model				dQW/QW(had.)	dQW/QW(total)
				2par	2par	3par	3par
				all	q2<.25	all	q2<.25
				(%)	(%)	(%)	(%)

super simple			0.7	1.4	3.2	3.8
GEs=rhos*tau				(1.3)		(3.8)
GMs=mus

new simple			2.3	3.2	5.1	6.2
GEs=rhos*tau				(1.3)		(3.8)
GMs=mus*G_D

linear
GEs=rhos*tau*G_D		1.4	2.0	4.0	4.7
GMS=mus*G_D				(1.9)		(4.7)

Galster
GEs=rhos*tau/(1+5.6*tau)*G_D	3.1	3.5	6.2	6.5
GMS=mus*G_D				(2.4)		(5.5)

nasty Galster			4.5	5.9	9.0	11
GEs=rhos*tau/(1+18.3*tau)*G_D		(3.1)		(7.3)
GMS=mus*G_D

nice Galster			2.3	2.4
GEs=rhos*tau/(1+1.8*tau)*G_D	(1.6)	(2.1)
GMs=mus*G_D

Chi-PT							7.5
GEs=rhos*tau				(3.8)		(6.9)
GMs=mus+rhos_m*tau					note: 4par

Lattice				2.4	3.4
GEs=rhos*tau			(0.6)	(1.3)
GMs=mus/(1+q2/0.54**2)

numbers in brackets include the constraint that mus = 0 exactly
(i.e. models the effect of input from Leinweber et al).

G_D is the dipole form 1/(1+q2/.711)**2

The former "lattice" model of Dong et al would give very similar
results to the "linear" form used above.