Sometimes it's not just an item in and of itself that provides irrefutable evidence as to its being genuine, but other similar extant items.
Case in point, this 1864 promissory note with $2 Probate of Will imperforate (Scott #R83a, cat. val. $5,000), along with a 20-cent Inland Exchange (Scott #R42c) and 5-cent Bank Check (Scott #R5c), that showed up on eBay and Hipstamp from seller momenstamps. Not incredibly material, but based upon the positioning of the manuscript cancels on the R42c and the R5c, I believe that the R5c was torn prior to it being affixed to the document and then manuscript canceled, not damage done after the fact.

Even beyond the catalog value of the stamp, examples of R83a on document are quite scarce. A February 1864 date is somewhat late for an imperf, but not egregiously so. The width across the top of the stamp is certainly large enough for an imperf, but the narrowness of the margins as you descend the stamp and the small bottom margin do give pause.
Back in 2007 at Matthew Bennett sale 319, lot #1841, this document sold for $856 including buyer premium. A relatively low amount at the time, presumably based on the uncertainty and the fact it did not have a cert.
So is it an R83a or a trimmed R83c?
Would this get a good cert standing on its own merits?
I could easily see it getting a declined opinion.
So... why then would I shell out good money for such an item without a cert?
If the above was all I had to go by, I would not have. However... I also have the document below, that I purchased from Eric Jackson in 2011. It had previously sold in 1997 for $1,680 at Spink Sale 23, lot #2506.
Same company. Same document type. Same 3 stamps. Same manuscript cancels, dated 5 days prior to the above. This R83a is large enough in all aspects to be clearly genuine.

Most importantly though for our discussion, the stamp on the second document is the sheet position immediately to the left of the stamp on the first; it contains the portion of the stamp that was cut away at lower left.
Below I show a combination image of the two R83a showing the perfect fit of the 2 stamps. As such, they each prove the other stamp genuine (even though the R83a on the second document does not require such substantiation).

Were I to try to get a PF cert on the first example, I would send BOTH documents in together for expertization, along with a print of the combination image, requesting that they be expertized in conjunction with one another, not standalone, and that they be issued consecutively numbered certificates.
It's interesting how one item can change the merits of another...
Why did I buy the top document if I already had a nicer example in the bottom one? As with matched bisect documents that are two halves of the same stamp, I like reuniting lost siblings if they can be proven such. I also like the notion that the bottom document lends credence to the top example; they work better as a pair than separately IMO. *shrug*