diff --git a/src/app/features/galaxy-system/system-orbits-renderer.spec.ts b/src/app/features/galaxy-system/system-orbits-renderer.spec.ts
index 0c01ec2..333be1e 100644
--- a/src/app/features/galaxy-system/system-orbits-renderer.spec.ts
+++ b/src/app/features/galaxy-system/system-orbits-renderer.spec.ts
@@ -442,8 +442,18 @@ describe('rotation without IAU elements', () => {
     expect(spinSense(spinning({ rotationPeriodHours: 23.934 }))).toBeGreaterThan(0.99);
   });
 
+  it('turns Nereid, as shipped, once in the 11.594 hours Kepler measured: a sixth of a turn in 1.93 hours', () => {
+    const shipped: BodyRecord[] = JSON.parse(readFileSync(`${process.cwd()}/src/assets/data/bodies.json`, 'utf8'));
+    const renderer = new SystemOrbitsRenderer(shipped.filter((body) => body.id === 'neptune' || body.id === 'nereid'), []);
+    const nereid = renderer.members.find((member) => member.id === 'nereid')!.marker;
+    renderer.update(DEFAULT_EPOCH_JD);
+    const start = nereid.quaternion.clone();
+    renderer.update(DEFAULT_EPOCH_JD + 11.594 / 6 / 24);
+    expect((nereid.quaternion.angleTo(start) * 180) / Math.PI).toBeCloseTo(60, 1);
+  });
+
   it('leaves a body with no published rotation still', () => {
-    // Titan: Horizons states no period for it, and an invented one would be a claim.
+    // Hyperion, which tumbles: an invented period would be a claim.
     const renderer = new SystemOrbitsRenderer([spinning({ rotationPeriodHours: undefined })], [], undefined, 1);
     renderer.update(DEFAULT_EPOCH_JD);
     const start = renderer.members[0].marker.quaternion.clone();
diff --git a/src/app/features/galaxy-system/system-orbits-renderer.ts b/src/app/features/galaxy-system/system-orbits-renderer.ts
index 4d4084e..aafff2f 100644
--- a/src/app/features/galaxy-system/system-orbits-renderer.ts
+++ b/src/app/features/galaxy-system/system-orbits-renderer.ts
@@ -297,9 +297,9 @@ const SPIN_AXIS = new THREE.Vector3(0, 1, 0);
 const HOURS_PER_DAY = 24;
 
 /**
- * How a body the IAU gives no rotational elements for is turned at a given date — Eris, Haumea
- * and Makemake, whose periods are measured (Makemake's only to a factor of two, see its spec in
- * `fetchSolarSystem.ts`) and whose poles are not: at its own sidereal rate, about
+ * How a body the IAU gives no rotational elements for is turned at a given date — Eris, Haumea,
+ * Makemake and Nereid, whose periods are measured (Makemake's only to a factor of two, see its
+ * spec in `fetchSolarSystem.ts`) and whose poles are not: at its own sidereal rate, about
  * its orbit's normal, backwards for a negative period. None of them has an obliquity, so none is
  * applied. The phase is arbitrary: each body starts at its elements' epoch in the shortest
  * rotation of +Y onto its axis, and turns from there. Exoplanets have no published rotation at
diff --git a/src/app/shared/models/body.model.ts b/src/app/shared/models/body.model.ts
index 25fe0f8..8fd579d 100644
--- a/src/app/shared/models/body.model.ts
+++ b/src/app/shared/models/body.model.ts
@@ -88,8 +88,10 @@ export interface BodyRecord {
    * Horizons gives a negative rate (Venus, Uranus), and the tilt of that axis from its orbital
    * plane — which past 90 degrees already says the turn is retrograde.
    *
-   * Absent where Horizons publishes neither — the view then leaves the body still rather than
-   * spinning it at an invented rate.
+   * For a locked moon the period is its orbit's, from the mean motion that carries it round; for
+   * a body whose source states none, the measured one its ETL spec carries (Eris, Nereid). Absent
+   * only for Hyperion, which tumbles — the view leaves it still rather than spinning it at an
+   * invented rate.
    */
   rotationPeriodHours?: number;
   obliquityDeg?: number;
diff --git a/src/assets/data/bodies.json b/src/assets/data/bodies.json
index 2fde121..fe24818 100644
--- a/src/assets/data/bodies.json
+++ b/src/assets/data/bodies.json
@@ -2130,7 +2130,8 @@
       "decDeg": 69.117
     },
     "orbitSource": "JPL SSD satellite mean elements, epoch 2000 Jan 1, within 11.2 degrees of Horizons from 1950 to 2100",
-    "parentBodyId": "neptune"
+    "parentBodyId": "neptune",
+    "rotationPeriodHours": 11.594
   },
   {
     "id": "proteus",
diff --git a/tools/etl/build.ts b/tools/etl/build.ts
index e0f8990..b5ceba1 100644
--- a/tools/etl/build.ts
+++ b/tools/etl/build.ts
@@ -204,6 +204,13 @@ const TRACK_OFFSET_CEILINGS_DEG: Record<string, number> = { hyperion: 23, nereid
  */
 const WITHOUT_ROTATIONAL_ELEMENTS = new Set(['hyperion', 'nereid', 'eris', 'haumea', 'makemake']);
 
+/**
+ * The one moon drawn still: Hyperion, whose page says "Rotational period = Chaotic". Every other
+ * moon without a lock has a measured day; Nereid's page states none, and it was drawn still until
+ * its K2 light curve's 11.594 hours was taken (see its spec).
+ */
+const TUMBLING = new Set(['hyperion']);
+
 /**
  * How far the IAU's day, 360 degrees over W's rate, may be from the one Horizons states, as a
  * fraction of it. Measured on this catalogue: at most 1.8e-5 (Jupiter's System III, 9.92492 hours
@@ -357,8 +364,12 @@ function validateBodies(bodies: BodyRecord[], horizonsOrbits: Map<string, Orbita
       assertCondition(parent !== undefined, `Moon ${body.id} has no valid parentBodyId.`);
       const orbitHours = (360 / body.rates.meanMotionDegPerDay) * 24;
       if (FREELY_SPINNING_MOONS.has(body.id)) {
-        // Hyperion tumbles and Nereid's page gives no spin, so they have none; Phoebe turns in
-        // 9.27 hours against a 550-day orbit. A lock here would be the rule below misapplied.
+        // Hyperion tumbles, and has no period; Nereid turns in 11.594 hours against a 360-day orbit,
+        // and Phoebe in 9.27 against 550 days. A lock here would be the rule below misapplied.
+        assertCondition(
+          body.rotationPeriodHours !== undefined || TUMBLING.has(body.id),
+          `Moon ${body.id} is drawn not turning, and is not known to tumble: its day was measured somewhere, find it.`
+        );
         assertCondition(
           body.rotationPeriodHours === undefined || Math.abs(body.rotationPeriodHours - orbitHours) > orbitHours * 0.1,
           `Moon ${body.id} does not keep one face to its planet, yet turns once in ${body.rotationPeriodHours} hours against an orbit of ${orbitHours}.`
diff --git a/tools/etl/fetchSolarSystem.ts b/tools/etl/fetchSolarSystem.ts
index cfe4239..984fb79 100644
--- a/tools/etl/fetchSolarSystem.ts
+++ b/tools/etl/fetchSolarSystem.ts
@@ -147,8 +147,10 @@ const BODY_SPECS: BodySpec[] = [
   { id: 'titania', name: 'Titania', horizonsCommand: '703', ...URANUS_MOON },
   { id: 'oberon', name: 'Oberon', horizonsCommand: '704', ...URANUS_MOON },
   { id: 'triton', name: 'Triton', kind: 'moon', horizonsCommand: '801', center: '500@899', parentBodyId: 'neptune' },
-  // Nereid's eccentric orbit, 0.75, cannot hold a face to Neptune; its page states no spin.
-  { id: 'nereid', name: 'Nereid', kind: 'moon', horizonsCommand: '802', center: '500@899', parentBodyId: 'neptune', spinsFreely: true, trackStepDays: 1 },
+  // Nereid's eccentric orbit, 0.75, cannot hold a face to Neptune. Its page states no spin, but
+  // Kepler's K2 light curve gives 11.594 +/- 0.017 hours, confirming the short periods measured
+  // from the ground (Kiss et al. 2016, MNRAS 457, 2908; arXiv:1601.02395). No pole is known.
+  { id: 'nereid', name: 'Nereid', kind: 'moon', horizonsCommand: '802', center: '500@899', parentBodyId: 'neptune', spinsFreely: true, rotationPeriodHours: 11.594, trackStepDays: 1 },
   { id: 'proteus', name: 'Proteus', kind: 'moon', horizonsCommand: '808', center: '500@899', parentBodyId: 'neptune' },
   // Pluto's section prints its epoch as 2000 Jan 1.0; JPL's current table gives Charon's as
   // 2000-01-01.5, and read at 1.0 Charon sat 27.8 to 28.2 degrees — half a day of its motion is