A couple of points on this. One is that SAAMI only specs outside case dimensions to ensure chamber fit. What the brass manufacturer does to the inside is up to him. The brass just has to perform in the intended chambering with a proof load charge. Proof loads are always fired in standard cases (so brass isn't helping the gun unduly), and standard brass has to survive that event (but just once). Since .38 Special proof loads (25,000-27,500 psi) are above +P (18,500 psi) and even the non-standard +P+ pressures thus far tested (22,000 psi on Fr. Frog's site), you can be sure standard brass is up to them for at least one firing, but case life could be pretty poor in some instances as far as tolerating reloading goes.
Actually, case life at those pressures is much better than you'd think. I have two Rossi 357mag leverguns and with Quickload's help, I've developed a great 38spl+P+ load specially set up to work well in leverguns. First of all, leverguns are case length sensitive so they cycle better with 357mag and long loaded 38spl. A standard 38spl has an OAL of 1.460" while a 357mag's OAL is usually 1.580". I've found that rounds with OALs of 1.500" and longer eliminates any case flipping and other feed failure issues in both my 20" Rossi carbine and 24" Rossi rifle.
With that understood, and with a very large supply of Zero 158grn JSP bullets, a few kegs of H110 powder, plenty of magnum primers, and a large supply of 38spl cases and few 357mag cases, I undertook to find a load that worked well in my leverguns with the components I had on hand even though there is no load data nor any recommendations to use H110 powder in any 38spl loads.
I understood that loading long will reduce pressure, that H110 likes to work in a high density environment, and that 38spl cases won't hold up long if overstressed. What I didn't know was what was a reasonable pressure limit for the 38spl brass as I wasn't concerned about the rifle being damaged because it's proofed for 357mag level pressures.
Anyway, I looked at the 38spl's remaining case volume when a 158grn bullet is loaded to a OAL of 1,500" and compared that volume to a 357mag case with the same bullet loaded to the recommended 1.580" OAL. Then I looked at the load data for H110 and noted that the starting was 15.0grns and max was 16.7grns with a similarly shaped 158grn XTP bullet.
Next, I calculated the percent reduction is case volume between the the long loaded 38spl and the normal 357mag and applied that percentage to the starting and max loads of H110 to get an idea of a target range. I wanted to maintain the same case density to ensure that the H110 burned completely in this hypothetical load.
Based upon that data, I selected a target powder load of 14.0 grns with that bullet for my Rossi Levergun only 38spl+P+ candidate. Then I loaded that information into Quickloads as well as information on a 357mag version using the same bullet to see what the pressure and velocity predictions would be.
Quickload didn't have the Zero JSP bullet in it's database but it had the Speer JSP in 158 grain. The Speer bullet has a length of 0.630" and is seated 0.275" deep. It turns out that the Zero bullet approximates the length of the Speer but sits a bit farther in a 0.290". The results are as follows:
In a 20" rifle barrel Quickload projected my 38spl+P+ load using 14.0grns of H110 at an OAL of 1.500" would produce a peak pressure of 20,411psi (vs the peak SAAMI spec for 38spl+P of 18,500psi) and 1,525fps while the .357 Magnum load with 16.7 grains of H110 at an OAL of 1.580" would produce a peak pressure of 32,400psi (vs it's SAAMI peak pressure of 35,000psi) and 1,798fps. I was encouraged by the max pressure level predicted for my 38spl+P+ load so I worked up 20ea of them and 20ea 357mag to take to the range with my chrono and actually measure the average velocities to compare with those predicted.
My thought was that if the actual velocities were close to the predicted velocities, then the actual pressures would also be close to their predictions. The actual velocities measured out of my 20" Rossi carbine for the 38spl+P+ averaged 1,503fps while the 357mag velocities averaged 1,799fps on a 78° day at 6,100ft above sea level.
Encouraged by the results, I next did a life test on 20ea 38spl cases with my 14.0grn loads. The cases were a mixture of Winchester, Remington, and Speer. Each case was loaded and fired 10 times and of the 20 cases tested, only 1 case failed after 9 reloads; it split. The case failure caused no noticeable issues with the rifle or even the load's performance. In fact, the split case wasn't even noticed until I got it home and tumbled it clean.
Note: Later I shortened some of the 38spl, 14.0grn loads to an OAL of 1.460" with a Quickload predicted peak pressure of 22,962psi to test what higher pressure's affect would be on case life. I ran 10 more 38spl cases for 10 reloads at this higher pressure and only had 1 case failure after the 10th reload. Again, it was a split case that produced no noticeable bullet performance issues or extraction issues.
From this data, I've surmised that my 38spl+P+ load is safe when fired from my Rossi M92 leverguns because, a) the peak pressure is far below the max pressure limits for a 357mag, and b) at even the elevated pressures of the 38spl+P+, the 38 cases proved durable enough to safely work in a Rossi M92 levergun for a reasonable number of reloadings.
Note: This load is above SAAMI specifications for a 38spl+P round and should not be undertaken lightly. You assume all risks of loss, liability, and injury should you elect to duplicate these loads in your firearm. Further, I've not tested, nor can I recommend the use of this load in any other firearm.