4.1. Methods and limitations
Star formation histories of resolved dwarf galaxies are commonly derived through photometric techniques. The most widely used method consists of sophisticated modelling of the observed color-magnitude diagrams (CMDs) through synthetic CMDs taking into account photometric errors, seeing, and crowding effects. For a recent review of procedures and techniques see Aparicio (1999). The methods are limited by the quality of the observations and by how closely theoretical evolutionary models reproduce observational features. For instance, Olsen (1999) notes that old red giant branches of evolutionary models may fit the observations poorly, which can lead to an underestimation of the contribution of the old population. Free parameters in modelling include the adopted initial mass function slope and the binary fraction.
Additional constraints can be imposed by using special types of stars as tracers of certain evolutionary phases. For instance, the presence of HB stars and RR Lyrae variables is a reliable indicator of an old population even when sufficiently deep main-sequence photometry is lacking. It is important to keep in mind that the age resolution that can be obtained is not linear and decreases strongly for older populations. Whereas young populations with short-lived, luminous massive stars can be accurately age-dated to within a few million years, the accuracy for the oldest, long-lived evolutionary phases is of the order of a few billion years. Relative ages of resolved old populations with high-quality, deep main-sequence photometry, on the other hand, can be established with a resolution of a Gyr or less through direct comparison with CMDs of Galactic globular clusters. In the following, ``young'' refers to populations with ages < 1 Gyr, ``intermediate-age'' denotes the age range from 1 Gyr to 10 Gyr, and ``old'' stands for ages > 10 Gyr.
Owing to the availability of 10-m class telescopes, spectroscopic measurements of stellar abundances are now feasible for individual supergiants and the brightest red giants in galaxies as distant as the M31 subgroup. Together with emission-line spectroscopy of H II regions, these data help to constrain the metallicity and metallicity spread in certain evolutionary phases. Still, accurate metallicity information as a function of time is lacking for almost all galaxies.
The increasing amount of data on internal kinematics and dwarf galaxy proper motions are beginning to constrain their dynamical history. Unfortunately accurate orbital data are not yet available for almost all of the Local Group galaxies, making it difficult to evaluate the suggested impact of environmental effects and interactions discussed later.