Fosbery’s Pump Shotgun: An AR15 Bolt in 1891

Armament Research Services (ARES) is a specialist technical intelligence consultancy, offering expertise and analysis to a range of government and non-government entities in the arms and munitions field. For detailed photos of the guns in this video, don’t miss the ARES companion blog post!

George Fosbery V.C. is best known in firearms circles for the Webley-Fosbery automatic revolver, but he experimented with several other firearms designs as well. This particular one never made it into commercial production, but it uses a bolt design very reminiscent of what Johnson and Stoner would use 50+ years later, with six independent locking lugs around the circumference of the bolt head.

This firearm actually began as a pump action rifle before being converted into a shotgun (using a Winchester barrel), which makes it all the more interesting. It was originally made circa 1891, and later converted from a rifle into a shotgun (hence the 1909 model barrel).


  1. The concept of a six lug bolt probably didn’t work out for the tech level of the era. I refer to commercial sales and production costs, of course. How much did one wish to spend on machining the bolt head? And yes, feel free to correct me if I’m wrong about the matter. In any case, many concepts can develop long before they can be put into production. The concept of hybrid cars was already around before the Great War. The tech level of the time wasn’t up to the task of making the hybrid car cost effective for manufacturing or maintaining (from the driver’s perspective, anyway)…

    • Machining bolt head? I’d say maybe tracer template; there were many innovations in machining popping up at that time. But for prototype, probably not.

      • If you were making the bolt and receiver 10 pairs at time it would take a good machinist about 5 or 6 days to make them on good tools and have them ready for heat treatment. Sixty hours of a machinist time in 1898 in the US was between $2.00 and $2.75 and hour Ten of then would have about 60 hours at $2.50 is $15 bucks each. It would be hard to sell an ugly $40 dollar gun when my gand day paid $35.00 for L.C. Smith Double barrle.

      • Small parts shaped like that from the era were usually made on a shaper (in my experience). The form would be ground into the tool and the cutting head lowered slightly after each stroke. Once the part was shaped to size, the part could be indexed. This requires one tool to be ground. Optionally the part(s) could have been milled or gang milled on a horizontal milling machine. That would entail much more work as you would need a rotating form cutter (ya know, with the multiple teeth and all). Yes, tracers existed and were used early on on the feed ramp of the 1903 Springfield at least as early as 1916.

    • “How much did one wish to spend on machining the bolt head? And yes, feel free to correct me if I’m wrong about the matter.”
      Back then economizing design wasn’t applied as heavily as later, see for example Repetierpistole M.7 as used by Austria-Hungary.

      “tech level of the time wasn’t up to the task of making the hybrid car cost effective for manufacturing or maintaining (from the driver’s perspective, anyway)”
      In fact first car to go over 100 km/h fast was electric-powered:

  2. Excellent schooling lesson. And yes, with its bolt it is very similar to my Winchester pump-gun, so the name on barrel is not far off. What strikes me is common sense G. Fosbery applied in form of lugs – rounded root and rounded crown, very much like some recent “high tech” DM/ sniper rifles.

    When comes to English attitude to shotgun shooting, it had been to my knowledge traditional to the last drop – with side by side guns and this is truly sportsman-like and fair to the game. Pity for Fosbery’s inventiveness which came on flat.

    • Making two shots is faster with a double barrel than with a pump action, so I don’t know about the fairness… For hunting waterfowl and grouse a double barrel is very good even today.

  3. The Mondragon straight pull of 1894 also had 6 lugs in the front.
    I always wonder why the bolt action rifles (Mauser, Moisin etc) of that timeperiod all just have 2 lugs.

    • “Moisin”
      Mosin (unless you intentionally used that name, to show disregard to that design)

      “all just have 2 lugs”
      Why to use more, when that number is enough?

    • One more thought… if you cock striker on closing or opening of the bolt, you have to energize striker spring. This is happening during locking/ unlocking portion of travel while bolt rotates. With 2 lugs (and as mentioned in other note downstream) the travel is relatively long which suits the purpose.

      • In battle and field conditions bigger rotation of handling bolt is a safety feature. The bolt might strike anywhere and leave its locked form. Autoloaders and other hand actuating actions have no big outside limbs like bolt actions. Hunting types might tolerate rather short bolt rotation since having lesser fatal risk than war conditions. IMHO.

  4. Im sorry for “Moisin”.

    “Why to use more, when that number is enough?”
    With 6 lugs you just have to rotate the bolt 30°, or?
    So you gain time.

    • With six vs. three or even too lugs, given same helix angle you end-up with shorter advance/ throw. If space in longitudinal direction was restricted then this approach is justified.

      In case of AR15 they wanted compact/ light weapon and with 22.5deg of rotation at around 45 degree helix you end-up with miniscule .31″ advance – very compact. If I recall it right the AR15 upper length is in neighbourhood of only 7 inch. This proved to be an issue while considering longer cartridges such 6.5 Creedmoor.

      If in future U.S. military was to implement round of greater potential than 5.56×45 it is bound to start with fresh design.

    • But on the other hand, if your 30 degree stroke is doing the same work (cocking the striker), it will take twice the torque of a 60-degree throw, or three times that of a 90-degree throw, to lift the bolt handle. You can mitigate this with a longer bolt handle, but of course that costs you the speed advantage you hoped to gain.

      (That’s all assuming cock-on-open designs, of course. Cock-on-closing is a whole other thing…)

      Basically, once you decide what force is acceptable for opening, that sets the actual distance the bolt knob will move; you can then pick any balance from a short handle with a long angular throw, to a long handle with a short angular throw, but since they all move the same distance, there’s no real speed-related reason to prefer one angle over another. (There’s other reasons — for example, the clearance issues when mounting a scope for large angles, or how the cocking cam will cause excessive friction for too small angles.)

      And Denny’s argument applies not just to the AR-15, but anywhere you’re using a helical cam to convert linear to rotary motion, whether that’s a gas-operated semi-auto, Fosbery’s pump-action repeater, or a straight-pull bolt action. So even though the Garand and Kalashnikov have two-lug bolts, they use relatively small angular throw, for the same reason.

    • “With 6 lugs you just have to rotate the bolt 30°, or?
      So you gain time.”
      Even if so, it is so small advantage that it is not worth implementing it to bolt-action repeating rifle (for general infantry use). Even straight-pull bolt repeating rifle seems to so little advantage that although adopted by few countries do not made bolt repeating 4-movement repeating rifles extinct from military service.

      • Probably because the main rival to the Mannlicher rotating bolt straight-pull action was the excellent Mauser type action, which was seen as more robust and dependable by many armies shopping for a bolt-action rifle at the time, and probably so. Mannlicher action was also typically coupled with the en bloc clip, which had some disadvantages compared to stripper clips, namely the open bottom of the magazine (although there were solutions for that) and the more expensive clips.

    • “So you gain time.”
      If you want time gain, I think Mauser Modell 66:
      is good example of improved speed 4-movement bolt-action repeating hunting rifle.
      Thanks to its design, it offer shorter stroke (distance between bolt handle in full open and full closed positions) over same-era same-cartridge repeating rifle.

      • It moves receiver ‘together with bolt’ (probably borrowed from self-loader) thus makes shorter built-in length (no return spring); so the effect is more less just visual. But then, you still have to get round into chamber and eject, just as any conventional action. Btw, it appears hand fed, round by round, which helps.

        Yeah I know, mechanics are catchy and nice to look at…. 🙂

      • To give you sample of my ‘practical’ thinking….. I was recently selecting small bore rifle. From .223Rem. down to .22LR. I really liked Browning T-bolt in .22WM – it’s slick and fine made by Miroku.

        Then I factored in availability/cost of ammo and other details such as what they want/ do not want to see on ranges in my province and settled for inexpensive and accurate Savage. I have mine with laminated stock and heavy profile barrel as you can see here:

        My preliminary experience with it shows there can be had lots of fun with simple rifle. Perfect at 100m/ yrds with both subsonic and supersonic ammo, no hearing protection needed. KISS principle in action 🙂

  5. From a machining viewpoint, multiple lug bolts can be generated in the same manner as, splines and gear teeth, on various sorts of shapers (check out the bilgram shaper, that is used to generate the correct gear tooth profile on bevel gears, where the profile is maintained as the tooth becomes smaller as it is followed towards the Centreline of the gear)

    And by hobbing, where the profile is generated by continuously rotating both the workpiece and a helical hobb cutter.

    In military bolt actions, the strength of chambering and extraction camming were generally seen as crucial. A full 90 degrees of bolt rotation allows for both substantial camming mechanical advantage and displacement, plus sufficient bolt rotation left over for seating the locking lugs,

    A Mauser 98 apportions 60° of bolt closure to chambering camming, and the remaining 30° to seating the locking lugs. Chambering camming achieves a displacement of quarter of an inch.

    sporting guns are unlikely to be used in such hot, dirty and dangerous situations, so, lesser camming gets used from time to time, however a sixty degree bolt rotation like a wetherby mk v, needs to fit both camming and locking into the rotation that a Mauser 98 uses for camming.

    Ideally, an action (bolt, pump or whatever else) should positively withdraw the firing pin tip as the bolt is unlocking, and keep it positively held out of the way until the bolt is fully locked. A spring alone is asking for trouble (in Olson’s case, a broken firing pin spring got him prison time) This doesn’t need much displacement to achieve, for example even a garand with its minimal bolt rotation, positively cams the firing pin tip into the bolt.

1 Trackback / Pingback

  1. SayUncle » Gun Porn

Leave a Reply

Your email address will not be published.