Adoption of high-density pear orchards during the last decades has resulted in a significant improvement in yield and fruit quality. However, such high-density orchards can only be maintained by the use of dwarfing rootstocks and appropriate growth control techniques ( Maas, 2005 ). In addition, full production is often achie­ved many years after orchard establishment and remains one of the main challenges in pear production ( Sansavini et al. , 2007 ). While a positive correlation exists among yield, light interception and tree density ( Palmer et al. , 1992 ), great vigor means diminution in light penetration, yield and fruit quality, and an increase in the cost of pruning and pest control ( Miller, 1995 ). Jauset et al. (2000) reported that the greatest incidence of pear psylla ( Cacopsylla pyri L., Homoptera: Psyllidae) tends to be associated with cultivars with the greatest vegetative development and highest concentrations of leaf nitrogen. Growth control in pear orchards is essential, not just to reduce competition between vegetative and reproductive sinks, but also to improve return bloom and select the best wood for production ( Costa, 2017 ).

Rootstocks are crucial for tree establishment but also to make trees more manageable through vigor control ( Sansavini & Musacchi, 2002 ). Pear orchards in North America are mostly planted on Pyrus seedling rootstocks, as quince ( Cydonia oblonga Mill.) rootstocks routinely suffer from winter damage, fire blight ( Erwinia amylovora Burill) infections and pear decline ( Westwood & Lombard, 1983 ; Lind et al. , 2003 ; Mitcham & Elkins, 2007 ; Robinson, 2011 ). On the other hand, 'Quince Adams', 'MC', and 'Sydo' are the most used rootstocks in Europe ( Deckers, 1992 ; Sansavini & Musacchi, 2002 ; Musacchi, 2008 ; Vercammen, 2014 ).

Italy and Spain are the most important pear producing countries in Europe ( Deckers & Schoofs, 2008 ). While 'Conference' is the second most important cultivar grown in Italy ( Deckers & Schoofs, 2008 ), it is the most important cultivar grown in Spain ( Iglesias & Casals, 2013 ), and in Northern Europe, with 80% of the acreage in the Netherlands ( Heijerman et al. , 2015 ), and 85% in Belgium ( Vercammen, 2014 ). 'Blanquilla' used to be the most important pear cultivar in Spain, but the great vigor of this cultivar and the problem to manage it after the chemical growth retardant chlormequat (CCC) got banned in 2001 challenged its viability. In other areas such as the Netherlands and Belgium, growers started to use mechanical methods like root pruning and trunk incisions to manage growth control on 'Conference' and 'Doyenné du Comice' ( Maas, 2005 ; Vercammen et al. , 2005 ).

Root pruning, by disturbing the shoot to root balance and limiting the gibberellin (GA) activity Saure, 2007 ), has been reported to reduce growth in apple ( Schupp & Ferree, 1988 ; Mitre et al. , 2012 ), grape ( Lee & Kang, 1997 ; Ferree et al. , 1999 ), and pears ( Vercammen et al. , 2005 ; Asin et al. , 2007 ; Maas, 2008 ; Janssens et al. , 2011 ). However, its use by fruit growers around the world is limited despite several promising reports ( Saure, 2007 ). One of the reasons to explain this scarce use is the imprecision of this technique, mostly due to the genetic diversity of the rootstocks, and the heterogeneity of environmental conditions ( Miller & Tworkoski, 2003 ). In addition, Schupp & Ferree (1988) reported the need to define pruning depth and distance from the trunk in order to optimize its performance.

Prohexadione-Ca (ProCa), a plant growth regulator that is absorbed by foliage and inhibits the biosynthesis of growth active gibberellins (GAs) ( Evans et al. , 1999 ), is one of the newest and most interesting growth retardant authorized for apple and pear ( Costa, 2017 ). Paclobutrazol, another plant growth regulator that inhibits GAs biosynthesis in the sub-apical meristem ( Hedden & Graebe, 1985 ), has been used in apple and pear as well ( Raese & Burts, 1983 ; Greene, 1986 ; Dheim & Browning, 1988 ; Sansavini et al. , 1988 ; Greene, 1991 ; Costa et al. , 1995 ). However, paclobutrazol has not been used in Italy until three years ago, when its authorization for pear orchards was requested and completed after most of the chemical plant regulators were banned ( Costa, 2017 ). Use of paclobutrazol, ProCa, deficit irrigation, root pruning, and summer pruning were studied during three years in a pear orchard ( Asin et al. , 2007 ). Nevertheless, all these strategies alone did not provide all of the CCC benefits at the same time, but each one contributed to growth control, return bloom, fruit set, and yield. Therefore, it seems than the goal to optimize growth control, return bloom and yield must be pursued with the integration or combination of several strategies. However, there is scarcity of studies where both applications of paclobutrazol and ProCa are compared together and in combination with root pruning.

The goal of this study was to assess the use of root pruning, paclobutrazol, and ProCa either alone or in combination to control growth and improve productivity on pear trees.

Final number of shoots per tree, or even the TCSA were barely affected by the treatments. On the other hand, all the treatments that were tested in this study had a positive effect on tree growth control through the reduction of shoot length. However, that reduction on shoot length or percentage of shoots shorter than either 60 cm or 40 cm differed depending on the treatment. The final pruning weight tended to be lower (~50% less) when root pruning was combined with either paclobutrazol or ProCa, whereas the highest values were for Control trees and ProCa. In terms of shoot length, root pruning alone had the least effect, and in some cases, there were not even significant differences with control trees. This suggests that the reduction of the root system caused by the root pruning was not enough to imply a significant growth control. Root pruning has been reported to limit the carbohydrate supply, creating a lack of balance between reproductive organs ( Khan et al. , 1998 ), and with stronger effect if it is performed closer to the trunk ( Schupp & Ferree, 1988 ). However, growth control was unexpectedly lower in this study than in a previous one ( Asin et al. , 2007 ), where root pruning was done farther from the trunk (30 cm vs 40 cm). In both experiments, pear trees presented similar growth, equal training system, rootstock, and orchard management. However, the current study was performed in a less fertile stony soil, suggesting that the absorbing roots were even more concentrated in the drip irrigation volume; thus, root pruning within a 10 cm difference was not enough to have a significant impact. Therefore, it is key to know where the main absorbing roots are located in order to estimate the percentage of absorbing roots that would be affected by the pruning. In addition, timing of root pruning can also affect vigor response. In our study, root pruning was performed in late winter, whereas higher vigor control would be attained if performed at bloom or later on the season. However, this was discarded in our conditions in order to not affect fruit set, already a limiting factor for pear growing in our area.

Applications of ProCa alone provided a growth reduction of ~18% in comparison to control trees. In previous studies, both root pruning and ProCa have been reported to reduce growth ( Costa et al. , 2001 ; Elfving et al. , 2003 ; Maas, 2005 ; Smit et al. , 2005 ; Vercammen et al. , 2005 ; Asin et al. , 2007 ; Mitre et al. , 2012 ). Deckers et al. (2005) reported a strong growth control of ProCa on 'Doyenné du Comice', whereas the same treatment had a slight effect on 'Beurré Alexander Lucas'. These different responses may be due to different vegetative habits and vigor inherited by the cultivar, which may imply different behaviors. In our case, applications of ProCa on 'Blanquilla' had a good effect on growth control, but the effect of paclobutrazol was stronger. In a previous study, we observed shorter shoots when applying paclobutrazol in comparison with ProCa, but the effect was quicker when ProCa was applied ( Asin & Vilardell, 2008 ). In addition, efficacy of paclobutrazol on growth control has been reported by Raese & Burts (1983) and Costa et al. (1995) in 'd'Anjou' and 'Blanquilla' pear orchards, respectively. However, there is scarcity of studies where both applications of paclobutrazol and ProCa are compared. Regarding that, Mouco et al. (2011) reported better growth control of paclobutrazol than ProCa in mango ( Magnifera indica L.). In our study, combination of root pruning plus ProCa had a similar effect on growth control than paclobutrazol alone, whereas the combination of paclobutrazol plus root pruning provided the highest control, with ~86% or ~71% of the shoots shorter than 60 cm or 40 cm, respectively. This is the first time that we study the combination of root pruning with either paclobutrazol or ProCa. Carra et al. (2017) reported 69% reduction in shoot growth when ProCa was combined with root pruning at one side of the trees; however, no significant differences were observed when ProCa was applied alone.

Furthermore than growth control, we also assessed how the combination of these different strategies may affect yield and fruit caliper. Overall, there were no significant differences with control trees regarding yield, crop load or yield efficiency, but in some years trees that were treated with ProCa had less values than when the combination of root pruning plus paclobutrazol was used. In addition, there was one year in which trees treated with paclobutrazol had larger fruits than control trees or than when root pruning or ProCa alone were used. Carra et al. (2017) reported lower fruit weight when ProCa was applied together with root pruning, but no significant differences with control trees were reported when was applied alone. Conversely, Smit et al. (2005) reported smaller fruit size when ProCa was applied, most likely due to increased fruit set. No significant differences regarding fruit set were observed in our current study. In a previous study, no significant differences for ProCa vs control trees regarding fruit weight were observed, whereas in some years paclobutrazol sprays provided larger fruits than either control, root pruning, or ProCa alone ( Asin & Vilardell, 2008 ).

In terms of fruit quality, none of the assessed strategies significantly affected either fruit firmness or soluble solids. Coinciding with our results, Southwick et al. (2004) and Maas (2008) did not see any effect of Proheaxdione-Ca or root pruning on fruit quality. Applications of paclobutrazol after petal fall have been suggested to improve fruit quality in apple ( Greene, 1991 ). Elfving et al. (2003) reported ProCa to slightly affect fruit quality in 'Barlet' pears, but those changes disappeared after two months of air storage. Byers & Yoder (1999) reported very little if any effect of ProCa on fruit firmness, soluble solids, starch content or even cracking on apple trees.

There were differences regarding return bloom depending on the treatment. When comparing with control trees, root pruning did not affect return bloom. Application of paclobutrazol and paclobutrazol plus root pruning tended to have higher return bloom than control trees; however, no significant differences were observed. On the other hand, applications of ProCa did reduce return bloom in some years when comparing with control trees, and differences with root pruning plus paclobutrazol were always significant. Coinciding with our results, reduction of return bloom in pear orchards by application of ProCa has been reported in other studies ( Sugar et al. , 2002 ; Gomand, 2003 ; Warnier, 2003 ; Asin et al. , 2005 ), and it seems to be related to the total amount of ProCa applied during the season in relation to the tree vigor. Risk to reduce re­turn bloom seems to be related to high-dose application ( Sugar et al. , 2002 ; Asin et al. , 2005 ; Deckers et al. , 2005 ). Nevertheless, there are some studies on pear in which no negative effects on return bloom were reported ( Southwick et al. , 2004 ; Asin et al. , 2007 ). Furthermore, the trend that we observed that paclobutrazol increased return bloom has been reported in previous studies ( Sansavini et al. , 1988 ; Rai & Bist, 1992 ; Asin et al. , 2007 ).

Combination of strategies for growth control is an interesting option in situations of vigorous orchards and/or cultivars. Williamson & Coston (1990) assessed the combination of root restriction and different levels of irrigation on a high-density peach orchard. Similar approaches have been performed in vineyards with the objective to enhance regulated deficit irrigation by combining it with cover crops ( Lopes et al. , 2011 ). Tworkoski & Glenn (2010) assessed grass competition combined with pruning methods to evaluate their effects on growth and yield in a peach orchard. Maas (2008) studied the combination of root pruning and trunk incisions with applications of ProCa or Ethrel on a vigorous 'Conference' pear orchard. More recently, Carra et al. (2017) tested different configurations of root pruning in combination with ProCa.

In a three-year study, we assessed performance of strategies that included root pruning, application of either ProCa or paclobutrazol, and combinations of them. Overall, all the different strategies that were assessed improved growth control in pear trees, with different grade depending on the strategy. To get satisfactory results by solely performing root pruning, other than timing (winter vs full bloom for instance) understanding of the soil characteristics and how this affects root distribution is key. We did not assess different root pruning configurations in this experiment, but similar results were attained when comparing with previous studies where root pruning was performed closer to the trunk. In situations where the absorbing roots are very close to the irrigation volume, little differences were observed when cutting 40 cm to 30 cm from the trunk. Solely applications of ProCa reduced growth in comparison to control trees, with similar results than the root pruning. Thus, use of ProCa seems to be a good alternative to root pruning, with similar growth control and increased yield (>60 mm), and with less damage to the roots. Then, paclobutrazol alone or the combination of ProCa with root pruning increased the growth control, and the highest reductions were observed when paclobutrazol was applied in combination with root pruning. In addition, yield, crop load, yield efficiency, fruit weight and return bloom seemed to be more affected by applications of ProCa than paclobutrazol. Therefore, use of paclobutrazol either alone or in combination with root pruning, depending on the desired grade of growth control, seemed to be most suitable for situations of high-vigor cultivars, with trouble on their management. The fact that use of paclobutrazol may be challenged again in the future, leave combinations of root pruning plus ProCa as the best shot for vigorous cultivars. In other situations of medium-low vigor, ProCa alone would be the best strategy. Further research should address relations between the amount of absorbing roots that are cut and the affection it has, especially in combination with paclobutrazol and ProCa. In addition, timing of root pruning may play an important role not only on vigor control but on vegetative response and fruit set, hence, they should be addressed as well.