In ecosystems, the rates of resource consumption by animals drive the flows of matter and energy. Consumption rates are known to vary according to consumer energy requirements, resource nutrient content and mechanical properties. The aim of our study is to determine how mechanical constraints, compared to energetic and nutritional constraints, explain the variation in leaf litter consumption rates by macrodetritivores. In particular, we focus on the impact of litter toughness. To this end, we propose a non-linear model describing leaf litter consumption rates of detritivore as a function of litter toughness. We also investigate a possible match between bite force and litter toughness, since consumer-resource co-occurrence is thought to be driven by the match between invertebrate mandibular traits and resource toughness. Our study was designed as follows: leaf litter from oak and hornbeam was exposed to field physical and microbial decomposition in aquatic and terrestrial ecosystems for selected time periods before it was offered to eight macrodetritivore taxa (three forest stream taxa and five forest soil taxa) in no-choice laboratory feeding experiments. Our findings show that, compared to energetic and nutritional constraints, mechanical traits have a greater impact on litter consumption rate by detritivores. After subtracting the contribution of the detritivore body mass, we report that litter consumption rates depend primarily on litter toughness. A sigmoid function is best suited to characterize the relationship between mass-independent consumption rate and litter toughness. We note that the parameters of our sigmoid model are taxon-specific, suggesting biomechanical thresholds and biological differences among taxa. Interestingly, we found no correlation with detritivore bite force, suggesting that food processing by detritivores does not only depend on mandibles strength.

Sites and experimental design

Our sampling sites were located south of Toulouse (France), in the Pyrenean Piedmont, an area characterized by a calcareous bedrock and continental climate. Leaf litter and detritivores were collected in oak-hornbeam forests. For each ecosystem type (low order streams and forest floors), we selected two sites, one dominated by oak (Quercus petraea) and the other by hornbeam (Carpinus betulus). Hornbeam sites were both located in a hornbeam coppice stand. The terrestrial oak site was a mature even-aged oak forest. As there was no stream crossing it, we selected a nearby stream running along a forest edge dominated by oak trees. Leaf litter was collected at the time of abscission and stored air-dried at room temperature. Leaf litter was also collected later, after various physical and microbial decomposition times, to produce litter with varying toughness values. In our feeding experiments, detritivore individuals collected at a site were only offered the locally dominant leaf litter (either oak or hornbeam). Furthermore, we collected detritivores in the field at the same time as partially decomposed leaf litter.

Leaf litter

Freshly fallen leaves of oak were expected to be tougher than those of hornbeam and the toughness of both leaf species was expected to gradually decrease with length of exposure in the field. Five grams of air-dried leaves were enclosed in 0.5-mm nylon mesh bags (20 × 15 cm), thus preventing macrodetritivores to enter the bags. In autumn 2020, oak and hornbeam litter bags were incubated at oak and hornbeam sites, respectively. They were arranged in three blocks in each site. In streams, bags were secured to iron sticks anchored in the sediment, whereas on forest floor, bags were laid flat onto the topsoil and secured with bamboo sticks driven into the soil. Assuming that leaf litter disappears faster in streams than in soils, we collected litterbags after 54 and 99 days of exposure in the streams, and after 82, 111, 152 (hornbeam only), 195 (oak only), 236 and 349 (oak only) days of exposure in terrestrial sites. At least one litterbag was retrieved from each block at each time point to cover a wide range of litter toughness. However, as hornbeam leaves decomposed more rapidly than oak leaves, only oak leaves bags were recovered from the terrestrial sites on the last sampling date. Partially decomposed leaves recovered from litter bags were rinsed with tap water to remove exogeneous particles. Discs were punched through the leaves with a 10-mm cork borer, avoiding central veins, to provide homogeneous food for detritivores. At each site, about 200 discs per sampling date were needed to assess litter traits (described thereafter) and perform feeding experiments. We also made around 300 discs from undecomposed oak and hornbeam leaves (t = 0). As the leaves were stored air-dried, they were soaked in water for an hour before being cut into discs. Leaf discs were freeze-dried, partitioned into sets of five (except for the first oak test for which we used sets of four discs, as consumption was very low), and weighed to the nearest 0.1 mg.

We assessed litter toughness by measuring the force required for a 2.2-mm diameter steel punch to penetrate a leaf disc laid flat on a perpendicular plane. We used a custom-made penetrometer, such as described in Graça et al. (2005), fitted to a digital force tester (CS225 Series, Chatillon®). This device provides high accuracy (± 0.1 N) when measuring penetration force. Litter toughness is equal to the penetration pressure (kPa), i.e., the penetration force divided by the cross-sectional area of the punch. Litter softness (kPa−1) is the inverse of penetration pressure. We calculated the average of 15 replicate measurements for each batch of leaf litter (from a same site at a given sampling date).
Litter N content was considered an indicator of its nutritional value, assuming that (1) detritivores would experience strong N limitation and (2) contents of other limiting nutrients (e.g. Phosphorus) would positively covary with N content. The latter assumption was found to hold true when comparing undecomposed oak and hornbeam leaves based on analyses of litter N, P, K, Ca and Mg.

Three replicates of 3–6 mg per leaf discs batch were grinded into powder by the means of a bead mill (FastPrep-24™ Classic Instrument). N content was quantified using a Total nitrogen and Organic Carbon analyzer (TOC L, Shimadzu) and values were reported on a dry weigh basis (mg N mg−1 litter), then expressed as a percentage.

Macrodetritivores

We collected macrodetritivores from natural accumulations of organic debris in the sites where litterbags were placed. We attempted to capture enough individuals of each taxon to ensure adequate levels of replication of experimental conditions (i.e., pairs of detritivore taxon and litter batch, at each time point). However, we did not achieve a fully-balanced design because the structure of detritivore assemblages differed between oak and hornbeam sites and population abundances exhibited large seasonal variations. We first conducted a feeding experiment with pre-weighed discs of undecomposed leaf litter (i.e., air-dried leaves used to fill litter bags) and detritivores collected in the field at litterbag installation. Feeding experiments were repeated whenever litterbags and detritivores were retrieved from the sites.

Individuals were assigned to coarse taxonomic categories in the field and were then identified at the lowest practicable taxonomic level, mostly genus, based on observations made on dead individuals under a dissecting microscope. We found case-bearing Trichoptera larvae and amphipods (Crustacea) at both stream sites, and Plecoptera larvae only at the hornbeam site. Amphipods and Plecoptera individuals were both assigned to a single genus, Gammarus and Capnia, respectively, and Trichoptera individuals to the Limnephilinae genus complex. Detritivores from the terrestrial sites were either Isopods (Philoscia and Porcellio) or Diplopods (Leptoiulus, Cylindroiulus, and Glomeris). Other genera (e.g., Polydesmus) were too scarce. Detritivores sorted by taxon and site of collection were held in plastic containers stored in the dark at 10 °C for 1–4 days. Containers with aquatic taxa were filled with permanently-aerated stream water.

Pre-weighed leaf discs were rewetted for one hour before they were offered to one individual of either detritivore taxon, except Capnia. For this small-sized detritivore, we pooled together up to five individuals of the same size to ensure accurate quantification of per capita consumption of leaf litter. We used 10-cm large plastic containers filled with 50 g of clean sand as experimental units. The sand was sprayed with 5 mL of water to maintain wet conditions required for terrestrial detritivores, whereas 200 mL of stream water was added to containers with aquatic taxa. Detritivores were starved for at least 18 h prior to the introduction of pre-weighed leaf discs. Experimental feeding time was 4 days and 2 days for terrestrial and aquatic detritivores, respectively, to account for their respective feeding rates. Feeding experiments were stopped earlier if disc area had been downsized by 75% (visual estimation). For each leaf litter species and ecosystem type, we set up five control experimental units with no detritivore, in order to correct for leaf mass loss caused by physical and microbial decomposition. Leaf mass loss in control containers did not exceed 10% of initial mass.

At the end of feeding trials, remaining leaf discs and fragments larger than 1 mm were freeze-dried and weighed to the nearest 0.1 mg. Detritivores were blotted dry with paper towels and weighed to the nearest 0.1 mg. For each taxon, wet body mass was converted into dry mass using a linear relationship established with freeze-dried individuals (n ≥ 12, p < 0.001, R2 ≥ 0.68). Five detritivore individuals per site, chosen to be representative of the diversity of conditions (temporal and geographical), were preserved in 70% ethanol, dissected and photographed under a stereomicroscope equipped with a digital camera (Olympus SZX10) to measure (0.01 mm accuracy) head width (H_W), mandible length (M_L), and width (M_W). Mandible length was measured between the incisor tip and the axis of rotation, identified by the condyles. Mandible width was measured between the adductor muscle insertion and the axis of rotation. Head width was measured just behind the eyes. Each taxon’s bite force (F index) was calculated as follows (Wheater and Evans 1989): F = H_W M_w / M_L

Note that, as head width H_W is a proxy for the force of the muscles involved in mandible motion, the unit of the F index is not relevant. A 2-D plot (F index versus dry body mass) shows morphological differences between dissected individuals.

Data analyses

Because it is more convenient to describe resource-mediated constraints on consumer performance by increasing functions (e.g., Monod equation, predator–prey function responses), we used the inverse of toughness (i.e., softness) as a predictor of litter consumption rates in statistical models. The conversion from toughness to softness made the data distribution less skewed, which helped to estimate parameters from linear and non-linear models.

Our dataset consisted of 291 values of litter consumption rate estimated on 63 pairs of detritivore taxon-litter batch. We calculated the mass-specific rate of litter consumption (?? in mg mg−1 day−1) as follows: ?? = (??−(???+?)) / ?Δ?, where ?? and ?? are the final and initial mass (mg) of the leaf discs, respectively; Δ? is the test duration; and B is the final dry body mass of the detritivore (mg). The equation term aMi + b was used to correct the initial litter mass for physical and microbial decomposition assessed in control containers, without detritivore. The coefficients a and b are the slope and intercept of the linear regression of final vs initial litter mass in controls (n = 5) for each condition (oak and hornbeam leaf discs incubated in terrestrial and aquatic habitats for each time point). As the lowest limit of quantification for litter consumption rate was estimated at 0.001 mg mg−1 day−1 (ca. 0.1 mg of litter weighing accuracy, divided by 100 mg (order of magnitude) of detritivore body mass for larger individuals), lower values have been rounded down to zero (n = 46). No litter consumption rate was calculated whenever a detritivore died during a feeding experiment (5% occurrence; n = 20). We have developed a linear model to assess the relative impacts of litter softness, litter N content and detritivore body mass on mass-specific litter consumption rate (?? > 0.001 mg mg−1 day−1, n = 245). Litter N content reflects nutritional constraints, whereas detritivore body mass reflects energy-mass allometry. The dependent variable and all the predictors were log-transformed to linearize the expected power-law relationships and to achieve assumptions of homoscedasticity and normal distribution of residuals.

Further analyses were carried out on mass-independent rate of litter consumption (?? in mg mg−1−c day−1), calculated as follows: ?? = ??∙?^−? , where ?? is the mass-specific rate of litter consumption, B is the final dry body mass of detritivores (mg), and ? is the estimate for the effect of body mass in the linear model. We fitted a random-effects model to assess the contribution of experimental factors (ecosystem type, detritivore taxon, leaf species, and sampling date) to ?? variability. A square root transformation was applied to the data to ensure residual normality and homoscedasticity. We assumed that litter consumption rate (??) evolves as a function of litter softness (S) according to a S-shaped curve, with the transition phase informing on mechanical traits of both the consumer and the resource. A three-parameter sigmoid function was fitted to data for each detritivore taxon, using a non-linear regression: ?? = ? / (1+???((?−?) / ?))?, ? and ? are the maximum consumption rate, the softness value at the inflexion point, and a scale parameter, respectively. To assess the relevance of sigmoid models, we compared the fit of the sigmoid model with that of a linear model using root mean square error (RMSE) and Akaike Information Criterion (AIC). Parameter estimates were extracted of each model in order to assess cross-taxon variability.

Graça MA, Bärlocher F, Gessner MO (2005) Methods to study litter decomposition: a practical guide. Springer Science & Business Media, Berlin

Wheater CP, Evans MEG (1989) The mandibular forces and pressures of some predacious Coleoptera.J Insect Physiol 35:815–820.https://doi.org/10.1016/0022-1910(89)90096-6