98.12.06 BTD:
	in rotate.f, it is *assumed* that phi0=0 when theta=0
	shouldn't phi0 be determined by orientation of a2?
	(though admittedly it is then necessary to make assumption
	about beta)?  should look into this...