Sample color location and even statistics are not enough to express if color differences are visually distinguishable. These color differences (ΔE*ab) are important to evaluate numerical and visual relations (CIE, 1995) and may be calculated by the distance between the two spots in tridimensional room defined as colorimetric parameter a*, b*, c* and L* mathematically described by Equation 4.

After being dried, the samples accompanied by their respective obtained residues from both extraction methods suffered losses during storage, but in spite of losses, it contains a higher residual ash content than that found by Silva et al. (2013) (0.51 g 100g-1). However, Lima et al. (2008) expressed higher values which were close to that found in this present work for jabuticaba Sabará (4.40 g 100 g-1 dry mass) and Paulista (2.88 g 100g-1) peels. Lenquiste et al. (2012) observed 3.52 g 100 g-1 in freeze-dried peels and the process applied in this work is efficient. Besides, ash values indicate quantity of mineral residue present in the peel (Marquetti, 2014). This process is also an alternative source of this element.

Jabuticaba peel demonstrated that it contains higher levels of flavonoids, differing statistically between cultivars when extracted by steam and crushing methods, and flavonoid contents reduced gradually from 0 to 135 days, showing oscillation on the 45th day in Clevelândia genotype by steam method, which is attributed to humidity. The degradation of these compounds is due to the conditions of high temperatures and humidity, as well as light, pH and others. All  necessary measures have been taken so that these interfering conditions did not cause any damage to analysis. However, with the advent of drying, there is a sign that these compounds are concentrated (Figure 1). The highest flavonoid contents were in Clevelândia genotype when extracted by steam method with 16.81 mg g-1, during 45 days of storage.

The differences between genotypes can be justified by the greater exposure of these peels to environmental factors, leading to greater incentives for the production of these secondary metabolites related to protection against abiotic stress, making them present higher or lower concentrations of flavonoids (Araújo, 2011). In a work performed by Marquetti (2014) with jabuticaba peel, the author emphasized that even after drying and dehydration, a great deal of these compounds remained and could be transferred to foods where they may work as bioactive ingredient. Abe et al. (2012), while evaluating various fruits as potential sources of bioactive compounds, observed lower contents (33.0 mg CE 100g-1) for jabuticaba and camucamu (Plinia dubia) (31.0 mg CE 100g-1). Dessimoni-Pinto et al. (2011) used jabuticaba from Diamantina – MG, from whose peel, there was production of jelly (87.80 mg CE 100 g-1 fresh mass) and considered as rich in flavonoids.

Total anthocyanin content found in Clevelândia genotype peel powder differed statistically (p<0.05) from Verê genotype (Figure 2), during storage and between extraction methods. Clevelândia genotype have higher concentrations of anthocyanins and such color is related to the hydroxyl, methoxyl and glycol groups present in these structures (Lee et al., 2005). Such contents demonstrated that both genotypes contain elevated concentrations of this compound. Due to jabuticaba peel color (pigmented, dark purple color), it could be predicted (Santos and Meirelles, 2009; Veggi et al., 2011).  a positive association between the Verê genotype’s SSC was also observed in steam extraction with their anthocyanin values, showing lower SSC concentrations and lower concentrations of anthocyanins. Anthocyanins are rarely found free in plant, being connected to various sugars (Francis, 2000). Therefore, this method of extraction may have been responsible for smaller anthocyanin values due to loss of organic compounds by temperature process. Previous studies, which looked forward to testing tropical fruits that were not traditional in Brazil reported better results of anthocyanin content for jabuticaba   Paulista    (58.1 mg   Cy-3-glicoside   100 g-1) (Rufino et al., 2010). In recent work, Böger (2013) while analyzing different ethanoic extracts of jabuticaba peel, obtained values that were similar to those present in this work (27.02 and 60.32 mg Cy-3-glicoside 100g-1), in same analytical conditions.

At 135 days, crushing extraction of Clevelândia genotype had lower levels in comparison with its early days. This degradation may occur by different mechanisms during storage, leading to formation of insoluble dimming products or colorless soluble products (Francis, 2000), but visually the samples showed no differences in this period.

CIELAB system attribute evaluation was performed due to the fact that Verê and Clevelândia genotype peels extracted by steam and crushing were subjected to dehydration under 70°C and stored at 0, 45 and 135 days (Table 2). Values of light coordinates (L*) vary from 0 to 100, where dark is close to 0 and light is close to 100. There was statistical difference in this parameter among fresh peels from Verê and Clevelândia with 23.46 and 20.25 light successively (in a 0 to 100 scale) determined by L* parameter. Where Verê’s tone is closer to white, in other words it is lighter, if compared with Clevelândia. Color intensity close to Verê, as defined by C parameter was 10.18 and Clevelândia 4.89, showing another statistical difference, showing that Verê has a more intense color when fresh, but when dehydrated by steam and crushing methods, did not differ statistically until 135 storage days.