Phenology and interspecific association of Forficula auricularia and Forficula pubescens in apple orchards

The European earwig Forficula auricularia L. (Dermaptera: Forficulidae) has been widely studied as a key predator of pests in temperate regions, but its phenology and behavior may differ in warmer areas such as the Mediterranean. Here we assessed the phenology, aggregation, and interspecific association of F. auricularia and Forficula pubescens Gené, the only two species found consistently in both ground and canopy shelters in Mediterranean apple orchards. In addition to F. auricularia and F. pubescens, three other earwig species, namely Labidura riparia Pallas, Nala lividipes Dufour and Euborellia moesta Gené, were found occasionally. The mature stages of F. auricularia were observed mainly from May to November in tree shelters and immature ones from October to June in ground shelters. Adult individuals of F. pubescens were observed year-round and nymph instars were detected from April to June in ground as well as in tree shelters. The suitability of the current degree-days models for temperate regions was evaluated for the prediction of European earwig phenology in a Mediterranean climate. Regarding interspecific association, F. auricularia and F. pubescens co-occurred in canopies without apparent competition. This study provides useful weekly data about the phenology of the two earwig species throughout the year that can be used to detect the key periods during which to enhance their populations in pip fruit orchards or to control them in stone fruit crops. Furthermore, our results are of relevance for the development of new phenological models of earwigs in Mediterranean areas where nymphs hibernate, a feature that makes current models inaccurate. Additional key words: biological control; Dermaptera; earwig; Forficulidae; Mediterranean; pest Abbreviations used: BB (Les Borges Blanques); DD (degree-days); FA (Forficula auricularia); FP (Forficula pubescens); IU (Ivars d’Urgell); MI (Miralcamp); MO (Mollerussa); N1 to N5 (1st to 5th nymph instars); RAA (Rosy apple aphid); WAA (Woolly apple aphid) Citation: Lordan, J.; Alegre, S.; Moerkens, R.; Sarasúa, M. J.; Alins, G. (2015). Phenology and interspecific association of Forficula auricularia and Forficula pubescens in apple orchards. Spanish Journal of Agricultural Research, Volume 13, Issue 1, e10-003, 12 pages. http://dx.doi.org/10.5424/sjar/2015131-6814. Received: 09 Sept 2014. Accepted: 12 Feb 2015 http://dx.doi.org/10.5424/sjar/2015131-6814 This work has one supplementary table that does not appear in the printed article but that accompanies the paper online Copyright © 2015 INIA. This is an open access article distributed under the Creative Commons Attribution License (CC by 3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Funding: This study was funded by the Spanish project Programa Nacional de Investigación y Desarrollo Agrario no AGL201017486 (AGR) Control integrado de plagas en frutales de pepita y hueso. Competing interests: The authors have declared that no competing interests exist. Correspondence should be addressed to Jaume Lordan: jaume.lordan@irta.cat; jaumelordan@gmail.com

as chlorpyrifos, deltamethrin, indoxacarb and spinosad, have been reported to have lethal effects on the European earwig (Peusens & Gobin, 2008;Peusens et al., 2010;Vogt et al., 2010;Fountain et al., 2013).Software applications and prediction models have been developed to optimize orchard management techniques geared to promoting the European earwig (Helsen et al., 1998;Moerkens et al., 2011;Belien et al., 2012Belien et al., , 2013)).However, these studies have been conducted in colder regions, and earwig phenology and behavior may differ in warmer areas such as the Mediterranean.
Here we assessed the phenology, aggregation, and interspecific association of F. auricularia and F. pubescens in Mediterranean orchards, with the aim to promote their populations in crops where they served as biocontrol agents but also to optimize their control in crops where they are pests.We evaluated the suitability of the current degree-days models for temperate regions to predict the phenology of the European earwig in a Mediterranean climate.
BB was an IRTA (Institute of Research and Technology, Food and Agriculture) experimental orchard of 'Fuji Kiku 8' apple grafted onto M9, planted in 2003, and trained to a central leader with 4 × 1.4 m spacing.MO was a commercial orchard of 'Golden Smoothee' apple grafted onto M9, planted in 1985, and trained to a double-axis system with 4 × 1.2 m spacing.IU was a commercial orchard of 'Golden Smoothee' apple grafted onto M9, planted in 1993, and trained to a central leader with 4 × 1.1 m spacing.MI was a commercial orchard of 'Golden Smoothee' apple grafted onto M9, planted in 2000, and trained to a central leader with 4 × 1.2 m spacing.
BB was sampled for 4 years (2010)(2011)(2012)(2013), MO andIU for 3 (2011-2013), and MI for 2 (2012)(2013).For each orchard from 2010 onwards, 10 shelters were set on the second scaffold limb of various trees (tree shelters).From 2012 onwards, 10 additional shelters were tied at the base of 10 supplementary trees in each or-However, due to their omnivorous diet, European earwigs can cause economic damage to stone fruit crops (Albouy & Caussanel, 1990;Kuthe, 1996;Grafton-Cardwell et al., 2003;Huth et al., 2011).In nectarines, the action threshold is considered when any trap in the orchard contains five or more earwigs (Hetherington, 2006), while in cherries, no predictive relationship between the number of earwigs in traps and the level of damage has been found (Allen, 2013).In addition, the frass produced by earwigs can negatively influence the aroma and flavor of some wines (Burdet et al., 2013).To control earwigs in conventional fruit production, growers spray orchards with commonly used pesticides such as chlorpyrifos and spinosad (Hetherington, 2006;Peusens & Gobin, 2008;Vogt et al., 2010;Fountain et al., 2013).In organic fruit production, alternative strategies, such as mass trapping and exclusion by setting glue around the base of tree trunks, are used (Hetherington, 2006;Alston & Tebeau, 2011;Saladini et al., 2012).
The European earwig forages at night and seeks shelter during the day (Albouy & Caussanel, 1990;Helsen et al., 1998).Given that these insects are important biocontrol agents, the use of additional shelters to enhance their populations has been assessed in apple, pear, and kiwifruit orchards (Solomon et al., 1999;Gobin et al., 2006;Logan et al., 2011).As earwigs have a univoltine life cycle, any disruption in their cycle in one year can have long-lasting repercussions on populations (Gobin et al., 2006;Peusens & Gobin, 2008;Peusens et al., 2010).To minimize negative effects on vulnerable life stages of earwigs, the prediction of their phenology will contribute to determining the precise timing for spray applications and soil tillage, thereby improving orchard management (Peusens et al., 2010;Belien et al., 2012Belien et al., , 2013;;Moerkens et al., 2012).For instance, common pesticides sprayed in orchards, such sine wave approximation (Rabbinge, 1976), using a lower threshold of 6°C and taking 1 st January as the biofix.These parameters were chosen following Helsen et al. (1998).

Data analysis
Data from April to July-when more earwigs were recorded-were used to compare among years within orchards.Replicates were the weekly mean number of earwigs in the 10 tree shelters.F. auricularia data were log-transformed and ANOVA assumptions (normality and homoscedasticity) were confirmed before analysis.Means were compared at the p = 0.05 level, and a Tukey HSD test was used to separate means.Due to heterogeneity of variance, F. pubescens data were analyzed by Welch's test.
To compare earwig species, data from April to July in tree shelters were used.Replicates were the weekly mean number of earwigs in the 10 tree shelters, and in this case they were compared within orchards by Welch's test.
Data from June and July-when more adults were recorded in tree shelters-were used to calculate and analyze the sex ratio for F. auricularia and F. pubescens within orchards.Data were log-transformed and analyzed by a nonparametric Wilcoxon test.Homogeneity of variance was also confirmed before each analysis.
Aggregation in shelters was evaluated by fitting data to Taylor's power law (Taylor, 1961): where S 2 is the variance, m is the sample mean, a is a sampling factor, and b indicates whether the population distribution is regular (b<1), random (b=1) or aggregated (b>1).For F. auricularia, the weekly mean data of the 10 shelters from June to July from all the years and orchards were used, while for F. pubescens the data used were from IU 2011-2012and MI 2012. Equation [1] was log-log transformed to estimate a and b.
To evaluate the interspecific association between F. chard (ground shelters) in an attempt to capture earwigs in younger stages that do not climb on trees.Following Lordan et al. (2014a), shelters were prepared by rolling a piece of corrugated cardboard into cylinders (12 cm height × 9 cm diameter), which were protected from rain and adverse conditions by a PVC tube (15 cm height × 9.5 cm diameter).Similar shelters have been used in studies of European earwigs elsewhere (Phillips, 1981;Helsen et al., 1998;Solomon et al., 1999;Burnip et al., 2002;Gobin et al., 2006;Logan et al., 2007;He et al., 2008;Moerkens et al., 2009).Every week throughout the year, the species, number, phenological stage, and sex of adult earwigs for each shelter were recorded, and earwigs were then released at the base of the assessed tree.Presence of wings was used to distinguish between F. auricularia and F. pubescens adults (Albouy & Caussanel, 1990).Cerci dimorphism was used to distinguish sex while size and number of antennal segments and the apparent wing buds on the 3 rd segment of the thorax were used to distinguish nymph stages (Albouy & Caussanel, 1990).

Evaluation of the degree-days models
The European earwig phenological degree-days model (Model) designed by Moerkens et al. (2011) was tested in our region.The daily minimum and maximum temperatures required to run the model were obtained from the closest automatic weather station of the Meteorological Service of Catalonia (Meteocat, Departament de Territori i Sostenibilitat, Generalitat de Catalunya).For BB, data were from the Castelldans station 8.5 km away, for IU from the Castellnou de Seana station 3 km away and for MO and MI from the Mollerussa station 0.5 km and 1 km away respectively.From 2011 onwards, daily soil temperatures at a depth of 5 cm were also available but only from the Mollerussa station, which is 12 km from BB and 10 km from IU; these distances were within the range used to construct the model described by Moerkens et al. (2011).Thus, the model was run with soil temperature data from MO for all the orchards.The model was checked for 2012-2013 based on the dates of first appearance and peak of each developmental stage observed in the field.
The sum of degree-days (DD) up to the first and maximum number of N3, N4 nymph instars, and adults was calculated for each orchard and year and compared with those reported by Helsen et al. (1998).The minimum and maximum temperatures from each weather station were used to calculate the effective temperature for each orchard and year.The effective temperature sum in DDs was calculated through the Cas varies from -1 to +1.A negative value shows competition, zero no interaction, and a positive value an association between species (Legendre & Legendre, 1984;Sauphanor & Sureau, 1993).
Data were analyzed using the JMP statistical software package (Version 9; SAS Institute Inc., Cary, NC, USA).

Phenology
F. auricularia was very common in all the orchards during the study period, whereas F. pubescens, although observed in all the orchards, was not captured all the years (Suppl.Table S1 [pdf online] and Fig. 1).Both species were found in tree and ground shelters (Suppl.Table S1).Higher numbers of F. auricularia than F. pubescens were observed in all the orchards (Figure 2).The abundance of F. auricularia did not change along the years in BB, IU or MI, whereas the population increased in MO over the years (F2,48=19.75;p=0.0001) (Figure 1).The abundance of F. pubescens decreased in IU (F2,20=35.44;p (F1,19=9.49;p=0.006) (Figure 1).
In addition to F. auricularia and F. pubescens, three other earwig species, namely Labidura riparia Pallas, Nala lividipes Dufour and Euborellia moesta Gené, were found occasionally but only in ground shelters.
F. auricularia was found throughout the year (Figure 3a-b and Suppl.Table S1).From January to June, N2, N3 and N4 instars were found in ground shelters.At the end of January the population peaked with an average of 3 N3 instar individuals (Figure 3b).The presence of the N4 instar rose from mid-March to the end of May, after which no more N4 instars were observed in ground shelters (Figure 3b).The presence of the N2 instar was intermittent during winter and early spring and more regular from May to June; however, the population peak was observed in December, with an average close to 3 individuals per ground shelter (Figure 3b).Adults were found in ground shelters from May to November, but their abundance was lower than that of nymphs (Figure 3b).In fact, adults were more abundant in the tree shelters than in the ground shelters (Figure 3a).In tree shelters, they were captured from April to November, but a greater presence of adults was observed from mid-May to the beginning of July, with a peak of 23 individuals per shelter (Figure 3a).N4 was the most abundant instar in tree shelters from the end of March to mid-May, with a population peak of 14 individuals per shelter in mid-May (Figure 3a).The N3 instar was also observed in tree shelters one month after the N4 instar was found.The abundance of the N3 instar was much lower, with an average of 3 individuals per tree shelter (Figure 3a).
Regarding F. pubescens, adults were found in ground shelters mainly from mid-January to April, and after that N2, N3, N4 and N5 instars were successively observed either in ground or tree shelters until July (Figure 3c-d).The N2 instar of F. pubescens was more common in ground shelters, while it was barely observed in tree shelters.In contrast, the N1 instar was not found in tree or ground shelters (Figure 3c-d).Adults of F. pubescens were observed from March to April and from June to December in canopies, with a maximum of 2 individuals per shelter (Figure 3c).
Captures dropped for both earwig species during molting into adults (Figure 3).In both species, the sex ratio was not significantly different from 1:1 (p>0.05,Wilcoxon test).

Aggregation behavior and interspecific association
The relationship between the variance and the mean was studied by Taylor's law.The distribution of F. auricularia in shelters was observed to be aggregated, as the b coefficient was higher than 1 in all the orchards (Table 1).On the other hand, for F. pubescens, the b coefficient was higher than 1 in IU, also indicating an tected in tree or ground shelters when running Moerkens' model (2011) (Table 2).Regarding Helsen's model (1998), the N3 instar was observed to appear at 215 DD; however, large differences between orchards were found (Table 3).Although smaller differences were observed for the N4 instar (264 DD) and adult stage (250 DD), there were no matches between observed and estimated dates (Table 3).We found some coincidences only when predicting the maximum number of N4 (613 DD) and adult individuals (1035 DD), with a aggregated distribution.In contrast, in MI this distribution could not be confirmed (Table 1).F. auricularia and F. pubescens showed mainly a positive association (Fig. 4).A few negative values were observed (Fig. 4).

Evaluation of the degree-days models
No matches among observed and estimated dates were found for any of the developmental stages de- in orchards according to Welch's test (p<0.05).
Earwigs per shelter range from to 29 days between observed and estimated date (Table 3).

Discussion
In our study the average number of F. auricularia was higher than that of F. pubescens, whereas in citrus orchards the opposite was observed (Romeu-Dalmau   In 2012, heavy hail seriously damaged the whole crop, thus destroying some of the natural shelters, such as fruit clusters.These circumstances increased the likelihood of earwigs using the artificial shelters.The following year, as consequence of the hail, the trees showed more vegetative growth but less crop load, thus also reducing the number of natural shelters.Moerkens et al. (2009) reported an increase in the number of adults in the shelters immediately after the harvest of pears, thereby pointing to the relevance of natural shelters.Regarding the variations observed for F. pubescens, although significant differences were observed, the variation was less than one individual per trap.
of our knowledge, been reported previously.The N1 instar was never observed.This could be attributed to the fact that this stage is very short and the nymphs probably remained in the nest with the female (Albouy & Caussanel, 1990).We found the N2 instar mainly in ground shelters from April to mid-May.After this time, the successive instars were also detected in tree and ground shelters.We found nymph instars only from April to July, thus indicating a single reproductive period per year.Similar observations were made by Romeu-Dalmau et al. (2011).
For both earwig species, after the peak numbers of last nymph instars (N4 or N5), we observed a population decline during molting into adults.Moerkens et al. (2009) proposed that this decrease was caused by competition for limited resources, such as hiding places and food, when the population increases, but also by an increase in cannibalism and intraguild predation, as insects are highly vulnerable during molting.
The distribution of F. auricularia in field shelters was clearly aggregated, coinciding with the findings of Sauphanor & Sureau (1993) in laboratory trials.In contrast, the distribution of F. pubescens in these shelters was not aggregated, although these authors found the opposite behavior in laboratory conditions.These differences may be due to the fact that F. pubescens was not abundant in field shelters, thus the opportunity to aggregate was lower than in lab trials, where more individuals per shelter were present.This observation is consistent with those of Taylor et al. (1978), who reported that in most species the degree of aggregation is density dependent.In addition, under laboratory conditions, the only shelters available were artificial.In contrast, the field provides a variety of alternative natural shelters, which may reduce the chance of detecting aggregation.
Both F. auricularia and F. pubescens were found together in only two orchards and with a low population of F. pubescens.On the basis of these data, we propose that there is only a tendency of F. auricularia and F. pubescens to associate or at least not to compete.The few negative values that we observed appeared only in months when the insects were barely found in the shelters.Sauphanor & Sureau (1993) observed a positive association, estimating a coefficient value of 0.75.High association values were observed when more earwigs were found in the shelters, thus resembling the conditions tested by Sauphanor & Sureau (1993) in lab trials.Even in the field, Debras et al. (2007) reported the absence of competition between F. auricularia and F. pubescens.We can assume that when these two earwig species are found in high numbers in the shelters, no competition occurs.This may be linked to high availability of food or to the different diet preferences of For both species, the presence of males and females was similar, with a sex ratio of 1:1, coinciding with observations made by Romeu-Dalmau et al. (2011) in citrus orchards.
Concerning earwig phenology, we found individuals throughout the year in apple orchards.The mature stages of F. auricularia were observed mainly from May to November in tree shelters and immature ones from October to June in ground shelters.Most published studies have been based on tree sampling, reporting the presence of F. auricularia individuals from May to October, with a May-June peak for N3 and N4 instars, and the abundance of adults in July (Lamb & Wellington, 1975;Phillips, 1981;Helsen et al., 1998;Gobin et al., 2008;Moerkens et al., 2009Moerkens et al., , 2011)).Romeu-Dalmau et al. (2011) also observed a longer active period in Mediterranean citrus orchards-an observation that coincides with our results.The decrease in tree shelter captures during the summer months may be explained by the increased availability of natural shelters during this period.For instance, Helsen et al. (1998) observed that an increase in the size of apples leads to greater numbers of earwigs in fruit clusters, thus reflecting the availability of alternative shelters in the tree canopy.
In our study, the N2, N3 and N4 instars of F. auricularia were not found in a consecutive order along the months of the year in tree or in ground shelters.These findings may indicate the coexistence of single brood and double brood strategies, as observed by Helsen et al. (1998), Gobin et al. (2008), and Moerkens et al. (2009) in pip fruit orchards in central-northern Europe.The single brood strategy has been reported to be more susceptible to cold temperatures than the double brood one (Moerkens et al., 2012).Therefore, depending on the strategy type prevailing in each area, distinct population fluctuations might be observed.Although low temperatures can be considered a crucial determinant of earwig mortality (Moerkens et al., 2012), in Mediterranean orchards nymphs were also found during winter, thereby indicating that earwig development in these conditions does not stop, as nymphs also hibernate.Due to these differences in phenology, the predictions of population dynamics through the available degree-days models are not appropriate in Mediterranean orchards.Adult individuals of F. pubescens were observed yearround (except in May in tree shelters) and nymph instars were detected from April to June in ground as well as in tree shelters.However, Romeu-Dalmau et al. (2011) observed individuals only from May to December.This finding could be attributed mainly to the sampling method, as the earlier earwig instars on the ground could not be detected by the beating technique used by those authors.
The occurrence of the different nymph instar stages of F. pubescens in apple orchards has not, to the best each species, which prevent interspecific competition.In this regard, Debras et al. (2007) observed that F. pubescens concentrated predation on younger pear psylla instars, whereas F. auricularia showed a preference for older ones.
The occasional presence of L. riparia, E. moesta, and N. lividipes can be explained by their low aggregation coefficient and, in some cases, solitary behavior (Albouy & Caussanel, 1990;Sauphanor & Sureau, 1993).The observation that these species were found only in ground shelters is consistent with their low appearance in literature as biocontrol agents in fruit orchards, as those surveys addressed mainly tree canopies.L. riparia, N.

Figure 1 .Figure 2 .
Figure 1.Number (mean ± SE) of (a) and (b) from April to July per year in the four orchards, BB, IU, MO and MI.Column bars marked with different letter HSD or Welch's tests (p<0.05).Note that y-axis scales are different.

Figure 3 .
Figure 3. Weekly mean earwig individuals per tree and ground shelters for (FA) and (FP) - Figure 4.F.auricularia and F. pubescens forIU and MI orchards (2011-2012).A negative value indicates active competition, zero no interaction, and a positive value an association between species.

Table 1 .
Taylor's parameters for each orchard and species; b indicates when the population in shelters was regular (b<1), random (b=1) or aggregated (b>1).

Table 2 .
stage according to the degree-days model (Model) and observations (Tree and Ground).

Table 3 .
earwig individuals for each developmental stage found in tree canopies.