Some spiders use sand grains, soil particles, prey remains, and plant debris to build detritus-based, bell-shaped cobwebs. The retreat and gumfooted lines in cobwebs are thought to have defensive and foraging functions. Extended phenotypic plasticity has been proposed as a means by which animals alter their phenotypes in response to  changing conditions. Animals may display extended phenotypic plasticity as a consequence of environmental variation. The aim of this study was to test both the extended phenotypic plasticity and the trade-off hypothesis by looking at whether a common cobweb spider Campanicola campanulata alter web architecture and behavioral investments in response to the availability of prey. When prey availability increases, spiders build webs higher up, build smaller and lighter retreats, and cobwebs with fewer gumfooted lines and smaller capture areas. The results showed that spiders invested less in both foraging and defensive investments in the presence of a large number of ant prey. These results suggested that the web architecture of C. campanulata exhibits a high degree of plasticity in response to variations in prey availability, but the behavioral investments did not involve trade-offs between foraging and defense.

Study species
Campanicola campanulata (Araneae: Theridiidae) is a cobweb-building spider that is widely distributed in Hubei, Zhejiang, and Guizhou province of China (Chen 1993). These spiders gather in cool but rain-proof habitats such as forests, farmlands or hillsides and mainly feeding on ants. We have observed that spiders often build their webs in places such as concave rocks, earthen walls, or under low-hanging branches. Each individual spider builds a complete cobweb that consists of anchor silks, dozens of sticky gumfooted lines, and a bell-shaped retreat made of silk, sand, pebbles, debris, leaves, or shells. 

The anchor silk, composed of strands of silk, connects the bell-shaped retreat to stone walls, tree branch or other substrates, providing stability and preventing wind damage to the web. The bell-shaped retreat, made of silk-wrapped sand grains, soil particles, prey remains and plant detritus, is usually suspended and lined with silk tunnels (Henschel and Jocqué, 1994). This structure significantly reduces the spider′s vulnerability to enemies and provides a safe space for resting, , mating, spawning, and caring for its young. It also offers protection against predators such as araneophagy spiders or birds. 

The capture web consists of a few to several dozen radial gumfooted lines that extend from the edge of the bell-shaped retreat and are fastened to coarse sand grains or fixed substrates. These lines primarily function to trap preys walking underneath the web and prevent the invasion of enemies. Once ants are intercepted by sticky gumfooted lines, the vibratory signal generated by the ant's struggle will be transmitted to the spider in the retreat. The spider then tear the lines away from the substrate and pull its prey up toward the retreat, making it difficult for the ant to get the leverage needed for escape (Argintean et al. 2006). 

Spider collection and maintenance
From 20 to 25 March 2022, we collected sub-adult (i.e., 1 molt before adulthood, ~2.4 mm in body length) C. campanulata from Guanyin Lake Scenic Area (31º 14′ 8.66′′ N, 114º 01′ 21.57′′ E) in Hubei Province, China, and raised them to adulthood in the laboratory. The spiders were maintained individually in a plastic box (length × width × height: 10 cm × 5 cm × 10 cm) with bamboo sticks attached to their inner sides to allow them to build webs. The front and back sides of the box were made of removable transparent Perspex glass. All spiders were kept in the laboratory with controlled environmental conditions (temperature: 25 ± 1℃; relative humidity: 80 ± 5%; photoperiod: 14 h:10 h L : D). The bottom of the box was covered with a layer of sand for the spider to collect to build a retreat, and a piece of sponge with absorbed water was placed at the bottom of the box to provide water for the spider. We fed each spider with 2 ants (Monomorium sp., ~2.2 mm in body length) every 2 days, and the developmental status (molting) of each spider was checked twice daily (09:00 and 21:00).

Experimental design and procedure
We selected sixty newly matured female spiders (~2 days after maturity) that have been well-fed for 3 days and were randomly divided into 4 treatments (N = 15 in each group) with different number of ants : blank control (CK)—0 ants, low prey availability (LA)—5 ants, medium prey availability (MA)—10 ants and high prey availability (HA)—20 ants, respectively. All the spiders were transferred to a box with a layer of sand grains at the bottom, and a wooden stick attached to the sides and the inner wall at the top to form a wooden frame. Sand grains were collected in the wild habitat for spiders to build the retreat, while wooden frames was used as a climbing fulcrum for spiders to build their webs.The body size (carapace width, body length, and body weight) of each spider was measured before the experiment. We measured the body condition parameters (carapace width, body length, and body weight) of each spider before the experiment. All spiders were provided with Monomprium ants and water ad libitum two days before the experiment and were given 36 h to allow them to build webs, none were fed during the experiment. All the trails were recorded using video camera (HDR-CX 680; SONY).

Latency and duration of web-building
After being introduced into the experimental setup, the spider will have a period of exploration of the new environment, during which spiders spin silk and fasten it to the appropriate sand grains. Then spiders climb to the top of the setup via sticks on both sides and select a suitable anchor point to begin transporting sand grains. The period from when spiders were introduced into the experimental setup until they begin to transport the first sand grain is defined as the latency of web-building. Due to the energy consumption of the spider to construct the retreat and the gunfooted lines (Blamires et al. 2014), the spider will have several rest times during the web-building, but the spider will not rest for more than 2 hours at a time (H. Zhang’s pers. obs.). Once the spider has rested for more than two hours, we consider the web-building process to be complete. The period between spiders begin to transport the first sand grain and when the spider finished web-building is called the duration of web-building. For each spider, we recorded the latency and duration of its web-building by analyzing the videos, because we believe that these parameters can more or less reflect the spider's decision on web-building behavior when different numbers of ants are present. Ants are not only dangerous prey but also potential predators for spider, and the roles of predator and prey may be reversed if one of them makes a mistake (Líznarová and Pekár, 2013; Peká rand Toft, 2015).

Web architecture measurements and behavioral investments quantification
For each web, we measured the parameters of web architecture including the length of anchor silk (ASL), the height of retreat (RH), the number of gumfooted lines (GLN), the height of the retreat center from the ground (CRH), the diameter of retreat entrance (ED), the length of each gumfooted line (GL), and the weight of the retreat (RW) (Figure 1D). Because the projection of the gumfooted lines on the ground is irregular, some gumfooted lines are near to the center, and the others are far away to the center, so we measured the length of the most inner (CRI) and the most outer (CRO) capture radius and use the average of the them as the average radius (CRA) to estimate the capture area. RW was measured after the other 8 web architecture parameters were measured. We cut the retreat off from the web and measured its weight to the nearest 0.01 mg using an electronic balance (FA1004N type, HANGPING). 

It is well known that the capture area and retreat volume in the cobweb are important indicators of the web-building strategy and behavioral investment in capturing prey and avoiding predators because spiders with larger capture areas will have higher rates of prey encounter and foraging success (Benjamin and Zschokke, 2003; Blackledge and Zevenbergen, 2007; Salomon 2007; Zevenbergen et al. 2008; Fischer et al. 2023; Sergi et al. 2020). While spiders with larger retreats are less likely to be discovered by predators and thus face lower  predation risk (Manicom et al. 2008). We used the same formula in previous studies to calculate the capture area and retreat volume (Zhang et al., 2023a; b) and also quantify the foraging investment (i.e., total length of gumfooted lines, GTL) and defense investment (i.e., energy consumption in building retreat, EC) during web-building (Zhang et al., 2023a; b).  

Statistical analysis
We checked the data for the normality and homogeneity of variance using Shapiro–Wilk and Levene tests, respectively. Since the data were not normally distributed, we used the Kruskal–Wallis test to compare the differences in web architecture and behavioral investments between treatments. If there was a significant difference in the results, Wilcoxon’s rank sum test with continuity correction was used to pairwise comparative the results between any 2 treatments. We performed all statistical analyses using R 4.4.2 (R Core Team 2025). All tests were 2-tailed, and the P-value for significance was set at < 0.05.