This study investigates the impact of light intensity on the physical defenses of sugar maple (Acer saccharum) saplings and how these defenses affect interactions with insect herbivores. Conducted at the Kenauk Nature Reserve in Quebec, the study involved 12 sites representing both shaded and sunlit environments. Leaf traits, including thickness, specific leaf area (SLA), water content, and toughness, were measured from 216 leaves collected in 2021 and 40 in 2022. To explore herbivore interactions, two Erebid caterpillars including the invasive spongy moth (Lymantria dispar) and the native white-marked tussock moth (Orgyia leucostigma) were used in controlled bioassays that assessed caterpillar feeding preference and performance. The results revealed that sunlit leaves were significantly thicker and had lower SLA compared to shaded leaves across both years. While water content was lower in sunlit leaves in 2021, it showed no significant difference in 2022. Leaf toughness, measured only in 2022, did not significantly differ between sunlit and shaded leaves. Caterpillar feeding preference tests indicated that white-marked tussock moth caterpillars consumed more sunlit leaves, showing a significant preference for them. However, spongy moth caterpillars did not exhibit a significant preference. In terms of performance, white-marked tussock moth caterpillars had a higher survival rate and larger pupal mass on sunlit saplings in 2021, but no significant differences were observed in laboratory feeding initiation assays with first-instar caterpillars in 2022. These findings suggest that a tree's investment in sun leaves to maximise photosynthesis is not necessarily followed by increased defense against herbivory. This research contributes to our understanding of how environmental factors like light can shape plant defense strategies and impact insect herbivore dynamics in temperate forests.

Materials and Methods

Study site and Plant species

This study was conducted at the Kenauk Nature property, an old-growth natural forest spanning two regions from the Outaouais River valley to the Laurentian Mountains in southwestern Quebec, Canada (45°42'N; 74°53'W) in 2021 and 2022. In the first year, we selected 12 sites in this maple/bitternut hickory domain of temperate deciduous forest (Quebec Ministry of Forests, 2023) with each site over 10 meters apart. Tree species composition was relatively uniform across all sites and site selection was made based on the similarity in tree size (1–2 m height) and ease of access to the branches.

Each site comprised one healthy sugar maple sapling in the shade under almost full canopy cover and one healthy sugar maple sapling in the sun with less canopy cover. Temperature, humidity, and light intensity were measured 10 times at 4-day intervals at different times of the day from late May to early July, and 3 more times in August at 10-day intervals during the bioassays in 2021. Canopy openness was measured once at each site in late August of the same year. We ensured the average value of light intensity and canopy openness showed meaningful differences between shaded and sunlit trees per site, but not between sites.

For visual clarity and to present all experiments and their respective timings in one place, an infographic has been included as Figure 1.

Measurements of leaf physical traits

In 2021, to evaluate the effect of light intensity on leaf defensive traits, leaf traits were measured three times. Once right before installing cages for a spongy moth performance test in late May, once in early June before conducting a spongy moth preference test in the lab, and last time before the performance bioassay and preference test with tussock moth in late June. Considering the completion of leaf growth, which leads to expected stability in leaf traits once leaves are mature, no further measurements were taken after this date. On each sampling date, 3 undamaged average-sized leaves per sapling were collected per site (In total 216 leaves = (3 leaves* 2 saplings* 12 sites) * 3 sampling dates). Leaves were weighed and their thickness was measured using a digital caliper when they were fresh. Subsequently, the leaves were pressed, labeled, and dried in an oven for 72 hours. Dried leaves were then weighed again, and all leaves were scanned to measure the leaf area using ImageJ software. All efforts were made to ensure that dried leaves remained intact and free from any folding or breaking to maintain measurement accuracy. This allowed us to determine water content and specific leaf area (SLA) for all sampled leaves. In early June of 2022, we used 10 of the 12 sites from the previous year to collect 20 pairs of leaves (two sunlit and two shaded leaves per site) for leaf trait measurements. This year we added leaf toughness measurement using a digital penetrometer burrowed from Janice Cooke's laboratory at the University of Alberta, Canada.

Insect defoliators

To explore potential differences in leaf palatability and susceptibility to insect herbivores between sugar maple saplings in shade and sun, we focused on insect leaf cutters, which are primarily lepidopteran caterpillars in sugar maple forests (Thomas et al., 2010). As a representative of these lepidopterans, we selected two common leaf-eating caterpillars, both generalists from the Erebidae family which are also feasible to rear in the lab. One early-season species, spongy moth caterpillar (Lymantria dispar) is an invasive outbreaking pest (Boukouvala et al., 2022). The other late-season species named white marked tussock moth (Orgyia leucostigma) is native and non-outbreaking (Hebert et al., 2017). For the spongy moth, egg masses were collected from Angrignon Park in Montreal, QC, CA in late January 2021 and tussock moth egg masses were obtained in diapause from Insect Production and Quarantine Laboratories (IPQL) in Ontario, CA.  Both species were allowed to hatch and were reared in our lab before being placed in the field. All were fed on an artificial Bell diet until they finished the first instar and then were fed on non-experimental red oak (Quercus rubra) and Norway maple (Acer platanoides) leaves prior to the experiment.

Caterpillars’ feeding Preference

Laboratory study: We conducted a preference test using two caterpillar species in their later instars, on sugar maple sapling leaves from shaded and sunny branches in 2021. The preference experiment was performed on June 17 with 4th instar spongy moth caterpillars (early season caterpillar) and on August 12th with 4th instar tussock moth caterpillars (late season caterpillar). From all previously mentioned 12 sites, 25 undamaged sunlit sugar maple leaves and 25 shaded sugar maple leaves were randomly harvested and each stored in a 2.0 ml microcentrifuge tube with a hole punched into the lid and filled with water. Leaf petioles were fixed and sealed in tubes using parafilm to avoid water loss. Subsequently, leaves were sorted by size, scanned, and their area measured using Image J. One sunlit and one shaded sugar maple leaf of similar size were placed on opposite ends of each of the 25 plastic containers (19cm L*13cm W), with a similar leaf area in each container. A 4th instar caterpillar was then introduced to each container, positioned equidistantly between the two maple leaves to allow free choice. After 48 hours, the caterpillars were removed. Then we measured the consumed leaf area by rescanning the leaves and calculating the area that was gone which was used as the measure for host preference (Fortin et al., 1997; DiTommaso & Losey, 2003; Foss & Rieske, 2003; Morrow et al., 2022). This whole process was repeated later in the season using white-marked tussock moth caterpillars.

Caterpillars' performance bioassay

Field study: To assess the performance of the caterpillars on sugar maple leaves, we installed one sleeve cage on a sunlit sapling and one on a shaded sugar maple sapling in all 12 sites in 2021. Then, 10 caterpillars in their late 3rd instar were placed in each cage (10 caterpillars*12 sites = 120 caterpillars per treatment) and left there for approximately 3 weeks. The cages were checked regularly to collect pupae and count the remaining caterpillars until they all pupated or died. All pupae were weighed to compare the performance of caterpillars on shaded and sun branches. This performance test was conducted twice, once in the early season (cages and caterpillars set on May 30 and collected on June 30), with spongy moth and once in the late season with white marked tussock moth (cages and caterpillars set on July 30 and collected on August 31), to coordinate with their natural life cycles.

Laboratory study: In 2022 we conducted an experiment with early instar tussock moth larvae to determine if they would be more affected by the differences detected in leaf physical defenses between the two maple saplings considering that mortality in lepidopteran first instar larvae is reported to be highest compared to the later stages (Zalucki et al., 2002). We were not able to get spongy moth larvae in 2022. The same protocol as the previous year was followed for obtaining tussock moth egg masses and for collecting and storing sugar maple leaves from sunlit and shaded trees in early July in Kenauk. Egg masses were hatched and reared on Bell diet for a few days till their late first instar stage of life. For this bioassay, a total of 52 caterpillars at their first instar larval stage were moved to the plastic containers (19cm L*13cm W) and were supplied with fresh sunlit maple leaves (26 containers * 2 first instar caterpillars per container), and the same number of caterpillars were fed on shaded sugar maple leaves. Due to our focus on the evaluation of the ability to establish feeding of young 1st instar leaf cutter insect herbivores, caterpillars were observed after 48, and their survival rate was recorded at this time point (Murakami & Wada, 1997b; Zalucki et al., 2001; DiTommaso & Losey, 2003).

Statistical analysis

All statistical analyses were conducted using R and RStudio (version 2024.04.1+748). Visualizations were generated using the ggplot2 package in R (Wickham, 2016).

Leaf traits: To examine differences in leaf traits such as specific leaf area, thickness, toughness, and water content between sunlit and shaded saplings, the average value measured for three leaves per sapling was used. Having normally distributed count data, we employed a linear mixed model with sunny and shaded treatments as fixed factors and site as a random factor for 2021 and 2022 datasets in separate models using the lme4 package (Bates et al., 2015). All model assumptions were eventually evaluated by using the DHARMa package (Hartig, 2022).

Preference test: To assess the effect of light intensity on third instar larval preference after 48 hours, the experiment was designed with each caterpillar fed in a separate container, each containing one shaded and one sunlit sugar maple leaf, with the consumed leaf area as the response variable. Since the data from each light condition represent paired observations from the same caterpillars, a paired t-test was employed. This statistical test accounts for the natural variability between individual caterpillars and provides a precise evaluation of the effect of light intensity on leaf area consumption by comparing the consumption of the two leaf types by the same individual. Test assumptions, including normality and equal variance, were evaluated using the Shapiro-Wilk test and Levene's test for each caterpillar species separately.

Performance bioassays: In the 2021 cage experiment, the proportion of surviving caterpillars per cage was calculated as the survival rate. This percentage data with a beta distribution correction was fit in a specialized package of betareg model (Cribari-Neto & Zeileis, 2010). Since the DHARMa package does not work for models with beta distribution, we used a Q-Q residual plot for the evaluation of our model assumptions. Pupal mass data, as a normally distributed continuous response variable with cages having possible effect as a random factor, was analyzed using a linear mixed-effects model (LMM) with lmer from the lme4 package.  In 2022, having 2 caterpillars per container made the survival data a better fit as a binomial distribution to be analyzed with a glm with a binomial family.

Due to a high incidence of spongy moth mortality caused by Nucleopolyhedrovirus (NPV), a common viral disease affecting insect larvae, only 4 individuals survived. As a result, their performance data was excluded from the analysis.

Finally, the correlations between caterpillars’ preference-performance measures and leaf traits, relationships between leaf trait variables, as well as differences between sun and shaded saplings, were analysed using separate Principal component analysis (PCA) for each experiment. For 2021 performance data, the average of three leaf trait measurements on May 20, June 10, and June 24 along with survival rate and pupal mass were used for PCA. For the preference experiments, we used early May leaf trait data for spongy moth and late June leaf trait data for tussock moth PCA analysis. Since the number of average leaf traits measured was less than the umber of leaves used for the preference test, missing data for PCA were handled using multiple imputations (van Ginkel, 2023).  Performance from 2022 data was not included in any PCA due to its lack of significant difference between treatments.