FWIW,
according to data posted at Ballistics By The Inch most 357 ammo seems to hit peak velocities out of about a 16" barrel, and that peak is approximately 400 fps or 20%-25% faster than the same ammo out of a 4" barrel.
Velocity Data
BBTI - Ballistics by the Inch :: .357 Mag Results
You have to take the BBTI data with a huge grain of salt, especially when it comes to barrel length for peak velocity.
The problem is their methodology. They shoot 3 rounds of each type of ammo in each barrel length, but shoot it across two chronographs placed in line to get 6 data points. That isn't 6 data points, it's still 3 data points now confused by 2 chronograph measurements a few fps apart. Even if it were 6 data points, that's about 1/3 rd the number you need to start getting reasonably reliable (repeatable) data.
Let's demonstrate this with the numbers for the Federal 158 grain load. First, let's look at the 7", 8", 9" and 10" data:
7" = 1517 fps, SD = 10.3 fps
8" =
1570 fps, SD = 14.6 fps
9" =
1563 fps, SD = 37.9 fps
10" = 1630 fps, SD = 13.1
You see a 53 fps increase between 7" and 8", then a 7 fps loss between 8" and 9" followed by a 67 fps increase between 9" and 10". The average velocity for just 3 shots at 8" was probably higher than would have been the case with a reasonable sample size of 10-15 rounds. Conversely, the average of the 3 rounds fired at 9" was probably lower than would be the case with a reasonable sample size.
You are also seeing the other effect of small sample size, where the standard deviation in velocity far exceeds the difference between velocities in adjacent barrel lengths, particularly in the 7", 8" and 9" length data.
Next let's look at the 11+ to 14" data. You see a 9 fps increases between 11" and 12" and a 20 fps increases between 12" and 13", and then a huge increase of 58 fps from 13" to 14", with a large SD in the 14" data.
Most of this is probably due to small sample size errors, but some of the data in some of the lengths may be due to variation in bore size along the bore. Tighter sections of bore will produce more velocity per inch than looser sections of bore, and BBTI isn't using high quality air gauged barrels hand lapped for consistency.
As background on Standard Deviation, that 23.4 fps SD means that the expected range in velocity over 100 shots with that load/barrel length would be +/- 3 SD, or in other words 6 times the SD, or 140.4 fps. Here, in the 3 shot data collected at 14" there was a range of 59 fps. Normally, you'd expect 68 out of 100 shots to fall within 1 SD of the mean, but the odds are that you'll get a shot 2 SD from the mean about 27 shots out of a 100 and 5 times out of 100 you'll get one that falls in the 3rd SD from the mean. Which means about 1/3rd of the time using 3 shot samples, you'll get a shot that screws up your average.
You see that here in the 14", 15" and 16" data and you are seeing it again in the 16", 17" and 18" data.
11" = 1643 fps, SD = 9.8 fps
12" = 1654 fps, SD = 12.1 fps
13" = 1674 fps, SD = 5.23 fps
14" = 1732 fps, SD = 23.4 fps
15" =
1708 fps, SD = 18.8 fps
16" = 1739 fps, SD= 19.5 fps
17" =
1689 fps, SD = 16.4 fps
18" = 1719 fps, SD = 20.5 fps
Even if all the shots fall within 1 SD at 18" you can expect a spread of 40 fps, which is greater than the difference between the 17" average velocity the 18" velocity. Similarly, at 16" the expected spread at 1 SD is still 39 fps, and that's again greater than the difference between the 17" velocity and the 15" velocity.
That means that even at the relatively low 1 SD, 68% confidence interval there is no significant difference between the 16" and 17" velocities or the 17" and 18" velocities.
In short, that means there is no evidence in the data that a peak velocity has occurred. And then of course BBTI data starts at 18" so they never collected data for the .357 Mag in longer barrels.
I happen to own Rossi 92 rifles in 20" short rifle and 24" rifle patterns and I happen to have collected 30 shot sample data at each barrel length with this same Federal load.
Interestingly enough I get an average velocity of 1807 fps in the 20" barrel and 1820 fps in the 24" barrel. That's only 13 fps over a 4" change in barrel length but:
1) it's reliable data;
2) it's still an increase per inch; and
3) it's a lot higher than the 16" and 18" BBTI data showing that there is indeed an increase past 16", and a large enough difference to offset any bore related differences.
----
Now...there is a point of diminishing returns for each cartridge and load combination where you get less velocity per inch, and that extra barrel length stops being worth the extra length and weight. In the case of the .357 Magnum, I'll agree 4" more barrel for 13 fps more velocity is not worth it (although balance of the rifle, a slightly lower SD in velocity, etc might still cause someone to choose a longer barrel).
In the .223 Stoner used an 18" tube in the AR-18 and AR-180, Ruger used an 18" barrel on the Mini-14, and the Israelis used an 18" barrel on the Galil, in large part because 18" is the sweet spot for the 55 gr .223. You can get more velocity at 20" and still more velocity at 24" and 26", but it's generally agreed that the rate of velocity increase per each additional inch of barrel starts to decline after 18". The velocity still increases, just not as much per inch as it did at shorter barrel lengths.
Conversely the rate of loss of velocity per each inch of reduction in barrel length increases as the barrel gets shorter than 18" in the .223.
Unfortunately, many shooters misunderstand what that means and start saying things that are totally incorrect, like "a .357 Magnum bullet will start slowing down in barrels longer than 16 inches", or more often " a .22 LR bullet will start slowing down in barrels longer than 16 inches."
Even in .22 LR that's not the case, I have 16", 18" 20" 24" 26" and even 28" .22 LR rifles and the velocity of even standard velocity ammo will continue to increase, just at progressively smaller amounts per inch. However, the SD also decreases in those longer barrels, so you get more consistent velocities, and that in combination with a longer sight radius is why you seen longer barrels on smallbore match and target rifles.
With the .22 rimfire, a .22 CB cap powered only by the primer compound might start slowing down in long barrel. However, for just about everything else, the distance in the barrel where the combined resistance of the bore and the pressure of the air in front of the bullet is greater than pressure of the gas behind the bullet is measured in yards, not inches.
