"bow"

[Rankin, Robert L]

Siouan has a good term for 'arrow' or some antecedant weapon also, presumably atlatl darts. 
 
Biloxi aNksapixti means 'real bow' (xti is the usual suffix for 'real' when there has been a replacement.).  ANksi, as David points out, is the 'arrow' root.  This leaves -api unexplained.  Note that Ohio Valley Siouan has no trace of the MVS pluralizer -api, so that interpretation is hors de combat.  I wonder if it could be from the Algonquian term Dave Costa cites just below?  I'm not 100% happy with the idea since Siouan usually just takes the first one or two syllables of those interminably long Algonquian words, but at least we have a look-alike.
 
Bob
 
> The other main term for the concept can be reconstructed variously as PA 
*/a?ca:pyi/ (usually 'bowstring') or */a?ta:pya/ 'bow' (the form found in 
Eastern). That has that same 'string' final as */me?tekwa:pyi/, but the 
front half is unfamiliar. That one *is* found in Eastern Algonquian (and 
Central), but it has several unresolved phonological glitches in the 
daughter languages. The fact that it doesn't reconstruct cleanly and that it 
has an odd geographic distribution means it's probably not a normal 
Proto-Algonquian etymon either, like it was formed later in the post-PA 
period and passed around the family after the languages were already 
dialectally differentiated.

> However, Proto-Algonquian does have a totally reconstructible form for 
'arrow' (when possessed): PA */ni:pi/ 'my arrow', */ki:pi/ 'your arrow', 
*/wi:pi/ 'his arrow', etc. It's found throughout the family. The basic stem 
is PA */-i:p-/, which is not internally analyzable and looks quite old.