Dakotan gw/gm

[Rankin, Robert L]

> The main cases of *kw with any cognate are wagmuN' 'squash' and igmuN'
'cat', where the cognates show up with gdh in Dhegiha, dw in Ioway-Otoe,
and c^Vw in Winnebago.  Both words are likely to be loans - or areally
widespread forms.  The 'squash' form has some resemblants in Algonquian,
but I am personally of the opinion that it does not originate there.
 
I believe the similarities between the Siouan and Algonquian 'squash' terms show too many matching phonemes to be coincidence, especially given their geographic distribution.  Since the Algon. is reconstructible and the Siouan is not, I'd assume the progression went from Siouan to Algonquian.  
 
Algonquian:            eemehkwaan-
Siouan prototype:       wi-  kwuN      (where *wuN often dissimilates to either waN or dhuN, and where we don't know the status of vowel length in the Siouan forms because it hasn't been recorded.)
 
At any rate, I gave John credit for the discovery before he changed his mind, and further discussion appears in:
 

Rankin, Robert L.  2006.  Siouan Tribal Contacts and Dispersions Evidenced in the Terminology for Maize and Other Cultigens.  Histories of Maize: Multidisciplinary Approaches to the Prehistory, Biogeography, Domestication, and Evolution of Maize.  In John E. Staller, Robert H. Tykot, Bruce F. Benz, eds.  The Histories of Maize II: Part I: North America and Northern Mexico, Chapter 44, pp. 564-578.  San Diego, N.Y.:  Elsevier.