MECHANICAL BEHAVIOUR OF GRAPEVINE WOOD AFFECTED BY Xylotrechus arvicola

The cerambycid insect Xylotrechus arvicola is considered a pest that affects the wood of the grapevine ( Vitis vinifera ) in the major wine areas of the Iberian Peninsula. The larva of this insect perforates the grape - vine wood, resulting in structural and biomechanical failure of the vine plants. Vine samples from wood damaged by X. arvicola larvae were picked up from different vineyards and grape varieties. Compressive and flexural tests were performed in order to assess the mechanical behaviour of the wood samples. Total length of the cracks in wood samples (TLCWS) that appeared on the surface of the grapevine wood samples after the mechanical tests was measured. Compressive strength (CS) and flexural strength (FS) decreased with the increase of the cross-sectional area (CSA) of both branches and trunks, regardless


INTRODUCTION
Insects are one of the most important causes of damage in woody crops (Sen et al. 2017), this resulting in significant economic losses (Ssemaganda et al. 2011). Wood boring beetles cause irreversible damage to forests, crops, and timber (Visitpanich 1994, Aukema et al. 2011. Woody plant species, like grapevines, are sensitive to being attacked by wood boring insects. There are insects among the European species, both polyphagous and monophagous, attacking different woody plant species (Svacha andDanilevsky 1988, Sama 2002).
The damage to grapevines wood is caused by larvae which perforate grapevine plants to feed on wood, making galleries within the plant for two years (Moreno 2005). Previous studies carried out with wood samples simulating the load conditions that vine wood bears on the field have shown that the vine wood affected by larvae is more sensitive to breakage and it breaks faster than undamaged wood (Rodríguez-González et al. 2019, Rodríguez-González et al. 2020. Other indirect damage is produced by adults emerging from holes, which are a direct infection access points for fungal diseases such us Diplodia seriata (De Not), Eutypa lata (Tul and Tul) or Phaeoacremonium minimun (Gams, Crous, Wingf., Mugnai) (García-Benavides et al. 2013). ´Tempranillo´ and ´Cabernet-Sauvignon´ varieties have a greater sensitivity to being attacked by X. arvicola (Ocete et al. 2002a, García-Benavides et al. 2013. In an affected vineyard, numerous broken branches could be seen due to a weakened wood structure caused by the galleries bored by larvae (García-Ruiz 2009). The only current cultural techniques available to control X. arvicola consist of removing the rhytidome of the grapevines (Peláez et al. 2006) or/and pruning branches below the affected area (Ocete et al. 2004), but these techniques are expensive and not sustainable (Peláez et al. 2006).
The present study would allow to estimate wood resistance under natural loads or/and under physical events that affect vineyards (wind, grape bearing, mechanical harvest) and the reduction in wood resistance of a grapevine affected by X. arvicola larvae.
Therefore, the objective of this study was to examine the mechanical behaviour of grapevine wood from trunks and branches affected by X. arvicola larvae using wood samples with different section area and from different V. vinifera varieties. Also, the total length of the cracks that appeared in the wood samples (TLCWS) after performing the mechanical tests was measured.

Grapevine wood samples
Grapevine wood samples (trunks and branches) of all three varieties infested with X. arvicola larvae or uninfested samples were randomly chosen to test the mechanical performance of wood.
Samples grouped by variety were then sorted into damaged or undamaged material according to external damage (galleries in pruning cuts and/or exit holes of adults and larval galleries on ends of samples) as described by Peláez et al. (2006).
Samples were collected from three severe pruned vineyards to rebuild the plant structure and avoid X. arvicola spreading, in 2017 and 2018 seasons. Two vineyards, located in 'Ribera Del Duero' (Peñafiel; Valladolid) and 'León' (Gordoncillo; León), both Protected Designation of Origin (PDO) in Castilla y Leon, Spain, were used for sampling. Sampled varieties in Peñafiel were 30-year-old ´Tempranillo´ and 32-year-old ´Cabernet-Sauvignon´ growing in loam-sandy soils. The sampled variety in Gordoncillo was 27-year-old ´Prieto Picudo´ growing in clay-loam texture soil. The vines in both locations, were spaced 3 m × 1,5 m. The vines trained into ´Trellis´ system, were consisted of two branches (1,0 m length each) and a trunk (0,7 m height).

Experimental conditions before mechanical tests were performed
Before mechanical tests Structural Round Timber Test Methods (EN 14251-2003), were used to choose only samples with the appropriate measurements to be analysed. Measurements from the diameter and length of both branches and trunks were taken for all the samples according to the EN 14251-2003. Rest of samples that does not fall in testing category according to EN 14251-2003 were discarded.
Eleven wood samples of ´Tempranillo´ and eight wood samples of ´Cabernet-Sauvignon´ varieties were evaluated when they were green so that water content of the samples was close to the one in the vineyard. However, the wood of the ´Prieto Picudo´ variety, could not be evaluated in fresh conditions because when collected it was dry. Moisture contents of the selected samples were determined as described in EN 14251-2003. Samples were dried to remove moisture from wood according to method described in Rodríguez-González et al. 2019 andRodríguez-González et al. 2020. Ten samples of ´Tempranillo´ and ´Cabernet-Sauvignon´ varieties were dried and moisture contents were measured using Equation 1.

Mechanical strength of grapevine wood samples
Two standard strength experiments: compressive tests for wood trunks and flexural tests for branches were performed in order to analyse the mechanical behaviour of grapevine wood, from selected V. vinifera varieties, damaged by X. arvicola larvae.
All grapevine samples, undamaged or damaged, fresh or dry, were tested with a hydraulic press (Figure 2a) according to the methodology described by Rodríguez-González et al. 2019, Rodríguez-González et al. 2020.

Test 1: Compressive strength (CS) of grapevine trunks in relation to cross sectional area (CSA)
Dimensions of trunks samples (fresh and dry) were measured before testing by adopting method described by Rodríguez-González et al. 2019, Rodríguez-González et al. 2020.
The existence of galleries and the number of adult exit holes on both ends of the sample were recorded for the damaged wood trunks samples.
Vertically placed trunks were used to simulate the compressive strength suffered by the vine trunks in field conditions, with both end surfaces cut perpendicularly to the axis of the sample (Figure 1a).

Test 2: Flexural strength (FS) of grapevine wood branches in relation to cross sectional area (CSA)
Before the mechanical strength was performed, different dimensions were measured in every of the branch samples (fresh and dry) according to the methodology described by Rodríguez-González et al. 2019, Rodríguez-González et al. 2020.
The existence of galleries and the number of adult exit holes on both ends of the sample were recorded for the damaged wood branches samples.
The branches, placed horizontally, rested in two roller supports 30 cm apart. Two concentrated loads were applied from the topside of the wood sample (four point bending test), to simulate the downward bending suffered by the vine branches in field conditions ( Figure 1b).

Test 3: TLCWS on wood branches and trunks
The cracks appearing on the surface of the grapevine wood samples after both compressive and flexural tests were measured from the fracture point. In order to make the cracks visible, talc powder was applied to the area and later on the excess of powder was shaken off. The cracks were then measured using a ruler and their lengths were added cumulatively to obtain the value of total length of the cracks in wood samples (TLCWS) (in mm) according to the methodology described by Persad et al. (2013).

Statistical analysis
In order to examine the effect of the cross sectional area (CSA) of grapevine wood samples (fixed factor) on compressive strength (CS) or flexural strength (FS) as a covariate an analysis of covariance (ANCOVA) was performed. The linear regression coefficients of the interaction CS (or FS) x CSA and FS x CSA were tested using an F-test. Data regarding the total length of the cracks of wood samples (TLCWS) were evaluated using analysis of variance (ANOVA) followed by a Fisher´s LSD test (significance at p ≤ 0,05). Analyses were conducted using SPSS software, version 24 software (IBM SPSS Statistics, 1968, Armonk, NY, USA).

Test 1: Mechanical strength of wood trunks (CS in relation to CSA)
Water content in fresh grapevine wood trunks of ´Tempranillo´ and ´Cabernet-Sauvignon´ varieties turned out to be 63,8 % and 57,0 %, respectively.
Up to a maximum of 6 exit holes/sample (Tempranillo) and 6 larvae galleries/sample (Cabernet-Sauvignon) were found in fresh damaged samples. Meanwhile, up to a maximum of 5 exit holes/sample (Tempranillo) and 2 larvae galleries/sample (Cabernet-Sauvignon) were found in dry damaged samples (Table 1).   Line´ is the trendline for ´Red Points´.

Mechanical strength of wood branches (FS in relation to CSA)
Water content in fresh grapevine wood branches of ´Tempranillo´ and ´Cabernet-Sauvignon´ varieties was 62,5 % and 59,6 %, respectively.
Up to a maximum of 3 exit holes/sample (Tempranillo) and 2 larvae galleries/sample (Tempranillo) were found in fresh damaged samples. Meanwhile, up to a maximum of 3 exit holes/sample (Tempranillo) and 3 larvae galleries/sample (Cabernet-Sauvignon) were found in dry damaged samples (Table 2).  The FS supported by undamaged fresh grapevine branches as a function of the CSA was not significantly different between undamaged and damaged wood branches. The linear regression coefficients of the FS x CSA interaction were significantly different (F=34,21; d.f.=1,17; P≤0,001) between undamaged and damaged grapevine. The FS in branches decreased with the increase in the CSA in both trunks (undamaged and damaged). Moreover, the FS values were lower in damaged wood than in undamaged wood (Figure 4a).
The FS supported by undamaged dry branches as function of the CSA was significantly high (F=5,455; d.f.=1,20; P=0,031) in comparison to damaged wood. The linear regression coefficients of the FS x CSA interaction were not significantly different between undamaged and damaged branches. For undamaged wood, the FS of branches decreased with the increase in the CSA, while for the damaged branches, the FS increased with the increase in the CSA (Figure 4b). Grapevine wood branches are usually subjected to severe pruning during the entire growth, which causes a non-uniform growth of the wood fibers. This promotes the appearance of knots in the wood, which directly affect the strength and resistance of the wood. This study shows how X. arvicola larvae negatively affect the mechanical behaviour of V. vinifera wood. These damages caused a decrease in the resistance of grapevine wood greater as its CSA increased and a longest TLCWS (regardless of the wood is fresh or dry, or if it subjected to compressive or flexural test´s). This phenomenon in woody species, was also observed when X. arvicola larvae attack Prunus pisardi Carrière, Koehne (Rosales: Rosaceae) trees, resulting in the weakness of the affected trees and/or the death or breakage of the affected branches during several years (Biurrun et al. 2007).
CS and FS decreased with the increase in the CSA irrespective of the part of the plant (branches and trunks), damage condition (undamaged and damaged) and water content (fresh and dry), significantly more accentuated in damaged than in undamaged wood. It was also verified through the linear regression coefficients of the CS x CSA interaction that significant differences between undamaged and damaged grapevine wood samples showed up in fresh trunks, dry trunks and fresh branches. The grapevine wood is subjected to severe pruning during the entire growth period to build up a defined training system. However, this pruning causes a non-uniform growth of the wood fibers, promoting the appearance of knots and cracks in the wood, which will directly affect the strength and resistance of the wood. Therefore, this fact will explain the decrease in both CS and FS as the CSA rises in undamaged wood samples because older wood has been subjected to more pruning cuts throughout its life. Moreover, a greater CSA implies that a more significant volume in timber could be attacked by X. arvicola larvae. This agrees with the decrease in both CS and FS as the CSA increases in damaged branches and trunks (fresh and dry). This would explain the downward slope of the trendline for the linear regressions obtained for fresh and dry trunks, and fresh branches, all whose wood samples had CSA greater than 2000 mm 2 , as Rodríguez-González et al. (2019) already observed in wood varieties. On the other hand, dry branches with a CSA lower than 2000 mm 2 in all the samples, did show a trendline with an upward slope for the linear regressions afore mentioned. Thus, wood samples with a smaller CSA had few galleries inside the wood, so the resistance of the wood was not severely damaged. Previous studies carried out by Rodríguez-González et al. (2020) only allowed us to conclude that the vine wood damaged by these larvae had a lower resistance and a higher breaking speed than undamaged wood.
X. arvicola larvae, like the larvae of other cerambycids, altered the mechanical behaviour of the trunk and branches of grapevine varieties. The effects of a continuous infestation of the grapevines by these larvae result in greater changes in the plants, for example, leaf development becomes scarce, the shoots are not very vigorous and productive the clusters are smaller, the flowers are less numerous, diminish their length, and they come off more quickly (Ocete et al. 2002b, Moreno et al. 2004). It has been stated that branches breakage due to physical properties can reduce plant fitness because of biomass and meristem loss, showing comparable results of biomechanic effects on Tsuga canadensis (Pinales: Pinaceae) produced by Adelges tsugae (Hemiptera: Adelgidae) (Soltis et al. 2014). Similar results were obtained by the effect of Monochamus galloprovincialis (Coleoptera: Cerambycidae) and Acanthocinus aedilis (Coleoptera: Cerambycidae) on Pinus sylvestris (Jankowiak and Rossa 2007). Adults of some cerambycid species select healthy trees for oviposition, whereas other choose unhealthy ones and, subsequent larval feeding and development can kill both kinds of trees (Allison et al. 2004).
Wood damages caused by X. arvicola larvae can be direct, caused by the reduction of vascular tissues of the plant which are ingested by larvae, resulting in a damaged grapevine with a lower resistance to bear crop loads, or indirect, because of the propagation of wood diseases in already affected wood that killed the vascular tissues of the wood. Thus, X. arvicola larvae damage grapevine wood and favor the propagation of grapevine trunk diseases through the emergency holes originated by X. arvicola insects on their way out of the wood, leading to the death of plant vascular tissue (Ocete et al. 2002a, García-Ruiz 2009. The fungal attack in ´Tem-pranillo´ and ´Cabernet-Sauvignon´, is more severe than in other varieties (Ocete et al. 2002a). The fungal symbionts of cerambycid beetles are endosymbiotic fungi, and they play an important role as suppliers of enzymes for degradation of organic matter, particularly wood (Buchner 1965, Dominik and Starzyk 1989, Jones et al. 1999. The fact that wood pathogens or diseases affect the biomechanical properties of woody species is already described in other genera, such as Pseudotsuga spp. (Hansen et al. 2000) and Pinus spp. (Jankowiak andRossa 2007, Drenkhan et al. 2006). According to Hauer et al. (1993), species affected by wood diseases accumulate a higher amount of dead wood, which results in a fragility, leading to a progressive death in the affected areas. An accumulation of dead wood caused by the attack of pathogens on branches or trunks predisposes the affected species to damage or breakage when they are subjected to external agents, including snow, wind and/or static loads such as the weight of grapes the grapevine wood is exposed at the time of harvesting (Detters et al. 2008, James andKane 2008).

Test 3: TLCWS on wood branches and trunks
Trunks under the CS, undamaged fresh showed the lowest TLCWS (30,71 mm), significantly lower than in damaged fresh (46,25 mm). In contrast, undamaged dry trunks also showed a significantly lower TLCWS (64,58 mm) than damaged dry trunks (86,66 mm). Dry wood trunks showed the greatest TLCWS (64,58 mm and 86,66 mm in undamaged and damaged, respectively), with values significantly greater than the respective ones in fresh trunks samples (30,71 mm and 46,25 mm in undamaged and damaged wood, respectively) ( Table  1).
Branches under the FS, undamaged fresh branches showed the lowest TLCWS (58,08 mm), significantly lower than in damaged fresh branches (85,62 mm), whereas undamaged dry branches also showed a significantly lower TLCWS (77,91 mm) than damaged dry branches (166,50 mm). Dry wood branches showed the greatest TLCWS (77,91 mm and 166,50 mm in undamaged and damaged, respectively), with values significantly greater than the respective ones in fresh branches samples (58,08 mm and 85,62 mm in undamaged and damaged, respectively) ( Table 3).
Damaged wood trunks (fresh and dry) evaluated on several V. vinifera varieties had a greater TLCWS than the undamaged trunks. A similar trend was observed in damaged branches (fresh and dry) compared to those undamaged. Moreover, the TLCWS, within the same damage condition (undamaged or damaged) and part of the plant (trunk or branch), was significantly higher in dry samples than in fresh samples. These branches and trunks turned out to be more vulnerable to mechanical stress where the grapevines are growing, as has been described for other woody species (Soltis et al. 2014). Rodríguez-González et al. (2019) showed that damaged wood of the ´Cabernet-Sauvignon´ variety could have their structural capacity reduced up to 62 % (compared to undamaged plants of the same variety) when subjected to the usual crop loads. In V. vinifera, the weight of the grapes, and the shaking produced by harvesting machines, could be two fundamental factors conditioning the wood resistance and therefore, the structural capacity of the grapevines damaged by X. arvicola larvae. TLCWS followed similar patterns, so relationships between fresh and dry wood samples (with a different water content) and different infestation levels could be established in future research. n = number of samples; v = grapevine variety; Different lowercase letters means significant differences between undamaged and damaged grapevine wood within the same water content of the wood (fresh or dry), part of the grapevine (trunk or branch) and mechanical strength (compressive or flexural) (p≤0,05). Different capital letters means significant differences between fresh and dry wood within the same damage condition (undamaged or damaged by X. arvicola larvae), part of the grapevine (trunk or branch) and mechanical strength (compressive or flexural) (p≤0,05).

CONCLUSIONS
Grapevine wood damaged by X. arvicola suffered a reduction in mechanical strength (compressive and flexural). CS and FS in wood tested samples decreased with the increase in their CSA, irrespective of the part of the plant, condition or water content. The interaction CS (or FS) x CSA showed significant differences between undamaged and damaged grapevine wood samples. Damaged wood trunk and branches samples had a greater TLCWS than undamaged samples, whereas TLCWS was higher in dry samples than in fresh samples. These damages in the wood would result in a higher vulnerability to the mechanical stress suffered by the grapevines in the field (heavy winds, crop load and plant canopy weights), and would lead to a higher exposure to the mechanical stress applied by harvesting (vibration) or pruning (traction) machines. Larvae of this insect altered the mechanical behaviour of grapevine wood and the mechanical strength of the wood, which was negatively affected when the CSA increased. Longer TLCWS was found in damaged wood.