A second comment on Sheldrake and his experiments. If psychic phenomena exist, and they can range from knowing when someone's calling you to dogs knowing when their owner is coming home, there would be lots of other, better ways to test for their existence. If one were to be a serious scientist about it, one would carefully design an experimental paradigm that tested for these phenomena in an experimentally favorable way, by which I mean a way that would allow one to collect large amounts of data in short periods of time (thus allowing useful statistics to be obtained), with few additional variables that are difficult to control for. It wouldn't resemble events that happen in normal life.

The fact that Sheldrake uses these experimentally unfavorable paradigms that closely match personal experience reflects the fact that he isn't being scientifically serious about testing these phenomena, but instead trying to convince nonscientific laypeople of the existence of them. It's simple to explain that dogs know when their owner is coming home because they have really good hearing, and it's also simple to explain the knowing-when-someone-is-calling-you thing by a recall bias (i.e. you remember when it happens, but you don't remember when it doesn't because there's no event to remember), but these phenomena that we've all experienced are superficially somewhat mysterious. Therefore, he claims that they are based on psychic phenomena, which resonates with an untrained person who hasn't thought deeply about these issues. However, he stops short of taking it to the next level and testing these things rigorously. I don't think this is an accident on his part.

My take on this as a working scientist is that the correct way to view these types of experiments is that dogs receive a large number of cues that their owner is coming home. For example, there are visual cues if the dog can see out the window, auditory cues if the dog can hear the car coming, or even olfactory cues if the dog can smell the person coming. There are also temporal cues (e.g. same time every day). Sheldrake believes that there are telepathic cues as well. The way that he tries to demonstrate this is by attempting to block ALL of the other cues the animal could be using. If the animal's performance is statistically above chance, he concludes there is another cue that the animal must be using, and that cue must be telepathic. However, this type of experimental paradigm is extremely tricky because there is no way to verify that he was successfully able to eliminate all of the non-telepathic cues. In fact, doing that is notoriously difficult in any behavioral experiments because animals are incredibly good at reading the most subtle cues. Any experienced behavioral experimentalist looking at this would say that it's most likely that Sheldrake was simply unable to isolate all of the other cues.

The way that he could do this experiment properly would be if there were a way to provide "psychic" interference that would remove the telepathic cue as well, thus bringing the animal's performance all the way down to chance. For example, if "psychic rays" can't penetrate lead, or if he could have some "psychic" stand in a distant place and provide mental interference to the dog. If the dog's performance dropped to chance, this would demonstrate that all other cues had been eliminated, and it really is the telepathic cue explaining the dog's performance. However, in the absence of this, any experiment of this type is not very convincing.

Finally, just to give an idea of how unsophisticated these experiments are relative to dog behavioral experiments even in the late 1800s, Pavlov's famous bell-food conditioning experiment involved constructing a concrete cylinder several stories high with the dog suspended from springs inside a soundproofed box, with a cannula surgically implanted inside the salivary gland and hooked up to a pressure transducer outside. There was also an elaborate system for transmitting the sound of the bell, food odor, etc. into the experimental chamber from outside.

It's like #9 said, there's a tradeoff between filament temperature and bulb life. A filament is a black body, meaning that it emits higher and higher energy photons as it gets hotter. Low temperature filaments emit most of their photons as invisible infrared, while higher temperature filaments emit more photons in the visible range. However, those bulbs don't last as long. Modern bulbs represent a tradeoff between efficiency and bulb life. So it's not really that big of an achievement to make a really inefficient bulb that lasts a really long time.

Halogen bulbs are better because there's a temperature-dependent chemical reaction inside the bulb that redistributes the metal atoms to the thin parts of the filament, allowing the bulbs to burn at higher temperatures and thus higher efficiencies. However, they require a UV shield because they're hot enough to emit UV photons which are higher energy than visible ones.

@bookyloo, #36:

In med school, one of my psychiatry attendings told us that there is an important distinction between two general subclasses of delusional patients: those who respond to their delusions in a rational, normal way, and those who don't. The ones that do bizarre things are much more common in an inpatient psych unit, and they're obviously likely to have more severe underlying pathology (e.g. delusions as a symptom of schizophrenia). Just because your relative's patient was not exhibiting irrational behavior aside from that belief does not mean that the belief was actually true. They could have had delusional disorder:

http://en.wikipedia.org/wiki/Delusional_disorder

The thing is, with persecutory delusions, the probability of "rogue CIA agents" being after someone is very, very low. If that person is convinced his wife is cheating on him, which is a common delusion, that's a situation where it's a lot more difficult to know. A common type of persecutory delusion I've seen in the elderly population in NYC is paranoia that the landlord is trying to drive the person out of a rent-stabilized apartment. Again, that's a situation where it's hard to be sure what's true and what's not. One thing I'd say is that if I were being chased by CIA agents, the last place I'd go is an inpatient psych unit. You're locked down and easy to find.

Finally, the classic definition of a delusion is not just a false belief, but a "fixed, false belief." In other words, the belief has to have a character to it that resists evidence to the contrary, or at least the strength of the belief has to be somehow disproportional to the evidence in favor of it.