Phenolic substances occur primarily in fruits and vegetables and in the seeds of certain pigmented cultivars of sorghum, millets, and legumes. One of the major difficulties encountered in polyphenol research is the lack of a standard quantitative method for the analysis of phenolics that would be suitable for a wide range of seeds, forage crops, and food products and under a variety of experimental conditions. Some methods measure "total phenol", which may not be a true index of the nutritional quality of foods and thus does not distinguish polyphenols of nutritional concern from other low-molecular-weight phenols that also occur naturally in these products. Tannic acid (a hydrolyzable gallotannin) is commonly used as a "reference standard", but this may be a questionable practice since its biological properties differ from those of tannins of flavonoid origin. Polyphenols of cereals and legumes are predominantly of the latter type. Also, commercially available tannic acid has been shown to be a mixture of four phenolic compounds, the relative proportions of which vary with the samples. Thus, the choice of a suitable standard for tannin analysis is also important. The quantitative extraction of the condensed tannins from plant tissue is always difficult, since it may be complexed to a carbohydrate or protein matrix which could be quite insoluble due to a high degree of polymerization. The literature on tannin methodology is diverse and at times conflicting. Currently available methods for tannin analysis range from simple colorimetric, UV spectrophotometric, chromatographic, and enzymic to more sophisticated and expensive nuclear magnetic resonance (NMR) techniques. None of these methods of analyses is completely satisfactory nor can it be applied to different food products with the same degree of success. This review covers physical and chemical methods for tannin analysis of different food products, the problems in analysis and interpretation of data, and future research needs in this area.