Adult host‐plant preference and larval performance in an oligophagous insect (Chnootriba elaterii)

The present research investigates host preferences and larval development of the melon ladybird Chnootriba elaterii (Rossi) (Coleoptera: Coccinellidae), considering seven Cucurbitaceae plant species (Citrullus lanatus, Cucumis melo, Cucumis sativus, Cucurbita pepo, Cucurbita moschata, Lagenaria siceraria and Luffa aegyptiaca), to determine whether mother host preference is related with larval performance. The damaged area on plant leaves due to insect feeding in dual‐choice experiments was used to evaluate food preferences. Behavioural experiments in a Y‐tube olfactometer tested the role of olfaction in host‐plant selection. Parameters such as development duration and morphometric characters of emerged adults were evaluated to estimate larval performance. Adult females selected watermelon and melon as their favourite food while firmly refusing loofah and calabash. Insects preferred melon over watermelon, but the larval development on melon was slower; moreover, watermelon allowed larvae to develop faster than other tested plants, while loofah did not allow larval development. Larvae fed and developed on melon, pumpkin, zucchini, cucumber and calabash, without strong differences. These results only partially supported the preference‐performance hypothesis suggesting that it could be context dependent. Contrary to expectation, volatile organic compounds (VOCs) belonging to the host, only partially could guide adult females in the olfactometer. This result suggests that long‐distance olfactory stimuli alone cannot provide sufficient information for host‐plant selection in melon ladybirds. Considering the economic relevance of C. elaterii, the results of the present paper answer some basic questions about host‐plant selection for this pest insect, potentially useful to improve control strategies in crops.


INTRODUCTION
In herbivorous insects, the main behaviours affecting fitness, such as feeding, mating, and oviposition, are influenced by plant-related cues (Awmack & Leather, 2002;Carrasco et al., 2015;Städler et al., 2002;Wang et al., 2022).The success of an oligophagous herbivorous insect, in particular, depends on its ability to locate the best host plant, use its nutrients and cope with its defence (Bernays & Chapman, 1994;Fernandez & Hilker, 2007).The selection of host plants by insects is often performed by adults, particularly in species with low mobility larvae, whose performance depends significantly on the mother's host choice.In this context, the relationship between female preference Silvana Piersanti and Valerio Saitta contributed equally to this work.and offspring performance is a central ecological question (Jaenike, 1978;Jones & Agrawal, 2019;Nardi et al., 2023); at an interspecific level, this is a critical factor in crop pest infestation (Berdegué et al., 1998).The 'preference-performance hypothesis' (PPH), also known as 'mother knows best' theory, has been widely investigated in the ecology of phytophagous insects and states that females preferentially lay their eggs on plants providing the best larval development (Menacer et al., 2021).This theory has been confirmed at intra and interspecific levels in many species, mainly Lepidoptera (Jones & Agrawal, 2019), but there are also several cases, some of them in Chrysomelidae Coleoptera (Jones & Agrawal, 2019), where the positive correlation between adult host preference and offspring performance does not exist (Berdegué et al., 1998;Hufnagel et al., 2017;Mayhew, 2001;Menacer et al., 2021;Valladares & Lawton, 1991).In particular, in insects that seem to be such poor mothers, this behaviour can be genuinely little adaptive, for example in exotic species that did not have time to adapt, or due to errors in cues perception (Mayhew, 2001).Still, it could be due to alternative adaptive strategies, such as mothers that choose host plants based on their own diet, selecting food that ensures their best survival and fecundity (Mayhew, 2001;Scheirs et al., 2000).
In this regard, the present study aims: (i) to test the ability of female C. elaterii in selecting host cucurbit plants using their VOCs in a Y-tube olfactometer; (ii) to investigate the relationship between female food preferences, tested by dual-choice behavioural experiments, and larval performance, measured in terms of larval mortality, developmental time and morphometric parameters of the emerged adults.Cucurbits selected for the experiments belong to major (watermelon, melon, cucumber, zucchini, pumpkin and calabash) and minor (loofah) crops economically relevant in Europe.These species are representative of five genera widespread in the phylogenetic cladogram of Cucurbitaceae (Chomicki et al., 2019) and have been previously tested in C. elaterii EAG recordings (Piersanti et al., 2022).

Insects
Adults of C. elaterii were collected into the field in Perugia, (Italy) (43 07 0 41.7 00 N 12 21 0 43.3 00 E), on plants of Ecballium elaterium (L.) A. Rich (Cucurbitaceae) in September 2021.Insects were kept in a culture room under controlled conditions (25 ± 2 C, 45 ± 15% RH, photoperiod LD 16:8 using lamps 3350 lm; 4000 K) inside a white net cage (30 Â 30 Â 30 cm) (Type 80.301, Vermandel, Hulst, the Netherlands).Insects in different developmental stages (eggs, larvae and adults) were maintained in separate cages for a maximum of 30 individuals per cage.They were fed with plants of C. melo var.inodorus (the honeydew melon, a cultivar not used in the experiments) and no free water was provided.Plants used as feed were grown in pots (as described in the next paragraph).Six plants with five healthy fully expanded leaves were provided in each cage per day.The eggs were laid on the leaf abaxial surface, and the plants with the eggs were gently moved into a new cage.As the larvae emerged, they were carefully moved in a cage with plants; mature larvae typically pupated on the top of the larval cage and emerged adults were moved into adult cages daily.Adults of both sexes were maintained in the same cage with a sex ratio of 50% and mated females (2 weeks old) were used in the experiments.Insect sex was determined under stereomicroscope (Optika SFX-33, Ponteranica, Italy) (Dieke, 1947;Tomaszewska & Szawaryn, 2016).

Plants
Seven species of Cucurbitaceae belonging to five genera (Table 1) were grown for the experiments.The cucurbit species were selected because they are ubiquitous in Italy and represent major (watermelon, melon, cucumber, zucchini, pumpkin, calabash) and minor (loofah) cucurbit crops economically relevant (Chomicki et al., 2019).In addition, these selected species represent both annual (watermelon, melon, cucumber, zucchini, pumpkin, loofah) and perennial (calabash) crops with different phylogenetic relationships within the family Cucurbitaceae (Chomicki et al., 2019).The same species used in this study have also been used in a previous paper on the chemical ecology of C. elaterii, aiming to explore antennal olfactory responses to plant volatiles (Piersanti et al., 2022).
Commercial seeds (Rosi Sementi, Italy) were used to obtain plants.The seeds were sown into individual plastic pots (Bamaplast, Pistoia, Italy) (8 Â 8 Â 9 cm for watermelon, melon, cucumber, calabash and loofah, and 9 Â 9 Â 13 cm for zucchini and pumpkin, which have a larger size and more rapid growth) in sterile topsoil (Patzer Einheitserde, Manna Italia, Bolzano, Italy).Plants were grown for approximately 30 days in a controlled climate chamber (25 ± 2 C, 45 ± 15% RH, photoperiod LD 14:10) equipped with lights with a photosynthetic photon fluence rate of 200 μmol m À2 s À1 placed above the foliage.Water was supplied by sub-irrigation.

Y-tube olfactometer
The Y-tube olfactometer consisted of a polycarbonate plate (10 mm thick) (Lexan™), with a Y-shaped space milled into the centre (stem 185 mm long; arms 125 mm long at 60 angle; internal section 20 mm), sandwiched between two glass plates (10 mm thick).
A 7-mm-diameter hole drilled through the polycarbonate plate into the end of each arm allowed air tube connections, and a similar hole drilled into the end of the stem permitted the airflow outlet.
Compressed medical-grade air (79% nitrogen and 21% oxygen by

Dual-choice chambers
To test the existence of a feeding preference in C. elaterii females, insects were starved for 12 h in laboratory conditions and tested in dual-choice experiments with all possible combinations of two plant species among the seven Cucurbitaceae selected for the experiments (Table 1).In each replicate, a couple of leaves, one from each selected species, both not excised from the plant, were exposed to one mated female, caged on the surface of the two leaves using a clip cage constituting a dual-choice arena, similar to those illustrated by Piersanti et al. (2020).The clip cage consisted of a modified plastic Petri dish (9 cm diameter; 1.5 cm high) with a mesh-covered hole (6 cm diameter) and the rim covered by a small sponge ring to avoid damage to the leaf.The equal surface (around 1000 mm 2 ) of the two leaves was exposed to the test insect inside the clip cage.Choice tests were performed on plants with five fully expanded leaves.The insects were allowed to feed for 5 h, then the arena with the insects was removed and the leaves were photographed by a digital camera (Koppace 21MP camera 1080P, Shenzhen, China) fitted with a 12.5-75 mm/F 1.8 zoom lens.The damage due to feeding was measured from digital images using the open-source image-processing program ImageJ (Schneider et al., 2012).We considered as feeding damages brown areas or holes on the green leaves, due to the insect bites.Considering the typical leaf damage due to C. elaterii, it was always possible to define easily the outline of the leaf.Ten to fifteen replicates were performed for each combination of cucurbit species.All tests were made between 7:00 AM and 6:00 PM (CEST), and the bioassay room was maintained at 25 ± 1.6 C, 50%-60% RH.

Larval development
Newly hatched larvae were individually gently removed from the brood and located on an experimental host plant, where they were maintained till pupation.The plants belonged to the seven species selected for the experiments (Table 1).The single newly hatched larva was confined in a clip cage similar to those described in the dualchoice chambers bioassays.Plants used in the experiments had four to five fully expanded leaves, and the insect arena did not prevent plant growth.Larvae were checked daily; dead insects were removed and their numbers were recorded, as well as pupae.When the leaf inside the clip cage was exhausted, the insect was gently moved to a different leaf (or position) on the same plant, taking a picture of the

Statistical analysis
In the Y-tube olfactometer tests, the time spent by C. elaterii in the two arms of the olfactometer was analysed by linear mixed models (LMM) with the plant species as fixed effect and insect nested within each plant pair as random effects to account for pseudoreplication (females tested towards the same plant pair).The significance of the fixed term in the model was determined using likelihood ratio tests (LRTs) comparing the model with and without the factor in question (Crawley, 2007).Because this approach compares models with different fixed-effect structures, maximum likelihood (ML) was specified in the models instead of restricted maximum likelihood (REML) (Crawley, 2007).Model fit was assessed with residual plots.The statistical analyses were carried out using the package lme4 (Bates et al., 2015) of R software, version 4.2.2 (R Core Team, 2020).In the dual-choice chambers tests, to analyse the difference in the leaf area damaged by the insects, the Wilcoxon signed-rank test was used because data were not normally distributed.The larval mortality recorded on the different plant species was compared using the Pearson Chi-square test followed by Goodman's post hoc procedure (Marascuilo & Serlin, 1988).
The two-way analysis of variances, considering the plant species and the sexes as factors, was used to compare data regarding the development time, the length of elytra and width of the pronotum of the adults, and the consumed leaf area in the larval development experiments.The Tukey unequal N HSD (honestly significant difference) post hoc test was used for post hoc comparisons.Plant species were compared for males and females when the sex factor was statistically different and the interaction effect between plant and sexes was not significant.When the difference between sexes was not different only plant species were compared (Statsoft, 2001).In the larval development experiments, plants that did not allow to complete the development were not included in the analysis regarding the development duration and the adult morphometric parameters.Before the analysis, the data normality was tested with Shapiro-Wilk test and the homogeneity of variances with Cochran C test.Data regarding post embryonic development duration and leaf damage produced by melon ladybird larvae were subjected to Box-Cox transformations to reduce heteroscedasticity (Sokal & Rohlf, 1998).

Dual-choice chambers
Considering all the 21 possible combinations of the 7 Cucurbitaceae used for the experiments (Figure 2 and Table 2), loofah was always significantly refused by the insects (in detail, it was never eaten, except one replicate with pumpkin and one with cucumber).Calabash elicited a very low feeding activity and was quite often significantly refused in all combinations.Insects did not eat either loofah or calabash in bioassays comparing the two plant species.Watermelon was preferred (larger surface of leaf damage) when compared with loofah, calabash, pumpkin and cucumber.Melon was preferred to loofah, calabash and watermelon.Pumpkin, zucchini and cucumber were chosen only when offered with loofah or calabash.In the remaining eight combinations, there was no significant choice by the insects; in fact, the consumed leaf areas were not significantly different.
F I G U R E 1 Time (s) spent by melon ladybirds in each arm of the Y-tube olfactometer during the bioassays.Strip chart with superimposed mean and 95% confidence intervals, each plant species is represented with a specific colour.Statistical parameters for each plant combination are reported in Table 2. Asterisk (*) indicates significant difference at p < 0.05 (linear mixed model).Each colour indicates one species of Cucurbitaceae.

Larval development
The mortality (%) among larvae fed with the different plan species was significantly different (χ 2 = 15.68;d.f.= 6; p = 0.0155), indeed no larvae completed development on loofah and all the insects tested died in a few days (Figure 3).In the other cucurbits, only 20%-30% of larvae died during the experiments, without significant differences between the plant species (χ 2 = 15.68;d.f.= 6; p = 0.0155) (Figure 3).= 5, 56; p = 0.6289) of adults obtained from larvae developed on the six cucurbit species were not different and no differences were also found for these parameters between the two sexes and in the interaction of the two factors.Loofah was not included because it did not allow to obtain any adult (Figure 5).The leaf damage was different among the plant species (F = 49.7;d.f.= 6, 77; p < 0.0001) but not between the two sexes (Sex: F = 0; d.f.= 1, 77; p = 0.9648) and their interaction (Plant species Â sex: F = 0.7; d.f.= 6, 77; p = 0.6599) (Figure 6).Larvae never ate loofah, then the leaf damage was zero.In the remaining cucurbits, the leaf damaged area was reduced in the watermelon and melon leaves compared with zucchini.The other cucurbits showed an intermediate leaf damage (Figure 6).

DISCUSSION
Olfaction in phytophagous insects is one of the major drivers used to detect the host-plant in the environment, both in long (Bernays & Chapman, 1994;Carrasco et al., 2015) and in short distance (Vaishampayan et al., 1975).
Notwithstanding many authors demonstrating the crucial role of olfaction in phytophagous Coleoptera (Jackson et al., 2005;Lewis et al., 1990;Rondoni et al., 2017;Stenberg & Ericson, 2007) (preferred host in dual-choice experiments) only when proposed in comparison with pumpkin; (iii) preferred calabash to pumpkin, notwithstanding the first one was largely refused in the dual-choice feeding tests.
These results were not expected given previous EAG investigations, demonstrating that antennae of C. elaterii can perceive a wide range of volatiles emitted by cucurbits (E2-exenal, heptanal, nonanal, decanal, methyl salicylate, (-) -linalool, Z3-hexen-1-ol, pentanol, hexanol, nonanol and octanol), and plant extracts from watermelon, melon, pumpkin, calabash, cucumber and zucchini but not loofah (Piersanti et al., 2022).We cannot exclude that olfactory cues may be important in host searching by melon ladybug, but probably in combination with other stimuli, as reported in many insects orders (Fernandez & Hilker, 2007;Harris & Miller, 1988;Prokopy & Owens, 1978;Saxena & Goyal, 1978;Vaishampayan et al., 1975).Similar results were found by Stenberg and Ericson (2007) also in Altica engstroemi (Sahlberg) (Coleoptera: Chrysomelidae), a monophagous chrysomelid that was guided to its host plant by visual and olfactory cues, but never responded to plant VOCs in Y-tube olfactometer.Very little is known about how olfactory cues interact with other stimuli in the central nervous system of herbivorous insects; still, undoubtedly, multimodal signals provide a degree of redundancy that can be advantageous in complex environments (Carrasco et al., 2015).In addition, unlike monophagous insects that can rely on specific volatiles to locate their best host plant, females of oligophagous insects may have high phenotypic plasticity and use multiple cues integration because there can be no common infochemicals indicative of plant quality in a range of host plants (Carrasco et al., 2015).Furthermore, the search for new habitats, unlike in other insects (Ferracini et al., 2023;Toepfer et al., 2006), is probably not a recurrent activity for C. elaterii whose adults are weak flyers with a limited dispersal ability (Liotta, 1964).
Based on the present results, loofah cannot be considered as host plant for C. elaterii; indeed, adult females do not feed on its leaves, and larvae do not develop on them.This is in agreement with EAG investigation (Piersanti et al., 2022), where loofah extract did not elicit any antennal response of adults.This plant is widely distributed in Africa and Asia, and fruits, traditionally used as sponges (Chomicki et al., 2019), are recently considered a promising sustainable and renewable biomass for new materials development (Hlel et al., 2021;Paglicawan et al., 2005;Su et al., 2018).The unpalatability of loofah could be due to some chemical defence, as suggested by the absence of EAG responses to loofah extract (Piersanti et al., 2022), which aligns with previous preliminary observations (El-Abdin & Siragelnour, 1991); otherwise, C. elaterii does not have any problem of adhesion to the loofah leaves (Saitta et al., 2022).For this reason, loofah could be used as an intercrop in cucurbit cultivations, developing interesting biocontrol strategies against the melon ladybug, similar to the protection of cabbage crops from Pieris rapae (L.) (Lepidoptera: Pieridae) using tomato, fennel and catnip (Wu et al., 2010).
Despite the uncertain role of olfaction in discriminating between cucurbits, female ladybirds show clear food preferences: watermelon is the most selected food in dual-choice chambers, where it was preferred in all the proposed combination, except those with zucchini and melon.Melon, at the same time, was preferred to watermelon, while calabash and loofah were firmly refused in all the combinations, and loofah was sampled by only 2 insects out of 70.Adult females regularly feed zucchini, pumpkin and cucumber but did not prefer them to other cucurbits, except loofah and calabash.These results agree with previous reports (Akandeh & Shishehbor, 2011;Awadalla et al., 2011;Wool & van Emden, 1981), where the watermelon was the preferred food, just after the snake cucumber (C.melo var.flexuosus).Our experiments did not test snake cucumber but the conspecific melon (C.melo var.cantalupensis), because it is a major European crop of a different variety.On this account, it is unsurprising that varieties of the same species result in different palatability and nutritional quality due to domestication and selection (Bayhan & Bayhyan, 2017;Chomicki et al., 2019).For example, in laboratory experiments with watermelon, the crimson tide cultivar was preferred by C. elaterii and guaranteed the shortest development time (Bayhan & Bayhyan, 2017).
This correlation is tighter for specialist insects than for generalist (Jones & Agrawal, 2019) and it is the main basic assumption of the preference-performance hypothesis (Menacer et al., 2021).This assumption is not always true (Carrasco et al., 2015;Gripenberg et al., 2010;Mayhew, 2001;Pappers et al., 2002;Thompson, 1988), as reported in Spodoptera exigua (Hübner) (Lepidoptera: Noctuidae), whose females prefer Chenopodium murale L. (Caryophyllales: Amaranthaceae) while larvae show better performance on Apium graveolens L. (Apiales: Apiaceae) (Berdegué et al., 1998).In our experiments, no relevant differences were noticed in larval mortality and morphometric parameters of the emerged adults comparing the host plants, except loofah where larvae never developed.Considering the developmental duration, larvae grew faster on watermelon, compared with other tested plants.Shorter development duration guarantee an earlier emersion of adults, then, probably, a reproductive advantage and a larger capacity for population increase (Nylin, 2001).For this reason, this parameter is considered a useful trait to characterise larval performance (Berdegué et al., 1998;Menacer et al., 2021).
Watermelon is also one of the favourite foods for adult females that, similar to other herbivore species, could perform host selection F I G U R E 5 Morphometric parameters of the emerged adults grown on the different cucurbit species used as host plants for developmental experiments: (a) length of elytra (mm) and (b) width of the pronotum (mm).Strip chart with superimposed mean and 95% confidence intervals.Among plant species, no significant differences are reported at p < 0.05 (two-way analysis of variances).
considering their food preferences (Mayhew, 2001).The adult preference for melon over watermelon, not related to better larval performance, could allow a direct advantage for the female, in terms of best survival or fecundity (Mayhew, 2001), all aspects that need to be better explored.Still, melon ladybird larvae regularly developed also on calabash, whose leaves were firmly refused by adults.This divergence may be due to plant mechanical defences.Indeed, the adhesion of insects to plant surfaces was largely studied (Gorb & Gorb, 2009;Prüm et al., 2012;Rebora et al., 2020aRebora et al., , 2020b;;Salerno et al., 2018Salerno et al., , 2020Salerno et al., , 2022;;Voigt et al., 2008) and a recent paper highlighted the difficult adhesion of adult C. elaterii on calabash, because of a dense layer of short and flexible trichomes creating a velvet surface (Saitta et al., 2022).
Larvae caged in our developmental experiments, unlike adults, could not move away from velvet leaves, and then regularly feed on them.
Insects are able to regulate nutrient intake when they are in the presence of imbalanced food sources; for example, studies carried out on grasshoppers show that they are capable of applying a compensatory feeding among poor and rich food sources (Berner et al., 2005).
Likewise, lower feeding damages performed by larvae on watermelon could be due to the higher nutritional quality of this plant; this agrees with adult preference in the dual-choice chamber and larval performance in terms of development duration.

CONCLUSIONS
In summary, food preference by melon ladybird females agrees with larval performance, considering the favourite crop of this pest, watermelon, and the non-host cucurbit, loofah, not allowing larval development.The lack of concordance between preference and performance concerning calabash, rejected by adults but able to guarantee normal larval development, is probably explained by the mechanical defences of the plant.The adult preference for melon over watermelon, not related to a better larval performance, can be considered maladaptive or the expression of an alternative adaptive strategy.It also suggests that the preference-performance of this insect is complex and needs further studies to be clarified.In addition, results obtained in Y-tube olfactometer show that olfaction alone probably is not sufficient to drive C. elaterii in host-finding process, or at least that this ability is context-dependent, based on the host combination.
To our knowledge, there are no data relating adult food preferences with larval performance in Epilachnini, or investigating the role of olfaction in host-plant selection from a behavioural perspective in melon ladybird.Elucidating the role of chemical cues in insect-plant interaction is crucial to developing effective and successful integrated pest management strategies, such as the push and pull crop and intercrop strategies (Khan et al., 2008), or the use of trap crops in collard (Piersanti et al., 2020).For example, our results suggest that loofah could be a promising 'push' species, to protect cucurbits crops from C. elaterii.This perspective could be interesting because several insecticides are recommended to control oligophagous pests; still, most of them are unsuitable or with low efficacy, owing to the emergence of resistant strains and toxicity to human and non-target organisms (Abdel-Moniem et al., 2004).In this context, the present results could significantly enhance the control of C. elaterii in organic cucurbit crops and introduce further studies to deepen the knowledge of the complex preference performance of these insects.
volume), regulated by flow meters (Matheson) and humidified by bubbling through bi-distilled water (Drechsel bottle 125 mL), flowed through both arms, creating an air stream of 70 mL min À1 per arm.The olfactometer arena was surrounded by a cardboard screen to minimise disturbance and was illuminated by four cool white fluorescent tubes (À3350 lm; 4000 K) above the device.Temperature in the olfactometer arena was 25 ± 1.6 C. Before entering the olfactometer arms, each air stream, controlled by flow meters, passed through a cylindrical glass chamber (Ø = 12 cm; h = 52 cm; 5.8 L) (Stereoglass S.r.l., San Martino in Campo, Italy) containing the plant responsible for odour sources.The plants used in the experiments were healthy, with five fully expanded intact leaves.The plant pot was wrapped with aluminium foil to minimise the odour from the soil.For each replicate, two plants of different species were located in the two glass chambers to produce alternative olfactory stimuli in the two arms of the olfactometer.For each bioassay, a mated female of C. elaterii, starved for 12 h in laboratory conditions, was introduced into the olfactometer at the entrance of the central stem.Insects were used only once to avoid any learning bias.The insect position (stem, right arm or left arm) was recorded for 5 min using EthoWatcher ® (Florian opolis, Brazil).The total time spent in each arm (mean ± SE) expressed in seconds was measured and analysed.Insects that remained in the stem for 5 min were not considered in the data analysis.Around 30 insects were tested for each pair of plant species (21 combination in total); for every 10 replicates, odour sources (plants) were replaced with new ones, and the glass chambers were washed with (in sequence) 10 mL of hexane (Sigma-Aldrich, St. Louis, MO), 10 mL of acetone (Carlo Erba, Cornaredo, Italy), 50 mL of distilled water and baked overnight (120 C).The olfactometer was cleaned and the odour sources were changed to alternate arms (to prevent any directional biases) every five replicates.The glass plates were cleaned with hexane, acetone and distilled water.In contrast, the polycarbonate part of the olfactometer was cleaned with a laboratory detergent and rinsed with hot tap water (approximately T A B L E 1 Cucurbitaceae species used in the experiments (correspondence between scientific name, variety and common name).
90 C, for 5 min) and distilled water.After washing, the polycarbonate plate was inverted.The odour sources were randomly assigned during bioassays, and the possible presence of bias was tested by running controls with both air streams passing through empty glass chambers (n = 18).All tests were performed between 7:00 AM and 3:00 PM (CEST), and the bioassay room was maintained at 25 ± 1.6 C, 50%-60% RH.
feeding damage.The total consumed leaf area for each insect, till pupation, was measured from the sum of the consumed areas in the daily digital images, using the open-source image-processing program ImageJ(Schneider et al., 2012).The pupated insects were moved individually in a plastic jar (30 mm diameter and 50 mm height) till emergence.When the adult emerged, as soon as the cuticle was hardened (with a darker colour), sex and biometric parameters (length of elytra and width of the pronotum) were recorded.The duration of postembryonic development (days) was calculated (ranging between 24 and 53 days).The individual development of 15-25 larvae for each host-plant species was observed.Experiments were run under controlled conditions between May and September 2021 (25 ± 2 C, 45 ± 15% RH, photoperiod LD 14:10).
No significant difference in the time spent by the ladybirds (t = 0.40; d.f.= 17; p = 0.6944) in each of the two arms in the control tests was recorded.Considering all the 21 possible combinations of the seven Cucurbitaceae used for the experiments (Figure1 and Table 2), only five cases showed a significant difference in the time spent by the ladybirds in the two arms.In detail, insects spent more time in the arm with volatiles emitted by calabash than in the arm with volatiles emitted by pumpkin, in the arm with volatiles emitted by watermelon than in the arm with volatiles emitted by loofah, in the arm with volatiles emitted by watermelon than in the arm with volatiles emitted by pumpkin, in the arm with volatiles emitted by melon than in the arm with volatiles emitted by pumpkin, and in the arm with volatiles emitted by cucumber than in the arm with volatiles emitted by zucchini.

F
I G U R E 4 After embryonic development duration (days), from neo hatched larva to the adult, on the cucurbit species used as host plants for developmental experiments.Strip chart with superimposed mean and 95% confidence intervals.Strips with different letters are significantly different at p < 0.05 (two-way analysis of variance, Tukey unequal N honestly significant difference); upper case letters are used for female and lower case letters are used for males.

F
I G U R E 6 Consumed leaf area (mm 2 ) produced by melon ladybird larvae feeding on the cucurbits proposed as host plants during the developmental experiments.Strip chart with superimposed mean and 95% confidence intervals.Strips with different letters are significantly different at p < 0.05 (two-way analysis of variances, Tukey unequal N honestly significant difference).Larvae never eat loofah.data curation; writingreview and editing.Manuela Rebora: Data curation; conceptualization; methodology; writingreview and editing.Gianandrea Salerno: Data curation; conceptualization; methodology;