A simple technique to assess resource use in dung beetle breeding studies

Current methods for identifying resource preferences in dung beetles are based on above‐ground trials. Although useful, these methods do not directly investigate resource provisioning of offspring below ground, missing an important part of dung beetle ecology. We tested the potential for UV‐fluorescent and non‐fluorescent coloured glitters to be used as markers for tracing the origin of dung incorporated into brood balls (dung shaped by parental beetles enclosing an egg), and so aid in a more complete understanding of resource use in dung beetles. We tested the effect of glitter addition on brood ball production in two species of tunnelling dung beetles, Onthophagus taurus and Euoniticellus fulvus. There was no effect of glitter addition on brood ball production during no‐choice tests for each species: both species made a similar number of brood balls, regardless of glitter presence or glitter colour. In a separate choice trial, O. taurus showed no preference for dung when presented with four dung pats containing four unique glitter colours. Here we show that glitter can be used as an effective marker of individual brood ball origin in tunnelling dung beetles. This method provides a useful tool for tracking below ground resource use and larval provisioning experiments in dung beetles.

Ecosystem services that result from this removal of dung include nutrient cycling, increased soil organic matter and control of pests and parasites, including flies and gastrointestinal nematodes (Menéndez et al., 2016;Veldhuis et al., 2018).This, and the fact that many dung beetles are considered environmental indicator species (Davis et al., 2004), has led to much research into this group globally (Bryan, 1973;Bryan & Kerr, 1989;Feehan et al., 1985;Kirk & Wallace, 1990;Ridsdill-Smith & Hayles, 1990;Sands & Wall, 2016).One focus area of this research has been identifying preferences of dung beetles for different dung types.
The development of a simple, cheap and effective neutral marker for tracking dung resource use would be beneficial to further improve ecological knowledge of dung beetle species.Tracking resources or individuals using indigestible markers has previously been used to understand a wide range of ecological processes in animals, including movement rates, territory use and species interactions within spatial and social networks (Delahay et al., 2000;Jackson et al., 1999;Rösner & Selva, 2005;Ruiz-Aizpurua et al., 2013).There has been frequent work on secondary seed dispersal using plastic beads, but less work on broader aspects of dung beetle ecology (Andresen, 2002;Andresen & Levey, 2004;Enari & Sakamaki-Enari, 2014;Feer et al., 2013;Koike et al., 2012;Manns et al., 2020;Santos-Heredia et al., 2011).Moczek and Cochrane (2006) used dung containing vermiculite to investigate the rate of intraspecific brood ball parasitism in Onthophagus taurus.They found that the markers did not have any effect on brood ball production.Furthermore, fine-grain vermiculite has also been used to track gut retention time in O. taurus and Digitonthophagus gazella (Rohner & Moczek, 2021).Additional research investigating secondary seed dispersal of dung beetles has utilised small plastic beads (2-5 mm) to understand the effect that seed size and soil type might have on depth of seed burial (Koike et al., 2012).However, these approaches have not fully explored the opportunities to develop a generalised neutral marker for dung resource tracking.
Here, we propose the use of non-toxic glitter as a generalised neutral marker for tracing the origin of dung resources used in the below ground creation of brood balls in dung beetle reproductive ecology.Glitter is made of polyethylene terephthalate (PET) and covered with aluminium (Yurtsever, 2019).It is non-toxic, chemically inert and small enough that it is unlikely to be removed from the resource by dung beetle activities.Glitter also comes in a range of distinct colours, which provides the ability to mark multiple resource types at once, allowing for multiple types of dung to be tested simultaneously (e.g., dung from animals on different diets, dung that contains anthelmintic residues or dung from different seasons).The ability to track and identify the origin of brood balls could assist in achieving a more thorough understanding of dung beetle ecology.In a series of experiments, we test the effects of glitter presence on dung beetle preferences and behaviour when creating brood balls and determine whether glitter can be used successfully as a marker to identify the origin of dung used for brood ball construction.

MATERIALS AND METHODS
All dung and dung beetles were collected from a dairy farm in Mardella, Western Australia.Dung was collected from cattle that had never been treated with anthelmintics.Dung was homogenised upon collection and stored at À20 C until required for the experiment.
Approximately 400 O. taurus and 100 Euoniticellus fulvus were collected via hand searching through fresh dung pats.Beetles were held in plastic containers (25 Â 25 Â 10 cm) filled with a 3-4 cm layer of moist builders' sand and provided ad libitum with dung until they were used in experiments (around 2 weeks).Beetles were selected at random from these containers for each experiment.Each beetle was used only once.Comparisons were made among four glitter colours: Gold, Silver, UV-pink and UV-yellow.These colours were found to be visually discernible from one another in a pilot test where 0.2 g of glitter was mixed with 10 g of dung.A UV torch was used for discriminating the UV glitter.

No-choice glitter test
Two dung beetle species (O.taurus and E. fulvus) were used to test whether glitter mixed in dung would be incorporated into their brood balls and whether there was any effect of glitter on brood ball production.A total of five glitter treatments were used, including a glitter-free control and the four colours of glitter (Gold, Silver, UV Pink and UV Yellow).Each replicate received 100 ± 10 g of dung.Replicates that were assigned a glitter treatment had 2 g of the chosen colour of glitter added, and thoroughly homogenised by hand for 30 s, or until the glitter was evenly distributed throughout the dung.No glitter was added to the glitter-free control dung.Because the measurement error in dung pat weight was far greater than the 2 g of glitter added, no additional dung was added to compensate for the lack of glitter in the glitter-free control.The five glitter treatments were replicated 10 times each for O. taurus (i.e., 50 separate experimental units with one female O. taurus each) and 8 times each for E. fulvus (i.e., 40 separate experimental units with one female E. fulvus each).
Each of the 90 experimental units consisted of a breeding chamber made from Polyvinyl Chloride (PVC) pipe (height: 30 cm; diameter: 9 cm), partly filled with moist builders' sand (22 cm deep) and dung (100 ± 10 g) placed on top.A single female dung beetle was added to each tube, and a perforated plastic film was secured over the top of the tube to avoid escape or movement between chambers.Breeding chambers were kept in a controlled temperature room (26 ± 1 C; photoperiod LD 12:12 h) for 5 days.Following the 5-day experimental period, breeding chambers were emptied, and the sand was sifted through a metal mesh (1 Â 1 mm apertures) to retrieve brood balls.
Brood balls were checked for the presence of glitter, and the number of brood balls was recorded.

Glitter choice test
The effect of the four selected glitter colours on dung choice by O. taurus was tested using 20 mesocosms set up in a glasshouse (Figure S1; temperature range from ca. 18 C minimum and 28 C maximum; daylight hours between 06:00 and 19:00 h).Insufficient E. fulvus individuals were available to carry out a similar glitter choice experiment on both species.Each of the 20 O. taurus mesocosms consisted of a Décor 60 L polyethylene bin (height: 51 cm, diameter: 40 cm), filled three quarters full (38 cm deep) with moist builders' sand and covered with a modified Décor 60 L dome lid (Décor, Dandenong South, Victoria).The centre of each lid was removed, and the external rim of the lid was used to secure a fine mesh over the bins.Four dung pats were formed from 240 ± 10 g of dung using 10 cm diameter moulds and then placed equidistantly from one another on the surface of the sand within each bin.
Each of the four dung pats within a mesocosm had one glitter colour (either Gold, Silver, UV Pink or UV Yellow) homogenised through it.This allowed the four dung pats to be distinguished from one another and linked to the origin of brood balls (Figure S1).The glitter colour assigned to each dung pat position was randomly allocated prior to mesocosm set up.A specific dung pat mould was assigned to a specific allocated glitter colour to reduce glitter cross-contamination.
Six male-female pairs of O. taurus were added to each of the 20 mesocosms (total 12 beetles per mesocosm; 240 beetles total) and allowed 7 days for brood ball production.After this time, any remaining dung on the surface of the sand was removed, and the sand was carefully sifted through a 1 Â 1 cm metal mesh on a mechanical sifter to retrieve brood balls.The number of brood balls within each mesocosm was recorded along with their associated glitter colours.

Statistical analysis
Data collected in both experiments were analysed using R v. 4.2.3 (R Development Core Team, 2020), in the RStudio v.4.1103 interface (RStudio Team, 2020).All R code used in the statistical analyses is provided in the supplementary material.No-choice glitter models were fitted using the 'lme4' package (Bates et al., 2015) while the glitter choice model was fitted using the 'glmmTMB' package (Brooks et al., 2017).

No-choice glitter test
Generalised linear models (GLMs) with a Poisson distribution were used to test the effect of glitter addition on the number of brood balls produced by O. taurus and E. fulvus.The response variable was the count of the number of brood balls produced, and the explanatory variable was the fixed categorical effect of 'glitter treatment' (with five levels).Initial tests using the 'simulateResiduals()' function from the 'DHARMa' package (Hartig, 2021) found significant overdispersion of model residuals in the Poisson GLMs.Therefore, GLMs with a negative binomial distribution were used to account for overdispersion in the data.In both the analysis for O. taurus and for E. fulvus, it was predicted that there would be no difference between the mean number of brood balls produced per female in control versus glitter colour treatments.

Glitter choice test
A multinomial generalised linear mixed model (GLMM) was applied to test the relative frequency of brood balls produced from each of four dung pats of differing glitter colours within the same mesocosm.This model takes the log-linear Poisson approach to multinomial analysis, in which the brood ball counts per glitter treatment group are used as the response variable and glitter colour is used as a categorical predictor in a Poisson GLMM.Glasshouse block, mesocosm and pat position were added as random effects in the GLMM (Figure S1).

No-choice glitter test
A total of 312 brood balls were produced by O. taurus, and 343 brood balls were produced by E. fulvus in the no-choice glitter trial.There was no significant difference in the average number of brood balls produced per female in the control treatment versus any of the glitter colour treatments, for either O. taurus (χ 2 = 0.737, DF = 4, p = 0.947 n. s.; Figure 1a and Table 1a) or E. fulvus (χ 2 = 1.757,DF = 4, p = 0.780 n. s.; Figure 1b and Table 1b).

Glitter choice test
A total of 1713 brood balls were retrieved from the 20 mesocosms in the glitter choice test.In all cases, there was only one glitter colour detected per brood ball, with no cases detected with mixed glitter colours in the same individual brood ball.In the multinomial GLMM analysis testing for variation in the relative frequencies of brood balls per dung pat at the mesocosm level, we did not detect an effect of glitter colour on O. taurus choice of dung pat for brood ball production (χ 2 = 4.770, DF = 3, p = 0.190 n. s., Table 2), after accounting for the baseline log-odds of 0.187 that beetles chose not to use any given dung pat at random (i.e., the inverse-logit of the zero-inflation estimate in Table 2).Regardless of glitter colour,  pat (± 95% confidence limits) overlapped the null expected frequency of 25% (Figure 2).

DISCUSSION
A critical limitation on studies of resource use in dung beetle ecology is the ability to trace the origin of dung used to construct brood balls.
To address this limitation, we explored the use of glitter as a marker in dung.The addition of glitter to dung did not impact the production of brood balls by O. taurus or E. fulvus during a no-choice experiment.
Glitter was still visible and identifiable post-brood ball manipulation.
Furthermore, when O. taurus pairs were simultaneously provided with four dung pats containing unique colours of glitter, brood ball production was similar for all glitter colours.Thus, our results demonstrate that glitter can be used as a marker, without bias to a particular colour.
The use of markers in dung is widespread and commonly used in domestic, captive and wild species to understand issues relating to health of individuals (Cabana et al., 2017;Fuller et al., 2011;Hogan et al., 2011), food sources (Újváry et al., 2014), territories (Delahay et al., 2000;Ruiz-Aizpurua et al., 2013) and population monitoring (Forin-Wiart et al., 2014).Faecal markers have also been used to study reproductive biology in mammals (Flacke et al., 2017).In these cases, glitter is often fed to a target individual or group with the glitter passing through the gastrointestinal system, and following elimination is then visible in the dung.This is the point at which glitter use (and the use of other markers) in relation to dung tracking has stopped, with few if any studies continuing to follow the movement of marked dung.
Yet there is a wealth of value in dung tracking further down the trophic system.Indeed, a whole-of-life approach may be possible, through mapping of ecological connections between plants, vertebrates and dung beetles using glitter as a marker to identify networks similar to networks between plants and pollinators (Dupont et al., 2014).

Previous incorporation of foreign objects into dung has not been
shown to have major effects on brood ball production or dung beetle functioning (Moczek & Cochrane, 2006;Santos-Heredia et al., 2011).
Industrial dyes and vermiculite have been used to understand brood ball parasitism (Moczek & Cochrane, 2006) and digestion in larvae (Rohner & Moczek, 2021), while the incorporation of small beads into dung has been used to gain insight into secondary seed dispersal (Enari & Sakamaki-Enari, 2014;Feer et al., 2013;Santos-Heredia et al., 2011).Glitter particle sizes can range from 50 to 6250 μm (0.05-6.25 mm), though most glitters are around 100-200 μm (Blackledge & Jones, 2007).Adult dung beetles often encounter nondigestible particulate matter (between 0.2 and 10 mm) in dung, which can be larger than common glitter sizes (Holter, 2016).Furthermore, adult beetles feed on even smaller particulate matter with around a maximum size of 60 μm (Holter et al., 2002).Impacts on brood ball production from the consumption of glitter by adult dung beetles or simply  the presence of glitter are therefore likely to be negligible.At a basic level, then, we readily expected dung beetles to incorporate glitter into their brood balls with no observed effect on the quantity of brood balls produced.However, we only tested for the effects of glitter on brood ball production for a maximum period of 5 days, and longer term tests would be important to carry out in the future.
Our present study focussed on the feasibility of using glitter as a method of tracking resource utilisation and movement in dung beetles.We did not consider other potential effects of glitter addition on brood ball size or larval development.These are of course important aspects of dung beetle ecology and would likely need to be considered if exploring questions beyond tracking resource use.
The glitter choice test found that the four selected glitter colours used did not affect dung preference for brood ball production based on relative proportions of brood balls (Figure 2).The study of vision in dung beetles has shown that some dung-rolling species are able to interpret an array of celestial cues to assist with orientation and direction (Dacke et al., 2003(Dacke et al., , 2014)).It is believed that they use a primitive, dichromatic colour vision (Yilmaz et al., 2022) to interpret celestial chromatic contrasts that range from an intense green light in the direction of the sun to intense UV light in the opposite direction (Coemans et al., 1994) to consistently maintain their rolling direction (El Jundi et al., 2015).Colour vision research, however, appears to be lacking in other functional groups of dung beetles such as tunnellers, which spend much time in subterranean tunnels where light is limited.While this study did not aim to understand colour recognition in dung beetle vision, it may be the case that either the dung beetles could not distinguish between different glitter colours when mixed into dung, or the colours simply did not influence their choice.Of course, given the wide diversity of available glitter colours, it is possible some other colours not tested here may influence dung beetles while producing brood balls.Further investigation into the colour perception of dung beetles could provide improved guidelines on suitable glitter colours.Furthermore, differences in colour perception are known to occur in humans and could potentially lead to identification errors during data collection.Prior testing of glitter colours to identify the most suitable colours that can be discriminated by all members of the research team is of importance.The use of reference samples and quality control checks between researchers could further reduce potential errors caused by inter-observer effects.
Some forms of glitter may be considered microplastic pollution, and there have been calls for more ethical use and legislation regarding glitter products, especially within the cosmetic industry.For example, there have been suggestions to reduce glitter use, through regulating or restricting the production and marketing of glitter or through providing suitable biodegradable alternatives (Anagnosti et al., 2021;Guerranti et al., 2019;Yurtsever, 2019).Some research into microplastic impacts on terrestrial invertebrates has been carried out in recent years.While the presence of glitter did not appear to impact adult dung beetles during brood ball construction, the potential implications of glitter ingestion during larval development are unknown.The impacts of microplastics on invertebrates have been shown to vary (Selonen et al., 2020;Silva et al., 2021), and examining the potential impacts of microplastics such as glitter on dung beetle larvae will be an important question for future research.
Importantly, the quantity of glitter used within dung is low in comparison to other uses of glitter (e.g., cosmetics and toys), particularly glitter introduced into domestic wastewater and thus concentrated into rivers and oceans (Geyer et al., 2017;Yurtsever, 2019) and as demonstrated here, could inform researchers of important ecological details that could be otherwise difficult to record.Biodegradable glitter alternatives, like those made from cellulose (Droguet et al., 2022), could be considered, though ecological impacts may not differ (Green et al., 2021).It would be beneficial to explore more natural-based alternative options especially as technology progresses.Further research into biodegradable alternatives will be important in the future use of this marking method.
This glitter marking method has one important caveat on its ecological application: the ability to locate and retrieve marked brood balls.This is simple in laboratory settings or bucket-style mesocosm experiments where substrates can be easily moved and sifted.
However, this is more difficult in the field, as dung beetles may relocate dung across a wider spatial area, both vertically and horizontally (particularly true for dung-rolling species).Yet there is a case for optimism, as researchers have been successful in the field with locating dung containing seeds and coloured plastic markers (Feer et al., 2013).Searching may be made easier when glitter-marked dung pats are geolocated, or when signs of dung beetle activity are clear.Further use of mesocosm and laboratory investigations may help to identify patterns to reduce search effort in the field.
In summary, glitter can be successfully used as a marker within dung allowing for successful tracing of brood ball origin.Potential uses in dung beetle studies include investigating daily brood ball production, identifying resource use across dung from distinct species, different seasons and different quantities or qualities of dung.Furthermore, this method could provide insight into brood ball spatial positioning relative to source origin or resource use difference based on individual age or size or under varying levels of competition when combined with marking of eggs from individuals (such as that seen in Moczek & Cochrane, 2006) Glasshouseblock was added to account for any variation in environmental conditions across the gradient of the glasshouse.Mesocosm accounted for the non-independence of the four dung pats within each mesocosm.Dung pat position accounted for the different positions of dung pats and glitter colours within mesocosms.Because of the log-link in the Poisson model, the relative differences among the model coefficients for glitter colour are exactly equivalent to the log-odds that relative frequency of brood ball production differs among glitter colours.If the relative frequency of brood balls produced from dung of each colour did not vary significantly (i.e., met the expected 0.25 proportion for each glitter colour), this would reflect no preference or bias for any of the four glitter colours.Additional random effects for the glasshouse block and dung pat position within mesocosm were added to the model.Glasshouse block was added to account for any possible gradients in environmental variation along the length of the glasshouse (i.e., difference in light levels from the front of the glasshouse to the rear; differences in temperature caused by the air-conditioning unit at the rear of the glasshouse and the door at the front).Dung pat position nested within mesocosm accounted for any possible behavioural effect caused by the position of each dung pat within mesocosms (see supplementary materials for a visual explanation).In model testing, significant zero-inflation of model residuals was detected, so a zero-inflated Poisson model was used (with a random intercept for zero-inflation, i.e., zi $ 1).
the model-predicted frequency of brood ball production per dung F I G U R E 1 Number of brood balls produced per female by (a) Onthophagus taurus (n = 10 replicates per glitter treatment group) and (b) Euoniticellus fulvus (n = 8 replicates per glitter treatment group).Control = dung with no glitter.Confidence intervals and means are the model-predicted effects from the generalised linear model (GLM) analyses.Grey circles represent raw data points.Points are offset for visual clarity.T A B L E 1 Model parameters for the no-choice glitter tests with (a) O. taurus and (b) E. fulvus.

F
I G U R E 2 The proportion of brood balls of each colour produced by Onthophagus taurus when given the choice between four 240 g dung pats containing glitter of differing colours in each mesocosm: Gold, Silver, UV-Pink and UV-Yellow.The dashed line at 25% represents the expected proportion under the null hypothesis of no effect of glitter colour on brood ball production.Error bars represent 95% confidence intervals.Confidence intervals and means are the model-predicted effects from the multinomial generalised linear mixed model (GLMM) analysis.Grey circles represent raw data points, with circle size scaled by the number of overlapping count values.Points are offset for visual clarity.
Conditional and zero-inflation model parameters for the glitter choice tests with O. taurus presenting treatment-specific parameters.
T A B L E 2Abbreviations: CL, confidence limits; p, p-value; SE, one standard error of the mean; z, z-value.
. Ultimately, this simple and inexpensive marking method could provide further insight into dung use by dung beetles and other dung-associated species, and perhaps the fate of dung in food chains.