Incidentally, two beehive boxes and a Miller-type hive top feeder toppled down a hillside last week. The boxes are about 5 years old, all stapled profusely, glued with Titebond III, treated with boiled linseed oil, and painted with exterior latex . The Miller-type feeder box built with butt-joints blew up entirely. A rabbeted-joint deep brood box tore out at the staples on one corner, but held at the other (rendered useless, unless I can build a corner and fit that in to the damaged box while not splitting the corner that held...?). The box on top that fell the farthest and hit by far the hardest was a finger-box joint deep brood box. Only the 3/8" rabbet where the frames rest broke out on one side, the joints all look like new (the repair will be easy).
Not a scientific test, just an accident, but somewhat telling about the strengths of the respective joints. I will be building with the finger box joints from now on! Had the finger-box-jointed box been made with that new arrangement that Brushy Mountain came up with, I may have had NO damage to that box!
I was just glad my suit was on and bee-tight! Those poor bees were NOT TOO HAPPY with that off-road ride...
One other aspect that could be affecting discussions on comparative joint strength is the wood grain at the joints. Try to avoid knots and wavy grain in the area of the joints - any joint will last longer if the grain is straight and strong around the joints, and I notice that glues all work better when the adjoining grains are about the same size - that is, all tight grain or all large grain, as opposed to large grain into small grain on the mating workpiece (I suspect that the reason is probably different shrinkage rates due to grain size). In fact, I notice that medium to large grains seem to glue better than tight grains (my suspicion here is that tight grains seem more prone to twist, warp, or cup than do medium and large grain woods - but that may well vary with different species), but this is years of subjective observation from building aircraft with spruce and cabinetry from hardwoods, not backed up by testing FBJ's made of sugar pine or Bald Cypress subjected to weather nor scientific study.
Three things are certain - tight joints, spreading glue on BOTH SURFACES of all joining surfaces, and strong clamping all make stronger joints. Two possible exceptions: 1) if a joint is made, clamped, then disassembled while the glue is still wet, then re-assembled without re-gluing, it may or may not be strong => therefore wipe clean, re-glue and re-spread the glue before re-clamping; 2) if a very wet sponge is used to clean up a small, thin glue joint, it may wash out the glue, rendering the joint weak => therefore squeeze out the excess water from the sponge before wiping off excess glue from a clamped joint.
One more thing is becoming clear - end grain cut oversize and sanded smooth and well-painted stays stronger after 5 years in the weather than does rough-sawn end grain given the same paint and exposed to the same weather. It seems that rough-sawn end grain gets many more splits and checks than sanded end grain. Sanded end grain seems to stay married to the paint better, probably deterring water from causing damage, but again, this is experience from my limited observation on sugar pine belt sanded to 120 grit, not proven through rigid study, unless someone knows of any such studies?...
I suspect that discussion regarding comparative strengths of joints is a case of apples + oranges + opinions = no real data until more thorough studies of destructive testing are made, so for now I'm going with my own experience. So far my experience has a strong bias for the finger box joints accurately cut, well-glued, firmly clamped, and made from straight, medium-sized grained, riff-sawn wood.
Hope this helps...