\documentstyle[paasms4,pptwocol]{preprint}
\input pub.sty
\begin{document}
\title{SPECTRAL CLASSIFICATION OF THE 30 DORADUS\\
~\\
STELLAR POPULATIONS}
\author{Nolan R.~Walborn\/\thanks{Visiting Observer, Anglo-Australian Telescope.}\morethanks{Visiting Astronomer, Cerro Tololo Inter-American 
Observatory, National Optical Astronomy Observatories, operated by the 
Association of Universities for Research in Astronomy, Inc., under
contract with the National Science Foundation.}
and J.~Chris Blades\/\samethanks{1}\\
\\
Space Telescope Science Institute\/\thanks{Operated by the Association 
of Universities for Research in Astronomy, Inc., under contract with the 
National Aeronautics and Space Administration.}\\
\\
3700 San Martin Drive\\
\\
Baltimore, MD  21218}

\tobe{Vol.\ 112, October 1997}{ApJ Supplements}
\recacc{13 January 1997}{7 April 1997}

\maketitle

\abstract{An optical spectral classification study of 106~OB stars within
the 30~Doradus Nebula has sharpened the description of the
spatial and temporal structures among the associated clusters.
Five distinct stellar groups are recognized: (1)~the central
early-O (Carina phase) concentration, which includes R136; (2)~a
younger (Orion phase) population to the north and west of R136,
containing heavily embedded early-O dwarfs and IR sources, the
formation of which was likely triggered by the central
concentration; (3)~an older population of late-O and early-B
supergiants (Scorpius OB1 phase) throughout the central field,
whose structural relationship, if any, to the younger groups is
unclear; (4)~a previously known, older still, compact cluster 3$'$
northwest of R136, containing A- and M-type supergiants (h and
$\chi$~Persei phase), which evidently affects the nebular dynamics
substantially; and (5)~a newly recognized Sco OB1-phase
association, surrounding the recently discovered Luminous Blue
Variable R143 in the southern part of the Nebula.  The intricacy
of this region and the implications for the interpretation of
more distant starbursts are emphasized.  The evidence is that the
formation of the 30~Dor stellar content was neither instantaneous
nor continuous, but rather that it corresponds to discrete events
at different epochs.

The average difference between the derived and calibration
absolute visual magnitudes of the stars is 0.05, indicating that
the classification, calibration, and adopted distance modulus
($V_0 - M_V = 18.6$) are accurate.  For 70 of the stars, either
the absolute value of that difference is $\leq 0.6$~mag, or they are subluminous
dwarfs or superluminous supergiants.  Many astrophysically
interesting objects have been isolated for further investigation.
Surprisingly in view of the presence of several O3 supergiants,
the mid-Of star R139 is identified as the most massive object in
this sample; it is located well along the 120~$M_{\odot}$ track,
very near the Humphreys-Davidson limit, and it is likely an
immediate LBV precursor.  This work can and should be extended in
three ways: (1)~higher-resolution and --S/N observations of many
of the stars with larger ground-based telescopes for quantitative
analysis; (2)~ground-based spectral classification of the
numerous additional accessible stars in the field; and
(3)~spatially resolved spectral classification of compact multiple
systems with the {\it Hubble Space Telescope}.}