate a C3 dung diet and are thus much closer to that of the eland dung (Table 2 and Figure 2).The δ 15 N values are 3.25‰ and 5.7‰ above eland dung but only 1.9‰ and 4.32‰ above bontebok dung.The evidence from δ 15 N of the beetles is less equivocal about the larval food source because of the small 1‰ difference in dung between the two herbivore species.
Overall the isotope results are compatible with eland being the main larval source of dung.Also bontebok dung is rare in the lowland fynbos habitat of C. argenteum.As these two dung beetle species are from different genera and are both attracted to C. argenteum seeds, these results indicate that the seed chemistry and deception by C. argenteum is not dung beetle species-specific.The deception appears most likely to be modelled on the chemistry of eland, rather than bontebok, dung.
Not much is known of the feeding biology of Scarabaeus or Epirinus beetles. 10For both species, we observed diurnal activity and no pair formation at seed burial sites.C. argenteum seeds are about the same size as bontebok droppings and about half the size of eland droppings.That many seeds were buried per site for S. spretus suggests that several pellets of eland dung are typically used for feeding or egg laying, whereas E. flagellatus only buried a single seed per burial event.This distinctive burial behaviour of the two species likely results in differential recruitment patterns for C. argenteum seedlings.Single E. flagellatus burials would lead to lower intraspecific competition between seedlings compared with the multiple burials by S. spretus.
Many other dung beetles occur at Potberg, for example, the millipedeeating Sceliages adamastor. 10The δ 15 N dung beetle values presented in Table 1 provide a framework to interpret those of S. adamastor to determine whether this beetle is an obligate insectivore.For example, δ 15 N values of an obligate millipede-eater should be a trophic level above herbivorous dung beetles such as S. spretus and E. flagellatus (i.e. they should have δ 15 N values greater than 7‰).The very large Addo flightless dung beetle (Circellium bacchus) also occurs at Potberg 10 ; elsewhere it feeds on elephant dung 10 but as there are no elephants at Potberg, its diet there is unknown and could too be clarified using the isotopic method.
Being flightless, C. bacchus is often killed on roads and our analysis of nine roadkill individuals (mean δ 13 C of -27.70‰ and δ 15 N of 2.39‰) indicates eland dung is also its major larval food source.
The fact that Ceratocaryum argenteum is an element of deep sand fynbos, 11 implies that sufficient quantities of large herbivore dung, such as that of eland, occurred in this vegetation.This would maintain the associated dung beetle species and the deceptive relationship between C. argenteum and these species.There is some debate as to whether large herbivores were once more common in fynbos and in this area of the Cape. 3 Our observation that C. bacchus, E. flagellatus and S. spretus utilise eland dung suggests that the eland is, and has been, a key species in this system and should be carefully managed as such.Finally, we suggest that there are now sufficient examples of seed dispersal by beetles for use of the term coleopterochory.This term would include primary dispersal such as that described above, as well as examples of beetle endozoochory 12 in which small seeds are swallowed, as well as incidental or secondary dispersal in dung or with fruit 1 .
E. flagellatus crawled out of the vegetation towards seed piles, with only an occasional individual flying in, whereas all S. spretus individuals flew towards the seed piles.It was clear, based on the direct flight or crawling paths of both species to the seeds, that the attraction is strongly chemical.A S. spretus beetle even flew into a paper bag containing seeds.All cases of S. spretus burial involved limited movement of seeds (<0.25 m) from seed piles, whereas E. flagellatus moved seeds up to 2 m. S. spretus beetles were observed to frantically bury up to three seeds (n=2) and often five or more seeds (n=4) per excavated hole (see the video in the supplementary material online).E. flagellatus was observed to only bury seeds individually, similarly to observations by Midgley et al. 1 Flies of the Sarcophagidae were frequently observed to settle on C. argenteum seeds (Figure

Figure 2 :Figure 1 :
Figure 2:The distribution of isotopes (δ 13 C and δ 15 N) of the dung of bontebok and eland as well as of three dung beetle species.

Table 1 :
Observations of dung beetles at Ceratocaryum argenteum seed stations at Potberg

Table 2 :
Isotope analyses of dung beetles and dung from Potberg