**3. Characterisation of fermented from melon fruits**

Fermentation of the juice, pws and paste was carried out at 20 °C. The substrates were inoculated with a commercial yeast (*Saccharomyces cerevisiae* UCLM 325) up to a concentration of approximately 106 cells/mL. The process was monitored daily by measuring residual sugars, and the end of fermentation was determined on the basis of the sugar consumption (OIV, 1969). The initial assimilable nitrogen was measured using the NitroGenius® kit.

Judging from the initial ºBrix, between 10,0-10,2 an alcohol degree of 5% (v/v) could be expected. Nonetheless, experimental data showed that the ethanol yield was acceptable only in juice and pws (4.2 % v/v), being very low (3.4 %) in the case of the paste, possibly due to the complexity of the structure of the fermentation media.

The fermented paste showed the highest values of acetic acid possibly as a consequence of a contamination by acetic bacteria. Adjusting the pH, successfully bacterial growth in all three melon wines produced was diminished. Under unadjusted pH conditions, the bacterial populations increased in the pws and paste substrates but decreased in the juice. This may be attributable to a higher level of contamination from the melon skins or to sluggish fermentation in a complex media like paste substratum.

In this kind of alcoholic beverages, concentrations of the volatiles has to be refereed to the ethanol content. Otherwise, the volatile composition is closely related to the type of substratum, the conditions of the fermentation and the yeast strain used. Respecting the major volatiles, acetaldehyde ranged widely from 243 to 1196 mg/L of ethanol, being highest in the pH-unadjusted substrates, possibly due to the action of spoilage microorganism (Silva et *al.*, 2000). Methanol is not a direct product of fermentation (Ribéreau-Gayon et *al*., 2000). Two types of fruit enzymes are able to act upon pectins to release methanol: polygalacturonases, by cleavage of the glycosidic bonds on the chains; and pectin-Methylesterases, by catalyzing hydrolysis of the esterified chemical function (Hernández Gómez et *al.*, 2003). The presence of high amounts of methanol in the wine fruit produced from the paste at both pH levels may be the result of the action of these enzymes in the skin.

In general the higher alcohols (HAs) quantified [1-Propanol, isoAmyl alcohols , 1-Butanol, and 2-Methyl-1-Propanol] were higher with pH adjusted, especially in the case of the wine made from the pws. 1-Butanol and 2-Butanol were not detected, a highly positive finding, because these two substances adversely affect the final aroma of the distillate. Total esters were higher in the pH-unadjusted wine made from paste than in the rest of the wines. Ethyl lactate was the main contributor to this high value and probably depends on the initial count of lactic acid bacteria, present in this kind of substratum (Briones et *al*. 2002).

When ANOVA statistical analysis was applied it was noticed that except for Ethyl acetate, there were differences in the volatiles for all the melon wine types, especially between the wines made from the paste and the rest.

Spirits and Liqueurs from Melon Fruits (*Cucumis melo* L.) 187

same time, the volatile compounds in the distillates obtained were compared with those in

The fermented juice was immediately distilled in a traditional 130-L "*alquitara"*, (reflux still) (Silva, Macedo & Malcata, 2000) filled to 70-80 % of capacity, equipped with a series of temperature sensors. Distillation flow rate was set at 170 mL/min, and the condenser was kept at below 21 °C throughout. The distillate was collected in volumes of 1 L each, except for the head fraction. The first distillation was stopped when the alcohol content in the volume collected had reached 8-10 % (v/v), which yielded a distillate with an alcohol content of 18.5-25 % (v/v), depending on the source substrate from which it had been made.

The second distillation was carried out in a traditional 30-L alembic copper still filled with 15 L of the first distillate. Distillation flow rate was set at 35-40 mL/min. The heads, 0.8 % of the distillate, were discarded, and distillation was stopped at 40 % (v/v), thus yielding a final distillate (hearts) of 58-69 % (v/v), again depending on the source substrate from which it had been made. The tails comprised the fractions from 40 % (v/v) to 5 % (v/v). For all the distillates the alcohol content of the last volume collected was between 9.2 and 11.8 % (v/v), and the final alcohol content was between 18.5 and 25 % (v/v). In the second distillation, the alcohol content decreased from 74.5 % (v/v) to 40.0 % (v/v) in the last volume collected. The total distillation time for all the fractions was around 4 h. The highest value was for the juice

After the second distillation the major volatiles present in the heads, hearts, and tails

Paste pH adjusted Paste pH unadjusted PWS pH adjusted PWS pH unadjusted Juice pH adjusted Juice pH unadjusted

hH t hH t hH t

Fig. 1. Evolution in major volatile compounds during the second distillation. Methanol. Higher alcohols: 2M1P, 2M1B, 3M1B, and 1-Propanol. Esters: Ethyl lactate, Ethyl acetate, and Ethyl butyrate (mg/L of EtOH). h, heads; H, hearts; t, tails. pws = paste without skins

Methanol was collected in nearly the same proportion in all the fractions, most likely due to the formation of azeotropic mixtures (Orriols, 1994). Nevertheless, methanol concentrations were higher in the distillates from the pH-unadjusted wines except for "juice" tails. High levels of methanol in the paste distillate were observed (pH-unadjusted). The higher alcohols (HA)

distillate (pH-adjusted) and the lowest for the paste (pH-unadjusted) distillate.

The head-fraction (200 mL), usually discarded, was not rejected.

fractions of the different spirits are depicted in Figure 1.

other commercially available spirits.
