This post is not normally how I do things as I generally try to avoid this much speculation, so I will preface this with a warning that this is extremely speculative and lacks any documents that directly confirms this theory. Instead, I will present a string of coincidences that might suggest that the United States adapted the British Flagpole device to produce the Fife thermonuclear secondary used in the Polaris A-2 W47Y2 warhead and in the Minuteman I/II W56 warhead. This is a topic I have been looking into for several years, and having only turned up circumstantial evidence for it, I have decided to throw it out into the wider community in the hopes that someone else may have a lead.
Figure 1 – Thermonuclear device Short Granite tested in Operation Grapple 15 May 1957. UK Crown Copyright. Source: [1, N. Plate 21]
For brevity, I will refer to University of California Radiation Laboratory (UCRL) which is now known as Lawrence Livermore National Laboratory as simply Livermore, as Livermore underwent many name changes over the years. I will also refer to Atomic Weapons Research Establishment (AWRE) which is now known as the Atomic Weapons Establishment (AWE) as simply Aldermaston as this is how it is often referred to.
For easy reference, I have provided the specifications for some US nuclear warheads of note below.
| Warhead | W47Y1 | W47Y2 | W56 | W59 |
| Yield (kt) | 500-600 | 1200 | 1200 | 800 |
| Bare weight (lb) | 717 | 733 | 600 | 550 (~500?) |
| Weight with RV (lb) | ~850 | ~850 | 831 | ??? |
| Diameter (inches) | 18 | 18 | 17.4 | 16.3 |
| Length (inches) | 46.6 | 46.6 | 47.3 | 47.8 |
| System | Polaris A-1/A-2 | Polaris A-2 | Minuteman I/II | Minuteman I |
| Service dates | Jun 1960-Oct 1967 | Feb 1963-Jun 1974 | Mar 1963-1993 | Jun 1962-Jun 1969 |
Sources: [2, pp. 218–221], [3], [4], [5], [6], [7], [8]
An Unclassified History of the W47 and W56 Warheads
I will discuss both warheads simultaneously as their histories are closely tied. I am also required to briefly discuss the W59 warhead.
The W47 warhead had its origins in Project Nobska, a broad mid-1956 US navy-hosted conference where leading scientists were invited to pitch various technological innovations to the navy for the purpose of solving various navy goals. Edward Teller — who alongside Earnest Lawrence founded Livermore — was one of the speakers invited. During the panel on Jupiter S (a solid fuelled missile similar to Jupiter for carriage on the proposed Fleet Ballistic Missile submarine) Teller suggested that by the planned entry-into-service date of 1965, a megaton yield, 12” diameter warhead with a total weight of 600 lb may be possible.[9, pp. 119–120], [10, p. 528]
Los Alamos scientists at the conference were sceptical as they believed that a yield of half a megaton was more likely in that timeframe,[9, p. 124] but the navy were not concerned by this as they just wanted a relatively high yield warhead in the 600 lb weight class and were happy to accept an interim design of only a few hundred kilotons with more being icing on the cake.[9, p. 125] This lead to the Nobska panel recommending that instead of Jupiter S, a missile 5 to 10 times lighter should be developed to carry a 600 lb warhead to a range of 1,000 to 1,500 miles.[9, p. 120]
Work on the new warhead began with a January 1957 request from the navy for design proposals. Both Los Alamos and Livermore contributed to the May 1957 feasibility study, but it was primarily a Livermore endeavour. The study stated that no existing warhead met the requirements, but that the warhead could be developed from devices already planned for testing. Total weight including re-entry vehicle would be no more than 835 lb.[4, pp. 8–9]
Sword of Armageddon lists two secondary designs for the W47 (later referred to as W47-Y1): Whistle and Piccolo.[11, p. 221] Shortly before the 1958 moratorium both secondaries were tested, with Whistle tested in Hardtack I Maple (213 kt), and Piccolo tested in Hardtack I Nutmeg (25.1 kt — likely a test of a thermonuclear mock-up), Aspen (319 kt), Redwood (412 kt), and Dogwood (397 kt).[11, pp. 258–361] The higher yield Piccolo design would end up being the secondary selected for the W47.[11, p. 354]
Possible W47 primaries were tested in Hardtack II, which included one-point safety tests (Mercury, Neptune and Titania), partial yield tests (Rushmore) and tests of the primary in a secondary stage mock-up (Evans and Blanca).[3, p. 410] Adams, the final test of Hardtack II was also to be a W47 test, but due to unfavourable winds the test could not be carried out on the planned date of 31st October 1958, the last day before the start of the 1958 nuclear testing moratorium. This lack of a proof test of a one-point safe primary forced designers to implement a mechanical safing device into the warhead.[9, pp. 134–135], [12, p. 62]
During the same time period in April 1958, warheads for Minuteman (then known as Weapon System WS-133A) were being considered. The intention was for the new missile to carry two warhead options: a lighter 350 lb warhead and a heavier a 550 lb warhead. The 350 lb warhead requirement was expected to be filled by a redesign of the W50 warhead for Nike Zeus and Pershing, while the 550 lb warhead requirement was expected to filled by reworking the W47 warhead.[5, pp. 8–9]
Little work was performed on the Minuteman warhead program during 1958 due to lack of funding, but by March 1959 a second feasibility study had been performed which concluded that if adapted to Minuteman the weight of the W47 design could be reduced and that other nuclear designs could be considered, but due to the testing moratorium, emphasis would be placed on designs that did not require further nuclear testing.[5, p. 9]
Concurrently, the W47 warhead was assigned the highest development priority as the first Polaris submarine was expected to enter service in April 1960.[4, pp. 14, 16] The warhead reached design release in December 1959, but the project was slowed down in January 1960 at the request of the navy due to delays in Polaris with the first warheads delivered June 1960.[4, p. 16]
In November 1959, the XW56-X1 nomenclature appears,[5, p. 5] with XW56 (or XW56-X0) being the initial warhead design and XW56-X1 being substantially different from the previous design. Knowing that the W47Y1 (and thus the XW56) had a yield of 500 or 600 kt, and that the production W56 (XW56-X2, which was the XW56-X1 design with some small improvements) had a yield of 1.2 Mt, it is clear that the X1 change is the replacement of the Piccolo secondary with Fife. The Sandia history of the W56 warhead describes the XW56-X1 as “an entirely new weapon taking advantage of all possible design improvements” and states that Livermore were willing to certify the weapon without nuclear testing.[5, p. 15] The language in the Sandia history makes it clear that the X1 design was being developed concurrently with the original X0 design as the topic of cancelling the X0 is discussed. Field command did not believe there was an advantage in using a mixture of “tested and untested” warheads, while Livermore did see some advantage to this, and the division of military application subsequently requested additional information on the matter in August 1960.[5, p. 16]
In December 1960 it is decided Livermore should focus efforts on the W56-X1 design using the Fife secondary and as such the XW56-X0 using the Piccolo secondary was cancelled.[5, p. 5] This was followed in January 1961 when the W59 warhead program — to be designed by Los Alamos — was instituted to design a warhead for the Skybolt missile, with second application for Minuteman.[6, pp. 5–6] Though not explicitly stated in the heavily redacted histories, it appears that the tested X0 design was cancelled and replaced with the tested XW59 design, while Livermore focused their efforts on the higher yield but untested XW56-X1. It’s not clear in what shot the W59 warhead (sometimes called “J-21”) was tested in but it was presumably a Hardtack I shot.
In September 1961 a first production date of February 1963 was set for the W56. One major design change was made in January 1962 which was to convert the design to use a different chopper-converter firing set and external initiation, which was designated XW56-X2, but the warhead would retain the W56 Mod 1 name in production.[5, pp. 20–21] First production unit was in March 1963[5, p. 6]
With the Soviets resuming nuclear testing in August 1961, the US resumed testing 15th September 1961 at low yields underground at the Nevada Test Site (Operation Nougat), followed by a high yield atmospheric test series at Kiritimati (Christmas) Island beginning April 1962 (Operation Dominic). Arkansas, the third test and first Livermore test of Dominic was of the XW56-X2 warhead. The fifth test of the series, Frigate Bird, was of an all-up test of the Polaris A-1 missile carrying an almost unmodified W47-Y1 warhead with a yield of 500 or 600 kt (the yield is disputed across different sources). A number of W47, W56 and W59 tests were carried out in Dominic, including a test of a W47 backfitted with the Fife secondary to produce the W47-Y2 warhead.[2, pp. 166–176] These tests are listed below.
| Name | Date | Warhead | Yield (kt) | Weight (lb) | Note |
| Arkansas | 02/05/1962 | XW56-X2 | 1090 | 600 | |
| Questa | 04/05/1962 | XW59 | 670 | 552 | Disappointing yield. Preshot prediction was 800 kt. |
| Frigate bird | 06/05/1962 | W47-Y1 | 500 or 600 | 717 | War reserve warhead modified with command destruct system and launched for Polaris A-1 missile. Detonated north-east of Christmas Island. |
| Swanee | 14/05/1962 | “Clean” W56 | 97 | 506 | Possibly a test of the Plowshare device tested in Storax Sedan. |
| Alma | 08/06/1962 | XW59 | 782 | 553 | |
| Harlem | 12/06/1962 | XW47-Y2 | 1200 | 733 | |
| Rinconada | 15/06/1962 | XW59 “Wall” | 800 | 507 | |
| Bluestone | 30/06/1962 | XW56-X2 “Prime” | 1270 | 564 | Highest yield to weight ratio known to have ever been tested. |
| Sunset | 10/07/1962 | XW59 | 800-1000 | 543 | Hansen gives the yield as 1 Mt, but a declassified document gives yields of 855 and 930 kt for Sunset[13] while another document gives the official yield of the 550 lb Minuteman warhead as 800 kt.[7, p. 20] |
Unless otherwise noted, data is sourced from [2, pp. 166–176]. Weights given are of the actual test device physics package weight and does not include things like fuzing and firing systems or ballistic case.
The development history of the W47 and W56 warheads continues until the late 1960s, with both weapons having new inherently one-point safe primaries installed while the W56 receives several rounds of hardening.
A Question of Risk
Obviously, the above is based on what has not been redacted in official reports, some attempt to read between those redactions, along with a small amount of speculation. Regardless, this leaves me with a very big question: why was Livermore willing to rely on an untested warhead for a significant fraction of the US nuclear arsenal? Not only a significant fraction of the arsenal, but also a significant fraction of their increasingly important second-strike capability.
A significant thing to note here is that the W59 acting as the tested warhead to compliment the W56 does not seem to have been what field command had in mind when the discussion came up. They were very clear about not wanting two different warheads and were even willing to adopt the untested warhead. It feels more like an afterthought, and even then, that might not have been the original intention as the W59 is significantly lighter than the W56, giving the lower performance Minuteman I more range.
Looking at the retirement and entry into service dates this theory may make sense, with Minuteman II being introduced in mid-1965, Minuteman I missiles being retired in late-1966, and Minuteman I missile replacement by Minuteman II ending in mid-1969 (the remaining 500 or so missiles were replaced by Minuteman III). Most but not all of the Minuteman I retirements in that period appear to be of the Mk5/W59 system with all Mk5/W59 missiles being retired by 1969.[14, p. 7] So as the more powerful Minuteman II came into service with its increased range, the advantage of the lighter warhead was lost.
This question could be solved if we knew precisely what they meant by “untested”. Untested could mean an almost clean sheet design with little part of the design being based on previously tested devices. It could also mean that the current configuration is untested, which itself could mean anything from there being slight changes between the tested and untested designs, to the secondary stage being previously tested using one primary design but is untested with the current primary design.
A good example of this “tested or untested” uncertainty can be found in the Proceedings of the 14th Meeting of the Fleet Ballistic Missile Development Program Steering Task Group (September 1959) — the steering group that led the development of Polaris — in which the idea of “cluster” warheads is discussed (which would eventually go on to be the Polaris A-3 missile carrying the W58 warhead). At the meeting, the warhead is described by a Lieutenant Commander Wertheim as “the 0.2 megaton unit that was tested last year” i.e. the Hardtack Olive device, but he notes that “only the thermonuclear unit has been tested — integrated into a warhead, it has not been tested”. A Mr Hawkins then asks if there is “general agreement that this was a tested unit?”, to which a Commander Backus replies that some DoD decisions throw that in doubt to which Mr Hawkins is confused.[15, p. 3]
Others in this meeting then go on to explain how the tested device had diagnostics equipment connected to it, how the tested device was not engineered for stockpiling and military service, and that after those changes, the DoD may no longer consider it a tested device. The following quote from Wertheim sums up the matter quite clearly:[15, p. 4]
“Largely, the DoD decisions seems to reflect the amount of risk they are willing to take, relative to the important of the system involved, and this is why it makes sense for us to proceed on tested designs, knowing that DoD is more likely to approve them through extrapolation.”
Therefore, for cornerstone weapons like Polaris and Minuteman — which were about to become the United States’ second strike capability — even small design changes may lead to a weapon being considered untested. While for less critical systems, slight changes might be acceptable. This also seems to rule out clean sheet designs quite firmly, which the Fife secondary appears to be if we assume Fife is a wholly US-indigenous design.
An Unclassified History of Operation Grapple
Operation Grapple — the British operation to test their newly developed thermonuclear weapons — consisted of four parts, those being Grapple (May to June 1957), Grapple X (November 1957), Grapple Y (April 1958) and Grapple Z (August to September 1958). All of the tests were carried out at Christmas Island. During this time, the UK also conducted Operation Antler at Maralinga in South Australia (September to October 1957) and some of those tests relate to Grapple. I have listed the Grapple and Antler tests below.
| Name | Date | Device | Yield (kt) | Note |
| Grapple 1 | 15/05/1957 | Short Granite | 300 | First attempt at thermonuclear test. Fizzle |
| Grapple 2 | 31/05/1957 | Orange Herald | 720 | Very large single stage device which was publicly reported as a thermonuclear test by the British government. Device was reportedly solid boosted, but boosting did not work. |
| Grapple 3 | 19/06/1957 | Purple Granite | 200 | Fizzle. Was an attempt to fix flaws in Short Granite device. |
| Antler 1 | 18/09/1957 | Pixie | 0.93 | Small, lightweight warhead. |
| Antler 2 | 25/09/1957 | Indigo Hammer | 6 | Lightweight warhead for Bloodhound missile and as primary for thermonuclear weapons. |
| Antler 3 | 09/10/1957 | Red Beard | 26.6 | Test of improved device for the Red Beard bomb and for use as a primary. |
| Grapple X | 08/11/1957 | Round A | 1800 | First successful British thermonuclear test. Used a “three layer” secondary and Red Beard primary.[1, pp. 152, 155] |
| Grapple Y | 28/04/1958 | Dickens | 3000 | Highest yield test by the UK. Device was intended to get more fusion yield compared to Round A.[1, p. 167] |
| Grapple Z1 | 22/08/1958 | Pendant | 24 | Solid boosted device for use as a primary. Used external neutron initiation.[1, pp. 182, 185] |
| Grapple Z2 | 02/09/1958 | Flagpole | 1210 | Scaled down Dickens device. Intended to be a 1-ton megaton weapon, and when fitted with a boosted primary was anticipated to be “immune” (preinitiation resistant). Used unboosted Indigo Hammer primary.[1, pp. 182, 187] |
| Grapple Z3 | 11/09/1958 | Halliard | 800 | Three-stage weapon. Heavy-casing version of device was tested at US request. Intended to be a preinitiation resistant design without boosting.[1, pp. 182, 186] |
| Grapple Z4 | 23/09/1958 | Burgee | 25 | Gas-boosted primary test.[1, p. 183] |
Data from [1].
Some notes on the above:
Much of what is known of Grapple comes from Britain and the H-Bomb by Lorna Arnold. The book is the official account of Grapple and Arnold was the official historian for the Atomic Weapons Establishment, who wrote the book with access to classified documents. I highly recommend reading it, and for those used to heavy handed redactions in declassified US documents you will find the book extremely candid.
Names are often attributed to the Antler shots, but the tests were actually only numbered. The names often attributed to the shots are from the locations the shot were fired at Maralinga.
In the first half of 1957, it was anticipated that two tests in Antler would be fired as thermonuclear mock-ups.[1, p. 142] It is unclear if this was actually carried out as the number of tests and devices to be fired changed over 1957. For example, in July 1957 five shots were planned, with yields ranging from 5 to 50 kilotons.[16, p. 7]
Pendant is often listed as Pennant, but according to Arnold this is in error and the correct name is Pendant, matching the theme of the other Grapple Z devices and official documents.
The Granites
The Short Granite device was very complicated, consisting of 14 shells ranging from 0.034 to 0.971 inches in thickness, with a lithium deuteride shell perhaps mentioned in singular form,[1, p. 141] and yielding a measly 300 kilotons. Purple Granite was similar to Short Granite, but contained extra HEU and had an outer layer of aluminium, but gave an even more disappointing 200 kiloton yield.[1, p. 146] Thought not explicitly given as failure points, various items to improve in follow up tests listed included reducing Taylor instability (Rayleigh-Taylor or R-T instability), increasing production of fast neutrons (I assume here they mean fusion neutrons and not fast fission neutrons), increasing compression and attempting to delay criticality (presumably of the spark plug). It’s specifically stated that they needed to “increase the thickness of lithium deuteride layer or layers” as part of this.[1, p. 152]
Round A
Following the failure of the Granite devices, a stepwise plan was developed to produce a workable thermonuclear device. First was that while they needed to reduce the fissile material requirements of the production device, they first needed to get the basic principles right and therefore should design their next device without concern for weight or fissile material consumption. They would then conduct three nuclear tests of either a five-layer or three-layer secondary, with different layers switched out for inert material for each test so that what affect this was having on the yield could be verified.[1, p. 153]
The plan eventually agreed was to fire a three-layer device as Round A. If neutron yield was good then a five-layer secondary — Round B — would be fired, and if the neutron was yield poor the issue might be in the HEU component so a device with inert fusion fuel — Round C — would be fired to confirm (though the language is not precisely clear, it appears to be referring to a spark plug as poor thermonuclear fuel heating from a low yield spark plug would lead to less fusion and less fusion neutrons). If Round C showed a low yield, a test of the primary may have to be conducted. A final option where fusion yield was poor but fission yield was good would mean the issue was due to layer mixing — the R-T mixing mentioned above — which would mean that the layered Granite concept might need to be abandoned.[1, p. 155]
But the above plan would ultimately not be carried out due to issues procuring lithium deuteride and in Aldermaston’s capacity to cast depleted uranium parts in time. This led to only Round A and C being manufactured and Round B being cancelled.[1, pp. 155–156]
If the numbers in Arnold’s book are to be believed, the implosion of the Round A primary took 70 μs, the primary detonation phase took 300 μs as the radiation case expanded by a few centimetres, and the secondary was compressed to one-twentieth of its initial volume in only 2 μs, as the “uranium-238 outer case crushed the thermonuclear fuel”. Total yield was 1800 kt, much better than the 1000 kt predicted before the shot. With the device having performed so well and no Round B to test, Round C was cancelled.[1, pp. 160–161]
While discussing the tests, Dr Ken Allen asked “But was it a real hydrogen bomb, or more like a double boosted bomb?”[1, p. 165]
Dickens
For the follow up for Round A, a number of options were considered. The first was to find a more ideal balance between different thickness of the layers in the secondary, under the assumption that too thick layers would moderate the fusion neutrons too much while too thin layers caused R-T instability (presumably referring to lithium deuteride layers as deuterium is a neutron moderator). Given what is currently understood about secondaries, it seems that the British were too concerned about this issue, and it almost certainly was central to the issues they had with the two Granite tests. A second option was to try improve compression and a third option was to try develop a secondary that combined fission and fusion fuels together (something that was believed would considerably delay the next planned test).[1, pp. 165–166]
The first option was ultimately chosen with a number of different secondary configurations being considered, including a five-layer device, a cylindrical device using lithium-7 and a spherical device using “a lot” of lithium-7 and less HEU, the latter two were suggested by Dr Ken Allen, and — due to issues simulating a five-layer device — the spherical device using large amount of Li-7 was adopted for Dickens.[1, p. 166] Ken Allen’s idea for Grapple Y is described as “a brilliant new idea” by Arnold,[1, p. 153] but this is not further explained and is presumably more brilliant than simply “more fusion and less fission”, as I will discuss later.
Predicted yield was 2 to 2.5 Mt, with an upper limit of 4 to 5 Mt.[1, p. 168] Actual yield was 3 Mt and was considered a resounding success.[1, p. 174]
Flagpole
Flagpole was a scaled down version of the Grapple Y Dickens device (which therefore originated in Ken Allen’s “brilliant idea” — though natural lithium is not mentioned) and intended as a one megaton in one ton device that would be immune to detonations from anti-ballistic missile systems.[1, pp. 175, 182] The actual yield of the test was 1210 kt.[1, p. 187]
Halliard
Halliard was a device intended to be a “immune” weapon that did not require boosting, and was developed in three variations: the three-stage heavy cased Halliard 1, the two-stage Halliard 2 (presumably with a heavy case as the thin case was a later thought) and the three-stage thin cased Halliard 3.[1, p. 183] The actual logic to select which device to be fired was complicated[1, p. 258], but as all three previous Grapple Z shots has been successful, the thin cased three stage Halliard 3 device would be fired. However, while preparations were underway for the test, it reached Christmas Island that during talks with the United States it was revealed that the US already had plenty of information on thin cased weapons, but that they were very interested in the three stage concept and would like the British to test the heavy cased three-stage device. It was therefore agree to test Halliard 1 in exchange for this data on radiation cases.[1, p. 186]
Predicted yield of the device was 750 kt, and the actual yield was 800 kt.[1, p. 188]
The First US-UK Exchanges
I will discuss this from both the US and British sides as they help fill in the redactions. The matter is further complicated by duplicate copies of the US documents, often with different redactions. The first meeting was held in August 1958 and the second in September 1958.
The first meeting began with an exchange of basic technical details of each other’s weapons, both in production and soon to be in production, describing their sizes, weights, yield, amount of nuclear material needed etc, followed by the British delegation giving a presentation on their current state-of-the-art. The US note following the presentation states that:[17, p. 14]
“During the first meeting it became obvious that the United Kingdom has achieved an advanced state of weapon research and development in both the fission and thermonuclear fields. Moreover, it appeared likely that certain advances made by the United Kingdom would be of benefit to the United States.”
Per the US reports, in the second meeting the US provided to the UK blueprints, technical specifications and experimental data for the B28 bomb and XW-44, XW45, XW47 and XW48 warheads, along with “[redacted] for our TX41 and TX46 weapons” (secondary perhaps?). The British provided information on a high-yield fission bomb (presumably Orange Herald), a 2200 lb thermonuclear bomb, a “small [redacted] device”, two boosted fission designs, a planned 1500 lb thermonuclear device and a proposed 6” artillery warhead.[17, p. 15], [18, p. 15]
On the British side, the delegation believed that General Starbird and General Herbert Loper were genuinely appreciative of these meetings and were “going as far as they [could]”.[19, Sect. E37]
The British delegation were impressed by the W28 warhead, but not initially so by the W47, describing it as:[19, Sect. E37]
“A scrappy assembly based on the [W28] warhead conception, but it is not yet clear to us how the various sections contribute and we have yet to find out (tomorrow) the depth of US understanding of this design.”
But in following meetings changed their opinion of it stating that:[19, Sect. E38]
“Now have more information on Mk47 design is more subtle than at first appeared. Severe difficulties encountered in getting down to this weight. A lot of calculation went into the design which is considered by United Sates to be sufficiently well understood for modifications to be made for weaponisation and production. United States considers only a minor increase in yield could be expected with this design.”
The US report on the meetings is very positive of British developments:[18, p. 16]
“While it does not appear that we are interested in taking any one United Kingdom weapon or device and weaponizing it for our use, there are specific developments which the United Kingdom scientists have made which hold a great deal of interest for us and which might offer advantages in our weapons systems.”
The British delegation got the same impression stating that “No doubt our technical achievements in thermonuclear warheads … have considerably impressed the United States”.[19, Sect. E38]
The Moratorium Years and the Resumption of Testing
As the Moratorium drew near, the Los Alamos and Livermore were at breaking point. Hardtack II had been planned as a series of four underground tests and seven one-point safety tests. But as the likelihood of a moratorium increased, Hardtack II turned into a series of 37 tests in total (19 weapons development and 17 one-point tests[17, p. 4]), quickly becoming a rushed operation intended to get everything the labs thought they needed tested before the door closed.
So, when the Soviets suggested they would not abide by the moratorium and General Alfred Starbird asked the labs to advise what else they could keep testing, the then director of Los Alamos Norris Bradbury was described as having “[come] close to rebellion” as it was time to stop for a while and “survey the situation”.[20, p. 107] Willard Libby — who was one of five AEC commissioners and the only commissioner-scientist — said that “there was a tremendous amount of data to be analysed” at the start of the moratorium. The data from Hardtack II was relatively easy to process, but data from Hardtack I was challenging as they struggled to understand the results from the secondaries tested.[20, p. 117]
So with all this going on, in February 1959 both Los Alamos and Livermore took the time to perform calculations on the Flagpole device, producing results similar to those obtained by the British.[1, p. 214] This was not a small feat and was no doubt challenging to work on when they had mountains of their own test data to work through.
In the same month a classification index for Fife II was distributed, meaning Fife (and Fife I) predated February 1959.[21]
At the April 1959 US-UK meeting, it was agreed to set up fifteen Joint Working Groups (JOWOGs), most notable of which was JOWOG 2 for a 500-600 lb megaton warhead lead by Livermore and JOWOG 3 for a 150 lb, 100 to 200 kt warhead headed by Aldermaston.[22] Interestingly, JOWOG 2 is one of the few JOWOGs described by Arnold in Britain and the H-Bomb.[1, p. 215]
But as Moratorium stuck around, things seemed to look bleak for the labs. At the mid-1959 Livermore stocktake, even the ever boastful Edward Teller believed that in the 600 lb warhead class, only 800 kt was possible without further nuclear testing, and that was with a “very substantial cost in oralloy”.[23, p. 3] A view that matches the above British reports in which they believe that only minor improvements in the W47 design were possible. The amount of nuclear material in the W47 using Piccolo was already extravagant, totalling 2.5 kg of plutonium, 60 kg of HEU and 36 kg of lithium-6 deuteride. For comparison, the megaton version of Red Snow — itself a clone of the W28Y1 — only used 1.6kg of plutonium, 11 kg of HEU and 16 kg of Li6D, so the W47Y1 required almost five times as much HEU and double the fusion fuel to produce half as much yield.[22]
Once testing resumed, both Livermore and Los Alamos were quick to test their 600 lb class warheads. The untested Fife was a massive and record-breaking success from the very first test, but Los Alamos’ J-21 design — which had a pedigree dating to before the moratorium — had some issues, only getting 670 kt instead of the 800 kt predicted before the test.[2, p. 166]
AWRE Reports
One interesting source of information I have found is the UK National Archives in which a relatively large number of closed Aldermaston reports are listed (that is to say that the documents are classified and only the title, date and short description are declassifiable). It’s doubtful that it is a complete list as I am sure that there are many documents whose titles are not even declassifiable, but it does give some insight into the direction Aldermaston was taking at the time.
I have done a number of keyword searches, compiled the results into a spreadsheet and then pruned the list down to what appears to be related to nuclear weapons design and analysis and not simply operation reports.
As a starting point, from 1960-1961 Fife was analysed by Aldermaston in reports titled “Fife calculations”[24] and “Calculations on the secondary of the latest FIFE design”.[25]
Searching for J-21 and XW59 gives four reports dated before the resumption of testing and four after testing, dating up to 1963. The only report that explicitly mentions J-21 is titled “Criticisms of J21 secondary implosion design”,[26] and another report titled “XW59 calculations resulting from the Questa shot” shows that Aldermaston had access to test data from the Questa shot.[27]
Up until the end of the moratorium, five reports relating to Flagpole calculations are publicly listed on the UK National Archives, including one that compares Aldermaston and Livermore calculations for the device.[28] A further nine reports on Flagpole are listed from 1963 to 1970, including a report on Flagpoles “Minus” and “Minus Plus”.[29]
The number of reports on J-21 and the XW59 are not surprising as the UK intended to use a copy of the weapon on US-procured Skybolt missiles and did later use the system under the name Simon in the WE.177B gravity bomb.[30] What is surprising is that there are 14 Flagpole reports compared to 14 reports for J-21, XW59 and Simon. Another good comparison is Dickens and Grapple Y which has 11 reports dating all the way up to 1970. Three of these reports appear to deal with devices that are “Grapple Y-like”, including the most recently listed report dated to 1970.[31], [32], [33]
Given that Dickens and Flagpole were never weaponised, it seems very strange that so many reports that involve a substantial amount of effort and computing time to create are dedicated to the topic.
The spreadsheet can be found here.
The Mad Theory
Livermore entered the Moratorium very pessimistic that further improvements in weapon yield — notably in 600 lb class weapons like their W47 — were possible, and yet by November 1959 Livermore had come up with the groundbreaking and untested XW56-X1 design using the Fife secondary (which predated Teller’s woes about not being able to make improvements) and by December 1960 confidence is so high in this design that the XW56-X0 design using the tested Piccolo secondary is cancelled.
One possible explanation for where Fife came from could be in if the Fife secondary was just a scaled up version of Tuba secondary that was eventually used in the W58, a theory posited by Graham Spinardi in From Polaris to Trident.[34, p. 213] Chuck Hansen rebukes this on the basis that Tube was spherical and Fife was cylindrical.[3, p. 431] I feel I must reject both of these positions as Spinardi’s position is based on Fife not being considered for service until the device was tested in 1962 — something that we have already established is not true — while Hansen’s position of a cylindrical secondary seems unlikely as we now know that spherical secondaries were one of the major developments that led to massive yield to weight ratio improvements. As the US was very new to spherical secondaries, I suspect that Hansen believed (and probably quite rightfully) that they would not have used Fife in such cornerstone systems if it were truly untested and therefore it had to have a close resemblance to some well-tested device or series of devices, hence it must have been cylindrical.
But, as I opened this discussion with, I would like to propose an alternative suggestion: Fife was derived from the British Flagpole device which was tested in Grapple Z2.
Firstly, some speculative specifications for Flagpole.
| Yield | 1200 kt | Same as Fife |
| Total Weight | 1500 lb | Bare warhead |
| Primary | Indigo Hammer | Unboosted. With higher yield than as tested in Antler |
| Primary weight | 400 lb | Based on weight of conventional warhead in Bloodhound Mk II missile |
| Radiation case | Heavy design | |
| Diameter | ~22 inches | Based on the diameter of the W31 warhead which is similar in weight to Indigo Hammer |
Regarding weight, in the first US-UK meeting the British presented two thermonuclear weapons: a 2200 lb device and a 1500 lb device. I assume that the heavier device was either Round A or Dickens, and the lighter device was Flagpole, with Flagpole being described as a “1-ton” weapon to mean that the assembled weapon with re-entry vehicle was 1 ton. In regards to the primary, on Wikipedia the book Cold War, Hot Science: Applied Research in Britain’s Defence Laboratories, 1945–1990 is cited stating that the Mk III Bloodhound designed to carry a nuclear warhead was based on the Mk II missile, which carried a 395lb warhead, which is assumed to be the weight of Indigo Hammer. I do not own a copy of this book to verify but I will assume that it is correct.
A Very Speculative Narrative
On the US side pre-moratorium, I suspect that the US started with solid secondary designs, with improvements to yield to weight ratio — disregarding rad cases for a moment here — coming from layered secondaries with air gaps carefully designed to produce high levels of compression through shockwave interference. Such a concept would be a natural outgrowth of levitated pits that the US pioneered before Ivy Mike. This would explain why the British delegation in those first meetings were not very impressed by the W47 warhead and why they treated it as if it was little improvement over the B28 design. It would be very difficult to spot at a glance how the different layers and mass ratios had been carefully tuned to produce the weapon.
On the British side pre-moratorium, I suspect they massively overthought the Granite series of devices, evident by Short Granite having 14 layers of material. The change to aluminium as the outer layer for Purple Granite makes me suspect that they were not using air gaps and the like, but instead were trying to design a graded ablator like what is sometimes seen in inertial confinement fusion experiments. Such an ablator would start with lower-Z material on the outside which are ablated by the initial lower temperature x-rays from the primary, with higher and higher-Z materials underneath ablated as the x-ray temp increases, creating multiple shockwaves and making the secondary compression more adiabatic.
[Post note: see the later section “Is There Another British Warhead?” for further evidence of this.]
But that did not work, so they cut the design down to something very basic with only three layers to produce Round A — so something like early solid US secondaries — which became their first successful thermonuclear test. But Aldermaston had already skipped over the basics and had bigger ideas in mind, so they reached and tried something new.
Both the US and UK aimed to improve compression as a key part of improving their thermonuclear weapons. The US was thinking about what they can do inside the secondary, while the British were thinking about how to improve compression by changing how the outside layers of the secondary ablated. Which is where I think Ken Allen’s “brilliant” idea comes in: was his idea to change the shape of the x-ray pulse? Did he come up with a way to modulate x-ray flow into the secondary? Did he invent the interstage?
I’m sure it was a relatively simple means of modulating the x-ray flow — perhaps something that produced only two x-ray pulses — but it was the start of a big idea, and later more sophisticated means of doing so would come about, such as those probably used in the Ripple device. So, Aldermaston tested Dickens which used this new concept and had a massive success, which lead to a follow up with a smaller and probably more refined design in Flagpole.
When the British presented this to the American scientists they were impressed by the idea and interested in learning more. But it was a radical idea and very different to the approached used in American weapons, which meant the Livermore and Los Alamos didn’t really have a feel for it, so they started by running their own simulations and calculations in early 1959. As was the practice when moving designs across the pond, the British device got a new American name: Fife.
But there were some big improvements that could be made. First was to replace the 400 lb primary the British were using with something lighter, and a few options in the 60 to 80 lb range were available (I won’t say any particular primary as there were lots to choose from at this time and Livermore were still trying to decide what to do regarding one-point safety) which immediately shaved 320 to 340 lb off the device weight plus more as the radiation case was reshaped to a slimmer 18” in dimeter. This gets the device down to 1150 lb or so.
On that topic, the next thing they probably did was thin down the extremely conservative radiation case to something much thinner. For example, if we assume rad case consisting of a cylinder 47” by 17.5” and a thickness of ¼” made from uranium, we have a rad case weight of about 240 kg or 530 lb. If this is reduced to 1/32” (similar in thickness to the W63 rad case I previously discussed), it only weighs 30.5 kg or 67 lb, a saving of 463 lb, bringing the total weight down to 687 lb. It’s more difficult to guess how much weight is saved trimming the diameter down due to the new primary as we don’t know the rad case shape, but it could easily have saved another 50 lb or more, bringing us down to the weight of the bare W56 warhead.
The above two weight reducing ideas are explicitly mentioned in regards to the already lightened Halliard 3 device during the September US-UK meeting,[19, Sect. E38] which I believe shows that the UK really were acting extremely conservatively in their radiation case design.
This device with a new primary and rad case was quite different from Fife, but the secondary itself was still Fife, so Fife II was born. But the US was not yet comfortable using Fife, and wouldn’t be until the moratorium dragged on and they ran more computer simulations of increasing complexity. But eventually, Livermore were confident enough in Fife.
One unredacted line in Sandia’s W56 history might support this:
“[Livermore] was thus prepared to certify the XW56-X1 design for stockpile on the basis that is assurance of performance was comparable to that of thermonuclear designs entered into stockpile in the past.”
The “assurance of performance” line is key here. They could only say that the design had comparable assurance if all the components where tested, but by their own admission the US did not test this device. Therefore, to be comparable it had to have been tested by someone else.
So, they had a primary based on previous US testing, radiation case data based on previous US testing, and a secondary based on British testing, all of which was then tied together with simulations using data and models developed from past US testing, and from this Livermore were willing to certify the untested design.
Eventually, the moratorium ends, and Livermore are able to prove their improved Fife secondary works. Meanwhile, Aldermaston are doing their own research and work, and having access to the test data from the Fife tests runs their own simulations for the purposes of testing their computer models and other general research. To them, Fife is Flagpole, so that’s the name they mostly use in their documents, but they need something to distinguish Fife II, so they call it “Flagpole Minus” perhaps to mean minus weight, and when the US develops a slightly heavier design with radiation hardening it gets named by Aldermaston “Flagpole Minus Plus” to indicate the lightweight version with some weight added back in.
Breadcrumbs in Britain and the H-Bomb
I initially suspected this idea may be true before I read Arnold’s book, with my confidence increasing as I read, but something felt off as I did as it felt like things were missing. I eventually noticed that the British Ministry of Defence owned the copyright to the book and learnt that Arnold was the official Aldermaston historian, with the book written with access to classified information. Obviously, this meant that once written, a declassification officer from MoD would have gone through and checked it for classified information and I am sure things were removed.
One example is from what follows the quote stating that Ken Allen came up with a brilliant idea for Grapple Y, stating simply “we shall see later what that idea was”.[1, p. 153] But beyond more fusion and less fission, the idea is never actually explained.
Another is in mentioning JOWOG-2 and the 500-600 lb megaton warhead program,[1, p. 215] and in mentioning that Livermore and Los Alamos ran their own calculations on Flagpole,[1, p. 214] but not following these mentions up with discussions of relevance.
From the above, I suspect that an earlier draft of the book went into more details on Allen’s innovations, and mentioned how Flagpole contributed to JOWOG-2, possibly even mentioning the US adaption of Flagpole for their own weapons. But this never made it past the censor, leaving only a few breadcrumbs behind.
Is There Another British Warhead?
[Post note: the mention of the Bradbury letter, extensions to Teller-Ulam and the British is something I only came across after the bulk of this was already written, and overall this post would have been better if originally written as a discussion of the W56, W47 and W38 as a whole, but it’s too late to revise the whole thing for that so I will leave it here as a post note.]
There is one other warhead that I have wondered has a similar British origin, which is the short-lived W38 warhead for Atlas E and F, and Titan I.
Developed by Livermore, the W38 was to be a parallel development of the Los Alamos W35 warhead but developed over a longer time period and offering improvements in yield over the W35. After the W35 gave only middling performance in Hardtack I and was cancelled (its role being taken over by the W49 which was a modified W28), the W38 program continued as a higher yield replacement for the W49. In a similar vein to the W56 program, the XW38-X1 appears on the scene in March 1959 and the original XW38-X0 is cancelled.[35]
However, unlike the W56, the W38 warhead was never tested.[20, p. 107] I suspect this was because both Polaris and Minuteman I were coming online at this time and all W38 warheads would be shortly retired, reducing urgency to test in Dominic, and by the time high-yield tests were carried out at Nevada the W38 was gone.
Some specifications:
| Warhead | W38 |
| Yield (kt) | 3750 |
| Bare weight (lb) | ??? |
| Weight with RV (lb) | 3309 |
| Diameter (inches) | 32 |
| Length (inches) | 82.5 |
| System | Atlas E/F and Titan I |
| Service dates | May 1961-May 1965 |
Sources: [8], [35]
If this theory has any merit, it would be an adaption of Dickens (Grapple Y), possibly using Li-6 instead of Li-7 to get 3750 kt yield.
There is an interesting line in Swords of Armageddon that could hint at this:[3, p. 265]
“The W-38 was based in part on technology of the W-47 POLARIS warhead.”
Which cites this citation:
“Letter dated August 30, 1961 to Honorable Clinton P. Anderson from Norris Bradbury, Director, LASL. In this document, Bradbury noted that both LRL and the British had “tried out an extension of the original Teller-Ulam concepts with moderate but hardly revolutionary success; a system of the latter sort is just beginning to appear in stockpile.””
Now, the dates here don’t quite make sense to say that the W38 was related to Fife as the W47, as this letter was written on the very last day of the nuclear test moratorium, but the author was Norris Bradbury the director of Los Alamos, who may not have known all of the details of Livermore’s programs, and the W56 and W47 were closely intertwined which muddies the waters. It’s also possible that the W47Y2 was already on the drawing board at this time.
So, either the W38 shares lineage with the Piccolo device, or in the Fife device, or in Fife and the Flagpole and Dickens devices. But the mention of the British lends some thoughts to the latter. Ideally, we would read this letter ourselves to confirm the wording, but unfortunately, only a small fraction of Chuck Hansen’s document collection has been digitised.
Some Final Thoughts
No one should look at the above and conclude that Fife was adapted from Flagpole. It is only a series of coincidences, strange decisions, odd timings and possible breadcrumbs, but nothing that actually confirms what is a very tall claim, that requires substantial evidence to back it up. Still, it is a hypothesis that perhaps be kept in mind when researching this period of American and British nuclear weapons development and is probably worth further investigation.
If you are aware of any further evidence for this, feel free to reach out.
References
Note: documents from the UK National Archives are not available online. If there is enough interest, I will try find some way to make them available.
[1] L. Arnold and K. Pyne, Britain and the H-Bomb. London: Palgrave Macmillan UK, 2001. doi: 10.1057/9780230599772.
[2] Chuck Hansen, Swords of Armageddon, vol. VII, 7 vols. 2007.
[3] Chuck Hansen, Swords of Armageddon, vol. VI, 7 vols. 2007.
[4] ‘History of the Mark 47 Warhead’, Sandia National Labs., Albuquerque, NM (USA), SC-M-67-679, Feb. 1968. [Online]. Available: https://osf.io/fytcp/
[5] ‘History of the Mark 56 Warhead’, Sandia National Labs., Albuquerque, NM (USA), SC-M-68-48, Feb. 1968. [Online]. Available: https://osf.io/fcmta/
[6] ‘History of the Mark 59 Warhead’, Sandia National Labs., Albuquerque, NM (USA), SC-M-68-51, Feb. 1968. [Online]. Available: https://osf.io/7j5qn/
[7] Director Central Intelligence, ‘DCI Briefing to Joint Chiefs of Staff’, 117940, Jul. 1963. [Online]. Available: https://commons.wikimedia.org/wiki/File:JCS_briefing_(July_30,_1963).pdf
[8] Carey Sublette, ‘List of All U.S. Nuclear Weapons’, Nuclear Weapon Archive. Accessed: Jun. 24, 2022. [Online]. Available: http://nuclearweaponarchive.org/Usa/Weapons/Allbombs.html
[9] S. Francis, ‘Warhead politics : Livermore and the competitive system of nuclear weapons design’, Thesis, Massachusetts Institute of Technology, 1995. Accessed: Jun. 28, 2022. [Online]. Available: https://dspace.mit.edu/handle/1721.1/10589
[10] Elliot V Converse III, Rearming for the Cold War 1945-1960, vol. I. Historical Office, Office of the Secretary of Defense, 2012. Accessed: Jul. 06, 2026. [Online]. Available: https://history.defense.gov/Portals/70/Documents/acquisition_pub/OSDHO-Acquisition-Series-Vol1.pdf
[11] Chuck Hansen, Swords of Armageddon, vol. IV, 7 vols. 2007.
[12] C. B. Tarter, The American Lab: An insider’s history of the Lawrence Livermore National Laboratory. in Johns Hopkins Nuclear History and Contemporary Affairs. Baltimore: Johns Hopkins University Press, 2018.
[13] R C Schneiderhan, ‘Shot Sunset, Shot Report’, EG and G, Boston, MH, J-15-3-17, Jul. 1962. Accessed: Jul. 13, 2021. [Online]. Available: https://www.osti.gov/opennet/detail?osti-id=16135561
[14] C H Builder, D C Kephart, and A Laupa, ‘The US ICBM Force: Current Issues and Future Options’, RAND Institute, R-1754-PR, Oct. 1975. Accessed: Jan. 18, 2026. [Online]. Available: https://generalstaff.org/WW3/Docs/RAND/R-1754-PR_ICBM_Force_OCT-1975.pdf
[15] ‘Proceedings of the Special Projects Office Steering Task Group: TASK II–Monitor and Sponsor the Fleet Ballistic Missile Development Program, Fourteenth Meeting–3 September 1959 [Includes Attachment]’, Sep. 1959. Accessed: Jul. 27, 2022. [Online]. Available: http://www.proquest.com/dnsa/docview/1679163232/abstract/88D14DA23F374573PQ/3
[16] ‘Atomic Weapons Research Establishment – Operation Antler – Summary Plan – Section A – Introduction – July 1957, Copy 76’, Canberra, 1957 1957. Accessed: Jun. 08, 2021. [Online]. Available: https://recordsearch.naa.gov.au/scripts/AutoSearch.asp?O=I&Number=1816933
[17] ‘QUARTERLY PROGRESS REPORT TO THE JOINT COMMITTEE ON ATOMIC ENERGY, PART III-WEAPONS, JULY-SEPTEMBER 1958 (DELETED)’, Atomic Energy Commission, LXXXI59618A, Sep. 1958. Accessed: Feb. 14, 2022. [Online]. Available: https://www.osti.gov/opennet/detail?osti-id=16091911
[18] W B McCool, ‘NOTE BY THE SECRETARY, SUBJECT: PART III – WEAPONS – QUARTERLY PROGRESS REPORT TO THE JOINT COMMITTEE ON ATOMIC ENERGY ( JULY-SEPT 1958 ) ( ENCL; REPORT RE SAME ) (DELETED)’, NV0073710, Nov. 1958. Accessed: Feb. 14, 2022. [Online]. Available: https://www.osti.gov/opennet/detail?osti-id=16091861
[19] ‘US/UK discussions at Sandia, September 1958’, 1960 1958. [Online]. Available: https://discovery.nationalarchives.gov.uk/details/r/C660497
[20] William E Ogle, ‘An Account of the Return to Nuclear Weapons Testing by the United States After the Test Moratorium 1958-1961’, US Department of Energy (USDOE), Nevada Operations Office., NVO291, Oct. 1985. Accessed: Jul. 16, 2021. [Online]. Available: https://www.osti.gov/opennet/detail?osti-id=16156584
[21] M F Moore, ‘Fife II Classification Index’, Classification Office, Sandia, ORF02160, Feb. 1959. Accessed: Jun. 09, 2025. [Online]. Available: https://www.osti.gov/opennet/detail?osti-id=16140716
[22] ‘Atomic Warheads Production Committee’, United Kingdom Atomic Energy Authority, 1965 1959. [Online]. Available: https://discovery.nationalarchives.gov.uk/details/r/C3204602
[23] Edward Teller, ‘MEMO TO A D STARBIRD, SUBJECT: MID-YEAR REVIEW OF THE LRL, LIVERMORE PROGRAM (DELETED)’, NV0125214, Jul. 1959. Accessed: Aug. 16, 2024. [Online]. Available: https://www.osti.gov/opennet/detail?osti-id=16182543
[24] ‘FIFE calculations’, Jan. 01, 1961. [Online]. Available: https://discovery.nationalarchives.gov.uk/details/r/C10891861
[25] ‘Calculations on the secondary of the latest FIFE design’, Jan. 01, 1962. [Online]. Available: https://discovery.nationalarchives.gov.uk/details/r/C10891979
[26] ‘Criticisms of J21 secondary implosion design’, Dec. 01, 1960. [Online]. Available: https://discovery.nationalarchives.gov.uk/details/r/C11508233
[27] ‘XW59 calculations resulting from the Questa shot’, Jan. 01, 1962. [Online]. Available: https://discovery.nationalarchives.gov.uk/details/r/C10892021
[28] ‘Comparison of time constants for FLAGPOLE calculated by AWRE and Lawrence Livermore…’, 1959. [Online]. Available: https://discovery.nationalarchives.gov.uk/details/r/C10891671
[29] ‘FLAGPOLE MINUS and FLAGPOLE MINUS PLUS’, 1969. [Online]. Available: https://discovery.nationalarchives.gov.uk/details/r/C10892658
[30] Richard Moore, ‘The Real Meaning of the Words: a Pedantic Glossary of British Nuclear Weapons’. Accessed: May 31, 2022. [Online]. Available: https://www.nuclearinfo.org/library/2010/real-meaning-words-pedantic-glossary-british-nuclear-weapons/
[31] ‘Pseudo GRAPPLE Y devices’, Jan. 01, 1962. [Online]. Available: https://discovery.nationalarchives.gov.uk/details/r/C10891988
[32] ‘Yield grid for “pseudo – GRAPPLE Y” devices and some optimum examples’, 1963. [Online]. Available: https://discovery.nationalarchives.gov.uk/details/r/C10892073
[33] ‘Anisotropic and isotropic explosion calculations on GRAPPLE Y – like object’, 1970. [Online]. Available: https://discovery.nationalarchives.gov.uk/details/r/C10892703
[34] G. Spinardi, From Polaris to Trident: The Development of US Fleet Ballistic Missile Technology. in Cambridge Studies in International Relations. Cambridge: Cambridge University Press, 1994. doi: 10.1017/CBO9780511559136.
[35] ‘History of the Mark 38 Warhead’, Sandia National Labs., Albuquerque, NM (USA), SC-M-68-47, Jan. 1968. [Online]. Available: https://osf.io/hdk49
Super Octopus 