It is quite apparent that the flavor deterioration which takes place during the unrefrigerated storage of canned orange juice is greater than that which is encountered during the similar storage of canned non-citrous fruit products. Observations made in this laboratory with canned Florida Valencia orange juice stored a t any temperature above 7O0F.(21"C.) may be described by saying that the predominant off-flavor which develops under such conditions is ' terebinthine " in character. The terebinthine factor is acrid and camphoraceous, and is often most clearly apparent as an after-flavor, perceived after swallowing the juice. Similar juice stored at temperatures which never exceed 70°F.(21"C.) also depart in time from a desirable fresh flavor, but in this case the off-flavor has a complex and variable quality which may be named "stale" for purposes of discussion, and which is conspicuous in such samples because a t the lower temperature the terebinthine factor does not become dominant. Such empirical observations indicate quite clearly that the over-all degradation in flavor during storage involves a large number of different chemical reactions, differing in regard to the temperature coefficient of reaction rate.It has been postulated in the past that flavor degradation is due to microbiological activity (30) o r to changes in nitrogen compounds (19), or to products resulting from a postulated initial oxidation of juice components. The primary oxidation products are not themselves off-flavored but are thought to constitute an unstable system which leads to the ultimate production, from the fatty components of the juice, of substances which impart to the flavor a rancid quality (20). It is probably fair to say that the proponents of these hypotheses were seeking explanations of the stale, rather than the terebinthine, flavor factors.

A rapid and precise polarographic method for the determination of tin in foods and biological materials has been developed, which is applicable to the determination of tin in concentrations as low as 0.5 part per million. The accuracy is of the order of *5% for concentrations in excess of 10 parts per million and somewhat less for smaller amounts. It has been demonstrated that moderate amounts of iron, copper, bismuth, cadmium, mercury, cobalt, and zinc may be present without interference. Lead leads to high results and a procedure in case lead is present is outlined. THE accurate determination of traces of tin in foods is of considerable interest and has many limitations.