Dear Ramana murthy,
In many Indian languages, the anusvāra (ṃ) sound represents a homorganic nasal, not just 'm' sound. For example, it is pronounced as the dental nasal /n/ (ನ್) before dental consonants, as the bilabial nasal /m/ (ಮ್) before bilabial consonants, etc. Such usage is found not just in Kannada, but in Telugu, Marathi and even in Hindi. Please note that in many north Indian languages, where loss of nasal sound is common, the anusvāra often simply represents the vowel nasalization, and therefore you might find the symbols of anusvāra and anunāsika (candrabindu) used interchangeably (although not in the South Indian scripts such as Kannada and Telugu).
Coming back to Kannada script:
aṅga is written as ಅಂಗ (aṃga) and pronounced as ಅಙ್ಗ (aṅga)
kañca is written as ಕಂಚ (kaṃca) and pronounced as ಕಞ್ಚ (kañca)
kaṇṭha is written as ಕಂಠ (kaṃṭha) and pronounced as ಕಣ್ಠ (
kaṇṭha)
danta is written as ದಂತ (daṃta) and pronounced as ದನ್ತ (danta)
pampa is written as ಪಂಪ (paṃpa) and pronounced as ಪಮ್ಪ (pampa)
Such simplified representation of nasal in nasal-consonant clusters using a single sign was found even in some of the earlier Prakrit inscriptions, and, I believe, is a brilliant innovation in simplifying the scripts by exploiting the homorganic nature of the nasal sounds.
Regards,
Suresh.