I have a 10 gallon tank which seems to exist only to aggravate me. It has a pair of GE 20 watt 6500K screw-in CFL bulbs, each with a pretty good DIY reflector. If I look up at the bulb I can see almost every surface of the coil or its reflection, so I think I am getting a substantial part of the bulbs light down into the tank. But, my results show beyond a doubt that this is low light intensity.

Those screw-in bulbs are a blob source, close to a spherical source, no where near a linear source. So, their intensity will drop off proportional to the inverse square rule. My bulbs are about one inch above the water line, and 10 inches from the substrate, where glosso is struggling to grow. The inverse square rule says I only have about 1% of the intensity at the substrate as at the surface!

If I had a T5 bulb, even a 20 watt one, I would get at least as much light at the water surface and would only lose light at the substrate proportional to the inverse of the distance (not squared), so I would have 10% of the light at the substrate as at the surface, ten times as much light as with the CFL bulbs.

Now I think this is the reason for the ineffectiveness of screw-in bulbs. It isn't restrike that kills them, it is the inverse square effect.

If I can find the parts without spending too much I will modify that Perfecto hood to use a linear bulb, just to test out this idea.