Pinus pinea subsp. volatiles

In control experiments before being exposed to M. galloprovincialis, nine components were detected by SPME analysis of P. pinea volatiles, Table 3. Like P. pinaster, comparable chromatographic profiles were obtained for each P. pinea tree prior to being exposed to insect damage. Isopropyl palmitate and limonene were the main volatiles detected before exposure to M. galloprovincialis, followed by hexadecanol and octadecanol. In contrast, during feeding, limonene was the most abundant volatile detected by SPME (F (8,90) = 7.32; p <0.0001), and consequently the component that showed the highest fold increase during insect feeding (128-fold increase) (F (8,90) = 7.88; p <0,0001), Fig. 3. Although there have been only a few studies of the volatiles emitted by P. pinea, limonene has been described as the main constituent of the EO or solvent extracts, from stone pine from different provenances (Rodrigues et al., 2017 and refs. cited therein). However, the two other compounds that showed the highest fold increase during M. galloprovincialis feeding, hexadecanol and octadecanol (44- and −17-fold increase, respectively), had not been reported previously. In situ sampling experiments using SPME-GC-MS with P. pinea from Italy have also shown limonene as the main volatile component in this species (Martini et al., 2010). These authors suggested that the high release rate of limonene by P. pinea may contribute to the avoidance of this tree by the sawfly Neodiprion sertifer (Geoffroy), because high concentrations of limonene were shown to repel pine sawfly females. Interestingly, while limonene was repellent at high release rates, lower release rates attracted the sawflies.

This differential response to different release rates of specific volatiles was also observed with other insects which infest conifers. For example, α-pinene was shown to be repellent to many insects at high release rates, but attractive to the same insects at lower concentrations (Annila and Hiltunen, 1977), including M. galloprovincialis (Pajares et al., 2004; Bonifácio et al., 2012). Based on this type of response of insects to volatile compounds, they may be used in lures. For instance, α-pinene is a component of the commercial lure for M. galloprovincialis (Pajares et al., 2010; Álvarez et al., 2014).

Moreover, it is also crucial to consider that the enantiomeric ratios of chiral compounds may be important in host finding, as was shown for the pine weevil (Hylobius abietis) (Wibe et al., 1998). These authors reported that the available enantiomers from seven monoterpenes (camphene, limonene, sabinene, α-pinene, β-phellandrene, β-pinene and δ-3-carene), elicited different responses in the pine weevil. Considering that the herewith assessed pine species are characterized by specific enantiomeric ratios of volatile compounds (Rodrigues et al., 2017), it would be relevant to assess if these enantiomers and their proportions elicit different responses in M. galloprovincialis.

In laboratory studies on shoot feeding, oviposition, and development of M. galloprovincialis on several pine species (Naves et al., 2006; Sanchez-Husillos et al., 2013), the beetles showed difficulties in feeding and breeding on P. pinea. In the current study, under non-choice confined conditions, the beetles were able to feed on umbrella pine, but under natural conditions, the beetle was never found on P. pinea (Rodrigues et al., 2015).

The wounds caused M. galloprovincialis feeding showed variable size, aspect, and surface area, as well as differing in the amounts of oleoresin exuded, Fig. 4–6F and G. Some of the debarked areas were just small round wounds, whereas others were more elongated with irregular contours.

In P. pinaster, on average, two wounds were found per tree, Fig. 4, with average dimensions of 11 mm × 3 mm (↨↔) and an average area of 35 mm 2. The wounds on trees from each chemotype were not significantly different in terms of the dimensions of the debarked area (F (1,70) = 0.037; p = 0.85).

In P. pinea, an average of four wounds per tree were found, Fig. 5, with average dimensions per wound of 13 mm × 3 mm (↨↔) and average area of 43 mm 2. These results can be partially explained by the bark being easily available for wounding, as P. pinea showed a lower number of branches than P. pinaster.