SENSORY CONTRIBUTION FROM EXTRACTS OF BOTANICALS TO DISTILLED ALCOHOLIC BEVERAGES

Extracts from botanicals were examined by sensory evaluation for rheir ability t o improve the flavour of 15% aqueous ethanol as an alcoholic beverage. More than 75% of the 35% aqilcou:, ::hano1 extractable constituents were extracted from the powdered botanicals o n refluxing in Soxhlet for 2 h. The extracts from cloves, cinnamon, "ingurupiyali", ginger, cardomum, nutmeg and mace improved the flavour of 15% aqueoushethanol a t concentrations of 8.2, 1.9, 3.3, 6.7, 2.5, 22.5, and 26.3 ppm respectively. The extracts of cummin seeds, coffee bcans and pepper did not contribute positively t o flavour of 15% aqueous ethanol at their most acceptable concentrations. The extracts of coriander, dill seed, fcnugreek, mustard and "valmee" did not contribute a regular flavour pattern when assessed s t a series of concentration and hence were eliminated. Cinnamon, cloves and cardomum provided the most preferred flavour contributions individually as well as in combinations of two at a time, at several concentrations.


Introduction
Flavouring of aqueous ethanol (spirits) with botanicals is practised to produce alcoholic beverages in the liquor industries today, avoiding the maturation treatment.In Sri Lanka, the production of alcoholic beverages from silent spirits is expanding in view of the decline in production of spirits from fermented palm saps, coconut (Cocos nucifera L.) and palmyrah (Borassus flu bellifer L,).
In this study extracts from locally available botanicals were added to silent spirits.The concentrations at which the extracts from botanicals rendered the spirits more acceptable, and their presence recognized were established by a sensory evaluation panel.
The order of preference of individual and combined botanicals, that rendered the spirits most acceptable was established.

Test Solutions
The ethanol test solutions were prepared by diluting potable spirit containing 96%(v/v) ethanol with glass distilled water t o 35% ethanol for the extraction of botanicals and 15% ethanol for sensory evaluation.All test samples were brought to the same colour by adding caramel.The potable spirits used in this study was purchased from State Sugar Corporation, Kantalai.The GLC trace of the spirits on a BP 20 capillary column and flame ionization detector gave only one peak (ethanol) and the estimated purity was 100%. .

Sensory Evaluation
PaHel: From a pool of 100 volunteers employed at the State Distilleries Corporation, Kandy, 30 volunteers were selected based on their ability to judge alyoholic beverages as examined by "Tringle Test" and the "Duo-trio test".
These volunteers who formed the panel were divided into 6 groups, so that each group consisted of sensory evaluators of different age, health status, education, smoking habits, type of employment and living area.
In sensory evaluation, the judges of the panel were required to rank the samples in a 'hedonic scale' indicating their preferences as : 1-like extremely : 2-like very much : 3-like moderately : 4-like slightly : 5-neither like nor dislike : 6-dislike slightly : 7-dislike moderately : 8-dislike very much and-9-dislike extremely?The duplicate and triplicate samples submitted in the preliminary experiments showed up to 5% variations in judgements.
All sensory evaluation experiments were carried out in late mornings, between 9.00 a.m.t o 11.00 a.m. and early evenings, between 2.00 p.m. t o 3.30 p.m. on week days avoiding Monday mornings and Friday evenings.Samples were served in clean 50ml clear glasses of the same type, in a quiet room with comfortable seats, artificial lighting from fluorescent tubes and containing air free from destructive odours.Between 6-9 samples were tested by a judge in a given sitting.Boiled water and facilities were provided to rinse the mouth prior to tasting each sample.

Extraction of Flavour Constituents from Botanicals
The botanicals were ground to pass through 500 mp sieve, in a Stein mill model M-2 for 3 min.at a speed of 7 .The moisture was estimated by drying l g samples at 105' c for 2 h. 3 T h e powdered botanicals equivalent t o 5g on dry weight basis, were extracted with 100 ml 35% aqueous ethanol in Soxhlet at a siphon rate of 8-10 cycles per h for 1h.The residue was reextracted thrice under similar conditions.A fraction from each e ~t r a c t (10 ml) was evaporated t o dryness over a water bath and then in a' vacuum oven at 70°C t o estimate the dry weight.
Solutions of ~otanical extracts were prepared by pooling the Soxhlet extracts from first and secona extractions and making to 72G na, with .35%aqueous ethanol.The resulting solutions, were diluted lC fold with 35% aqueous ethanol and stored at 4OC pending evaluation.

Establishment of Unpleasant and Detection Threshcr~ Zs
A series of test samples of 15% ethanol containing the extract.^oC each of the botanicals in the order of inc~easing concentration was submitted t o the judges in the six groups t o assess the concentrations of individual botanicals at which the spirits become unpleasant and the concentration at which the added botanical could be identified.(The botanical used was not disclosed t o the judges).The concentration at which 50% of the judges assessed spirits as unpleasant was established as unpleasant threshold.The concentration at which 50% of the judges identified the added botanical co,-rectly was established as the detection threshold.

Establishment of Most Acceptable Concentrations
A series of nine samples for each of the botanicals, within the range decided by the unpleasant threshold and by the preliminary experimerts, was submitted t o the judges for ranking in the nine point hedonic scale.The samples submitted, randomized among and within the groups were coded using numbers or letters following no specific seauence.Histograms were prepared for each of the rank (1-9) and for each btanical by plotting the number of judges givikg the rank against the concentration of the botanical.The most acceptable concentration was established as the concentration preferred by the highest rlumber of judges from the .node of the frequency distribution.

Statistical halysis
Ncn-parametric techniques -iuere used for the analysis of data as application of parametric proceedures of analysis would not be valid for data based on ranks.The setting up oT this study was a. randomized complete block where the judges were taken as blocks and the botanical extracts as treatments.The overall significant differences among the treatments were assessed by Priedman's test.2In establishing the most acceptable concentrations of botanicals the data were analysed usi~lg Wilcoxon rank-sum test for comparisons in pairs.2

Comparison of Botanicals for Order of Acceptability at.Their Most Acceptable Concentrations
, , The ethanol test solutions flavoured with individual botanicals at their most acceptable concentration were compared together with 15% ethanol containing no extracts and the order of preference was established based on tht ranking by the judges in the nine point hedonic scale.
A second set of ethanol test solutions containing a combination of extracts from two botanicals at a time, each contributing 50% of the most acceptable concentration was prepared.The 45 different combinations of test solutions grouped into five sets were initially ranked by the judges in the hedonic scale.The most acceptable combination of extracts in the test solutions from each of these sets were next ranked together in a six point hedonic scale by the judges t o select the order of preference.
A third set of ethanol test solutions containing a combination of extracts from two botanicals at a time, one contributing 25% of it's most acceptable concentration and the other contributing 75% of it's most acceptable concentration was prepared.The ninety combinations produced were grouped into 10 sets for ranking as before.The most acceptable two samples for each set, together with a control of 15% ethanol (total of 21 combinations) was divided into 3 groups of 7 each and ranked by judges in a seven point hedonic scale.The final order of preference was established based on ranking by the judges in these 3 stages.
2.9 Overall Order of Preference of Botanicals rhe overall order of preference of botanical extracts in 15% ethanol solution was established by ranking the most acceptable two samples from each of the above sets, containing the individual botanicals extracts, combination of extracts from two botanicals at 50% most acceptable concentration and the combination of extracts from two botanicals at 25% and 75% most acceptable concentrations.The ranking by the judges was done in a seven point hedonic scale.

Extraction of .Flavour Constituents
Aqueous ethanol (35%) was used in extracting the botanicals as the extracted material should be totally soluble giving a clear liquid at the marketstrength of spirits.It was Tound,that 85% of the aqueous ethanol extractable material could be obtained by extracting in Soxhlet for 2 h.except in case of ginger where only 70% was extracted.The extraction for 2 h in Soxhlet was therefore fixed as the standard.The botanicals used for extraction contained.lessthan 10% moisture except for "valmee"(Tab1e 1).In establishing the sensory evaluation panel reliability was achieved as far as possible by eliminating those who failed t o detect flavour differences when tested by "Triangle" and "Duo-trio" tests.The uniformity and wide range of sensory responses was maintained by grouping the judges so that each group contained people of diverse habits.

Unpleasant and Detection Thresholds
The botanical extracts in test solutions could contribute a decoction-likeflavour t o the spirits at certain concentrations causing a disadvantage.The recognition of the botanical added to improve the flavour may sometimes devalue the product and discourage the consumer.The concentrations at which these features would become obvious was establised by sensory evaluation (Table 1).It is noteworthy that the concentration at which the flavoured test solutions could be detected was higher than the concentra-'tions at which the presence of flavour constituent from the botanicals render the test solutions unpleasant except in case of ginger.

'3;3 Most Acceptable anc cent rations
Of the extracts from the 15 botanicals examined, the test solutions flavoured with coriander, dill seed, fenugreek, mustard and "valmee" did not show a normal distribution pattern when the number of judges giving any given rank was plotted against the concentration.They were eliminated from further tests as their most acceptable concentration cannot be established (Table 1).Of the extracts from the balance 10 botanicals evaluated' cloves, cinnamon, "ingwupiyali", ginger, cardomum, nutmeg and mace were judged t o improve the flavour of 15% ethanol at their most acceptable concentrations, with cloves and cinnamon providing the best flavours (Table 2).Of the 45 combinations of spirits obtained by combining extracts from two botanicals at a time, at 50% of their most acceptable concentrations, only 5 combinations contributed to the flavour of 15% ethanol positively (Table 3).Cardomum, cinnamon and cloves between themselves were judged to provide the most preferred combinations.Of the 90 possible combinations evaluated as combined extracts from two botanicals,uone contributing at 25% of its most acceptable concentration, and the ~t h e r contributing at 75% of the most acceptable concentration, 1 9 combinations were judged to provide a better flavour than the control, 15% ethanol (Table 4).Combinations oi" cloves, cinnamon and cardomum between themselves were judged t o be the best.Extracts from these three botanicals appeared to occur inore frequently in combination with others in the 1 9 more preferred combinations.Table 4.The decrease in order of acceptability of combinations of ex- tracts from botanicals at 25% (A) and 75% (B) of their most acceptable concentrations (mg extract / 1).

OveraU Preference of Added Botanical Extracts
When the two most acceptable botanicals or combinations of botanicals from each of the above sets (Tables 2, 3 and 4) were compared, combination of cloves and cinnamon was judged the best (Table 5 ).It is noteworthy that in all experiments the extracts from the botanicals cloves, cinnamon and cardomum were judged t o make the best contributions towards flavour individually as well as in the two different combinations examined.

Conclusion
Zx%racts from cloves, cinnamon and cardomum could be, used in improving the flavour characteristics of silent spirits to produce a processed alcoholic beverage more acceptab!.e than the beverages produced by using the spirits alone.

Table 1 .
The. moisture percentage and the concentrations at which the botanical extracts are detectable, unpleasant and most acceptable in 15% ethanol.
not identifyable up t o 70 nigll

Table 2 .
The decrease in order of acceptability of extracts from individual botanicals in 15% ethanol and their most acceptable concentrations.(mg extract / 1).

Table 3 .
The decrease in order of acceptability of combination of extracts from botanicals at 50% of their most acceptable concentration.

Table 5 .
The most acceptable concentrations of extracts from individual botanicals and combinations of botanicals arranged in the decreasing order in 15% ethanol.