Re: Electronic Fat test for food?

"Crude fat" or "fat" is a somewhat difficult test to perform.

One traditional way is to chop up the food and then extract it with an 
organic solvent. The solvent is then driven off under heat and the 
remaining "fat" weighed. Dividing this number by the weight of the sample 
gives the % fat in the food. Examples include the Soxhlet and Goldfisch 
methods. These methods are considered the "gold standard" for measuring 
fat.

Another way is to release the fat by acid hydrolysis (digestion with 
acids). The fat goes to the top and is measured in a calibrated tube 
attached to the special bottle used. Examples include the Babcock and 
Gerber methods.

You can also extract with an organic solvent and then use a high-accuracy 
hydrometer to measure the specific gravity of the extract. This method is 
not used much any more because of health and ecological problems with 
solvents. 

Nowadays, food companies typically use techniques based on instrumental 
methods. For example, near-infrared absorption is used on several types of 
foods to get moisture, fat and protein simultaneously. Differential 
nuclear magnetic resonance absorption (similar to hospital MRI techniques) 
is used to measure "fat" directly. The equipment for these tests is fairly 
expensive, costing about $20,000-$40,000.

Another method is to measure moisture, protein and salt and subtract from 
100% to get fat. This works for foods with no carbohydrate content, such 
as meat.