FISHERY BULLETIN: VOL. 69, NO. 1 



process and by a modified two-step process in 

 which the silica-gel procedure was eliminated 

 but in which the collection of distillate during 

 molecular distillation was restricted. Such a 

 procedure greatly simplified refining but resulted 

 in a menhaden-oil product containing consider- 

 able burnt-flavor component. Flavoring ma- 

 terials giving best results with the latter oil 

 did not necessarily give optimum results with 

 the silica gel refined oil. 



Of the flavoring materials found to be most 

 successful for the ordinarily refined procedure, 

 those with a fruity-type flavor — especially 

 Cherry Bouquet and Fruity Bouquet — masked 

 the burnt flavor better than did such flavors as 

 root beer oil, methyl salicylate, or anethol. 

 None of these compounds was able to mask 

 the burnt flavor completely when added to the 

 freshly distilled oil. After 6 weeks of storage 

 at 40° F (4° C) samples masked with Cherry 

 Bouquet or Fruity Bouquet had no burnt flavor, 

 but considerable remained when root beer oil 

 or methyl salicylate was used, and an interme- 

 diate amount remained when anethol was used. 



Flavoring materials liaving good masking 

 properties and giving promise for good consumer 

 acceptability were as follows: anethol, anise 

 oil, cinnamon oil, lemon oil, Lemon Juice Flavor, 

 methyl salicylate, orange oil, root beer oil. Imi- 

 tation Oil Rum, Imitation Oil Wild Cherry, and 

 Tutti Frutti Oil. 



When considerable burnt flavor is present in 

 menhaden oil, such as occurs when the two-stage 

 modified refining process is used, none of the 

 masking agents completely obscures this burnt 

 flavor. Certain fruity flavors such as Cherry 

 Bouquet or Fruity Bouquet are the most efl'ective 

 in masking this flavor, with methyl salicylate or 

 root beer oil being less effective. 



EXPERIMENT 2-CONSUMER-TYPE 

 TESTS 



The next stage of the investigation was to 

 determine consumer hedonic rating for the sev- 

 eral masking suljstances that had been found 

 to be almost equally suitable for disguising the 

 objectionable flavors developing in menhaden 

 oil during storage. 



MATERIALS 



The same materials— both menhaden oil and 

 flavor additives — were used in this phase. 



The flavors used were those that in Experi- 

 ment 1 had given the best masking results. Any 

 flavors that were used in Experiment 1 and that 

 were not approved for food additive use by the 

 U.S. Food and Drug Administration were elim- 

 inated in this part of the work. Table 2 shows 

 the identity and source of these substances. 



Table 2. — Flavor components rated in the consumer test. 



Flavor moterial 



Suppliers^ 



Anethol, U.S. P. 



Anise oil, U.S. P. 



Cinnamon oil, U.S. P. 



Lemon oil, U.S.P. 



Imitation Lemon Juice Flovor #5I.I24A 



Methyl salicylote 



Orange oil, U.S.P_. 



Root beer oil 



Imitation Oil Rum #017 



Imitotion Oil Wild Cherry #12009 



Tutti Frutti Oil 



1,4 

 3 

 3 

 4 

 2 



1,4 

 4 

 3 

 4 

 4 

 4 



1 — Fellon Chemicals Co., Inc., Brooklyn, N.Y. 



2 - Firmenich, New York, N.Y. 



3 — Florosynth Laboratories, Inc., New York, N.Y. 



4 - Frilzche Brothers, Inc., New York, N.Y. 



METHODS 



A modified molecular distillation procedure 

 was used. This procedure eliminated the ne- 

 cessity for carrying out the time-consuming 

 treatment with silica gel. The oil was distilled 

 only until the pot residue amounted to 10 % 

 (as contrasted to 3 ^r in the usual methods). 

 This decreased amount of distillation resulted 

 in much less burnt flavor (although in consider- 

 ably more burnt flavor than when a silica-gel 

 treatment was used). The oil could be used, 

 however, directly without silica-gel treatment 

 when sufficiently effective masking agents were 

 used. 



Oils were stored at (1) 40° F (4° C) (2) at 

 —20° F (—29° C) for 4 to 12 weeks. 



The panel for the consumer-type test consisted 

 of Bureau of Commercial Fisheries personnel 

 who had no previous experience in taste testing. 

 The 15 to 20 participants were asked to rate 



220 



