This is a forum devoted to phonetics, not to historical linguistics, so this discussion is slightly OT (we could continue it on Cybalist). The PIE syllabic nasal *n has the following _regular_ reflexes:
Indo-Iranian a, Baltic in, Slavic In, Greek a, Latin en, Welsh an, Germanic un, Armenian an, etc.
PIE *h2juh1nk^os > Skt. yuvas'a-, Lat. iuvencus, MWel. ieuanc, Gmc. *juwungaz > *jungaz (Eng. young).
PIE *mntis > Skt. mati-, OCSl. (pa-)mInt-, Latin ment-, Gmc. *mundiz (Eng. mind).
PIE *newn > Skt. nava, OPr. nevin-ts, Gk. (en-)ne[w]a, Lat. novem (< *nowen, cf. no:nus '9th'), Gmc. *niwun/*nigun (OE nigon > Eng. nine).
PIE *n- (negative prefix) > Skt. a-, Gk. a-, Lat. in- (< *en-), Gmc. *un- (Eng. un-), Arm. an-
In Indo-Iranian (as well as Greek) the syllabic nasal became a nasalised vowel (*a~) which then became denasalised, eventually yielding *a. Note that e.g. the negative prefix in Greek and Indo-Iranian is <a-> before consonants but <an-> before vowels (the nasal survived as a hiatus-filler). You can't see the final nasal anywhere in Indo-Aryan (including Vedic), Iranian (including Avestan) or Nuristani, because the vocalisation took place already in their common ancestor and all the historically known forms derive from Proto-IIr *nava. This form (or its derivative *nava-ka) underlies all the Indic words you list below. The loss of the unstressed final vowel produced forms like <nau>/<nav>, and the resulting diphthong has been smoothed to <no:> or <na:> in many languages. There are very similar parallel developments in Iranian. In some Indo-Aryan languages the initial nasal has caused the progressive nasalisation of the vowel, producing <no~>, <nau~> or <nam>.
The final <-h> in several Iranian languages is results from language-specific developments within Iranian (the Proto-Iranian form was *nava without a shadow of a doubt). I know little about Khotanese sound changes, but if you can wait till tomorrow I'll check that up for you.
----- Original Message -----
From: kalyan97
Sent: Wednesday, February 27, 2002 3:43 PM
Subject: [phoNet] Re: Phonetic change lo -- no in some languages

--- In phoNet@y..., "Piotr Gasiorowski" <gpiotr@i...> wrote:
> I don't rule it out... The form of <lo> rules out an early loan;
the borrowed word would have had to be something like <no:>, i.e.
_Modern_ Indo-Aryan.

The variety of phonetic forms for the numeral nine in the Indo-Aryan
languages is remarkable; what is the basis for constructing a
syllabic nasal ending for PIE *newn. (ending with a syllabic nasal).
You say that it is vocalised in Indo-Iranian; is it in Avestan?

In none of the I-A languages, not even in the Dardic languages, do we
find an ending with a syllabic nasal.

There is an intriguing lexeme nyoh in Khotanese which
connotes `nine'. [Whence y? –h from Persian? Belvalkar Vol. 94]. Is
it possible that n- in nyoh was derived from loh? The transformation
lohe > nohe is attested in Santali. lohe, nohe = v.a.disregard,
disown, disobey (Santali); nohe = is not (B.)(Santali.lex.) lo = nine
(now often heard)(Santali); lo (desi); noe (B.)(Santali.lex.Bodding) 

nava = nine (RV.Pali); n.ava (Pkt.); nau (D.); no, nu_ (Ash.); nu~_
(Wg.); nu_ (Pr.); no~_ (Dm.); na_h (Tir.); nawa, na_u, na_ (Pas'.);
nu_ (Shum.); nu~_ (Gaw.); no_ (Kal.); nyoh (Kho.); nah, num (Bshk.);
nom (Tor.); nau_ (Kand.); nau_ (Mai.); no_u (Sv.); nau, nu_, nu~_
(Phal.); nau~, na_u_ (Sh.); nav, nau, nam, na_u (K.); na_va (S.);
no~_, nau_, nao~ (L.); nau~, nau (P.); nao (WPah.); nau, no (Ku.);
nau (N.); na (A.B.Or.); naa (Or.); nau (Bi.Mth.Aw.H.); nam (H.); nova
(Omarw.); nav (G.); nav, nau_ (M.); nav (Kon.); nava (Osi.); namaya
(Si.); nuva (Md.)(CDIAL 6984). Navaka consisting of 9 (RV Pra_t.);
collection of 9 (R.); n.avaga (Pkt.); nomu (K.); nawwa_, namma_ (H.);
navvo the 9 in cards (G.)(CDIAL 6985).