Introduction
Thrips (Thysanoptera: Thripidae) are best known as phytophagous pests threatening various agroecosystems worldwide (Mound Reference Mound2005; Silva et al. Reference Silva, Hereward, Walter, Wilson and Furlong2018; Rodriguez-Saona et al. Reference Rodriguez-Saona, Vincent and Isaacs2019) including in North America. However, besides being pest species, thrips are a highly diversified insect order (Mound Reference Mound2009). Climate change is expected to enhance major pest thrips fitness (Reitz et al. Reference Reitz, Gao, Kirk, Hoddle, Leiss and Funderburk2020) and globally expand thrips species geographical distribution by modifying their local faunal habitats (Park et al. Reference Park, Mo, Lee, Lee, Lee and Cho2014; He et al. Reference He, Lin, Qian, Li, Xi, Yang and Gui2017).
Thrips occur in almost all terrestrial ecosystems and are usually associated with both dead (fungus decomposers) and living vegetation (phytophagans and predators; Mound Reference Mound2018). About one-quarter of thrips species depend upon flowering plants for feeding and reproduction (Mound Reference Mound2009). In agroecosystems, flower-feeding species often use both cultivated and uncultivated vegetation, creating small-scale migratory dynamics that favour crop colonisation by pest thrips from adjacent wild host plants (Mound and Tuelon Reference Mound and Tuelon1995; Silva et al. Reference Silva, Hereward, Walter, Wilson and Furlong2018). Wildflower vegetation in crop field margins can therefore be an important source of pest thrips in agricultural landscapes (Pearsall and Myers Reference Pearsall and Myers2001; Northfield et al. Reference Northfield, Paini, Funderburk and Reitz2008), and monitoring thrips distribution and plant use in peripheral habitats could be critical to prevent pest thrips outbreaks. However, identification delay or lack of local taxonomic updates are frequent issues complicating pest thrips management (Reitz et al. Reference Reitz, Gao and Lei2011; Sabahi et al. Reference Sabahi, Fekrat and Zakiaghl2017; Wu et al. Reference Wu, Tang, Zhang, Xing, Lei and Gao2018). Most studies on thrips have primarily focused on economic losses and control of common pest species in crops (Moreira et al. Reference Moreira, de Oliveira, de Morais Oliveira, de Souza and Breda2014; Renkema et al. Reference Renkema, Evans and Devkota2018), which represent only approximately 1% of all thrips species (Morse and Hoddle Reference Morse and Hoddle2006). Consequently, thrips local community composition beyond crop pests, mainly associated with wild host plants near crops, remains neglected. Thrips communities in agroecosystems require more attention because their local diversity quickly evolves through changes in land use and crop production practices (Mound Reference Mound2018). Flowers are appropriate sampling units to describe thrips fauna in agroecosystems because they allow thrips screening in both the crops and the surrounding wild vegetation (Northfield et al. Reference Northfield, Paini, Funderburk and Reitz2008).
Thrips are pests in small fruit and berry crops and are frequently reported to attack raspberries (Rosaceae) (Leach and Isaacs Reference Leach and Isaacs2018) and blackberries (Moraceae) (Rhodes and Liburd Reference Rhodes and Liburd2017), as well as strawberries (Rosaceae), which are particularly susceptible (Mound Reference Mound2009). With specialised sucking mouthparts, thrips larvae and adults pierce plant cells to suck their contents (Bournier Reference Bournier1983). Thrips feeding and oviposition on strawberries create characteristic net-like russet discolouration called bronzing (Koike et al. Reference Koike, Zalom and Larson2009). Thrips pollen feeding and oviposition on strawberry flowers cause less injury, although high populations damage stamens and limit pollen maturation (Steiner and Goodwin Reference Steiner and Goodwin2005). Strawberry bronzing due to thrips is widely reported in the United States of America (Dara et al. Reference Dara, Peck and Murray2018; Renkema et al. Reference Renkema, Evans and Devkota2018), Europe (Linder et al. Reference Linder, Terrettaz, Antonin and Mittaz2006; Sampson and Kirk Reference Sampson and Kirk2013), the Middle East (Coll et al. Reference Coll, Shakya, Shouster and Steinberg2007), and Australia (Steiner and Goodwin Reference Steiner and Goodwin2005). Main pest thrips issues in strawberry involve the invasive Frankliniella occidentalis (Pergande) (Coll et al. Reference Coll, Shakya, Shouster and Steinberg2007; Sampson and Kirk Reference Sampson and Kirk2016), whereas F. tritici (Fitch) seems to be dominant in eastern North America (Matos and Obrycki Reference Matos and Obrycki2004; Northfield et al. Reference Northfield, Paini, Funderburk and Reitz2008), possibly migrating yearly into Canada from southern locations via prevailing winds (Lewis Reference Lewis1991). Thrips damage is predicted to increase due to the higher temperatures and atmospheric CO2 concentration that accompany climate change (Reitz et al. Reference Reitz, Gao, Kirk, Hoddle, Leiss and Funderburk2020), including at high latitudes (Parikka and Tuovinen Reference Parikka and Tuovinen2014). Thrips are key pests in Canada’s strawberry-producing provinces (Tellier Reference Tellier2021; Ontario Ministry of Agriculture, Food and Rural Affairs 2022). Thrips species diversity remains overlooked in Canada (Foottit and Maw Reference Foottit, Maw, Langor and Sheffield2019) and has not been evaluated in Québec since the extensive work of Chiasson (Reference Chiasson1986), who described thrips–host plant relationships in Canada based on widely collected thrips across the country and maintained her own collections in Ontario and Québec.
In the present study, we investigated the abundance and diversity of thrips in strawberry agroecosystems (i.e., field and adjacent uncultivated margins) on Orléans Island, near Quebec City, a region known for strawberry production. Based on preliminary work in 2016 and 2017 (C. Cloutier, unpublished data), sampling covered the entire strawberry production season in 2018 in both cultivated and wild areas, with emphasis on potential relations between these environments. After describing the flower thrips community and its phenology in the strawberry agroecosystem, we tested whether wildflower resources and their associated thrips affected the pest thrips assemblage in the strawberry fields. We expected that, despite the presence of various thrips species associated with wildflowers, F. occidentalis and F. tritici would likely predominate. In addition, we presumed that those thrips species may not overwinter on Orléans Island because of the region’s cold winters and thus may recolonise local strawberry fields annually, arriving after the spring thaw and early springtime period. Finally, we assumed that local wildflower resources and the densities of the main pest thrips species on frequently used wildflower species would influence pest thrips densities that we would observe in strawberry flowers due to their net movement from wild hosts towards crops.
Materials and methods
Study sites and seasonal partitioning of sampling
Weekly sampling of day-neutral strawberries on raised beds covered with plastic mulch was conducted on Orléans Island, near Quebec City, Québec, Canada during the 2018 strawberry production season. Initially, our study also covered the 2016 and 2017 seasons (June–October), but data were not included in the final analyses due to very low thrips abundance (C. Cloutier, unpublished data). Samples were collected weekly from mid-May until mid-October 2018 at 13 sites. Each site consisted of a day-neutral strawberry field (approximately 83 m long × 80 m wide) that was planted with “Seascape” cultivar and one of its uncultivated field margins (an approximately 3-m-wide field border) that was naturally colonised by wildflowers. Sites were distributed among four farms that applied conventional strawberry crop management, which includes the use of pesticides, the main control method for thrips generally (Reitz et al. Reference Reitz, Gao, Kirk, Hoddle, Leiss and Funderburk2020) and in strawberry (Steiner and Goodwin Reference Steiner and Goodwin2005). Sampled uncultivated margins were selected so as to standardise the width as well as the degree of vegetation cover while minimising the influence of access roads and the nearby proximity of crops other than strawberries.
The day-neutral Seatascape cultivar (Fragaria × ananassa Duchesne) continuously produces fruits from early summer until autumn. Typically, strawberry fields are exploited during a two-year rotation cycle, two-year-old plants being destroyed due to disease susceptibility after early summer production. The thrips sampling season was divided into three periods based on the seasonality of crop production and thrips abundance in 2018. The first period, referred to as May–June (18 May–03 July), corresponded to harvest in two-year-old fields in early season. The second period, or July–September (10 July–17 September), corresponded to harvest in one-year-old fields, with sampling starting as soon as the first flowers appeared in all sampled sites; this was the main fruit-production and high thrips–abundance season. In the third period, or October (24 September–16 October), fruit harvest in one-year-old crops was reduced, and thrips abundance declined. During this period, fields were sheltered at night with polyethylene covers until permanent sheltering for winter. Each of the three periods also presented a distinct wildflower profile in the uncultivated field margins. Four sites were sampled in May–June, six sites were sampled in July–September, and three sites were sampled in October because of logistical constraints.
Early season thrips emergence
Emergence traps were randomly installed in the strawberry fields and their uncultivated margins. Each trap consisted of an inverted opaque plastic box (31 cm × 28 cm × 14 cm) with a 150-mL collecting vial filled with water and a drop of liquid soap. Traps were set up on 8 May and checked weekly during May and June 2018, with four traps deployed in the strawberry field and four traps in the adjacent uncultivated margin at each site (total sampling effort of 12.5 m2/site in each environment). Trapped specimens were collected once each week. After each sampling event, traps were randomly moved 2 m away from their earlier position to increase trapping probability.
Thrips captures on flowers and fruit damage monitoring
Sampling events for thrips in strawberry flowers and wildflowers from May to October occurred at weekly intervals. In each site, the strawberry field–uncultivated margin interface was subdivided lengthwise into 10-m-wide strips that extended into both environments for concurrent sampling. During a sampling event at each site, one strip was randomly selected among the eight available for sampling. The field was sampled with 1-m2 quadrats at intervals of 2, 7, 11, 15, and 19 m from the field–margin interface, each corresponding to strawberry rows (0.76 m wide, spaced by 0.61 m). In each field quadrat, strawberry flowers were counted, and a maximum of five flowers were collected in one vial. Ripe and unripe strawberry fruits were also counted, and one unripe fruit was collected. In line with the field transect, uncultivated field margins were similarly sampled in four 1-m² quadrats that were randomly selected within the 10-m-wide strip. In each quadrat, flowers were counted for each wildflower species, and a maximum of five flowers per species was collected. In the case of plants with compound inflorescences, such as vetch (Vicia sp.) (Fabaceae) or goldenrod (Solidago spp.) (Asteraceae), the entire raceme was collected and considered as one flower unit. The total vegetation area sampled at each site during a sampling event amounted to 20 m² in the strawberry field and 4 m² in the field margin. Samples were stored at 4 °C until dissection under a stereomicroscope (Zeiss, Oberkochen, Germany) to detect thrips. Adult and larval thrips were counted and stored in 45% ethanol to prevent muscle stiffening. Strawberry fruits were observed under a stereomicroscope to identify bronzing that could be attributable to thrips, and scoring of bronzing was ranked for severity from 1 (low) to 5 (high), based on the proportion of the fruit that was affected by bronzing.
Thrips and flower identification
In flower, fruit, or trap samples containing fewer than 100 adult thrips, all individuals were identified to species. In those samples with more than 100 or 200 adult thrips, a fraction (50% or 25%, respectively) of all individuals were randomly subsampled, slide-mounted in 45% ethanol, and examined with an Olympus B × 41 compound microscope (Markham, Ontario, Canada). Adults were sexed and identified using morphological keys by Mound and Kibby (Reference Mound and Kibby1998), Stannard (Reference Stannard1968), and Hoddle et al. (Reference Hoddle, Mound and Paris2012). Specimens were sent for identity validation to Agriculture and Agri-Food Canada’s Canadian National Collection of Insects, Arachnids and Nematodes (Ottawa, Ontario, Canada), and voucher specimens were deposited at the Laboratoire d’expertise et de diagnostic en phytoprotection (Québec Ministry of Agriculture, Fisheries and Food; Québec, Canada; https://www.mapaq.gouv.qc.ca/fr/Productions/Protectiondescultures). Because we were mainly interested in Terebrantia thrips diversity, Tubulifera were identified only to genus. Immature thrips were not identified to species due to taxonomic difficulties, even for dominant species identified in our system (Skarlinsky and Funderburk Reference Skarlinsky and Funderburk2016). All flowering plants were identified to species (Clemants and Gracie Reference Clemants and Gracie2006; Ministère de l’Agriculture, des Pêcheries et de l’Alimentation and Ministère des Forêts, de la Faune et des Parcs 2019).
Statistical analyses
Analyses were performed using R software, version 3.6.1 (https://cran.utstat.utoronto.ca/bin/windows/base/old/3.6.1). Redundancy analyses were used to assess how variations of thrips diversity and abundance in strawberry (i.e., the response variables) might be related to their abundance and diversity on wildflowers in field margins (i.e., the explanatory variables). Redundancy analysis is particularly suitable for statistical testing at the community level (Legendre and Legendre Reference Legendre and Legendre2012; Borcard et al. Reference Borcard, Gillet and Legendre2018). Because we expected that regional, local, and seasonal factors (e.g., phenology and wildflower community composition) might be important, site and wildflower species variables were initially considered as explanatory variables.
Analyses were performed using the “rda” function from the vegan package (Oksanen et al. Reference Oksanen, Simpson, Blanchet, Kindt, Legendre and Minchin2019). Considering that wildflower diversity and abundance vary greatly in time and space, thereby preventing systematic sampling of five flowers per species, analyses were performed on mean thrips densities per flower unit. As data contained a high proportion of collected samples with no (0) thrips (71%), the Hellinger transformation was applied to response variables (Borcard et al. Reference Borcard, Gillet and Legendre2018). Inflating colinearity among the explanatory variables (thrips–wildflower associations (n = 20)) was avoided by using the function “vif.cca”. Important explanatory variables were then forward selected with the “ordistep” function based on the Akaike information criterion, using the 5% threshold for inclusion (Blanchet et al. Reference Blanchet, Legendre and Borcard2008). The predictive value of the final model was evaluated using the “Rsquare Adj” function. The proportion of variance explained by each explanatory variable included in the final model was examined using “varpart” function.
Results
From the 44 species of flowering plants sampled, including strawberry, about 10 000 flowers were collected and more than 12 000 thrips specimens were extracted, of which 73% were adults. A total of 11 thrips species were identified from the margins and fields (Fig. 1). Only one species belonged to the Tubulifera suborder: this was an undetermined Haplothrips sp. Most species were phytophagous thrips of the Thripidae family, with the exception of two predatory Aeolothrips spp. (Aeolothripidae). Sampled only on strawberry flowers and fruits, immature thrips were first- or second-instar larvae, with very few pro-pupae and pupae recognisable to their wingtips (Bournier Reference Bournier1983), observed only four times in the strawberry flowers.
Thrips species relative abundance (%) during the entire sampling season and by sampling environment. Total adult thrips per sampling environment (Ntot) and per thrips species in both environments (n) are shown in brackets. In the legend, thrips species are in descending order of relative abundance, detailed on the right for Halothrips sp., O. biuncus and A. fasciatus. Aeolothrips crassus and Chirothrips manicatus were two additional species anecdotally observed in margins.
Orius spp. (Hemiptera: Anthocoridae), a genus that includes common thrips predators occurring in crop fields in North America (Kelton Reference Kelton1963), were rarely found in flowers. Nineteen were caught in wildflowers, and 37 were captured in strawberry flowers. Among other thrips natural enemies, Chrysopidae (Neuroptera) and Coccinellidae (Coleoptera) were occasionally observed in the wild and cultivated vegetation, but no specimens were found in the strawberry fruit or flower samples.
Thrips emergence in early season
Close to 92% of adult thrips emergence in the spring (May–June period) occurred in the uncultivated field margins. Nine species emerged from margins, and four from the fields (Table 1; Fig. 2). The first captures in emergence traps occurred on 15 May (Supplementary material, Table S1) when daily mean air temperature approached 10 °C. Thrips trehernei (Priesner), F. tritici, F. intonsa (Trybom), and T. atratus (Haliday) were the first species to emerge in the margins (Supplementary material, Table S1). During May–June, the above four species emerged in margins but rarely in the field. The undetermined Haplothrips sp. was the most numerous species to emerge, first in early June, with peak emergence in late June (Supplementary material, Table S1). Overall, five species emerged in numbers exceeding 10 individuals, 97% of which were females, with rare males of T. trehernei, F. tritici, Odontothrips biuncus (John), and T. tabaci also emerging (Table 1). Notably F. tritici and F. intonsa were among the few thrips to emerge from the strawberry fields (Supplementary material, Table S1).
Emerged adult thrips captures in the May–June period. Data presented are total adult thrips numbers (n = 4 sites) for a total sampling effort of 12.5 m² in both uncultivated margins and strawberry field.
Total captured thrips adults from two sampling environments during May–June (18 May–3 July). Data are mean number of thrips per trap (± standard error; n = 4 sites) for a total sampling effort of 12.5 m² in both uncultivated margin and strawberry field.
Flower–thrips community composition and phenology in strawberry and wildflowers
Diversity and abundance of thrips in flowers varied according to sampling environment (margins versus strawberry fields) and period (Fig. 3). In strawberry flowers, thrips diversity was largely dominated by F. intonsa and F. tritici, representing more than 95% of all adult thrips collected throughout the season (Fig. 3). Other species collected on strawberry were T. tabaci (2%) and T. atratus (1%), with even fewer Haplothrips sp., T. trehernei, T. varipes, and O. biuncus collected (Supplementary material, Tables S2, S3, and S4).
Thrips species relative abundance (%) in margins and strawberry fields, by sampling period. Numbers of thrips per species (n) as well as total of all identified adults, including unrepresented species (TOTAL), are indicated. Only species with more than 100 collected adults are represented, except in the October period where all individuals are represented. Number of sites varied with periods, but sampling effort in each site at each sampling event was fixed at 4 m² in margin and 20 m² in strawberry field.
On flowers of two-year-old strawberry plants in the May–June period, F. intonsa and F. tritici reached their maximum densities around 3 July, particularly F. intonsa (Fig. 4). During the July–September period, on flowers of 1-year-old strawberry plants, densities of F. intonsa (especially females) peaked twice, around mid-August and again in early September. In contrast, F. tritici had a single peak around 20 August (Fig. 4). Density of thrips larvae on strawberry flowers started to rise in mid-July and peaked around mid-August (Fig. 4); no thrips were observed in early September.
Mean density of Frankliniella intonsa and F. tritici by sex, and of unidentified thrips larvae, by sampling date. Density of larvae, which we could reasonably assume mostly derived from Frankliniella reproduction, shown only for strawberry flowers. Data presented are mean thrips/10 flowers ± standard error. Number of sites varied with periods, but sampling effort in a site for each date was fixed at of 4 m² in margin and 20 m² in a strawberry field.
Comparatively, in wildflowers, Frankliniella spp. densities per flower unit were low to moderate until August. Frankliniella tritici densities peaked sharply in uncultivated margins, also around 20 August, the only time when its densities exceeded those of F. intonsa. Both species were found at low densities in strawberry fields and margins until sampling ended in mid-October.
Bronzing attributable to thrips feeding was rarely observed. Only 27 adult thrips were caught on unripe fruits, which were mostly female F. intonsa (n = 16) and F. tritici (n = 6), with single mentions of T. tabaci (both sexes), T. atratus, and Aeolothrips fasciatus (Linnaeus) (one female each). No immature thrips were found on strawberry fruits during the May–June and October periods, but 37 larvae were detected under sepals in July–Sept, mostly in mid-July.
On wildflowers, the three dominant species were F. tritici (31% of total abundance) F. intonsa (29%), and T. trehernei (25%; Fig. 1; Supplementary material, Tables S2, S3, and S4), which together accounted for approximately 85% of all adult thrips collected on wildflowers. Other major species were T. atratus (Haliday), T. tabaci, and T. varipes (Hood) (Fig. 1). Sex ratio was generally skewed in favour of females, except for F. tritici and T. trehernei, which had nearly 1:1 male:female ratios.
Females of Haplothrips sp. and Odontothrips biuncus were rarely collected in high numbers on wildflowers (> 50), they represented only 3% and 4% of total abundance, respectively. Another minor thrips species collected in uncultivated margins was A. fasciatus. Anecdotally, single records of female Aeolothrips crassus (Hood) and Chirothrips manicatus (Haliday) were observed, respectively, in the July–September period on Vicia cracca, and in the May–June period on Taraxacum officinale (Weber ex F.H. Wiggers).
Target species F. intonsa and F. tritici were collected on 18 May on Tussilago farfara (Asteraceae) and Taraxacum officinale flowers and shortly after on strawberry flowers (21 May). By the end of May–June, 10 species had been collected on strawberry flowers, among which three numbered fewer than 10 individuals (Supplementary material, Table S2). In July–September, when sampling switched to one-year-old strawberry, total species richness was also 10, including nine species that were previously found in May–June in two-year-old fields (Supplementary material, Table S3). However, total thrips abundance was more than four times higher than in May–June. In October, total species richness decreased to five, few thrips being caught other than F. intonsa (n = 6), F. tritici (n = 23), and T. trehernei (n = 10; Supplementary material, Table S4), especially on strawberry flowers (6 of 45 captured specimens). The last captures of F. intonsa and F. tritici on strawberry flowers occurred 2 October and again two weeks later in margins on Sinapis arvensis (Linnaeus) (Brassicaceae), where females of these thrips species coexisted until 9 October.
Use of wild floral resources in margins by pest and nonpest thrips species
Adult and larval thrips densities per floral unit of each host plant species varied during the season (Fig. 4), as observed in 2017 in preliminary surveys (C. Cloutier, unpublished data). Several thrips species were mostly or exclusively associated with wildflowers, e.g., T. trehernei, Haplothrips sp., and O. biuncus, whereas others, in particular Frankliniella species, occurred on both wild and strawberry flowers (Figs. 2 and 3). Flower species diversity gradually increased during the sampling season, with the most abundant wildflower hosts (e.g., Vicia cracca) blooming in late summer, from August onwards (Fig. 5). Some wildflowers appeared as key hosts for specific thrips species, and particular thrips–wildflower associations prevailed. The main uncultivated thrips host plant overall was Vicia cracca Linnaeus, which sustained approximately 34% of the total adult thrips sampled in margins and hosted 10 of the 11 thrips species present, particularly T. trehernei, F. tritici, F. intonsa, and O. biuncus (Fig 6; Supplementary material, Fig. 1). Together with V. cracca, flowers of Sinapis arvensis, Sonchus asper (Linnaeus), Cichorium intybus (Linnaeus) (Asteraceae), Taraxacum officinale, and Oenothera biennis (Linnaeus) (Onagraceae) supported approximately 80% of all adult thrips. Notably, F. intonsa, F. tritici, T. trehernei, and T. varipes shared much of their large wildflower host range, which was most clear in July–September (Fig. 6B). Wildflower species also varied in their contribution to thrips immature development, the wildflower hosts mostly used by larvae in May–June and July–September being Vicia cracca and Leucanthemum vulgare (Lamarck) (Supplementary material, Tables 2 and 3; Fig. 4).
Flowering phenology of the most common uncultivated wildflowers bordering strawberry fields in 2018 on Orléans Island. Wildflower species shown (N = 17) in descending order of seasonal abundance represent approximately 90% of all flowers examined for thrips presence.
Relative abundance of main thrips species on major wildflower species in margins: A, during May–June (18 May–27 June) on wildflowers hosting more than 10 adult thrips and for thrips species with more than 10 individuals; B, during July–September (10 July–17 September) for wildflowers hosting more than 70 adult thrips, for thrips with more than 50 individuals. Numbers of thrips (nT) and flower units (nF) in brackets, excluding thrips larvae, not identified to species. In legend for each period thrips species listed in descending order of relative abundance.
Relationships between thrips in strawberry and margins
Possible causal relationships between thrips in field margins and strawberry crops were revealed by redundancy analysis using the observed thrips–wildflower associations in margins (n = 20) that we considered as potential explanatory variables (Supplementary material, Tables 2, 3, and 4) of the abundance of main thrips species in strawberry. Colinearity between the 20 explanatory variables was deemed minor, as revealed by vif.cca (colinearity index < 10). The two explanatory variables retained in the redundancy analysis model (ordistep; 5% significance threshold) were the density of F. intonsa on S. arvensis and the density of F. tritici on Vicia cracca (Fabaceae) (Table 2; Fig. 7).
Analysis of variance table for explanatory variables considered in redundancy analysis model of thrips abundance in strawberry, and thrips abundance on wildflowers in field margins (with site as a random factor). Variables are ranked in order of importance based on permutational selection.
The final redundancy analysis model was highly significant (F 4,96 = 3.702; P = 0.007), with an adjusted R-square value of 0.072 and both redundancy analysis axe 1 (P = 0.040) and axe 2 (P = 0.038) being significant (above the 5% level). The marginal effects of explanatory variables “Frin_wild mustard” (P = 0.0108) and “Frtr_tufted vetch” (P = 0.0370) were both significant (Table 2).
Figure 7 shows the redundancy analysis model fitted to the data for the whole season. Note that angles between variables (eigenvectors) of thrips densities in strawberry fields and on their major host plants in the field margins reflect their correlations. Centroids for each month (blue dots) reflect how well, as the season advanced, densities of F. intonsa and F. tritici in strawberry crops correlated with their densities in the field margins on main host plants, according to the model.
Redundancy analysis (RDA) correlation Triplot on density of thrips Frankliniella intonsa and F. tritici in strawberry fields over the season, as explained by their density on wildflowers in field margins. Red lines represent thrips density in strawberry, and green lines represent thrips density on wildflower hosts retained as significant explanatory variables. Blue dots are the fit to the redundancy analysis model of data for each month.
Discussion
Thrips community composition and phenology
We found 11 thrips species in the strawberry agroecosystems of Orléans Island. Most of the species are already known from flora of North America (Chiasson Reference Chiasson1986), with F. intonsa being newly reported for eastern Canada (Nakahara and Footitt Reference Nakahara and Foottit2007). Surprisingly, the strawberry pest F. occidentalis was absent. Thrips richness and abundance on wildflowers varied, with a few flowering plants supporting the majority of the community. Our results found local wildflowers and sampling sites influenced thrips abundance on strawberry.
In the uncultivated field margins, the thrips community was more diversified than in the strawberry crops but was dominated by three species. First, T. trehernei predominated, being found on 25 wildflowers. Widespread in North America, this species probably originated from Europe and was previously reported from Québec (Nakahara Reference Nakahara1994) as a polyphage on Taraxacum spp. (Asteraceae), Sonchus spp. (Asteraceae), and Oenothera spp. (Onagraceae) (Chiasson Reference Chiasson1986; Nakahara Reference Nakahara1994). Second in importance, F. tritici was found on 29 wildflowers. As a native species (Stannard Reference Stannard1968; Hoddle et al. Reference Hoddle, Mound and Paris2012), F. tritici was known in flowers from disturbed habitats such as Sonchus spp., Trifolium spp. (Fabaceae), or Aster spp. (Asteraceae) (Chiasson Reference Chiasson1986; Chellemi et al. Reference Chellemi, Funderburk and Hall1994). The third main species in the field margins, F. intonsa, was previously reported from British Columbia, Canada on Prunella vulgaris (Lamiaceae) (Chiasson Reference Chiasson1986). It has been found in Washington State, United States of America, from various flowers, weeds, and other plants since 1972 (Nakahara and Foottit Reference Nakahara and Foottit2007). Presumably from Asia, it spread to Europe where it occurs on Sinapis (Brassicaceae), Trifolium, or Ranunculus (Ranunculaceae) flowers (Atakan and Uygur Reference Atakan and Uygur2005; Hoddle et al. Reference Hoddle, Mound and Paris2012). In the present study, it was found on 26 wildflowers.
Numerous other, less abundant thrips were found in the field margins. Thrips varipes was present in all periods. It is widely distributed in the United States of America (Chiasson Reference Chiasson1986; Nakahara Reference Nakahara1994) on flowers of Ranunculaceae (Stannard Reference Stannard1968). We found it on 12 wildflowers, mainly on Sonchus asper. Odontothrips biuncus was found almost exclusively on V. cracca, in agreement with its known association to Fabaceae (Mound and Kibby Reference Mound and Kibby1998). Few T. atratus were observed despite being a widespread polyphagous thrips in the Northern Hemisphere on Compositae (Asteraceae) flowers. It was already known in Canada and United States of America (Chiasson Reference Chiasson1986; Nakahara Reference Nakahara1994). Haplothrips sp. was mainly found on Leucanthemum vulgare (Asteraceae). Several Haplothrips spp. were known from Québec as pollen feeders on Compositae (Chiasson Reference Chiasson1986; Mound and Kibby Reference Mound and Kibby1998). Few T. tabaci and A. fasciatus were observed, both previously known from Québec flora (Chiasson Reference Chiasson1986). Thrips tabaci is cosmopolitan, being found on field or greenhouse crops and wild vegetation (Reitz et al. Reference Reitz, Gao and Lei2011). Aeolothrips fasciatus is a facultative predator and pollen feeder (Mound and Kibby Reference Mound and Kibby1998). A single female A. crassus, a rare species in United States of America (Stannard Reference Stannard1968), was found. Likewise, a single female of Chirothrips manicatus was found, a species that is widespread in temperate areas and associated with Poaceae, probably disseminated via the seed trade (Hoddle et al. Reference Hoddle, Mound and Paris2012). Frankliniella intonsa and F. tritici predominated in strawberry fields. Frankliniella intonsa was previously reported to damage strawberry in Europe (Buxton and Easterbrook Reference Buxton and Easterbrook1988; Linder et al. Reference Linder, Terrettaz, Antonin and Mittaz2006; van Kruistum and den Belder Reference van Kruistum and den Belder2016) and in Japan (Fujiwara Reference Fujiwara2022). In Denmark, it is the major pest species in sheltered strawberry, ahead of T. tabaci (Nielsen et al. Reference Nielsen, Sigsgaard, Kobro, Jensen and Jacobsen2021). In Asia, F. intonsa is a crop pest that also exploits wild vegetation (Wang et al. Reference Wang, Mound and Tong2019). A few T. tabaci and T. atratus occurred on strawberry in the present study. Thrips tabaci is common in Italy (Gremo et al. Reference Gremo, Bogetti and Scarpelli1997), England (Buxton and Easterbrook Reference Buxton and Easterbrook1988), the Netherlands (van Kruistum and den Belder Reference van Kruistum and den Belder2016), Denmark (Nielsen et al. Reference Nielsen, Sigsgaard, Kobro, Jensen and Jacobsen2021), Brazil (Pinent et al. Reference Pinent, Nondillo, Botton, Redaelli and Pinent2011), and Australia (Steiner and Goodwin Reference Steiner and Goodwin2005).
Frankliniella intonsa and F. tritici may overwinter on Orléans Island: they were found at the beginning of the strawberry production season and until October. Our results approximate what was observed for F. occidentalis in orchards in British Columbia, Canada. Springtime emergence of females occurred in wild areas, followed by migration towards orchards (Pearsall and Myers Reference Pearsall and Myers2000). In Florida, F. tritici overwinters as active, reproducing adults (Chellemi et al. Reference Chellemi, Funderburk and Hall1994; Northfield et al. Reference Northfield, Paini, Funderburk and Reitz2008). Winter activity can be excluded on Orléans Island in natural habitats, where winter temperatures fall well below freezing for many months. In western Canada, thrips may overwinter in soil as mated females (Pearsall and Myers Reference Pearsall and Myers2000). Local overwintering in soil on Orléans Island would be consistent with the early arrival of Frankliniella spp. males from mid-June. In Europe, Frankliniella thrips could survive natural winter conditions (Trdan et al. Reference Trdan, Bergant and Jenser2003), including in strawberry (Sampson et al. Reference Sampson, Bennison and Kirk2021). Abundance of F. intonsa and F. tritici in strawberry flowers from mid-May and their predominance thereafter suggest that the unidentified larvae captured in strawberry fields in July and August in the present study were likely current season offspring.
When sampling switched to one-year-old strawberry in July, high densities of both Frankliniella spp. were observed, suggesting their capacity to colonise new fields (Rodriguez-Saona et al. Reference Rodriguez-Saona, Polavarapu, Barry, Polk, Jörnsten, Oudemans and Liburd2010) or an attraction towards younger crops (Bournier Reference Bournier1983; Fernandes and Fernandes Reference Fernandes and Fernandes2015), although long-distance migration cannot be excluded.
Voltinism is unclear for F. intonsa in our data, with at least two density peaks. However, one generation of F. tritici most likely occurred. Thrips larval density was negligible until mid-July and peaked once, in August. However, sampling resolution and potential thrips movement between fields (Fernandes and Fernandes Reference Fernandes and Fernandes2015) prevent confirmation of the actual generation timing. Only one F. occidentalis generation has been reported on nectarine in British Columbia (Pearsall and Myers Reference Pearsall and Myers2000). Appearance of males of both Frankliniella species in the present study relatively early suggests sexual reproduction until September. Males of these species seem to be rare in autumn in Québec, which is unlike F. occidentalis in British Columbia (Pearsall and Myers Reference Pearsall and Myers2000).
Well-known F. occidentalis (Pergande) was absent from our samples, despite its cosmopolitan distribution (Kirk and Terry Reference Kirk and Terry2003) and known association with strawberry (Sampson and Kirk Reference Sampson and Kirk2016). Strawberry and other suitable hosts in field margins at Orléans Island (clovers, dandelions, goldenrods) would likely sustain it locally (Chellemi et al. Reference Chellemi, Funderburk and Hall1994; Northfield et al. Reference Northfield, Paini, Funderburk and Reitz2008). Low F. occidentalis densities in the eastern United States of America have been attributed to competitive exclusion by F. tritici at larval stages (Paini et al. Reference Paini, Funderburk and Reitz2008). Both interference and exploitation competition were found between F. intonsa and F. occidentalis, favouring F. intonsa (Bhuyain and Lim Reference Bhuyain and Lim2019). To our knowledge, strawberry agroecosystems on Orléans Island represent the first environment where coexisting F. intonsa and F. tritici are reported. The interactions between these two species and the potential cumulative impacts on F. occidentalis should be further studied because co-occurrence of two competitively superior Frankliniella spp. could negatively affect F. occidentalis under field conditions, resulting in “asymmetrical occurrence,” as Bhuyain and Lim (Reference Bhuyain and Lim2019) suggest. Future investigations could target F. occidentalis in protected crops on Orléans Island, such as strawberries, raspberries, and blueberries grown under tunnels, which are potential habitats for winter survival in semiprotected contexts, as Sampson et al. (Reference Sampson, Bennison and Kirk2021) described.
Influence of wild floral resources on strawberry thrips
Variations in thrips richness and density were observed between wild and strawberry flowers and between wildflowers in field margins. Distinction between a diversified thrips community in uncultivated field margins and one dominated by two species in strawberry crops illustrates the influence of local floral diversity over strawberry thrips community composition. This appears to be similar to the thrips community in tomato agroecosystems (Chellemi et al. Reference Chellemi, Funderburk and Hall1994). In the thrips–tomato system, a small portion of wildflowers within the field margins sustained the thrips community.
Thrips abundance on uncultivated vegetation is widely considered a positive determinant of thrips abundance in adjacent crops (Mound and Tuelon Reference Mound and Tuelon1995; Pearsall and Myers Reference Pearsall and Myers2001; Silva et al. Reference Silva, Hereward, Walter, Wilson and Furlong2018). We intuitively expected F. intonsa and F. tritici density on major wildflower hosts to correlate with their density in strawberry (Northfield et al. Reference Northfield, Paini, Funderburk and Reitz2008; Silva et al. Reference Silva, Hereward, Walter, Wilson and Furlong2018). Our redundancy analyses results indicate possible causal effects between densities of F. intonsa and F. tritici in field margins and their densities on strawberry flowers (Table 2). The abundance of F. intonsa in strawberry in July was partly explained by its abundance in July on Sinapis arvensis in margins (Fig. 6). Similarly, the abundance of F. tritici in strawberry in June, August, and September correlated with its abundance on V. cracca in field margins. This supports our prediction of a relation between thrips abundance in margins and in strawberry, even though the predictive power of the redundancy analysis model was modest (7.2%) – probably due to the multiple ecological factors at work – over a long seasonal gradient (Borcard et al. Reference Borcard, Gillet and Legendre2018).
Conclusion
Observations in strawberry agroecosystems on Orléans Island revealed a community of 11 flower-inhabiting thrips. This community mostly encompasses species already recorded from North American flora and is dominated by the endemic F. tritici and the exotic F. intonsa. This newly mentioned species for eastern Canada appears as a major thrips pest in day-neutral strawberry. Data suggest that both F. intonsa and F. tritici can overwinter under natural conditions in the agroecosystems of Orléans Island. Females emerged early in the season, mostly from uncultivated areas, established quickly in strawberry crops, and were present until late in the season. Thrips species richness was higher in uncultivated margins than in strawberry fields, but only six wildflower species (of the 44 present) sustained high thrips densities. Field site affected thrips density, which may be linked to farming practices and local wildflower profiles. Additional investigations are warranted to determine the mechanisms shaping the thrips community assemblage in strawberry crops. Our results contrast with the vision of a strawberry production system dominated by F. tritici in eastern North America (Matos and Obrycki Reference Matos and Obrycki2004). New information regarding thrips species in strawberry, their abundance in wild vegetation, and their overwintering potential should help pest decision making in strawberry (Coll et al. Reference Coll, Shakya, Shouster and Steinberg2007). The first record of F. intonsa in Québec raises awareness about thrips as emerging pests in cold temperate latitudes. Identification of F. tritici and F. intonsa as the major pests could lead to more efficient pest control in strawberry (Lim and Mainali Reference Lim and Mainali2009).
Supplementary material
The supplementary material for this article can be found at https://doi.org/10.4039/tce.2023.15.
Acknowledgements
This work was supported by Natural Sciences and Engineering Research Council of Canada (NSERC) via the Research and Cooperative Development programme (RCD) grant number CRDPJ 491223 – 2015 awarded to C. Cloutier, and cofunded by four partner producers: farms Onésime Pouliot, François Gosselin, Polyculture Plante, and M&P Vaillancourt, which the authors thank for their collaboration and providing access to strawberry fields. The authors thank Eric Maw (CNC) for thrips species identity verification and Gaétan Daigle (professional statistician in the Département de Mathématiques et statistiques, Université Laval) for statistical validation. The authors also warmly thank Philippe Gagnon, B.Sc. Biology student, for help with sampling and sorting thrips, and Annabelle Firlej, Chercheure en entomologie fruitière, IRDA, Saint-Bruno-de Montarville, QC, Canada, and Professeur Line Lapointe, Département de Biologie, Université Laval, for useful comments on the manuscript.
Competing interests
The authors declare they have no competing interests involving authors of this paper.
Open access
