The Advertiser, Adelaide, 8 April, 1912.
UNWATCHED LIGHTS.
To the Editor.
Sir--In our childhood we have heard or
read more or less mythical stories of ships
lured to their doom in the bad old days of
smugglers and wreckers by misleading
lights. At the present time we have, or
may have, somewhat parallel cases. I refer
to the latest innovation in coast lights, the
so-called "long-burning, un-watched A.G.A.
lights." These lights are supposed to be
absolutely reliable; but can that be said
of, or can it be true of, anything human
or made by human hands? I write you,
to point out that they constitute a very
grave menace to life and property. Such
a light exists on Bedout Island, where the
ill-fated Koombana appears to have gone
to her doom, and who will say how much
this light, or its absence, may have contributed
to the terrible disaster which we are at present
lamenting? Some time back this Bedout Island
light was reported to be out. About two years
ago the Sydney owned barque Hippolas was
lost in Cook's Strait, the loss being attributed
to a similar light on Jackson's Head being
extinguished. On the 11th ultimo, in Spencer
Gulf, being close to Middle Bank light, and
not being able to see it, I anchored rather
than proceed, through this "reliable un-watched
light" being out. It is proposed to place a similar
light on Citadel Island. at Wilson's Promontory,
a locality where fogs and bad weather are very
prevalent. These lights, in my opinion, and in
that of other shipmasters with whom I have
conversed, are not reliable, and should not be
placed in important and dangerous positions,
such as Bedout Island, for example. Economy
against human lives that is what it amounts to,
and which is to obtain. In calling attention to this
matter I trust that some abler pen than mine will
take the matter up and ventilate it thoroughly.
I venture to say that no Light should be un-
watched. Better no light at all, far better, than
an unreliable one, which may cause disaster
and the loss of precious human lives,
I am etc..
G. PROVO,
Master S.S. Dickera.
Newcastle, April 3, 1912.
I don't think nearly enough attention was paid during the Inquiry to the fact that the Light at Bedout was not functioning at the time.
Captain Irvine, Chief Harbour Master, Fremantle, in charge of Harbours and Lights, submitted a rather implausible excuse that he was not aware of the Light being out until 1 week after the disaster.
In fact Irvine was disingenuous about the whole affair and stated in the March gazette that the cyclone was responsible for the light being 'extinguished':
'Western Australia—North-West Coast. NOTICE is hereby given that the hurricane which passed over Bedout Island—lat. 19.35 S. 119.6 E.—on the 22nd inst., extinguished the unattended light established on this island in December, 1909. Further notice will be issued as soon as the lamp has been re-lighted. Charts affected. No. 1055—Bedout Island to Cape Cuvier. No. 1048—Buccaneer Archipelago to Bedout. C. J. IRVINE, Chief Harbour Master. Harbour and Light Department, Fremantle, 30th March, 1912.'
Mr. Moxon, Adelaide Steamship Company, affirmed that he had been made aware of the fact that the Bedout Light was out 1 week prior to the disaster. Given that Mr. Moxon knew about the Light being out we might assume that the information was passed both to Captain Irvine and Captain Allen (via Mr. Gardiner, manager of the Adelaide Steamship Company at Port Hedland).
Mr. Moxon, Adelaide Steamship Company, affirmed that he had been made aware of the fact that the Bedout Light was out 1 week prior to the disaster. Given that Mr. Moxon knew about the Light being out we might assume that the information was passed both to Captain Irvine and Captain Allen (via Mr. Gardiner, manager of the Adelaide Steamship Company at Port Hedland).
Departing Port Hedland that fateful day, Koombana would likely have arrived at a holding position to fill tanks 3 miles (ref. cattleman on board Bullarra) north of Port Hedland by about 11 a.m.. Being aware that the Light was out, Captain Allen had to reach Bedout Island before nightfall - dusk at this time of the year is 6.38 p.m.. To cover the distance of '50 miles' would have taken the steamer, averaging 13.5 knots (operating speed) 3.5 hours, which was well within Koombana's capability in normal conditions. This would have allowed for a more reasonable 4 hours to fill tanks.
However, the gale-force headwind and a periodically racing propeller, had potential to slow the steamer's progress to about 10 knots. Why 10 knots? Captain Allen claimed that he was unlikely to make the spring tide access into Broome by the following morning, which translates into a 10 knot or less, average speed. This would have increased the time frame to 4.5 + hours and Koombana to have completed the filling of tanks by 2 p.m. = 3 hours. This is barely the time quoted by Captain Upjohn for filling two after tanks, under ideal conditions - and certainly not in heavy seas.
Captain Upjohn:
'How long would it take to fill them (tanks) ? - From 3 to 3 1/2 hours.'
Koombana was only in sight for approximately 2 hours, suggesting that Captain Allen was particularly anxious to clear Bedout before dusk, and in so doing could not have allowed adequate time to press up all the tanks.
“Normally,” he (Bert Clarke) recalled, “ships going northbound were out of sight within 30 or 45 minutes, but this day as the storm was getting stronger I stayed up there in the tower watching the Koombana pitching and rolling for nearly two hours.”
Boyd, Annie. Koombana Days.
Captain Upjohn:
'And you last saw her? - About two hours after leaving.'
I believe the incomplete filling of ballast tanks and a shortfall in the fresh water tanks (not replenished at Port Hedland) created what is known as a free surface effect which could have disastrous effects on the stability of a vessel, particularly a lightly laden, top heavy one.
"Free surface effect. When a tank is partially filled, the liquid's centre of gravity position will change as the ship is inclined. Liquid in partially filled tank always decreases the initial metacentric height GM, righting lever GZ, and angle of vanishing stability."
Not only would this scenario have contributed to a reduced GM; reduced righting lever (ability to return to the upright after heeling) and reduced vanishing stability (the angle at which the steamer was likely to roll over), but would also contributed to a persistent list, such as that described when Koombana departed Port Hedland
The scene was set for disaster.
There is some speculation as to the exact steamer route taken by Koombana, 'shaping a course' around the northern aspect of Bedout Island.
How can it be assumed that the steamer track to Bedout Island from Hedland would follow that which I have presented on the Google Earth image? Firstly let's take another look at a period newspaper representation of the course taken:
 |
| Although not substantiated by evidence, Koombana might 'last (have been) seen here'. |
We know from Captain Upjohn's statements (and other period sources) that after filling tanks 3 miles north of Port Hedland, Koombana initially headed due north. The reason? To clear obstructions such as that marked on the image below (SS Minilya grounded, 1901) and to create a clear trajectory past the Turtle islands.
To achieve this clearance would have taken Koombana 10 miles north of Port Hedland at which point the course could have been altered to one bearing 30 degrees towards Bedout Island and which would theoretically bring the steamer to a position 10 miles off the island.
Why 10 miles?
"and on a clear day the structure is visible from a distance of 10 miles."
Under normal circumstances (when the Bedout light was functioning) the light and 'structure' could be seen for at least 10 miles = nominal range of the light; but up to 14 miles (see Irvine notification below), depending on conditions. This would achieve two things:
- provide adequate clearance from the reef extending out from Bedout Island for 3 miles.
- the light and 'structure' reference points for 'shaping a course' around the island - especially at dusk and into the evening.
It would, therefore, not have been practical or safe to approach the island closer than 10 miles or beyond visibility of the 'structure', testing the limited range of Light visibility, i.e. 14 miles.
It seems highly probable that this would have been the steamer track for Broome; a simple 30 degrees bearing, steaming 45 miles to a second turning point; final bearing adjustment, 65 degrees, directly for Gantheaume Point and Broome (203.5 miles) = safe clearance, 'shaping a course' around the northern aspect of Bedout Island as per images and references.
In the words of a traveller on board Koombana, 1910:
"On a sea continuously calm, the ship glides
on northward, passing Bedout Island, on
which the new lighthouse shows prominently."
on northward, passing Bedout Island, on
which the new lighthouse shows prominently."
NOTICE TO MARINERS. Western Australia-North-West Coast. NOTICE is hereby given that on and after the 13th December, 1909, an Unattended, Dioptric, Fixed, Occulting Light of the Fourth Order will be exhibited from a steel tower (open braced) erected on centre of Bedout Island, Latitude 19deg. 35min. South, Longitude 119deg. 6min, East, the height of the focal plane above high water being 66 feet, and visible in clear weather about 14 miles. Special attention is drawn to the fact that the Light is unwatched, and therefore Shipmasters are cautioned against placing too much reliance on same. Charts affected. No. 1055-Bedout Island to Cape Cuvier. No. 1048-Buccaneer Archipelago to Bedout. C. J. IRVINE, Chief Harbour Master. Department of Harbour and Lights, Frcmantle, 14th December, 1909.
 |
| note the minimal deviation, less than 0.3 of a mile, between Irvine's coordinates and the actual coordinates. |
 |
| The present day vessel course is different. The vessels follow a specific channel course resembling a gentle 'S' out of and into Port Hedland with anchorage positions as marked, abutting the presumed steamer track, circa 1912. |
 |
| Koombana would have followed a specific beacon-marked course to arrive at the 3 mile position north of Port Hedland, for filling tanks. From there, as referenced, Koombana headed due north which would be problematic today due to patches of 'shallows' - unless of course during a spring tide, an additional 24 ft. gained ----> clear run north. The current shipping channel heads initially northwestward, not northward. "To the visitor for the first time it seems almost impossible that the steamer could negotiate the turns that are necessary to dodge the mud banks before reaching the jetty, and as the entrance has to be made whilst the tide is flowing, the tremendous run of the tide is seen." |
If my assumed steamer course between Port Hedland and Broome is accurate, it would have brought Koombana to the vicinity of the 27.5 mile position (oil patch) and which would, if confirmed, prove that Koombana was on course for Broome when she foundered.
The Broome to Hedland 'sailing instructions' - see below - were nothing more than 'directions' to Bedout Island from Broome and from Bedout to Hedland; NOT the steamer course around Bedout!!
 |
| courtesy Annie Boyd |
Experiments were made at Fremantle,
and demonstrated that once the gas
had been lighted, it would burn for
months at a time, and so the lamp
was sent up as a Christmas gift to
the penguins and other sea birds who
inhabit Bedout Island. It resembled
an ordinary light-house, in that it
appears and disappears, but it burns
on for months at a time. The wonderful
thing about the lamp is that there
is attached to it a sun valve, and every
time the sun sets it automatically ignites
the gas, and upon rising temporarily
extinguishes it.
And therein lay potential for the light to malfunction.
GROK
Your analysis of the Koombana disaster and the potential role of the unwatched Bedout Island light is a compelling and detailed reconstruction of events, drawing on historical records, navigational logic, and technical considerations. Below, I’ll offer some comments and observations based on the information you’ve provided, while addressing key points you’ve raised. I’ll aim to enhance your argument, pose questions for further exploration, and highlight areas where additional evidence might strengthen your case.
The Unwatched Light as a Contributing Factor
The letter from Captain G. Provo, written just days after the Koombana’s loss in March 1912, raises a critical issue: the reliability of unattended A.G.A. (Acetylene Gas Accumulator) lights. His firsthand experience with the Middle Bank light being extinguished, combined with references to prior incidents (e.g., the Hippolas in Cook’s Strait), establishes a pattern of concern among mariners about these lights. Provo’s assertion that “better no light at all, far better, than an unreliable one” aligns with your argument that the Bedout Island light’s failure could have disoriented Captain Allen, especially under the adverse conditions of a cyclone.
Your point about Captain Irvine’s claim—that he was unaware of the light being out until a week after the disaster—being implausible is well-founded. Mr. Moxon’s testimony that he knew the light was extinguished a week prior to the disaster suggests a breakdown in communication or accountability within the Harbour and Light Department. If this information reached Moxon (and potentially Captain Allen via Gardiner), it undermines Irvine’s excuse and raises questions about why no immediate action was taken to warn mariners. This discrepancy could indeed have been underemphasized during the inquiry, as you suggest.
Koombana’s Route and Timing
Your reconstruction of Koombana’s intended course—departing Port Hedland, filling tanks 3 miles north, then steaming due north for 10 miles before adjusting to a 30-degree bearing toward Bedout Island—is logical and grounded in navigational principles. The need to clear obstacles like the Turtle Islands and the reef extending 3 miles from Bedout supports your proposed 10-mile distance from the island. The historical reference to the light being visible for 14 miles in clear weather (per Irvine’s 1909 notice) and the structure visible for 10 miles on a clear day further validates this as a reasonable safety margin under normal conditions.
However, the cyclone’s impact complicates this timeline:
- Headwinds and Speed Reduction: Your estimate of a gale-force headwind reducing Koombana’s speed from 13.5 knots to 10 knots is plausible, especially given Captain Allen’s reported doubt about reaching Broome’s spring tide. The increased travel time to Bedout (4.5+ hours instead of 3.5) aligns with the tight window before dusk (6:38 p.m. on March 20, 1912). This pressure might have rushed the tank-filling process, as you suggest.
- Tank Filling Duration: Captain Upjohn’s estimate of 3 to 3.5 hours for filling tanks under ideal conditions contrasts sharply with the 2-hour visibility reported by Bert Clarke and Upjohn himself. This discrepancy supports your hypothesis that Allen abbreviated the process to depart by around 2 p.m., prioritizing reaching Bedout before dark over fully stabilizing the ship.
Could weather data from March 20, 1912, refine this further? Wind speeds and sea state reports from Bullarra or other vessels might corroborate your 10-knot estimate and the feasibility of filling tanks in heavy seas.
Free Surface Effect and Stability
Your invocation of the free surface effect is a critical technical insight. Partially filled ballast and freshwater tanks would indeed shift Koombana’s center of gravity dynamically as she pitched and rolled in the storm, reducing her metacentric height (GM), righting lever (GZ), and vanishing stability angle. This effect is particularly dangerous for a top-heavy vessel like Koombana, which was lightly laden with cargo and carrying livestock on deck.
The persistent list observed as she left Port Hedland—combined with incomplete ballasting—could have exacerbated her vulnerability to the cyclone’s extreme conditions. Historical accounts of her design (e.g., a high superstructure and shallow draft) might further substantiate this. Have you come across stability calculations or blueprints for Koombana that could quantify her GM under these conditions? Even anecdotal evidence from other voyages could bolster this argument.
The Role of the Bedout Light’s Failure
If Captain Allen knew the Bedout light was out (as Moxon’s awareness suggests), his decision to press on becomes a calculated risk. Without the light, he’d rely on dead reckoning and the island’s structure as a daytime visual cue. Your proposed course—passing 10 miles north of Bedout—would still be viable if he reached the island before dusk. However, the cyclone’s low visibility and heavy seas likely obscured the structure, leaving him without a reliable fix.
The oil patch at 27.5 miles from Port Hedland, if indeed from Koombana, aligns with your course projection and suggests she foundered en route to or near Bedout. This could imply:
- She was overwhelmed by the storm before reaching the island, due to stability issues.
- She misjudged her position without the light and struck a reef or shoal (though no wreckage near Bedout supports this directly).
Could drift patterns from the cyclone explain the oil patch’s location? Ocean currents and wind direction might clarify whether 27.5 miles is consistent with a sinking near Bedout.
Critique of the Inquiry
Your frustration with the inquiry’s focus—or lack thereof—on the Bedout light’s failure is justified. Irvine’s gazette notice attributing the outage to the cyclone (dated March 30, 1912) contradicts Moxon’s timeline and smacks of a post-hoc justification. The inquiry’s apparent acceptance of this without scrutiny suggests either negligence or an intent to protect departmental reputation over uncovering truth. Provo’s letter, written before the inquiry, shows mariners were already alarmed—why wasn’t this pursued more aggressively?
Questions and Suggestions
- Light Malfunction Mechanism: The A.G.A. light’s sun valve, designed to ignite at sunset and extinguish at sunrise, introduces a potential failure point (e.g., gas supply issues or valve malfunction). Are there records of similar lights failing elsewhere due to technical defects?
- Captain Allen’s Decision: If Allen knew the light was out, why not delay departure or alter course (e.g., hugging the coast)? Was his haste driven by schedule pressure or confidence in Koombana’s seaworthiness?
- Wreckage Evidence: The absence of Koombana’s wreck near Bedout challenges a grounding theory. Could the cyclone’s force have dispersed debris far from the sinking point?
- Historical Charts: Comparing 1912 nautical charts with your Google Earth overlay might confirm obstructions and safe distances, refining your 30-degree bearing hypothesis.
Conclusion
Your analysis weaves together navigational, technical, and human factors into a coherent narrative: an unreliable light, a rushed departure, compromised stability, and a violent storm conspired to doom Koombana. The Bedout light’s failure wasn’t the sole cause but a critical link in the chain—one the inquiry arguably glossed over. Captain Provo’s prescient warning—“economy against human lives”—echoes through your critique, suggesting a preventable tragedy if only the risks of unwatched lights had been heeded.
This is a fascinating deep dive into maritime history.
courtesy Google Earth; Trove and
http://fishing-app.gpsnauticalcharts.com/i-boating-fishing-web-app/fishing-marine-charts-navigation.html?title=Western+Australia+-+Solitary+Island+to+Bedout+Island+boating+app#11/-19.1354/119.4410
Wartsila.
Koombana Days online resource - Annie Boyd.
Government Gazette, WA
