Most languages of South Asia make a distinguish between dental and retroflex consonants, most commonly [t] vs. [ʈ], [d] vs. [ɖ] and [n] vs. [ɳ]. The distinction is consistently made in scholarly transliterations, where retroflex consonants are marked with an underdot, for example [ṭ] for [ʈ]. But things get a bit murky when it gets to common everyday romanisation. A common practice in South India is to make avail of a following letter h to indicate a dental consonant: th for [t] vs. t for [ʈ]. This is almost always applied to t, less often to d and never, as far as I know, to other consonants. This practice isn't followed in North India, probably because an h is already employed to mark aspirated consonants (t for [t] vs. th vs. [tʰ]). In the south such consonants are either absent (as in Tamil), or much less frequent.
