Two recent blog posts by Carol Beuchat at http://www.instituteofcaninebiology.org/blog/ entitled "Vulnerable breeds: how small is too small? and "The trouble with Terriers" are worth a read. (Her April Fools post is also good.) In general I agree with her analysis and conclusions.
To double-check her calculations, I used version 1.94 of my Norwich pedigree database. For dogs in my database born in a particular year, I counted the number of distinct sires (Nm) and the number of distinct dams (Nf), and calculated the effective population size (Ne). The results since 1990 are as follows:
So the numbers are roughly in line with what Carol Beuchat claims for Norwich.
However I have three reservations about this calculation.
1. The American Kennel Club (AKC) only lists dogs in its studbooks when they *first* sire or whelp a litter. (Every other registry of which I am aware lists all dogs registered.) Since the vast majority of information in my database comes from studbooks, my database does not list all AKC registered Norwich. Thus the number of sires and dams for any given year is an underestimate.
2. My database is only complete for Norwich from the UK, Canada, Germany, and Switzerland. There are several other European countries that I would like to add that would make the database more complete. So again the number of sires and dams for any given year is an underestimate.
3. While the definition of effective population size is easy to understand, the derivation of the formula of effective population size is not easy. So most people must take the formula on faith. I dislike such formulas.
In spite of my reservations, I believe her conclusions are correct - Norwich are a small gene pool ... perhaps too small to survive without some infusion of new genes.
For an interesting contrary viewpoint, see http://rufflyspeaking.net/is-opening-the-akc-stud-book-a-good-idea/