*2.3. Statistical Data Elaboration*

The results of the analyses were elaborated as mean ± standard deviations of three sampling years for two harvesting times. Significant differences (*p* < 0.05) were obtained by one-way analysis of variance (ANOVA) and multivariate analysis. Pearson's coefficient was used to study the correlation among qualitative parameters of olive oils. SPSS Software (Version 15.0, SPSS Inc., Chicago, IL, USA) was used for statistical elaboration.

#### **3. Results**

Mean results of the olive oil yield during the three years of study for the minor olive varieties are reported in Figure 1. Drupes of Tonda di Filogaso and Ottobratica Calipa possessed similar oil content at the first sampling (27–29% d.m.), whereas Ciciarello and Ottobratica Cannavà differed for less abundant oiliness (18–19%). In the following harvesting period, the oil content remained significantly similar in Tonda di Filogaso cv, whereas it tended to increase with the highest result in Ottobratica Calipa (44%). The ripening index varied among varieties and harvesting months, as Supplementary Materials shows (Table S1).

The results of principal qualitative parameters of oils, as three-year means, are illustrated in Tables 1 and 2.

During the three years and at the second production in particular, that is November, a large variability in free acidity was observed in olive oils from the same accession, except in the Ottobratica Cannavà clone (0.55 ± 0.13). The range observed in oils produced in October was 0.31–0.57%; at November, it tended to increase in all of the samples, exceeding the 0.8% in some years, except for Ottobratica Cannavà oils. The other productions were affected probably by a different varietal response to some negative environmental factors linked to a specific year (Tables S2–S6): 2018 for Ottobratica Calipa and Ciciarello (total acidity >1%), and 2019 for Tonda di Filogaso (total acidity near 2.5%), as evidenced by the values of the standard deviations. The observed low quality of Ottobratica olive oil produced in November (Table 2) was confirmed by previous works [28,29], and reflected the origin of its name, strictly linked to its optimal ripening in the month of October. It is interesting to note that for one of the two Ottobratica-related clones, Ottobratica Cannavà,

the free acidity inside the legal limit of 0.8% in that period expressed a positive result of the performed new genetic duplication.

**Figure 1.** Olive oil yield at two harvesting times of the four studied minor varieties, with Ottobratica cv used as the reference for Calipa and Cannavà clones. Values are the means of 2017, 2018, and 2019. Different letters show significant differences at *p* < 0.05 by Tukey's post hoc test.

**Table 1.** Principal chemical parameters of olive oils of Tonda di Filogaso (TF), Ciciarello (C), Ottobratica Calipa. (O. CLP) Ottobratica Cannavà (O. CNV) and Ottobratica (O) accessions.


The data are presented as means ± standard deviations. \*\* Significance at *p* < 0.01; \* significance at *p* < 0.05; n.s., not significant; a, ab, b see Figure 1. 42


**Table 2.** Fatty acid compositions of olive oils of Tonda di Filogaso (TF), Ciciarello (C), Ottobratica Calipa. (O. CLP) Ottobratica Cannavà (O. CNV) and Ottobratica (O) accessions.

The data are presented as means ± standard deviations. \*\*, \*, n.s. see Table 1; a, ab, b see Figure 1.

Peroxide values of oils were in the range of 2.50–5.58 mEq O2/kg, with significant differences between the two harvesting times only in Ottobratica Cannavà oils. Higher peroxide values were noted in the oils of clones compared to those from Ottobratica cv in October, whereas an opposite result was detected for the productions of November (6.63 mEq O2/kg in oils from Ottobratica). Spectrophotometric indices denoted olive oil productions of good quality at both harvesting times without significant differences, with the only exception of Tonda di Filogaso olive oils.

The major fatty acid in olive oil is oleic acid; in our study, its content varied with significance from 68.15% (Tonda di Filogaso oils in November) to 75.79% (Ciciarello oils in October). The followed principal detected fatty acids were palmitic acid (C16:0), quantified from 13.23% to 15.74%, and then linoleic acid (C18:2) that varied from 5.96 to 10.06%; both components were similar among minor varieties. The stearic acid (C18:0) was significantly higher in the oils of Tonda di Filogaso obtained in November (3.44%) than in the other samples. The other fatty acids that significantly varied among the samples were C16:1, among the unsaturated ones, from 0.99 to 1.64% and C20:0, among the saturated ones, from 0.34 to 0.46%, as evidenced by the Tukey's post hoc test (*p* < 0.05) elaboration (data

not shown). Olive oils from Ciciarello showed the highest oleic/linoleic (11–13) and monounsaturated/polyunsaturated acid (MUFA/PUFA) (10–12) ratios, confirming the previously discussed results for fatty acid quantification. The antioxidant compositions of olive oil from minor accessions is reported in Table 3.

**Table 3.** Antioxidant composition and activity of Tonda di Filogaso (TF), Ciciarello (C), Ottobratica Calipa. (O. CLP) Ottobratica Cannavà (O. CNV) and Ottobratica (O) accessions.


Amount expressed as mg/kg for TChl (total chlorophylls), TCa (total carotenoids), TT (total tocopherols), TP (total polyphenols), and % inhibition/mg for DPPH and ABTS assays. The data are presented as means ± standard deviations; \*\*, \*, n.s. see Table 1; a, ab, b, c see Figure 1.

> Significant (*p* < 0.01) differences of pigment amounts were observed among the samples; the olive oils of Ciciarello and Tonda di Filogaso obtained in October were the richest in chlorophylls (10.95 ± 3.41 mg/kg and 10.21 ± 5.05 mg/kg, respectively). Ciciarello olive oils were even the richest in total carotenes (7.16 ± 2.28 mg/kg). The oils extracted in November showed reduced pigment amounts and, in particular, a major

reduction was observed in Ottobratica Calipa olive oils (TCL: 2.53 ± 1.23 mg/kg and TCA: 2.26 ± 0.90 mg/kg). ANOVA data elaboration showed variations for pigments between harvesting times, except in the oils of the Ottobratica Cannavà clone (Table 3).

Chlorophyll and carotenoid amounts were also significantly (*p* < 0.05) higher than those resulted in the cultivar population of reference (Ottobratica cv).

Comparing the total mean amounts of polyphenols quantified at two harvesting times, no significant differences were noted.

Among productions in November, Ottobratica Cannavà oils possessed higher mean phenolic antioxidant amount than the other clone, Ottobratica Calipa, and Ottobratica cv. The total tocopherols detected in the oils from minor accessions were in the range of 225–595 mg/kg; Ottobratica Cannavà olive oils were the richest for this typology of antioxidants (with the only observed significant variation between harvesting times), followed by Ottobratica Calipa olive oils, whereas lower amounts were detected in those from Tonda di Filogaso (224–227 mg/kg) as confirmed by literature [30]. A significant decrease in TT content was observed only in oils from Ottobratica Cannavà extracted in November. In the other productions, the total tocopherols remained constant.

The antioxidant activity of the oils was analyzed by the reaction against two radicals, DPPH and ABTS. The obtained results denoted a higher response with the second antioxidant assay (23.29–45.67%) than the DPPH radical (13.80–34.71%). The largest differences among the varieties were significantly observed in the oils produced in October for ABTS assays (Ottobratica Cannavà > Ottobratica Calipa > Ciciarello = Tonda di Filogaso).
