The spelling of names is wildly inconsistent in historical documentation. This is primarily because our 18th and 19th century ancestors lived in a culture with less literacy, and in an age when printing was so hard or so expensive that individuals might have gone years without seeing their names in print.
The agencies that we would like to consider official sources—church records, census returns, etc—offer no help. Again, the written word was less common, and its accuracy had less importance. The Golden Book of the Minorcans, for instance, dealt with a fairly finite group of family names, yet the spelling of those names varies from one entry to another, even on the same page. The census, which was taken by an interview process, has an astonishing rate of error—not only spelling the names, but the names themselves. One must remember that the purpose of the census was to count heads, not to label them (the early census didn't even use house numbers, just family units).
Gravestones are formal records that every generation took seriously, and provide the most reliable source. But they are not perfect. For one, inscriptions were dependent on the veracity and literacy of the survivors. Gravestones might have been used as an opportunity to permanently revise a historical inconvenience (an illegitimacy, for instance) or perpetuate family folk lore. In an less literate society, inscriptions were undoubtedly left to clergymen—qualified scribes, but also the same men who managed birth and marriage records with little consistency.
I personally maintain original source spellings in all of my research notes, but the high rate of variation makes the telling of a narrative with those names confusing. For this genealogy, therefore, I am using one spelling of every name across generations. In some cases, I have confidence that this it not the way the individual spelled the name in her/his lifetime, but I think it serves my purpose, which is clarity.
I am using the "-y" version for all the Italian families (Bonelly, Leonardy, Pacetty) that came to New Smyrna in 1768. This form was very common in hand-written documents in Saint Augustine's early history, but far less popular in the 20th century. Most present day Italian descendants use "-i" (Leonardi, Pacetti) and our family members had begun to make that transition as well. For the Minorcan women, I am taking my best guess—in these cases, it is less about spelling than identifying the name at all.
Those patient enough to read this whole history will appreciate the irony in having to even have this discussion. Our complicated ethnic ancestry is so dominated by women that by marriage their Mediterranean names and spellings disappeared from the family tree by the 1930s. The last Saint Augustine branch in the family were four girls named Jones. Spelling undisputed.