Zulu, and the other https://en.wikipedia.org/wiki/Nguni_languages as user@6726 reports in his answer there, capitalise the root of proper nouns, and leave their prefixes uncapitalised; this results in camel-case like isiZulu and uDavida. When and how was this convention introduced?
I'm building on https://linguistics.stackexchange.com/a/27752/17064, @user6726's answer to the related question.
The very earliest renderings of Zulu in Roman script, as @user6726 reports, used normal Roman capitalisation conventions. @user6726 reports that
In Bryant's 1848 article "The Zulu language", citation names are given as e.g. Umpandi, Unyokhana, and Grout 1849 "The Zulu and other dialects of southern Africa" likewise gives Untaba, Ubalekhile.
On the other hand, Bishop John Colenso's grammar written in the same year uses the modern camel-case convention. And little hints I see in Wikipedia and Google show that inconsistency in Zulu capitalisation persisted for a while.
What seems to be an explanation occurs in the 1868 book Nursery Tales, Traditions, & Histories of the Zulus by Henry Callaway, which uses the older capitalisation, and rejects camel-case:
There are two modes of writing—one adopted by Dr. Colenso and Dr. Bleek, in which a number of small words is run together; and the other, that adopted by the American missionaries and others, in which there is, perhaps, the opposite mistake of unnecessary division.
As regards the first, I am quite unable to see anything to recommend it, or even to conceive the reason of its adoption. Why should we write ngabebabopa, "they ought to bind them"; and not nga be ba bopa, "ought they them bind?" Why should we run the Zulu words together, when we write the English ones apart? (p. ii)
A difficulty, too, has been felt as regards the capital letters; and we find consequently in printed books some ugly anomalies, such as a capital in the middle of a word, and paragraphs beginning with a small letter. This has arisen apparently, in part, from the error of not regarding the prefix as an essential part of a noun, and so giving the nominal root an undue prominence; and, in part, from our not being accustomed to those initial changes upon which grammatical inflection so much depends in the Zulu language. But to use the capital letters to distinguish nominal roots is a novelty in writing (p. iii)
So, Grout wrote Zulu affixes as separate words, Colenso as affixes (something Callaway disagreed with). And (I assume!) Colenso also came up with the camel-case orthography, since he was in effect joining up what Grout had written as two words (u Davida, n Kosi, isi Zulu) into single words (uDavida, nKosi, isiZulu).