Thesis (PhD)--Stellenbosch University, 2014. / ENGLISH ABSTRACT: Brandy has a high alcohol content of at least 36% alcohol by volume and a complex volatile chemical structure. Due to this, the sensory evaluation of brandy is challenging to execute and must be carefully managed to avoid panellists’ sensory and mental fatigue. Although rapid sensory profiling methods, such as projective mapping (PM), are commonly used for fast moving consumer goods, it was unclear whether these methods could accommodate the difficulties associated with brandy evaluation. The primary objective of this study was to validate PM, also called Napping®, as a reliable tool for brandy evaluation. Two variations of PM were tested: global napping (GN) where products are evaluated based on overall sensory perception, and partial napping (PN) where the product perception is broken down in to separate sensory modalities, in this case appearance, aroma and in-mouth sensations. Several practical aspects of PM were investigated in order to optimise the method for brandy evaluation and to gain understanding into practical methodological aspects that have not been fully understood at the onset of this study. The results showed that both GN and PN delivered reliable results, but that PN was more reproducible and better suited to larger sample sets (10 products). The concept of in-mouth sensations was found to be ineffective in extracting useful information on mouthfeel differences in the product set and that retronasal flavour should be separated from basic tastes and mouthfeel. A verbal instruction to the panellists was sufficient to obtain reliable information on mouthfeel differences; it was not necessary to use black glasses or nose-clips to eliminate the influence of appearance and flavour perception. In response to the insufficient reproducibility of GN, the sorting method was tested and validated as a suitable screening method that delivered reliable product maps of larger sample sets (10 products) in one replication, compared to GN which has to be replicated several times. A recent hypothesis states that the elongated horizontal dimension of a rectangular PM tasting sheet could be used to elicit more prominent product differences, while tasting sheets without an elongated axis, such as circles or squares, could be used to bring forth more subtle differences. This hypothesis was challenged by testing rectangular, square and round tasting sheet shapes. Although differences were observed between the shapes in terms of product configuration and panellist performance, the practical value of using different tasting sheet shapes to obtain specific information could not be established. The impact of high alcohol content and product complexity on panellist performance in PM was tested and it was concluded that product complexity did influence the panellists’ performance, particularly for high alcohol products. It was also found that prior knowledge of a products’ high alcohol content influenced the panellists’ descriptive language to include more alcohol-related terms. A new panellist performance measure was developed, namely the Relative Performance Index (RPI). This measure can be used to monitor trained panellists’ performance in the PM task more effectively to thereby ensure reliable results. The outcomes of this study extended the brandy sensory evaluation toolbox. Practical measures were identified that can be used to overcome the challenges associated with the sensory evaluation of complex high alcohol products. / AFRIKAANSE OPSOMMING: Met ‘n alkohol inhoud van mintens 36% alkohol per volume en ‘n komplekse vlugtige chemise samestelling is brandewyn ‘n uitdagende produk om sensories te evalueer. Maatreëls moet in plek gestel word om te verhoed dat sensoriese paneellede nie uitgeput raak nie om sodoende betroubare resultate te verseker. Hoewel vinnige sensoriese toetsmetodes, soos projeksiekartering (PM) en sortering, gereeld gebruik word vir die evaluasie van vining bewegende verbruikersgoedere was dit weens die voorafgenoemde redes onduidelik of die metodes geskik sou wees vir brandewynevaluasie. Die oorhoofse doelwit van hierdie projek was om PM, ook genoem Napping®, te valideer as ‘n geskikte metode vir die sensoriese evaluering van brandewyn. Twee variasies van PM is ondersoek: oorhoofse kartering (GN) waar die produkte op grond van ‘n algehele sensoriese waarneming evalueer word, en gedeeltelike kartering (PN), waar die verskillende sensoriese waarnemingsmodaliteite afsonderlik gemeet word. In hierdie geval is voorkoms, aroma en algehele binne-mondse ervaring gemeet. Verskeie aspekte van GN en PN is ondersoek ten einde die metodes vir brandewynevaluasie te optimiseer asook lig te werp op praktiese metodologiese aspekte waaroor daar nog onsekerheid was by die aanvang van hierdie studie. Die resultate het getoon dat beide GN en PN betroubare inligting oor brandewyn kan oplewer, maar dat PN meer herhaalbaar en beter geskik is vir groter produkstelle (10 produkte). Dit is egter bevind dat die konsep van algehele binne-mondse waarneming nie doeltreffend was om mondgevoelverskille tussen produkte uit te lig nie. Smaak- en mondgevoelwaarnemings moet liefs onderskei word van binne-mondse geurwaarneming. Dis bevind dat dit voldoende is om hierdie onderskeid deur skriftelike instruksies tot die paneel te bewerkstelling en dat die gebruik van swart glase en neusknippies nie nodig was om betekenisvolle inligting oor mondgevoel te bekom nie. Na aanleiding van die swak herhaalbaarheid van GN, is sortering getoets en gevalideer as ‘n geskikte siftingsmetode wat betroubare produkkaarte van groot produkstelle (10 produkte) kan oplewer met slegs een evaluasie, teenoor GN waarvoor meer as een sessie benodig word. ‘n Onlangse hipotese stel dat die langer horisontale dimensie van reghoekige PM proekaarte gebruik kan word om meer prominente verskille tussen produkte uit te wys, terwyl proekaarte sonder verlengde asse, soos sirkels en vierkante, gebruik kan word om meer subtiele verskille uit te wys. Hierdie hipotese is bevraagteken deur reghoekige, vierkantige en ronde PM proekaarte teenoor mekaar te toets. Hoewel die resultate vanaf die verskillende proekaarte effens verskil het in terme van produk konfigurasie en die paneel se taakverrigting, kon die praktiese waarde van die gebruik van produkkaarte met verskillende vorms om spesifieke inligting te bekom nie bevestig word nie. Die impak van hoë alkohol en produkkompleksiteit op paneellede se PM taakverrigting is ondersoek. Dit is bevind dat produkkompleksiteit wel hul taakuitvoering beinvloed, en dat hierdie invloed groter is vir hoë alkohol produkte. Dit is ook gevind dat paneellede se beskrywende taalgebruik beïnvloed om meer alkoholverwante woorde in te sluit word indien hul kennis dra van ‘n produk se hoë alkohol inhoud. ‘n Nuwe vaardigheidsberaming vir die meet van paneellede se taakverrigting is ontwikkel, naamlik die Relatiewe Taakverrigtingsindeks (RTW). Dit kan gebruik word om paneellede se vaardigheid in PM meer doeltreffend te monitor en sodoende betroubare resultate te verseker. Die uitkomste van die studie het die poel beskikbare metodes vir die sensoriese evaluering van brandewyn suksesvol verbreed en praktiese maatreëls is geidentifiseer om die uitdagings van ‘n komplekse hoë alkohol produkte te oorkom.