2006ApJ...647..984H

Chemical evolution models and Monte Carlo simulation techniques have been combined for the first time to study the distribution of blue compact galaxies on the N/O plateau. Each simulation comprises 70 individual chemical evolution models. For each model, input parameters relating to a galaxy's star formation history (bursting or continuous star formation, star formation efficiency), galaxy age, and outflow rate are chosen randomly from ranges predetermined to be relevant. Predicted abundance ratios from each simulation are collectively overplotted onto the data to test its viability. We present our results both with and without observational scatter applied to the model points. Our study shows that most trial combinations of input parameters, including a simulation comprising only simple models with instantaneous recycling, are successful in reproducing the observed morphology of the N/O plateau once observational scatter is added. Therefore, simulations that include delay of nitrogen injection are no longer favored over those that propose that most nitrogen is produced by massive stars, if only the plateau morphology is used as the principal constraint. The one scenario that clearly cannot explain plateau morphology is one in which galaxy ages are allowed to range below 250 Myr. We conclude that the present data for the N/O plateau are insufficient by themselves for identifying the portion of the stellar mass spectrum most responsible for cosmic nitrogen production.