There are two schools on this, and I alternate between them. In this case, it's a piece of test equipment that I use almost daily and it's frequently being power-cycled on and off. (The main reason I don't leave it on all day is fear that the selenium could blow while I'm out shopping or something, and fill the house with noxious fumes.) Put it this way, replacing the rectifier cannot do any harm -- if I do it properly -- and removes a fault source, however minor, in future decades. I'll also, eventually, add the fuse as suggested by N2EY. I'm 57 years old and fully intend to keep using my vintage test equipment for the rest of my life! That's what I thought too, it would be very low in normal usage. For some reason I don't have anything bigger than 1/2W in those low values; that would probably be OK at a pinch but might as well super-size it to 2W. After all it's a piece of test equipment, you want your resistor to run cool so that the B+ stays steady. Yes of course. Call me strange, but resistors are cheap and I usually prefer to use one rather than several. I often use parallel/series combinations while experimenting; then, when the value is determined, I buy a single resistor of the right value. No logic behind this at all, I agree.