370 FOOD SCIENCE APPLICATIONS 



representative samples of many different lots of fish representing such 

 differences as fish caught at different times of year, in many different 

 geographical areas, and in different years would be required. In such an 

 instance, a minimum sample of six fish from each of as many as 100 

 different batches representing the different seasonal and geographical 

 variables would be needed to establish the range and average proximate 

 composition. 



Because of the relatively great heterogeneity of the composition of fish, 

 proper sampling is often the major criteria as to whether the analysis is 

 meaningful. Errors in achieving representative samples are usually many 

 times greater than those involving the analytical procedure itself. Careful 

 consideration of proper sampling is, therefore, mandatory if results hav- 

 ing any real meaning are to be obtained. 



After selection has been made of a suitable number of fish for a repre- 

 sentative sample, the problem of mixing the individual fish to give a 

 suitable sample for chemical or other analysis is posed. Where the entire 

 fish is to be sampled, the problem of reducing the flesh, skin, and bones 

 to a homogeneous mass presents a problem involving use of suitable 

 grinding and mascerating techniques. Large bones (in head or backbone) 

 can, in particular, cause difficulties. Where cooking of the fish will not 

 affect the resulting analyses, a steam processing to soften the bones will 

 often solve this problem. Otherwise, rather powerful, large grinders of the 

 hammermill or similar types are needed for preliminary grinding of bone, 

 skin, and flesh. Skin may pose a problem with some types of grinders 

 since it has a tendency to strip off and wind around conveyer screws. 



For reducing fish flesh, free of skin and bone (or whole fish which has 

 been given a preliminary treatment in a suitable mill) to an impalpable, 

 homogeneous state, suitable for analytical use, a grinder adapted to cut- 

 ting up pieces of tendon and connective tissue prevalent in fish is needed. 

 For this purpose cutting-type grinders, such as are sometimes used in 

 making sausages, are preferable to the type employing a wheel revolving 

 against a plate with small holes (such as the ordinary kitchen meat 

 grinder). The cutting type is better adapted to homogenizing small pieces 

 of connective tissue, skin, and tendon. 



For very soft materials such as fish liver and, in some cases, skin- and 

 bone-free fish flesh, the blender type of equipment is suitable especially 

 where a slurry is to be made and water or other liquid can be added. 



An alternate method which sometimes has proved to be useful in sam- 

 pling fish for chemical analysis is the use of ''press juice." The fish (usu- 

 ally the flesh) is placed in a small hydraulic press inside canvas and 

 pressure exerted. The resulting fluid is used for analysis. Use of press 

 juice is usually satisfactory for determining water soluble components. 



