Consider the following example.
Possible Signs Signs Present Diagnoses Beliefs
aaaaaaa fever Mastitis 0.992
bbbb changes in milk Metabolic Diseases 0.007
cccccccccccc sudden milk drop Anaplasmosis 0.001
ddddddd wasting    

After entering 4 pieces of evidence,   the belief network  evaluates  only  4 diseases as having non-zero beliefs. Of these, mastitis has such a high belief that it appears by far the most likely scenario.

By contrast, consider the following scenario:
Possible Signs Signs Present Diagnoses Beliefs
aaaaaaa anaemia Babesiosis 0.510
bbbb fever Acute Trypanosomiasis 0.388
cccccccccccc lethargy ECF 0.090
ddddddd diarrhoea Metritis 0.008
eeeeeeeee inappetance Chronic Tryps 0.005
ffffff sudden death    

Although  6  pieces  of  evidence  have  been  entered,  and  most diseases  have been excluded from consideration,  of  those  which remain possible, 2 diseases are broadly plausible,  2 cannot be discounted  (since they have low,  but not minute belief ratings), and 2 can probably be discounted as highly unlikely. It would be very wrong to interpret this screen as stating that the correct diagnosis is babesiosis.  Rather,  babesiosis and acute trypanosomiasis should be regarded as likely causes of the observed signs; ECF and anaplasmosis as possible causes and a clinical sign  or  laboratory test  should be sought to distinguish between these different possibilities.