Development and thermal activity thresholds of European mirid predatory bugs

A R T I C L E I N F O

Dicyphine species are true omnivores, and prefer hairy plants because of morphological and behavioural adjustments (Voigt et al., 2007;Ingegno et al., 2008Ingegno et al., , 2011Ingegno et al., , 2016. The presence of insect carrion entrapped by glandular trichomes might also attract them to some of these plants, as observed for other predatory bugs (Krimmel and Pearse, 2013). During spring, they move from wild host plants used as winter refuges to agricultural crops (Gabarra et al., 2004;Balzan and Moonen, 2014;Balzan, 2017). Their crop colonizing ability and their predatory efficacy are affected by several factors, such as the presence and abundance of host plants surrounding the crop, the suitability of the prey and host plant as food source, and the climatic conditions.
In Europe, many endemic dicyphine species can be considered as potential biological control agents (BCAs) (Messelink et al., 2015;Ingegno et al., 2017aIngegno et al., , 2019Biondi et al., 2018). The species richness of this tribe may offer opportunities to select the most suitable candidates for biological pest control in specific crops or at certain climatic conditions. Some greenhouse crops for example in northern Europe are grown during winter at low temperatures, which can reduce the activity or survival of some species of natural enemies (Hatherly et al., 2005(Hatherly et al., , 2008Hughes et al., 2009). The same situation occurs in the Mediterranean area where extreme hot temperatures during summer can hamper the performance of natural enemies (Thomson et al., 2010;Aguilar-Fenollosa and Jacas, 2014). Moreover, the rate and the duration of insect development strongly depend upon the temperature to which the insects are exposed Lykouressis, 2000, 2002;Sengonca, 2001, 2003;Sanchez et al., 2009). The temperature below which no measurable development occurs is the threshold of development (Campbell et al., 1974). In studies of thermal tolerance, particular behaviours are linked to physiological limits to temperature. Although the critical thermal maximum (CT max ) may occur at different temperatures in relation to the species, the behavioural response (i.e. the onset of spasms) is the same across a diversity of taxa, representing an excellent index and standard for evaluating the thermal requirements and physiology of an organism.
To determine thermal tolerance, two major methods are commonly used: (1) the static method, which measures the time to death at constant test temperatures, and (2) the dynamic method, which increases or decreases test temperatures until an end point is reached (Lutterschmidt and Hutchison, 1997). However, within the range of temperatures in which insects can survive there is a more restricted zone in which insects are capable of normal locomotory function (Mellanby, 1939). In temperate regions, where conditions are rarely severe enough to cause direct mortality, these movement thresholds can be an indication for activity at lower or higher temperatures, which is very important for selecting natural enemies for various climatic conditions (Hughes et al., 2010). Traits often examined include non-lethal measurements such as chill coma temperature (Mellanby, 1939;Colhoun, 1960;Gaston and Chown, 1999), chill coma recovery (Ayrinhac et al., 2004;Macdonald et al., 2004;Castañeda et al., 2005) and motility across a range of temperatures (Allen, 2010;Hughes et al., 2010). A method for rapid non-lethal dynamic measurement of thermal tolerance in small insects has been developed by using a temperaturecontrolled block in which multiple individuals can be observed without direct handling or disturbance with camera recording for retrospective analysis (Hazell et al., 2008). The thermal thresholds for activity observed with this technique can be a good indication for heat or cold tolerance.
In this study, we determined the developmental time at different temperatures (static method) and the thermal thresholds for motility (dynamic method) of the following six species: Dicyphus bolivari Lindberg [formerly described as Dicyphus maroccanus Wagner (Sanchez and Cassis, 2018)], Dicyphus eckerleini Wagner, Dicyphus errans (Wolff), Dicyphus flavoviridis Tamanini, M. pygmaeus and N. tenuis. These species showed potential as BCAs for pest control in vegetable crops and some are already commercialized for augmentative biological control. Therefore, information on their thermal requirements will be useful to select species in relation to climatic conditions.

Insect collection and rearing
Colonies of D. eckerleini were started from individuals collected in Hungary on Geranium spp. This species was reared on small tobacco plants Nicotiana tabacum L. in transparent plastic cylinder cages (H 27 cm, Ø 25 cm; JET 107 PM, Jokey plastic GmbH, Sohland, Germany) with the drilled lid covered with a fine net mesh. Colonies of D. errans and D. flavoviridis originated from individuals collected in the Alpine valleys in Piedmont (NW Italy) on various herbaceous plants. These two latter species were reared on Solanum nigrum L. and tobacco plants, respectively, in net cages (60 × 60 × 90 cm; Kweekkooi, Vermandel, The Netherlands). Two D. bolivari strains were collected in two Spanish regions, namely in Valencia area (V) on tomato and in Barcelona area (B) on Solanum luteum Mill. Macrolophus pygmaeus and N. tenuis were provided by Koppert B.V. (Berkel en Rodenrijs, The Netherlands) from populations originally coming from South France and Spain, respectively. Colonies of D. bolivari, M. pygmaeus and N. tenuis were reared on pods of the flat bean Phaseolus vulgaris L. in transparent plastic cylinder cages (H 27 cm, Ø 25 cm; JET 107 PM, Jokey plastic GmbH, Sohland, Germany) with the drilled lid covered with a fine net mesh. All dicyphine rearings were supplied with eggs of Ephestia kuehniella Zeller (Lepidoptera: Pyralidae) mixed with dehydrated and decapsuled cysts of Artemia sp. (Anostraca: Artemiidae) (Entofood, Koppert B.V. Berkel en Rodenrijs, The Netherlands) and kept at 25 °C, 70% RH, 16:8 h L:D.

Development at different temperatures (static method)
The development times at six constant temperatures (15, 20, 25, 30, 35, 40 °C) were assessed for all the tested species (D. eckerleini, D. errans, D. flavoviridis, two D. bolivari strains, M. pygmaeus and N. tenuis) in apposite climate chambers (70% RH, 16:8 h L:D). A newly emerged nymph ( < 1 day old) was placed in a single plastic medicine cup (H 40 mm, Ø min 25 mm, Ø max 30 mm), closed with a drilled cap covered with a fine mesh. In each container, a gerbera leaf disk (Ø 25 mm) was provided with E. kuehniella eggs on a 15 mm high Gypsum plaster layer wetted with water to maintain humidity. Nymphs were checked every day removing the exuviae. Leaf disks, provided with food, were replaced every 2 days. Emerged adults were sexed, weighed with an analytical balance (Mettler Toledo AE100) and length was measured by using a biocular microscope. Fifteen replicates were performed for each species/strain and temperature combination.

Tolerance at different temperatures (dynamic method)
For recording the thermal thresholds for activity for both low and Table 1 Duration ( Ingegno, et al. Biological Control 152 (2021) 104423 high temperatures, we used the method developed by Hazell et al. (2008). Temperature was controlled with a refrigerating/heating circulator (F25-ME, Julabo, USA Inc.) by pumping heated or cooled fluid through channels bored into a thermostated aluminium block (130 × 60 × 20 mm). Insects were placed in a circular arena (Ø 25 mm, 75 mm high) in the middle of this block. The whole block was covered with a layer of double glass and placed in a block of extruded polystyrene foam for insulation (Styrodur®, BASF). Insect behaviour in the arena was recorded through the glass with a digital video-camera (Leica DFC425, GmbH, Wetzlar, Germany) connected to a stereomicroscope. Temperatures at which a visible physiological response (such as paralysis, spasm, trembles, etc.) was evident were recorded as key temperatures (KTs). For each species/strain, 10 males and 10 females divided in groups of five individuals were tested.

Chill tolerance
To check the chill tolerance, the temperature was reduced from the rearing temperature (25 °C) to 10 °C at an initial decreasing rate of 1.0 °C min −1 , and it was further lowered from 10 °C to -8°C at a rate of 0.20 °C min −1 . Temperature was then increased back to 10 °C at a rate of 0.25 °C min −1 and to 25 °C at a rate of 1.0 °C min −1 . The total experimental running time was about 3 h, and pictures were taken every minute, resulting in a total of 180 pictures for each run. A fan positioned on the upper surface of the cover provided a continuous air flow across it to prevent condensation below 0 °C.

Heat tolerance
To check the heat tolerance, the temperature was increased from laboratory temperature (20 °C) until 30 °C at an initial increasing rate of 0.5 °C min −1 , and it was further increased from 30 °C to 48 °C at a rate of 0.25 °C min −1 . Temperature was then decreased back to 20 °C at a rate of 1.0 °C min −1 . The total experimental running time was about 2 h, and pictures were taken every minute, resulting in an total of 120 pictures for each run.

Statistical analyses
Analyses of the developmental time was only based on temperatures that allowed at least one individual per species/strain to reach adulthood. The survival rate in each developmental stage was referred to the initial number of individuals used in the trial, and differences in survivorship were analysed by a General Linear Model (GLM) with a binomial distribution and logit link function. Temperature was considered as a fixed factor in order to avoid a linear or defined relationship between temperature and insect development. To compare the duration of developmental time within developmental stages among species and temperatures, a two-way analysis of variance (ANOVA) was performed, since data were normally distributed. Length and weight of adults emerged in the developmental trials were compared by perfoming multivariate ANOVA with temperature, sex and species as fixed factors and considering all the interactions among them. Then, within each species body weight and length were analysed by performing a one-way ANOVA and differences among developmental temperatures were separated with Tukey's tests (P < 0.05).
Differences in temperature tolerance, both for chill and heat, among species were investigated with Generalized Linear Mixed Models (GLMMs) with a normal distribution and an identity link function, and an unstructured covariance matrix. Species and sex were analysed as fixed effects, KT as a repeated measure. Sex × species, sex × KT, species × KT interactions were also analysed. Models including 100.0 ± 0 100.0 ± 0 100.0 ± 0 100.0 ± 0 87.5 ± 8.26 B.L. Ingegno, et al. Biological Control 152 (2021) 104423 different predictors were compared by sequentially subtracting the predictors from the full model and, subsequently, by calculating the Akaike Information Criterion (AIC) (Akaike, 1978) and selecting the best model. Differences among treatments were considered at a 5% level of confidence, using pairwise Bonferroni protected t-tests. All statistical analyses were performed with the statistical package SPSS 24 (IBM Corp., Armonk, NY, USA).

Development at different temperatures (static method)
None of the tested species/strain was able to develop until adulthood at 40 °C (Table 1). At 35 °C, only strain B of D. bolivari and N. tenuis reached adulthood. Dicyphus eckerleini reached the 4th juvenile instar, but all other species could stopped the development at the 2nd or 3rd instar (Table 1).
Among temperatures that allowed at least one individual per species to reach adulthood, the survival rate did not differ in the first four instars, but it was different in the 5th instar (df = 3; χ 2 = 12.615; P = 0.006) as well as the interaction between species × temperature (df = 18; χ 2 = 34.181; P = 0.012). In particular, at 30 °C, both D. bolivari strains showed a higher mortality in the last juvenile instar compared to the other species and temperatures (Table 2).

Tolerance at different temperatures (dynamic method)
Mirids responded to changes in temperature with an increase or decrease in walking speed. Then, when a critical temperature was reached, the insect lost locomotory function (critical thermal temperature). A further temperature increase or decrease made the insect completely immobile, i.e. the temperature at which there was the last subtle movements (coma temperature). By reversing the temperature trend from low temperatures, a recover of subtle motility (coma recovery) and of locomotory function (activity recovery) could occur.

Chill tolerance
From video recording, four KTs were recorded to evaluate the cold tolerance for each individual: KT1-stop walking (critical thermal minimum, CT min ), KT2-stop moving (chill coma level), KT3-start moving (chill coma recovery), KT4-start walking (activity recovery).
A GLMM with all fixed factors included gave the best maximum likelihood (AIC = 2,528.041), and significant differences among all the fixed effects (df = 38, 590; F = 162.808; P < 0.001).

Heat tolerance
From video recording, two KTs were recorded to evaluate the heat tolerance for each individual: KT5-stop walking (CT max ), KT6-stop moving (hot coma level); no recovery was possible since all the individuals died. A GLMM with all fixed factors included gave the best maximum likelihood (AIC = 2,528.041), and significant differences among all the fixed effects (df = 22, 278; F = 20,050.774; P < 0.001).

Discussion
Previous studies on insect thermal biology focused mainly on direct lethal effects of low temperatures and on overwintering strategies. However, temperature can influence other species-specific thermal thresholds, such as development, walking or flying skills reducing or preventing movement. Climatic conditions close to these thresholds can indirectly increase mortality, because of the inability of species to find food resources or to escape from predation or parasitism (Renault et al., 2002;Turnock and Fields, 2005;Pitts and Wall, 2006;Hatherly et al., 2008). In fact, as the temperature decreased or increased until a limit threshold (i.e. KT1, KT5), we observed that the individuals lost the ability to move their limbs in a coordinated manner and were therefore unable to walk. Since in all the trials, no recovery from the hot coma was possible, this threshold, unlike chill coma, is more likely an irreversible state leading to mortality (Hazell et al., 2010;Hughes et al., 2010). Consequently, the study of sublethal thermal tolerance traits can help to determine the potential efficacy of a BCA across a range of temperatures (Hughes et al., 2010).
In temperate zones, Honek (1996) reported a significant trend for decreasing average lower temperature threshold with increasing geographical latitude based on literature data of 335 insect species. Therefore, the presence of different dicyphine species on tomato crops, depending on geographic zone, can probably be explained by a different tolerance to lower and higher temperatures. The knowledge of these thermal limits can help in the selection of the most suitable BCAs in relation to climatic conditions. Based on our results, we can separate  two main groups of species, one more adapted to a temperate climate (D. errans, D. eckerleini and D. flavoviridis) and the other more adapted to a warmer climate (D. bolivari, M. pygmaeus and N. tenuis). Our data shows a clear relationship between the minimum and maximum thermal thresholds. The species less vulnerable to lower temperatures were more vulnerable to higher temperatures and the opposite; species more vulnerable to lower temperatures were less vulnerable to high temperatures. These findings confirm the general theory that each species of insect can only develop over a limited range of temperatures, the so-called thermal window (Dixon et al., 2009).
In all treatments, D. eckerleini and D. errans were the only species having the CT min (stop walking) at temperatures below 0 °C. Both species showed similar developmental spans and were able to tolerate lower temperature compared to the other species, suggesting that they can be considered as suitable BCAs in cropping areas with cooler temperatures. This is consistent with the latitudinal distribution of D. errans, a species that is widespread throughout Europe (except in Finland, Latvia, Lithuania, northern Ireland, northern Russia and Slovakia) and western Asia (Kerzhner, 1999). Similarly, D. eckerleini is spread in southeastern Europe and in the Middle East (Kerzhner, 1999) and also tolerates low temperatures, maybe because of more extreme winter conditions in eastern Europe than western Europe. Also D. flavoviridis showed a good cold resistance (KT1 approx 1 °C), but at the same time a long developmental time at 15 °C (about 2 months). This is probably related to its bigger body size compared to the other tested species. This species is reported in Italy, France and Switzerland, and is mainly spread in wet piedmont environments (Kerzhner, 1999;Ingegno et al., 2008;Matocq and Streito, 2013). Moreover, females are brachypteran so they are not able to cover long distance in their life span (Tamanini, 1949). This characteristic could be usefully exploited for a spot release to prevent natural escaping in the crop surroundings especially in new areas. Establishment of non-native generalist species outside the glasshouse may have undesirable effects on native populations, either by competition or predation (Hatherly et al., 2005;van Lenteren et al., 2006). However, this risk should not be overestimated because the establishment of most mirids strongly depends on the presence of suitable host plants in the agroecosystem. At the moment, little is known about the actual predatory effectiveness of these more cold tolerant species, since predation trials were performed only for D. errans so far (Ingegno et al., 2013(Ingegno et al., , 2017bIngegno et al, 2017a). Therefore, further research is needed to know the potential of D. eckerleini and D. flavoviridis for their possible use as BCAs in temperate areas.
Our experiments showed that the three species D. bolivari, M. pygmaeus and N. tenuis were able to tolerate higher temperatures than the other three tested species, which makes them suitable candidates as BCAs in crops under warmer temperatures. Native to the Palaeartic ecozone, M. pygmaeus is distributed from Finland to Algeria, the Azores Islands and Tajikistan (Sanchez et al., 2012), and is abundant in the Mediterranean region depending on geographic area, season and crop cycle (Hamdi and Bonato, 2014). Also N. tenuis is a common species in Mediterranean region but it has its northern limit in North Italy; moreover, it is spread also in Africa, Central America, South Asia and Oceania, and it is reported in many crops mainly from localities with a mild climate (Sanchez et al., 2009). Recently, this species also established unintentionally in many Dutch tomato greenhouses (GM, personal observation). Dicyphus bolivari has been reported in Morocco and in Spain where it was found preying on T. absoluta (Abbas et al., 2014). Despite its distribution, Pérez-Hedo and  suggested that the optimal temperature range of D. bolivari is below that of the commercialized species M. pygmaeus and N. tenuis. This is in contrast with our data that showed no significant differences in the CT max for both D. bolivari strains and in the CT min only for D. bolivari V compared to the other two species mentioned above. We also observed that D. bolivari B and N. tenuis were the only species able to reach adulthood at the constant temperature of 35 °C. However, the mortality at 30 °C was higher for both D. bolivari strains than the other species and temperatures, which might limit their performance under hot climatic conditions.
Both D. bolivari strains showed similar duration in their nymphal development, which was for most temperatures longer compared to the other tested species, except for D. flavoviridis. Our data on development times at 25 °C are generally consistent with what obtained in previous studies (Perdikis and Lykouressis, 2000;Urbaneja et al., 2005;Sanchez et al., 2009;Ingegno et al., 2011Ingegno et al., , 2017bIngegno et al., , 2019Abbas et al., 2014;Mollá et al., 2014), with exception of D. flavoviridis and D. eckerleini, which were here investigated for the first time. A longer juvenile developmental time may increase the total prey consumption per individual. Indeed, Mollá et al. (2014) reported three-fold predation values on the total number of E. kuehniella eggs preyed by nymphal instars of D. bolivari compared to those consumed by N. tenuis and M. pygmaeus. However, it remains to be clarified whether this trait is positive, because a longer generation time also results in a slower numerical response, which may be detrimental for biological control in the short term (Abrams and Matsuda, 1996;Symondson et al., 2002). So far, only a few studies have been performed on D. bolivari biology and behaviour (Abbas et al., 2014;Pérez-Hedo and Urbaneja, 2014;Ingegno et al., 2019), although more information might actually be available in the literature because of a possible misidentification with the species Dicyphus tamaninii Wagner (J.A. Sanchez, personal communications).
In the chill tolerance experiment, N. tenuis showed a significantly higher activity recovery temperature (about 13 °C) compared to the other species. This result could explain its low chill tolerance reported by other authors who observed its inhability to survive through winter in cold climates (Hughes et al., 2009). By using the same refrigerating/ heating circulator model to investigate the thermal activity threshold of N. tenuis, Hughes et al. (2010) registered different KTs. The main differences were in KT2 and KT4, which in our experiment were lower and higher, respectively (i.e. ΔKT2: −4°C; ΔKT4: +3°C), while the heat KTs (i.e. KT5 and KT6) were similar (Table 3). Beside the origin of the mirid populations and their previous acclimatation, this could be due to a more stressful biologic consequence caused by heat. In fact, often the hot coma temperature and the upper lethal limit do not differ significantly (Hazell et al., 2010;Hughes et al., 2010). The chill coma temperature registered by Hughes et al. (2010) for N. tenuis was higher than all the KT2 registered in our experiment, in which M. pygmaeus showed the highest KT2. Despite the different trend curves applied in the two studies, this record suggests a minor cold resitance for M. pygmaeus compared to the other tested species. Nevertheless, in NW Italy during the cold season, nymphs can be found overwintering with adults (Ingegno et al., 2009), underlying the establishment potential of these mirid predators. Moreover, it has to be considered that experimental conditions were extremely forced while, in enviromental conditions, temperature fluctuations are less extreme concerning intensity and duration, allowing a more likely successful establishment.
Our results showed interestingly that in all species females and males differed in their cold and heat tolerance. Males and females were more vulnerable to cold and to heat, respectively. This difference in thermal tolerance might be explained by different size between males and females. The effect of sex on insect cold hardiness has been often ignored but can play an important role in the supercooling ability of an organism (Renault et al., 2002). Carrillo et al. (2004) found no difference in the supercooling points between males and females of Diamesa mendotae Muttkowski (Diptera: Chironomidae). On the contrary, highly significant differences between sexes in the supercooling point temperature distributions were found in Alphitobius diaperinus (Panzer) (Coleoptera: Tenebrionidae): males were characterized by lower supercooling point temperatures than females (Salin et al., 2000). Although the differences in body size and in fat content or metabolism [i.e. females have different nutritional needs than males (Stockhoff, 1993;Mira, 2000;Telang et al., 2001;Moreau et al., 2003;Colinet et al., 2007)] are evident, the exact mechanisms behind thermal requirements remain unclear. Whether body size can be a good indicator for thermal tolerance might be interesting for further studies. Larger sized and smaller sized species in this study seemed to be better adapted to lower and higher temperatures, respectively. Moreover, manipulation of rearing temperatures can be worth to explore since it can influence both phenotypic characteristics, such as size and efficiency of BCAs as reported for generalist parasitoids of the genus Aphidius (Hymenoptera: Braconidae) (Ismail et al., 2014;Jerbi-Elayed et al., 2015a, 2015b. We think that our study can provide useful basic information to predict which mirid predatory species can be suitable in relation to environmental conditions, especially in glasshouses where thermal settings are known and controlled during the crop season. Besides thermal thresholds for activity, the choice of the best mirid predator should obviously include also other criteria such as predatory efficacy, developmental time, prey and plant preference. In fact, host plants may influence the behaviour of both prey and predator, and their use as companion plants or in crop consociation can play a fundamental role in pest control as reported for D. errans (Ingegno et al., 2017c). Some dicyphines are stenophagous or even plant specific, while others are polyphagous, and some can cause damage by feeding on plants (i.e. N. tenuis), producing necrotic rings in vegetative parts and flower abortion depending on the geographical region (Sanchez, 2008;Arnó et al., 2010;Castañé et al., 2011;Biondi et al., 2016;Naselli et al., 2016). In conclusion, the thermal threshold studies for activity of arthropod natural enemies are a useful tool to get a rapid impression of their suitability as BCAs for certain climatic conditions in addition to population dynamic experiments.