So no simple twofold rule will work. Of course this seems a terrible waste, using a 9 watt resistor where a 2 or even a 1 watt resistor would probably work just fine. But designing for the worst case is just that; it is not economical or space preserving.
    Capacitors often bear two voltage ratings on their bodies: a DC rating and a lower AC rating. A sinusoidal wave peaks at 1.4 times its equivalent DC voltage. Consequently, if the capacitor only bears one voltage rating, assume that it refers to the DC rating and de-rate the capacitor when using in an AC application. Here is an extreme example, fearing high frequency noise, an audiophile places a capacitor across the outside power transformer leads. The power supply yields 400 volts, so a 400 volt capacitor is soldered in place. But as the power supply comprises a full-wave center-tapped diode-transformer arrangement,  the capacitor sees 800 volts peak, not 400 volts. Even worse, if a large choke connects directly to the diodes (choke loading), then the peak voltage across the leads is 1130 volts, not 400 volts.
    When a capacitor couples one stage to another, two voltages develop across its leads.  When the amplifier runs idle, the static DC voltage difference between stages must be tolerated; but when the amplifier is first turned on, the tubes cold, the full B+ voltage must be withstood.

Safety Issue Three:
What happens when a component only partially fails?
    The previous example was straightforward, but here is an example that is not. When resistors fail, they usually burn up and become open circuits. Not always, though. Bulk foil resistors are made by depositing a metal film on a glass substrate and then laser trimming away at the foil to bring about the right resistance value. Close examination reveals that these resistors often are made up of two parallel resistors: one large resistor at a slightly higher value than it's

nominal value and one small resistor at some much higher value. It is this second resistor that is trimmed away until the desired total parallel value achieved.
    For example, a 2k resistor might contain a 3k and a 6k resistor in parallel. Now for the what if. What happens when only one of the internal resistor burns up and opens? The only possible answer is that the resistor's effective value increases. Usually, this does not harm any series components. But what if the failing resistor is part of a high voltage divider whose output feeds a voltage sensitive device. If the nominally 2k resistor becomes a 6k resistor, the voltage division skyrockets. (I have seen this very failure take out a solid-state module on a Loftin-White amplifier. The cathode voltage climbed from 150 to 400 volts.) 

Safety Issue Four:
What happens when a fuse blows?
   When a fuse blows? Isn't that like asking, what happens when the seatbelt restrains? Yes, it is and both are good questions. The fuse we are concerned with is not the one in series with primary winding of the power transformer, but the one within the DC portion of the power supply. With a uni-polar power supply, the fuse's blowing only saves the amplifier...well, usually that's the case. What if a solid-state high voltage regulator is used? This regulator may work perfectly when the amplifier is turned off, but fail when the fuse blows. The problem lies in the disappearance of its normal voltage gradient and discharge path when the fuse blows. Normally, the voltage regulator only sees the voltage difference between its output and input voltage. But the  fuse blows, the difference increases and reverses polarity. The large charge held in the regulator's output capacitors must now discharge through the regulator's reverse biased diodes and transistors.
      A fuse blowing can even damage high voltage solid-state diodes. The power transformer secondary winding's voltage