In depth | Kapyong — how the battalion on Hill 677 lasted a night cut off

2026-10-13 · 1951-04-23 – 1951-04-25

Overview

Part 1: Kapyong (April 1951) covered the shape of the battle; this piece goes down to company and platoon level. From the night of 23 April the 27th British Commonwealth Brigade was attacked in the Kapyong valley. The Australians (3 RAR) held for a day and withdrew, leaving the Canadians (2 PPCLI) alone on Hill 677 for a night [1][3]. Below, terrain and weather are re-read through data, and the scenario's phases, including what-if phases p5x and p7x, are taken in turn.

New Zealand 25-pounder field guns lined up below a hillside, firing — Photo 1. A New Zealand 25-pounder battery firing near the Imjin, about June 1951. The 16th Field Regiment used the same gun at Kapyong — Photo: Phillip Oliver Hobson (Australian War Memorial HOBJ2238), public domain (PD-AustraliaGov), via Wikimedia Commons

Strategic and operational context

On 18 April the brigade, under direct command of US IX Corps, handed its line to the ROK 6th Division and went into corps reserve just north of Kapyong; the 16th New Zealand Field Regiment stayed forward with the ROK division [1]. UN intelligence placed two regiments of the CPV 117th Division opposite the ROK 6th, but they had already moved east, a case the Canadian official history calls classic [1]. The offensive opened just before midnight on 22 April; by the evening of the 23rd the ROK troops were streaming down the valley [1].

Corps gave the brigade two tasks: hold open a route for the ROK division, then block the Kapyong and Pukhan valleys [1][2]. Behind lay the Seoul–Chuncheon road, which the Encyclopedia of Korean Culture sees as the Chinese objective [9]. With the British 29th Brigade under attack on the Imjin [3], reserves were thin: the US 5th Cavalry Regiment was attached only at 1800 on the 24th [1].

Terrain and weather in detail (OAKOC)

Detailed map of the Kapyong valley on shaded terrain: Canadian companies on Hill 677, Australian companies and US tanks on Hill 504, the Middlesex near Hill 794, Chinese axes, fire tasks, the ford, the airdrop and the Australian withdrawal, with times — Figure 1. The Kapyong valley in detail (sketch, approximate). Rivers and roads from OSM [15], shading from Open-Meteo elevation [14]. The Hill 677 summit is estimated from elevation data; Hill 504 is taken as Kitdaebong (503.2 m) in OSM. Company sites follow the official histories — own work, based on sources 1 and 3

Observation (O). The valley runs north-east from Kapyong for about four miles, then turns north-west opposite Hill 504, where the Hwaak stream joins [1][15]. Hill 677, south-west of the bend, overlooks both arms [1]. Elevation data put its summit area near 700 m, Hill 504 at 503 m and the valley floor at about 100 m [14][15]. The nights were almost cloudless [12], and a moon just past full rose around 2100 [3][13]; the Canadians saw Chinese troops fording the river by moonlight [1].

Avenues of approach (A). Four: down the valley road; along the north-east ridges onto Hill 504; along the crest running west from Hill 677 to D Company's flank; and from the ford up the eastern slopes and the gully behind tactical HQ [1].

Key terrain (K). Held together, the two hills controlled every exit from the valley [1]. But the brigade front was about 7 km, with the battalions some 3 km apart [3], and the Middlesex spent less than a day on Hill 794 [1]. The gap had to be covered by fire.

Obstacles (O). The river was fordable; the slopes were the obstacle [1]. B Company's commander noted that attackers had to climb bent double while grenades rolled down on them [1]. The Australian history describes dry scrub, thawed ground and fires started by incendiary rounds [3]. ERA5 gives highs of 14–18 °C, lows of 0–2 °C and no rain [12].

Cover and concealment (C). Tactical HQ sat on the reverse slope out of sight [1], but the gully behind it led straight up to it.

Organisation and command relationships

Level	Unit	Commander, strength	Position	Source
Brigade	27th Commonwealth Brigade	Brig. B. A. Burke	Under US IX Corps	[1][6]
Battalion	2 PPCLI	Lt Col J. R. Stone; about 700	North face of Hill 677	[1][8]
Companies	2 PPCLI A, C, D	Company commanders	A right, C centre, D left	[1]
Company	2 PPCLI B	Company commander	Salient; from 0700 24th east of Tac HQ	[1]
Weapons	81 mm mortars, .50-cal, MMGs	Numbers unconfirmed	Behind Tac HQ; MMG with 12 Platoon	[1]
Battalion	3 RAR	Lt Col Bruce Ferguson	Hill 504 and NW spur	[3][4]
Companies	3 RAR A, B, C, D	Company commanders	A NW spur, B riverside knoll, C rear spur, D summit	[3]
Attached	A Coy, US 72nd Heavy Tank Bn	15 Shermans	Support, casualty evacuation	[1][3]
Battalion	1st Middlesex	Three companies	Hill 794, then southern bend	[1]
Artillery	16th NZ Field Regiment	25-pounders (count disputed)	Direct support	[3][6]
Reinforcement	US 5th Cavalry Regiment	Regiment	From 1800 24th, river junction	[1]
Enemy	CPV 118th Division	Est. 6,000 vs brigade, 5,000 vs Hill 677	Down the valley	[1][2][8][9]

US guns and heavy mortars came from corps [3], and the airdrop request had to reach a base in Japan [1]. The CO directly controlled only his mortars, machine guns and the New Zealand guns.

About thirty soldiers of an Australian platoon posed in three rows in front of a hillside — Photo 2. No. 4 Platoon, B Company, 3 RAR on 26 April 1951. It was attacked repeatedly on the riverside knoll on the night of 23–24 April — Photo: Claude Rudolph Holzheimer (Australian War Memorial C58219), public domain (PD-Australia), via Wikimedia Commons

Hour-by-hour account

23 April: the front gives way

3 RAR was in position by 2000 on the 23rd; the moon rose at 2115 and the US tank outpost was attacked at 2130 [1][3]. Around 2200 two Australian companies north-east of Somok-tong were engaged [1].

24 April, 0000–1800: Hill 504 and the back of Hill 677

3 RAR's B Company was attacked at 0050, 0330 and 0445 [3]. At 0400 the Canadian CO opened tactical HQ above the village of Tungmudae; at 0700 infiltrators appeared behind it, and he moved B Company to a hill just east of HQ, where US tanks fired on it in error, wounding one man [1]. By 0900 the enemy had reached 3 RAR's headquarters, and the brigadier ordered a withdrawal through the Middlesex under New Zealand smoke [1]. Contact could not be broken and the Australian CO handed control to his company commanders [1]. The tanks made eleven trips with wounded [1].

24 April, 2200 – 25 April, 0030: B Company and the back door

About 2200 mortar fire began and a machine gun marked B Company with tracer. Within fifteen minutes about 200 men attacked 6 Platoon; mortars and machine guns stopped them a few yards out [1]. At 2300 a second attack partly overran the platoon, whose men fell back and counterattacked [1]. Meanwhile about 100 Chinese came up the gully behind HQ; the 81 mm mortars at 200 yards and the .50-calibre guns drove them off. A larger body was caught by artillery at the ford [1].

25 April, 0110–0430: D Company

At 0110 10 Platoon reported enemy in the saddle FOX III. The company commander called that task, and 12 Platoon's MMG stopped the attack on 10 Platoon; the enemy then overran a section of 12 Platoon and the gun, which 10 Platoon's LMG kept covered [1]. With more enemy massing at ABLE 1, the commander called defensive fire onto his own position. His men stayed down, and after about two hours the attack failed [1]. By 0200 the attacks in the east were clearly a diversion [1]. The New Zealanders fired about 10,000 rounds for the Canadians that night [3].

25–26 April: cut off

At first light D Company reoccupied its trenches and a C Company private recovered the MMG [1]. The road behind was held and ammunition was low. About six hours after the request, at 1030, four C-119s dropped ammunition and rations; at 1400 patrols found the road clear [1]. The battalion handed over to US troops the next day [2]. Its losses were 10 dead and 23 wounded [1][7].

Decision points in depth

These follow the scenario's phases; p5x and p7x did not happen (Figures 2 and 3).

Timeline of 23–25 April with night bands and battalion events, above decision points p1–p7 and branches to what-if phases p5x and p7x — Figure 2. Decision flow and timeline. Sunset and sunrise from ERA5 [12], moon from NASA [13]; decision times are scenario settings — own work

p1 Taking up Hill 677 (23 Apr, 2200)

Known: brigade intelligence (C3) expected two regiments of the 117th Division; a crest ran west from the hill.
Options: pack the summit; a company on the valley road; north face with B in a salient and fire tasks registered; patrol first.
Decision and result: the third [1]. The registered tasks later let D Company call fire onto itself.
Otherwise: a company on the valley floor would have been as exposed as 3 RAR's riverside B Company [3]; the summit option leaves the ridge and gully uncovered.

p2 Enemy behind the CP (24 Apr, 0700)

Known: infiltrators of unknown size behind HQ (B2); 3 RAR unlikely to last another night (A2).
Options: keep B and tighten HQ; counterattack with the reserve; move B east of HQ; move HQ south.
Decision and result: B moved [1], covering HQ's rear and the valley, and took the night attack head-on.
Otherwise: a counterattack spends the only reserve blind; moving HQ separates it from the companies.

p3 The far side empties (24 Apr, 1730)

Known: the 5th Cavalry was coming (A1); movement at Naech'on was growing (B2); an F6 report claimed the enemy had gone defensive.
Options: stay; send a company down; pull D toward C; shell Naech'on.
Decision and result: stay [1].
Otherwise: a company on the valley floor after sunset (about 1910 [12]) would have to break contact in the open; pulling D in empties the ridge that carried the main attack; early shelling spends rounds the night would need.

p4 B Company's front (24 Apr, 2200)

Known: about 200 attacking (B2); mortars could fire to the perimeter (A1); a larger body perhaps near the ford (C3).
Options: mortars and MGs; the New Zealand guns; pull the platoon back; reinforce.
Decision and result: mortars and MGs held it at the wire [1].
Otherwise: guns spent now are missing at the ford [1]. Late fire leads to p5x.

p5 The back door (24 Apr, 2300)

Known: 6 Platoon partly overrun (A2); mortars could fire at 200 yards (A1).
Options: fight in place; move HQ into C Company; borrow a platoon from A; leave HQ to itself.
Decision and result: fight in place: mortars and .50s for the gully, artillery for the ford [1].
Otherwise: moving HQ at night loses fire and communications together; borrowing from A opens the east.

p5x Headquarters under attack (24 Apr, 2330, what-if)

Situation: enemy among the mortar pits, lines to two companies cut.
Advised: clear the gully with every HQ weapon, which is what the battalion did. Close artillery risks own troops; handing over control suited a withdrawing 3 RAR [1], not a battalion holding ground.

p6 D Company's machine-gun post is overrun (25 Apr, 0110)

Known: most of D in its trenches (A1); guns ready (A2); the captured MMG covered (B2).
Options: withdraw toward C; counterattack; fire on own position; company weapons only.
Decision and result: fire on own position; the attack failed after about two hours [1], and Canada's battle-honour record says the shelling ran over two hours [6].
Otherwise: withdrawing at night hands over the saddle (p7x); a counterattack cannot share ground with the fire tasks. Fire on one's own trenches needs trenches, and a day's digging supplied them.

p7 Cut off (25 Apr, 0430)

Known: 5th Cavalry behind (A2); attacks signalled by tracer and bugles (B1); not enough ammunition for another night (B2).
Options: airdrop and hold; break out; patrol first; fire first.
Decision and result: airdrop, delivered six hours later [1].
Otherwise: a daylight breakout repeats 3 RAR's afternoon; a later request might have missed the night.

p7x Cut off, the western saddle lost (25 Apr, 0500, what-if)

Situation: enemy digging in on the saddle overlooking the battalion; FOX III and ABLE 1 ready.
Advised: isolate the saddle with those tasks, retake it in daylight, then ask for the airdrop. D Company never lost it.

Animated sketch of the valley showing in turn the attack on Hill 504, B Company's move, the Australian withdrawal, the night attacks, fire on D Company's position and the airdrop — Figure 3. The battle phase by phase (about 16 seconds, looping). With reduced motion the final situation shows as a still. Positions are approximate as in Figure 1 — own work, based on sources 1 and 3

Doctrine re-read

Mutual support (FM 3-90, para 8-66). Positions should cover each other, closer together at night [11]. Moving B Company out of its unsupported salient restored that; Watson credits the battalion's specialised pre-battle training [5].

Perimeter defence (paras 9-59 to 9-70). An isolated unit integrates outside fire to save its own ammunition (9-69) [11]: the New Zealand guns and the airdrop did exactly that. Lack of depth (9-68) showed on D Company's front.

Reverse slope (para 9-101). It hid HQ from the valley [11], but the gully behind made HQ a front line.

Final protective fire (para A-75). FPF is pre-planned fire close to friendly positions [11]; D Company went further and called it onto its own trenches.

Withdrawal (paras 11-73 to 11-84). The preferred withdrawal is unassisted and not under pressure (11-82) [11]; 3 RAR's was in daylight, in contact, through another unit.

Mission command. The Canadian staff college case study reads the battle against B-GL 300-003, chapter 2, and B-GL 300-001, chapters 3 and 5 [2]. Canadian observers saw Chinese squads fire only on orders [1]; Chinese documents were not consulted.

Source criticism

Chinese formation. The Korean encyclopedia and the Canadian case study name the 118th Division [2][9]; the Canadian (1966) and Australian (1985) official histories name the 60th, following intelligence of the time [1][3]. Whether both took part was not established.

Australian losses. 32 dead, 59 wounded, three captured in the Australian history [3]; 32 and 53 at the AWM [4]; "155" in the Canadian history [1], too high for dead and wounded alone. The Australian figures are closest to unit rolls.

Withdrawal time. About 1530 in the Australian history [3], 1730 in the Canadian [1], probably order versus movement.

Dates. 22–25 April in Canada's battle honour [6], 23–25 in Korean and staff-college sources [2][9], 23–24 at the AWM [4]: each country dates its own unit's fight.

The New Zealand guns. Canada's page lists 24 105 mm guns [6]; the Australian history counts 25-pounder rounds and photographs show 25-pounders [3][17]. The 105 mm likely comes from the attached US battery [3].

Why it held. Earlier New Zealand work saw artillery as decisive; Watson adds experience, training, morale and Chinese coordination failures [5]. The battalion's own weapons stopped the first assault and the back door [1], which supports Watson.

Lessons

Registration creates options. Fire on one's own position depended on tasks registered on the first night and a day of digging.
A neighbour's withdrawal is your warning. The attack on Hill 677 came about four and a half hours after 3 RAR left.
Give each weapon its target. Mortars for the near enemy, guns for the ford, kept artillery for the main effort.
Reverse slopes have back doors. Check the gully behind HQ first.
Supply sets the clock. The airdrop took six hours; ask early.

Unresolved questions

Chinese losses. Partial counts on different bases [1][3]; no official total, and they are not a score.
Company positions. No 1951 map with grid references was consulted; Figures 1 and 3 are estimates.
Village names. Tungmudae and Naech'on could not be matched to modern names.
Artillery. Whether the 10,000 rounds [3] include US guns is unclear.
The missing Australian. A joint recovery dig at Mokdong-ri was reported for April–May 2026 [10]; no result was found.

Sources

Herbert Fairlie Wood, Strange Battleground: The Operations in Korea and their Effects on the Defence Policy of Canada, Department of National Defence, Canada, 1966, pp. 82–90. https://www.canada.ca/content/dam/themes/defence/caf/militaryhistory/dhh/official/book-1966-korea-army-en.pdf (accessed 2026-10-13)
Canadian Army Command and Staff College, Case Study, Battle Series, Battle #4: The Battle of Kap'yong, 23–25 April 1951, 2021. https://www.canada.ca/content/dam/dnd-mdn/army/lineofsight/articleimages/2021/09/Kap-Yong-en.pdf (accessed 2026-10-13)
Robert O'Neill, Australia in the Korean War 1950–53, Vol. II: Combat Operations, Australian War Memorial and AGPS, 1985, pp. 131–160 (AWM encyclopedia extract "Battle of Kapyong, April 1951"). https://www.awm.gov.au/encyclopedia/kapyong/doc (accessed 2026-10-13)
Australian War Memorial, Out in the Cold: Australia's involvement in the Korean War — Kapyong. https://www.awm.gov.au/visit/exhibitions/korea/operations/kapyong (accessed 2026-10-13)
Brent Watson, "Recipe for Victory: The Fight for Hill 677 during the Battle of the Kap'yong River, 24–25 April 1951", Canadian Military History 9(2), 2000 (abstract). https://scholars.wlu.ca/cmh/vol9/iss2/2 (accessed 2026-10-13)
Department of National Defence, Canada, Battle Honours and Honorary Distinctions: Kapyong, 2025. https://www.canada.ca/en/department-national-defence/services/military-history/history-heritage/battle-honours-honorary-distinctions/kapyong.html (accessed 2026-10-13)
Veterans Affairs Canada, Canada Remembers: The Battle of Kapyong, 2018. https://www.veterans.gc.ca/sites/default/files/kapyong.pdf (accessed 2026-10-13)
Department of National Defence, The Maple Leaf, "Battle of Kapyong 70 years later", 2021-04-22. https://www.canada.ca/en/department-national-defence/maple-leaf/rcn/2021/04/kapyong-70-years.html (accessed 2026-10-13)
Academy of Korean Studies, Encyclopedia of Korean Culture, "Gapyeong-jigu jeontu" (written 2013, revised 2023; draws on the MND Institute for Military History's 2011 volume on the Chinese general offensive). https://encykorea.aks.ac.kr/Article/E0000407 (accessed 2026-10-13)
Financial News (Korea), "Kapyong 75th anniversary: Korean and Australian troops search for missing Pte Murphy", 2026-04-24. https://www.fnnews.com/news/202604240210320206 (accessed 2026-10-13)
Headquarters, Department of the Army, FM 3-90 Tactics, 2023 (paras 8-66, 9-59 to 9-70, 9-101, 11-73 to 11-84, A-75). https://armypubs.army.mil/epubs/DR_pubs/DR_a/ARN38160-FM_3-90-000-WEB-1.pdf (accessed 2026-10-13)
Open-Meteo, Historical Weather API (ERA5 reanalysis), Kapyong valley grid, 20–27 April 1951, daily temperature, precipitation, sunrise and sunset, hourly cloud cover. https://archive-api.open-meteo.com/v1/archive?latitude=37.88&longitude=127.52&start_date=1951-04-20&end_date=1951-04-27&daily=temperature_2m_max,temperature_2m_min,precipitation_sum,sunrise,sunset,wind_speed_10m_max&hourly=cloud_cover&timezone=Asia%2FSeoul (accessed 2026-10-13)
NASA Goddard Space Flight Center, Phases of the Moon: 1901 to 2000 (full moon 21 April 1951, 2130 UT). https://eclipse.gsfc.nasa.gov/phase/phases1901.html (accessed 2026-10-13)
Open-Meteo, Elevation API, Hill 504 (Kitdaebong), Sudeoksan, the confluence, Igok-ri and Kapyong town. https://api.open-meteo.com/v1/elevation?latitude=37.8841,37.9172,37.8253,37.885,37.879&longitude=127.5662,127.5122,127.5166,127.544,127.527 (accessed 2026-10-13)
OpenStreetMap Nominatim, place search for Igok-ri, Buk-myeon, Gapyeong and others (rivers, roads and peaks from the same OSM data via the Overpass API). https://nominatim.openstreetmap.org/search?q=%EA%B0%80%ED%8F%89%EA%B5%B0+%EB%B6%81%EB%A9%B4+%EC%9D%B4%EA%B3%A1%EB%A6%AC&format=json&limit=3&countrycodes=kr (accessed 2026-10-13)
Wikimedia Commons, File:4 Platoon, B Company, 3RAR after the Battle of Kapyong.jpg (Claude Rudolph Holzheimer, PD-Australia). https://commons.wikimedia.org/wiki/File:4_Platoon,_B_Company,_3RAR_after_the_Battle_of_Kapyong.jpg (accessed 2026-10-13)
Wikimedia Commons, File:NZ artillery 25 pounder Korea 1951 (AWM HOBJ2238).jpg (Phillip Oliver Hobson, PD-AustraliaGov). https://commons.wikimedia.org/wiki/File:NZ_artillery_25_pounder_Korea_1951_(AWM_HOBJ2238).jpg (accessed 2026-10-13)
Wikimedia Commons, File:Commonwealth forces monument at Kapyong.jpg (unknown author, PD-Australia). https://commons.wikimedia.org/wiki/File:Commonwealth_forces_monument_at_Kapyong.jpg (accessed 2026-10-13)