There appears to be a malfunction in the software. There are several duplicate sources for his daughter Hannah. If you delete them, hints will appear in the research help. There is no way to get rid of them.
@Dean, Barton David
I am just another 'lowly' User/Patron ...
Just in passing ...
Short Answer: You DO NOT 'get rid of them', you "Attach" ALL of them ...
I would humbly suggest, that there is NO problem/issue ...
I just took a 'look' at Josiah STAFFORD ( MZW5-ZTL ).
[ In particular the "Sources" 'Tab': https://www.familysearch.org/tree/person/sources/MZW5-ZTL ]
I would humbly suggest, that those "Sources" are NOT "Duplicates"; as, you suggest/consider; but, are, in fact, simply ADDITIONAL "Indexings" of the SAME "Record" (ie. "Event"); as, there are NO "Duplicate" URLs.
Such is common; but, maybe, not to the extent, here; which, is exacerbated, by BOTH, "Christening" in 28 January 1798; and, subsequent "Burial" on 13 March 1798, just over a Month and a Half later.
"Records" (ie. "Events") can; and, are often, "Indexed" MORE than ONCE, for whatever reason (and, there are many).
Plus, the "Records" (ie. "Events"), MAY even be "Indexed", by MORE than ONE organisation.
[ eg. 'FamilySearch'; "Ancestry_com"; "FindMyPast_co_uk" [ FindMyPast_com" ]; "MyHeritage_com"; etc ... ]
And, many of those DIFFERENT "Indexings", of the SANE "Record" (ie. "Event"), can find their way into 'FamilySearch'.
Sometime the "Indexing", comes from the actual "Parish Registers"; whereas, other times, the "Indexing", comes from "Bishops' Transcripts"; plus, some can be ADDITIONAL "Copies", that are in the 'Keep' of various "Record Custodians".
That Said ...
As, NONE of those "Sources" are, in fact, a "Duplicate" URLs; and, 'Yes', as you have discovered, if "Detached", they will REAPPEAR; as, a "Record Hint" (AGAIN), such is WHY they NEED to be "Attached".
I am not concerned that there are similar "Indexings" of the SAME "Record" (ie. "Event").
In fact, I like to say ... the more, the merrier ...
What I do, is put ALL the Sources in 'Date' Order; and, the, for the similar "Sources" for the SAME "Record" (ie. "Event"), I simple ADD "Reference x of y", to each; so that, it is easy to 'see', that there are a number of "Sources" for the SAME "Record" (ie. "Event").
Many Users/Patrons, do not like the fact, that there can often be, ADDITIONAL "Indexings" of the SAME "Record" (ie. "Event").
Personally, I really think, for many such Users/Patrons, that such equates to just unnecessary work, on their part; as, they need to "Attached" ALL those "Sources".
Genealogy/Family History is NOT a "Sprint" ...
Genealogy/Family History is a "Marathon" ...
GOOD Genealogy/Family History takes BOTH, 'Time'; and, 'Effort'.
Many Users/Patrons these days, do not want to do, the necessary 'hardy yards', taking the, 'Time'; and, 'Effort'
[ They just want it done - quick; and, easy ... ]
Just my thoughts.
Whilst Brett gives a correct explanation of the situation you have encountered, I have always held a different opinion to him regarding sources that, in many cases, contain completely identical information.
I have examples involving different filming of the same material, attaching which offers me no benefit (due to the identical data) and serves to "clog up" the Sources sections of the individuals referenced. It was FamilySearch's intention, some years back, to "retire" some of these records (i.e. everything in common with another source except its URL reference), but this proved to be too difficult a challenge. Firstly, there was the problem of exactly how a person (or computer program) could weed-out these "unneeded" duplicates and, secondly, the closely connected problem that the retired "duplicate" just might contain some additional detail its "counterpart" did not include.
So whilst "in principle" I would like to be able to discard / merge / detach any sources that truly have identical origins and data, in practice I see the reasons why this is just not going to happen. Unfortunately, this issue will no doubt continue to be an annoyance to many of us, as (for reasons explained) there is no real prospect of the matter being addressed.
However, for the future, it would be good if any items planned to become part of a new indexing project could be checked to ensure the same work has not already been undertaken. This would save time and effort that could be dedicated to getting truly new material online, rather than adding this duplicate data.