I’ve done my homework! I have compared 4 types of methacrylates and 2 leading epoxies on nearly 180 test coupons. These coupons were prepared in a scientifically controlled manner with detailed notes taken. The coupons were all subjected to accelerated heat aging at 150 deg F and thermal cycling from 20 deg. F. to 150 deg F. for the purpose of arriving somewhere at the bottom of the reliability bathtub failure rate curve. The testing was divided up into 180 degree peel, shock vibration, and lap shear. This work has consumed perhaps 80 hours of planning and execution over the course of the past 5 months. I have come to some definite conclusions. Preparation: Must wipe original finish with acetone or MEK first to achieve a clean surface, then 100% abrade both mating surfaces using a high speed sanding disk or flapper wheel of approximately 100 grit. Smooth or even smooth etched surfaces consistently failed against the abraded surfaces. I personally wouldn't trust a Scotched Brite surface either. Freshly abraded surfaces out performed abraded surfaces that had subsequently been phosphoric acid etched and vigorously rinsed with distilled water and a brush then dryed. The adhesives materials just seem to like sticking to fresh virgin aluminum. The etch and rinse process seems to only have acted as a contaminant. Application: To achieve an even layer of adhesive, trowel the material on using a 24 tpi hacksaw blade as the trowel. Use a coarser blade to apply more material for gap filling, or a finer blade only for very precise fits. Apply the adhesive to both surfaces before mating. To prevent excessive squeeze out resulting in spots where there is no more adhesive, two methods can be used. Trowel adhesive to one surface and let it cure prior to a second application to the other surface, then mate. This only works with adhesives that are viscous enough to hold the plowed field texture until setup. Proseal and some methacrylates perform well in this way. Epoxy just flows out smooth. Use some form of micro spacer such as micro balloons or nylon tule netting fabric, which is about a .009” thick, just right for an adhesive bond joint. I used the tule. Be aware that too much micro balloons will displace the adhesive, thereby reducing the strength. The balloons of course have no strength of their own. Adhesives: Most epoxies and some methacrylates are brittle and are therefore intolerant and fail in the presence of vibration and parent material deformation under load. The thing to look for here is the material’s elongation percentage. Whereas 7% is brittle, which is bad, and 50% is very tough, which is good. An X-acto knife taken to 7% material will make chips to the surface; 50% material can be sliced with moderate force. Methacrylates do tend to bond more readily to aluminum than epoxies. Methacrylates do stick pretty good to smooth aluminum, but it sticks even better to the rough texture better, despite what the salesmen tell you. I know, because I've seen the difference with my own eyes. Failure Modes: There are two types, Interfacial which is bad, and Cohesive which is good. Interfacial is where the adhesive debonds from the aluminum. Cohesive failures are where the fracture passes through the adhesive itself. Interfacial failures always occur at lower load levels than cohesive failures. Interfacial failures are a result of poor surface preparation, vibration, or deformation of the substrate. If everything is done reasonably well, the cohesive strength of epoxies and methacrylates far out perform standard AN –3 and –4 rivet spacing schedules. It's the interfacial strength that is the Achilles heel of the brittle epoxies and some methacrylates. My standard test coupons bond joints were .75” x 1.5”. When adhesive joints are made per the best practices and materials of that described here, I have been able to shear groups of three AN –4 rivets consistently when using .065” material; and groups of four –3, and groups of two –4 rivets all of which tear through .023” 2024 T3. My most recent bonds remain unbroken, as their strength is coming to the limits of my rather ad hoc pull test set up. Can I break a group of four –4 rivets? I don’t really need to. If I can consistently break two at a time, then I have a quality process. It doesn’t matter how strong the maximum strength of your joints can be. What matters is, is how reliable the minimum strength is. Incidentally, I found that on thin sheet metal, the rivets would tear through the sheet, and the thicker sheet would shear the rivets; all the while, the tougher adhesive joints held. When joints of dissimilar material thickness were pulled apart in peel, the interfacial failures always occurred on the thinner material, the side that deformed. The top performing adhesive I used was Partite 3750 from Parson Adhesives, followed by Extreme 310 and Extreme 5375HS. The brittler stuff was Partite 5140 (which was surprising because they advertised it as being rubberized), and the two epoxies, PTMW ES6228 and Solar EA-0504 So there you have it, the fruit of my labor over the past 5 months. I am confident that I will be able to skin by plane using Partite 3750 or something similar to it. At this point I consider myself somewhat of a self-made expert on this subject. I am however soliciting further light and knowledge, and any gotchya’s that may be awaiting anyone considering adhesives in lieu of rivets for secondary structural attachment.