diff --git a/docs/now/2026-09-24-published-test-the-1-untested-function-in-specs-fpga-testbench-memory-.md b/docs/now/2026-09-24-published-test-the-1-untested-function-in-specs-fpga-testbench-memory-.md
new file mode 100644
index 0000000000..de1a70b34c
--- /dev/null
+++ b/docs/now/2026-09-24-published-test-the-1-untested-function-in-specs-fpga-testbench-memory-.md
@@ -0,0 +1,11 @@
+# NOW -- Test the 1 untested function in specs/fpga/testbench/memory_tb.t27 (published 2026-09-24)
+
+## A bee's work on #4693, published from `queen-4693` (Closes #4693)
+
+- The branch changes 1 file(s): `specs/fpga/testbench/memory_tb.t27`.
+- `git diff --stat origin/master...queen-4693` reads: 1 file changed, 8 insertions(+), 1 deletion(-)
+- This entry is written by the publisher, not by the bee. A pull request must
+  add exactly one `docs/now/` entry and a bee has no way to know that: its brief
+  names a boundary file and acceptance criteria, and `docs/now/` is neither.
+- What this entry does NOT establish: that the work is correct. The gates on the
+  pull request judge that, and they are the same gates every other change meets.
diff --git a/specs/fpga/testbench/memory_tb.t27 b/specs/fpga/testbench/memory_tb.t27
index bfe3950af2..8c11fcc387 100644
--- a/specs/fpga/testbench/memory_tb.t27
+++ b/specs/fpga/testbench/memory_tb.t27
@@ -129,8 +129,15 @@ module Memory_Testbench {
         // After second tick, clock should be false again
         tick();
         invariant clk == false;
+}
+    
+    test test_on_comb_returns_mem_read {
+        reset();
+        mem_write(0x100, 0xDEADBEEF);
+        var val : u32 = on_comb(0x100);
+        invariant val == 0xDEADBEEF;
     }
-
+    
     bench bench_memory_bandwidth {
         reset();
         var i : u32 = 0;