Human settlements and urbanization are increasing globally with more than half of the Earth’s terrestrial surface being impacted by humans. This development has resulted in numerous anthropogenic stressors including nocturnal sensory pollution (i.e. light pollution), which is a key driver of insect declines. Nocturnality is hypothesized to reduce predation risk from visually-guided diurnal predators. More than half of all insect species, and 80% of Lepidoptera, are estimated to be nocturnal. Predation rates on insects are likely a result of habitat, time of day, and the local predator composition. We investigated how predation rates on plasticine moth replicas differed between urban and rural sites, and between night and day. Visually matching paper-winged, clay-bodied replicas of the white-lined sphinx moth, Hyles lineata, were placed in a natural area within the city of El Paso, Texas, and in remote Chihuahuan Desert with minimal human disturbance. These replicas were checked during dawn and dusk for three days. Predation rates were significantly lower at night than during the day regardless of location, and predator composition differed between sites. Insectivorous birds were the primary diurnal predators in both locations, whereas nocturnal predators were represented primarily by insects at the rural site and by mammals at the urban site. These findings support the hypothesis that visually-guided predators, such as birds, exert higher predation pressures during the day, and supports the hypothesis that insect biodiversity, especially of predaceous insects, is affected by urbanization.

Plasticine Replicas

To quantify predation rates and predator composition between time of day and urbanization, we developed visually accurate plasticine replicas of adult Hyles lineata. Eight H. lineata moths in pristine condition were selected from the University of Texas at El Paso Biodiversity Collections. Following Troscianko & Stevens (2015) and Da Cunha et al. (2024), we quantified wing coloration with a UV/VIS Samsung NX1000 camera. We photographed moth specimens both dorsally and ventrally with a UV-suitable gray standard (Spectralon; Labsphere; 40% reflectance) under clear skies between 13:15 and 14:15 in February and March. We then printed moth replica wings on White, Letter Paper (Office Depot® Multi-Use Printer & Copy Paper, White, Letter 8.5" x 11) using a Xerox C310 printer. Photographs of moth specimens and paper wings were compared for spectral match. We adjusted the brightness and RGB values of the paper wing reflectance using the photo editing software General Image Manipulation Program (GIMP) to ensure a close spectral match to the moth specimens. This process was repeated until reflectance values were similar between moth wings and paper wings, see Figures 1 and 2. Paper wings were then cut to the correct shape and size, and inserted into 4.5 grams of non-toxic brown clay (Brown Craft Smart Oven Baked Polymer) molded to the shape of H. lineata’s body. Finally, twist ties were used to attach clay replicas to branches.

Study Sites and Replica Deployment

For moth deployment we selected one urban site within El Paso, TX - Knapp Land Nature Preserve (KLNP; 31.864035, -106.465106) - and one rural site, Indio Mountains Research Station (IMRS; 30.776681, -105.016751), which is approximately 180 km from El Paso, TX. KLPN is 143 hectares of protected Chihuahuan Desert with high human disturbance whereas IMRS consists of 161 km²of undisturbed Chihuahuan Desert.. Additionally, KLNP is adjacent to Franklin Mountains State Park, which consists of 9,182 hectares of preserved Chihuahuan Desert. However, both KLNP and adjoining Franklin Mountains State Park are centrally located within El Paso and receive high levels of light and noise pollution, as well as other urban impacts.

During dusk on June 8, 2023, 200 moth replicas were placed on native plants at each site, 400 in total, at least 50 meters apart from other replicas. Replicas were then checked every dawn (approx. 04:58-06:00) and dusk (approx. 20:10-21:12) for attack marks for 72 hours. Attacks recorded at dawn indicated a nocturnal attack and attacks at dusk indicated a diurnal attack. All attacked replicas were removed from the study and photographed. We categorized attacks as bird, mammal, insect, or unknown based on the indentations left on the clay replicas (see Low et al., 2014 and  supplementalfigure 1).

Statistical Analysis

Differences in survival probabilities after 72 hours between the two sites were tested using Kaplan-Meier survival analysis with the log-rank test ‘survival’ package in R (Therneau & Others, 2015). Missing replicas were incorporated into the Kaplan-Meier survival analysis as censored individuals (Seymoure et al., 2024). Chi-square analyses were used for testing differences in attack rates between night and day, as well as predator composition. For the temporal niche Chi-square analysis, we corrected for a longer daytime by multiplying the total number of attacks (140) by the proportion that was daytime (15hr/24hr, 0.625) and nighttime (9hr/24hr, 0.375) resulting in 87.5 and 52.5 expected attacks for day and night, respectively.