diff --git a/doc/bugs/annex_merge__breaks_git_repository__33__/comment_4_c818e240e0d0ed8da3733e574694cc44._comment b/doc/bugs/annex_merge__breaks_git_repository__33__/comment_4_c818e240e0d0ed8da3733e574694cc44._comment
index 4e25ea3ec6..39a8f71865 100644
--- a/doc/bugs/annex_merge__breaks_git_repository__33__/comment_4_c818e240e0d0ed8da3733e574694cc44._comment
+++ b/doc/bugs/annex_merge__breaks_git_repository__33__/comment_4_c818e240e0d0ed8da3733e574694cc44._comment
@@ -6,8 +6,4 @@
 Spotchecked a few other OpenNeuro datasets in the same numeric range and
 they seem ok, so this may have been a 1-off problem. It would be good to
 check all 1.1k datasets.
-
-One possible way this might happen: git-annex is performing a transition,
-and gets interrupted after making the commit starting the new git-annex
-branch, but before it can graft in the export trees.
 """]]
diff --git a/doc/bugs/annex_merge__breaks_git_repository__33__/comment_5_0998f6d67fc6327ecbbb3c23ad7f5275._comment b/doc/bugs/annex_merge__breaks_git_repository__33__/comment_5_0998f6d67fc6327ecbbb3c23ad7f5275._comment
new file mode 100644
index 0000000000..5e9fb1bc0d
--- /dev/null
+++ b/doc/bugs/annex_merge__breaks_git_repository__33__/comment_5_0998f6d67fc6327ecbbb3c23ad7f5275._comment
@@ -0,0 +1,16 @@
+[[!comment format=mdwn
+ username="joey"
+ subject="""comment 5"""
+ date="2024-06-03T17:28:02Z"
+ content="""
+performTransitionsLocked, when `neednewlocalbranch = True`,
+first writes the new git-annex branch, and then calls
+regraftexports which adds a second commit onto it.
+
+In the window before regraftexports finishes, interrupting git-annex will
+leave the repository in this state.
+
+There may be some other way this could happen, but that seems like a likely
+cause. It needs to avoid updating the git-annex branch ref until it's
+grafted the exports into it.
+"""]]