In this paper the author considers a more realistic approach to the periodic Anderson model by including spin-orbit and crystal field effects using the slave-boson technique. Looking firstly at the degenerate case with appropriate spin orbit coupled bands, he shows that while for T=0 he obtains the same results as the 1/N approach there are differences when he extends to finite T or when he goes beyond the mean-field theory. He considers also crystal field splitting within the 1/N approach and show that this gives an anisotropic susceptibility in qualitative agreement with experiment. The inclusion of spin-orbit coupling again does not affect the results at T=0.