44 1. Introduction Although olives are grown in other world regions, such as California, Australia and Argentina, the most important production areas are found in the Mediterra- nean basin where olive finds its best growing conditions, in particular in Spain, Portugal, Italy, Greece, Albania, North Africa and the Middle East. Olive tree has had tra- ditional importance in the Mediterranean region since ancient times (Loumou and Giourga, 2003) and its culti- vation retains importance in this area for its social, envi- ronmental and economic value. Currently a large part of the total world olive plantations are found in the Medi- terranean basin. The olive tree defines the Mediterranean region within the Holarctic Kingdom (Ubaldi, 2003) and it is considered one of the most typical species of this area, where it represents a very important element in de- fining the “identity” of the rural landscape. The ecological function of rural landscapes and the promotion of multifunctional agriculture is an important topic in agricultural and agri-environmental policy with- in the European Union (Gerowitt et al., 2003). Vegetation related to olive orchards plays an important ecological function that can be efficiently used to improve the mul- tifunctional role of olive growing in the Mediterranean region (Margaris, 1980). For instance, Saavedra (1998) reports over 500 species in the olive area of Córdoba province. In a selection of plantations in western Anda- lucía, 75 plant species were recorded prior to the spring cultivation (Rodenas et al., 1977). In Greece, tradition- ally managed olive groves have been identified as impor- tant habitats that often support a rich ground flora, which may include species with habitats threatened by land-use changes (Allen et al., 2006). Surveys in several olive orchards of south-western Al- bania report more than 80 species belonging to 14 botani- cal families (Huqi et al., 2009). In Italy, Viggiani (2009) reports more than 50 species as typical of olive groves, most of them having also ethno-botanical importance. The number could be even higher considering the species that can be found along field margins, i.e. roads, stone walls and other traditional human infrastructures typical of olive orchard landscapes. The presence of a significant number of plant species in olive groves offers favourable conditions for a multi- tude of animals such as arthropod fauna, reptiles, mam- mals and birds (Beaufoy, 2000; Loumou and Giourga, 2003). This is due not only to primary production in the food chain, but also to the provision of cover and repro- duction sites (Marshall et al., 2003). Potts et al. (2006) assessed the biodiversity value of six common habitats on the Greek island of Lesvosmic. They found that man- Ground cover management strategies in an Apulian oil-producing olive grove: agronomic and ecological assessment proposals M. Fracchiolla, D. Caramia, C. Lasorella, P. Montemurro Dipartimento di Scienze Agroambientali e Territoriali, Università degli Studi di Bari, Via G. Amendola, 165/a, 70126 Bari, Italy. Key words: chopping, cover crop, herbicides, Olea europaea L., soil management, weeds. Abstract: Several studies have pointed out that ground flora in olive groves, such as in any orchard, should ideally combine adequate positive effects on the agro-environment with only marginal negative competitive effects on the olive plants. This paper reports the results of an experiment carried out in an irrigated olive orchard (cv. Leccino), located in the area of Savelletri, Puglia (southern Italy), regarding the effects of ground flora as a consequence of different man- agement techniques. An aggregate index is proposed, able to provide a comprehensive evaluation of flora from both an ecological and agronomic point of view. Four different weed control strategies were compared: A) seeding, every other year, of a cover crop (Vicia sativa L.) chopped in springtime; B) weed control using a mixture of a systemic herbicide and a residual herbicide; C) weed control using a systemic herbicide only; D) chopping. The results revealed that the differ- ent management practices largely influenced the ground cover values in each study year, but not the yield. Ground cover features, assessed both from an agronomic and ecological point of view varied in particular, as was well reflected by the applied index, which proved to easily and effectively describe the flora features in different plots. Adv. Hort. Sci., 2013 27(1-2): 44-54 Received for publication 27 February 2013 Accepted for publication 21 March 2013 45 aged olive groves had the highest diversity of bees, com- parable with natural habitats such as oak woodlands and pine forests. Natural flora in olive orchards and the related fauna are also important sources of food for many species of birds, with consequent internationally important effects related to the migration of these animals (Guzman Alva- rez, 1999). For example, in southern Italy about 6% of ol- ive groves are included in “Natura 2000 habitats” (Birds Directive 92/43/EEC). From an agronomic point of view, natural flora is often able to enhance pest control because it can be an alternative host or direct food source for ben- eficial organisms (Marshall et al., 2003; Norris, 2005). In addition, ground cover can positively affect the diversity of soil biota, improving the soil ecosystem function. This effect was shown in a trial conducted in a rain-fed olive orchard, located in south-eastern Spain by Moreno at al. (2009) in which covered soils exhibited greater bacterial biomass and diversity, as well as higher microbial functional diversity than non-covered soils. Conservative flora management can also increase CO 2 fixation and enhance the capacity of olive orchards to accumulate significant amounts of biomass and humus (Sofo et al., 2004; Palese et al., 2005). Vegetation cover also has an important function in significantly reducing soil erosion (Hernandez et al., 2005), one of the most serious and widespread envir- onmental problems in many areas of the Mediterranean region (Pastor Muñoz-Cobo and Castro, 1995) where olive groves are often located in marginal soils and on steep slopes (Gomez et al., 2003; Francia Martínez et al., 2006). However, olive tree vegetation and yield can be sig- nificantly damaged if weed flora is not correctly man- aged, especially under rain-fed conditions. In addition to competing with olive for water, nutrients and - at early crop stages - even for light, weeds may also hamper olive picking. Moreover, during summer, dead weed residues can catch fire and seriously damage olive plants in cases where the residues are abundant. Traditional soil management is based on tillage, keep- ing the soil bare of vegetation all year round. This prac- tice, in addition to undoing the potential benefits of nat- ural flora, is also labour-intensive and expensive, hence it must be considered not sustainable. Several experiments show that it is possible to obtain the same or better pro- ductive results by adopting practices, such as chemical weeding or mowing, reducing or eliminating soil tillage and maintaining weed flora density at a level that is not dangerous for olive plants. Some significant results are reported by Huqi et al. (2009) in Albania, Montemurro and Mastropirro (1995), Montemurro et al. (2002) and Toscano et al. (2004) in southern Italy, Hernandez et al. (2005) in central Spain, Pastor Muñoz-Cobo (1990; 1991) in Spain, and Kabourakis (1999) in Greece. These alternative strategies could represent a remark- able sustainable approach for the maintenance of the en- vironment both in intensive systems, mitigating the envi- ronmental impact of olive growing, and in low-intensity farming systems located in marginal areas. In these lat- ter areas, reduced tillage could reduce management costs and contribute to preventing abandonment of these groves and preserve natural and cultural resources (Duarte et al., 2008). The effects could be beneficial on a large portion of the European territory; olive groves occupy approximately 5.4 million hectares, or about 4% of the utilisable agricul- tural area (Source: European Community). Each soil management system provides different con- ditions of the growth for weeds. Tillage destroys the an- nual flora, but can also create favourable conditions for new germinations and, moreover, it benefits perennial weeds by fragmenting and scattering vegetative repro- ductive organs such as rhizomes, tubers, bulbs and sto- lons. Foliar herbicides, such as glyphosate, are able to control both annual and perennial plants, but they could exert a selection pressure on tolerant or resistant species. Residual herbicides, such as oxyfluorfen, keep the soil weed-free for a longer period of time. Generally, chemi- cal weed control can cause a simplification of the flora spectrum with fewer species that are often more prob- lematic to manage. Mowing can encourage those species that are able to re-sprout after cutting. Cover crop (living mulches) may contribute significantly to weed suppres- sion providing early soil coverage and reducing the num- ber of established weed seedlings. Research proposal Weed flora, also as a consequence of different man- agement practices, can have both positive and negative effects on olive orchards as well as on the agro-ecosys- tem. Several studies have focused specifically on the eco- logical importance of natural flora in olive groves; other experiments have been performed in order to suggest the best control practices and to reduce the negative effects of weeds. It is reasonable to suppose that ground flora in olive groves, such as in any orchard, should ideally com- bine adequate positive effects on the agro-environment with only marginal negative competitive effects on the olive plants. The objective of the current work is to report data on flora communities established as a consequence of dif- ferent management techniques and to suggest an aggre- gate index able to give a comprehensive evaluation of flora, both from the ecological and agronomic point of view. The effects on olive production and oil yield are also considered. 2. Materials and Methods The experimentation was carried out between No- vember 2005 and December 2010 in an irrigated olive orchard located in the area of Savelletri (Puglia -southern Italy) made up of 11-year-old cv. Leccino plants, vase shape trained and spaced 7 x 7 m. The soil that hosted the orchard was loamy-textured (16.9% clay - 35.8% 46 silt - 47.3% sand), with a moderate presence of shallow pebbles (7.5 to 25 cm in size). The trial involved the comparison of the following four different weed control strategies: i) Seeding, every other year, of a cover crop (Vicia sativa L.) chopped in springtime. The subsequent infestation was controlled by chopping (VE); ii) Weed control using a mixture of a glyphosate-based systemic herbicide and a residual her- bicide containing oxyfluorfen, at a rate of 1.08 and 0.12 l ha-1 respectively (GLY + OX); iii) Weed control using glyphosate only at a rate of 1.08 l ha-1 (GLY); iv) Chop- ping (TRI). Herbicides were diluted in a water volume of 400 l ha-1 and applied with a hand-pump spray bottle, equipped with flat fan nozzles. Weeds were chopped using a shred- der. The vetch was sown broadcast at the rate of 80.0 kg ha-1, burying the seed by a shallow harrowing. The different management strategies of ground flora were applied following the general principle of applica- tions so as to keep the orchard fully free of natural flora in the peak vegetative growth period, i.e. in spring-sum- mer, when the flora reached a mean height of about 10-15 cm. The dates of weeding operations for each treatment are detailed in Table 1. The other agronomic and plant protection practices were applied using the techniques commonly used in the research area. The experimental plots, covering an area of 441.00 m2 (21 x 21 m), were arranged in the field following a randomised block design, with four replicates; for flora surveys a central test area of 196.0 m2 was used including the four plants on which vegetation production surveys were conducted. Flora surveys and data processing Plot flora surveys were run in April and October of each year and in the two peak growth periods of weeds. In these surveys, made prior to the execution of scheduled weed control operations, species were divided into the following two groups: a) species distributed uniformly in the test area; b) species represented either by solitary plants or distributed in restricted patches. Afterwards, for each species of the first group, a percent ground cover value related to the reference test area was estimated. These data have been used to calculate the Specific Con- tribution (CS), dividing the ground cover of each single species by the total cover (sum of the covers attributed to each single species) and multiplying it by 100. More- over, the presence of each botanical family was obtained by summing the percent ground covers of each species belonging to it. Nomenclature refers to Pignatti (1982). To provide an estimate of the ground cover features of each treatment, an index defined as Ground Cover Qual- ity Index (GCQI) was proposed. This index is calculat- ed by the following formula: GCQI = [ ∑ i (CSi x V i ) ]/6, where V i is a total score assigned to each species having a uniform distribution in the test area, calculated by sum- ming the values assigned to the following parameters: - ability to cover the soil and protect it from erosion processes (0 = negligible; 1 = average; 2 = good); - general ability to improve/preserve the chemical and physical soil properties through biomass production, nitrogen fixation or development of bunched roots (e.g. grass plants) that increase its porosity (0 = neg- ligible; 1 = good); - competitive ability against the orchard (0 = very competitive; 1 = normal; 2 = negligible); - flammability in summer periods (0 = plant that leaves much dry biomass easily flammable; 1 = thin plant that produces little biomass potentially flam- mable or whose biomass is easily degraded prior to the warm season or that remains green in summer periods). It follows that the value of could range between 0 and 6 and that, by the indicated formula, the Spe- cific Contribution (CS) of each species, based on their morphologic and eco-physiological features, can take a weight varying between 0 and 6 times its value. The specified parameters are proposed as a general indication, based on the specific needs of the test area. This does not exclude the possibility of using other ones based on other needs related to different conditions (e.g. aesthetic contribution, ability to be intermediate hosts of predators or of hyper pests, etc.). This work presents only the flora data and the relevant analyses for 2006, 2008 and 2010, i.e. only in the pres- ence of vetch (VE treatment) and in heavy years, consid- ering olive alternate bearing. Surveys on olive plants For each of the four plants included in the test area, plants were tested for mean shoot growth recorded be- tween April and October of each year, by selecting four shoots per plant arranged along the four cardinal direc- Table 1 - Calendar of applied practices in different ground cover management strategies Strategies Practices 2005 2006 2007 2008 2009 2010 Nov April Oct April July Oct April Oct April Oct Nov April Oct 1) VE Sowing of Vicia sativa x x x Chopping x x x x x x x x 2) GLY + OX Chemical weeding x x x x x x x x x x 3) GYI Chemical weeding x x x x x x x x 4) TRI Chopping x x x x x x x x x x x x 47 tions. The data concerning each plot was thus obtained as an average of 16 values (four shoots per plant x four plants). At the beginning and end of the trial, the trunk diameter of each plant included in the plot area was also measured, and the mean growth occurred in that period was obtained by difference. Olive harvesting was carried out in alternate years, i.e. in December 2006, 2008 and 2010. The fruits produced by the four plants of the test area were weighed. The oil yield was measured on a randomly chosen 2 kg sample from all olives harvested in each plot. The applied pro- cedure complied with the guidelines of Annex XV of the EC Reg. No 2568/1991. All data were submitted to variance analysis and the means were compared using Duncan’s test. Climate pattern Figures 1 and 2 show the climate pattern observed during the experiment. Each year, in accordance with the climate pattern of the test area, the hottest months were June, July and August, whereas the coldest ones were December, January and February. The highest posi- tive deviations (from +2.0 to +3.6°C) were recorded for the mean temperatures of January, June, July and August of 2007, January 2008 and November 2010; the high- est negative difference was observed in February 2009 (-2.0°C compared to the plurennial mean). As to rainfall, the rainiest years were 2009 and 2010 with values ex- ceeding the plurennial mean of the area (577.2 mm) by 243.5 mm and 162.1 mm, respectively. In those years, the months that deviated most from the average were Janu- ary (+110.7 mm), March (+57.5 mm) and October (+64.6 mm) in 2009, and May (+72.9 mm) and October (+147.3 mm) in 2010. 3. Results Table 2 lists the species found during the experiment. A total of 60 were identified; only 34 had a uniform distri- bution in the plots. The results obtained with regard to the uniformly distributed species are addressed in this section. Spring flora surveys In 2006 (Table 3), the statistically lowest total ground cover values were observed in treatments GLY and GLY+OX with 35.6 and 33.9%, respectively, followed by vetch-sown plots with 91.5%. The highest mean num- ber of species, equal to 23.0, was recorded in vetch-sown areas, whereas the highest number of families (10.0) was observed in the plots subjected to chopping. Table 4 re- veals that in the TRI treatment the statistically highest mean ground cover values were found for Gramineae, Compositae and Leguminsae, equal to 113.9-12.8 and 12.5% respectively, whereas the lowest values were re- corded in treatment GLY (10.2% Graminae and 0.4% Leguminosae), GLY+OX (13.3% Graminae, 5.7% Com- positae and 0.1% Leguminosae) and VE (6.1% Composi- tae). As to single species (Table 5), the highest specific contributions were calculated in chemically weeded plots for Malva sylvestris L. with 16.7% in treatment GLY+OX and 12.8% in GLY; these values were significantly higher than those observed in the plots of treatments VE and TRI. For Avena sterilis L., Bromus sterilis L. and Lolium rigidum Gaudin, in chopped or vetch-sown plots higher specific contributions were observed than in chemically weeded plots. On the contrary, the specific contribution of Hordeum murinum L. and Setaria verticillata (L.) Beauv. was significantly higher in the GLY+OX treat- ment. With regard to the Ground Cover Quality Index in treatments TRI and VE, the observed values (52.4 and 55.1 respectively) were shown to be statistically higher than those calculated for chemically weeded treatments (Table 3). In 2008, the highest values of total infestation and mean number of species (Table 3) were found in chopped (83.7% - 18.0) and vetch-sown plots (85.3% - 18.0), whereas the number of families was lower in treatment VE. The family Gramineae had a ground cover equal to 63.8% in treatment TRI, which is statistically higher than Fig. 1 - Mean monthly temperatures recorded during the trial and pluriannual means (1951-2001). Fig. 2 - Monthly rainfall recorded during the trial and pluriannual means (1951-2001). 48 Table 2 - Species found in experimental plots (z) Treatments VE GLY+OX GLY TRI 2006 2008 2010 2006 2008 2010 2006 2008 2010 2006 2008 2010 Adonis aestivalis L. + Anagallis arvensis L. X X X X X X Anthemis arvensis L. X X X X X Arum italicum Miller. + + + Asparagus acutifolius L. + + + + + + + + + + + + Asphodelus fistulosus L. + Aster squamatum (Sprengel) Hieron. + + + + + + + + + + Avena sterilis L. X X X X X X X X X X X X Bellardia trixago (L.) Ali. X X X X X X X X Briza maxima L. + Bromus sterilis L. X X X X X X X X X X X X Calendula arvensis L. X X X X X X X X Capsella bursa-pastoris (L.) Medicus X X X X X Catapodium rigidum (L.) Hubbard + + + + Cerinthe major L. + Chrysanthemum segetum L. X X X X X X X X X X X X Convolvulus arvensis L. X X X X X X Conyza canadensis (L.) Cronq. X X X X X X X X X X X Cynodon dactylon (L.) Pers. X X X X X X X X X X Digitaria sanguinalis (L.) Scop. + + + + Diplotaxis erucoides (L.) DC. X X X X X X X X X X X X Diplotaxis muralis (L.) DC. X X X Dittrichia viscosa Greuter. + + Erodium malacoides (L.) L’Hér. X X X X X X X X X X X X Euphorbia chamaesyce L. + + + + + + + Galactites tomentosa Moench. + + + + Geranium molle L. + + + + + Heliotropium erupaeum L. X X X X X X X X Hippocrepis unisiliquosa L. + Hordeum murinum L. X X X X X X X X X X X X Lactuca serriola L. + Lamium purpureum L. X Lolium rigidum Gaudin X X X X X X X X X X X X Lotus ornithopodioides L. + + Malva sylvestris L. X X X X X X X X X X X X Medicago hispida Gaertner X X X X X X X X X X X X Melilotus indica (L.) All. + Mercurialis annua L. + + Muscari neglectum Guss. + + + + Ononis natrix L. + + + + Oxalis pes-caprae L. X X X X X X X X X X X Papaver rhoeas L. X X X X X X X X X X Phalaris paradoxa L. X X X X X X X X Portulaca oleracea L. X X X X Raphanus raphanistrum Strobl. + + + + + Scorpiurus muricatus L. X X X X Serapias sp. + + Setaria verticillata (L.) Beauv. X X X X X X Sherardia arvensis L. + + + Solanum nigrum L. X X X X Sonchus oleraceus L. X X X X X X X X X X X X Sonchus tenerrimus L. X X X X Tetragonolopus purpureum Moench. + + Trifolium campestre Shreber X X X X X X X Trifolium fragiferum L. + + + + + + Trifolium repens L. X X X X X X X X X X X X Trifolium scabrum L. X X X X X X X X X X Trifolium tomentosum L. X X X X X X X X X X Valerianella eriocarpa Desv. + + + + + + + + + + Verbascum sinuatum L. + + + + + Total species with uniform distribution (n.) 27 22 22 25 18 22 24 21 25 26 24 25 Total others (n.) 6 6 12 5 5 10 7 8 11 9 9 15 (z) X = Uniformly distributed species; + = Others: solitary plants or restricted to patchy areas. 49 Table 4 - Ground cover (%) of the botanical families found in spring surveys Weeds Crop year 2006 Crop year 2008 Crop year 2010 GLI GL+OX TRI VE GLY GLIY+OX TRI VE GLY GLY+OX TRI VE Gramineae 10.2 c 13.3 c 113.9 a 68.5 b 22.7 d 40.0 c 52.4 b 63.8 a 23.2 b 22.4 b 30.0 b 51.9 a Compositae 9.1 b 5.7 c 12.8 a 6.1 c 8.2 a 7.4 a 7.3 a 5.9 b 21.5 a 20.0 a 6.7 b 6.6 b Leguminosae 0.4 c 0.1 c 12.5 a 7.3 b 1.4 d 2.9 c 16.2 a 6.7 b 12.0 b 11.1 b 63.5 a 10.5 b Cruciferae 1.6 a 1.6 a 1.0 b 0.6 b 1.6 b 1.7 ab 0.5 c 1.9 a -- -- -- -- Primulaceae 2.6 a 0.0 c 1.9 b 0.0 c -- -- -- -- 4.5 b 0.0 d 5.7 a 1.9 c Scrofuliaraceae 0.0 b 0.0 b 2.0 a 0.1 b 0.0 0.1 0.6 a 0.0 1.9 a 0.5 b 0.5 b 0.5 b Convolvulaceae 5.1 4.9 4.9 4.7 2.3 a 2.2 a 0.0 b 0.0 b -- -- -- -- Geraniaceae -- -- -- -- 2.0 b 1.9 b 2.9 a 2.6 ab -- -- -- -- Boraginaceae -- -- -- -- 0.9 b 1.7 a 2.4 a 1.9 a -- -- -- -- Labiatae 0.0 0.5 0.0 0.0 -- -- -- -- -- -- -- -- Malvaceae 4.5 b 5.6 a 3.2 c 0.0 d 5.7 a 2.6 b 1.0 c 0.6 c 7.0 a 3.6 b 0.5 c 2.5 b Papaveraceae 0.1 c 0.3 c 1.3 b 3.0 a 0.4 b 0.0 b 0.4 b 2.0 a 4.0 a 0.5 b 0.5 b 0.0 b Portulacaceae 1.9 1.8 2.0 1.5 -- -- -- -- -- -- -- -- (z) Values that do not have a letter in common are significantly different at 0.05 P (Duncan’s test). Table 3 - Total ground cover, number of families and species, agro-ecological indices in spring surveys Ground cover management strategies (2) Crop year 2006 Crop year 2008 Crop year 2010 GLY GLY+OX TRI VE GLY GLY+OX TRI VE GLY GLIY+OX TRI VE Total ground cover (%) 35.6 c 33.9 c 155.7 a 91.5 b 45.3 c 61.5 b 83.7 a 85.3 a 74.2 b 58.2 c 107.5 a 74.5 b Botanical families (n.) 8.5 b 8.0 b 10.0 a 7.7 b 9.0 a 9.0 a 9.0 a 8.0 b 7.0 a 4.5 c 5.5 b 6.0 b Species 16.0 c 15.5 c 20.7 b 23.0 a 17.0 b 17.0 b 18.0 a 18.0 a 16.2 a 12.5 b 15.7 a 15.7 a GCQI 53.2 B 49.6 C 52.4 A 55.1 A 53.1 C 53.9 C 61.0 A 54.6 B 57.5 B 58.4 B 80.9 A 57.3 B (z) Values that do not have a letter in common are significantly different at 0.01 P (capital letter) or at 0.05 P (small letter) (Duncan’s test). the values recorded for the other strategies under con- sideration (Table 4). The Leguminosae, instead, showed the highest cover value (16.2%) in chopped plots. The data presented in Table 5 point out that the statistically highest CS value of A. sterilis was found in vetch plots VE (18.1%), whereas that of B. sterilis was higher with only chopping, where the lowest specific contribution of L. rigidum was also calculated. Within Leguminosae, the most represented species were Medicago hispida Gaert- ner, Trifolium repens L. and Trifolium tomentosum L., whose specific contributions were higher in chopped plots. Conyza canadensis (L.) Cronq. and M. sylvestris, instead, showed the significantly highest CS in the ex- perimental plots weeded only by Glyphosate. As for the Ground Cover Quality Index (Table 3), the highest sta- tistical value was calculated in chopped plots (61.0), fol- lowed by treatment VE (54.6). In 2010, the statistically highest total cover percentage (Table 3) was observed in the chopped plot (107.5%), fol- lowed by vetch plots (74.5%). The lowest mean number of families and species (4.5 and 12.5, respectively) was re- corded in OX treatment. The data included in Table 4 point out that the most significant cover value of Graminae and Leguminosae was found, respectively, in the plots with vetch (51.9%) and in TRI treatment (63.5%). As for single species (Table 5), A. sterilis and B. sterilis showed the sta- tistically highest mean values of CS in the VE treatment (21.3 and 22.0% respectively); the CS of L. rigidum and Trifolium campestre Shreber were found to be, instead, the lowest in statistical terms in treatments TRI (4.2%) and GLY+OX (0.0%). With regard to the GCQI (Table 3), the statistically highest mean value (80.9) was observed in chopped plots among all compared treatments. Autumn flora surveys In 2006, the statistically highest infestation value (Ta- ble 6) was observed in chopped plots (34.6%). The statis- tically lowest mean number of species and families, 13.0 and 9.0 respectively, was recorded in the TRI treatment. In TRI and VE treatments the mean cover values of Gramine- ae species were 22.4 and 21.5% respectively, statistically higher values compared to treatments GLY and GLY+OX (Table 7). As to specific contributions, Table 8 shows that L. rigidum is the species with the highest mean data of all the monitored species; more specifically, it accounted for 63.5% of cover in the VE treatment and 61.1% in TRI, both values being significantly higher than those observed in chemically weeded plots. The highest GCQI (Table 6), 50 equal to 54.9, was calculated in the plot submitted to chop- ping; this data is statistically different from that observed in the other treatments. In 2008, the statistically highest values of total cover percentage and number of species (Table 6) were observed in chopped plots (49.6% and 13.7%, respectively). For the number of families (Table 6), which varied between 6.0 and 6.5, no sharp differences were found. The statistically highest mean cover values of Gramineae (21.8%), Legu- minosae (5.0%) and Compositae (14.5%) were recorded in the TRI treatment (Table 7). The data shown in Table 8 point out that the highest specific contributions were found for: a) A. sterilis in the treatment weeded by the mixture of Glyphosate and Oxyfluorfen (22.9%); b) B. sterilis (12.6%) and Oxalis. pes-caprae L. (14.6%) in VE; c) C. ca- nadensis (47.0%) in the plots weeded by Glyphosate only; d) Heliotropium erupaeum L. in both chemically weeded plots; and e) M. hispida in the plots submitted to chopping Table 5 - Specific contributions (%) calculated for the species found in spring surveys Weeds Crop year 2006 Crop year 2008 Crop year 2010 GLI GLI+OX TRI VE GLI GLI+OX TRI VE GLI GLI+OX TRI VE Anagallis arvensis L. 7.3 a (z) 0.0 c 1.2 b 0.0 c -- -- -- -- 6.1 a 0.0 c 5.3 ab 2.6 b Anthemis arvensis L. 0.0 b 0.0 b 1.2 a 0.0 b -- -- -- -- 4.8 a 0.9 b 0.4 b 3.3 a Avena sterilis L. 3.3 c 3.9 c 20.2 a 17.1 b 0.5 c 2.1 c 13.0 b 18.1 a 0.6 c 5.1 b 3.8 b 21.3 a Bellardia trixago (L.) Ali. 0.0 c 0.0 c 1.3 a 0.1 b 0.0 c 0.2 b 0.7 a 0.0 c 2.7 0.9 0.5 0.7 Bromus sterilis L. 3.0 c 3.9 c 19.9 a 16.5 b 23.5 c 25.6 b 29.2 a 17.9 d 9.4 c 12.7 b 14.3 b 22.0 a alendula arvensis L. -- -- -- -- 0.3 b 0.2 b 0.2 b 1.1 a -- -- -- -- Capsella bursa-pastoris (L.) Medicus 0.0 b 0.0 b 0.0 b 0.2 a 0.2 0.2 0.1 0.2 -- -- -- -- Chrysanthemum segetum L. 3.2 b 4.8 a 1.3 c 1.0 c 5.9 a 4.4 b 3.2 c 3.2 c 6.2 a 8.5 a 1.1 b 4.0 a Conyza canadensis (L.) Cronq. 9.8 a 11.0 a 2.9 b 0.5 c 3.2 a 0.4 b 0.0 b 0.0 b 12.7 b 16.2 a 0.9 c 0.9 c Convolvulus arvensis L. 14.3 a 14.7 a 3.2 b 5.1 b 5.1 a 3.7 b 0.0 c 0.0 c -- -- -- -- Cynodon dactylon (L.) Pers. 14.8 a 14.4 a 3.1 b 0.2 c 3.2 a 2.5 b 0.4 d 2.1 c -- -- -- -- Diplotaxis erucoides (L.) DC. 4.4 a 4.8 a 0.6 b 0.5 b 4.4 a 3.1 b 3.5 b 3.0 b -- -- -- -- Hordeum murinum L. 0.3 d 6.3 a 2.1 c 3.5 b 2.1 2.8 2.8 2.2 2.6 bc 0.8 c 4.6 b 5.4 a Lamium purpureum L. 0.0 b 1.3 a 0.0 b 0.0 b 1.3 c 14.0 a 13.6 ab 12.9 b -- -- -- -- Lolium rigidum Gaudin 5.6 c 6.4 c 27.1 b 34.7 a 24.9 a 24.8 a 6.9 b 25.8 a 16.7 a 17.8 a 4.2 b 20.0 a Malva sylvestris L. 12.8 b 16.7 a 2.1 c 0.0 c 12.7 a 4.3 b 1.2 c 0.7 c 9.5 a 6.3 a 0.5 b 3.4 a Medicago hispida Gaertner 0.7 b 0.1 b 1.9 a 0.0 b 1.9 d 3.9 b 6.6 a 2.9 c 4.1 7.9 5.2 2.0 Papaver rhoeas L. 0.3 b 0.9 b 0.8 b 3.3 a 1.0 b 0.0 c 0.5 bc 2.3 a 5.5 a 0.8 b 0.4 b 0.0 b Phalaris paradoxa L. 0.0 b 0.0 b 0.0 b 0.1 a 0.0 b 0.0 b 0.0 b 0.1 a 1.6 a 1.7 a 0.9 b 1.6 a Portulaca oleracea L. 5.5 a 5.3 a 1.3 b 1.7 b -- -- -- -- -- -- -- -- Setaria verticillata (L.) Beauv. 1.7 b 4.3 a 0.7 b 2.4 b -- -- -- -- -- -- -- -- Scorpiurus muricatus L. 0.0 c 0.0 c 0.2 b 1.1 a -- -- -- -- 0.7 0.0 0.4 0.0 Sonchus oleraceus L. 12.5 a 1.1 d 2.9 c 5.1 b 8.8 a 6.9 b 5.4 c 2.5 d 5.5 b 8.5 a 3.8 b 1.0 c Trifolium campestre Shreber 0.3 b 0.1 b 1.5 a 1.8 a -- -- -- -- 3.4 a 0.0 b 5.2 a 3.4 a Trifolium repens L. 0.1 b 0.1 b 1.3 a 1.4 a 1.1 c 0.8 c 5.2 a 1.8 b 0.6 c 1.2 c 11.8 a 3.4 b Trifolium scabrum L. 0.1 b 0.1 b 1.6 a 1.9 a 0.0 c 0.0 c 1.4 a 0.9 b 1.9 b 2.7 b 21.9 a 2.0 b Trifolium tomentosum L. 0.1 b 0.1 b 1.5 a 1.8 a 0.0 c 0.0 c 6.3 a 2.2 b 5.5 b 7.8 b 14.9 a 3.4 b (z) Values that do not have a letter in common are significantly different at 0.05 P (Duncan’s test). Table 6 - Total ground cover, number of families, number of species and agro-ecological indices in autumn surveys Crop year 2006 Crop year 2008 Crop year 2010 GLI GLI+OX TRI VE GLI GLI+OX TRI VE GLI GLI+OX TRI VE Total ground cover 23.4 c (z) 23.6 c 34.6 a 31.9 b 19.1 c 20.7 c 49.6 a 39.8 b 8.2 b 9.1 b 24.0 a 22.0 a N° of families 10.0 a 10.0 a 9.0 b 9.7 a 6.5 6.5 6.0 6.2 6.7 ab 5.7 bc 5.5 c 7.5 a N° of species 14.0 a 14.0 a 13.0 b 13.7 a 8.7 c 11.5 b 13.7 a 11.7 b 9.7 7.7 8.2 10.7 GCQI 53.3 b 52.7 b 54.9 a 53.6 b 42.2 c 55.7 a 54.2 a 52.6 b 58.8 62.0 58.4 56.8 (z) Values that do not have a letter in common are significantly different at 0.05 P (Duncan’s test). 51 (10.1%). The GCQI with the lowest statistical value was in the treatment weeded by Glyphosate only (42.0), whereas the values calculated in the other treatments were not sta- tistically different from each other (Table 6). In 2010, the highest total cover values were observed in TRI and VE treatments, with 24.0 and 22.0%, respectively (Table 6). The highest mean number of families (7.5) was recorded for treatment VE. As to the number of species, statistical analysis did not point out any reliable difference between the values of different strategies that ranged be- tween 7.7 for treatment GLY+OX and 10.7 for VE. The data in Table 7 show that Graminae and Compositae had a statistically higher mean cover value in TRI and VE plots, whereas the cover values of Leguminosae species did not show any remarkable difference between each other. As to single species (Table 8), the highest CS values in statisti- Table 7 - Ground cover (%) of the botanical families found in autumn surveys Weeds Crop year 2006 Crop year 2008 Crop year 2010 GLI GLI+OX TRI VE GLI GLI+OX TRI VE GLI GLI+OX TRI VE Graminaceae 11.1 b (z) 11.6 b 22.4 a 21.5 a 1.1 d 6.3 c 21.8 a 9.3 b 1.3 b 1.7 b 4.3 a 5.9 a Compositae 2.5 a 2.7 a 2.9 a 1.5 b 10.8 b 6.1 c 14.5 a 11.5 b 2.2 b 3.5 b 6.6 a 6.6 a Leguminosae 1.2 c 1.1 c 3.3 a 1.7 b 0.1 c 0.3 bc 5.0 a 0.6 b 0.9 1.0 1.5 1.2 Cruciferae 2.1 a 2.2 a 1.3 c 1.9 b 0.3 b 0.5 b 0.2 b 6.0 a 0.1 0.1 0.2 0.7 Convolvulaceae 2.6 2.4 2.4 2.3 -- -- -- -- -- -- -- -- Geraniaceae 0.5 b 0.5 b 0.5 b 1.0 a 0.0 0.0 0.0 0.1 1.0 0.2 0.4 1.0 Boraginaceae 0.6 0.6 0.8 1.0 5.2 b 5.1 b 5.1 b 6.5 a -- -- -- -- Malvaceae 1.3 a 1.0 ab 0.9 b 0.7 b 0.1 0.1 0.0 0.0 1.6 a 0.4 b 0.2 b 0.7 b Oxalidacee 1.0 a 1.0 a 0.1 b 0.0 b 1.5 d 2.2 c 2.9 b 5.8 a 0.6 d 2.1 c 10.2 a 5.5 b Primulaceae -- -- -- -- -- -- -- -- 0.3 0.3 0.5 0.4 Solanacee 0.4 a 0.5 a 0.2 b 0.2 b -- -- -- -- -- -- -- -- (z) Values that do not have a letter in common are significantly different at 0.05 P (Duncan’s test). Table 8 - Specific contributions (%) calculated for the species found in autumn surveys Weeds Crop year 2006 Crop year 2008 Crop year 2010 GLI GLI+OX TRI VE GLI GLI+OX TRI VE GLI GLI+OX TRI VE Anagallis arvensis L. -- -- -- -- -- -- -- -- 4.1 0.5 1.9 2.0 Avena sterilis L. -- -- -- -- 0.0 c (z) 22.9 a 5.0 b 3.7 b -- -- -- -- Bromus sterilis L. -- -- -- -- 0.0 c 1.3 b 11.6 a 12.6 a 2.4 c 1.5 c 13.3 b 25.0 a Calendula arvensis L. -- -- -- -- 8.9 b 1.7 c 10.6 b 16.3 a 4.3 b 1.1 b 20.1 a 16.0 a Chrysanthemum segetum -- -- -- -- -- -- -- -- 3.8 0.0 0.9 2.5 Conyza canadensis (L.) Cronq. 6.5 b 7.6 a 4.4 c 1.4 d 47.0 a 3.6 c 8.6 b 0.0 c 11.0 a 12.9 a 0.1 b 1.6 b Convolvulus arvensis L. 10.9 a 10.3 a 6.9 b 7.2 b -- -- -- -- -- -- -- -- Cynodon dactylon (L.) Pers. 5.4 a 5.3 a 3.6 b 3.9 b 5.9 a 5.3 a 2.2 b 0.4 c 3.5 0.0 1.1 0.0 Diplotaxis erucoides (L.) DC. 4.9 a 5.0 a 3.3 b 5.5 a 1.4 2.4 0.3 13.8 1.5 1.2 0.9 3.2 Diplotaxis muralis (L.) DC. 4.2 a 4.1 a 0.4 b 0.5 b 0.1 bc 0.0 c 0.2 b 1.3 a -- -- -- -- Erodium malacoides (L.) L’Hér. 2.1 b 2.0 b 1.4 c 3.1 a 0.1 0.0 0.0 0.1 12.4 a 2.5 b 1.7 b 4.6 b Heliotropium europaeum L. 2.7 2.7 2.4 3.1 27.5 a 24.5 a 10.3 c 16.4 b -- -- -- -- Hordeum murinum L. -- -- -- -- 0.0 b 0.7 b 10.1 a 0.1 b -- -- -- -- Lolium rigidum Gaudin 42.3 b 43.9 b 61.1 a 63.5 a 0.0 b 0.2 b 9.6 a 0.2 b 10.0 ab 16.5 a 3.7 bc 1.6 c Malva sylvestris L. 5.4 a 4.1 b 2.5 c 2.4 c 0.5 0.7 0.0 0.0 19.9 a 4.1 b 1.1 b 3.2 b Medicago hispida Gaertner 0.9 c 0.9 c 7.0 a 2.0 b 0.3 c 1.4 b 10.1 a 1.6 b 12.2 12.0 6.1 4.2 Oxalis pes caprae L. 4.3 a 4.1 a 0.0 b 0.3 b 7.7 c 10.9 b 5.9 c 14.6 a 8.5 c 25.0 b 42.6 a 25.8 b Setaria verticillata (L.) Beauv. -- -- -- -- 0.0 b 0.0 b 5.6 a 6.3 a -- -- -- -- Sonchus oleraceus L. -- -- -- -- 0.5 d 24.2 a 10.1 c 12.6 b 2.9 5.4 0.9 4.0 Sonchus tenerrimus L. 4.2 4.1 3.8 3.1 -- -- -- -- 3.3 b 16.9 a 5.2 b 5.2 b Solanum nigrum L. 1.9 a 2.2 a 0.6 b 0.4 b -- -- -- -- -- -- -- -- Trifolium campestre Shreber 4.3 3.7 2.5 3.5 -- -- -- -- -- -- -- -- (z) Values that do not have a letter in common are significantly different at 0.05 P (Duncan’s test). 52 cal terms were observed: a) for B. sterilis (25.0%) in the plots of treatment VE; b) for Calendula arvensis L. both in VE (16.0%) and TRI (20.1%); c) for C. canadensis in both chemically weeded plots; and d) for O. pes-caprae in the chopped treatment (42.6%).The GCQI (Table 6) ranged between 62.0 in treatments GLY+OX and 56.8 in treat- ment VE and did not show any significant differences be- tween the values of the strategies being compared. Finally, Table 2 lists the other species found during the experiment, with low cover percent values and a non-uniform distribu- tion. The occurrence of these 26 species increased during the years for all treatments, although none of them attained a uniform distribution over time. Vegetation surveys Shoot growth and trunk diameter did not show any signif- icant differences between the mean values measured for the compared treatments (Table 9) during the study three years. Olive production and oil yield As shown by the data in Table 10, no statistical differenc- es were observed between the mean recorded values during the study period for the different strategies being compared with regard to olive production and oil yield per plant. 4. Discussion and Conclusions In this work diachronic analysis was not carried out given the limited number of years under study, however synchronic comparisons have supplied data that can lead to some interesting conclusions. The different management practices employed in the study largely influenced the ground cover values in the various years, both quantitatively and qualitatively, but not the yield. In all surveys, the most represented families, both in terms of ground cover and number of species, were Graminae, Leguminosae and Compositae. With regard to single species, significant specific con- tributions were recorded for A. sterilis, B. sterilis, C. ca- nadensis, L. rigidum, M. sylvestris, O. pes-caprae and Tri- folium spp. The largest differences, in terms of total ground cover, were observed between the chemically weeded plots and plots submitted to chopping only or sown with vetch. Moreover, at the time of surveys, no important differences were found between the treatment with the systemic herbi- cide only and the plot supplied also with the residual herbi- cide; this is presumably due to the low application rate of oxyfluorfen. In particular, in spring and autumn surveys, the highest total infestation was found in chopped and vetch- sown plots. The latter, although covering the whole plot area at spring surveys, has only partially limited the growth of weeds, especially grasses (Graminae). The number of species having a uniform distribution in springtime in the plots controlled by chopping or through the sowing of the cover crop rarely exceeded the value observed under different management practices. The number calculat- ed in autumn, instead, was virtually equal for all strategies. The species that showed the highest specific contribu- tions in chemically weeded plots include C. canadensis and M. sylvestris, which might be related to the fact that those species are tolerant to the applied rates of herbi- cides or maybe, for C. canadensis, resistant to Glyphosate (Montemurro, 2008; Herbicide Resistance Action Com- mittee, 2012). In the two other conditions, in general, no single species was found to be markedly present. The most influenced families in spring surveys were Le- guminosae and Graminae; in particular, the latter seemed to be facilitated by vetch sowing, whereas the former was aided by chopping, conditions that were more evident in spring than in autumn surveys when all differences in gen- eral seemed to be less marked. Table 9 - Shoot and trunk growth measured during the trial Treatments Shoots (cm) Trunk diameter (cm) April-October 2006 April-October 2008 April-October 2010 April 2006-October 2010 VE 15.9 14.2 15.0 3.0 GLI + OX 13.5 16.6 14.0 3.5 GLI 16.4 17.3 13.4 3.9 TRI 13.1 16.6 14.8 3.1 Table 10 - Olive production per plant and oil yield Production per plant (Kg) Oil yield (% ) 2006 2008 2010 2006 2008 2010 GLI 13.6 11.7 14.8 18.0 19.1 21.6 GLI+OX 13.9 10.8 15.0 19.0 20.0 21.7 TRI 15.8 11.1 15.6 19.0 19.0 22.3 VE 12.8 10.9 15.1 17.9 20.1 20.9 53 The Ground Cover Quality Index calculated in spring 2006 and 2008 was on average higher in the chopped or vetch-sown plots. In 2010 the value calculated in chopped plots was markedly higher than in the other treatments, which, instead, did not show any differences for this pa- rameter. This would suggest a shifting, over time, of veg- etation towards a higher quality composition which was more accentuated in the case of the strategy involving chopping only. These effects are well summarized by the applied index, which indicates that for the same weed spe- cies assortment influenced the ground cover quality. Differences in the flora composition, both quantitative- ly and qualitatively, did not affect olive yield or vegetation, since weeds were however controlled during the plants’ critical periods. What varied was above all the ground cover features, assessed both from an agronomic and eco- logical point of view. From our perspective, this feature is well reflected by the applied index, which easily and effectively described the flora features in different plots. In this regard it should be said that this index is obviously influenced by the value attributed to each species that may vary in relation to the objectives of weed management, as previously mentioned. In our case emphasis was placed on competition, protection from erosion and on the capacity to preserve or even increase fertility: these features coin- cide with the objectives that olive growers normally try to achieve, especially in our areas. In other situations differ- ent parameters could be applied, namely by varying the index numerically while still keeping its functional mean- ing. Moreover, in the case under study, the GCQI value was largely influenced by the total ground cover because none of the weeds found in the trial was assigned a zero score (V i ). In the event that undesired species were found in among the cover composition, the index would certainly have been less dependent on total ground cover. Finally, since different cover crop and ground flora management practices seemed to give results, in terms of yield, that were not different from each other, a long-term approach could be applied for their selection. Currently it seems possible to prefer a ground cover management strat- egy that enables a sustainable use of olive agro-ecosystems and emphasizes the different roles of wild flora, including landscaping. This keeps in mind the fact that in 2006 the Puglia Regional Government enacted a law regarding the protection and enhancement of monumental olive trees and of the olive agro-ecosystems of its region (L.R. N. 39 del 03/10/2006) and that the location where the trial was conducted falls within the areas of highest density of an- cient and traditional olive tree landscapes. The value of this area was also further declared by its inclusion among the “High Nature Value Farmland” areas (European Envi- ronment Agency, 2004). References ALLEN H.D., RANDALL R.E., AMABLE G.S., DEVEREUX B.J., 2006 - The impact of changing olive cultivation prac- tices on the ground flora of olive groves in the messara and Psiloritis regions, Crete, Greece. - Land Degrad. Develop., 17: 249-273. BEAUFOY G., 2000 - The environmental impact of olive oil production in the European Union: practical options for improving the environmental impact. - Report produced by the European Forum on Nature Conservation and Pastoral- ism and the Asociación para el Análisis y Reforma de la Política Agro-rural. Published by European Commission, Environment Directorate-General. Available online at: http:// ec.europa.eu/environment/agriculture/pdf/oliveoil.pdf DUARTE F., JONES N., FLESKENS L., 2008 - Traditional olive orchards on sloping land: sustainability or abandon- ment? - Journal of Environmental Management, 89: 86-98. EUROPEAN ENVIRONMENT AGENCY, 2004 - High nature value farmland characteristics, trends and policy challenges. - Office for Official Publications of the European Communi- ties, Luxembourg, pp. 32. Available online at: http://www. eea.europa.eu/publications/report_2004_1. FRANCIA MARTíNEZ J.R., DURáN ZUAZO V.H., MAR- TíNEZ RAYA A., 2006 - Environmental impact from moun- tainous olive orchards under different soil-management sys- tems (SE Spain). - Science of the Total Environment, 358: 46-60. GEROWITT B., BERTKE E., HESPELT S.K., TUTE C., 2003 - Towards multifunctional agriculture - weeds as ecological goods? - Weed Research, 43(4): 227-235. GOMEZ J.A., BATTANY M., RENSCHLER C.S., FERERES E., 2003 - Evaluating the impact of soil management on soil loss in olive orchards. - Soil use and management, 19: 127- 134. GUZMAN ALVAREZ J.R., 1999 - Olivicolture ed ecologia: la situazione in Spagna. - Olivae, 78: 41-49. HERBICIDE RESISTANCE ACTION COMMITTEE, 2012 - The international survey of herbicide resistant weeds. - http://www.weedscience.com. HERNANDEZ A.J., LACASTA C., PASTOR J., 2005 - Effects of different management practices on soil conservation and soil water in a rainfed olive orchard. - Agricultural Water Management, 77: 232-248. HUQI B., DHIMA K., VASILAKOGLOU I., KECO R., SAL- AKU F., 2009 - Weed flora and weed management in estab- lished olive groves in Albania. - Weed Biology and Manage- ment, 9: 276-285. KABOURAKIS E., 1999 - Codice di pratiche per i sistemi eco- logici di produzione oleicola in Creta. - Olivae, 77: 35-45. LOUMOU A., GIOURGA C., 2003 - Olive groves: “The life and identity of the Mediterranean”. - Agriculture and Hu- man Values, 20: 87-95. MARGARIS N.S., 1980 - Structure and dynamics of Mediterra- nean type vegetation. - Portugaliae Acta Biologica, 16: 45-58. MARSHALL E.J.P., BROWN V.K., BOATMAN N.D., LUT- MAN P.J.W., SQUIRE G.R., WARD L.K., 2003 - The role of weeds in supporting biological diversity within crop fields. - Weed Research, 43: 77-89. MONTEMURRO P., 2008 - Conyza canadensis negli oliveti, un’espansione da contrastare. - Terra e Vita, 27: 68-70. MONTEMURRO P., FRACCHIOLLA M., GUARINI D., 2002 - Results of a chemical weed control trial in an olive oil or- chard. - Acta Horticulturae, 586: 397-400. 54 MONTEMURRO P., MASTROPIRRO A., 1995 - Un biennio di ricerche sulla gestione della vegetazione infestante nella col- tura dell’olivo (Olea europeaea var. sativa L.) da olio in ir- riguo in Puglia. - Atti Conv. Naz. “L’olivicoltura Mediterra- nea: Stato e prospettive della coltura e della ricerca” - Rende (CS), 26-28 January, pp. 425-433. MORENO B., GARCIA-RODRIGUEZ S., CANIZARES R., CASTRO J., BENITEZ E., 2009 - Rainfed olive farming in south-eastern Spain: Long term effect of soil management on biological indicators of soil quality. - Agriculture, Ecosys- tems and Environment, 131: 333-339. NORRIS R.F., 2005 - Ecological bases of interactions beteween weeds and organisms in other pest categories. - Weed Sci- ence, 53: 909-913. PALESE A.M., CELANO G., PETRILLO G., GRAZIANO D., XILOYANNIS C., 2005 - Gestione del suolo negli oliveti e conservazione delle risorse naturali. - L’Informatore Agrario, 38: 41-45. PASTOR MUÑOZ COBO M., 1990 - La non lavorazione e altri sistemi di lavorazione ridotta nella coltivazione dell’olivo. - Olivae, 34: 18-30. PASTOR MUÑOZ COBO M., 1991 - La non lavorazione e altri sistemi di lavorazione ridotta nella coltivazione dell’olivo (continuazione e fine). - Olivae, 35: 35-49. PASTOR MUÑOZ COBO M., CASTRO J., 1995 - Sistemi di manutenzione del suolo ed erosione. - Olivae, 59: 64-74. PIGNATTI S., 1982 - Flora d’Italia. - Ed. Agricole, Bologna. POTTS S.G., PETANIDOU T., ROBERTS S., O’TOOLE C., HULBERT A., WILLMER P., 2006 - Plant-pollinator biodi- versity and pollination services in a complex Mediterranean landscape. - Biological Conservation, 129(4): 519-529. RODENAS L.M., SANCHO R.F., RAMIREZ D.L., BERNAL- DEZ G.F., 1977 - Ecosistemas del area de influencia de Sevil- la. - Monografia 18. Doñana: Prospeccìon e inventario de Eco- sistemas, ICONA, Madrid, Spain (Cited by Beaufoy, 2000). SAAVEDRA M., 1998 - Flora del olivar y manejo de herbici- das. - Paper presented at Universidad Internacional de Anda- lucía, Baeza, Spain (Cited by Beaufoy, 2000). SOFO A., PALESE A.M., XILOYANNIS C., MONTANARO G. E MASSAI R., 2004 - Il ruolo della frutticoltura nella mitigazione dell’effetto serra. - L’informatore Agrario, 44: 27-31. TOSCANO P., BRICCOLI-BATI C., GODINO G., DE SI- MONE C., RAGLIONES M., LORENZONI P., ANGELINI R., ANTONUCCIO S., 2004 - Effetti agronomici e pedologi- ci di due diverse tecniche di gestione del suolo in un oliveto collinare del Meridione d’Italia. - Olivae, 102: 21-26. UBALDI D., 2003 - Flora, fitocenosi e ambiente: elementi di geo- botanica e Fitosociologia. - Clueb Bologna, Italy, pp. 334. VIGGIANI P., 2009 - La Flora spontanea. - In: PISANTE M., P. INGLESE, and G. LERCKER L’ulivo e l’Olio. - Coltura & Cultura. Collana ideata e diretta da Angelini Renzo. Bayer Crop Science. ART Servizi Editoriali S.p.A., Bologna, Italy, pp. 784.