The purpose of this study was to determine the proanthocyanidin content of bananas at the green, the ripe, and the overripe stages. Three bananas from each of three hands at each stage of ripeness, making a total of 27 bananas, were freeze-dried. The ground tissue was dispersed in methanol and the proanthocyanidin in the slurry was converted to anthocyanidin by heating in 0.8 N HCl-n-butanol containing iron (FeSO₄ . 7H₂0). The major anthocyanidin was identified as delphinidin with a minor pigment as cyanidin. Two sub-samples of each banana were analyzed in triplicate and the absorbance of anthocyanidin was read at a wavelength of 545 mμ. The proanthocyanidin content, expressed as milligrams of delphinidin chloride per gram of dried banana, averaged 19.7 for the green, 16.0 for the ripe, and 13.7 for the overripe. Differences in the proanthocyanidin content of the bananas due to ripeness were statistically significant. Moisture content averaged 71.3% for the green, 73.0% for the ripe, and 74.9% for the overripe. Proanthocyanidin content on a fresh weight basis, expressed as milligrams of delphinidin chloride per gram of banana, averaged 5.63, 4.33, and 3.44 for the green, ripe and overripe, respectively. The proanthocyanidin content of the bananas in this study was considerably higher than values previously reported. Analysis of the pulp rather than the extractable proanthocyanidin and a more acidic developer and one which contained iron were major factors contributing to the higher values. While the proanthocyanidin content decreased significantly with ripeness of the bananas, two other groups of workers reported decreases that were proportionately greater. Their data was based on extractable proanthocyanidin and polymerization of these compounds as the fruit ripens could account for their lower values. Conversion of unextractable polymers in the pulp to soluble anthocyanidin could account for the higher values in the present study. A high proportion of the variance for treatment mean was due to hand and relatively little due to banana on the hand, suggesting that fewer bananas from more hands could have been analyzed. A better method than skin color for assessing ripeness is suggested. / Graduation date: 1969