Thesis (MSc Food Sc )--Stellenbosch University, 2003. / ENGLISH ABSTRACT: The effect of drying conditions on the retention of harpagoside, a glucoside iridoid found in the indigenous Southern African plant Harpagophytum procumbens (Devil's Claw), was investigated. Additionally, the feasibility of using near infrared spectroscopy (NIRS) as a rapid and nondestructive analytical tool for the prediction of the iridoids harpagoside (HS), 8-p-coumaroyl harpagide (8pCHG) and moisture contents, as well as the HS:8pCHG ratio in Devil's Claw was also studied. Due to the purported medicinal activity ofthe tuberous secondary roots, Devil's Claw is of commercial interest with a well established export market in Germany. The stability of HS during processing has, however, never been investigated and the current study evaluated the effect of three drying methods (sun, tunnel and freeze-drying), as well as three tunnel-drying temperatures (40°C, 50°C and 60°C, dry bulb temperature) on HS content. Methanolic extracts (70% methanol-water) were analysed by high-performance liquid chromatography (HPLC) and a multiple extraction test (n = 12) indicated a 91.70% recovery for HS. The comparative drying results showed significant (P < 0.05) differences between tunnel and freeze drying on the one hand and sun-drying on the other, with the latter being the most detrimental technique. In comparing tunnel-drying temperatures, at a constant relative humidity (RH) of30%, drying at 40°C showed the lowest retention ofHS, differing significantly (P < 0.05) from that at 50°C. Drying at 60°C did not differ significantly (P > 0.05) from either of the other two conditions, but also resulted in lower retentions ofHS content than drying at 50°C. Overall, freeze-drying showed excellent results for HS retention (1.565% ± 0.394%; mean ± standard deviation), but economic constraints limit the commercial application of this drying method. Although further study is required to optimise tunnel-drying conditions, this method indicated acceptable and practical results (1.526% ± 0.396%), especially with conditions set to 50°C and 30% RH (1.750% ± 0.438%). Future research should concentrate on elucidating the degradation mechanisms of HS during processing, transport and storage of dried Devil's Claw root. The analysis of dried, ground Devil's Claw root by NIRS was referenced to HPLC analyses of the same methanolic extracts used in the drying studies. The spectra were generated on two different NIRS instruments by measuring diffuse reflectance of the powder whilst calibrations were performed by partial least squares regression. Both independent data set validation and full cross-validation were used to evaluate the performance and predictive abilities of the various models. Excellent moisture content (MC) calibrations were developed with a standard error of prediction (SEP) ofO.24% (range: 2.44% to 10.43%) and correlation coefficient (r) of 0.99. For the given MC range, the model performed very well against the standard error of laboratory (SEL) ofO.14%. Considering the sensitivity of the HPLC reference method (SEL = 0.035%), the HS content calibration (range: 0.693% to 2.244%) performed admirably with an SEP ofO.134%. The available range was, however, relatively small and future NIRS studies should incorporate a larger range if possible. The 8pCHG model (range: 0.069% to 0.290%) performed similarly, with a standard error of cross-validation of 0.028% and SEL of 0.007%. By inclusion of a second species, H. zeyheri, both the 8pCHG content and especially HS:8pCHG ratio (range: 1.84 to 34.48) calibrations were evaluated as possible methods of distinguishing between species. With 8pCHG values of ca. 0.9%, good separation was achieved for the first model but, contrary to previous studies, the ratio was found to be less successful. Principle component analysis of the spectra, however, showed great potential as a qualitative tool for this purpose. It is believed that NIRS can become an invaluable tool for the division of Devil's Claw into export classes, depending on HS content. This content is, however, dependant on the drying conditions used and drying parameters should be optimised to ensure the best possible quality and retention ofHS. / AFRIKAANSE OPSOMMING: Die invloed van drogingskondisies op die behoud van harpagosied, 'n iridoïed glikosied wat in die inheemse Suider-Afrikaanse plant, Harpagophytum procumbens (Duiwelsklou) voorkom, is ondersoek. Die uitvoerbaarheid van naby infra-rooi spektroskopie (NIRS) as 'n vinnige en niedestruktiewe analitiese metode vir die voorspelling van die inhoud van iridoïede, harpagosied (HS), 8-p-kumaroïel-harpagied (8pCHG) en vog, sowel as die HS:8pCHG verhouding in Duiwelsklou is ook ondersoek. Die beweerde medisinale aktiwiteit van die knolagtige sekondêre wortels van Duiwelsklou het gelei tot 'n gevestigde uitvoermark in Duitsland, en is dus van kommersiële belang. Die stabiliteit van HS gedurende prosessering is nog nooit ondersoek nie. Die huidige studie het die invloed van drie drogingsmetodes (son-, tonnel- en vriesdroging) en drie tonneldrogingstemperature (40°C, 50°C en 60°C, droëbol temperatuur) op die HS inhoud evalueer. Metanolies ekstrakte (70% metanol-water) is met hoë-druk vloeistof chromatografie ("HPLC") geanaliseer en 'n veelvoudige ekstraksie toets (n =12) het 'n 91.70% herwinning van HS getoon. Betekenisvolle (P < 0.05) verskille is gevind tussen tonnel- en vriesdroging enersyds en sondroging andersyds, met laasgenoemde die nadeligste metode. Die vergelyking van tonneldrogingstemperature, by 'n konstante relatiewe humiditeit (RH) van 30%, het getoon dat droging by 40°C die minste HS behou het en betekenisvol (P < 0.05) verskil het van droging by 50°C. Alhoewel die droging by 60°C nie betekenisvol verskil het van die ander twee kondisies nie, het dit laer HS waardes as droging by 50°C getoon. Vriesdroging het uitstekende resultate ten opsigte van die behoud van HS gelewer (1.565% ± 0.394%; gemiddelde ± standaard afwyking), maar ekonomiese beperkings plaas noodgedwonge limiete op die toepaslikheid van hierdie drogingsmetode. Verdere studie om tonneldrogingskondisies te optimiseer word is nodig (1.526% ± 0.396%), maar veral droging by 50°C en 30% RH (1.750% ± 0.438%) het prakties aanvaarbare resultate gelewer. Toekomstige navorsing moet daarop gemik word om die spesifieke degradasie meganismes van HS gedurende die droging, vervoer en opberging van Duiwelsklouwortel te verklaar. Die NIRS ontleding van gedroogde, gemaalde Duiwelsklou wortel is vergelyk met HPLC analises van dieselfde metanolise ekstrakte wat tydens die drogingstudies gebruik is. Twee verskillende NIRS instrumente is gebruik om spektra deur diffuse weerkaatsing in die poeier te meet en kalibrasies is daaropvolgend met gedeeltelike minste vierkante ("PLS") regressie gedoen. Beide onafhanklike data stel validasie en volle kruisvalidasie is gebruik om die prestasie en voorspellingsvermoë van die onderskeie modelle te evalueer. Uitstekende voginhoud (VI) kalibrasies met 'n standaard voorspellings fout (SVF) van 0.24% (klasinterval: 2.44% tot 10.43%) en 'n korrelasie van 0.99 is gevind. Die model het goed vergelyk met die standaard laboratorium fout (SLF) van 0.14%, oor die gegewe klasinterval. Met inagneming van die sensitiwiteit van HDVC as verwysingsmetode (SLF = 0.035%), het die HS inhoud kalibrasie (klasinterval: 0.693% tot 2.244%) aanvaarbaar gepresteer met 'n SVF van 0.134%. Die beskikbare klasinterval was egter relatief klein en toekomstige NIRS studies moet gebruik maak van 'n wyer klasinterval indien moontlik. Die 8pCHG model (klasinterval: 0.069% tot 0.290%) het soortgelyk gepresteer, met 'n standaard kruisvalidasie fout van 0.028% en SLF van 0.007%. Met die insluiting van monsters van 'n tweede spesie, H. zeyheri, is beide die 8pCHG inhoud en HS:8pCHG verhouding (klasinterval: 1.84 tot 34.48) kalibrasies as moontlike metodes vir onderskeiding tussen die spesies geëvalueer. Met'n 8pCHG inhoud van ca. 0.9% het die eerste model 'n goeie skeidingsvermoë getoon. In teenstelling met vorige studies egter, het die verhoudingsmodel minder sukses getoon. Hoofkomponent analise ("PCA") van die spektra was egter baie belowend as 'n kwalitatiewe onderskeidingsmetode. Volgens hierdie studie kan NIRS 'n waardevolle tegniek wees vir die verdeling van Duiwelsklou in uitvoer klasse op grond van HS inhoud. Hierdie inhoud is egter grotendeels afhanklik van die drogingskondisies en drogingsparameters wat verder verfyn behoort te word om 'n produk van die hoogste kwaliteit met behoud van HS te verseker.