Figures 2 and 3 summarise the subtreatment responses shown by Treatments 1 and 2.

Head-space analysis resulting from L. bonariensis feeding damage indicated de novo synthesis of the HIPV phenylacetaldehyde; of the VOCs, the weevil-damaged ryegrass also showed significant increases in the monoterpenes (Z)-β-ocimene, (E)- β-ocimene, along with the diterpene neophytadiene ( Figure 4 ). Weevil feeding damage also increased release rates of the green leaf volatiles 2-hexenal and cis-3-hexen-1-ol, but these increases were comparatively small ( Figure 4 ; Supplementary Material ).

There is an extensive body of literature that considers the interactions between insects and plants via various combinations of VOCs and HIPVs. Significantly, the HIPV mix can vary in quality and quantity compared with the VOC blends released by undamaged hosts (Schoonhoven et al., 2005). Plant HIPVs are typically more readily detected by responding species than VOCs, with the insect-induced blend conveying additional information about, for example, the presence of conspecific aggregations of host plants or intense competition for plant resources. Significant to this study, such characteristics may influence insects’ foraging decisions, leading to attraction to or repellency from the HIPV source (e.g., Kalberer et al., 2001; Meiners et al., 2005; Ballhorn et al., 2013; Aartsma et al., 2017; Menacer et al., 2023). This study has shown that diapausing populations of L. bonariensis, (Treatment 1) were, in the main, strongly attracted to feeding damage by conspecifics. Conversely, in the case of the reproductive populations, there was little if any response.

In general, the patterns of plant attractiveness found in this study are similar to those found in other Listronotus spp. McGraw et al. (2011), using olfactometry, found that the annual bluegrass weevil L. maculicollis (Kirby) was attracted to damaged Poa annua L. Justus (2019) reported that L. oregonensis was attracted to carrot and parsley in both olfactometer studies and field trappings. Other curculionid species where attraction occurs include the tree-of-heaven trunk weevil, Eucryptorrhynchus brantii (Harold) (Sun et al., 2023), the tea weevil Myllocerinus aurolineatus (Voss) (Sun et al., 2010), the plum curculio Conotrachelus nenuphar (Herbst) (Leskey and Prokopy, 2001), the boll weevil Anthonomus grandis Boheman (e.g., Magalhães et al., 2012), the black vine weevil O. sulcatus (Van Tol et al., 2004), and the cranberry weevil Anthonomus musculus Say (Szendrei et al., 2011).

The emission of the HIPVs triggered by L. bonariensis damage of the L. multiflorum plants ( Figure 4 ) in this study was typical of plant HIPV emissions measured elsewhere (e.g., Dudareva et al., 2006; Farré-Armengol et al., 2017). Furthermore, Figure 4 shows how the detected insect-elicited responses were much greater than those of the controls indicating little (if any) effect resulting from the initial handling of the test plant material. Of particular interest amongst the HIPV treatment responses was the de novo detection of phenylacetaldehyde ( Figure 4 ) which is known to occur in response to insect feeding across a range of plant species (e.g., El-Sayed et al., 2008, 2016, 2018; Gutensohn et al., 2011). Zhang et al. (2023) have shown that phenylacetaldehyde induced by the rice water weevil, Lissorhoptrus oryzophilus Kuschel feeding on rice seedlings is attractive to the incipiently reproductive spring populations of the weevil when migrating from their overwintering sites into the emerging rice crop (Saito et al., 2005). This has the effect of attracting more conspecifics onto the plants (Zhang et al., 2023). Significantly, these workers then determined that the L. oryzophilus responsiveness to phenylacetaldehyde ceased when the weevils entered their summer reproductive phase. This life history is strikingly similar to that found in L. bonariensis except that in New Zealand the adults diapause in the pasture itself (e.g., Goldson, 1981b). As with L. oryzophilus the present study has shown that late diapausing/spring L. bonariensis populations showed strong responses to the HIPVs including phenylacetaldehyde ( Figure 4 ) whereas, as with L. oryzophilus, this almost completely disappeared in the summer reproductive populations ( Figure 3 ). This result explains the lack of L. bonariensis responsiveness found by Hennessy et al. (2022) who worked on only reproductive populations of the weevil. The adaptive implications of this seasonal pattern are discussed below. Notably many of the c.100 species of Listronotus occupy the similar semiaquatic habitats as L. oryzophilus (e.g., Blatchley and Leng, 1916; Lloyd, 1966; O’Brien, 1977).

The cessation of attraction to the VOC/HIPV treatments shown by diapausing L. bonariensis (Treatment 1) to damaged ryegrass and frass after the addition of 20 conspecifics ( Figure 2 ) is consistent with the generally observed population dispersion as L. bonariensis populations increase. Barlow et al. (1994), conducted an analysis of field populations of L. bonariensis in the South Island of New Zealand and showed that the adult weevils were aggregated at low densities but trended towards randomness when approaching c. 400 m^-2. Barker and Addison (1989) found a similar pattern in the North Island of New Zealand and McNeill et al. (1998) tested weevils confined to field plots and again, found evidence of negative density dependent oviposition. Probably also related to this, Prestidge et al. (1987) found that there was never more than 3% of ryegrass tillers in any sample showing multiple L. bonariensis ovipositional sites. Pilkington and Springett (1988) noted that the proportion of weevil ovipositional holes plugged with frass increased with increasing weevil density and suggested that there was an epideictic pheromone associated with this. Goldson et al. (2011) using a life-table approach, found that the realised fecundity of the emergent first summer generation of L. bonariensis was inversely related to the density of adults per ryegrass tiller. Such population density dependent responses in L. bonariensis populations could go some way towards explaining the very strong negative influence that caging has on L. bonariensis reproductivity, including the resorption of oöcytes (Goldson, 1981b; Vereijssen and Goldson, 2016).

More generally, the pattern of population density-related behaviour in L. bonariensis is similar to the observation of Tansey et al. (2005) who showed that on leafy spurge (Euphorbia spp. near esula), the HIPVs arising from flea beetle, Aphthona nigriscutis Foudas, damage ranged from attractive to repellant depending on their concentration. Similarly, Ballhorn et al. (2013) noted that lima bean plants (Phaseolus lunatus L.), when emitting low HIPVs, attracted males of the chrysomelids Gynandrobrotica guerreroensis Jacoby and Ceratoma ruficornis Olivier to facilitate mating. At higher concentrations of the HIPVs, they became repellent (Ballhorn et al. (2013).

Studies elsewhere have shown that the sex and physiological condition of insects influence their responses to HIPVs (e.g., Bernays and Chapman, 1994; Davidson et al., 2006; McGraw et al., 2011; Ballhorn et al., 2013). However, sparse evidence for this in this contribution is probably not definitive in that when the L. bonariensis population was partitioned into the various physiological categories for analysis, small sample sizes became problematic.

The L. bonariensis behaviour observed in this contribution may well reflect relict aspects of the natural history of L. bonariensis in its native range, in a way similar to that found by Goldson and Emberson (1980) and Goldson (1981b) when considering the species’ diapause seasonality. These workers showed that the New Zealand weevil populations entered diapause throughout the country triggered by a critical photoperiod of 12.3:11.7 hours L:D. While of no apparent adaptive advantage in New Zealand, Goldson and Emberson (1980) concluded that this behaviour had occurred as a result of evolutionary pressures in the species’ native range. Published peer-reviewed literature relating to the biology and ecology of L. bonariensis (and M. hyperodae) in South America is sparse (e.g., Loan and Lloyd, 1974; Ahmad, 1978). However, there exist nine, sometimes untitled, mimeographed reports to the Commonwealth Institute of Biological Control (CIBC and its antecedents) written by DC Lloyd and later, Lloyd and Ahmad, between 1945 and 1973. This detailed compendium is based on field and laboratory studies describing the ecology and biology of L. bonariensis at various altitudes and latitudes in temperate South America. In the Bariloche area in Argentina (altitude 890m; latitude 41.13°S; longitude 71.3°W) the reproductive season of L. bonariensis was initially observed by Lloyd and Ahmad (1972) and Loan and Lloyd (1974) to extend from October to December, with adult weevils emerging in February overwintering and becoming reproductive in the following spring. Subsequently, Ahmad (1978) extended this reproductive season of L. bonariensis from September to February, noting that the weevil overwintered in diapause, as found in New Zealand by Goldson (1981b). Lloyd and Ahmad (1972) also observed a second summer generation of the weevil at latitudes c. 35^°S in the Buenos Aires and Montevideo areas and asserted that this life history probably occurred in latitudes 31°S to 28°S. This pattern is also very close to that which has been observed in New Zealand (e.g. Goldson, 1981b; Barker et al., 1989; Goldson et al., 2011). Moreover, Lloyd (1966) also noted that a third L. bonariensis generation was known to occur in favourable conditions such as those in Santiago, Chile (altitude 520m, latitude 33.45°S, longitude 70.68°W) again this is similar to that observed in the warmer northern parts of New Zealand (e.g., Barker et al., 1989). Although Lloyd (1966) reported a very wide distribution of L. bonariensis in South America, latitudes 17°S and 51°S and from sea level to 3660 m, he and Lloyd and Ahmad (1972) contended that the weevil’s native range is likely to be in the Andean intermontane valleys known as ‘mallines’, which comprise cool wet meadows ( Figure 5 ). These ecosystems extend from c. 36° to 56°S (e.g., Mazzoni and Rabassa, 2013) and have been comprehensively described (e.g., Squeo et al., 2006; Mazzoni and Rabassa, 2013; Curcio et al., 2023). The native Gramineae found in these ecosystems include Festuca spp., Poa lanuginosa Poir. and Hordeum comosom J.Presl (Gaitán et al., 2011) but notably, there were no Lolium spp. in the indigenous ecosystem (Frakes, 1973). Further, Mazzoni and Rabassa (2013) have also observed that variability within the mallines ecosystem occurs in very small areas, depicting high intrinsic habitat heterogeneity. Based on these descriptions it can be argued that suitable L. bonariensis host plants in the weevil’s native range were likely to have been sparsely and widely distributed, supporting low weevil population densities (Goldson, personal observation). It is probable that the weevil has subsequently become more widely distributed, as described by Lloyd (1966), with the arrival of ‘European’ agricultural systems based on exotic host plant species, especially Gramineae, which are more palatable to the weevil than the native host plants such as Paspalum gayanum E. Desv. and Sertaria geniculatea (Lam.) P. Beauv (Lloyd, 1966).

Williams et al. (1994) used Random Amplified Polymorphic DNA (RAPD) analyses of the various weevil populations collected in South America, Australia and New Zealand to determine that the New Zealand populations were likely to have originated in the River Plate area. Later, Harrop et al. (2020) used genotyping-by-sequencing (GBS) analysis to determine the genetics of L. bonariensis populations distributed throughout New Zealand. These workers concluded that a likely scenario is that there were two separate introductions from the same South American source population into regions north and south of New Zealand’s Main Divide and that thereafter there had been some migration between the two populations. While the exact locations of where the New Zealand populations came from have yet to be definitively resolved, the observations made by the CIBC researchers are probably correct, given that Bariloche is relatively close to the mallines ecosystems. That the weevil evolved in these ecosystems is also arguably reflected by the life history of its parasitoid M. hyperodae, which has been the subject of extensive research into what are taken to be its own relict characteristics, given its relatively recent introduction into New Zealand (e.g., Goldson et al., 1990, 1995; McNeill et al., 1998; Phillips and Baird, 2001; Phillips and Kean, 2017). Under laboratory conditions, the parasitoid has been shown to have low fecundity of 48 ± 8 eggs and a long life span of 21 ± 4 days (Goldson et al., 1995) coupled with very high searching efficiency (Barlow et al., 1993). Related to this and significantly when weevil hosts are abundant, 71% of the parasitoid’s ovipositional effort occurs within the first three days after eclosion and with this, 41% of the parasitoid’s lifespan occurs after the cessation of oviposition. Taken together, these characteristics point to M. hyperodae being adapted to spending a large proportion of its life searching for patchy and sparse distributions of L. bonariensis; i.e., the parasitoid’s ovipositional effort is time-limited (Phillips and Kean, 2017) and as discussed by Goldson et al. (1995) this again reflects features of the mallines ecosystem.

The absence of olfactometer responsiveness of the reproductive populations of L. bonariensis (Treatment 2) may, at first sight be counterintuitive, but this may reflect the nature of the weevil’s primitive ecosystem. In view of the mallines habitat’s latitude and altitude, there are short growing seasons. For example in Pilcaniyeu, which is at northern end of the mallines distribution (altitude; 900 m latitude 41.12°S; longitude 70.72°W), the growing season is only 2.1 months (67 days), from around 18 December to around 22 February (https://weatherspark.com/y/26504/Average-Weather-in-Pilcaniyeu-Argentina-Year-Round) Given this limitation, the short growing season could well lead to extensive L. bonariensis mortality owing to the population’s incomplete development prior to the onset of cold conditions. Under such conditions there would therefore be adaptive value in L. bonariensis commencing reproductive development as soon as temperatures permit in the early spring. This would optimise the chances of the still pre-reproductive L. bonariensis populations locating sparsely distributed host plants with conspecific populations required for feeding and reproduction. Arguably, with L. bonariensis, this requirement could possibly be effected by the damaged plant’s emission of phenylacetaldehyde ( Figure 4 ). As discussed above, such ecology is very reminiscent of the phenylacetaldehyde-stimulated L. oryzophilus spring migration into rice crops at a time of ovarian development after which responsiveness declines in the summer generation (Saito et al., 2005; Zhang et al., 2023).

Listronotus bonariensis is known to be able to disperse rapidly into resown pastures in New Zealand (e.g., Prestidge et al., 1989; Popay et al., 2011) in a manner similar to the observation of Justas and Long (2019), working with L. oregonensis using baited traps and (2011) observed movement of L. maculicollis towards mechanically damaged turfgrasses (McGraw et al., 2011). While the situation is different in New Zealand, it is also interesting to note that after diapause in the spring (even with an obvious abundance of host plants available), those L. bonariensis with developed flight musculature were 20 times more likely to fly than at any other time of the year (Goldson et al., 1999).

The cessation of the damaged plants’ attractiveness when 20 weevils were added ( Table 2 ) is again consistent with the relict behaviour hypothesis, whereby the species needs to avoid the overexploitation of the widely distributed plant hosts. Typically, L. bonariensis larval mining of one larva destroys 3 - 8 ryegrass tillers during instar development (Ferguson et al., 2019). This limitation of conspecific recruitment is reminiscent of the findings of Tansey et al. (2005) and Ballhorn et al. (2013), where insect responses to HIPVs varied according to their intensity.

This study has demonstrated unequivocally that diapausing L. bonariensis is significantly attracted to conspecific frass and feeding damage on L. multiflorum foliage. This may, in part, be the result of the only identified de novo synthesised HIPV, phenylacetaldehyde, which initially attracted the weevils to the damaged plants. Once established on their host plants, the weevils had become reproductive and ceased to respond to the phenylacetaldehyde. These characteristics were very similar to those observed in the closely-related L. oryzophilus (Saito et al., 2005; Zhang et al., 2023) in rice fields.

That the loss of attractiveness to weevil feeding damage and frass when 20 L. bonariensis were added to the test sample can be interpreted to be a relict response to prevent the over exploitation of a scarcely distributed host plants.

Taken together, the results of this olfactometric study strongly suggests L. bonariensis relict behaviour reflecting the species adaptation to the short growing season and sparsely distributed host plants in the species’ mallines native range. Such responses by diapausing L. bonariensis are in marked contrast to the minimal response shown by reproductive weevils.