We reared 81 individuals of Spodoptera littoralis (Lepidoptera: Noctuidae) on an artificial diet provided by the Max Planck Institute for Chemical Ecology (Jena, Germany). We maintained all larvae under identical laboratory conditions in September 2024 at 20–24°C and 40–50% relative humidity.

We weighed pupae at least 24 hours after pupation. Adult emergence occurred predominantly at night; we weighed adults the following morning, after complete wing expansion and meconium excretion, to obtain live adult weight from newly emerged, unfed individuals. We then kept adults without food until natural death. After death, we air-dried individuals for 10 days and weighed them again to obtain dry adult weight. Live and dry adult weights represent distinct physiological states of the adult stage, and do not constitute repeated measurements of the same condition. In none of our experiments we did apply oven-drying because it is not common practice in insect curation. Our aim was to quantify body mass changes under realistic curatorial and laboratory conditions, rather than to define an absolute dry mass standard. We recorded sex for all Spodoptera individuals. In females, we assessed egg presence by direct observation of eggs laid prior to death and classified females that did not lay eggs as lacking visible egg deposition.

We collected 71 moth individuals belonging to 13 families at the Naturschutzstation Dr. Helmut Steuer, Unteres Schwarzatal, Thuringia, Germany in July 2024 (50.668°N, 11.222°E (DD)). Six specimens could not be reliably identified to family level because of poor condition and are therefore reported as N/A.

We collected all specimens alive. After death, we stored specimens in a sealed container maintained at near-saturated humidity (~100%) to minimise post-mortem desiccation. We weighed specimens the following morning to obtain fresh dead weight. We then air-dry specimens for 10 days under laboratory conditions (20–24 °C, 40–50% relative humidity) and weighed them again to obtain dry weight.

We analysed 30 geometrid moth specimens collected in Peru in 2023 as part of the ANDIV project (Holzmann et al. 2025). Specimens had died shortly after capture, been allowed to dry, and were subsequently stored frozen at -18°C in glassine bags inside plastic containers. Specimens were transported to Germany under dry conditions and stored at the Phyletisches Museum (Jena, Germany), where they remained frozen at -18°C for several months prior to analysis.

We weighed specimens in September 2024 under laboratory conditions (20–24 °C, 40–50% relative humidity). We measured body weight at three preparation stages: (1) frozen weight, measured immediately after removal from frozen storage; (2) rehydrated weight, measured after rehydration; and (3) prepare dry weight, measured ten days after mounting and air drying.

We rehydrated specimens in a sealed plastic container at approximately 100% relative humidity for 24 hours following standard protocols (Walker and Crosby 1988; Krogmann and Holstein 2010), allowing specimens to become sufficiently flexible for handling and mounting. Rehydrated weight was recorded at this stage, as specimens are commonly handled and weighed during rehydration in museum workflows. Prior to mounting, we recorded the weight of the mounting pin and subtracted it from the final prepared (dry) weight.

We conducted all statistical analyses in R v4.4.0 (R Core Team 2023). To avoid pseudoreplication arising from repeated measurements of the same individuals, we analysed body weight using linear mixed-effects models, with individual identity included as a random intercept in all datasets.

We analysed body weight on the log-transformed scale, which improved normality and homoscedasticity of model residuals. For each dataset, we included the relevant weight state as a fixed effect:

life stage and adult condition (pupa, live adult, dry adult) in the laboratory dataset;

preservation condition (fresh dead adult vs dry dead adult) in the German field dataset; and

specimen preparation stage (frozen, rehydrated, prepared dry) in the Peruvian dataset.

Where applicable, we included additional fixed effects (sex, egg presence, or family).

We fitted models using restricted maximum likelihood (REML) and assessed the significance of fixed effects using Type III ANOVA with Satterthwaite or Kenward–Roger approximations for degrees of freedom. To quantify differences among levels of fixed effects, we estimated model-based marginal means and performed Tukey-adjusted pairwise comparisons. To estimate fresh or rehydrated body weight from dry specimens, we calculated correction factors using two complementary approaches across datasets. We first quantified the mean ratio between weights measured under different conditions and estimated uncertainty using non-parametric bootstrap confidence intervals. We then fitted regression-based calibration models relating log-transformed weights measured under different conditions to assess size dependence and quantify prediction uncertainty.

We produced all figures using ggplot2 (Wickham 2016), with individual-level data overlaid on boxplots to visualise variation across stages, conditions, and treatments. Analyses used the R packages lme4/lmerTest (Kuznetsova et al. 2017), emmeans (Lenth 2023), dplyr (Wickham et al. 2019), patchwork (Pedersen 2020), and reshape2 (Wickham 2007).

The dataset included 81 individuals, comprising 35 females (43%) and 46 males (57%). Body weight differed markedly among developmental and physiological stages (Fig. 1).

Pupal weight ranged from 0.190 to 0.419 g (mean=0.268 g, SD = 0.044). Live adult weight ranged from 0.071 to 0.247 g (mean = 0.137 g, SD = 0.033), while dry adult weight ranged from 0.022 to 0.109 g (mean = 0.041 g, SD = 0.015). Thus, individuals lost substantial weight during metamorphosis and experienced an additional pronounced reduction in body weight after death and drying (Table 1). Females consistently weighed more than males across all stages. Female pupae averaged 0.281 g, compared to 0.259 g in males. Live adult females averaged 0.154 g, whereas males averaged 0.124 g. After drying, females retained higher weight (0.047 g) than males (0.036 g). These descriptive patterns indicate strong effects of developmental stage and sex on body weight (Fig. 1; Table 1).

Mixed-effects models confirmed these patterns statistically (see Suppl. material 1: table SS1). Developmental stage had a strong effect on body weight (p < 0.001), and sex also significantly affected weight (p = 0.001), with a significant Stage × Sex interaction (p = 0.001; Table 2). In contrast, egg presence did not significantly influence body weight and did not interact with developmental stage (p > 0.5).

Estimated marginal means showed a consistent ranking of stages in both sexes, with pupae being heaviest, followed by live adults and then dry adults. All pairwise comparisons among stages within each sex were statistically significant (Tukey-adjusted p < 0.001; Suppl. material 1: table SS2).

Regression-based calibration further showed a strong positive relationship between preserved and live adult weight (p < 0.001). The slope of this relationship was significantly less than one, indicating size-dependent mass loss, with smaller individuals losing proportionally more mass during preservation. Sex and egg presence did not significantly affect this relationship (Table 3).

Fresh dead body weight ranged from 0.0078 to 0.421 g (mean = 0.085 g, SD = 0.092), while dry body weight ranged from 0.0026 to 0.183 g (mean = 0.033 g, SD = 0.038). Drying therefore resulted in a substantial reduction in body mass across all specimens (Fig. 2).

Mixed-effects models accounting for paired measurements within individuals confirmed a strong effect of measurement condition on body weight. Dry specimens weighed significantly less than fresh specimens (estimate = −1.01 on the log scale, p < 0.001; Table 4), corresponding to an average reduction of approximately two-thirds of fresh mass.

Estimated marginal means showed a clear separation between fresh and dry conditions across specimens (Table 5), with dry body weight consistently lower than fresh body weight.

Descriptive summaries showed substantial variation in absolute body weight among families in both fresh and dry conditions (Suppl. material 1: table SS3). These differences reflect variation in overall body size across taxonomic groups.

When family was included in the analysis, several family-level estimates differed from the reference level (Suppl. material 1: table SS4). However, the reduction in body weight from fresh to dry condition was consistent across families, and no family-specific drying effects were identified. Family-level results are therefore reported descriptively.

Body weight differed markedly among preparation stages (Fig. 3). Rehydrated specimens showed higher body weight than both frozen and prepared dry specimens, whereas prepared dry specimens consistently had the lowest weights.

Mixed-effects models confirmed a strong effect of preparation stage on body weight (Table 6). Estimated marginal means revealed a consistent ranking across specimens, with rehydrated weight highest, followed by frozen weight and prepared dry weight lowest (Suppl. material 1: table SS5). All pairwise differences among preparation stages were statistically significant.

Across all three datasets, dry specimens consistently underestimated body weight relative to live, fresh, or rehydrated conditions. Correction factors derived from paired measurements showed clear and reproducible patterns, with relatively low uncertainty.

In the laboratory dataset, live adult body weight was on average 3.53 times higher than dry adult mass (SD = 0.78), based on paired measurements within individuals. Bootstrap resampling produced a narrow 95% confidence interval of 3.36–3.70 (Table 7). Sex and egg presence had negligible effects on this relationship. A simplified correction equation can therefore be used to estimate live body weight from dry specimens. When uncertainty needs to be quantified, regression-based predictions with confidence intervals are recommended. The most parsimonious correction equation was: Live = exp(−0.457) × dry^0.464.

In the German field dataset, fresh dead body weight was on average 2.77 times higher than dry body mass (SD = 0.44; n = 71). Bootstrap resampling yielded a 95% confidence interval of 2.67–2.87 (Table 7), indicating consistent mass loss during drying. Regression analysis of log-transformed weights revealed a nearly proportional relationship between fresh and dry mass (slope = 0.95, intercept = 0.79, R² = 0.98, p < 0.001). The slope was slightly below one (β = 0.92), indicating weak size dependence. No significant differences among families were detected (F_10.47 = 1.23, p = 0.30). A single correction equation therefore describes the relationship across families: Fresh = exp(0.788) × dry^0.946.

In the Peruvian geometrid moth dataset, rehydrated body weight was on average 2.32 times higher than prepared dry mass (SD = 0.24; n = 30). Bootstrap resampling produced a 95% confidence interval of 2.23–2.40 (Table 7), indicating low uncertainty. Regression analysis showed a strong relationship between prepared dry and rehydrated weight (R² = 0.996, p < 0.001). The slope (β = 0.93) was close to one, again indicating weak size dependence. The corresponding correction equation was: Rehydrated = exp(0.537) × dry^0.934.

In Spodoptera littoralis, body mass declined by approximately half between the pupal and adult stages, reflecting water loss, meconium excretion, and shedding of the pupal case during metamorphosis (Jungreis et al. 1982; Molleman et al. 2011). Subsequent desiccation reduced adult mass by an additional ~70% from live to dry adult, resulting in an overall loss of ~85% from pupa to dry adult. These findings underscore the extent to which both developmental processes and specimen preparation influence body weight in Lepidoptera (Chown et al. 2011; Gilbert 2011; Bogusz et al. 2024). The pupal–adult weight relationship is treated as species-specific and exploratory and is restricted to the laboratory-reared S. littoralis dataset; it is not used for correction factors or extrapolation across species.

Sexual dimorphism in body mass, with females consistently heavier than males, aligns with patterns documented across multiple Lepidoptera lineages (Allen et al. 2011; Stillwell et al. 2014; De Camargo et al. 2015). Egg presence did not result in higher female weight, suggesting that S. littoralis allocates most reproductive resources gradually after emergence. This pattern is consistent with an income-breeding strategy, in which adult feeding and ongoing resource acquisition contribute substantially to egg production (Tammaru and Haukioja 1996; Javoiš et al. 2011; Teder 2014; Davis et al. 2016), in contrast to capital-breeding strategy of groups such as Saturniidae with non-feeding adults (Correa-Carmona et al. 2022).

Freshly collected moths from Germany lost, on average, approximately 63% of their fresh mass after air-drying. This magnitude of mass loss is consistent with previous reports of water loss following insect death (Noble-Nesbitt 1969; Benk et al. 2020). The resulting ratio, with dry weight representing about 37% of fresh weight, closely matches estimates reported for Lepidoptera by García-Barros (2025). Our regression analyses further showed that dry weight strongly predicted fresh weight, with only weak size dependence and no detectable family-level effects, indicating that a single correction equation adequately describes the drying effect across taxa included in this dataset.

These results highlight why uncorrected dry weights underestimate energetically and functionally relevant traits such as energy reserves, wing loading, and flight capacity (Rogers et al. 1976; Sample et al. 1993; García-Barros 2015).

In the Peruvian dataset, rehydration increased specimen weight by approximately 61% relative to prepared dry mass, yet rehydrated individuals remained substantially lighter than their estimated fresh condition. Rehydrated moths therefore represented only a partial recovery of pre-drying mass (Sanford et al. 2011; Thorat and Nath 2018). Regression analyses showed that prepared dry weight was an excellent predictor of rehydrated weight, with near-proportional scaling and minimal size dependence.

Incomplete recovery probably reflects limited water infiltration into internal tissues and irreversible structural changes caused by desiccation. During drying, intracellular water loss can collapse muscle fibres and damage lipid membranes, reducing porosity and altering capillary action (Abdel-Maksoud et al. 2011; Chown et al. 2011; Sheikh et al. 2017). Consequently, rehydration primarily affects outer cuticular layers and hemolymph spaces rather than deeply embedded tissues (Oliver et al. 2020; O’Donnell 2022). Frozen specimens may retain residual moisture, contributing to greater structural stability and more moderate mass loss compared to fully air-dry individuals (Neisskenwirth 2020).

We do not interpret rehydrated body weight as an estimate of live or fresh body mass. Instead, it represents a workflow-specific reference stage, as rehydration is routinely applied prior to mounting museum specimens. At this stage, specimens are less fragile than when fully dry, making weighing more practical while acknowledging its biological limitations.

Each dataset represents a common preservation workflow in Lepidoptera research, including laboratory rearing, fresh field collection, and museum specimen preparation (Walker and Crosby 1988; Krogmann and Holstein 2010; Neisskenwirth 2020). Analysed together, these datasets allow reconstruction of a realistic sequence of body-mass change from fresh or live adults through drying and subsequent rehydration (Sanford et al. 2011; Bogusz et al. 2024).

Across workflows, preservation consistently resulted in substantial mass loss (Fig. 4). Dry specimens retained approximately 30–40% of fresh or live body mass, whereas rehydrated specimens recovered only part of this loss, reaching approximately 60–65% of fresh mass.

Regression-based calibration showed that mass change during preservation followed predictable scaling relationships. In the laboratory dataset, regression slopes below one indicated size-dependent mass loss, with smaller individuals losing proportionally more mass than larger individuals. In contrast, slopes in the German and Peruvian datasets were close to one, indicating largely proportional mass change across specimen sizes. Despite this variation in size dependence, all workflows exhibited strong and systematic relationships between dry and fresh or live body mass. This suggests that preservation-induced water loss and partial water uptake during rehydration follow systematic and quantifiable patterns across taxa and preparation methods, consistent with previous studies on mass change in preserved insects (Sanford et al. 2011; Bogusz et al. 2024).

Fig. 4 integrates these results into a conceptual framework linking developmental transitions and preservation workflows. The diagram illustrates the average multiplicative correction factors derived from paired measurements, showing that fresh or live adult mass can be estimated from dry specimens using correction factors of approximately 2.8–3.5×, while rehydrated mass can be estimated from prepared dry specimens using a factor of approximately 2.3×. Together, these results demonstrate that preservation-induced mass loss and partial mass recovery during rehydration follow consistent and quantifiable patterns across Lepidoptera, supporting the use of standardized correction factors to approximate fresh or live body weight from preserved material.

Accurate estimates of body weight are central to insect ecology, as mass influences a broad range of physiological, behavioural, and life-history traits (Peters 1986; Kalinkat et al. 2015). Fresh weight best reflects the functional mass available for flight, metabolism, and reproduction, and is therefore essential for estimating traits such as wing loading, metabolic rate, and dispersal capacity (Beck and Kitching 2007; Boggs 2009). Body weight also integrates environmental and nutritional influences, including larval diet quality, adult feeding behaviour, and responses to temperature and humidity (Chown and Gaston 2010; Molleman et al. 2011; Padda and Stahlschmidt 2022).

Because fresh specimens are rarely available for large-scale surveys or museum-based studies, correction factors are essential to avoid systematic underestimation of ecologically relevant traits (Rogers et al. 1976; Sample et al. 1993; García-Barros 2025). Our results show that preservation can reduce body mass by more than half, highlighting the magnitude of bias introduced when dry weight is used uncorrected. Applying correction factors, reporting specimen condition, and documenting preservation workflow will substantially improve the comparability and reliability of trait-based analyses (Chown and Gaston 2010; Molleman et al. 2011; Gallagher et al. 2020; Keller et al. 2023).

A fully continuous dataset spanning all developmental and preservation stages is logistically unfeasible for most Lepidoptera. Our multi-dataset approach therefore represents a practical compromise that captures the principal sources of body weight variation using workflows that reflect common field and curatorial practices. Specimens were air-dry under laboratory conditions rather than oven-dry, which may allow some residual moisture and limits direct comparability with studies using highly standardized drying protocols at high temperatures.

The correction factors presented here should be considered as order-level approximations rather than species-specific physiological models. They are intended to improve the interpretation of preserved specimens and do not replace direct measurements of fresh weight.