The tested varieties were shown to support the mirid development, from fourth instar to adult, during seven days under all tested temperatures, except two cases at the lowest tested temperature, i.e. Delizia and Marinda at 15 °C. Data also suggest that the damage caused by the omnivorous mirid N. tenuis is positively correlated with increasing temperatures, with an optimal temperature comprised between 25 and 35 °C. 

Although the variety does not play a statistically signifcant role on the number of necrotic rings, the highest number of rings was observed on average on Rovente, while the lowest on Marinda at all temperatures, and as a general trend at 30 °C and 15 °C the highest and lowest damage for all varieties were, respectively, observed. The temperature data are partly in accordance with data by Sánchez (2008a). 

The damage reduction observed at 35 °C in some varieties could be due to the fact that this high temperature exceeds the optimum range of temperature that for this mirid was assessed between 20 and 30 °C (Sánchez et al. 2009). It is worth mentioning that the feeding behaviour of the nymphs at lower temperature was unexpected, as N. tenuis is considered being the most thermophiles species among the Dicyphinae in the Mediterranean area (Sánchez et al. 2009). 

Indeed, on the base of relevant scientifc reports, we estimated that at 15 °C the damage should not have occurred since this temperature approximates the lower nymphal development threshold assessed in previous studies [11.7 °C in Sánchez et al. (2009) and 12.9 °C in Hughes et al. (2009)].
Moreover, this temperature is considered detrimental to the mirid development, survival and reproduction (Sánchez et al. 2009). Finally, in previous studies it was observed that the feeding activity increased proportionally to the temperature (Sánchez 2008; Sánchez et al. 2009; Vacante and Tropea Garzia 1994).
This clearly suggests that the higher damage registered at higher temperatures in our study is related to an increased feeding activity rather than other potential factors. Despite its relevance for plant protection, few studies were carried out to assess the infuence of thermal regime on the feeding behaviour of N. tenuis and often irrespectively to varieties.
In Sánchez (2008), the damage intensity was correlated with prey availability (e.g. whitefies) and temperature but only at 20 and 30 °C. The impact of the mirid feeding activity on tomato yield was assessed in order to defne an economic threshold in greenhouse condition (Sánchez and Lacasa 2008), but only for one variety and in this case regardless of temperature regime. More recently, the infuence of tomato variety on the damage by N. tenuis was evaluated on fve commercial varieties and in greenhouse conditions within a wider feld trial aimed to set up various biological control strategies (Cabello et al. 2012).
The progression of damage induced for each variety was followed during the entire growing cycle, and only in partial agreement with our fndings, a signifcant diference among varieties was evidenced. In our work, tomato variety signifcantly afected the nymphal development while not the damage, although numerical diferences were observed. This discordance is to be related to the diferent experimental conditions (variable temperature and prolonged permanence of the mirid in the feld trial), to the prey availability, the potential intraguild interactions, and most likely also to the diferent methods of damage evaluation that, indeed, was not clearly referred in the text of Cabello et al. (2012). In the same study, N. tenuis progression number of nymphs and adults per leaf were recorded and a correlation between damage and nymphs was evidenced; data showed that the cluster fruit tomato type sufered more feeding damage compared to the others.

In the present study, we observed that at constant temperature, there is no apparent correlation between fruit commercial types, plant damage and nymphal development. Indeed, the highest number of rings was observed (although without any signifcant diference) in Rovente and Delizia, round and marmande types, respectively, while the highest number of adults was recorded in Cikito and Motekino, date plum and piccadilly types, respectively.
It was postulated that the diferences observed within diferent varieties could be correlated with plant nutrients of each variety, which may afect the feeding behaviour of the mirid (Cabello et al. 2012).
On the base of data acquired in our experiment, we can confrm that the nymphal performance can vary with the variety but without any clear correlation with fruit types. 

However, it is noteworthy that the highest damage level was observed on the two varieties Rovente and Faustyno both with round fruit type. Indeed, in these two varieties more severe ring induction was observed and consequently leaf or stem drying derived. As a consequence, research trials in feld conditions are thus needed to provide reliable indications and to draw a risk assessment of N. tenuis. This is particularly relevant because thermal regimes, characterized by wide daily variations, in Mediterranean cold greenhouses occur.
According to our data, N. tenuis is potentially able to induce plant tissue necrosis in protected crops throughout the year. Damage and development observed at 15 °C indicate that nymphs are capable to injure tomato plants even during cooler periods of the cultivation cycle. It is known that this mirid demonstrates a wide thermal adaptability (Sánchez et al. 2009), and its distribution area includes Africa, Asia,  Southern Europe and the Mediterranean basin.
Data collected in our experiment at 15 °C, where a low number of necrotic rings as well as moulted adults were observed, are not in accordance with previous laboratory studies carried out on the biology of the mirid, surprisingly showing much longer fourth and ffth nymphal development time at 15 °C (Hughes et al. 2009). However, the discordance could be explained by the diferent N. tenuis strains tested and the diferent feeding substrates adopted in the two experiments. Indeed, Hughes et al. (2009) employed a commercial strain that was reared on tobacco leaves and E. kuehniella eggs as factitious prey, while the strain used in our study is a strain locally collected in the feld reared on sesame plants (Biondi et al. 2016). Based on these considerations on the thermal range in protected crops and the data collected in our experiment, there are not enough elements to identify a variety suitable for each transplant period and/or fruit commercial type. Indeed, although some varieties have shown to be less susceptible at certain temperatures, this cannot eliminate the risk of necrotic rings induced by the mirid. 

The common occurrence of temperatures between 20 and 30 °C, corresponding to the thermal optimum for the mirid (Sánchez et al. 2009), increases the risk of damage, since all the varieties tested have shown similar susceptibility under these temperature conditions. However, although several studies reporting N. tenuis feeding activity on the plants also provide the costs in terms of yield reduction (Arnó et al. 2010), fower and fruit abortion (Sánchez and Lacasa 2008; Sánchez 2009), most of them do not make a thorough estimation of the result of omnivore-mediated plant protection (Puentes et al. 2018). 

Our work aimed to evaluate the infuence of tomato variety and temperature on the herbivory of N. tenuis. The initial hypothesis was that data on the susceptibility of diferent tomato varieties to the mirid feeding damage could be useful to draw a risk analysis across seasons, in order to manage tomato cultivations during the year, so to mitigate the feeding damage by the mirid either released or spontaneously present. To the best of our knowledge, the combined infuence of temperature and tomato variety under controlled conditions was evaluated in this study for the frst time.
The results show that N. tenuis damage and development are infuenced by tomato variety and temperature, although with varying intensities. These preliminary fndings are useful for the selection of suitable tomato varieties for the optimized N. tenuis exploitation in IPM programmes. In addition, predator risk assessment on crop varieties under diferent temperature regimes may support tomato industry in selecting the most compatible varieties for IPM tomato packages that include N. tenuis as a benefcial antagonist.
For example, the variety Shiren proved to sustain the mirid development at tested temperatures, and at high temperature, it had a signifcantly lower damage; the varieties Delizia and Marinda are not suitable varieties for N. tenuis at low temperatures despite having damage.
Moreover, the varieties Cikito and Pixel apparently are more susceptible at high temperature and then are not suitable for summer cultivation; at the same time, the varieties Faustino and Shiren should be adapted for summer crops. However, further laboratory and feld studies are needed to better entangle other aspects of the tomato variety suitability for this mirid. 

For example, new knowledge should be gained for detecting the potential diferences in volatile attractivity towards N. tenuis of the various tomato varieties both as healthy and as infested plants, that is crucial for host plant location by the mirid after artifcial releases or during natural colonization processes. Moreover, the potential damage found in the present study needs to be confrmed with the concomitant availability of prey.