Tropical root crops, of which the sweet potato is an important representative, constitute an underexploited resource of developing countries. They can be used as food for both human and animal consumption and their starch is a source of industrial raw material. This review will consider recent reports on sweet potatoes, the physicochemical properties of their starches in comparison with other starches, and the possible causes of variation in these characteristics.

Starch was extracted from the roots of four different cassava cultivars (Munihot esculentu Crantz) grown in Colombia and harvested on four different occasions with a view to determining whether there was any relationship between the physicochemical properties of the starch and the texture of the cooked roots. The organoleptic quality of the cooked roots was evaluated by a trained taste panel and found to vary in terms of both their glassy texture and their hardness. Studies of the gelatinisation behaviour of the starches by DSC and amylography (with a Brabender amylograph), of the pasting properties by viscography and of the viscoelastic characteristics by a Bohlin rheometer demonstrated some differences between the physicochemical properties of the starches but insufficient to account for the pronounced textual differences.

The starch contents of three cultivars of cassava were determined in cooked and uncooked samples with and without removal of alcohol-soluble solids. Hydrochloric acid and amyloglucosidase methods of hydrolysis were used. Hydrolysates were analysed using glucose-specific (glucose oxidase) and non-specific (ferricyanide reduction) methods. Statistical analysis of the results obtained indicated that there were no significant differences between the methods of hydrolysis and sugar analysis. The increased cost of using an enzymic method of starch analysis appears not to be justified for the determination of starch in cassava, unlike a number of other starch sources. The starch values obtained by enzymic hydrolysis of uncooked samples were significantly lower than for cooked samples. The percentage of uncooked cassava starch susceptible to enzymic degradation varied with cultivar from 14.0 to 46.88%. In view of the use of unprocessedlungelatinised cassava chips or partly gelatinised cassava pellets for animal feed, further research into cultivar variation of starch susceptibility to enzymic degradation is required.

Rapid post-harvest physiological deterioration of cassava roots (Munihot esculentu Crantz) appears to be due essentially to wound responses, comparable to those observed in other plant storage organs. The wound responses observed include: increased activity of phenylalanine ammonia lyase, an enzyme associated with phenol biosynthesis; increased activity of peroxidase and polyphenol oxidase; formation of phenolsipolyphenols including leucoanthocyanidins, catechins, scopoletin and condensed tannins, and often the formation of a wound periderm. In cassava, the responses did not remain localised at wound surfaces in roots when held at low storage humidity but spread through the roots causing a discolouration of the vascular tissue and storage parenchyma. Roots stored at high humidity showed a more typical wound response with localised production of phenols and periderm formation.

Four varieties of cassava plants were cultivated at CIAT in Colombia and harvested on four different occasions with a view to evaluating whether there was any relationship between the texture of the cooked roots and the constituent starch. The organoleptic qualities of the cooked roots showed considerable differences in ‘glassiness’ and hardness of texture, both between varieties and times of harvest. The starch granules were isolated and examined by a variety of techniques. The size of the granules did vary with the seasons but there were only minor differences in the x‐ray crystallinity, the elution patterns of iso‐amylase‐debranched material (including the amylose content), the proximate composition of the starch and the time course of digestion by glucoamylase. None of these provided a basis for explaining the differences in cooked texture.