Of course, stripped lowers need a few extra parts before they'd be ready for use with an upper - so another $100, at a bare minimum, would be spent on a lower parts kit and a stock assembly.

Several months ago, however, a very affordable complete AR-15 lower became available - one that can currently be found for about $120 online.

"There has to be a catch," you say. Well, yes, there is. It's a composite lower.

Called the PlumCrazy C15 (but referred to by most as the "plum crazy lower"), it weighs 1lb, 11 ounces with an M4 type stock that weighs 7.5 ounces. For comparison purposes, an assembled aluminum lower with a lower parts kit and receiver extension tube installed - using a carbine buffer and spring - weighs 1lb, 11 ounces. In other words, the C15 assembly is about 7.5 ounces lighter than a standard lower assembly.

As if having the lower made from a composite material wasn't enough, the fire control group is also polymer. The receiver has also been thickened - slightly - in certain areas.

Certain components, however, remain standard metal fare - the bolt catch, for example, and the castle nut.

The receiver extension tube, too, is not polymer. It appears to be a "regular" commercial spec, 6 position, slant back tube - and it was indexed properly, which impressed me. It's worth noting that the slant back tubes can be an ounce or two heavier than mil spec tubes - and Colt/BCM type M4 stocks are also about an ounce lighter than the C15 stock - so if weight reduction is the ultimate goal, ditch the stock and RET for lighter milspec components.

The supplied buffer was anodized black (I would hazard a guess that the anodizing was not Type III, due to some wear patterns - but I could be wrong) and weighed 2.8 ounces.

The receiver endplate is unstaked, but this is probably not a huge deal to the average purchaser of the C15.

The magwell features a nice flare, and the trigger guard is molded as part of the receiver. It's straight - not bowed like the "enhanced" trigger guards out there. The finishing work - removal of flashing and so on, which is done by hand - appeared, to me, to be well done.

Fit with a variety of upper receivers was extremely tight, even after repeated "mating" cycles. In addition, the pivot pin is flush with the surface of the receiver, requiring more than just finger pressure to pop out. I found this to be fairly annoying.

In actual use, I found the trigger to be quite nice, with negligible creep and a fairly light pull weight. I attempted to use it with the Spike's 5.45 upper - polymer components would be ideal for low-maintenance use with corrosive ammunition - but after doubling on the first trigger pull, the hammer spring wasn't powerful enough to reliably fire the hard primers of the surplus ammunition.

With a 5.56 upper, I experienced no further issues with the fire control group, and a cursory inspection left me puzzled as to the exact cause of the double.

I then used the C15 lower with a .22 conversion, and believe that I found its true calling. While most of my shooting revolves around the defensive use of semi-automatic firearms, an afternoon with a brick of cheap .22 is undeniably fun. Using the C15 in conjunction with a light upper - say, a 16" A1 profile barrel with a Troy TRX 9" handguard - would allow an for an exceptionally light firearm that I would consider to be ideal for introducing new shooters to the AR15 platform, especially in .22LR guise.  With a heavier upper, however, the removal of 8 ounces resulted in a forward CG shift that I didn't particularly like. Of course, those seeking a lightweight AR are not likely to stop their weight removal efforts at the lower, so this is probably not a huge issue.

Although "torture testing" the lower would seem to be right up my alley, the lower was a loaner, and I had no intention of breaking it. I may, in the future, purchase one with the intentions of finding its breaking point, but spending $120 on something that I intend to break isn't exactly an attractive option at this point, when I could spend the same money on items that would provide more "testing value."