In an investigation into the production of a r o m in roasted coffee, some of the principal volatilc constituents have been determined. In the staling of coffee only small losses of aldehydes, acetone, and volatile phenols occurred and there was no loss of pyridine or of furfural, despite the volatility of these constituents. Apart from a conversion of diacetyl to acctyl-methyl carbinol, no change indicating a loss of aroma was discovered. When coffee was prepared for drinking, some losses of volatile constituents were detected. The substance, or substances~ g-iving the characteristic aroma to roisted coffee is as yet unidentificd. It is probable that other less volatile substances, that are soluble in light petroleum and are readily oxidizablc, arc involvcd in the flavour of roasted coffee.
It is well-known that raw coffee beans contain no aromatic substances, but that the characteristic aroma of coffee is produced during the roasting process. A great deal of work on the chemical composition of coffee and on the changes that take place during roasting has been done in the past. It is now known that ::asted coffee contains no essential oil, but that the volatile oil" (which has been loosely referred to as %affeol") consists of a mixture of organic compounds produced by the decomposition of larger molecules during roasting.
This paper includes the results of part of an investigation on some of the volatile constituents that are produced when coffee is roasted, and also on the changes that take place during staling. It was decided to determine, at different stages of roasting, the main volatile constituents that have been identified in roast coffee, and for which suitable micro-methods were available. In this way it was possible to follow the changes taking place during roasting. By repeating these determinations after the samples of roasted coffee had been allowed to become stale by storing them in non-airtight containers, it was also possible to detect any changes in the amounts of these substances which take place during staling. The investigation was also extended to an examination of the changes in some of the volatile constituents during the preparation of coffee for drinking.
Review of the literature
By examining the condensates obtained from coffee roasters, Bernheimer,' Jaeckle? and Moriari and ScocciantP detected formic acid, acetic acid and higher fatty acids, acetone, furfural, ammonia, methylamine, trimethylamine, pyridine and homolopes, pyrrole, hydroquinone, and resorcinol. Jaeckle? referred to an unidentified substance containing sulphur," and Bernheimer claimed to have discovered a substance which he believed to be a methyl derivative of saligenin.
By examining roasted coffee itself, Erdmann,' Bertrand and Weisweiller,6 Sayre,O Grafe: Schmalfuss and Barthmeyer? Pmcott et U Z . , ~~ Staudinger and Reichstein,"* and Johnston and F~ey,'~ detected formic acid, acetic acid, n-valeric acid, ethyl alcohol, acetaldehyde, diacetyl, acetylmethyl carbinol, acetylpropionyl, diethyl ketone, furane, furfural, furfuryl alcohol, furfuryl acetate or formate, phenol, guaiacol, pvinyl guaiacol, pvinyl catechol, eugenol, sylvestrene, vanillone, hydrogen sulphide, methyl mercaptan, methane thiol, furfuryl mercaptan (2-fury1 methane thiol), dimethyl sulphide, pyridine and homologues, pyrazine, N-methyl pyrrole, and n-heptacosane, etc. In addition, reference was made by Erdrnad to "an oil containing nitrogen, with a characteristic odour of coffee," and by Johnson and Frey13 to 'I an oil, turning reddish-brown