> I answered this at some length a few days ago. The tradition > before this century (dating back to Euclid's Elements) was indeed > to count 1 as a prime, and even Lehmer's 1914 list of prime numbers to > 10 million does so. But the many inconveniences it causes have > led people this century to put it into a new category, and call > it a "unit" rather than a prime. > > John Conway >
Some of this presisted into this century. N. Lehmer (father of D. Lehmer) has published a count of primes up to some very large x, sometime in the early 20's. When people checked his results on a machine, they were off by one. It took a while to realize that he counted 1 as a prime.
In connection with this, there is a silly anecdote about late E.G. Strauss. People (non-mathematicians) would perpetually ask him for an opinion of why is 1 not a prime. One day he replied: Look, the primes are the buiding blocks of inetgers. But is is plain that 1 is not a brick.