We examine phantom dark energy models derived from a scalar field with a negative kinetic term for which V(phi) approaches infinity asymptotically. All such models can be divided into three classes, corresponding to an equation of state parameter w with asymptotic behavior w -> -1, w -> w_0 < -1, and w -> infinity. We derive the conditions on the potential V(phi) which lead to each of these three types of behavior. For models with w -> -1, we derive the conditions on V(phi) which determine whether or not such models produce a future big rip. Observational constraints are derived on two classes of these models: power-law potentials with V(phi) = lambda phi^alpha (with alpha positive or negative) and exponential potentials of the form V(phi) = beta e^{lambda phi^alpha}. It is shown that these models spend more time in a state with Omega_m ~ Omega_phi than do corresponding models with a constant value of w, thus providing a more satisfactory solution to the coincidence problem.

The dark energy that appears to produce the accelerating expansion of the universe can be characterized by an equation of state p=w\rho with w<-1/3. A number of observational tests have been proposed to study the value or redshift dependence of w, including SN Ia distances, the Sunyaev-Zel'dovich effect, cluster abundances, strong and weak gravitational lensing, galaxy and quasar clustering, galaxy ages, the \lya forest, and CMB anisotropies. The proposed observational tests based on these phenomena measure either the distance-redshift relation d(z), the Hubble parameter H(z), the age of the universe t(z), the linear growth factor D_1(z), or some combination of these quantities. We compute the evolution of these four observables, and of the combination H(z)d(z) that enters the Alcock-Paczyznski anisotropy test, in models with constant w, in quintessence models with some simple forms of the potential V(\phi), and in toy models that allow more radical time variations of w. Measurement of any of these quantities to precision of a few percent is generally sufficient to discriminate between w=-1 and w=-2/3. However, the time-dependence predicted in quintessence models is extremely difficult to discern because the quintessence component is dynamically unimportant at the redshifts where w departs substantially from its low-z value. Even for the toy models that allow substantial changes in w at low redshift, there is always a constant-w model that produces very similar evolution of all of the observables simultaneously. We conclude that measurement of the effective equation of state of the dark energy may be achieved by several independent routes in the next few years, but that detecting time-variation in this equation of state will prove very difficult except in specialized cases.

A time variation in the Higgs vacuum expectation value alters the electron mass and thereby changes the ionization history of the universe. This change produces a measurable imprint on the pattern of cosmic microwave background (CMB) fluctuations. The nuclear masses and nuclear binding energies, as well as the Fermi coupling constant, are also altered, with negligible impact on the CMB. We calculate the changes in the spectrum of the CMB fluctuations as a function of the change in the electron mass. We find that future CMB experiments could be sensitive to |\Delta m_e/m_e| \sim |\Delta G_F/G_F| \sim 10^{-2} - 10^{-3}. However, we also show that a change in the electron mass is nearly, but not exactly, degenerate with a change in the fine-structure constant. If both the electron mass and the fine-structure constant are time-varying, the corresponding CMB limits are much weaker, particularly for l < 1000.