No effort was made to counteract the color factor since tasters 

 had to know the species of salmon to make an intelligent evaluation 

 of the flavor. Any prejudice for the more highly colored salmon, 

 as red or king salinon, would still influence the flavor score, even 

 though the color variable had been controlled if the taster knew the 

 species of salmon being tested. 



The average flavor ratings obtained at the ei^t examination per- 

 iods may represent seme degree of bias. The factor of color and 

 species bias seems to be ruled out by the fact that the second and 

 seventh examination of king salmon and the seventh examination of red 

 salmon brought out as great spreads in flavor scores as were I'eported 

 for chum salmon, wiiich is generally classed as the least popular 

 salmon on the basis of color and species. 



The flavor scores for the king salmon samples show a more consist- 

 ent trend in favor of the lot to vrtiich alkali process oil was added. 

 Lower scores were given to the samples containing the boiling process 

 oil. There was a general drop in the total average scores from 9.1 at 

 the first to 7.0 at the last examination. 



The pink salmon samples were judged equally often in favor of the 

 control lot. There was very little to choose between the ratings of 

 the two lots containing added oil. Again the total average scores 

 per examination declined from an initial 7.9 to 6.8 at the eighth 

 examination. 



For Coho salmon, the lot with added oil prepared by the boiling 

 process made the best showing with the taste panel. The pack, contain- 

 ing the alkali process oil was the least favored. The same decline in 

 total average scores was noted, with a change from 8.9 to 6.8 in the 

 flavor rating. 



The control saunples were scored highest in six of the eight exami- 

 nations of the red salmon packs. The lot containing the boiling 

 process oil was almost as consistently rated least acceptable. The 

 decline in flavor score over the entire period of the tests was con- 

 siderably less than in the previous three species; initial average 

 score was 7.7 and final score 7.2. 



The chum salmon series resulted in top flavor rating for the 

 control samples. There was little to choose between the packs having 

 added oil. The total average flavor score of 7.2 at the first examina- 

 tion declined to 6.3 at the eighth examination. 



Summary 



Experimental packs of each of the five species of salmon were pre- 

 pared with added salmon head oil made by two procedures: 1. by alkali 



^s 



