What I found fascinating was its apparent ammo preferences. Now, this particular pistol is one of a handful of prototypes here in the U.S. The "production" of them was more like gunsmiths hand-fitting parts from short CNC runs off the computer-driven milling machines. So the accuracy and the acceptance of ammo of this particular pistol should not be taken as the norm for the ones you'll get.

This one does not like hardball--a situation I'm certain Beretta will correct, as that is mostly what the regular guns will get fed. Give this gun 230-grain full metal jacket ammo and you'll get groups four inches or larger at 25 yards. With some brands I tried, groups were a lot worse.

Give this particular .45 Storm high-speed hollowpoints and it is one happy pistol. The prototype gun didn’t much care for hardball ammo, though.

At first I thought it was me, and I checked myself by switching to a high-end 9mm target gun I had along on the same range trip: five shots into just over an inch at the same distance. So back to the PX4 Storm in .45 ACP I went: a five-inch group. Ouch. Out of equal parts frustration and desperation, I tried some high-speed hollowpoints. Wow. Five shots of Black Hills 185-grain JHPs went into three inches on the first group, and once I'd gotten accustomed to the brisker recoil, I could do better than that all the time.

Why the difference? I can only speculate. Rotating-barrel designs use barrel rotation counter to the torque created by the bullet engaging the rifling. (Barrel rotates left, torque pushes right, or vice-versa.) If the bullet is slow, the barrel may have time to rotate enough to change the fit of the barrel muzzle to the slide opening. Or not.

The PX4 Storm uses frame rails of steel, cast in place in the injection-molded frame. The cam block which controls barrel rotation is a separate steel block that's machined to fit the barrel and frame and rests in place in the frame. When I stripped the PX4, the barrel feels snug in its front-to-back fit in the slide, but I could feel a small amount of movement in the barrel-to-slide muzzle fit. When I rotate the barrel to its unlocked position, the muzzle fit gets a bit looser.

As a gunsmith, I can't help but wonder what re-fitting the muzzle end of the barrel would do for accuracy. An hour with a lathe to fit an oversize sleeve and then custom-fitting it to the slide could turn this or any PX4 Storm into a real tack-driver. Or not. Only experimenting will tell.

For all of my grumbling about accuracy, I was able to slam down the falling plates as fast with the .45 ACP Storm as I was with the 9mm or .40 Storms. Part of that was the certainty of the .45 in action: hit a plate, and you know it is going over. An edge hit with a .45 does as well as a center hit. With the .40, not so much, and with the 9mm you really have to follow-through to be certain of a center hit and a falling plate. Which is the same kind of attraction in defensive uses that the .45 has: It works better.