In Argentina, approximately 13,000 hectares of garlic are planted annually, with the majority (90 %) situated in Mendoza province, and the remaining area distributed in San Juan, Córdoba and Buenos Aires. The commercial types are purple, red, white and early white garlic, which are grown in the following percentages: 67 % purple, 24 % red, 5.5 % white, 2.7 % early white and 0.8 % of other garlic types (IDR (Instituto de Desarrollo Rural), 2020). The international trading of Argentine garlic involves more than 30 countries; however, only Brazil, France and the USA import more than 80 % of national production (Burba, 2022).

In Argentina, the main pests affecting garlic crops are the onion thrips, Thrips tabaci Lindeman (Thysanoptera: Thripidae), the stem nematode, Ditylenchus dipsaci (Kühn) (Tylenchida: Anguinidae), and the wheat curl mite, Aceria tulipae (Keifer) (Acari: Eriophyidae). Garlic production in Mendoza province is mainly carried out by small and medium independent producers who inadequately use pesticides (Burba, 2022).

Frankliniella platensis De Santis has recently been reported in garlic plants (de Borbón, 2018). This species was described by De Santis in 1966 from specimens collected in false garlic flowers, Nothoscordum gracile (Dryand ex Aiton) Stearn (Asparagales: Amaryllidaceae) (De Santis, 1966). There were no more reports of this species until 1996, when it was found in flowers of three-cornered leek, Allium triquetrum L. (Asparagales: Amaryllidaceae) (Carrizo, 1996). Additionally, in 2012, it was recorded in Brazil from specimens collected in pitfall traps (Cavalleri & Mound, 2012). Frankliniella platensis was not reported to cause damage to commercial garlic crops; however, it may pose a potential threat to various bulb species. This insect became relevant when it was intercepted in 2015 in the Netherlands on Tulbaghia violacea Harv. (Asparagales: Amaryllidaceae) bulbs from Brazil (National Plant Protection Organization, the Netherlands, 2015). Subsequently, this insect was found in Mendoza, Argentina, infesting this host in the agroecological orchard of the Agricultural Experiment Station (EEA) Mendoza, INTA (de Borbón, 2018).

Thrips tabaci is commonly found causing damage to garlic plants in Argentina. Both larvae and adults of this species typically inhabit the base of the sheathed leaves and the inner part of young leaves near the central rib (Zamar et al., 2007). They primarily undergo pupation and diapause in the soil, making them difficult to observe, especially in the early months of cultivation. However, this thrips poses a threat to production only when the attack occurs very early.

Like other insects, thrips locate their host plants using visual and olfactory cues such as color, shape, size and plant-associated volatiles (Frey et al., 1994; Mainali & Lim, 2010; Pobożniak et al., 2021; van Tolet al., 2020). Studies have evaluated the oviposition preference of Thrips tabaci for different potato cultivar colors (Westmore et al., 2019). Additionally, positive correlations have been observed between concentrations of reducing sugars and T. tabaci abundance, while negative relationships have been noted between the total phenolic content and T. tabaci damage in various onion cultivars (Pobożniak et al., 2022). The architecture of onion plants also influences thrips abundance, with the abundance of T. tabaci being correlated to the number of leaves and the angle of leaf insertion (Loges et al., 2004).

While extensive knowledge exists regarding various aspects of the biology of T. tabaci, our understanding of F. platensis remains limited. Only a few hosts are documented, and there is a complete absence of information regarding its biological cycle.

We hypothesize that F. platensis is unlikely to establish itself in commercial garlic crops. Although Tulbaghia plants may act as a source of infestation on garlic and affect different cultivars to varying degrees. In the absence of this source of infestation, it is anticipated that F. platensis will not be able to survive on garlic cloves or establish itself on commercial garlic crops.

The objectives of this study were to (1) verify the absence of F. platensis in some commercial garlic crops, (2) confirm natural infestation on select garlic cultivars located close to Tulbaghia plants infested with F. platensis and to compare with T. tabaci infestation, (3) assess the absence of F. platensis in bulbs and cloves after harvest, (4) evaluate the reproductive capacity of F. platensis in plants of different cultivars under controlled confinement conditions, and (5) determine the oviposition preferences of F. platensis between six different garlic cultivars and two other Amaryllidaceae species in bioassays.

Table II summarizes the goodness-of-fit measures for statistical models, categorized by trial type and thrips species. All results from the Generalized Linear Models (GLMs) were significant, with the exception of the date-block interaction for T. tabaci (Table III).

Table II Table II Goodness-of-fit measures for statistical models, categorized by trial type and thrips species. Table II Goodness-of-fit measures for statistical models, categorized by trial type and thrips species. Notes

N= number of replicates, AIC= Akaike Information Criterion, BIC= Bayesian Information Criterion, logLik= Log-Likelihood

Sampling of Frankliniella platensis in Commercial Crops of Argentina

A total of 148 thrips individuals were collected from 11 samples taken from garlic-producing areas of Argentina. No adults or larvae of F. platensis were found. Of the thrips collected, 90 % were identified as Thrips tabaci adults, while the remaining 10 % belonged to the following species: Frankliniella occidentalis (Pergande), Frankliniella sp., and Limothrips cerealium (Haliday). All larvae found were identified as T. tabaci.

Table III Table III Output of Generalize Lineal Model (GLM), categorized by trial, thrips species, and fixed effects Table III Output of Generalize Lineal Model (GLM), categorized by trial, thrips species, and fixed effects Notes

Null Deviance, Residual Degrees of Freedom, Residual Deviance, and p-value (χ 2)

Frankliniella platensis was not detected in the sampled commercial garlic crops in Argentina. This thrips species was only observed on garlic plants located near Tulbaghia violacea-infested hedge in an agroecological garden.

There is limited data available concerning the biology of Frankliniella platensis. While some host plants are known, information about its breeding capabilities within these hosts is lacking. This study contributes to increasing our current knowledge about this thrips species.

Despite the presence of an infestation source of F. platensis near the garlic crop, a low number of individuals of the species was found on plants (Fig. 2). In contrast, Thrips tabaci populations were lower in the early stages of the crops but increased later in the season, particularly in spring (Fig. 3). The occasional presence of F. platensis on garlic in Mendoza, Argentina, as reported previously, could be attributed solely to the proximity of garlic plants to Tulbaghia plants infested with this thrips. Although F. platensis can complete its life cycle on garlic, its establishment in commercial crops is considered unlikely.

Most F. platensis specimens are wingless, with only a few found to be winged. Wind may play a significant role in facilitating the non-selective arrival of wingless specimens to garlic plants. Therefore, differences observed between cultivars could be attributed to intrinsic characteristics of the cultivars affecting breeding. This distinguishes F. platensis from T. tabaci, as all adults of the latter species are winged, allowing for more selective host choices.

In the case of F. platensis, the first two blocks closest to the infestation source exhibited significantly higher infestation levels than the remaining blocks (Fig. 1). However, on the second sampling date carried out during spring, no differences were observed between the first four blocks. This suggests a gradual spread of F. platensis to more distant blocks.

In this study, no specimens of F. platensis were found in garlic bulbs or individual cloves after harvest. This suggests that F. platensis is unlikely to affect the garlic trade, contrasting with what has been observed with Tulbaghia plants. Notably, F. platensis was intercepted in the Netherlands on Tulbaghia bulbs imported from Brazil (National Plant Protection Organization, the Netherlands, 2015), highlighting the need for implementing control measures related to Tulbaghia trade in Argentina.

Rearing F. platensis in cages showed that this species can develop on garlic plants, reaching the prepupal and pupal stages in the leaves. Similar observations have been made in both garlic and Tulbaghia plants in the field. This adaptation to perennial plants with linear leaves may provide a refuge for the pupal stage in the leaf insertion zone throughout the year. No thrips were found on re-sprouted garlic plants in the cage assay, likely because there were no pupae buried in the pots and all present thrips were removed when the leaves were cut. In contrast, T. tabaci, being a polyphagous species, primarily pupates in the soil (Deligeorgidis & Ipsilandis, 2004).

There is no evidence to suggest that F. platensis is present in commercial garlic crops. In these crops, leaves senesce almost simultaneously, bulbs are harvested annually and new cloves are replanted. Consequently, there are no plants left during the summer, potentially eliminating alternative hosts for F. platensis. Although it is unknown if this thrips buries as pupae in the soil, it is possible that all stages are eliminated with the harvest of garlic plants. Furthermore, F. platensis is adapted to certain perennial Amaryllidaceae such as Tulbaghia violaceae. In these plants, the leaves do not senesce simultaneously, whereas the opposite occurs in garlic.

The white garlic cultivars, Nieve and Unión consistently exhibited the highest preference for breeding by F. platensis in both field and cage trials. Conversely, in the oviposition bioassays, early white and purple cultivars were the most preferred. Various hypotheses could be proposed to explain why the oviposition bioassays contrast with field and greenhouse trials.

The leaf disk method for evaluating oviposition preference may not accurately reflect thrips breeding in the field as suggested by the results of the assays, which did not align with those of crop and cage trials. One possible explanation is that the number of inserted eggs depends on the physiological conditions of the leaves. For example, younger cucumber leaves are preferred for oviposition by Frankliniella occidentalis over older ones (de Kogel et al., 1997). A plant grown in a pot in a greenhouse may not accurately represent field conditions.

The relationship between plant architecture and thrip densities has been studied in onions, where closer leaf insertion angles and a greater number of leaves correlate with higher Thrips tabaci densities (Loges et al., 2004). It is possible that Frankliniella platensis prefers garlic plants with an architecture similar to its primary host, Tulbaghia.

The different results in field trials, greenhouse cage assays and bioassays, may be attributed to the influence of various variables such as plant architecture, the physiological state of the plants -specifically of the leaf discs- and thrips migration, among others. Moreover, these assays address different questions. In bioassays, conditions are more controlled and only the preference of cultivars for F. platensis to lay eggs on leaf discs produced in the greenhouse can be evaluated. Conversely, cage breeding trials assess preference and development in whole plants under greenhouse conditions. Finally, field trials examine the behavior of cultivars under actual field cultivation conditions.

Frankliniella platensis was not detected in commercial garlic crops during the sampling period. This thrips preferred the white cultivars Nieve and Unión, in field trials. Frankliniella platensis showed significantly lower breeding rates compared to Thrips tabaci, except on those cultivars. Furthermore, F. platensis was not found in garlic bulbs or cloves after harvest. The preference for these white cultivars was further confirmed in cage trials conducted in the greenhouse. In leaf disk bioassays, the cultivars Killa and Morado were the most preferred for oviposition, while Tulbaghia was preferred over garlic or Agapanthus.

Our findings suggest that F. platensis is unlikely to become a significant pest of garlic, although Tulbaghia plants may serve as a source of infestation for garlic crops by this thrips species.