Knowledge of the primordial lithium abundance (Li(p)) helps to constra
in cosmological parameters and to discriminate between models of primo
rdial nucleosynthesis; the best current probe of Li, is the set of lit
hium abundances measured in old halo stars, provided that stellar Li d
epletion mechanisms are taken into account. Standard stellar evolution
ary models show little depletion (Deliyannis, Demarque, & Kawaler), co
nsistent with standard big bang nucleosynthesis (BBN), whereas models
with rotational mixing deplete about an order of magnitude (Pinsonneau
lt, Deliyannis, & Demarque), possibly consistent with inhomogeneous BB
N, or other nonstandard scenarios. One way to test the rotational mode
ls is through their prediction that a small dispersion can result in t
he lithium abundances if there is a significant dispersion in the init
ial angular momenta. For all available extreme halo dwarf data, we hav
e performed a detailed analysis in the (b - y) color-lithium equivalen
t width plane, which eliminates many of the uncertainties associated w
ith the transformation to the T(eff)-abundance plane. Furthermore, we
have used two independent means to compute the uncertainties: one that
incorporates detailed information about the individual S/N and releva
nt instrumentation, and one that utilizes empirically derived uncertai
nties from multiple observations; our conclusion is the same in either
case. Using chi2 tests, we find evidence for a small dispersion (of o
rder +/-20% about the mean) in the Li equivalent width at fixed color,
with a high degree of confidence. Interpreted according to usual stan
dard model atmospheres theory, this translates directly into a small d
ispersion in the Li abundances themselves. This dispersion is consiste
nt with, but does not uniquely support, the rotational depletion. We h
ave also corrected the colors for reddening, and the evidence for disp
ersion remains strong. One star, BD 23-degrees-3912, lies more than 50
% above the lithium plateau.