Crosman 114 – Part 1

by B.B. Pelletier

I’m on the road driving to the Roanoke Airgun Expo today, so I’m asking the old hands to help me with the comments. I’m taking an extra travel day both ways to ease the strain. Naturally, I’ll give you a report of this year’s show.

I made the picture big so you can really see the rifle. Crosman’s 114 is what little boys’ dreams are made of.

Crosman’s 114 pellet rifle may have been the precursor to the modern precharged rifle. It’s a single-shot. It stores bulk CO2 gas and liquid in a reservoir under the barrel, and it has the same general look of the Benjamin Discovery. It also has a simple rear aperture sight.

Wacky Wayne saw this 114 and asked me how it shot. To tell the truth, I never shot this one before now, but the last 114 I owned shot just like a PCP.

The 114 was the .22-caliber companion to the 113, a .177 caliber bulk-fill CO2 rifle made between 1950 and 1955. Crosman had limited success with their first CO2 rifle, the CG that launched league shooting in the late 1940s. Oh, there was an earlier version of the bolt-action repeating model 118 (model 117) that existed in the 1930s, before the war, but it never went anyplace. So, the 113/114 rifles were really Crosman’s first successful foray into the world of commercial CO2 air rifles. The repeating 118 came out two years later.

At the same time these rifles were selling, Crosman also made four CO2 pistols–the 111 and 112 that have 8-inch barrels and the shorter 115 and 116 with 6-inch barrels. With all of the Crosman guns from this era, the lower number is the .177 and the higher is the .22.

The bulk-fill guns came with a CO2 tank that holds 10 oz. of liquid and gaseous CO2. The tank connects to the guns at the front of the reservoir (under the muzzle) and opens to fill the gun. I’ve recorded about 70 shots on my other 114 and 50 shots from a 111 I used to own, so I would expect this one to do about the same.

Back when these rifles were new, an owner was expected to send his CO2 tank through the Postal Service (those were simpler days) to get it refilled. Today, most serious airgunners fill their own CO2 tanks from a 20-pound bulk tank they have on hand. That large tank is sometimes used as a fire extinguisher, and I once used one of the three bulk tanks I own to put out a fire in a stolen Mustang that the thief abandoned in front of my house.

There’s another, even easier way to fill bulk-fill guns that I’ll discuss in Part 2.

The 114 is simplicity itself. A small brass receiver is attached to a brass tube that houses both the action and the reservoir. The valve is a straightforward, knock-open type, and the barrel can be made of brass or steel, with brass preferred because of condensation.

The breech is sealed at the rear, not by an o-ring, but by a metal bulge on the bolt that’s pushed into a mating flare at the rear of the barrel. That’s all that keeps the pressurized gas from coming back at your face like a supercharged glaucoma test. It does work, as long as the shooter remembers to engage the locking slot for the bolt.

The trigger couldn’t be simpler. Just a lever that gets out of the way of the hammer. The safety is equally simple–a crossbolt that blocks the trigger.

The finish is black paint on the metal and a thin shellac finish on the wood. Most of these guns have been refinished numerous times by now, so original finish on a gun is always a plus. The Blue Book of Airguns has a lot of special things to look for on these models–things that can boost the value because of rarity, so don’t start carving on the stock before you check to see if you have something special.

The rifle is just over 38 inches long with a 21.5-inch barrel. It weighs 4.75 lbs., which makes it a light, handy rifle. The stock is without comb or cheekpiece, so it’s quite ambidextrous, though the cocking-bolt handle is on the right side of the receiver.

One neat tidbit is that all three CO2 guns came with power adjusters. Even back in 1950, Crosman engineers knew what shooters wanted. Imagine a power adjuster two decades before the first affordable chronographs!

The rear sight is an aperture that adjusts via an oval slot for windage; and for elevation, the eyepiece loosens to slide up and down. It couldn’t be simpler, and it mimics the inexpensive peep sights seen on single-shot youth rimfires of the day.

This is action central. The bolt works conventionally. The knob just below adjusts the power through tension on the hammer spring. The peep sight is held in place by friction screws. And the safety is a crossbolt through the stock.

When we come to Part 2, I’ll show you how this little rifle is filled.

51 thoughts on “Crosman 114 – Part 1

  1. Morning B.B.,

    Have a safe trip to and from the show, sir. I can see some of my 180 in this gun's features.

    "Supercharged glaucoma test", my eye doctor will enjoy your word smithing. He's an old flint lock shooter.

    Mr B.

  2. BB –

    Please don't leave us hanging with little more than "…to put out a fire in a stolen Mustang that the thief abandoned in front of my house."

    There must be more to it than this. If you get a chance, please share the rest of the story with us!

  3. Hi BB,

    Hope you have a good trip. Bring back some goodies we can drool over.

    That Crossman looks real interesting. I only wish I had owned one "back then". But then I would have needed a job to pay for the CO2 and pellets.

    And you for sure know us kids had much better things to do than work "back then".

    My dad never had many guns, nor did he want us to shoot those he had nor did he teach us how to shoot.

    My first gun was a Daisy bb gun which I used to terrorize the local sparrow population. Of course they were never in any real danger as I probably managed to kill like one for every 200 shots. In fact, I was astonished when I did hit one.

    After I was 16 dad deemed that I was old enough to use regular firearms. So I got a job and bought my first Marlin 39 A lever action. Killed many a squirrel with that gun and a RWS pellet gun I also bought.

    I have been shooting and reloading ever since and love every minute. Only now it is hard to find a place to shoot. Back then I lived on a farm and the nearest house was like 3 – 4 miles "down the road". Now I have to drive 25 miles to a commercial range or to get "out in the country" where my relatives live.

  4. Phil,

    There isn't that much more to the flaming car story. It was about 11 pm on a cool summer's eve. The windows were open, & we were asleep when a bright, flaming Mustang came driving down the street. I woke up & thought the sun was coming up over the horizon til I realized what was happening. I woke up Tom, and he immediately sprang into action. Before our feet even hit the floor, our neighbor was standing outside next to the car. We saw the driver open the car door & take off down the street. Our neighbor stood & watched the car burn. Apparently, it was hypnotic.

    By this time, Tom had already run into his office, picked up a gigantic CO2 tank & ran out to the car. He laid down on the street, rolled the CO2 tank under the car & opened the valve. CO2 shot up into the engine compartment, dousing most of the fire. He pulled the tank out from under the car & continued to hose CO2 onto the car's hood. Several more sprays of CO2 under the car finally extinguished the bonfire.

    We called the police, who took statements, told us the car had been reported stolen & towed the car.

    The end 🙂


  5. BB,

    Interesting old gun. I was just given an old Crosman in .22 that has some family resemblance. So far I haven't been able to identify it's model. It's a multi-pump and is labelled Crosman "22". The rear of the bolt has a knurled knob to pull back on, with a little pin on the knob that locks the bolt when it's rotated to the right. The front sight looks very similar to your 114, and the rear sight is a simple stamped steel unit. There is no safety. It still shoots quite well. Is there a site that has photos of the old models to compare against?


  6. PurcHawk,

    I haven't sloughed off the burning car episode. Compared to other episodes in the insane neighborhood where we used to live years ago, a burning, stolen car ranks very low on the excitement scale. Having a SWAT team surround your house & tell you that you have to leave because they've been sent to arrest your neighbor…now that's just a bit more memorable & exciting!

    Tom mentioned in his original blog report that he used the CO2 tank to extinguish a burning car…no other details. I edit all his blogs before they're published, and I filled in the details because I thought they were interesting.


  7. B.B.

    "League shooting" with airguns? That sounds interesting.

    Edith, I've heard of B-17 crew members flood burning engines with CO2 to put out the fires, but they couldn't very well fail to take action. I believe that I would have let the car burn and stayed put.


  8. Matt61,

    In these situations, you never know how you'll act. While we saw the driver run away from the burning car, we didn't know if someone else was in the car. We didn't know the driver was a thief until the police told us after the fact. We also didn't know how far down the street he'd run until we got outside. For all we knew, he ran away to go roll in some dirt to put out the flames licking his body. Hard to tell from our vantage point in our bedroom.

    I believe we did the right thing.


  9. Matt,

    RE: Short answer you question in yesterday's blog about the average 30 5-shot groups and 7 30-shots groups having the same relative standard deviation

    The short answer is that if you measured the radial standard deviation by calculating the distance from each shot to the group's center, then you be 100% efficient in your use of the data. When you only calculate group size, then you are less than 100% efficient. So there isn't as much information in a 5 shot as a 30 shot group but a 5 shot group uses almost all of the available information. When you work it all out 30 5-shot group sizes have the same relative standard deviation as 7 30-shot groups.

    It is important to understand the statistics because you get the same information with 150 shots as opposed to 210 shots.

    Also I'll throw into the discussion fliers again. With 30 5-shot groups you have enough data to calculate a decent std dev using the data alone. So you have flexibility in using a 95%CI to throw out any groups which had an unusually large group size due to a flier. With only 7 measurements of 30-shot groups it will be much harder to detect a bad group due to a flier.

    There is another way to look at this wrinkle. Think about dividing 150 shots into smaller groups.

    Since there are the same number of shots in each 5-shot group, then the average of using all 150 shots in one calculation, is the same as the average of the 30 5-shot groups. However if you calculated the standard deviation of over all 150 shots, then it won't be the same as the average standard deviation of the 30 groups. The standard deviation over the 150 shots would be based on 149 degrees of freedom (lose one degree because you measured mean with same data). However the 30 groups would each have a mean measurement. So the average standard deviation of the 30 groups is based on 120 degrees of freedom. So the average standard deviation from 30 5-shot groups has the same % error as the standard deviation of 1 131-shot group. So by using group size, you lose information.

    For a more details on efficiency of the Gaussian distribution see my comments in yesterday's blog.


  10. BB,
    Classic is the only word for this one. I don't know why anyone would want to touch that stock, as it looks perfect as is.

    You and BB did do the right thing.

    I see your issue with the scope. 10X is hard to use offhand, and even 6X takes some real patience. You can see why there is a market (though quickly drying up) for lower powered scopes. 2.5-3X is pretty nice offhand and in hunting situations. I don't have much experience with red-dots, but they must work pretty well for those types of applications.

  11. Edith,

    Putting out the fire in a burning car is a very brave thing to do. I guess I myself would have trouble in the wee hours putting myself at serious risk for the sake of people who may or may not be at risk. But I am glad there are people who are willing and would know what to do.

    Herb, I have no doubt that more calculations can be done, but it seems that after shooting one 30 shot group and thirty 5 shot groups that the results are relatively stable. Going further down this road seems like a matter of pouring more effort in for less return. Anyway, additional testing will have to be done by someone else besides me. 🙂


  12. Edith:

    Such an exciting neighborhood I could do without. It's never a good sign when a SWAT team shows up.

    I was in no way criticizing your editing. I am a professional book editor, and I greatly admire both Tom's writing and your editing thereof.

    I meant my remarks to be ironic and light-hearted, forgetting for the moment that comments, forums, and e-mail will seldom support irony.

    My compliments on your continued success with this wonderful blog/forum. I think it is unique on the Web.

  13. PurcHawk,

    No apologies needed. No offense was taken. In fact, I had to go back & re-read your & my comments to see why you responded the way you did. But, I agree that irony, sarcasm & humor can be difficult to convey.


  14. Matt,

    I agree that the 5-shot data should be stable since it is based on 30 groups.

    However I'm confused as to what result that you think is "stable" for the 30-shot group. You estimated the size as 1.50 inches. subtracting 0.177 for c-t-c your measurement was 1.32 inches. That is too large compared to the 0.70 inches that a thirty shot group should be based on your 5 shot data.

    I came up with an estimate of 0.80 inches which I do think is a pretty reasonable interpretation of the data.

    Just to be sure I did do the formal analysis on the group size from your target image. There are not values for a 29-shot group in the tables, so I just did a linear interpolation between 28 and 30. Not quite right, but it should be close enough.

    Mean(29) = 0.5*(4.734+4.788)= 4.761

    95% CI low = 0.5*(3.605+3.678)= 3.642

    95% CI high = 0.5*(6.113+6.170) = 6.142

    So a 95% CI for an individual group based on one group of 29 shots is:
    Low = 0.80/4.761*3.642 = 0.62
    High = 0.80/4.761*6.142 = 1.03

    Since the group size of:
    1.37-0.18= 1.19
    is larger than 1.03 there is reason to throw out the one shot and assume that the best estimate for the group size (c-t-c) is 0.80 -0.18/+0.23 inches at the 95% CI.

    I like it when theory and experiment agree. It seems to me that your data is consistent with the theoretical predictions. The theoretical calculations using the tables of Taylor et al offer a much richer analysis than simply assuming a Gaussian distribution. You can of course assume that group data is Gaussian, but you'd be handicapping yourself.

  15. Herb,
    That's another good example of using the math to properly cull a flier, I think. My belief is that if you are going to use large (>10?) shot groups, that has to be part of the process, just because shooters are human and ammo is imperfect. The purpose is not to exaggerate the accuracy of the rifle and ammo. but to characterize it fairly.

  16. One other note on my last statement. I don't mean to imply that we should throw out fliers when we're recording groups for match scores or bragging rights (where shooter skill is the point), only when trying to characterize the rifle and ammo, such as when BB tests a rifle with large groups. Example, a 20 shot group with 18 inside .5" and 2 outliers blowing the group up to 2".

  17. BG_Farmer,

    I understand what you're saying. You want to use the mathematical analysis to get as much information that you can.

    Like I said earlier, experiments lead to more questions. So knowing that you had 2 fliers in 20 shots, as you proposed, is a "different problem." I shoot so poorly that I'd certainly assume that the problem was me.

    A match shooter with great technique should probably assume that 2/20 or 10% of the pellets are "bad." He would then need to design an experiment to quantify the % fliers better. He certainly wouldn't want to go to a match with ammo where he suspected that 10% of the rounds will be fliers.

    So like BB takes the problem apart, Is the barrel fouled? Can you sort the pellets? Is it as simple as the particular tin of pellets is bad? Are all the pellets of this brand as bad? Questions within questions…


    I do wish to thank you for the questions and especially for the work you but into the experiment.

    I knew group size would not be Gaussian, but I couldn't figure out what the distribution was. It wasn't till I found the paper by Taylor et al that I finally made sense of all of it. Nobody has found a particular distribution that fits!

    The other part that bugged me was it seemed like there should be some sort of relationship between the group size for multiple shots and the variability of individual shots. Analogous to the interplay between the standard deviation and the range for the Gaussian distribution. Of course if you just assume that the distribution for group size is Gaussian, you lose that interrelationship.

    So again thanks. This has been a lot of fun to me. As you can tell, I'm a math nut. I'm pretty good with statistics, but I'd never done this sort of analysis before. This was certainly a learning experience for me, and I really enjoyed learning something new.


  18. Herb and BG_Farmer, maybe a better word than "stability" is "fuzziness." For instance, one could define a flier based on a 95% confidence interval. But why 95%? Because it's a standard in laboratories? That's an arbitrary choice. Why not 90% or 99%?

    The bigger question underlying the whole shooting test in the first place is how do we know the theoretical models are correct? They're good approximations, but even the language of the Taylor paper was qualified saying that the chi distribution fit the data relatively better than the lognormal distribution. But the fit wasn't exact. Even the data, as I understood it, was not based on real shooting but on a computer simulation which was itself based on some assumptions. We don't really know how it would compare to actual shots. It wouldn't be the first time that elaborate scientific models just have not been correct for reasons unknown. Given this uncertainty, the precision of calculations beyond a certain point seems hard to justify.


  19. Disco owners,
    Mounted a bi-pod on my Disco (the one recommended by the tech at PA).

    This is GREAT for bench shooting off the deck and an expected added bonus is some nose weight. The Disco with a big scope is nose light. So hanging a bi-pod on it it helps with off hand shooting. If you've been thinking about it just do it.

    Another bonus is you'll have to add sling pins for the bi-pod so the Disco is then sling ready.

    The bi-pod removes in attaches in a snap. So it is not a burden to add on and take off.


  20. BG_Farmer, Matt, Herb,
    Been following your math thread. Now that you have it all settled…

    Please consider you may have better success in measuring chrony results of a given gun rather than target results. Take a 40-shot string and eight 5-shot strings and look at the results.

    What do you think?


  21. DB,

    Which bi-pod did PA recommend?

    BB, I look forward to seeing you at the show this weekend. I'm inclined right now to bring that Crosman Model 99 I bought through an ad I placed in the local pennysaver paper. I want to let you look at it and give me your opinion on it's condition. I "think" it ranks around 90-95% but am curious to see what you think.


  22. DB,
    Herb and Matt did all the work; I just interjected a stupid question or two.

    Are you talking about a chrony string from a springer or pumper or PCP in its sweet-spot? With unregulated PCP's, there will be a large-scale "non-random" variation from valve-lock to sweet-spot to running out of air, but the variations inside the sweet spot might be interesting.

    Note that I don't own either a PCP or a chrony, but I've seen enough shot-strings on-line to get a feeling for them:)!

  23. DB,

    Heh heh, nice try, but my general feeling is that the same statistical dynamics should apply. 🙂 After all, velocity is related to accuracy. More importantly, the normal distribution in all its variations applies to any situation with many variables each of which has a range of values around a mean. In other words, most complex situations. It has no special attachment to the distribution of pellets. It shouldn't be hard to run a few tests for those with chronographs. There are a lot of shot strings already in the blog. Anyway, my prediction is that the distributions will be exactly the same.

    I must say that I feel quite empowered after the statistics discussion. I was reading an article tonight about the Kimber 8400 police tactical rifle chambered in .300 Winchester. The reviewer's only data was that his best group was 3 shots in .43 inches at 100 yards. Says nothing. Borderline fraud. In my own string of thirty 5 shot groups, there was one group where I soared with the Olympians at 4 MOA. But I spent an equal amount of time–5 shots–down in the cellar at 20 MOA.


  24. DB,

    To clarify, I'm claiming that the statistical distributions within a velocity spread should be distributed normally more or less. Outside this range, there will be little variation. In other words, my answer is very similar to BG_Farmer's.


  25. Matt,

    The 95% is somewhat arbitrary. If the distribution is a Gaussian distribution, then cutting off the tails at +/-2.5% won't change the mean or standard deviation to by a significant amount. Trimming the tails does protect against outliers (fliers) since group size is particularly sensitive to fliers.

    Statistical tables are generally calculated for 95% and 99% and a few other values. If you use 95% or 99% isn't as critical as the fact that you apply the criteria consistently. You shouldn't throw out fliers on one target at 99% and a second at 95%.

    The 95% is also somewhat based on notion that you will have a limited amount of data. If there were a million measurements, then throwing out 5% or 50,000 measurements would certainly be odd.

    The statistical experiment is easy to simulate. You assume the X and Y coordinate are normally distributed and that both have a standard deviation of 1. You then generate random numbers between 0 and 1. You back-calculate how many standard deviations from the mean that number for the cumulative probability would be. So 0.500 would be zero, or the mean. About 0.16 would be -1 std dev away from mean, and 0.84 would be about +1 std dev from the mean.

    So you generate random pairs of X-Y numbers which are the standard deviations from the mean. You then group the random pairs into "shot groups." You find the two points the furthest apart and calculate the "group size" for that "group."

    You simulate thousands of groups. You can then plot a histogram of the data to get a distribution. You then try to find a mathematical function which "fits" the simulated data.

    So the model isn't just absolutely arbitrary. There is a very reasonable set of assumptions behind the model.

    Just like the fact that "group size" isn't Gaussian, in truth the model surely doesn't ABSOLUTELY describe reality. The model is closer than assuming a Gaussian, and perhaps reasonable enough that only an unreasonable amount of shooting would be required to show that the model was wrong.

    Who is going to shoot 100,000 shots and measure the X-Y deviations? The barrel would be worn by 100,000 shots thus changing the distribution.

    The 100,000 shots aside, it would be reasonable to also measure the ratio of the width of groups to the thickness (An aspect ration like an ellipse). On average, the groups should be circular. An aspect ratio other than 1 would indicate that the assumption that either X and Y errors were not equal (eg horizontal or vertical stringing) or that the X-Y positions were not independent (eg, stringing at 45 degrees to horizontal or vertical). So you can keep peeling back the assumptions in the model ad infintum.

    There is also the question of the random number generator. All sorts of testing can be done to see that the random number generator does indeed generate numbers which are random for the type of testing being done. Of course "random numbers" are not truly random. The numbers are generated by a computer program. If you know what the program is, and you know the "seed" value, then you can calculate what the next "random" number will be.

  26. DB,

    RE: Chrony numbers

    You really shouldn't start with numbers then try to see what you can figure out. The real question is what do you want to know? Then you try to design an experiment around the question.

    For example a "string" for a PCP would have different questions and assumptions than a "string" for a springer as BG_Farmer pointed out.

    I worked at a semi-conductor plant doing chemical analysis. The production line was forever having problems. The problems that got sent to us never had the appropriate data to analyze the problem even though the production line was trying to make every measurement that could be automated. We were always trying experiments to figure something out.

  27. I have Crosman 114 in new condition, I've had it for about 20 years a buddy of mine found it on a job site he was working on, he was doing some demo work on,it was in the rafters ,the rifle was in a canvas and leather case perfect size for the rifle,if some one could tell me where I could get a co2 cylinder it would be appreciated. Thanks LC.

  28. LC,

    as you read the article, you understand this is a bulk fill CO2 gun. You need a bulk tank which is sold at Pyramydair


    or where paintball guns are sold. However, you will also probably need an adapter and hose and fitting for that rifle. Pyramydair will be able to help you out.

    Fred PRoNJ

  29. I have what I believe to be is a crosman model 113 with no trigger or tank. Its been sitting in a closet for about 16 years and was never once shot from the time it was purchased. I was wondering if anyone rebuilds them or if there is somewhere I can find parts for it. Any help is greatly appreciated.

  30. Mr. Pelletier I really hate to bother you again especially about the same rifle but could you possibly pull the trigger Assembly out of your rifle or find a picture of the trigger assembly. I went to the website you pointed me to and although it was a great help I have decided to try to find another 113 or 114 to make one good rifle or try to find/make a trigger assembly. I took the bolt and firing assembly out today and although the seals are hard all I believe I need to fire it is a air tank or adapter and a trigger : ). If it helps you at all I can give you a picture of what I’m working with. Once again any information is greatly appreciated.

Leave a Reply