A useful reference model for chemical evolution theory is
the so-called "simple
model", involving progressive conversion of gas to stars in a
closed box. In this model, stars either live
forever or die instantaneously, returning enriched material to the
interstellar medium (ISM). The mass of freshly produced metals per unit mass
of long lived stars formed is the "yield", p. Then the mass fraction
of metals, Z, in the gas increases as the gas fraction,
µ 
Mgas / Mtot, decreases,
according to Z = p ln µ-1
(Searle & Sargent
1972).
An interesting way to express
measured abundances is the "effective yield", defined
in terms of the simple model of chemical evolution: Z(O) = peff (O) lnµ-1, etc.
Oxygen, neon, and sulfur, among other elements, are "primary" products of stellar nucleosynthesis, whose yields depend relatively little on the abundances in the progenitor star. On the other hand, nitrogen largely results from secondary production, so that the yield is proportional to the abundance of C and O in the progenitor. This leads to an increase in N/O with increasing O/H, above a minimum N/O that corresponds to the primary contribution to nitrogen production. This is shown by observations of H II regions in spiral and irregular galaxies for O/H above a threshold value ~ 0.2 solar (e.g., Talbot & Arnett 1974; Garnett 1990, 2001b). Iron is believed to be produced in part by type Ia supernovae that require time (~ 109 yr) to evolve to the point of explosion (see review by Wheeler, Sneden, & Truran 1989). The more massive stars producing elements such as O explode effectively instantaneously, and therefore the O/Fe ratio is an indicator of the timescale on which a population of stars was formed.
The number of main sequence stars as a function of abundance in the solar neighborhood disagrees with the simple model in the sense that there are too few metal poor stars (van den Bergh 1962; Audouze & Tinsley 1976). This "G dwarf problem" is an important constraint on models for chemical evolution of the Galaxy. One solution is that continuing infall of metal poor gas causes the abundances to approach an asymptotic value at which most stars are formed (Larson 1972).
Pagel (2001) gives a succinct discussion of results involving stellar abundances in the halo, "thick disk", and "thin disk" of the Milky Way. The high O/Fe ratio in the halo and thick disk suggests a brief formation period, early in the history of the Galaxy. The thin disk has lower O/Fe and a metallicity distribution consistent with prompt initial enrichment (perhaps from the thick disk) combined with infall of extragalactic gas, with a hiatus in star formation between the thick and thin disk formation of several billion years. See Pagel (2001) and Chiappini, Matteucci, & Gratton (1997) for further discussion and references.