Gotta admit this is a new one on me....interesting though
 
https://www.premierguitar...ticles/Light_Bulb_Gear

Interesting article, would have been cool if he had thought about going into fluorescent bulbs or LED bulbs as well - and if those wouldn't work to explain why. Since LEDs are used in some distortion circuits might have been interesting.
 
I thought the Morley stuff was just tangential filler. I think it would have been more relevant to have gone into optical gear like the LA2A.  
 
 

The lamp he has here is the humble tungsten filament.  The higher the rated wattage of the lamp, the lower the resistance (so the filament carries more current and burns brighter). Fluorescents and LEDs are completely different. Fluorescents are gas discharge and need a striking voltage (starter) to initiate the illumination. This is unlikely to happen with speaker signals. LEDs are one-way devices - they won't support AC signals, like the output of the amp.  Diodes (non-light-emitting) are used in effects units because they allow the signal to drive to full conduction in one direction of current flow, and thereby saturate (peak) the signal on that direction of the AC.  In FX units there's likely be another diode the opposite way round to allow the other half of the signal to saturate .. and so produce the desired distortion.
 
The experiments here are interesting.  Like the warning box says - if you're using amps with valve output stages with transformer coupling, beware the increased impedance introduced by the lamp.  This could blow that expensive transformer. Transistor outputs won't be affected.
 
He mentions supply lag*.  This is where the demands of the amp circuit require more power as the signal levels increase. With a basic non-regulated supply this will cause the power supply unit to drop a few volts, causing clipping to occur earlier than otherwise expected. So he may want to repeat the experiments by putting the lamp in the power supply from the PSU to the main amp. I haven't thought the details of that through yet, but it may be worth a try.
 
jdf
 
*Edit - supply sag - not lag 

after you get done with lightbulbs try a toaster element!

jerrydf
The lamp he has here is the humble tungsten filament.  The higher the rated wattage of the lamp, the lower the resistance (so the filament carries more current and burns brighter). Fluorescents and LEDs are completely different. Fluorescents are gas discharge and need a striking voltage (starter) to initiate the illumination. This is unlikely to happen with speaker signals. LEDs are one-way devices - they won't support AC signals, like the output of the amp.  Diodes (non-light-emitting) are used in effects units because they allow the signal to drive to full conduction in one direction of current flow, and thereby saturate (peak) the signal on that direction of the AC.  In FX units there's likely be another diode the opposite way round to allow the other half of the signal to saturate .. and so produce the desired distortion.
 
The experiments here are interesting.  Like the warning box says - if you're using amps with valve output stages with transformer coupling, beware the increased impedance introduced by the lamp.  This could blow that expensive transformer. Transistor outputs won't be affected.
 
He mentions supply lag*.  This is where the demands of the amp circuit require more power as the signal levels increase. With a basic non-regulated supply this will cause the power supply unit to drop a few volts, causing clipping to occur earlier than otherwise expected. So he may want to repeat the experiments by putting the lamp in the power supply from the PSU to the main amp. I haven't thought the details of that through yet, but it may be worth a try.
 
jdf
 
*Edit - supply sag - not lag 

Excellent explanation. Well done.