Over the course of your life time, the average person eats X spider/insects/whatever whilst they are sleeping.
X tends to vary according to who you ask. Is this just an urban legend?
This is an utter fabrication.
It is thought to have started with an article in PC Professional Magazine regarding ridiculous facts circulating via e-mail.
In a 1993 PC Professional article, columnist Lisa Holst wrote about the ubiquitous lists of "facts" that were circulating via e-mail and how readily they were accepted as truthful by gullible recipients. To demonstrate her point, Holst offered her own made-up list of equally ridiculous "facts," among which was the statistic cited above about the average person's swallowing eight spiders per year, which she took from a collection of common misbeliefs printed in a 1954 book on insect folklore. In a delicious irony, Holst's propagation of this false "fact" has spurred it into becoming one of the most widely-circulated bits of misinformation to be found on the Internet.
However, the Snopes article itself has been called into doubt, with claims that Snopes made up the reference.
For a sleeping person to swallow even one live spider would involve so many highly unlikely circumstances that for practical purposes we can rule out the possibility. No such case is on formal record anywhere in scientific or medical literature.
[...] I remain unconvinced that a spider would visit a huge breathing monster and enter its mouth.
Unless a spider is so small that it wouldn't realize that the "hole" is the mouth of a large predator, it's probably unlikely that a spider would crawl into it.
The Straight Dope claims that one person may accidentally swallow a large number of spiders after an egg sac bursts:
Put it all together, and it would be a miracle for a spider to end up in anyone's mouth while they're sleeping, except for one rare circumstance--when a spider egg sac hatches indoors. At that point, you can have hundreds of microscopic spiders, a millimeter long or less, leaping into the air in a short time span (under an hour total) and trying to ride the air currents to freedom.
However, these events should be rare and the author argues that average number of spiders swallowed being so high seems doubtful:
After all, most people breathe while they sleep (at least I do) and spiders, like virtually all arthropods, flee from breath. After all, there are lots of vertebrates that EAT arthropods, and if you're an arthropod and something is breathing on you, it's not a good idea to stick around. Simple enough.
So, it seems unlikely, but the arguments given in this particular article aren't conclusive.
The 1992 book Basic Statistics for the Behavioral Sciences Student Workbook and Study Guide says:
b.This is a very low p, indicating that, over the long run, we can expect to swallow spiders only 5 out of every 10,000 times we sleep.
c. Since p is not zero, we do expect to swallow spiders sometimes, and we don't know when it will happen.
This appears to be just a hypothetical statistics problem, but it could have been easily misunderstood as actually indicating that people swallow a certain number of spiders.