0706.1939/ms.tex
1: %\documentclass[manuscript]{aastex}
2: \documentclass[12pt,preprint]{aastex}
3: %%
4: %% fdraft: rtr dsb tmb 9 march 2007  whoopie
5: %% draft8: rtr new latex macros etc
6: %% draft7: tmb rtr dsb 8 March 2007: final version?
7: %% draft6: tmb rtr dsb 7 March 2007
8: %% draft5: tmb  Mar 2007 mostly language and typos
9: %% draft4: dsb  Nov 2006
10: %% draft3: tmb  2 august 2006
11: %% draft2: dsb  ca. 2003
12: %% draft1: dsb 
13: %
14: % Author-defined commands
15: %----------------------------------------------------------------------
16: % ----- define organization
17: \newcommand{\nraoblurb}{The National Radio Astronomy Observatory is
18: a facility of the National Science Foundation operated under cooperative
19: agreement by Associated Universities, Inc.}
20: % ----- units --------------
21: %           \arcdeg !!!! use AASLaTeX macro!!!!
22: \newcommand{\degree}{\ensuremath{\,^\circ}}
23: \newcommand{\degper}{\rlap.{^{\circ}}}
24: \newcommand{\arcmper}{\rlap.{^{\prime}}}
25: \newcommand{\arcsper}{\rlap.{^{\prime\prime}}}
26: %
27: \newcommand{\rgal}{\ensuremath{{R_{\rm gal}}}}
28: \newcommand{\gyr}{\ensuremath{\,{\rm Gyr}}}
29: \newcommand{\mhz}{\ensuremath{\,{\rm MHz}}}
30: \newcommand{\ghz}{\ensuremath{\,{\rm GHz}}}
31: \newcommand{\kel}{\ensuremath{\,{\rm K}}}
32: \newcommand{\K}{\ensuremath{\,{\rm K}}}
33: \newcommand{\mk}{\ensuremath{\,{\rm milliK}}}
34: \newcommand{\m}{\ensuremath{\,{\rm m}}}
35: \newcommand{\cm}{\ensuremath{\,{\rm cm}}}
36: \newcommand{\percc}{\ensuremath{\,{\rm cm^{-3}}}}
37: \newcommand{\kpc}{\ensuremath{\,{\rm kpc}}}
38: \newcommand{\pc}{\ensuremath{\,{\rm pc}}}
39: \newcommand{\kms}{\ensuremath{\,{\rm km\, sec^{-1}}}}
40: \newcommand{\msun}{\ensuremath\,M_\odot}
41: \newcommand{\s}{\,s}
42: \newcommand{\microns}{\ensuremath{\, \mu {\rm m}}}
43: \newcommand{\ev}{\,ev}
44: \newcommand{\jy}{\,Jy}
45: \newcommand{\jyb}{\ensuremath{\rm \,Jy\,beam^{-1}}}
46: \newcommand{\mjy}{\,mJy}
47: \newcommand{\mjyb}{\ensuremath{\rm \,mJy\,beam^{-1}}}
48: \newcommand{\microjy}{\,\ensuremath{\mu}Jy}
49: \newcommand{\microjyb}{\ensuremath{\rm \,\mu Jy\,beam^{-1}}}
50: %
51: \newcommand{\kmsp}{\ensuremath{{\,{\rm km\, sec^{-1}\, pc^{-1}}}}}
52: \newcommand{\EM}{\ensuremath{{\,{\rm pc}\,{\rm cm}^{-6}}}}
53: \newcommand{\kapi}{\ensuremath{\kappa_i}}
54: % ----- chemical symbols 
55: \newcommand{\hp}{\ensuremath{{\rm H}^+}}
56: \newcommand{\he}[1]{\ensuremath{^#1{\rm He}}}
57: \newcommand{\heo}[1]{\ensuremath{^#1{\rm He}^{\rm o}}}
58: \newcommand{\hep}[1]{\ensuremath{^#1{\rm He}^+}}
59: \newcommand{\hepp}[1]{\ensuremath{^#1{\rm He}^{++}}}
60: \newcommand{\her}[1]{\ensuremath{^#1{\rm He}/{\rm H}}}
61: \newcommand{\hepr}[1]{\ensuremath{^#1{\rm He}^{+}/{\rm H}^+}}
62: \newcommand{\heppr}[1]{\ensuremath{^#1{\rm He}^{++}/{\rm H}^+}}
63: % ----- he-3 commands
64: \newcommand{\hethree}{\ensuremath{{}^3{\rm He}}}
65: \newcommand{\hefour}{\ensuremath{{}^4{\rm He}}}
66: \newcommand{\hepthree}{\ensuremath{{}^3{\rm He}^+}}
67: \newcommand{\lir}[1]{\ensuremath{^{#1}{\rm Li}/{\rm H}}}
68: % ----- recomb lines
69: \newcommand{\hal}{\ensuremath{{\rm H}91\alpha}}
70: \newcommand{\hbet}{\ensuremath{{\rm H}114\beta}}
71: \newcommand{\hgam}{\ensuremath{{\rm H}130\gamma}}
72: \newcommand{\heal}{\ensuremath{{\rm He}91\alpha}}
73: \newcommand{\hebet}{\ensuremath{{\rm He}114\beta}}
74: \newcommand{\hegam}{\ensuremath{{\rm He}120\gamma}}
75: \newcommand{\calx}{\ensuremath{{\rm C}91\alpha}}
76: \newcommand{\cbet}{\ensuremath{{\rm C}114\beta}}
77: \newcommand{\cgam}{\ensuremath{{\rm C}130\gamma}}
78: %
79: \newcommand{\halk}{\ensuremath{{\rm H}70\alpha}}
80: \newcommand{\hbetk}{\ensuremath{{\rm H}88\beta}}
81: \newcommand{\healk}{\ensuremath{{\rm He}70\alpha}}
82: \newcommand{\hebetk}{\ensuremath{{\rm He}88\beta}}
83: \newcommand{\calk}{\ensuremath{{\rm C}70\alpha}}
84: \newcommand{\cbetk}{\ensuremath{{\rm C}88\beta}}
85: % ----- misc 
86: \newcommand{\expo}[1]{${10^{#1}}$}
87: \newcommand{\nexpo}[2]{\ensuremath{#1 \times 10^{#2}}}
88: \newcommand{\hi}{H~{\sc i}}
89: \newcommand{\hei}{He~{\sc i}}
90: \newcommand{\hii}{H~{\sc ii}}
91: \newcommand{\heii}{He~{\sc ii}}
92: \newcommand{\y}[1]{\ensuremath{y_{#1}}}
93: \newcommand{\yp}[1]{\ensuremath{y_{#1}^{+}}}
94: \newcommand{\ypp}[1]{\ensuremath{y_{#1}^{++}}}
95: \newcommand{\ngc}[1]{NGC~#1}
96: \newcommand{\sgr}[1]{Sgr\thinspace #1}
97: %
98: % Define \lsim and \gsim per AAS LaTeX macros
99: %
100: \newcommand{\gsim}{\ensuremath{\gtrsim}}
101: \newcommand{\lsim}{\ensuremath{\lesssim}}
102: %
103: \def\rtrcom#1{{\bf RTR: #1}}
104: \def\tmbcom#1{{\bf TMB: #1}}
105: \def\dsbcom#1{{\bf DSB: #1}}
106: %-------------------------------------------------------------------
107: 
108: \begin{document}
109: 
110: %\slugcomment{draft7 tmb+rtr+dsb: \today}
111: \shorttitle{$^{3}$He in the Milky Way}
112: \shortauthors{Bania et al.}
113: 
114: \title{$^{\bf 3}$He in the Milky Way Interstellar Medium: Ionization Structure}
115: 
116: \author{T. M. Bania\altaffilmark{1}, Dana S. Balser\altaffilmark{2},
117: Robert T. Rood\altaffilmark{3}, T. L. Wilson\altaffilmark{4}, \& 
118: Jennifer M. LaRocque\altaffilmark{5}}
119: 
120: \altaffiltext{1}{Institute for Astrophysical Research, Department of Astronomy,
121: Boston University, 725 Commonwealth Avenue, Boston MA 02215, USA. 
122: (bania@bu.edu)}
123: \altaffiltext{2}{National Radio Astronomy Observatory, 520 Edgemont Rd., 
124: Charlottesville, VA 22903, USA.}
125: \altaffiltext{3}{ P.O. Box 400325, Astronomy Department, University of Virginia, 
126: Charlottesville VA 22904-4325, USA.}
127: \altaffiltext{4}{ESO Room 422, Karl-Schwarzschild-Str. 2, 85748
128: Garching, Germany}
129: \altaffiltext{5}{Saint Michael's College, Colchester VT 05439}
130: 
131: \received{}
132: \revised{}
133: \accepted{}
134: \cpright{AAS}{2007}
135: 
136: \begin{abstract}
137: The cosmic abundance of the \he3\ isotope has important implications
138: for many fields of astrophysics.  We are using the 8.665\ghz\
139: hyperfine transition of \hep3\ to determine the \her3\ abundance in
140: Milky Way \hii\ regions and planetary nebulae.  This is one in a
141: series of papers in which we discuss issues involved in deriving
142: accurate \her3\ abundance ratios from the available measurements.
143: Here we describe the ionization correction we use to convert the
144: \hepr3\ abundance, \yp3, to the \her3\ abundance, \y3.  In principle
145: the nebular ionization structure can significantly influence the \y3\
146: derived for individual sources.  We find that in general there is
147: insufficient information available to make a detailed ionization
148: correction.  Here we make a simple correction and assess its validity.
149: The correction is based on radio recombination line measurements of
150: \hp\ and \hep4, together with simple core-halo source models.  We use
151: these models to establish criteria that allow us to identify sources
152: that can be accurately corrected for ionization and those that
153: cannot. We argue that this effect cannot be very large for most of the
154: sources in our observational sample.  For a wide range of models of
155: nebular ionization structure we find that the ionization correction
156: factor varies from 1 to 1.8.  Although larger corrections are
157: possible, there would have to be a conspiracy between the density and
158: ionization structure for us to underestimate the ionization correction
159: by a substantial amount.
160: \end{abstract}
161: 
162: \keywords{\hii\ regions --- ISM: abundances --- radio lines: ISM}
163: 
164: \section{INTRODUCTION}
165: 
166: \subsection{The $^{\bf 3}$He Experiment}\label{sec:he3}
167: 
168: The abundance of the light isotope of helium, \he3, is astrophysically
169: important.  Knowing the \her3\ abundance ratio can be used to test the
170: theory of stellar nucleosynthesis; it gives important information
171: needed to evaluate models of Galactic chemical evolution; it can help
172: constrain Big Bang Nucleosynthesis.  For over two decades we have used
173: the 8.665\ghz\ hyperfine transition of \hep3\ to derive the \her3\
174: abundance in the interstellar medium (ISM) of the Milky Way.  Our
175: \he3\ sources, planetary nebulae and \hii\ regions, are distributed
176: throughout the Milky Way's disk, from the Galactic Center to the
177: outermost regions.  There is no other \he3\ spectral transition
178: available that can be used to probe transgalactic paths with the
179: sensitivity and accuracy of the \hep3\ hyperfine line.
180: 
181: We observed \hep3\ using the Max-Planck-Institut f\"{u}r Radioastronomie
182: (MPIfR) 100\m\ telescope and the National Radio Astronomy Observatory
183: (NRAO)\footnotemark[1] 140\,Foot telescope and Very Large Array (VLA).
184: \footnotetext[1]{\nraoblurb} With the 100\m\ and VLA we primarily studied
185: planetary nebulae \citep{BBRW97, balser06} whereas the 140\,Foot was used mostly
186: for \hii\ regions \citep[hereafter Paper I]{BBRWW97}.
187: \defcitealias{BBRWW97}{Paper I}
188: %
189: In addition to the \hep3\ observations, measurements were also made of
190: a variety of radio recombination line transitions (RRLs) of H, \he4,
191: and C, as well as of the thermal continuum emission from the nebulae.
192: Paper I discussed the observations and the measurement errors.  It is
193: a summary of the status of the observations as of 1996 March.  We
194: continued the \hep3\ program on the 140\,Foot until it was
195: decommissioned in 1999 July.  (The last scientific spectrum observed
196: at the 140\,Foot telescope was the \hep3\ scan that finished at
197: 08:12:20 EDT 19 July 1999.) The final paper in this series
198: \citep[Paper IV]{RBB04} summarizes the results of the \hep3\ 
199: experiment for all the sources observed with the NRAO 140\,Foot
200: telescope during the period 1982--1999.  Paper IV compiles the
201: observed properties of the \hep3\ emission and gives the final \her3\
202: abundances derived for the NRAO 140\,Foot sample of
203: \hep3\ \hii\ regions.
204: 
205: The quantity of astrophysical interest is the \her3\ abundance ratio by
206: number which we define as \y3.  This ratio provides information about
207: both stellar \citep{Charbonnel98} and Galactic chemical evolution
208: \citep{Romano03} and constrains cosmological models during the era of
209: primordial nucleosynthesis \citep{Yang84, BRB02}.  The species directly
210: accessible to observation, however, are \hep3\ and \hp.  The
211: collisionally excited \hep3\ hyperfine transition directly measures the
212: total column density of \hep3\ atoms along the line-of-sight within the
213: telescope's beam.  To determine the column density of \hp, either radio
214: continuum or hydrogen RRLs can be used.  At 8.7\ghz\ the radio continuum
215: emission in \hii\ regions and planetary nebulae is primarily due to
216: thermal free-free emission which is proportional to the emission
217: measure, $EM = \int \, n_{e}^{2} \, d\ell$, where $n_{e}$
218: is the electron density and $d\ell$ is the differential path length
219: through the ionized nebula.  The H RRL emission is also proportional to
220: the emission measure.  Thus neither the RRL nor the radio continuum data
221: for H$^{+}$ are probing the total proton column density (or mass) in a
222: straightforward manner.  {\it In order to determine the \hepr3\
223: abundance ratio, it is therefore necessary to model the nebula's density
224: structure.\/} Models are required because in most cases detailed
225: information on nebular density structure is not observationally
226: accessible.  Nebular density models and the \hepr3\ abundance ratios
227: derived from them are discussed by \citet[hereafter Paper II]{BBRW99}.
228: \defcitealias{BBRW99}{Paper II}
229: %
230: In particular, Table 5 of Paper II lists the adopted \hepr3\ abundance
231: ratios for the 21 \hii\ regions discussed in Paper I.
232: 
233: Using the observables, \hep3\ and \hp, to derive a \her3\ abundance
234: ratio, \y3, is, however, a two step process. The first step is to
235: determine the source density structure and calculate \hepr3\ (Paper
236: II).  The next step, the topic of this paper, is to determine the
237: source ionization structure and to use this information to convert
238: \hepr3\ into \her3.  This requires an understanding of the ionization
239: properties of each ionized nebula.  Specifically, one needs to know
240: the ionization structure of both the He and H throughout each nebula.
241: Because the first ionization potentials of H (13.6 eV) and He (24.6
242: eV) are nearly a factor of two different, in principle ionization can
243: significantly influence \y3\ for individual sources.  We find that in
244: general there is insufficient information available to make a detailed
245: ionization correction.  Here we make a simple correction and assess
246: its validity.
247: 
248: \subsection{Helium Ionization Structure in H\,{\bf\small II} Regions}
249: \label{sec:struc}
250: 
251: Several diagnostics are commonly used to study the helium ionization
252: structure in \hii\ regions.  The most direct is to measure the
253: fraction of \heo4, \hep4, and \hepp4\ within the \hii\ (${\rm H}^{+}$)
254: region.  Since both \he4\ and \he3\ have essentially the same
255: ionization potential the \hep3\ and \hep4\ emission should come from
256: identical zones within the nebula.  The \hep4\ and \hepp4\ emission
257: can be measured using recombination lines.  Unfortunately there is no
258: direct way of measuring the amount of neutral helium within the \hii\
259: region.  In some cases the spectral types of the ionizing OB stars can
260: be identified from optical photometry and spectroscopy.  The helium
261: ionization properties are then derived using stellar atmosphere models
262: to calculate the expected escaping flux \citep[e.g.,][]{Vacca96},
263: together with photoionization models of the nebula to determine the He
264: ionization structure \citep[e.g.,][]{Rubin84}.  At optical wavelengths
265: spectral transitions of several other atomic species are also observed
266: to help constrain these models.  The most detailed analyzes have been
267: made for the Orion nebula in order to determine the total \her4\
268: abundance ratio \citep{Mathis91, Baldwin91, Rubin91, Pogge92,
269: Esteban98, Blagrave06}.  Nevertheless, the limiting factor in
270: determining accurate total helium abundances in Orion is in converting
271: \hepr4\ to \her4.
272: 
273: Most \hii\ regions are ionized by several stars and may thus have a
274: complex geometry. Because of this little direct information may be
275: known about the radiation field.  A variety of diagnostics have been
276: developed to probe the ionization structure.  \citet{Vilchez88}
277: developed a radiation softness parameter based on the fine structure
278: lines of O and S ions that is not very sensitive to chemical
279: composition \citep[see also][]{Shields78, Mathis82,
280: Mathis85}. \citet{Armour99} use infrared transitions of Ne and Ar ions
281: to determine the ionization structure.  In general \hii\ regions
282: ionized by a hard radiation field produce the most accurate \her4\
283: abundance ratios since all of the helium within the \hii\ regions will
284: be ionized.  For example, metal poor blue compact galaxies with \hii\
285: regions ionized by hard radiation fields are used to measure
286: primordial \her4.  The ionization structure of these objects has been
287: extensively studied \citep{Ballantyne00, Viegas00, Gruenwald02,
288: Sauer02, Peimbert02a, Izotov07}.  In some cases the total \her4\
289: abundance ratio will be less than \hepr4\ by a few percent,
290: significant for cosmological implications, since the \heii\ zone is
291: larger than the \hii\ zone.  If clumping exits in these \hii\ regions,
292: however, the determined \y4\ values will be underestimated by a few
293: percent \citep{Mathis05}.  In the Galaxy the high excitation \hii\
294: regions M17 and S206 have been used instead of Orion to determine the
295: total \her4\ abundance ratio 
296: \citep{Peimbert93, Esteban99, Deharveng00, Balser06}.
297: 
298: Many of our sources are located throughout the Galactic disk and are
299: totally obscured at optical wavelengths by dust.  Furthermore, most of
300: the optically visible \hii\ regions in our sample are low emission
301: measure \hii\ regions in the outer Galaxy.  The low EM renders useless
302: the Orion-type detailed models since the spectral diagnostics used to
303: constrain them are not available.  Fortunately, unlike many of the
304: light elements, \her3\ abundances accurate to $\sim$\,10\% can yield
305: important astrophysical conclusions \citep{Wilson94} .  Because of
306: this we adopt here a simple ionization structure model that is
307: constrained primarily by H and \he4\ radio recombination line
308: observations.  We then use numerical models to assess the accuracy of
309: this approach for the \hii\ regions in our sample.  Our goal is to
310: identify a subset of sources in our nebular sample wherein we can
311: derive \her3\ abundances accurate to $\sim$\,10\%.
312: 
313: \subsection{${\bf {}^{\bf 3}He}$ Ionization Correction}
314: \label{sec:defkap}
315: 
316: We seek an ionization correction factor to derive the \her3\ abundance
317: from the \hepr3\ abundance gotten from the \hep3\ observations.  We
318: define the ionization correction factor, \kapi, to be
319: $y_{3} \equiv \kappa_{i}\,y_{3}^{+}$, where $y$ is the He/H
320: abundance ratio by number, the subscript denotes the isotope, and the
321: superscript is the ionization state.\footnotemark[2]
322: %
323: \footnotetext[2]{Thus the notation used throughout is: \y3\ = \her3;
324: \yp3\ = \hepr3; \y4\ = \her4; and \yp4\ = \hepr4.}
325: %
326: The ionization correction factor is determined by using the H and
327: \he4\ radio recombination line observations.  We assume that the \hep3\
328: emission traces the \hep4\ emission, that the amount of neutral and
329: doubly ionized helium is negligible, and a canonical value for the
330: total \her4\ abundance, \y4. The \her3\ abundance ratio is then given by
331: %
332: \begin{equation}\label{eq:defkap}
333: { 
334: y_{3} = \biggl(\frac{y_{4}}{y_{4}^{+}}\biggr)\,y_{3}^{+} 
335: = \kappa_{i}\,y_{3}^{+}\,\biggl(\frac{y_{\rm 4GAL}}{0.10}}\biggr).
336: \end{equation}
337: %
338: Here $y_{\rm 4GAL}$ is the actual \he4/H abundance by number in the
339: Milky Way.  There are, of course, some problems with this simple
340: model.  \hep3\ and \hep4\ are measured using transitions that are
341: sensitive in different ways to density structure within the nebula.
342: The \hep3\ is observed using a collisionally excited hyperfine
343: transition that is sensitive to the column density and thus
344: proportional to the electron density, $n_{e}$.  The \hep4\ is observed
345: using recombination transitions that are proportional to the emission
346: measure or $n_{e}^{2}$.  Thus if the \hii\ region has large density
347: fluctuations the different transitions may probe significantly
348: different material.  For example, consider a simple two-component
349: nebula with a small, very dense core and a larger, diffuse halo.
350: Under certain physical conditions the halo will dominate the \hep3\
351: emission while the core will dominate the \hep4\ emission.  If this is
352: coupled with ionization structure between the core and halo then the
353: simple formula in equation (1) will be incorrect.
354: 
355: It has also proven difficult to measure accurately the \her4\
356: abundance, \y4, in the Galaxy. \he4\ cannot be directly measured in
357: the Sun and must be inferred from theoretical stellar evolution models
358: and helioseismology.  Measurements of \he4\ in \hii\ regions using
359: recombination lines require an ionization correction as discussed
360: above.  A canonical value of $\y4 = y_{\rm 4GAL} = 0.1$ is adopted
361: here although there is evidence that this often cited value may be too
362: high (see \S\ref{sec:discuss}).  For this reason we have parameterized
363: equation (1) with the $y_{\rm 4GAL}$ factor so that our ionization
364: corrections can be easily scaled should an accurate Milky Way \y4\
365: value be derived in the future.
366: 
367: \subsection{Radio Recombination Line Observations}\label{sec:rrl}
368: 
369: Here we use high signal-to-noise ratio recombination line spectra
370: taken at two different spatial resolutions to probe for any
371: significant ionization structure.  For our \hii\ region sample we have
372: obtained H and \he4\ RRL data simultaneously for several different
373: transitions near 8\ghz\ with a spatial resolution of 3\farcm5 (see
374: Paper I).  These observations are briefly reviewed here in
375: \S\,\ref{sec:Xband}.  Because they sample the same \hii\ region 
376: zone as the \hep3\ spectra, these data are primarily used to determine
377: \yp4\ in equation (1).  Most \he4\ RRL data are from single-dish
378: telescopes which typically have spatial resolutions \gsim\ 1\arcmin.
379: Measuring accurate \yp4\ abundance ratios with single-dish telescopes
380: is difficult because non-random frequency structure is produced in the
381: instrumental baselines owing to reflections from various parts of the
382: telescope structure \citep[see][]{Lockman82}.  Improvements in
383: receiver technology have provided better stability and enhanced
384: signal-to-noise.  This, together with a better understanding of the
385: instrumental effects, has enabled more accurate measurements of \yp4\
386: (Peimbert et al. 1992a; Paper I).
387: 
388: Using these improved techniques we made additional observations at
389: 18\ghz\ with a spatial resolution of 1\farcm5 for a subset of our \hii\
390: region sample (\S\,\ref{Kband}).  The smaller beam of the 18\ghz\ RRLs
391: typically probes the more compact, dense components, whereas the larger
392: 3\farcm5 beam of the 8\ghz\ RRLs is more sensitive to the extended,
393: diffuse material.  Radio interferometers can measure the \hep4\
394: distribution within \hii\ regions at spatial resolutions \lsim\
395: 1\arcmin.  Although interferometers do not detect the diffuse emission
396: because of the zero-spacing flux problem, this information can be
397: provided by single-dish measurements.  To our knowledge only five \hii\
398: regions in our sample have been observed in H and \he4\ RRL emission
399: with radio interferometers: W3, \sgr{B2}, W43, W49, and W51.
400: 
401: Since the entire nebula can be imaged, the best optical data for this
402: study are observations made with Fabry-Perot spectrophotometers.
403: \citet{Caplan00} derived oxygen and helium abundances for 34 \hii\
404: regions using the ESOP Fabry-Perot instrument.  For many objects the
405: entire nebula was probed (the largest diaphragm was 4\arcmin\
406: 27\arcsec).  The \he4\ abundances from this survey are discussed by
407: \citet{Deharveng00}.  There are four \hii\ regions that are in our sample:
408: S206, S209, S212, and S252.
409: 
410: Although the 8\ghz\ RRLs are primarily used to determine \yp4\ the other
411: radio and optical \hep4\ data can not only provide an independent
412: confirmation of the \hepr4\ abundance, but also indicate whether the
413: assumptions in equation (1) are valid.  In \S\,\ref{sec:models}
414: numerical models are used to explore the physical conditions under which
415: the assumptions in equation (1) begin to fail.  This information is then
416: used in \S\,\ref{sec:kappa} to determine a value for the ionization
417: correction factor, \kapi, for each source. A discussion of the results
418: is in \S\,\ref{sec:discuss} and a summary of the paper in
419: \S\,\ref{sec:summary}.  
420: 
421: \section{$^{\bf 4}$He$^{\bf +}$ OBSERVATIONS}\label{sec:data}
422: 
423: \subsection{X-band (8\,GHz) Radio Recombination Lines}
424: \label{sec:Xband}
425: 
426: The X-band RRL observations are discussed in \S{2} of Paper I.  The
427: \hal\ and \heal\ transitions were observed in every \hii\ region.
428: Additional, higher order, transitions were also observed at nearby
429: frequencies with essentially the same spatial resolution (3\farcm5).
430: These data were used to constrain the nebular models in Paper II.
431: Here we only use the higher quality 91$\alpha$ and 114$\beta$
432: transitions of H and \he4.
433: 
434: Calibration details are discussed in \S{2.1} of Paper I.  We used the
435: planetary nebula \ngc{7027} to calibrate the flux density scale for
436: each observing epoch.  The absolute calibration is judged to be within
437: $\sim$\,10\%; the internal consistency is $\sim$\,5\%.  From these
438: observations we use only the ratios of the line areas to calculate
439: \hepr4; the calibration is therefore not a significant source of
440: uncertainty.
441: 
442: A much larger source of uncertainty are the standing waves produced by
443: reflections from various parts of the telescope structure (see \S{2.2}
444: in Paper I and references therein).  To minimize the amplitude of
445: these standing waves, observations are made with focus offsets
446: alternating between $\pm \lambda_0/8$ every 60\s.  Cancellation of the
447: standing waves is, however, not perfect and the spectral baselines
448: must be modeled.  A 12$^{\rm th}$ order polynomial baseline is 
449: used to model this non-random frequency structure.  The necessity for
450: using a high-order baseline and the selection of the appropriate order
451: to be used is discussed in \citet{Balser94} and Paper I.  The
452: results of Gaussian fits to each transition are given in Table 4 of
453: Paper I.
454: 
455: \subsection{K-band (18~GHz) Radio Recombination Lines}
456: \label{Kband}
457: 
458: Observations of H and \he4\ RRLs at K-band were made with the NRAO 140
459: Foot telescope during the periods: 1995 November, 1995 December, and
460: 1996 January.  The 140 Foot telescope has a half-power beam-width of
461: 1\farcm5 at 18\ghz.  The radiometer consisted of a dual circularly
462: polarized HEMT receiver located at the Cassegrain focus of the 140 Foot
463: telescope.  The system temperature on cold sky under good weather
464: conditions was $\sim 50$\kel.  The spectrometer consisted of the NRAO
465: model IV 1024 channel autocorrelator (AC).  The AC sampled four
466: independent 20\mhz\ wide frequency bands (``quadrants'') of 256 channels
467: each.  The velocity resolution at 18\ghz\ is 1.3\kms.  The 20\mhz\
468: bandwidth allowed both the H and \he4\ transitions to be observed
469: simultaneously within each quadrant.  For most of the observing the AC
470: was configured with the \halk/\healk\ and the \hbetk/\hebetk\ RRLs in
471: two quadrants each, simultaneously sampling orthogonal polarizations in
472: order to maximize our spectral sensitivity.  In some of the brighter
473: objects the 100$\gamma$ and 110$\delta$ transitions of H and \he4\ 
474: were observed.
475: 
476: Spectra were taken using the total power observing mode.  First a
477: reference spectrum (OFF) was taken at a position offset 6 minutes in
478: right ascension from the nebula and then the nebula itself (ON) was
479: observed.  Each position had an integration time of 6 minutes.  Local
480: pointing corrections were determined every hour.  After each pointing
481: a continuum cross scan, in right ascension and declination, was made
482: on the \hii\ region to measure the radio continuum emission.
483: 
484: The radiometer system temperature was measured using noise tubes
485: calibrated by connecting the receiver to matched hot and cold loads.
486: An ambient temperature absorber was used for the hot load and the
487: zenith sky for the cold load.  Since this procedure has many
488: uncertainties and is not always performed before each observing
489: session we have re-calibrated the system temperature by using the
490: observed continuum intensity of the unresolved planetary nebula
491: \ngc{7027}.  \ngc{7027} was observed several times during each
492: observing session during good weather near transit; it has a flux
493: density of 6\jy\ at 18\ghz\ \citep{Peng00}.
494: 
495: There are two effects that can modify the calibration as a function of
496: time.  First, at these frequencies the aperture efficiency of the 140
497: Foot telescope changes with elevation as gravity distorts the primary
498: reflector; an elevation gain correction must therefore be made.  We
499: measured the \hii\ region continuum intensity every hour for each
500: source in order to determine the appropriate gain correction.  Second,
501: the atmospheric opacity is significant and depends on the water vapor
502: content.  We made a tipping scan several times during each observing
503: period to measure the opacity.  Since we are interested in calculating
504: the \hepr4\ abundance ratio, which depends on the ratio of the line
505: intensities, the calibration effectively cancels.  All intensities
506: quoted here are therefore antenna temperatures calibrated using only
507: the \ngc{7027} data.
508: 
509: We used the techniques discussed in \S{2.2} of Paper I to model the
510: K-band spectral baselines, except that a 6$^{\rm th}$ order polynomial
511: was used instead of a 12$^{\rm th}$ order.  Although the AC
512: configuration was the same, yielding a 20\mhz\ bandwidth with 256
513: channels, only $\sim 12$\mhz\ is required to cover the frequency range
514: between the H and \he4\ lines.  We have therefore reduced the total
515: number of terms in the polynomial baseline model.
516: 
517: \subsection{The ${\bf {}^{\bf 4}He^{\bf +}\,/\,H^{\bf +}}$ Abundance Ratio}
518: \label{Yplus}
519: 
520: Example RRL spectra are shown in Figure~\ref{fig:h90h70}.  Only 10
521: objects from the original sample of 21 were observed at K-band.  The
522: data were smoothed to a velocity resolution of 8.1\kms.  The vertical
523: lines flag the expected locations of the H, He, and C lines for the
524: specified RRL transition (91$\alpha$ or 70$\alpha$).
525: Table~\ref{tab:data} summarizes the results of Gaussian fits to the
526: K-band spectra. Listed are the source name, the RRL transition, the peak
527: intensity and full-width half-maximum (FWHM) line-width together with
528: their associated errors, the total integration time, the resulting
529: r.m.s.  noise of the spectral baseline, and the quality factor of the
530: line.  The quality factor is determined using the same criteria as in
531: Paper I.  These include the signal-to-noise ratio, the structure of the
532: baseline, and the crowding of spectral lines.  We do not include any
533: calibration error in assessing the quality factor since we only use line
534: ratios in this paper.
535: 
536: The \hepr4\,, \yp4, abundance ratios calculated from the RRL data using
537: the line areas are summarized in Figure~\ref{fig:QMn} which plots the
538: \yp4\ abundance ratio as a function of the principal quantum number,
539: $n$, of the RRL transition.  The 70$\alpha$, 88$\beta$, 91$\alpha$, and
540: 114$\beta$ transitions are shown.  The Gaussian line parameters in Table
541: 4 of Paper I and in Table 1 here are used for the X-band and K-band
542: transitions, respectively.  The errors are determined by propagating the
543: uncertainties in the Gaussian fit to the line intensity and width.
544: 
545: For some nebulae there are real differences among the \yp4\ ratios
546: derived from the various RRL transitions.  This is to be expected, of
547: course, since the K-band spectra probe, on average, the denser, more
548: compact nebular interior whereas the X-band spectra measure a much
549: larger region, perhaps including a low density halo.  These
550: statistically significant differences are clearly seen in
551: Figure~\ref{fig:yplus} which compares in various ways the \yp4\ ratios
552: derived from the \hal\,(diamonds) and \halk\,(circles) spectra for the
553: 10 K-band sources (see \S\,\ref{sec:discuss}).
554: 
555: \section{SYNTHETIC NEBULAR MODELS}\label{sec:models}
556: 
557: We use the formalism defined in \S\ref{sec:defkap} together with
558: numerical models for the nebulae to explore the range of validity of
559: this approach to the ionization correction.  The \hii\ regions in our
560: sample span a wide range of physical conditions. At one extreme there
561: are large, complex \hii\ regions such as W49 that are ionized by many
562: OB-type stars and have multiple compact components in their density
563: structure.  At the other extreme are \hii\ regions such as S206, which
564: are smaller, much less complex sources, that are probably ionized by a
565: single OB-type star.  The simple ionization structure correction model
566: discussed in \S\,\ref{sec:defkap} could well not be very accurate for
567: sources like W49.  It is important, however, to understand under what
568: conditions this model breaks down and what impact this has on deriving
569: \her3\ abundance ratios.  
570: 
571: Paper II explored the effect of source density structure on the
572: derivation of the \her3\ abundance.  We used RRL and continuum data to
573: constrain the models.  Both RRL and continuum emission intensities are
574: proportional to $n_{e}^2$ and so they trace the densest components
575: of the nebulae.  We found that our sources could be grouped into two
576: broad categories.  In one class were ``complex,'' W49-like sources
577: that required models with complex density structure and non-LTE
578: radiative transfer in order to reproduce the observations.  In the
579: other class were ``simple,'' S206-like sources whose observations
580: could be reproduced by models with LTE radiative transfer and
581: homogeneous density structure.
582: 
583: We adopt a simple core-halo model for the nebular density and ionization
584: structure.  From the models we calculate simulated observations and
585: assess the effect of the model structure on our ability to recover
586: accurately the \her3\ abundance using our simple ionization structure
587: correction.  We use a computer program called NEBULA to model the \hii\
588: region (see Balser 1995 and Paper II for details about NEBULA). The
589: model nebula can have arbitrary density, temperature, and ionization
590: structure.  The gas consists of only hydrogen and helium.  We calculate
591: the radiative transfer through the model nebula for the radio continuum,
592: recombination line, and \hep3\ hyperfine line emission.  Since we shall
593: be comparing these synthetic spectral lines to the \S\,\ref{sec:data}
594: data taken at 8 and 18\ghz, we calculate the model antenna temperature
595: for each species by convolving the model sky brightness temperature
596: distribution with the telescope beam and multiplying by the telescope
597: main beam efficiency.  The telescope beam is assumed to be a Gaussian
598: with a half-power beam-width (HPBW) of 1\farcm5 or 3\farcm5.  These are
599: the HPBW sizes of the 140\,Foot telescope beam at 18\ghz\ and 8\ghz,
600: respectively.
601: 
602: All model nebulae are located at a distance of 8\kpc\ which is roughly
603: the average distance to our \he3\ sources.  The models are comprised
604: of nested, spherical, homogeneous, isothermal components.  For
605: simplicity, all model components have a fixed, 8,000\K\,, electron
606: temperature that is typical for Galactic \hii\ regions.  For all
607: models it is assumed that the total abundance ratios are
608: $\y3 \equiv \nexpo{1}{-5}$ and $\y4 \equiv 0.10$.
609: Table~\ref{tab:NEBULA} summarizes the model parameters used as input
610: for the NEBULA program.
611: 
612: Model A is a one-component, low density, $n_e = 100\percc$, nebula
613: with an angular size of 3\farcm5 (8.14\pc).  All other models are
614: two-component (core-halo) nebulae which consist of a 1\farcm5
615: (3.49\pc) core embedded in a larger 3\farcm5 nebula.  Listed for each
616: component is the angular size, the electron density, the fraction of
617: the component \hii\ mass relative to the total mass, the \yp3\
618: abundance ratio, and the \yp4\ abundance ratio.  Altogether, the
619: models span a core-halo density contrast ratio from 1:1 to 5:1, and a
620: core mass fraction from 8\% to 30\%.  That is, all models have
621: halos that contain more mass than the cores by a factor of either
622: $\sim 2$ or $\sim 10$.  These parameters span the range of the
623: Galactic \hii\ region physical properties derived from single-dish RRL
624: observations \citep[e.g., ][]{Quireza06,Quireza07}.
625: 
626: The NEBULA code calculates the following observable quantities: the
627: continuum antenna temperature at 8.665\ghz\ and the antenna
628: temperatures of the \hep3, \hal, \heal, \halk, and \healk\
629: transitions.  Properties of the synthetic observations that result
630: from the Table~\ref{tab:NEBULA} models are summarized in
631: Table~\ref{tab:models}.  Listed are the continuum and line antenna
632: temperature for the \hep3\ transition, the \yp3\ (\hepr3) abundance
633: ratio, the \yp4\ (\hepr4) abundance ratio and the values of \kapi\ and
634: \y3\ calculated using the equation (\ref{eq:defkap}) Ansatz.  All of
635: these quantities are determined for X-band (8\ghz).  The last column
636: lists the \yp4\ model values at K-band (18\ghz).  The \yp4\ values are
637: calculated from the H and \he4\ line parameters produced by NEBULA;
638: they are the ratio of the line areas.
639: %
640: The \yp3\ value is derived from the NEBULA models assuming a single,
641: homogeneous, fully ionized sphere where ${y}_{3}^{+} \propto
642: T_{\rm L}\,\Delta{v}/T_{\rm C}^{1/2}$.  This analytical expression 
643: approximates the numerical calculation to within a few percent 
644: \citep[see][]{Balser95}.  This is our standard way of
645: deriving the \her3\ abundance ratio (Paper II).  Here we apply this
646: analysis to the Table~\ref{tab:NEBULA} nebular models and use the
647: formalism developed in \S\ref{sec:defkap} to derive the ionization
648: correction factor, \kapi.
649: 
650: Model A1 is a uniform nebula with no density or ionization structure.
651: The NEBULA model abundances in Table~\ref{tab:models}, \yp3\ and \yp4,
652: exactly match the input abundances, \y3\ = \nexpo{1}{-5} and \y4\ =
653: 0.10.  This of course must be the case since Model A1 is constructed
654: to match our analysis assumptions.  It is a single, homogeneous, fully
655: ionized sphere and the NEBULA Model A1 gives an ionization correction of
656: $\kapi = 1$ as it should.
657: 
658: In Model A2 we simulate a one-component nebula that has 30\% of all
659: the helium neutral.  For this model the NEBULA \he4\ RRL synthetic
660: spectra together with equation (\ref{eq:defkap}) give an ionization
661: correction of $\kappa_i = 1.4$ which in turn produces \y3\ =
662: \nexpo{1.0}{-5}, the model input abundance.  Thus, for a one-component
663: model nebula the observables are sufficient to derive ionization
664: corrections that can accurately recover the input structure.
665: 
666: Model B has a core-halo density structure where the core component has
667: a higher density by a factor of 5, but the halo has a factor of
668: $\sim 2.3$ times more mass.  Both core and halo are fully ionized:
669: Model B has no ionization structure so we cannot derive any correction
670: factor.  The NEBULA Model B \yp3\ abundance thus underestimates the
671: true \yp3\ abundance by a factor of 1.3.  In Paper II we discuss
672: density structure corrections for the \hepr3\ abundance ratio
673: derivation.  There we model nebular density structure and estimate the
674: density correction factor for each \hii\ region in our sample.
675: Applying these techniques to Model B recovers, as expected, the input
676: \yp3\ abundance.
677: 
678: Models C and D are constant density nebulae with core-halo ionization
679: structure.  The halo mass in both models is $\sim 11.5$ times that of
680: the core.  Model C has a fully ionized core and a halo that is
681: under-ionized by 50\%.  Model D reverses this ionization structure.
682: That these models have ionization structure is revealed by the
683: different NEBULA values for the X and K-band \yp4\ abundance ratios
684: (see Table~\ref{tab:models}).  Since most of the mass resides in these
685: halos the NEBULA \hepr3\ abundance ratio is much less for Model C with
686: its under-ionized halo (\yp3\ = 0.57) than for model D (\yp3\ = 0.96).
687: The NEBULA results give ionization corrections of \kapi = 1.82 and
688: 1.05 for models C and D, respectively.  In both cases
689: Equation~\ref{eq:defkap} yields \y3\ = \nexpo{1.0}{-5}, the model
690: input value.  For these two models our simple ionization correction
691: reproduces the correct \her3\ abundance ratio.
692: 
693: Models E and F combine both density and ionization structure in a
694: core/halo geometry.  Both have high density cores (5:1 density
695: contrast) and halo masses that are $\sim2.3$ times that of the
696: cores.  Model E has a fully ionized core and a halo that is
697: under-ionized by 50\%.  Model F reverses this ionization structure.
698: Once again the different NEBULA values for the X and K-band \yp4\
699: abundance ratios reveal the presence of nebular ionization structure.
700: For model E the NEBULA \yp3\ value is 0.53, roughly a factor of two
701: too low.  Yet the NEBULA \yp4\ abundance ratio implies only a small
702: ionization correction factor, \kapi\ = 1.18, which gives a corrected
703: \y3\ abundance of only \nexpo{6.2}{-6}, 62\% of the input value.
704: The situation is much better for Model F with its fully ionized halo.
705: The NEBULA values for \yp4\ and \yp3\ together give an ionization
706: correction factor of \kapi\ = 1.56 and a \y3\ abundance that is within
707: 2\% of the input value.  
708: %
709: 
710: These models show that the simple ionization correction given in
711: Equation~\ref{eq:defkap} will be reasonably accurate for nebulae with
712: core/halo geometries.  Our correction begins to break down whenever the
713: low density component contains most of the mass and is under-ionized.
714: This is because the \hep3\ and RRL emission arises from gas located in
715: different zones of the nebula and these regions do not have the same He
716: ionization structure.  
717: 
718: In sum, the simple ionization correction given by equation 
719: (\ref{eq:defkap}) provides a reasonable value for the total \y3\
720: abundance ratio except for the Model E \hii\ region that has an
721: under-ionized halo containing a significant fraction of the total
722: mass.  Clearly more complicated density and ionization structures
723: exist in real \hii\ regions.  The simple models discussed here
724: nevertheless illustrate the main issues.  The simple ionization
725: correction breaks down when ionization structure exists between
726: significantly different density components.  True variations of \y4\
727: and \y3\ structure within the \hii\ region will complicate the
728: analysis.  Nonetheless, because of our large X-band beam that in many
729: cases covers the entire nebula, emission from dense, compact
730: components tends to be diluted.  We thus derive an average value of
731: \y3\ for each object.
732: 
733: \section{IONIZATION CORRECTION FACTOR}\label{sec:kappa}
734: 
735: Here we assess the ionization structure correction for our \he3\
736: sample of 21 Galactic \hii\ regions.  In Paper II we classify nebulae
737: into two categories: (i)``simple'' \hii\ regions dominated by
738: low-density, extended structure that is well fit by a single-component
739: uniform density sphere model; and (ii) ``complex'' \hii\
740: regions that have multiple high density components and thus require a
741: more sophisticated model.  The simple \hii\ regions are: G133.8, S206,
742: S209, S212, S228, S235, S252, S311, S76, S90, S156, \ngc{7538}, and
743: S162; the complex \hii\ regions are: W3, \sgr{B2}, G1.1, M17S, M17N,
744: W43, W49, and W51.  Simple \hii\ regions characteristically have low
745: emission measure.  For complex \hii\ regions that have significant
746: ionization structure we will underestimate \kapi\ as discussed in
747: \S\ref{sec:models}.  We therefore can only derive a lower limit
748: to the \her3\ abundance for these nebulae.  
749: 
750: We used various techniques to assess the excitation state of the
751: nebulae.  Table~\ref{tab:excitation} summarizes the excitation
752: properties for our \hii\ region sample.  Listed for the radio data are
753: the \hepr4\ abundance ratio, \yp4, derived using only the \hal\ and
754: \heal\ RRLs, the H-ionizing luminosity, $N_{\rm L}$, determined from the radio
755: continuum data in Paper I, and the spectral type of a single ZAMS star
756: that can produce the observed luminosity.  Spectral types are derived
757: from stellar atmosphere models: both the \citet{Vacca96} and \citet[in
758: parentheses]{Panagia73} models are listed.  Similar results from optical
759: data are taken from the literature.  The spectral types are based on
760: photometry and spectroscopy; the earliest type star thought to ionize
761: each nebula is listed in Table~\ref{tab:excitation}.  Overall, the radio
762: and optical spectral classifications are in good agreement.
763: 
764: We plot the ionization state, \yp4, of the 14 nebulae for which we
765: have sufficient information as a function of the log of the H-ionizing
766: luminosity, $N_{\rm L}$, in Figure~\ref{fig:Nlc}.  There is no discernible
767: trend connecting \yp4\ and $N_{\rm L}$.  Except for their systematically
768: lower $N_{\rm L}$ values (recall they are low emission measure nebulae), the
769: simple sources span the same \yp4\ range as do the complex nebulae.
770: Values of $N_{\rm L}$ determined from the radio continuum include the
771: contribution of H-ionizing photons from all of the early-type stars
772: within the \hii\ region complex.  Therefore low emission measure does
773: not necessarily imply that the radiation field is soft. For example
774: the radiation field of the low emission measure Rosette nebula is
775: dominated by O4 stars compared to the rather weak ionization field of
776: the high emission measure Orion nebula.
777: 
778: Below we discuss each \hii\ region separately in more detail.  An
779: ionization correction factor, \kapi, is determined for each object.
780: For consistency we only use data as of 1996 March (cf., Papers I and
781: II).  The \kapi\ values can be multiplied by the \hepr3\ abundance
782: ratios in Table 5 of Paper II to yield the \her3\ abundance.
783: \citet{RBB04} derive \her3\ abundances for all the \he3\ nebulae using
784: the entire 1982--1999 140 Foot dataset.  
785: 
786: {\it W3 ---} Although W3 is optically obscured it has been extensively
787: studied at radio wavelengths.  Our higher spatial resolution K-band data
788: give \yp4\ = 0.081 whereas the X-band data yield \yp4\ = 0.076.
789: Both data sets are of very high quality.  One interpretation of these
790: results is that the more compact components of W3---in particular
791: W3A---have a higher excitation than the outer regions probed by the
792: larger X-band beam.  This is sometimes called the geometric effect where
793: \yp4\ ratios increase with higher resolution observations.
794: 
795: Interferometers can probe the RRL emission of the nebular gas at high
796: spatial resolution.  Extant observations include: 76$\alpha$ and
797: 110$\alpha$ data from the Very Large Array (VLA) and the Westerbork
798: Synthesis Radio Telescope (WSRT) \citep{Roelfsema91} and 92$\alpha$ from
799: the VLA \citep{Roelfsema92, Adler96}.  The \hepr4\ abundance ratio
800: varies considerably among the different components ($0.06 \le $ \yp4\ $
801: \le 0.20$).  Values of \yp4\ higher than 0.10 are thought to be due to
802: local enrichment from a nearby evolved object.  Alternatively, radiative
803: transfer effects may cause one to overestimate the true \y4\ abundance
804: \citep{Gulyaev97}.  High \yp4\ values in other \hii\ regions are
805: associated with evolved early-type objects \citep[e.g.,][]{Balser01}.
806: W3 is rather unusual in our sample since the total mass of ionized
807: gas is small (20$\msun$; paper I) and therefore mass loss from local
808: objects can significantly pollute the surrounding material.  We
809: nevertheless adopt $\yp4 = 0.076$ as a reasonable value for most of the
810: mass in W3.  Because W3 has complex density structure and there is
811: uncertainty about the ionization structure, we adopt an ionization
812: correction of $\kappa_i \ge 1.32$.
813: 
814: {\it G133.8 ---} Also known as W3N, G133.8 is the northern most region
815: of the W3/W4 complex and has a slightly more negative velocity (by 7
816: \kms) than the other components of W3.  Infrared and radio continuum
817: observations of G133.8 indicate that this region is ionized by a
818: luminosity equivalent to an O7 star
819: \citep[Table~\ref{tab:excitation};][]{Schraml69, Thronson84, Carpenter00}.  
820: The X-band RRLs measure \yp4\ = 0.082, roughly consistent with the
821: ionization properties predicted from the continuum observations.  We
822: adopt an ionization correction of $\kappa_i = 1.22$.
823: 
824: {\it S206 ---} This outer Galaxy \hii\ region is thought to be ionized
825: by a single star with spectral type O4--O6.  Fabry-Perot
826: spectrophotometer data show that S206 contains essentially no neutral
827: helium with \yp4\ = 0.10 \citep{Deharveng00}.  Our X-band measurements
828: give a slightly lower value of \yp4\ = 0.092.  Recent high sensitivity
829: RRL data with the Green Bank Telescope yield \yp4\ = 0.085 \citep{Balser06}.  
830: Since we are correcting 140 Foot \he3\ data, however, we adopt
831: $\kappa_i = 1.09$.
832: 
833: {\it S209 ---} This \hii\ region has the largest Galactocentric
834: distance, $\rgal = 16.9$\kpc, in our entire sample and thus is an
835: important object.  Spectrophotometry reveals an O9 and two B1 stars
836: exciting S209 \citep{Chini84}; this suggests that S209 might contain
837: some neutral helium.  Optical \he4\ observations show that \yp4\ 
838: ranges from 0.084 \citep{Vilchez96} to 0.12 \citep{Deharveng00}.  The
839: \heal\ transition gives \yp4\ = 0.083.  We adopt \yp4\ = 0.083
840: producing an ionization correction of $\kappa_i = 1.20$.
841: 
842: {\it S212 ---} \hep3\ was not detected toward S212 and the \he4\ data
843: are of poor quality.  S212 is nonetheless a potentially important
844: object with $\rgal = 14.2$\kpc.  The nebula is ionized by an
845: early-type star, O5.5--O8.  Optical observations measure $\yp4 =
846: 0.088$--0.10; most of the helium should be ionized.  We adopt $\yp4 =
847: 0.088$ which gives an ionization correction of $\kappa_i = 1.14$.
848: 
849: {\it S228 ---} \hep3\ was not detected toward S228 and the \heal\ line
850: is only a probable detection.  As is the case for S212, this outer
851: Galaxy nebula is potentially a good \he3\ candidate.  Based on both
852: radio and optical data, however, this region is ionized by a late-type
853: O-star; the helium is probably not fully ionized within the \hii\
854: region.  Because \hep3\ and \hep4\ have not been detected we have a
855: direct measure of neither the \her3\ abundance nor the ionization
856: correction.
857: 
858: {\it S235, S76, S90 ---} These physically small, shell-like nebulae were
859: observed because they have a morphology that is similar to that of
860: W3\,A.  The anomalously high \her4\ abundance ratio in W3A might be
861: caused by local enrichment from nearby, evolved objects.  This might
862: also explain the high W3 \her3\ abundance ratio \citep{Olive95}.
863: 
864: \hep3\ was not detected in either S235 or S76.  Although we do not
865: have good quality \hep4\ data it is clear that the helium in these
866: nebulae is significantly under-ionized (Paper I).  Both radio and
867: optical data show that these two \hii\ regions are ionized by late
868: O-type or early B-type stars.  Given these soft radiation fields even
869: a high \her3\ abundance ratio would be difficult to detect.
870: 
871: In S90, both \hep3\ and \hep4\ are detected.  Although an O9.5 star is
872: optically identified near S90, \citet{Crampton78} argue that this star
873: does not excite the nebula. The radio continuum data suggest an earlier
874: type object, O8--O6.5, roughly consistent with the observed value of
875: $\yp4 = 0.070$.  We adopt an ionization correction factor of $\kapi =
876: 1.43$.
877: 
878: {\it S252 ---} Also known as \ngc{2175} or W13, this diffuse \hii\
879: region extends over 25\,\arcmin\ on the sky and contains several
880: compact components.  The nebula is ionized mainly by an O6--O6.5 star
881: centered on the \hii\ region \citep{Haikala95}.  Our observed position
882: lies on a bright rim located at the western edge of the nebula
883: \citep{Felli77}.  Because of the low emission measures in S252 there
884: is only a probable detection of \heal\ (Paper I).  \hep4\ is, however,
885: detected at optical wavelengths.  \citet{Shaver83} measured $\yp4 =
886: 0.051$ near our position.  Towards the center of the \hii\ region this
887: value increases with $\yp4 = 0.091$--0.10 \citep{Shaver83,
888: Deharveng00}.  We adopt an ionization correction factor of $\kappa_i =
889: 1.96$.
890: 
891: {\it S311 ---} S311 is thought to be ionized by an O6--O6.5 star
892: located north of the radio emission peak \citep{Persi87}.  The
893: asymmetric density distribution suggests a blister type geometry
894: \citep{Albert86}. \heal\ emission is detected: $\yp4 = 0.071$.  We
895: adopt an ionization correction factor of $\kappa_i = 1.41$.
896: 
897: {\it \sgr{B2} ---} \sgr{B2} is the most complex source in our sample.
898: It is important because it is located near the Galactic Center (GC).
899: The ionized gas is observed at high spatial resolution with radio
900: interferometers \citep[e.g.,][]{Martin72, Benson84, Gaume90, Gaume95}.
901: There are at least 57 separate, compact components associated with
902: \sgr{B2} \citep{DePree96}.
903: 
904: The \hepr4\ abundance in \sgr{B2} is known to be anomalous
905: \citep[see][and references within]{Lockman82}.  The radio continuum
906: emission implies a very high excitation but the \hepr4\ ratio is at
907: times reported to be below 0.06 (e.g., Table~\ref{tab:excitation}).  In
908: the literature the \sgr{B2} \her4\ abundance is reported as
909: underabundant, overabundant, or normal, depending on various effects
910: such as the geometric effect, selective photo-absorption by dust, or the
911: external maser effect.  These uncertainties are compounded by complex
912: spectral baselines produced by reflections from the telescope structure
913: for the single-dish observations.
914: 
915: Some insight can, however, be gleaned from high resolution
916: interferometric observations of RRL emission.  \sgr{B2} was observed
917: with the VLA in 76$\alpha$ emission \citep{Roelfsema87}; 110$\alpha$
918: emission \citep{Mehringer93}; and 66$\alpha$ emission \citep{DePree95,
919: DePree96}.  The \hepr4\ ratio varies from 0.05 to 0.17 among the various
920: compact components; the average value is 0.10 \citep{Roelfsema87,
921: DePree96}.  The component F bright cluster of sources is the sole
922: extremely low \hepr4\ ratio, $\yp4 \le 0.05$, component
923: \citep{DePree96}.
924: 
925: Most of the radio continuum flux density at X-band arises from the
926: compact components towards \sgr{B2} \citep{Balser95a}.  The \sgr{B2} F
927: components have the highest peak emission measures. Since the RRL
928: emission is proportional to emission measure, the \hep4\ RRL emission is
929: strongly influenced by the F components.  It is therefore not surprising
930: that, depending on frequency and spatial resolution, different \yp4\
931: values are measured in \sgr{B2}.
932: %
933: Because toward \sgr{B2} there is ionization structure coupled with very
934: complex density structure within the 3\farcm5 beam of the 140 Foot
935: telescope, using equation (1) is inappropriate.  For this source we can
936: only adopt a limit to the \her3\ abundance ratio.
937: 
938: {\it G1.1 ---} Better known as \sgr{D}, this \hii\ region is located
939: only 1\arcdeg\ from the GC but is much less complex than \sgr{B2}.
940: Although we have placed its location at the GC there is evidence
941: suggesting that it might be located outside the molecular nuclear disk
942: \citep{Lis91, Mehringer98, Blum99a}.  High resolution radio continuum
943: images reveal several compact components within a diffuse, extended
944: halo \citep{Liszt92, Mehringer98}.  A single O7 star can ionize the
945: core \citep{Liszt92}.  This is consistent with narrow band imaging at
946: 2.17\micron, Br$\gamma$, and 2.06\micron, and \hei\
947: \citep{Blum99}.  If the diffuse gas is included in the analysis an O4
948: star is required to account for the excitation
949: (Table~\ref{tab:excitation}).  The \hepr4\ ratio of 0.064 implies the
950: presence of some neutral helium in the nebula.  This is also
951: consistent with the infrared observations of \citet{Blum99}.
952: Initially G1.1 was classified as complex because of scant data;
953: additional RRL data supports a simple homogeneous model.  We therefore
954: adopt the ionization correction factor of $\kappa_ i = 1.56$.
955: 
956: {\it M17 ---} The Omega nebula, also known as M17, S45, \ngc{6618}, and
957: W28, is a well studied inner Galaxy \hii\ region.  The nebula spans over
958: 10\,\arcmin\ in angular size and consists of two main components: M17
959: north (M17N) and M17 south (M17S).  The Omega nebula is complex: high
960: resolution radio continuum images reveal this density structure directly
961: \citep{Felli84}.  We were unable to model the density structure in Paper
962: II.  The complex velocity field produces non-Gaussian profiles in both
963: M17N and M17S (Paper I; also see Joncas \& Roy 1986).
964: 
965: It is nonetheless clear that most of the helium is ionized.  High
966: extinction makes optical identification of the ionizing stars difficult,
967: but infrared spectrophotometry shows at least five O3-O6 stars
968: \citep{Hanson95}.  The large number of early O-type stars and the high
969: \hepr4\ abundance ratios observed at both radio and optical wavelengths
970: suggest that $\kappa_i \approx 1$.  Using the \heal\ transitions we
971: adopt an ionization correction factor of $\kapi = 1.09$ and 1.06 for
972: M17S and M17N, respectively.  
973: 
974: {\it W43 ---} W43 is a large, optically obscured \hii\ region located
975: only 4.6\kpc\ from the GC.  Because \hii\ regions are sparse between
976: the GC and 4.5\kpc\ \citep{Lockman96}, W43 probes the \her3\ abundance
977: near the outer edge of this \hii\ region desert.  W43 contains low
978: density, diffuse gas and many compact components
979: \citep[e.g.,][]{Subrahmanyan96, Balser01}.  Near infrared observations
980: show a star cluster at the center of the \hii\ region consisting of a
981: WN7 star and two O-type stars \citep{Blum99}.  The total H-ionizing
982: luminosity is considerable, requiring at least an O4 star for the
983: entire region (Table~\ref{tab:excitation}).
984: 
985: We studied the helium ionization properties of both the diffuse and
986: compact gas in W43 using the 140 Foot telescope and the VLA
987: \citep{Balser01}.  Using RRL emission we imaged the \hepr4\ abundance
988: ratio for both the compact and diffuse components.  The average
989: \hepr4\ ratio of the entire dataset is $\langle\yp4\rangle = 0.077 \pm
990: 0.01$.  We find no significant variations of the \hepr4\ abundance,
991: even for positions observed at the edge of the nebula with the 140
992: Foot.  Thus although W43 is a complex nebula, the ionization
993: properties do not appear to vary within this object so we can derive
994: an accurate ionization correction factor.  We adopt $\kappa_i = 1.30$
995: based on the average \yp4\ abundance.
996: 
997: {\it W49 ---} W49 is one of the most luminous \hii\ regions in the
998: Galaxy, consisting of a ring of ultra compact \hii\ regions (UC \hii\
999: regions), each containing at least one O-type star
1000: \citep[e.g.,][]{Welch87}.  These UC \hii\ regions make up only
1001: one-third of the total H-ionizing luminosity observed \citep{Conti02}.
1002: %
1003: The \heal\ RRL gives a \hepr4\ abundance ratio of $0.079 \pm 0.0052$,
1004: consistent with the K-band data.  Higher resolution VLA H and He RRL
1005: images find \yp4\ ranging from $0.05 \pm 0.02$ to $0.18 \pm 0.06$ in
1006: 13 objects with an average of $\langle\yp4\rangle = 0.11 \pm 0.01$
1007: \citep{DePree97}.  This significant ionization structure combined with
1008: considerable density structure makes any equation (\ref{eq:defkap})
1009: type ionization correction inappropriate.  We therefore adopt
1010: $\kappa_i \ge\ 1$.
1011: 
1012: {\it W51 ---} W51 is a bright, complex \hii\ region with an angular
1013: extent of 2\,\arcdeg\ on the sky.  Radio continuum observations reveal
1014: several O4--O6 stars associated with the main components
1015: \citep{Mehringer94}.  Mid-infrared imaging and spectroscopy suggests
1016: stars with spectral types $\sim$\,O9 \citep{Okamoto01}.  The X-band
1017: \hep4\ RRL data give a \yp4\ abundance of $0.079 \pm 0.0024$.  The
1018: \healk\ transition appears to be slightly lower with \yp4\ $\sim
1019: 0.07$.  High resolution VLA observations of 92$\alpha$ RRLs yield
1020: \yp4\ values ranging from $0.08 \pm 0.02$ to $0.14 \pm 0.06$ with
1021: an average of $\langle\yp4\rangle = 0.096 \pm 0.015$ \citep{Mehringer94}.
1022: Although most of the helium appears to be ionized, given the
1023: complexity of this object we adopt an ionization correction factor of
1024: $\kappa_i \ge\ 1$.
1025: 
1026: {\it S156 ---} S156 is a compact \hii\ region excited by an O6.5-type
1027: star.  We have only an upper limit for the \hepr3\ abundance.  Although
1028: we have detected \hep4\ at X-band the accuracy is not very good.
1029: Optical observations yield $\yp4 = 0.081$ \citep{Deharveng00}.  We adopt
1030: an ionization correction factor of $\kappa_i = 1.23$. 
1031: 
1032: {\it \ngc{7538} ---} Also known as S158, \ngc{7538} is centrally diffuse
1033: with a bright rim to the west \citep{Israel77}.  An O7 star located at
1034: the center of the nebula is responsible for the excitation
1035: \citep{Deharveng79}.  Optical data yield $\yp4 = 0.10$ \citep{Lynds86}, a
1036: value higher than our RRL results.  Both the X-band and K-band data give
1037: \yp4\ $\sim 0.08$, although the \hebet\ transition produces a slightly
1038: higher value of 0.10.  We adopt an ionization correction factor based 
1039: on the X-band data of $\kappa_i = 1.19$.
1040: 
1041: {\it S162 ---} Sometimes called the Bubble nebula, S162 consists of a
1042: circularly shaped shell with bright emission to the north observed in
1043: H$\alpha$ and the radio continuum \citep[Paper I]{Barlow76, Israel73}.
1044: The exciting star has been classified as an O6.5-type, consistent with
1045: the radio continuum estimates.  We do not detect \hep3\ emission.
1046: Although we do detect the \heal\ RRL, the quality factor is poor.  We
1047: cannot make an ionization correction so therefore adopt $\kappa_i \ge
1048: 1$.
1049: 
1050: \section{DISCUSSION}\label{sec:discuss}
1051: 
1052: Our nebular ionization correction, \kapi\,, is the simplest possible
1053: approach to the problem.  It makes several key assumptions: (1) the
1054: \he4\ abundance is \her4\ $\equiv 0.10$ for all Galactic \hii\ regions;
1055: (2) all the He is singly ionized---there is neither neutral nor doubly
1056: ionized He within the nebulae; and (3) the \hep3\ and \hep4\ ions
1057: occupy identical volumes within the nebulae.  Measuring the total
1058: helium abundance is difficult since there is no direct way to observe
1059: neutral helium.  We do not expect any significant doubly ionized
1060: helium since the radiation field from Galactic O-type stars is not
1061: hard enough.  For example, M17 is ionized by several stars classified
1062: between O3--O6 \citep{Hanson95} and is one of the Galactic \hii\
1063: regions with the highest degree of ionization.  Yet only a very low
1064: upper limit, $^3{\rm He}^{++}/{\rm H}^+ < 8 \times 10^{-5}$, is measured for 
1065: \hepp4\ in M17 \citep{Peimbert92b}.  In fact, to our knowledge \hepp4\ 
1066: has never been detected in any Galactic \hii\ region.
1067: 
1068: A canonical value for the Galactic \he4\ abundance of $\y4 = 0.1$ is
1069: typically used in the literature.  Most of the \he4\ was produced
1070: during the era of primordial nucleosynthesis.  Therefore the variation
1071: of \he4\ in the Galaxy, caused by stellar and Galactic evolution, is
1072: expected to be small because $\Delta Y / \Delta Z \sim 1$, where $Y$
1073: and $Z$ are the helium and metal abundances by mass
1074: \citep[e.g.,][]{Chiappini02}.  Direct measurement of the total \her4\
1075: abundance ratio in the Galaxy is difficult. The \he4\ abundance cannot
1076: be determined using either the Solar photosphere or in meteorites.
1077: Measurements of \he4\ in Galactic \hii\ regions have to correct for
1078: ionization structure just as we are trying to do here for \he3.
1079: 
1080: Solar \he4\ abundances are determined from theoretical stellar
1081: evolution models with \y4\ $\sim 0.1$ \citep[e.g.,][]{Anders89}.
1082: \citet{Peimbert93} suggested that the high excitation \hii\ region M17
1083: was the best object to measure \he4\ directly and they also determined 
1084: $\y4 \sim 0.1$.  There is mounting evidence, however, that \y4\ is
1085: less than 0.1 in the Galaxy.  More recent calibration of the \he4\
1086: Solar abundances yield \y4\ values lower by as much as 10\%
1087: \citep{Grevesse96, Grevesse98, Basu04, Asplund05}.  Recent estimates
1088: of \y4\ in M17 that include temperature fluctuations reduce \y4\ by
1089: about 5\%, consistent with improved RRL data \citep{Peimbert02,
1090: Quireza06}.  Furthermore, RRL observations of the lower metallicity
1091: Galactic \hii\ region S206, which is also expected to contain little
1092: or no neutral helium, give \y4 = 0.085 \citep{Balser06}.
1093: 
1094: Nevertheless, we adopt $\y4 \equiv y_{\rm 4GAL} = 0.1$ in equation (1)
1095: for our ionization correction factor analysis.  We judge that for our
1096: simple sources the ionization correction should be good to about 10\%.
1097: At this level of accuracy all of the additional ionization effects 
1098: considered by \he4\ abundance analyses \citep[e.g.,][]{Izotov07} 
1099: can be neglected.  
1100: 
1101: \section{SUMMARY}\label{sec:summary}
1102: 
1103: The significance of our results for both primordial \he3\ and chemical
1104: evolution of \he3\ lies in the fact that the abundances we find are
1105: quite low. Hence we have paid particular attention to factors that
1106: might lead us to measure a lower abundance than that which is actually
1107: present.  In this paper we conclude that it is unlikely that large
1108: ionization corrections are required for our high excitation state
1109: \he3\ nebulae.  We use the \yp4\ abundance to select for high
1110: excitation \hii\ regions.  We determined an ionization correction
1111: factor for the Paper I nebular sample. There would have to be a
1112: conspiracy between the density and ionization structure for these
1113: ionization corrections to be underestimated by a substantial amount.
1114: %
1115: Our synthetic nebular models (\S\ref{sec:models}) do show that large
1116: ionization corrections are possible.  Such models, however, require
1117: that there be rather specific density and ionization structure
1118: configurations.  Our simple ionization correction fails when \yp4\
1119: changes significantly between the low density gas where \hep3\
1120: dominates the emission and the high density gas where \hep4\
1121: dominates.  The more complex the source the larger our potential error
1122: in determining the ionization correction.  We find both numerically
1123: and observationally that the simple sources give the best ionization
1124: corrections (\S\ref{sec:kappa}).
1125: 
1126: For the sources that we classify as simple and which also have well
1127: determined line parameters, the ionization correction given in
1128: \S\ref{sec:kappa} should be accurate. Indeed, the major uncertainty is
1129: likely to be the value adopted for $y_{\rm 4GAL}$.  Even for more complex
1130: sources our simple ionization correction should be reasonable for any
1131: nebula that does not have large fluctuations in excitation. We have
1132: shown this to be the case for the complex nebula W43 where we have
1133: single-dish and interferometry data for both radio recombination line
1134: and continuum emission (\S\ref{sec:kappa}).
1135: 
1136: For a subset of our \he3\ sources, we use both 8\,GHz and 18\,GHz RRL
1137: data to derive \yp4.  The two sets of observations probe different
1138: regions of the nebula, yet we in general find very similar values for
1139: \yp4.  This is consistent with the notion that for our sources
1140: selected for high excitation there are not large ionization
1141: corrections.  
1142: 
1143: We can use this simple ionization correction analysis because unlike
1144: many of the light elements \her3\ abundances accurate to $\sim 10\%$
1145: can yield important conclusions for both the primordial \he3\ and the
1146: chemical evolution of \he3\ \citep{Wilson94}.  This is in stark
1147: contrast to the situation for \he4\ where the \her4\ abundance needs
1148: to be determined to an accuracy of $\lsim 1$\% if one is to draw
1149: cosmologically significant inferences \citep[e.g.,][]{Izotov07}.
1150: Furthermore, because the majority of our \he3\ nebulae are not
1151: optically visible it is not likely that we will ever have sufficiently
1152: detailed information about their physical properties to make a
1153: significally better correction.
1154: 
1155: \acknowledgments{We thank the staff of NRAO Green Bank for their help,
1156: support, and friendship.  The \he3\ research has been sporadically
1157: supported by the National Science Foundation (AST 97-31484; AST
1158: 00-98047; AST 00-98449).  This research has made use of NASA's
1159: Astrophysics Data System and the SIMBAD database, operated at CDS,
1160: Strasbourg, France.}
1161: 
1162: %\clearpage
1163: 
1164: \begin{thebibliography}{}
1165: 
1166: \bibitem[Adler et al.(1996)]{Adler96}
1167: Adler, D. S., Wood, D. O. S., \& Goss, W. M. 1996, \apj,
1168: 471, 871
1169: 
1170: \bibitem[Albert et al.(1986)]{Albert86}
1171: Albert, C.~E., Schwartz, P.~R., Bowers, P.~F., \&
1172: Rickard, L.~J. 1986, \aj, 92, 75
1173: 
1174: \bibitem[Anders \& Grevesse (1989)] {Anders89}
1175: Anders, E., \& Grevesse, N. 1989, Geochim. Cosmochim. Acta, 53, 197
1176: 
1177: \bibitem[Armour et al.(1999)] {Armour99}
1178: Armour, M., Ballantyne, D.~R., Ferland, G.~J., Karr, J., \& Marin,
1179: P.~G. 1999, \pasp, 111, 1251
1180: 
1181: \bibitem[Asplund et al.(2005)] {Asplund05}
1182: Asplund, M., Grevesse, N., \& Sauval, A.~J. 2005, in ASP Conf. Ser. 
1183: 336, Cosmic Abundances as Records of Stellar Evolution and Nucleosynthesis,
1184: ed. T.~G. Barnes III \& F.~N. Bash (San Francisco: ASP) 25
1185: 
1186: \bibitem[Baldwin et al.(1991)]{Baldwin91} 
1187: Baldwin, J.~A., Ferland, G.~J., Martin, P.~G., Corbin, M.~R.,
1188: Cota, S.~A., Peterson, B.~M., \& Slettebak, A.\ 1991, \apj, 374, 580
1189: 
1190: \bibitem[Ballantyne et al.(2000)] {Ballantyne00}
1191: Ballantyne, D.~R., Ferland, G.~J., \& Martin, P.~G. 2000, \apj, 536, 773
1192: 
1193: \bibitem[Balser(1995)]{Balser95} 
1194: Balser, D.~S.\ 1995, PhD dissertation, Boston University
1195: 
1196: \bibitem[Balser(2006)] {Balser06}
1197: Balser, D.~S. 2006, \aj, 132, 2326
1198: 
1199: \bibitem[Balser et al.(1994)]{Balser94} 
1200: Balser, D.~S., Bania, T.~M., Brockway, C.~J., Rood, R.~T., 
1201: \& Wilson, T.~L. 1994, \apj, 430, 667
1202: 
1203: \bibitem[Balser et al.(1995)]{Balser95a}
1204: Balser, D. S., Bania, T. M., Rood, R. T., \& Wilson, T. L.
1205: 1995, \apjs, 100, 371
1206: 
1207: \bibitem[Balser et al.(1997)]{BBRW97} 
1208: Balser, D.~S., Bania, T.~M., Rood, R.~T., \& Wilson, T.~L.\ 1997, 
1209: \apj, 483, 320 
1210: 
1211: \bibitem[Balser et al.(1999)]{BBRW99} 
1212: Balser, D.~S., Bania, T.~M., Rood, R.~T., \& Wilson, T.~L.\ 1999, 
1213: \apj, 510, 759 (Paper II)
1214: 
1215: \bibitem[Balser et al.(2006)]{balser06}
1216: Balser, D.~S., Goss, W.~M., Bania, T.~M, \& Rood, R.~T. 2006, \apj,
1217: 640, 360
1218: 
1219: \bibitem[Balser et al.(2001)]{Balser01}
1220: Balser, D.~S., Goss, W. M., \& De Pree, C. G. 2001, \apj, 121, 371
1221: 
1222: \bibitem[Bania et al.(1997)]{BBRWW97} Bania, T.~M., Balser, D.~S., Rood, R.~T., 
1223: Wilson, T.~L., \& Wilson, T.~J.\ 1997, \apjs, 113, 353 (Paper I)
1224: 
1225: \bibitem[Bania et al.(2002)]{BRB02} Bania, T.~M., Rood, R.~T., 
1226: \& Balser, D.~S.\ 2002, \nat, 415, 54 
1227: 
1228: \bibitem[Barlow et al.(1976)]{Barlow76}
1229: Barlow, M.~J., Cohen, M., Gull, T.~R. 1976, \mnras, 176, 359
1230: 
1231: \bibitem[Benson \& Johnston(1984)]{Benson84}
1232: Benson, J.~M., \& Johnston, K.~J. 1984, \apj, 277, 181
1233: 
1234: \bibitem[Basu \& Anita (2004)] {Basu04}
1235: Basu, S., \& Anita, H. M. 2004, \apj, 606, L85
1236: 
1237: \bibitem[Blagrave et al.(2007)] {Blagrave06}
1238: Blagrave, K.~P.~M., Martin, P.~G., Rubin, R.~J., Dufour, R.~J.,
1239: Baldwin, J.~A., Hester, J.~J., \& Walter, D.~K. 2007, \apj, 655, 299
1240: 
1241: \bibitem[Blum \& Damineli(1999)]{Blum99a}
1242: Blum, R.~D., \& Damineli, A. 1999, \apj, 512, 237
1243: 
1244: \bibitem[Blum et al.(1999)]{Blum99}
1245: Blum, R.~D., Damineli, A., \& Conti, P.~S. 1999, \aj,
1246: 117, 1392
1247: 
1248: \bibitem[Caplan et al.(2000)]{Caplan00}
1249: Caplan, J., Deharveng, L., Pe$\tilde{\rm n}$a, M., Costero,
1250: R., \& Blondel, C. 2000, \mnras, 311, 317
1251: 
1252: \bibitem[Carpenter et al.(2000)]{Carpenter00}
1253: Carpenter, J.~M., Heyer, M.~H., \& Snell, R.~L. 2000, \apjs, 130, 381
1254: 
1255: \bibitem[Charbonnel(1998)]{Charbonnel98} 
1256: Charbonnel, C.\ 1998, Space Science Reviews, 84, 199 
1257: 
1258: \bibitem[Chiappini et al. (2002)] {Chiappini02}
1259: Chiappini, C., Renda, A., \& Matteucci, F. 2002, \aap, 395, 789
1260: 
1261: \bibitem[Chini \& Wink(1984)]{Chini84}
1262: Chini, R., \& Wink, J.~E. 1984, \aap, 139, L5
1263: 
1264: \bibitem[Conti \& Alschuler(1971)]{Conti71}
1265: Conti, P.~S., \& Alschuler, W.~R. 1971, \apj, 170, 325
1266: 
1267: \bibitem[Conti \& Leep(1974)]{Conti74}
1268: Conti, P.~S., \& Leep, E.~M. 1974, \apj, 193, 113
1269: 
1270: \bibitem[Conti \& Blum(2002)]{Conti02}
1271: Conti, P.~S., \& Blum, R.~D. 2002, \apj, 564, 827
1272: 
1273: \bibitem[Crampton(1971)]{Crampton71}
1274: Crampton, D. 1971, \aj, 76, 260
1275: 
1276: \bibitem[Crampton et al.(1978)]{Crampton78}
1277: Crampton, D., Georgelin, Y.~M., \& Georgelin, Y.~P. 1978,
1278: \aap, 66, 1
1279: 
1280: \bibitem[Deharveng et al.(1979)]{Deharveng79}
1281: Deharveng, L., Lortet, M.~C., \& Testor, G. 1979, \aap,
1282: 71, 151
1283: 
1284: \bibitem[Deharveng et al.(2000)]{Deharveng00}
1285: Deharveng, L., Pe$\tilde{\rm n}$a, M., Caplan, J., \& Costero,
1286: R. 2000, \mnras, 311, 329
1287: 
1288: \bibitem[DePree et al.(1995)]{DePree95}
1289: De Pree, C.~G., Gaume, R.~A., Goss, W.~M., \& Claussen,
1290: M.~J. 1995, \apj, 451, 284
1291: 
1292: \bibitem[DePree et al.(1996)]{DePree96}
1293: De Pree, C.~G., Gaume, R.~A., Goss, W.~M., \& Claussen,
1294: M.~J. 1996, \apj, 464, 788
1295: 
1296: \bibitem[DePree et al.(1997)]{DePree97}
1297: De Pree, C.~G., Mehringer, D.~M., Goss, W.~M 1997, \apj, 482, 307
1298: 
1299: \bibitem[Esteban et al. (1998)] {Esteban98}
1300: Esteban, C., Peimbert, M., Torres-Peimbert, S., \&
1301: Escalante, V.  1998, \mnras, 295, 401
1302: 
1303: \bibitem[Esteban et al.(1999)]{Esteban99}
1304: Esteban, C., Peimbert, M., Torres-Piembert, S., \&
1305: Garc\'{i}a-Rojas 1999, Rev. Mexicana de Astron. y Astrofis., 35, 65
1306: 
1307: \bibitem[Feinstein \& V\'{a}zquez(1989)]{Feinstein89}
1308: Feinstein, A. \& V\'{a}zquez, R.~A. 1989, A\&AS, 77, 321
1309: 
1310: \bibitem[Felli et al.(1977)]{Felli77}
1311: Felli, M., Habing, H.~J., \& Israel, F.~P. 1977, \aap, 59, 43
1312: 
1313: \bibitem[Felli et al.(1984)] {Felli84}
1314: Felli, M., Massi, M., \& Churchwell, E. 1984, \aap, 136, 53
1315: 
1316: \bibitem[Fich \& Silkey(1991)]{Fich91}
1317: Fich, M., \& Silkey, M. 1991, \apj, 366, 107
1318: 
1319: \bibitem[Gaume \& Claussen(1990)]{Gaume90}
1320: Gaume, R.~A., \& Claussen, M.~J. 1990, \apj, 351, 538
1321: 
1322: \bibitem[Gaume et al.(1995)]{Gaume95}
1323: Gaume, R.~A., Claussen, M.~J., De Pree, C.~G., Goss,
1324: W.~M., \& Mehringer, D.~M. 1995, \apj, 449, 663
1325: 
1326: \bibitem[Georgelin(1975)]{Georgelin75}
1327: Georgelin, Y.~M. 1975, Th\`{e}se d$^\prime$\'{E}tat,
1328: Univ. de Provence (Univ. d$^\prime$Aix-Marseille I)
1329: 
1330: \bibitem[Georgelin et al.(1973)]{Georgelin73}
1331: Georgelin, Y. M., Georgelin, Y. P., \& Roux, S. 1973,
1332: \aap, 25, 337
1333: 
1334: \bibitem[Grevesse et al.(1996)] {Grevesse96}
1335: Grevesse, N., Noels, A., \& Sauval, A.~J. 1996, in ASP Conf. Ser. 99,
1336: Cosmic Abundances, ed. S. S. Holt \& G. Sonneborn (San Francisco: ASP),
1337: 117
1338: 
1339: \bibitem[Grevesse et al.(1998)] {Grevesse98}
1340: Grevesse, N., Sauval, A.~J. 1998, Space Sci. Rev., 85, 161
1341: 
1342: \bibitem[Gruenwald et al.(2002)] {Gruenwald02}
1343: Gruenwald. R., Steigman, G., \& Viegas, S.~M 2002, \apj, 567, 931
1344: 
1345: \bibitem[Goy(1980)]{Goy80}
1346: Goy, G. 1980, A\&AS, 1980, 42, 91
1347: 
1348: \bibitem[Gulyaev et al.(1997)] {Gulyaev97}
1349: Gulyaev, S. A., Sorochenko R. L., \& Tsivilev, A. P. 1997,
1350: Astronomy Letters, 23, 165
1351: 
1352: \bibitem[Haikala(1995)]{Haikala95}
1353: Haikala, L.~K. 1995, \aap, 294, 89
1354: 
1355: \bibitem[Hanson \& Conti(1995)]{Hanson95}
1356: Hanson, M.~M., \& Conti, P.~S. 1995, \apj, 448, L45
1357: 
1358: \bibitem[Hunter \& Massey(1990)]{Hunter90}
1359: Hunter, D.~A., \& Massey, P. 1990, \aj, 99, 846
1360: 
1361: \bibitem[Israel(1977)]{Israel77}
1362: Israel, F.~P. 1977, \aap, 59, 27
1363: 
1364: \bibitem[Israel et al.(1973)]{Israel73}
1365: Israel, F.~P., Habing, H.~J., \& de Jong, T. 1973, \aap, 27, 143
1366: 
1367: \bibitem[Izotov et al.(2007)]{Izotov07}
1368: Izotoz, Y.~I., Thuan, T.~X., \& Stasi\'nska, G. 2007, 
1369: astro-ph/0702072
1370: 
1371: \bibitem[Joncas \& Roy(1986)] {Joncas86}
1372: Joncas, G., \& Roy, J.-R. 1986, \apj, 307, 649
1373: 
1374: \bibitem[Lockman \& Brown(1982)]{Lockman82}
1375: Lockman, F.~J., \& Brown, R.~L. 1982, \apj, 259, 595
1376: 
1377: \bibitem[Lockman et al.(1996)]{Lockman96}
1378: Lockman, F.~J., Pisano, D.~J., \& Howard, G.~J. 1996,
1379: \apj, 472, 173
1380: 
1381: \bibitem[Lis(1991)]{Lis91}
1382: Lis, D.~C. 1991, \apj, 379, L53
1383: 
1384: \bibitem[Liszt(1992)]{Liszt92}
1385: Liszt, H.~S. 1992, \apjs, 82, 495
1386: 
1387: \bibitem[Lynds \& O'Neil, Jr.(1986)]{Lynds86}
1388: Lynds, B.~T., \& O'Neil Jr., E.~J. 1986, \apj, 306, 532
1389: 
1390: \bibitem[Martin \& Downes(1972)]{Martin72}
1391: Martin, A.~H.~M., \& Downes, D. 1972, Astrophysical Letters, 11, 219
1392: 
1393: \bibitem[Mathis (1982)] {Mathis82}
1394: Mathis, J.~S. 1982, \apj, 261, 195
1395: 
1396: \bibitem[Mathis (1985)] {Mathis85}
1397: Mathis, J.~S. 1985, \apj, 291, 247
1398: 
1399: \bibitem[Mathis \& Rose(1991)]{Mathis91} 
1400: Mathis, J.~S., \& Rosa, M.~R.\ 1991, \aap, 245, 625
1401: 
1402: \bibitem[Mathis \& Wood(2005)] {Mathis05}
1403: Mathis, J.~S., \& Wood, K. 2005, \mnras, 360, 227
1404: 
1405: \bibitem[Mehringer(1994)]{Mehringer94}
1406: Mehringer, D.~M. 1994, \apjs, 91, 713
1407: 
1408: \bibitem[Mehringer et al.(1993)]{Mehringer93}
1409: Mehringer, D.~M., Palmer, P., Goss, W.~M., \&
1410: Yusef-Zadeh, F. 1993, \apj, 412, 684
1411: 
1412: \bibitem[Mehringer et al.(1998)]{Mehringer98}
1413: Mehringer, D.~M., Goss, W.~M, Lis, D.~C., Palmer, P., \&
1414: Menten, K.~M. 1998, \apj, 493, 274
1415: 
1416: \bibitem[Miller(1968)]{Miller68}
1417: Miller, J.~S., 1968, \apj, 151, 473
1418: 
1419: \bibitem[Moffat et al.(1979)]{Moffat79}
1420: Moffat, A.~F.~J., Fitzgerald, M.~P., \& Jackson,
1421: P.~D. 1979, \aaps, 38, 197
1422: 
1423: \bibitem[Moreno \& Chavarr\'{i}a-K.(1986)]{Moreno86}
1424: Moreno, M.~A., \& Chavarr\'{i}a-K., C. 1986, \aap, 161, 130
1425: 
1426: \bibitem[Okamoto et al.(2001)]{Okamoto01}
1427: Okamoto, Y.~K., Kataza, H., Yamashita, T., Miyata, T.,
1428: \& Onaka, T. 2001, \apj, 553, 254
1429: 
1430: \bibitem[Olive et al.(1995)]{Olive95}
1431: Olive, K.~A., Rood, R.~T., Schramm, D.~N., Truran, J.,
1432: \& Vangioni-Flam, E. 1995, \apj, 444, 680
1433: 
1434: \bibitem[Panagia(1973)]{Panagia73}
1435: Panagia, N. 1973, \aj, 78, 929
1436: 
1437: \bibitem[Peimbert(1993)] {Peimbert93}
1438: Peimbert, M. 1993, Rev. Mexicana Astron. Astrofis., 27, 9
1439: 
1440: \bibitem[Peimbert \& Peimbert(2002)]{Peimbert02}
1441: Peimbert, M., \& Peimbert, A. 2002, Rev. Mexicana
1442: Astron. Astrofis. Ser. Conf., 14, 47
1443: 
1444: \bibitem[Peimbert et al.(2002)]{Peimbert02a}
1445: Peimbert, A., Peimbert, M. \& Luridiana, V. 2002, \apj, 565, 668
1446: 
1447: \bibitem[Peimbert et al.(1992a]{Peimbert92a}
1448: Peimbert, M., Rodr\'{i}guez, L. F., Bania, T.~M., Rood,
1449: R.~T., \& Wilson, T.~L. 1992a, \apj, 395, 484
1450: 
1451: \bibitem[Peimbert et al.(1992b)]{Peimbert92b}
1452: Peimbert, M., Torres-Piembert, S., \& Ruiz, M.~T. 1992b,
1453: Rev. Mexicana Astron. Astrofis., 24, 155
1454: 
1455: \bibitem[Peng et al. (2000)] {Peng00}
1456: Peng, B., Kraus, A., Krichbaum, T. P., \& Witzel, A. 2000, A\&AS, 145,
1457: 1
1458: \bibitem[Persi et al.(1987)]{Persi87}
1459: Persi, P., Ferrari-Toniolo, Shivanandan, K., \&
1460: Spinoglio, L. 1987, A\&AS, 70, 437
1461: 
1462: \bibitem[Pogge et al. (1992)] {Pogge92}
1463: Pogge, R. W., Owen, J. M., \& Atwood, B. 1992, \apj, 399, 147
1464: 
1465: \bibitem[Quireza et al.(2006a)]{Quireza06}
1466: Quireza, C., Rood, R.~T., Balser, D.~S., \& Bania, T.~M., 2006a, \apjs,
1467: 165, 338
1468: 
1469: \bibitem[Quireza et al.(2006b)]{Quireza07}
1470: Quireza, C., Rood, R.~T., Bania, T.~M., Balser, D.~S. \& 
1471: Maciel, W.~J., 2006b, 653, 1226
1472: 
1473: \bibitem[Roelfsema et al.(1987)]{Roelfsema87}
1474: Roelfsema, P.~R., Goss, W.~M., Whiteoak, J.~B., Gardner,
1475: F.~F., \& Pankonin, V. 1987, \aap, 175, 219
1476: 
1477: \bibitem[Roelfsema \& Goss(1991)]{Roelfsema91}
1478: Roelfsema, P.~R., \& Goss, W.~M. 1991, A\&AS, 87, 177
1479: 
1480: \bibitem[Roelfsema et al.(1992)]{Roelfsema92}
1481: Roelfsema, P.~R., Goss, W.~M., \& Mallik, D.~C.~V. 1992,
1482: \apj, 394, 188
1483: 
1484: \bibitem[Romano et al.(2003)]{Romano03}
1485: Romano, D., Tosi, M., Matteucci, F., \& Chiappini, C.\ 2003, 
1486: \mnras, 346, 295
1487: 
1488: \bibitem[Rood et al.(2007)]{RBB04} 
1489: Rood, R.~T., Bania, T.~M., Balser, D.~S., \& Wilson, T.~L.\ 
1490: 2007, \apj, in preparation (Paper IV) 
1491: 
1492: \bibitem[Rubin(1984)]{Rubin84} 
1493: Rubin, R.~H.\ 1984, \apj, 287, 653
1494: 
1495: \bibitem[Rubin et al. (1991)] {Rubin91}
1496: Rubin, R. H., Simpson, J. P., Haas, M. R., Erickson, E. F.
1497: 1991, \apj, 374, 564
1498: 
1499: \bibitem[Sauer \& Jedamzik (2002)] {Sauer02}
1500: Sauer, D., \& Jedamzik, K. 2002, \aap, 381, 361
1501: 
1502: \bibitem[Schraml \& Mezger(1969)]{Schraml69}
1503: Schraml, J, \& Mezger, P.~G. 1969, \apj, 156, 269
1504: 
1505: \bibitem[Shaver et al.(1983)]{Shaver83}
1506: Shaver, P.~A., McGee, R.~X., Newton, L.~M., Danks, A.~C., \&
1507: Pottasch, S.~R. 1983, MNRAS, 204, 53
1508: 
1509: \bibitem[Shields \& Searle (1978)] {Shields78}
1510: Shields, G.~A., \& Searle, L. 1978, \apj, 222, 821
1511: 
1512: \bibitem[Subrahmanyan \& Goss(1996)]{Subrahmanyan96}
1513: Subrahmanyan, R. \& Goss, W.~M. 1996, \mnras, 281, 239
1514: 
1515: \bibitem[Thronson et al.(1984)]{Thronson84}
1516: Thronson, Jr., H.~A., Schwartz, P.~R., Smith, H.~A.,
1517: Lada, C.~J., Glaccum, W., Harper, D.~A. 1984, \apj, 284, 597
1518: 
1519: \bibitem[Vacca et al.(1996)]{Vacca96} 
1520: Vacca, W.~D., Garmany, C.~D., \& Shull, J.~M.\ 1996, \apj, 460, 914
1521: 
1522: \bibitem[Viegas et al.(2000)] {Viegas00}
1523: Viegas, S.~M., Gruenwald, R., \& Steigman, G. 2000, \apj, 531, 813
1524: 
1525: \bibitem[V\'{i}lchez \& Esteban(1996)]{Vilchez96}
1526: V\'{i}lchez, J.~M., \& Esteban, C. 1996, \mnras, 280, 720
1527: 
1528: \bibitem[V\'{i}lchez \& Pagel(1988)]{Vilchez88}
1529: V\'{i}lchez, J.~M., \& Pagel, B.~E.~J. 1988, \mnras, 231, 257
1530: 
1531: \bibitem[Wilson \& Rood(1994)]{Wilson94} 
1532: Wilson, T.~L., \& Rood, R.~T. 1994,\ \araa, 32, 191
1533: 
1534: \bibitem[Welch et al.(1987)]{Welch87}
1535: Welch, W.~J., Dreher, J.~M., Jackson, J.~M., Terebey, S., \&
1536: Vogel, S.~N. 1987, Science, 238, 1550
1537: 
1538: \bibitem[Yang et al.(1984)]{Yang84}
1539: Yang, J., Turner, M.~S., Steigman, G., Schramm, D.~N., \&
1540: Olive, K.~A. 1984, \apj, 281, 493
1541: 
1542: \end{thebibliography}
1543: 
1544: \clearpage
1545: 
1546: %
1547: % Include tables
1548: %  
1549: \begin{deluxetable}{llrrrcrcc}
1550: \tablecolumns{9}
1551: \tablecaption{K-band (18\ghz) Radio Recombination Line Parmeters
1552: \label{tab:data}}
1553: \tablehead{
1554: \colhead{}  & \colhead{} &
1555: \colhead{$T_{\rm L}$}  & \colhead{$\sigma(T_{\rm L})$} &
1556: \colhead{$\Delta v$}  & \colhead{$\sigma(\Delta v)$} &
1557: \colhead{$t_{\rm intg}$}  & \colhead{$RMS$} & Quality \\
1558: \colhead{Source}  & \colhead{Transition} &
1559: \colhead{(mK)}  & \colhead{(mK)} &
1560: \colhead{(${\rm km}\,{\rm s}^{-1}$)}  & \colhead{(${\rm km}\,{\rm s}^{-1}$)} &
1561: \colhead{(hr)}  & \colhead{(mK)} & Factor
1562: }
1563: \startdata
1564: 
1565: W3.....
1566:  & \halk\   & 579.63 &    0.76 &   27.44 &    0.04 &   11.4 &    0.97 & A \\ 
1567:  & \healk\  &  60.85 &    0.88 &   20.94 &    0.36 &   11.4 &    0.97 & B \\ 
1568:  & \hbetk\  & 160.73 &    0.31 &   27.80 &    0.06 &   13.0 &    0.61 & A \\ 
1569:  & \hebetk\ &  17.30 &    0.36 &   21.21 &    0.53 &   13.0 &    0.61 & B \\ 
1570: 
1571: G133.8.....                                 
1572:  & \halk\   &  64.99 &    0.47 &   30.06 &    0.25 &   23.6 &    0.48 & B \\
1573:  & \healk\  &   6.04 &    0.55 &   22.94 &    2.56 &   23.6 &    0.48 & C \\ 
1574:  & \hbetk\  &  19.63 &    0.24 &   30.14 &    0.42 &   22.2 &    0.51 & C \\ 
1575:  & \hebetk\ &   1.41 &    0.36 &   13.01 &    4.05 &   22.2 &    0.51 & D \\ 
1576: 
1577: S206.....
1578:  & \halk\   &  29.24 &    0.15 &   27.32 &    0.16 &   26.0 &    0.50 & B \\ 
1579:  & \healk\  &   3.19 &    0.16 &   21.89 &    1.33 &   26.0 &    0.50 & C \\ 
1580:  & \hbetk\  &   8.29 &    0.10 &   26.61 &    0.38 &   26.0 &    0.43 & C \\ 
1581: 
1582: S209.....
1583:  & \halk\   &  22.24 &    0.15 &   29.70 &    0.23 &   30.8 &    0.39 & B \\ 
1584:  & \healk\  &   1.74 &    0.17 &   23.07 &    2.71 &   30.8 &    0.39 & D \\ 
1585:  & \hbetk\  &   7.20 &    0.14 &   29.22 &    0.67 &   30.4 &    0.41 & C \\ 
1586: 
1587: M17N.....
1588:  & \halk\   & 387.81 &    3.28 &   33.11 &    0.32 &    8.0 &    1.97 & A \\ 
1589:  & \healk\  &  39.74 &    3.60 &   27.69 &    2.94 &    8.0 &    1.97 & B \\ 
1590:  & \hbetk\  & 111.53 &    1.04 &   33.27 &    0.36 &    8.0 &    0.89 & A \\ 
1591:  & \hebetk\ &  11.93 &    1.21 &   25.32 &    3.16 &    8.0 &    0.89 & B \\ 
1592: 
1593: M17S.....
1594:  & \halk\   & 835.20 &    5.25 &   35.29 &    0.26 &    4.8 &    2.16 & A \\ 
1595:  & \healk\  &  77.02 &    5.41 &   34.25 &    2.98 &    4.8 &    2.16 & B \\ 
1596:  & \hbetk\  & 245.60 &    1.46 &   35.91 &    0.25 &    4.8 &    1.57 & A \\ 
1597:  & \hebetk\ &  22.27 &    1.51 &   34.97 &    2.96 &    4.8 &    1.57 & B \\ 
1598: 
1599: W43.....
1600:  & \halk\   & 241.82 &    0.45 &   33.34 &    0.07 &   21.0 &    0.94 & A \\ 
1601:  & \healk\  &  19.37 &    0.47 &   31.37 &    0.98 &   21.0 &    0.94 & B \\ 
1602:  & \hbetk\  &  71.83 &    0.19 &   32.55 &    0.10 &   21.0 &    0.73 & A \\ 
1603:  & \hebetk\ &   6.34 &    0.21 &   26.81 &    1.09 &   21.0 &    0.73 & C \\ 
1604: 
1605: W49.....
1606:  & \halk\   & 417.38 &    0.60 &   29.69 &    0.05 &   20.2 &    0.85 & A \\ 
1607:  & \healk\  &  41.31 &    0.69 &   23.94 &    0.65 &   20.2 &    0.85 & B \\ 
1608:  & \hbetk\  & 107.21 &    0.28 &   29.55 &    0.09 &   20.2 &    0.76 & A \\ 
1609:  & \hebetk\ &  10.67 &    0.34 &   21.55 &    1.29 &   20.2 &    0.76 & B \\ 
1610: 
1611: W51.....
1612:  & \halk\   & 971.22 &    0.58 &   29.53 &    0.02 &    8.4 &    2.80 & A \\ 
1613:  & \healk\  &  87.97 &    0.66 &   23.44 &    0.21 &    8.4 &    2.80 & A \\ 
1614:  & \hbetk\  & 248.53 &    0.32 &   29.77 &    0.04 &    8.4 &    1.15 & A \\ 
1615:  & \hebetk\ &  23.75 &    0.36 &   25.65 &    0.86 &    8.4 &    1.15 & B \\ 
1616: 
1617: NGC7538.....
1618:  & \halk\   & 126.41 &    0.23 &   27.59 &    0.06 &   25.2 &    0.52 & A \\ 
1619:  & \healk\  &  13.17 &    0.26 &   21.90 &    0.53 &   25.2 &    0.52 & B \\ 
1620:  & \hbetk\  &  37.87 &    0.14 &   28.22 &    0.12 &   24.4 &    0.45 & B \\ 
1621:  & \hebetk\ &   3.49 &    0.17 &   20.59 &    1.35 &   24.4 &    0.45 & C \\ 
1622: 
1623: 
1624: \enddata
1625: %\tablenotetext{a}{}
1626: \end{deluxetable}
1627: 
1628: 
1629: \clearpage
1630: 
1631: \begin{deluxetable}{lcccccccccc}
1632: %\tablewidth{16cm}
1633: \tablecolumns{11}
1634: \tablecaption{NEBULA Model Parameters$^{\rm a}$\label{tab:NEBULA}}
1635: \tablehead{
1636: \colhead{}  & 
1637: \multicolumn{5}{c}{\underline{~~~~~~~~~~~~~~~Component 1~~~~~~~~~~~~~~~}} &
1638: \multicolumn{5}{c}{\underline{~~~~~~~~~~~~~~~Component 2~~~~~~~~~~~~~~~}} \\
1639: \colhead{Model}  & \colhead{$\theta_s$} & \colhead{$n_e$} & \colhead{$M_f$} & 
1640: \colhead{\yp3}   & \colhead{\yp4}       & \colhead{$\theta_s$} & 
1641: \colhead{$n_e$}  & \colhead{$M_f$}      & \colhead{\yp3}  & \colhead{\yp4} \\
1642: \colhead{}       & \colhead{(\arcmin)}  & \colhead{(cm$^{-3}$)} & \colhead{} & 
1643: \colhead{(10$^{-5}$)} & \colhead{}      & \colhead{(\arcmin)} & 
1644: \colhead{(cm$^{-3}$)} & \colhead{}      & \colhead{(10$^{-5}$)} & \colhead{}
1645: }
1646: \startdata
1647: 
1648: A1 & 3.5 & 100 & 1.00 & 1.0 & 0.10 & 
1649:      \nodata  & \nodata  & \nodata  & \nodata  & \nodata  \\
1650: A2 & 3.5 & 100 & 1.00 & 0.7 & 0.07 & 
1651:      \nodata  & \nodata  & \nodata  & \nodata  & \nodata  \\
1652: B  & 1.5 & 500 & 0.30 & 1.0 & 0.10 & 3.5 & 100 & 0.70 & 1.0 & 0.10 \\
1653: C  & 1.5 & 100 & 0.08 & 1.0 & 0.10 & 3.5 & 100 & 0.92 & 0.5 & 0.05 \\
1654: D  & 1.5 & 100 & 0.08 & 0.5 & 0.05 & 3.5 & 100 & 0.92 & 1.0 & 0.10 \\
1655: E  & 1.5 & 500 & 0.30 & 1.0 & 0.10 & 3.5 & 100 & 0.70 & 0.5 & 0.05 \\
1656: F  & 1.5 & 500 & 0.30 & 0.5 & 0.05 & 3.5 & 100 & 0.70 & 1.0 & 0.10 \\
1657: \enddata
1658: \tablenotetext{a}{The distance to all model nebulae is 8\kpc.}
1659: \end{deluxetable}
1660: 
1661: 
1662: \clearpage 
1663: 
1664: \begin{deluxetable}{lccccccc}
1665: \tablecolumns{8}
1666: \tablewidth{0pt}
1667: \tablecaption{NEBULA Model Results\label{tab:models}}
1668: \tablehead{
1669: \colhead{}  & 
1670: \multicolumn{6}{c}{\underline{~~~~~~~~~~~~~~~~~~~~~~~X-band$^{\rm a}$~~~~~~~~~~~~~~~~~~~~~~~}} &
1671: \multicolumn{1}{c}{\underline{K-band$^{\rm b}$}} \\
1672: \colhead{Model}  & \colhead{$T_C$} & \colhead{$T_{\rm L}$} & \colhead{\yp3} & 
1673: \colhead{\yp4} & \colhead{$\kappa_i$} & \colhead{\y3} & \colhead{\yp4} \\
1674: \colhead{}  & \colhead{(K)} & \colhead{(mK)} & \colhead{(10$^{-5}$)} & 
1675: \colhead{} & \colhead{} & \colhead{(10$^{-5}$)} &\colhead{}
1676: }
1677: \startdata
1678: 
1679: A1 & 0.93 & 1.10 & 1.00 & 0.100 & 1.00 & 1.00 & 0.100 \\
1680: A2 & 0.93 & 0.81 & 0.72 & 0.070 & 1.43 & 1.03 & 0.070 \\
1681: B  & 3.00 & 1.50 & 0.77 & 0.100 & 1.00 & 0.77 & 0.100 \\
1682: C  & 0.93 & 0.64 & 0.57 & 0.055 & 1.82 & 1.04 & 0.067 \\
1683: D  & 0.93 & 1.10 & 0.96 & 0.095 & 1.05 & 1.01 & 0.082 \\
1684: E  & 3.00 & 1.10 & 0.53 & 0.085 & 1.18 & 0.62 & 0.096 \\
1685: F  & 3.00 & 1.30 & 0.65 & 0.064 & 1.56 & 1.01 & 0.053 \\
1686: 
1687: \enddata
1688: \tablenotetext{a}{$\nu = 8$\ghz; HPBW = 3$\arcmper$5.}
1689: \tablenotetext{b}{$\nu = 18$\ghz; HPBW = 1$\arcmper$5.}
1690: \end{deluxetable}
1691: 
1692: 
1693: \clearpage 
1694: 
1695: %\begin{deluxetable}{lcccccc}
1696: %\tablecolumns{7}
1697: %\tablewidth{15cm}
1698: %\tablecaption{H\,II Region $^{3}$He/H Abundances\tablenotemark{a}
1699: %\label{tab:abundances}}
1700: %\tablehead{
1701: %\colhead{Source} & 
1702: %\colhead{LTE U$\rho$S\tablenotemark{b}} & 
1703: %\colhead{$\kappa_s$\tablenotemark{c}} & 
1704: %\colhead{$\sigma_t$\tablenotemark{d}} &
1705: %\colhead{$\kappa_i$\tablenotemark{e}} & 
1706: %\colhead{$\sigma_i$\tablenotemark{f}} & 
1707: %\colhead{\y3\ Abundance\tablenotemark{g}} 
1708: %}
1709: %\startdata
1710: %
1711: %W3           & 2.3       & 3.43        & 28      & $\ge 1.32$  & \nodata & $\ge 10.$     \\
1712: %G133.8       & 1.7       & 1.18        & 15      & 1.22        & 9.1     & $2.4 \pm 0.42$ \\
1713: %S206         & 1.9       & 1.25        & 20      & 1.04        & 3.2     & $2.5 \pm 0.51$ \\
1714: %S209         & 0.90      & 1.21        & 18      & 1.20        & 4.5     & $1.3 \pm 0.24$ \\
1715: %S212         & $< 0.89$  & \gsim\ 1    & \nodata & 1.20        & \nodata & \nodata       \\
1716: %S228         & $< 2.2$   & \gsim\ 1    & \nodata & $\ge 1$     & \nodata & \nodata       \\
1717: %S235         & $< 1.4$   & \gsim\ 1    & \nodata & $\ge 1$     & \nodata & \nodata       \\
1718: %S252         & 0.98      & \gsim\ 1    & 12      & 1.96        & \nodata & \gsim \,1.9     \\
1719: %S311         & 1.8       & \gsim\ 1    & 22      & 1.41        & \nodata & \gsim \,2.5     \\
1720: %\sgr{B2}     & 1.1       & 1.74        & 21      & $\ge 1$     & \nodata & $\ge 1.9$     \\
1721: %G1.1         & 0.78      & \gsim\ 1    & 16      & 1.56        & \nodata & \gsim \,1.2     \\
1722: %M17S         & 1.3       & $\ge 1$     & 15      & 1.09        & \nodata & $\ge 1.4$     \\
1723: %M17N         & $< 0.37$  & $\ge 1$     & \nodata & 1.06        & \nodata & \nodata       \\
1724: %W43          & 0.69      & 1.57        & 21      & 1.30        & 7.7     & $1.4 \pm 0.31$ \\
1725: %S76          & $< 4.0$   & \gsim\ 1    & \nodata & $\ge 1$     & \nodata & \nodata       \\
1726: %W49          & 0.43      & 2.86        & 31      & $\ge 1$     & \nodata & $\ge 1.2$     \\
1727: %W51          & 1.6       & 1.50        & 16      & $\ge 1$     & \nodata & $\ge 2.4$     \\
1728: %S90          & 1.5       & \gsim\ 1    & 16      & 1.43        & \nodata & \gsim \,2.1     \\
1729: %S156         & $< 7.9$   & \gsim\ 1    & \nodata & 1.23        & \nodata & \nodata       \\
1730: %\ngc{7538}   & 1.3       & 1.20        & 15      & 1.19        & 4.0     & $1.9 \pm 0.29$ \\
1731: %S162         & $< 0.91$  & 1.21        & \nodata & $\ge 1$     & \nodata & \nodata       \\
1732: %\enddata
1733: %
1734: %\tablenotetext{a}{All abundances are \expo{-5} by number.}
1735: %\tablenotetext{b}{Uniform density sphere LTE model (U$\rho$S) \her3\ 
1736: %abundance from Paper II.}
1737: %\tablenotetext{c}{Density structure correction factor from Paper II.}
1738: %\tablenotetext{d}{Total percent uncertainty from \he3\ observables and
1739: %the uncertainty in $\kappa_s$ (see Paper II).}
1740: %\tablenotetext{e}{Ionization structure correction factor 
1741: %(see text).}
1742: %\tablenotetext{f}{Ionization correction uncertainty percent error;
1743: %includes errors in the Gaussian fits to the observed H and \he4\ RRL
1744: %parameters.}  
1745: %\tablenotetext{g}{\y3\,=\,$\kappa_s\,\times\,\kappa_i\,\times$\,LTE (U$\rho$S). 
1746: %The percent uncertainty is $(\sigma_t^2 + \sigma_i^2)^{1/2}$.}  
1747: %\end{deluxetable}
1748: 
1749: 
1750: %\clearpage 
1751: 
1752: \begin{deluxetable}{lccccc}
1753: \tablecolumns{6}
1754: %\tablewidth{16cm}
1755: \tablecaption{\hii\ Region Excitation Properties\label{tab:excitation}}
1756: \tablehead{
1757: \colhead{} & \multicolumn{3}{c}{\underline{~~~~~~~~~~~~~~~~~~~~~~Radio\tablenotemark{a}~~~~~~~~~~~~~~~~~~~~~~}} &
1758: \multicolumn{2}{c}{\underline{~~~~~~~~~~~~~~~~~Optical\tablenotemark{b}~~~~~~~~~~~~~~~~~}} \\
1759: \colhead{Source}   & \colhead{\hepr4} & \colhead{$N_{\rm L}$} &
1760: \colhead{Spectral} & \colhead{\hepr4} & \colhead{Spectral} \\
1761: \colhead{}         & \colhead{}       & \colhead{ (s$^{-1}$)} &
1762: \colhead{Type}     & \colhead{}       & \colhead{Type} }
1763: \startdata
1764: 
1765: W3           & $0.076 \pm\ 0.0033$ & 49.28 & O6.5 (O5.5) & \nodata & \nodata \\
1766: G133.8       & $0.082 \pm\ 0.0075$ & 48.96 & O7.5 (O6) &   \nodata & \nodata \\
1767: S206         & $0.092 \pm\ 0.0029$ & 48.95 & O7.5 (O6.5) & 0.10 & O4--O6 \\
1768: \nodata & \nodata &\nodata & \nodata & (8) & (7,8,13,17,20) \\
1769: S209         & $0.083 \pm\ 0.0037$ & 49.82 & O3 (O4) & 0.084--0.12 & O9 \\
1770: \nodata & \nodata & \nodata & \nodata & (8,25) & (3) \\
1771: S212         & \nodata           & 48.67 & O8.5 (O7) & 0.088--0.10 & O5.5--O7 \\
1772: \nodata & \nodata & \nodata & \nodata & (8,11,25) & (3,17,20) \\
1773: S228         & \nodata             & 47.95 & B0.5 (O9.5) & \nodata & O8--B0 \\
1774: \nodata & \nodata & \nodata & \nodata & \nodata & (3,17) \\
1775: S235         & \nodata             & 47.86 & B0.5 (O9.5) & \nodata & O9.5--B0 \\
1776: \nodata & \nodata & \nodata & \nodata & \nodata & (13,17) \\
1777: S252         & \nodata             & 48.67 & O8.5 (O7) & 0.051 & O6--O6.5 \\
1778: \nodata & \nodata & \nodata & \nodata & (24) & (5,12,14,15,19) \\
1779: S311         & $0.071 \pm\ 0.0038$ & 48.93 & O8 (O6.5) & \nodata & O6--O6.5 \\
1780: \nodata & \nodata & \nodata & \nodata & \nodata & (6,10,14,23) \\
1781: \sgr{B2}     & $0.059 \pm\ 0.0182$ & 50.57 & $<$ O3 ($<$ O4) & \nodata & \nodata \\
1782: G1.1         & $0.064 \pm\ 0.0066$ & 49.91 & O3 (O4) & \nodata & \nodata \\
1783: M17S         & $0.092 \pm\ 0.0110$ & 50.46 & $<$ O3 ($<$ O4) & \nodata & O3--O6 \\
1784: \nodata & \nodata & \nodata & \nodata & \nodata & (16) \\
1785: M17N         & $0.094 \pm\ 0.0089$ & 49.75 & O4 (O5) & 0.10 & O3--O6 \\
1786: \nodata & \nodata & \nodata & \nodata & (9,22) & (16) \\
1787: W43          & $0.068 \pm\ 0.0052$ & 50.42 & $<$ O3 ($<$ O4) & \nodata & WN7 \\
1788: \nodata & \nodata & \nodata & \nodata & \nodata & (2) \\
1789: S76          & \nodata             & 47.76 & $>$ B0.5 (O9.5) & \nodata & \nodata \\
1790: W49          & $0.079 \pm\ 0.0052$ & 50.87 & $<$ O3 ($<$ O4) & \nodata & \nodata \\
1791: W51          & $0.079 \pm\ 0.0024$ & 50.47 & $<$ O3 ($<$ O4) & \nodata & \nodata \\
1792: S90          & $0.070 \pm\ 0.0067$ & 48.87 & O8 (O6.5) & \nodata & O9.5 \\
1793: \nodata & \nodata & \nodata & \nodata & \nodata & (13,14) \\
1794: S156         & \nodata             & 48.91 & O8 (O6.5) & 0.081 & O6.5--O7 \\
1795: \nodata & \nodata & \nodata & \nodata & (8) & (1,12,17) \\
1796: \ngc{7538}   & $0.084 \pm\ 0.0034$ & 49.25 & O6.5 (O5.5) & 0.10 & O7 \\
1797: \nodata & \nodata & \nodata & \nodata & (18) & (13,21) \\
1798: S162         & \nodata             & 48.79 & O8 (O6.5) & \nodata & O6.5 \\
1799: \nodata & \nodata & \nodata & \nodata & \nodata & (4) \\
1800: \enddata
1801: 
1802: \tablenotetext{a}{Based on the Paper I X-band radio data that are 
1803: appropriate for the \hep3\ ionization correction.  The
1804: \hepr4\ abundance ratio is determined using the \hal\ and \heal\ 
1805: RRLs (only quality factors A and B are considered).  The H-ionizing
1806: luminosity, $N_{\rm L}$, is calculated from the radio continuum
1807: emission.  The spectral type is determined from $N_{\rm L}$ and a
1808: single star ZAMS model (Vacca et al. 1996; Panagia 1973, in
1809: parentheses).}
1810: \tablenotetext{b}{Based on optical data from the literature.  The
1811: references are listed in parentheses.}
1812: \tablerefs{
1813: (1) Barlow et al. (1976);
1814: (2) Blum et al. (1999);
1815: (3) Chini \& Wink (1984);
1816: (4) Conti \& Alschuler (1971);
1817: (5) Conti \& Leep (1974);
1818: (6) Crampton (1971);
1819: (7) Crampton et al. (1978);
1820: (8) Deharveng et al. (2000);
1821: (9) Esteban et al. (1999);
1822: (10) Feinstein \& V\'{a}zquez (1989);
1823: (11) Fich \& Silkey (1991);
1824: (12) Georgelin (1975);
1825: (13) Georgelin et al. (1973);
1826: (14) Goy (1980);
1827: (15) Haikala (1995)
1828: (16) Hanson \& Conti (1995);
1829: (17) Hunter \& Massey (1990);
1830: (18) Lynds \& O'Neil Jr. (1986);
1831: (19) Miller (1968);
1832: (20) Moffat et al. (1979); 
1833: (21) Moreno \& Charvarr\'{i}a-K. (1986)
1834: (22) Peimbert et al. (1992b);
1835: (23) Persi et al. (1987);
1836: (24) Shaver et al. (1983); 
1837: (25) V\'{i}lchez \& Esteban (1996).
1838: }
1839: 
1840: 
1841: \end{deluxetable}
1842: 
1843: 
1844: \clearpage
1845: 
1846: \begin{figure}
1847: \includegraphics[angle=-90,scale=0.65]{f1a.ps}
1848: \includegraphics[angle=-90,scale=0.65]{f1b.ps}
1849: \caption{Spectra of the 91$\alpha$ X-band (8\ghz) and the
1850: 70$\alpha$ K-band (18\ghz) radio recombination line emission for the
1851: \hii\ regions \ngc{7538} and W43.  The vertical lines flag from
1852: left to right the C, He, and H transitions.  (The additional flag in
1853: the 91$\alpha$ spectra is the H\,154$\epsilon$ transition.)  The
1854: intensity scale is in units of milliKelvins.  The LSR velocity
1855: increases from left to right at 1.3\kms\ per channel.
1856: \label{fig:h90h70}}
1857: \end{figure}
1858: 
1859: \clearpage
1860: 
1861: \begin{figure}
1862: \includegraphics[angle=-90,scale=0.60]{f2.ps}
1863: \caption{The \hepr4, \yp4, abundance ratio plotted as a function of the
1864: principal quantum number, $n$, of the RRL transition.  The 70$\alpha$, 
1865: 88$\beta$, 91$\alpha$, and 114$\beta$ transitions are shown.  
1866: Abundances are plotted only for transitions that have a quality
1867: factor of A or B (see Table I and Paper I).
1868: \label{fig:QMn}}
1869: \end{figure}
1870: 
1871: \clearpage
1872: 
1873: \begin{figure}
1874: \includegraphics[angle=-90,scale=0.60]{f3.ps}
1875: \caption{Comparison of the \yp4\ ratios derived from the \hal\ 
1876: and \halk\ spectra for the 10 K-band sources in the order listed in
1877: Table~\ref{tab:data}.  Only abundances for quality factor A or B are
1878: plotted.  {\it Top Left:\/} The
1879: \hal\,(circles) and \halk\,(diamonds) \yp4\ ratios for each nebula.
1880: {\it Top Right:\/} The \yp4\,(\hal)\,$-$\,\yp4\,(\halk) difference.
1881: {\it Bottom Left:\/} The fractional \yp4\ deviation:
1882: (\yp4\,(\hal)\,$-$\,\yp4\,(\halk))\,/\,\yp4\,(\hal).  {\it Bottom
1883: Right:\/} The \yp4\,(\hal)\,/\,\yp4\,(\halk) ratio.
1884: \label{fig:yplus}}
1885: \end{figure}
1886: 
1887: \clearpage
1888: 
1889: \begin{figure}
1890: \includegraphics[angle=-90,scale=0.70]{f4.ps} \caption{The
1891: \hepr4, \yp4, abundance ratio plotted as a function of the H-ionizing
1892: luminosity, $N_{\rm L}$.  The \yp4\ values are calculated using only the \hal\
1893: and \heal\ transitions.  Only transitions that have a quality factor of
1894: A or B (see Paper I) are shown.  Circles denote simple sources;
1895: triangles identify complex sources (see text).
1896: \label{fig:Nlc}}
1897: \end{figure}
1898: 
1899: 
1900: 
1901: 
1902: 
1903: \end{document}
1904: