1: % ****** Start of file apssamp.tex ******
2: %
3: % This file is part of the APS files in the REVTeX 4 distribution.
4: % Version 4.0 of REVTeX, August 2001
5: %
6: % Copyright (c) 2001 The American Physical Society.
7: %
8: % See the REVTeX 4 README file for restrictions and more information.
9: %
10: % TeX'ing this file requires that you have AMS-LaTeX 2.0 installed
11: % as well as the rest of the prerequisites for REVTeX 4.0
12: %
13: % See the REVTeX 4 README file
14: % It also requires running BibTeX. The commands are as follows:
15: %
16: % 1) latex apssamp.tex
17: % 2) bibtex apssamp
18: % 3) latex apssamp.tex
19: % 4) latex apssamp.tex
20: %
21: %\documentclass[twocolumn,showpacs,preprintnumbers,amsmath,amssymb]{revtex4}
22: \documentclass[prl,showpacs,amsmath,amssymb,twocolumn,aps]{revtex4}
23: %\documentclass[preprint,showpacs,preprintnumbers,amsmath,amssymb]{revtex4}
24:
25: % Some other (several out of many) possibilities
26: %\documentclass[preprint,aps]{revtex4}
27: %\documentclass[preprint,aps,draft]{revtex4}
28: %\documentclass[prb]{revtex4}% Physical Review B
29: \usepackage{graphicx,graphics}% Include figure files
30: \usepackage{epsfig}% Include figure files
31: \usepackage{dcolumn}% Align table columns on decimal point
32: \usepackage{bm}% bold math
33: \usepackage{natbib}
34: \usepackage{times}
35: \usepackage{ulem}
36: \def\kmm#1 {{\bf [KMM:~ #1]~}}
37: \def\new#1 {{\bf #1 }}
38: \def\cut#1 {\sout{#1} }
39: %\nofiles
40:
41: \newcommand{\pmn}{PMN~J0134$-$0931}
42: \newcommand{\al}{\ensuremath{\alpha}}
43: \newcommand{\dal}{\ensuremath{\Delta \alpha/ \alpha}}
44: \newcommand{\gp}{\ensuremath{{g}_{p}}}
45: \newcommand{\lqcd}{\ensuremath{\Lambda_{\rm QCD}}}
46: \newcommand{\y}{\ensuremath{{m}_{e}/{m}_{p}}}
47: \newcommand{\dmu}{\ensuremath{\Delta \mu/\mu}}
48:
49:
50: \newcommand{\beq}{\begin{equation}}
51: \newcommand{\eeq}{\end{equation}}
52: \newcommand{\noi}{\noindent}
53: \newcommand{\lb}{\left(}
54: \newcommand{\rb}{\right)}
55: \newcommand{\lsb}{\left[}
56: \newcommand{\rsb}{\right]}
57:
58:
59: \begin{document}
60: \title{Constraints on changes in fundamental constants from a cosmologically
61: distant OH absorber/emitter}
62:
63: \author{N.~Kanekar$^1$}
64: \email{nkanekar@aoc.nrao.edu}
65: \author{C.~L.~Carilli$^1$}
66: \author{G.~I.~Langston$^1$}
67: \author{G.~Rocha$^2$}
68: \author{F.~Combes$^3$}
69: \author{R.~Subrahmanyan$^4$}
70: \author{J.~T.~Stocke$^5$}
71: \author{K.~M.~Menten$^6$}
72: \author{F.~H.~Briggs$^{4,7}$}
73: \author{T.~Wiklind$^{8}$}
74: \affiliation{$^1$National Radio Astronomy Observatory, Socorro, NM 87801, USA}
75: \affiliation{$^2$Cavendish Laboratory, Cambridge CB3 0HE, UK}
76: \affiliation{$^3$Observatoire de Paris, F-75014, Paris, France}
77: \affiliation{$^4$Australia Telescope National Facility, Epping, NSW 1710, Australia}
78: \affiliation{$^5$University of Colorado, Boulder, CO 80309, USA}
79: \affiliation{$^6$Max-Planck-Institut f\"ur Radioastronomie, 53121 Bonn, Germany}
80: \affiliation{$^7$Australian National University, ACT 2611, Australia}
81: \affiliation{$^8$Space Telescope Science Institute, Baltimore, MD 21218, USA }
82:
83:
84: \date{\today}
85:
86:
87: \begin{abstract}
88: We have detected the four 18cm OH lines from the $z \sim 0.765$ gravitational lens
89: toward \pmn. The 1612 and 1720~MHz lines are in conjugate absorption and
90: emission, providing a laboratory to test the evolution of fundamental
91: constants over a large lookback time. We compare the HI and OH~main line absorption
92: redshifts of the different components in the $z \sim 0.765$ absorber and the
93: $z \sim 0.685$ lens toward B0218+357 to place stringent constraints on
94: changes in $F \equiv g_p \lsb \alpha^2/\mu \rsb^{1.57}$. We obtain
95: $\lsb \Delta F/F\rsb = (0.44 \pm 0.36^{\rm stat} \pm 1.0^{\rm syst}) \times 10^{-5}$,
96: consistent with no evolution over the redshift range $0 < z \lesssim 0.7$. The
97: measurements have a $2 \sigma$ sensitivity of $\lsb \dal \rsb < 6.7 \times 10^{-6}$
98: or $\lsb \dmu\rsb < 1.4 \times 10^{-5}$ to fractional changes in
99: $\alpha$ and $\mu$ over a period of $\sim 6.5$~Gyr, half the
100: age of the Universe. These are among the most sensitive constraints on changes in $\mu$.
101: \end{abstract}
102:
103: \pacs{98.80.Es,06.20.Jr,33.20.Bx,98.58.-w}
104: \maketitle
105: {\it Introduction.}--- A fairly generic feature of modern higher-dimensional
106: theoretical models is that fundamental constants like the fine structure constant
107: $\alpha$, the electron-proton mass ratio $\mu \equiv m_e/m_p$, the proton
108: gyromagnetic ratio $g_p$, etc, depend on the scale lengths of the extra
109: dimensions of the theory. In the current theoretical framework, it is implausible that
110: these scale lengths remain constant, implying that $\alpha$, $\mu$, etc
111: should vary with time. The detection of such changes provides an avenue to
112: probe new and fundamental physics.
113:
114: Null results have been obtained in all terrestrial studies of evolving
115: constants, with atomic clocks and isotopic abundances in the Oklo natural
116: fission reactor providing the tightest constraints on fractional
117: changes in the fine structure constant ($(1/\alpha)\lsb \Delta \alpha/{\Delta t}
118: \rsb < 4 \times 10^{-15}$ per year, over three years \cite{peik04}, and
119: $\dal < 1.2 \times 10^{-7}$, over $\sim 1.8 \times 10^9$ years \cite{damour96},
120: respectively). However, terrestrial measurements only probe fairly small
121: fractions of the age of the Universe; astrophysical techniques
122: are needed to examine the possibility of variations at earlier times
123: (e.g. \cite{murphy03,srianand04,kanekar04,tzanavaris05,ivanchik05}).
124: It is these techniques that provide tantalizing evidence for changes in $\alpha$;
125: the ``many-multiplet'' method, applied to Keck telescope optical
126: spectra, gives $\dal = (-5.4 \pm 1.2) \times 10^{-6}$ over the redshift range
127: $0.2 < z < 3.7$ \cite{murphy03}. However, a similar technique, applied to Very
128: Large Telescope spectra, yields a conflicting result,
129: $\dal = (-0.6 \pm 0.6) \times 10^{-6}$, over $0.4 < z < 2.3$
130: \cite{srianand04}. Independent techniques are clearly needed as systematics
131: appear to play a significant role in the current results.
132:
133: The four 18cm radio OH lines have very different dependences on $\alpha$,
134: $\mu$ and $g_p$ and their redshifted frequencies can hence be compared to each
135: other and to those of the HI 21cm hyperfine line or HCO$^+$ rotational lines
136: to measure any variation \cite{chengalur03,darling03}.
137: Even more interesting is the case of conjugate emission/absorption by the 18cm
138: satellite OH lines, detected in a single cosmologically distant object (at $z \sim 0.247$
139: toward PKS1413+135; \cite{kanekar04,darling04}). Here, the pumping mechanism
140: guarantees that the two lines arise from the same gas; a comparison can thus be made
141: between the 1720 and 1612~MHz redshifts without concerns about systematic motions
142: between the clouds in which the different lines arise.
143:
144: Only four redshifted OH main absorbers are currently known \cite{kanekar02}, with
145: high resolution data only available for the $z\sim 0.685$ lens toward
146: B0218+357 \cite{kanekar03}. Similarly, only one redshifted conjugate
147: OH satellite system is known,
148: at a relatively low redshift, $z \sim 0.247$; this corresponds to a lookback
149: time of $\sim 2.9$~Gyr, not much earlier than the time range probed
150: by the Oklo reactor \footnote{Throughout this paper, we use an LCDM cosmology,
151: with $\Omega_m = 0.3$, $\Omega_\Lambda = 0.7$ and $H_0 = 70$~km/s~Mpc$^{-1}$.}.
152: We report here the detection of all four 18cm OH lines in a new conjugate
153: system at $z \sim 0.765$, corresponding to a lookback time of $\sim 6.7$~Gyr.
154:
155: {\it Spectra and results}---
156: The redshifted OH 18cm lines from the $z \sim 0.765$ lens toward \pmn~
157: were observed simultaneously with the Green Bank Telescope (GBT) in October, 2004,
158: and January, 2005, with an observing resolution of $\sim 1$~km/s (after Hanning
159: smoothing). A separate
160: GBT observation in January, 2005, provided a high resolution spectrum in the
161: redshifted HI~21cm line, originally detected by \cite{kanekar03b}. The
162: HI~21cm and main line OH spectra (smoothed to resolutions of $\sim 5$ and
163: $\sim 14.5$~km/s, respectively, and resampled) are shown in Figs.~\ref{fig:hi-oh}[A]
164: and [B], while the three panels of Fig.~\ref{fig:oh-sat} show the 1720 and
165: 1612~MHz satellite spectra and the sum of 1720 and 1612~MHz optical depths
166: (all at a resolution of $\sim 4.7$~km/s).
167: \pmn~is unresolved by the GBT beam; the above optical depths are
168: the ratio of line flux density to continuum flux density for
169: each transition (using the low optical depth limit).
170: All spectra have a root-mean-square optical depth noise of $\sim 0.0018$, per
171: $\sim 5$~km/s channel.
172:
173: \begin{figure*}[t!]
174: \centering
175: \epsfig{file=fig1a.eps,height=3.3in,width=3.3in}
176: \epsfig{file=fig1b.eps,height=3.3in,width=3.3in}
177: \vskip -0.1in
178: \caption{GBT HI~21cm ([A], $\sim 5$~km/s resolution) and OH main line
179: ([B], $\sim 14.5$~km/s resolution) spectra
180: toward \pmn, with negative optical depth $(-10^3 \times \tau)$ plotted against
181: barycentric frequency, in MHz. The solid line shows the three-gaussian fit to each
182: spectrum. The vertical lines in each figure indicate the locations of the three
183: components (marked A, B and C), with the dashed and dotted lines in [B]
184: showing the 1667 and 1665 components, respectively. }
185: \vskip -0.2in
186: \label{fig:hi-oh}
187: \end{figure*}
188:
189: \begin{figure}
190: \centering
191: \hskip -2.5 in \epsfig{file=fig2.eps,height=4.0in,width=5.0in}
192: \vskip -0.1in
193: \caption{OH satellite line spectra ($\sim 5$~km/s resolution), with negative optical depth ($-10^3 \times
194: \tau $) plotted against barycentric velocity, in km/s, relative to $z = 0.76355$,
195: the redshift of component C.
196: [A] 1720~MHz transition, redshifted to $\sim 975.6$~MHz; [B] 1612~MHz transition,
197: redshifted to $\sim 912.2$~MHz;
198: [C] Sum of 1612 and 1720~MHz spectra; this spectrum is
199: consistent with noise. Note that the feature at $\sim -30$~km/s
200: is not statistically significant. The satellite lines peak
201: $\sim 10$~km/s blueward of component C (the dashed vertical line),
202: which has the lowest redshift of the three spectral components.}
203: \vskip -0.2in
204: \label{fig:oh-sat}
205: \end{figure}
206: Besides the above, the redshifted HCO$^+$ 2--1 line was observed with the
207: IRAM 30m telescope and the Australia Telescope Compact Array (ATCA), the 6~cm ground state
208: H$_2$CO doublet lines with the GBT and the 2~cm first rotationally
209: excited state H$_2$CO lines
210: with the Very Large Array and the GBT. None of these transitions were detected,
211: down to $3\sigma$ limits of $\tau < 0.07$ for the HCO$^+$ line and
212: $\tau < 0.002$, $\tau < 0.005$ for the ground and excited state H$_2$CO lines,
213: respectively. This is
214: curious as every other redshifted OH absorber has also been detected in
215: HCO$^+$ absorption (e.g.~\cite{wiklind98,kanekar02}).
216: While the structure of the background source is very different at the HCO$^+$
217: and OH frequencies, implying that small-scale structure in the molecular cloud
218: could be an issue, it is still surprising that none of the
219: OH components, at very different velocities from each other, show any trace of
220: mm-wave absorption.
221:
222: The HI~21cm profile has three fairly clear components, two of which are blended
223: and well-separated from the third. Similarly, both the 1667 and 1665~MHz OH lines
224: show two clearly resolved components, with the lower redshift one somewhat
225: asymmetric, suggesting that it is blended. While the redshifts of the two strong
226: 21cm and main OH components are similar, the OH satellite lines are
227: blue-shifted by $\sim 10$~km/s relative to the lowest redshift (highest frequency)
228: component in Fig.~\ref{fig:hi-oh}[A] and [B]. This is reminiscent of the situation in the
229: other redshifted conjugate OH system, PKS1413+135 \cite{kanekar04}. In both cases,
230: the sum of the satellite frequencies is different from the sum of the main
231: line frequencies; since the two sums have the same dependence on $\alpha$, $\mu$
232: and $g_p$ \cite{chengalur03,darling03}, the satellite and main OH lines must
233: arise in different gas.
234:
235: The sum of the 1720 and 1612~MHz optical depths in Fig.~\ref{fig:oh-sat}[c]
236: is consistent with noise; the satellite lines are thus conjugate
237: with each other. Such conjugate behavior arises due to competition between the
238: intra-ladder $119\mu$ and cross-ladder $79\mu$ decay routes to the OH ground state,
239: after the molecules have been pumped by collisions or far-infra-red radiation into
240: the higher excited states (e.g. \cite{langevelde95}). The fact that the 1720~MHz
241: line is seen in emission and the
242: 1612~MHz, in absorption, implies that the intra-ladder decay route dominates. This,
243: and the requirement that the $119\mu$ transition be optically thick
244: for the 1720~MHz line inversion \cite{langevelde95}, yields the constraint
245: $3.5 \times 10^{15} \lesssim N_{\rm OH} \lesssim 3.5 \times 10^{16}$~cm$^{-2}$
246: on the OH column density.
247:
248: The 1720~MHz line luminosity is $L_{\rm OH} \sim 3700 L_\odot$, making this
249: the brightest known 1720~MHz megamaser by more than an order of magnitude. It is
250: also the most distant OH megamaser; for comparison, the furthest known 1667~MHz
251: megamaser is at $z \sim 0.265$ \cite{baan92}. This is also the first case
252: of conjugate OH satellite emission/absorption in a ``normal''
253: galaxy; all previous cases were objects containing an active galactic nucleus
254: (e.g. Cen~A; \cite{langevelde95}), where the OH level populations might have been
255: affected by nuclear activity. This is interesting as it suggests that such conjugate
256: behaviour might not be as rare as earlier expected and hence, that it might be used
257: as a tool to probe both spatial and temporal changes in fundamental constants.
258:
259: {\it Constraining changes in fundamental constants.}---
260: A comparison between the redshifts of different spectral lines to measure
261: changes in fundamental constants involves the assumption that the lines have
262: no intrinsic velocity offsets from each other. This is true even for comparisons
263: between lines of the same species as different transitions might be excited
264: under different physical conditions and thus, in different spatial locations.
265: This appears to be the case with the main and satellite OH lines here, implying
266: that one cannot compare their redshifts to estimate changes in the
267: different constants.
268:
269: However, as in the case of PKS1413+135 \cite{kanekar04,darling04}, the conjugate
270: nature of the satellite lines ensures that they arise in the same
271: physical region and, crucially, that systematic velocity offsets are not an
272: issue. The different dependences of the sum and difference of the 1720~MHz and 1612~MHz
273: frequencies on $\alpha$, $\mu$ and $g_p$ then allows us to estimate
274: changes in the quantity $G \equiv g_p \lsb \alpha^2/\mu \rsb^{1.85}$
275: \cite{chengalur03,kanekar04}. The low signal-to-noise ratio of the 1720
276: and 1612 spectra of Fig.~\ref{fig:oh-sat} precludes such an estimate at the
277: present time. However, the high redshift of the system implies that
278: it is an excellent target for deep integrations in the satellite lines, enabling
279: a precision measurement of changes in $\alpha$, $\mu$ and $g_p$ in the
280: future. For example, observations with the Square Kilometer Array,
281: a next generation radio telescope, should be able to detect
282: fractional changes $\Delta G / G \sim 5 \times 10^{-7}$ in this system, implying
283: a sensitivity of $\dal \sim 1.4 \times 10^{-7}$. This is similar to
284: the sensitivity of the Oklo measurement but with fewer assumptions and out to
285: a far larger lookback time of $\sim 6.7$~Gyr. In addition, this sensitivity
286: would be obtained from a single system, unlike the many-multiplet method, which
287: requires a large number of absorbers to average out systematic effects.
288: A comparison between the results from the conjugate
289: systems in PKS1413+135 and \pmn~ will thus allow one to probe true spatio-temporal
290: changes in the above constants (instead of merely averaging over spatial effects),
291: especially since the two sources are very widely separated on the sky.
292:
293: Comparisons between the HI 21cm and main OH lines suffer from the drawback of
294: possible systematic velocity offsets between the two species.
295: However, a tight correlation has been found between the velocities of HCO$^+$ and nearest
296: HI~21cm absorption in the Galaxy, with a dispersion of only $\sim 1.2$~km/s
297: \cite{drinkwater98}. Further, the velocities of Galactic OH and HCO$^+$ absorption
298: have been found to be remarkably similar (\cite{liszt00}; see Figs.~4
299: and 5 of \cite{liszt00}). OH and HI~21cm velocities should thus also be well-correlated
300: and, in fact, the dispersion in this correlation is likely to be less than that between
301: the HCO$^+$ and HI~21 velocities, as the spatial structure of the background source is
302: quite similar at the nearby OH~18~cm and HI~21cm frequencies. It should thus be possible
303: to use a comparison between main OH and HI~21cm absorption from a statistically
304: large number of redshifted systems as an independent probe of any evolution
305: in $\alpha$, $\mu$ and $g_p$ \cite{chengalur03}. We next apply this technique
306: to the absorbers
307: toward \pmn~ and B0218+357, albeit using a more conservative dispersion of 3~km/s
308: between OH and HI intrinsic velocities, characterising internal motions within a
309: molecular cloud.
310:
311: The HI~21cm profile of Fig.~\ref{fig:hi-oh}[A] has the highest signal-to-noise
312: ratio of our spectra (at the same resampled resolution of $\sim 1$~km/s)
313: and has three clear components. We hence used a three-gaussian template to locate
314: the peak redshifts of the different 21cm absorption components, with the amplitudes,
315: positions and widths of the gaussians all left as free parameters. A similar
316: three-gaussian fit was then carried out to the smoothed and resampled 1665 and
317: 1667~MHz spectra, with the difference that the velocity widths here were
318: fixed to those obtained from the 21cm fit and only the amplitudes and
319: positions of individual components left as free parameters. The 1665
320: and 1667~MHz fits were carried out simultaneously, to account for the
321: possibility of blending between the components. Note that the
322: original spectral resolution was $\sim 1$~km/s in all cases and no additional
323: components were seen in any of the spectra. The possibility that strong narrow
324: components might be blended in the lower resolution spectra of Figs.~\ref{fig:hi-oh}~[A]
325: and [B] can thus be ruled out.
326:
327: Figs.~\ref{fig:hi-oh}[A] and [B] show the fitted gaussians as solid lines,
328: overlaid on the HI and OH spectra. The three fitted 21cm components have peak redshifts
329: $z_{\rm HI-A} = 0.764938 \pm 0.000015$, $z_{\rm HI-B} = 0.763881 \pm 0.000033$ and
330: $z_{\rm HI-C} = 0.763515 \pm 0.000014$, while the sums of the 1665 and 1667~MHz
331: frequencies have peak redshifts $z_{\rm OH-A} = 0.764871 \pm 0.000042$,
332: $z_{\rm OH-B} = 0.763852 \pm 0.000019$ and $z_{\rm OH-C} = 0.763550 \pm 0.000010$.
333: Comparing these redshifts, component by component, we obtain
334: $\lsb \Delta F/F\rsb_{\rm A} = (-3.7 \pm 2.8) \times 10^{-5}$,
335: $\lsb \Delta F/F\rsb_{\rm B} = (-1.6 \pm 2.1) \times 10^{-5}$ and
336: $\lsb \Delta F/F\rsb_{\rm C} = (2.0 \pm 1.0) \pm 10^{-5}$, where
337: $F \equiv g_p \lsb \alpha^2/\mu \rsb^{1.57}$ and only statistical errors
338: have been included. A weighted average of these values gives $\lsb \Delta F/F \rsb =
339: (0.86 \pm 0.86) \times 10^{-5}$. Note that a comparison using an entirely unconstrained
340: six-Gaussian fit to the OH spectrum gives the weighted average $\lsb \Delta F/F \rsb =
341: (2.25 \pm 0.84) \times 10^{-5}$. Similarly, the HI~21cm and main OH redshifts
342: in the $z \sim 0.685$ lens toward B0218+357 are $z_{\rm HI} = 0.684676 \pm 0.000005$
343: \cite{carilli00} and $z_{\rm OH} = 0.684682 \pm 0.0000056$ \cite{chengalur03},
344: giving $\lsb \Delta F/F \rsb_{\rm D} =
345: (3.5 \pm 4.0) \times 10^{-6}$. All the above values are consistent with
346: the null hypothesis of no evolution in the different constants. Combining
347: results from the two absorbers (using the constrained fit and a weighted average),
348: we obtain $\lsb \Delta F/F\rsb = (0.44 \pm 0.36^{\rm stat} \pm 1^{\rm syst}) \times 10^{-5}$,
349: over $ 0 < z \lesssim 0.7$, where the second error is due to velocity offsets
350: between HI and OH lines, assuming a velocity dispersion of 3~km/s.
351: Of course, four measurements are far too few for a reliable estimate of
352: this error. However, the fact that the OH and HI redshifts are in reasonable
353: agreement within the measurement errors in all four cases (two of which have
354: errors $\lesssim 1 \times 10^{-5}$) suggests that systematic velocity offsets do
355: not dominate the accuracy of the measurement.
356:
357: The strong dependence of $F$ on $\alpha$ and $\mu$ ($F \propto \alpha^{3.14}$
358: and $F \propto \mu^{-1.57}$) implies a $2 \sigma$ sensitivity of
359: $\lsb \dal \rsb < 6.7 \times 10^{-6}$ or $\lsb \dmu\rsb < 1.4 \times 10^{-5}$
360: to fractional changes in $\alpha$ and $\mu$ from $z \sim 0.7$ (i.e. a
361: lookback time of $\sim 6.5$~Gyrs) to today, where we have added the errors in
362: quadrature (these sensitivities are not independent as we measure $\lsb \Delta F/F \rsb$).
363: Assuming linear evolution, these correspond to $2\sigma$
364: sensitivities of $(1/\alpha) \lsb \Delta \alpha/ \Delta t \rsb < 1.1 \times
365: 10^{-15}$~yr$^{-1}$ or $(1/\mu) \lsb \Delta \mu/\Delta t \rsb < 2.1 \times
366: 10^{-15}$~yr$^{-1}$, among the best present sensitivities to changes in
367: $\mu$. For comparison, \cite{ivanchik05} obtains
368: $\lsb \dmu \rsb < 1.48 \times 10^{-5}$ for $0 < z \lesssim 2.75$ while
369: \cite{tzanavaris05} obtains $\lsb \dmu\rsb \lesssim 2 \times 10^{-5}$
370: for $ 0 < z < 2.0$, at $2\sigma$ level, using redshifted optical lines.
371: The present radio analysis is not affected by two important sources
372: of systematic error in the optical regime (wavelength
373: calibration and relative isotopic abundances; e.g. \cite{murphy03}),
374: although, of course, the optical observations probe a larger redshift range.
375: The primary source of the error in our technique is likely to lie in the
376: velocity dispersion
377: between OH and HI lines and perhaps in blending between weak narrow
378: spectral components. We do not feel that these dominate the present
379: results; deeper observations in the OH and HI lines should help
380: quantify their effects.
381:
382: While the size of the radio sample is still small, surveys are being carried
383: out that will significantly increase the number of known redshifted OH, HI and
384: HCO$^+$ absorbers. Comparisons between radio lines are thus likely to provide
385: an important independent constraint on changes in fundamental constants in the
386: future.
387:
388:
389: \begin{acknowledgments}
390: We thank Bob Garwood and Jim Braatz for help with the data analysis.
391: The National Radio Astronomy Observatory is operated by Associated Universities,
392: Inc., under cooperative agreement with the National Science Foundation. The
393: Australia Telescope Compact Array is part of the Australia Telescope, funded by
394: the Commonwealth of Australia for operation as a National Facility managed by CSIRO.
395:
396:
397: \end{acknowledgments}
398:
399: \bibliography{ms4}
400:
401:
402:
403: \end{document}
404: