In the plant physiological literature fructose is classified by some as an excellent source of carbohydrate while others have reported poor utilization or outright toxic effects of this hexose. Experiments have shown that fructose media sterilized by filtration are clearly beneficial for the growth of intact plants of Cruciferae. Upon autoclaving fructose decomposes more rapidly than glucose or sucrose. The plants suffer on the autoclaved media not because of the reduction of the sugar level but because of the toxicity of some of the degradation products accumulated. The decomposition takes place in several steps and may follow different paths depending on the composition of the rest of the media and on the duration and/or temperature of heating. The first products of the breakdown displaying an absorption maximum around 230 nm are apparently harmless. On prolonged heating the absorption maximum shifts to around 285 nm signifying the appearance of 5-hydroxymethyl-2-furaldehyde and hydroxyacetyl furan. Both of these compounds are toxic in the 10-4 M range, especially the first one. By thin layer chromatography ten or more dinitrophenyl hydrazones of the decomposition products can be separated. Thus the ‘fructose effect’ in higher plants is apparently quite different from catabolite repression, common in microorganisms. The toxicity of autoclaved fructose is not relieved by cyclic adenosine monophosphate. In crucifers the adverse effect of the fructose media had no connection with the chelating effect of this sugar. The acidification of the medium during autoclaving was of only minor importance. Though 5-hydroxymethyl-2- furaldehyde is primarily responsible for the deleterious effects of autoclaved fructose, a number of other decomposition products compound the toxicity. The selective effects of autoclaved fructose on plant metabolism are pointed out and the usefulness of plants for the detection of low levels of these toxic furans and derivatives is indicated.