Fear of predation can affect important ecosystem processes by altering the prey traits' expression that, in turn, regulates the quantity and quality of nutritional inputs to soil. Here, we aimed to assist in bridging a knowledge gap in this cascading chain of events by exploring how risk of spider predation may affect grasshopper prey performances, and the activity of various microbial extracellular enzymes in the soil. Using a mesocosms field‐experiment, we found that grasshoppers threatened by spider predation ate less, grew slower, and had a higher body carbon to nitrogen ratio. Herbivory increased activity of all microbial extracellular enzymes examined, likely due to higher availability of root exudates. Predation risk had no effect on C‐acquiring enzymes but decreased activity of P‐acquiring enzymes. We found contrasting results regarding the effect of predation on the activity of N‐acetyl‐glucosaminidase and leucine arylamidase N‐acquiring enzymes, suggesting that predation risk may alter the composition of N‐inputs to soil. Our work highlighted the importance of soil microbial enzymatic activity as a way to predict how changes in the aboveground food‐web dynamics may alter key ecosystem processes like nutritional‐cycling.

Site description and experiment design

We conducted our field experiment at Jilin Songnen Grassland Ecosystem National Observation and Research Station, Changling, China (44°45′ N, 123°45′ E). The climate in this site is semi‐arid continental and monsoon with average annual precipitation ranging from 280 to 400 mm. This system is dominated by the grass L. chinensis, and its main insect herbivore E. unicolor grasshopper. Wolf spiders are common predators of grasshoppers in this system.

We tested how the risk of spider predation regulates grasshoppers' trophic function and subsequently soil microorganism enzymatic activity and stoichiometry, by setting a randomized block experiment with three‐trophic‐level treatments: (i) plants (control), (ii) plants and grasshoppers (herbivore), and (iii) plants, grasshoppers, and spiders (predation risk; Figure 1). We constructed 10 blocks of three 0.25‐m²‐basal area, 0.9‐m‐high mesocosms covered by 0.8 mm × 1.2 mm mesh size screen. All mesocosms were constructed over naturally growing vegetation of similar composition, in which L. chinensis accounted for about 75% of the soil surface coverage. The blocks were at least 10 meters apart. Before stocking the mesocosms, we removed all naturally occurring invertebrates using a vacuum cleaner. Grasshopper density in this field site reaches a maximum of 20–30 per m² (Zhong et al., 2014). Thus, we stocked each mesocosm with eight third-instar grasshoppers and added one wolf spider to each predation mesocosm. We glued the spider chelicerae with quick‐drying glue to prevent them from preying upon grasshoppers (Hawlena & Schmitz, 2010). Previous work in our own study system, and elsewhere have shown that spiders with glued chelicerae can remain active for more than 2 months (Barton, 2010; Hawlena et al., 2012). Our experiment lasted from 10 August to 20 September 2020.