Stefan Michalski's contribution to a continuing discussion on the DistList,
I try to stay out of this flashbulb thing but it has dogged me throughout years of lighting seminars, and I want it to go away. Besides, right now isn't a good time for our profession to go to the wall with yet another piece of revered wisdom that turns out to be built on quicksand.
There have been reliable articles on an identical problem, the effects of flash tubes in copying machines, that have already gone through this issue, and laid it to rest (such as Jim Hanlan's about twenty years ago).
The underlying principle for any such calculation is reciprocity, the notion that light effects are cumulative in a simple additive manner. Reciprocity has been found to hold with reasonable accuracy in several industrial studies of textile light-fading, in at least one copying machine flash tube study, and most recently in a paper I just accepted for the upcoming ICOM-CC conference, by Saunders and Kirby of the National Gallery London ("Light induced damage: Investigating the reciprocity principle," ICOM-CC 1996 Preprints, p. 87-90). It is far and away the best study to date, on several historically important colorants, and finds reciprocity holds very nicely. While it is true, as someone noted, that some chemically measured effects linger in the dark (e.g. free radicals have been shown to take many hours to decay in photodegradation studies of wool) there is no evidence that this is a significant correction to estimates based on simple reciprocity. (Ezrati has a paper in the same conference preprints, showing intermittency has no significant effect on dye fading: "L'usage de la lumière naturelle en muséographie," p. 36-40.) So, the rest is arithmetic, and units.
An electronic flash on a camera is typically sized to use f8 for a film of 100ASA at a subject distance of 3m (10ft). From photo handbook data, this is equivalent to a light dose at the artifact of about 30 lux seconds (lx-s). An alternate route to this estimate is via the Illumination Engineering Handbook data: xenon flash tubes for photography range from 10 to 200 joules rating. Given efficacy of about 50 lumens per joule, a wide-angle reflector throwing the light forward into about 1/4 of the sphere, this gives a range of 20 lx-s for a little built-in flash to 400 lx-s for big fat studio tubes.
For convenience, round up to 50 lx-s for each amateur. Assuming the gallery lighting is the lowest most museums can tolerate, 50 lux (5 foot candles), then each flash adds the equivalent of one second of normal gallery exposure. So, 300 amateur flashes a day is equivalent to adding five minutes to the display day. In order to actually increase damage by 10% on a ten-hour day, one would need to experience 3600 flashes per day. Two large professional flashes would raise the ante a little: they would need 225 flashes a day to add 10%. For museums at 150 lux (15 footcandles) these numbers become 10,000 amateurs, or 700 pros, every day. To actually double fading would need 100,000 amateurs a day. Most museums would kill for those attendance figures! As for the UV wrinkle, xenon is used because it has a spectrum very close to daylight (6,000K). Given typical glass tubes and plastic diffusers, the UV ratio will be a little higher than properly filtered light, but UV type damage is far from the Achilles heel of artifacts at controlled light levels; it is color fading, and UV is not the issue here.
In other words, flash may very well be banned for reasons of copyright, or as a disturbance to the act of contemplation (my personal vote) but there is no preservation reason. I think the ban started originally because flash bulbs (and their precursors, the open magnesium flash) were a genuine fire hazard, and an explosion hazard (hot fragments) and a garbage problem. Of course, tripods, hot studio lamps, and bulky equipment are still hazards, and a photography policy is still necessary, but please don't wave the red flag of conservation over flashcameras.