Question: fsolve stopping criteria

What are the stopping criteria for fsolve?
I cannot find anything in the help page and there seems to be no way of adding an optional argument to fsolve about errors.

I was initially surprised by the results of the first two fsolve commands below:

fsolve(x->1,0.4); #OK, returns unevaluated
fsolve([1,1],{x,y}); #OK, returns NULL

I assume that in the first two examples the criterion used is that at some point in the process the iterates [x(n+1),y(n+1)] and [x(n),y(n)] are close enough together and the difference between results from the two is small enough (clearly 0).

Please Wait...