My paternal Aunt LaWanda explained to me back in 1977 that family tradition states that the first spelling of our surname was "Denis" as opposed to "Dennis". The additional "n" was entered later and never changed.
No explanation for this change is offered for this diverse spelling. I present that it is possible that the Denis was from French Canadian who drifted southward down the Atlantic Coast. I am unable to bring any evidence to the table, but believe that the conflict between French and English settlers may have pressed the family to insert the second "n".
The first Dennis ancestor was William Dennis (1797). No middle name is given. His birthplace is uncertain, but the birthplace of one of his sons, William Granville Dennis (1820), indicates Virginia was his birthplace. This may also apply to at least one other child, Almarine (1816), as well.. Other children include Jonathan (1829), Sampson David 1836), George (1831), Joshua B (1838)., Chesley (1839), and Malinda (1840). There is the possibility that another son, Eli (1826). In review all show in the 1850 Census of Morgan County, AL show their birthplace as Tennessee, except for William Granville Dennis (alternative: Granville William). William married Ruth A. Pettit (1797) in Knox County on 9 Feb 1818.
This leaves some disparity in believing Almarine was the son of William and Ruth, inasmuch as he was born in 1816 in Tennessee. The evidence would be supported by the marriage date recorded in Knox County, TN on 08 Feb 1818. The date is recorded twice in the marriage book. Once on the 8th, the other on the 9th of the same month.. The identity of Almarine and his family line is uncertain, but probably may be a younger brother of William or a cousin.