On Stirling's Numbers