diff --git a/image_analysis/geom_refinement/PostRefine.cpp b/image_analysis/geom_refinement/PostRefine.cpp
index 681bca17..7e952786 100644
--- a/image_analysis/geom_refinement/PostRefine.cpp
+++ b/image_analysis/geom_refinement/PostRefine.cpp
@@ -149,6 +149,12 @@ PostRefineResult PostRefineRotationGeometry(const std::vector<IntegrationOutcome
         // in one pass moved 56 bytes per element through every level of a comparison sort, on one
         // thread, over tens of millions of reflections.
         {
+            // Not a total order: two partials of one reflection on one image still tie, as they did
+            // before this was bucketed. What makes the result reproducible is the scatter below
+            // rather than the comparator - the prefix lays each bucket out chunk by chunk and a
+            // chunk is a contiguous span of the gathered order, so a bucket reaches std::sort in
+            // global gather order whatever the thread count. Ties therefore resolve the same way on
+            // every run and at every -N; they are simply not resolved by rank.
             const auto part_less = [](const Partial &a, const Partial &b) {
                 if (a.h != b.h) return a.h < b.h;
                 if (a.k != b.k) return a.k < b.k;