In order to test the reliability of determinations of the shapes of dark-ma
tter haloes of the galaxies, we have made such measurements for the Milky W
ay by two independent methods. First, we have combined the measurements of
the overall mass distribution of the Milky Way derived from its rotation cu
rve and the measurements of the amount of dark matter in the solar neighbou
rhood obtained from stellar kinematics to determine the flattening of the d
ark halo. Secondly, we have used the established technique based on the var
iation in thickness of the Milky Way's H I layer with radius: by assuming t
hat the H I gas is in hydrostatic equilibrium in the gravitational potentia
l of a galaxy, one can use the observed flaring of the gas layer to determi
ne the shape of the dark halo.
These techniques are found to produce a consistent estimate for the flatten
ing of the dark-matter halo, with a shortest-to-longest axis ratio of q sim
ilar to 0.8, but only if one adopts somewhat non-standard values for the di
stance to the Galactic centre, R-0, and the local Galactic rotation speed,
Theta(0). For consistency, one requires values of R-0 less than or similar
to 7.6 kpc and Theta(0) less than or similar to 190 km s(-1). The results d
epend on the Galactic constants because the adopted values affect both dist
ance measurements within the Milky Way and the shape of the rotation curve,
which, in turn, alter the inferred halo shape. Although differing signific
antly from the current IAU-sanctioned values, these upper limits are consis
tent with all existing observational constraints.
If future measurements confirm these lower values for the Galactic constant
s, then the validity of the gas-layer-flaring method will be confirmed. Fur
ther, dark-matter candidates such as cold molecular gas and massive decayin
g neutrinos, which predict very flat dark haloes with q less than or simila
r to 0.2, will be ruled out. Conversely, if the Galactic constants were fou
nd to be close to the more conventional values, then there would have to be
some systematic error in the methods for measuring dark halo shapes, so th
e existing modelling techniques would have to be viewed with some scepticis
m.