Identification of the promising olive (Olea europaea L.) cultivars based on morphological and pomological characters

Abstract Olive (Olea europaea L.) is an ancient tree and can tolerate drought very well. In the present study, morphological and pomological diversity of 24 olive cultivars (5–15 replications for each cultivar, 243 trees in total) was evaluated. There were significant differences among the cultivars studied based on the characters recorded. The CV was more than 20.00% in 46 of 50 characters measured. Leaf length ranged from 27.07 to 78.54 mm, and leaf width varied from 5.42 to 23.06 mm. Ripening date ranged from late‐August to early‐October. Fruit length ranged from 13.04 to 33.72 mm, fruit diameter varied from 10.24 to 23.71 mm, fruit weighted from 0.97 to 9.61 g, and the range of fruit flesh thickness was 1.63–7.65 mm. There was high variability in terms of fruit color, ranging from light green to black. Hierarchical cluster analysis (HCA) performed based on the mean of replications with Euclidean distance and Ward method grouped the cultivars into two major clusters. Differences in many of the morphological traits were observed across the cultivars. These sets of data were used to identify unique and desirable cultivars morphologically. The present research demonstrates that local olive cultivars have unique characteristics that differentiate them from imported cultivars. Thus, local cultivars provide novel genetic resources that should be conserved.

The cross-pollinating nature of olive and its secular history contributed to determine a wide germplasm biodiversity with a large number of more than 1200 cultivars present in the main olive oil producing countries (Bartolini et al., 2005). This genetic diversity could be an important resource for the development of modern olive culture toward typical olive oil and fresh productions. This richness in terms of available biodiversity, however, often has determined some drawbacks in the management and identification of the plant material to distinguish between cultivars, and this has been further complicated by the frequency of homonyms and synonyms (Hegazi et al., 2012).
Many researchers observed that different cultivars are morphologically variable based on geographical locations and under various plant growth management practices (Grati et al., 2002;Youssefi et al., 2011).
The present research aimed to investigate the phenotypic characterizations of olive cultivars from Gilvan area in Zanjan province/ Iran.

| Plant material
Morphological and pomological diversity of 24 olive cultivars (5-15 replications for each cultivar, 243 trees in total) was evaluated at a collection in Gilvan area in Zanjan province/Iran. Gilvan area is located at 36º44′20′′N latitude, 48º53′42′′E longitude, and 1080 m height above sea level. The cultivars were between 10 and 12 years old and were healthy and in full fruiting stage. The orchard management operations, including nutrition, irrigation, and pest and disease control, were performed regularly and uniformly for the cultivars.

| The characters evaluated
Fifty morphological and pomological traits were used to evaluate phenotypic diversity (Table 1). A total of 50 adult leaves and 50 mature fruits per cultivar were randomly selected and harvested. The traits related to dimensions of leaf, fruit, and stone were measured using a digital caliper. A digital scale with an accuracy of 0.01 g was used to measure the weight of fruit and stone. The qualitative traits (Table 2) were visually examined and coded according to the olive descriptor (UPOV, Barranco et al., 2000).

| Statistical analysis
Analysis of variance (ANOVA) was performed to evaluate the variation among cultivars based on the traits measured using SAS software (SAS Institute, 1990). Principal component analysis (PCA) was used to investigate the relationship between cultivars and determine the main traits useful in cultivars segregation using SPSS software. Hierarchical cluster analysis (HCA) was performed using Ward's method and Euclidean coefficient using PAST software (Hammer et al., 2001). The first and second principal components (PC1/PC2) were used to create a scatter plot with PAST software. Also, independent traits affecting the fruit weight as a dependent trait were determined through multiple regression analysis (MRA) using the "linear stepwise" method with SPSS software.  (Lazovic et al., 2018;Peres et al., 2011;Rotondi et al., 2011). The pictures of leaves and fruits of the studied olives are shown in Figure 1.

F I G U R E 1 The pictures of leaves and fruits of olive cultivars studied
Here, fruit weight was considered as a dependent variable and then the direct and indirect effects of each independent variable on this key trait were calculated using MRA (Table 3). The MRA showed that fruit weight was found to be associated with 18 characters. Fruit weight showed the highest positive standardized beta coefficient (β) value with stone weight (β = 0.61, p <.000). Thus, this key variable is one of the main traits accounting for fruit weight and should be considered in breeding programs.

TA B L E 3
The traits associated with fruit weight in the olive cultivars as revealed using MRA and coefficients

TA B L E 4 (Continued)
The PCA was used to understand the relationships among the cultivars. The first 14 PCs explained 75.80% of the total variance (Table 4). The PCA has been used in the evaluation of olive germplasm (Bandelj et al., 2002;Cantini et al., 1999;Hannachi et al., 2008;Hosseini-Mazinani et al., 2004;Lazovic & Adakalic, 2020;Lazovic et al., 2018;Strikic et al., 2009;Trentacoste & Puertas, 2011;Uylaser et al., 2008;Zaher et al., 2011). The first three PCs explained 31.46% of the total variance observed. The characters, including fruit length, fruit diameter, fruit weight, fruit flesh thickness, stone length, stone diameter, and stone weight, were positively correlated with PC1, explaining 14.72% of the total variance. Fruit size morphology is the product of complex genetic and environmental character (Strikic et al., 2009). Five characters, including fruit shape, fruit apex shape, fruit base shape, fruit nipple shape, and stone shape, were placed into the PC2, representing 10.74% of the total variance. The PC3 explained 6.00% of the total variance and showed positive correlations with tree growth vigor, tree height, and trunk diameter. Results obtained agreed with previous PCA of morphological characters in olive cultivars grown in different olive areas (Cantini et al., 1999;Lavee & Wonder, 2004;Lazovic et al., 2018;Ozkaya et al., 2006;Taamalli et al., 2006;Trentacoste et al., 2010;Zaher et al., 2011).
In addition, the scatter plot created based on the PC1 and PC2, accounted for 25.46% of the total variance (Figure 2), showed that the cultivars with close proximity were more similar in terms of effective traits in PC1 and PC2 and were placed in the same group.
The scatter plot showed that residuals of the majority of cultivars bounce randomly around 0.00 line forming the horizontal band. This suggests that the variances in the error terms are equal and the relationship among the cultivars is linear. However, few outliers were observed among the cultivars evaluated, which might be due to their extreme values for particular traits.
Besides, the HCA performed based on the mean of replications with Euclidean distance and Ward method (Figure 3) grouped the cultivars into two major clusters. The first cluster (I) was divided into three subclusters. Subcluster I-A consisted of six cultivars.
Subcluster I-B included 12 cultivars, while subcluster I-C included 2 cultivars. The second cluster (II) included four cultivars. Furthermore, according to an analysis based on replications of cultivars (Figure 4), the studied cultivars were placed into four groups. The mean values of most important fruit traits for the studied olives are shown in Table 5.
The present study confirms previous studies in other countries on the importance of measuring morphological and pomological traits (Cantini et al., 1999;Lavee & Wonder, 2004;Lazovic et al., 2018;Ozkaya et al., 2006;Taamalli et al., 2006;Trentacoste et al., 2010;Zaher et al., 2011), which successfully classified cultivated olives. Furthermore, the evaluation of agronomic traits may be difficult since it may take as long as 10 years to reach reproductive maturity (Suarez, et al., 2011). Hannachi et al. (2008) found that there was a genetic basis in olive cultivars related to fruit size and probable fruit use.

| CON CLUS ION
The identification of olive cultivars and their area of origin are very important to expand cultivation of those commercial varieties with superior products that are best adapted to specific local environmental conditions. Differences in many of the morphological traits were observed across the cultivars. These sets of data were used to identify unique and desirable cultivars morphologically. Stable phenotypic traits were used to discriminate between use of fruit as well as cultivar origins (local or introduced). This research demonstrates that local olive cultivars have unique characteristics that differentiate them from imported cultivars. Thus, local cultivars provide novel genetic resources that should be conserved.

F I G U R E 3
Ward cluster analysis of the studied olive cultivars based on the morphological and pomological traits by Euclidean distances

ACK N OWLED G M ENTS
None.

CO N FLI C T O F I NTE R E S T
The authors declare no conflict of interest.

E TH I C S S TATEM ENT
Research involving Human Participants and/or Animals: None.

I N FO R M E D CO N S E NT
None.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the corresponding author upon reasonable request.