Biological soil crusts (BSCs) are communities of microorganisms that control ecosystem functions in drylands. Despite their importance, little is known about how trophic interactions affect BSC function. We conducted a series of mechanistic experiments to tease out the direct (i.e., consumption) and indirect (i.e., fecal and mucus deposition) pathways by which crustivore (i.e., consume BSCs) and detritivores affect BSC functions— complemented by a manipulative field experiment exploring the integrative effect of these pathways. We showed that detritivore feces, mucus, and grazing increased the BSCs CO2 efflux. Detritivore feces also increased BSC %N. Crustivorous snail feces increased BSC CO2 efflux, and their mucus decreased BSC %C and %N. In the field, both snail species increased BSC %C and did not affect BSC CO2 efflux. Combined, our findings highlight that macro-invertebrate consumers exert top-down regulation on BSC function, opening the door for a whole new avenue of trophic research.

Study site and species

The Avdat Research station (30°47’02” N, 34°46’09” E; hereafter, Avdat) is in the central Negev highland and receives ~93 mm of rainfall year^-1 across 19-42 days (Israel Meteorological Survey, 2021; Station: 253052).  Avdat is characterized by 1-2 mm thick cyanobacteria dominated BSC and has multiple common snail species, including Xerocrassa simulate and Sphincterochila prophetarum, which reach densities of 12 snails m^-2 (Rinehart et al. 2021).  Xerocrassa simulate (hereafter, detritivore) consumes plant litter and BSCs (Ward and Slotow 1992), while S. prophetarum (hereafter, crustivore), feed exclusively on BSCs (Shachak and Brand 1981, Appendix S1). A single crustivore can consume 203 – 577 mg of BSC day^-1 (Appendix S1). Both snail species break their aestivation and become active after substantial precipitation events when the ground becomes wet. Then, snails begin foraging on BSCs leaving behind feces and mucus trails.

Mechanistic laboratory assays

Laboratory grown BSCs. We grew laboratory BSCs by adding 25 ± 1.0g of soil, followed by 25 ± 1.0g of BSCs to 120 mm Ø petri dishes. We used laboratory generated BSCs to minimize natural variation in BSC function.  This method has been employed previously in BSC experimentation (Doherty et al. 2015, Rinehart et al. 2021). Soils and BSCs were collected from Avdat. Laboratory-reared BSCs were watered with DI water every-other day for 80 d then left dormant in an incubator at 16°C with 40% humidity and a 16:8 light cycle (light intensity: 240Mmol m^2-1 s^-1) until used.

Fecal deposition. Snail crustivores and detritivores collected from Avdat were placed inside circular plastic containers in groups of three, by species. All containers were provided BSCs and plant litter (a 50:50 mix of Hammada scoparia and grasses) ab labium then placed in an incubator at 18˚C with 40% humidity and a 16:8 dark: light cycle. Every other day, we collected any feces produced by snails and refreshed all food resources. Collected feces were freeze-dried and their dry mass (DM) determined. Crustivores generated 77 ± 7.9 mg DM (mean ± 1SE) of feces day^-1, while detritivores only generated 13 ± 0.2 mg DM of feces day^-1. This difference in fecal production is not due to differences in snail biomass (crustivores: 1.29 ± 0.1 g WM; detritivores: 1.24 ± 0.2 g WM). We allocated fecal samples, by consumer species, to laboratory BSCs in the following three consumer treatments: crustivore, detritivore, and no snail (n = 15).  We added 130 ± 10 mg DM of homogenized detritivore feces to detritivore BSC replicates and 770 ± 10 mg DM of crustivore feces to crustivore BSC replicates. We did not standardize the DM of feces added across species because we wanted to know the mean impact of feces snail^-1. No feces were added to no snail BSC replicates. All BSCs were watered every-other day with 3.5 ml of DI water and housed in an incubator at 18˚C with 40% humidity and a 16:8 dark: light cycle. Ten days into our watering regime we paused our study for 30 d due to Covid-19. After the lockdown, we resumed our watering regime for eight days before measuring the CO₂ efflux, %C, and %N of each BSC. This delay should not affect our results as BSCs are inactive when dry (Weber et al. 2016).

Mucus deposition. Snail crustivores and detritivores collected from Avdat were randomly assigned to plastic containers in groups of 14 individuals by consumer species (n = 7). Containers with snails were maintained in a climate-controlled room at 14-17°C with an 11:13 light: dark cycle. Snails were incubated in an 11:13 hour light cycle, rather than the 16:8 hour light cycle used for BSC incubations, to further promote movement and mucus deposition. Snail consumers in all containers were provided BSCs and H. scoparia ad labium. Mucus was extracted from snails weekly for three weeks. During each extraction, snails in each container were housed on a transparent plastic sheet and wetted with 1ml of DI water daily. After 48 hour, we scraped mucus off the transparent plastic sheets with razor blades— using 14 ml of DI water per container to facilitate scraping. Collected mucus was stored at -80°C until all extractions were complete. After the third extraction, we freeze-dried 1.0 ml mucus samples to determine their DM. The remaining mucus was homogenized by snail consumer species and allocated to laboratory BSCs in the following three consumer treatments: crustivore, detritivore, and no snail control (n = 20-21). Mucus harvested from snails was allocated, by species, to laboratory BSCs in the following three consumer treatments: crustivore, detritivore, and no snail control (n = 20-21). We watered crustivore and detritivore BSCs with 1.5 ml of homogenized crustivore or detritivore mucus, respectively, and 8.5 ml of DI water every-other day. No snail BSCs were watered with 10 ml of DI water every-other day. All BSCs were housed in an incubator set at 18˚C and 40% humidity with a 16:8 dark: light cycle. After 14 d, we quantified the CO₂ efflux, %C, and %N of each BSC.

Grazing and mucus deposition. We manipulated snail consumer treatment (crustivore, detritivore, or no snail) within circular plastic containers (Ø = 25 cm; height = 15 cm). In crustivore and detritivore treatments, we included three snail crustivores and detritivores, respectively, all collected from Avdat and starved for 6 d prior to the assay. No snail treatments received no snails. All containers received a single pre-weighed laboratory reared BSC. Containers were then placed in a climate-controlled room at 18˚C with a 16:8 dark: light cycle. We watered all treatments with 3 ml of DI water daily to promote snail activity. All feces generated by snails were removed daily before watering occurred. After 4 d, we measured the biomass loss of BSCs, CO₂ efflux, %C, and %N of each BSC.

Field experiment

At Avdat, we deployed ten blocks of three 0.25 m² circular enclosures on natural BSCs. In each block, we added five crustivore snails to a randomly chosen enclosure, five detritivore snails to another, and the third enclosure was a no snail control. The enclosures were constructed from flexible PVC tubing (40 mm Ø) anchored in place with metal pegs and painted with antifouling paint. Antifouling paint is used to restrict gastropod movement in marine communities (Bracken et al. 2011) and effectively restricted snail movement in our system. We used antifouling paint covered enclosures to minimize climatic effects of typical enclosures. All enclosures started with no vascular plants but were equipped with 1) a fake plant constructed of plastic-coated wire and plastic leaves, 2) 1.5 ± 0.1g of H. scoparia litter, and 3) a rock of standard size (~ 6 cm long, ~ 4 cm wide). The fake plant and rock were included to provide refuge and aestivation sites for the snails. We monitored all enclosures weekly and replaced any snails that escaped the enclosure (no ‘wild’ snails entered the enclosures). After 113 d, we collected five, 120 mm Ø BSCs samples from each plot. BSCs samples were housed in an incubator at 16°C with 40% humidity and a 16:8 light: dark cycle until CO₂ efflux, %C, and %N were quantified. For all our BSC responses we averaged the five subsamples collected from each enclosure.   

BSC function measurements

Biomass loss. We obtained the initial dry biomass of BSCs added to each experimental replicate. Upon completion of the study, we let BSCs dry for 24 hours in an incubator at 18˚C with 40% humidity and a 16:8 dark: light cycle (light intensity: 240Mmol m^2-1 s^-1) before obtaining the final dry biomass of each BSC. Biomass loss was then calculated for each BSC (Initial Dry Biomass – Final Dry Biomass = Biomass Loss).

CO₂ efflux. We quantified the BSC CO₂ efflux by placing watered (3 ml of DI water) BSCs in airtight plastic chambers [155mm x 155mm x 61mm (length x width x height) Lock & Lock HPL 823; www.s-d.co.il]. We flushed chambers with CO₂-fress air at a rate of 2L minute¹ for a total of 5 minutes. The BSCs were then incubated in their flushed, airtight chambers at 18°C with no light for a total of 48 hours. After the dark incubation, we used a LI-7000 CO2/H2O infrared gas analyzer (IRGA; LI-COR, Inc.) with a designated self- manufactured injection system to quantify the amount of CO₂ released by each BSC sample. We quantified the CO₂ efflux of BSCs used in our fecal deposition, mucus deposition, and grazing laboratory assays at the start and end of each assay, allowing us to calculate the consumer-mediated change in BSC CO₂ efflux (i.e., Final CO₂ efflux – Initial CO₂ efflux = Change in CO₂ efflux). We used the change in CO₂ efflux for all statistical analyses of our laboratory assays. In our field study (see “Long-term field study” in Methods section of main text), we only quantified BSC CO₂ efflux at the end of the study (i.e., final BSC CO₂ efflux) since our method of measuring CO₂ efflux required destructive sampling. Because we collected five BSC sub-samples from each of our experimental plots, we pooled the results of these sub-samples and used these pooled means for all our statistical analyses.

Total carbon and nitrogen. After completion of CO₂ efflux assays, we homogenized five, 0.8 cm³ systematically collected (“x” shape with samples at the ends of each arm and center) sub-samples from each BSC sample. We randomly selected 12 of our 20 BSCs in our Grazing and mucus deposition laboratory assay for total carbon and nitrogen analysis, in all other laboratory assays we quantified total carbon and nitrogen in all BSCs (Table S1). For our field study, we further homogenized subsamples collected from each of the five BSC samples harvested within each plot to generate a single homogenized sample from each field plot. All homogenized BSC samples were then freeze-dried for 48 hours and packaged (29-30ug sediment sample^-1) for analysis of total C and N at the University of Georgia’s Center for Applied Isotope Studies. We then calculated the %C and %N of each samples and used these percentages for all statistical analyses.

Treatments are consistant across all datasheets, where Crustivore is biological soil crusts exposed to crustivorous snails and their relative byproducts (i.e., feces or mucus), Deteritivore is biological soil crusts exposed to deteritivorous snails and their relative nyproducts (i.e., feces or mucus), and No Snail are biological soil crusts exposed to no snail consumer byproducts (i.e., served as our control for snail effects). 

Missing Values: 

In the "Feces Nutrient Composition" dataset, the %N values for Crustivore are listed as "Below Detection Limit" because their %N content was below 0.2%-- the detection limit of University of Georgia Isotope Laboratory's equipment. 

In the "Grazing Assay" dataset, we report biomass loss and CO₂ efflux for 20 replicates and %C, %N, and C:N for 12 replicates. We do not have all 20 values for our stoichiometry measurements because one of our plates was unreturned from the University of Georgia Isotope Laboratory.