Previous
abstract 
Next abstract
I present UBV\/ photometry for 2400 stars in the OB associations of 30~Doradus in the LMC, and new spectroscopic classifications for 54 stars including 23 O stars. The entire catalog (exclusive of the dense core cluster R~136) is photometrically complete to $V = B = 18$~mag and $U = 17$~mag.
From these data, I have determined the effective temperatures and bolometric magnitudes of the stars and placed them on the theoretical H-R diagram. Using stellar evolution models, I then bin the stars by mass to obtain the initial mass function (IMF). The IMF shows marked curvature, flattening out to lower masses, even for masses above which the photometry is complete. Best estimates of the IMF slope yield values of $\Gamma = -1.3$ to $-1.5$ for ${\cal M} \ge 12~{\cal M}_\odot$, where the Salpeter slope is $\Gamma = -1.35$. However, there are indications that different regions of 30~Doradus have different IMF slopes, perhaps the result of sequential star formation effects.
Calculations of the Lyman continuum photon luminosity, $N_{\rm Ly}$, lead to the result that $\sim 3.2 \times 10^{51}$~photons s$^{-1}$ are being produced by the observed stars in 30~Doradus. This is equivalent to the ionization luminosity of 440 O7~V stars or 77 O5~V stars. Less than 60\% of this luminosity comes from the 150 stars with spectroscopic classifications, implying that there still remain a large number of OB stars yet to be observed spectroscopically. This, along with stellar evolution effects, could explain the IMF curvature and the apparent deficiency of stars in the most massive bins of the IMF.
For an ionization-bounded H\,{\sc ii} region, the H$\alpha$ luminosity due to the Lyman continuum photons produced by the stars would be $L_{{\rm H} \alpha} \approx 4.5 \times 10^{39}$~erg~s$^{-1}$. This value agrees with the luminosity determined from the observed H$\alpha$ flux, but is a lower limit since the effects of W-R stars and the unresolved cluster of R~136 have not been included.
