astro-ph0410457/ms.tex
1: \documentclass[12pt,preprint]{aastex}
2: 
3: %\usepackage{natbib,psfig}
4: %% If you want to create your own macros, you can do so
5: %% using \newcommand. Your macros should appear before
6: %% the \begin{document} command.
7: %%
8: %% If you are submitting to a journal that translates manuscripts
9: %% into SGML, you need to follow certain guidelines when preparing
10: %% your macros. See the AASTeX v5.0 Author Guide
11: %% for information.
12: \renewcommand{\floatpagefraction}{0.9}
13: \renewcommand{\textfraction}{0.00}
14: \newcommand{\vdag}{(v)^\dagger}
15: \newcommand{\myemail}{kmb@astro.ox.ac.uk}
16: \newcommand{\gtsim}{\mbox
17: {{\raisebox{-0.4ex}{$\stackrel{>}{{\scriptstyle\sim}}$}}}}
18: \newcommand{\ltsim}{\mbox
19: {{\raisebox{-0.4ex}{$\stackrel{<}{{\scriptstyle\sim}}$}}}}
20: \newcommand{\mc}{\multicolumn}
21: %% You can insert a short comment on the title page using the command below.
22: 
23: \slugcomment{Revised version \today}
24: 
25: \shorttitle{Changes in jet speed and cone angle in SS433}
26: \shortauthors{Blundell \& Bowler}
27: 
28: \begin{document}
29: 
30: \title{Jet velocity in SS433: its anti-correlation with
31:   precession-cone angle and dependence on orbital phase}
32: 
33: \author{Katherine M.\ Blundell\altaffilmark{1} and Michael G.\
34:   Bowler\altaffilmark{1}} 
35: 
36: \altaffiltext{1}{University of Oxford, Department of Physics, Keble
37:   Road, Oxford, OX1 3RH, U.K.}
38: 
39: \begin{abstract}
40: We present a re-analysis of the optical spectroscopic data
41: on SS\,433 from the last quarter-century and demonstrate
42: that these data alone contain systematic and identifiable
43: deviations from the traditional kinematic model for the
44: jets: variations in speed, which agree with our analysis
45: of recent radio data; in precession-cone angle and in
46: phase. We present a simple technique for separating out the
47: jet speed from the angular properties of the jet axis,
48: assuming only that the jets are symmetric. With this
49: technique, the archival optical data reveal
50: that the variations in jet speed and in precession-cone
51: angle are anti-correlated in the sense that when faster jet
52: bolides are ejected the cone opening angle is smaller.  We
53: also find speed oscillations as a function of {\em orbital}
54: phase.
55: \end{abstract}
56: 
57: \keywords{stars: binaries: stars:  individual: (SS\,433)}  
58: 
59: \section{Introduction}
60: \label{sec:intro}
61: 
62: In a recent paper \citep{Blu04} we presented the deepest yet radio
63: image of SS\,433, which revealed an historical record over two
64: complete precession periods of the geometry of the jets.
65: Detailed analysis of this image revealed systematic deviations from 
66: the standard kinematic model \citep{Mar84,Eik01}.  Variations
67: in jet speed, lasting for as long as tens of days, were needed to
68: match the detailed structure of each jet.  Remarkably, these
69: variations in speed were equal, matching the two jets
70: simultaneously.
71: 
72: The Doppler residuals to the kinematic model
73: show little variation with the precessional phase of the jets and this
74: observation rules out variations in jet speed {\em alone} as the
75: source of the residuals \citep[e.g.\, ][]{Kat82a,Eik01}.  Very little
76: phase variation is obtained if the pointing angle jitters
77: \citep{Kat82a} but there is no evidence excluding symmetric speed
78: variations of the magnitude reported in \cite{Blu04} superposed on
79: pointing jitter, Fig\,\ref{fig:varywithphase}.
80: Thus our findings from the radio image led us to re-analyse the
81: archival optical data.  G.\ Collins II and S.\ Eikenberry (with kind
82: permission of B.\ Margon) made available their compiled
83: datasets, published in \cite{Col00} and used in \cite{Eik01}.  We use
84: the Collins' compilation (available at
85: http://www-astro.physics.ox.ac.uk/$\sim$kmb/ss433/) because of its
86: higher quoted precision, but very similar results are obtained from
87: Margon's. 
88: 
89: \section{Speed and angular variations from the optical data}
90: \label{sec:variations}
91: 
92: If the precessing jet axis of SS\,433 traces out a cone of semi-angle
93: $\theta$ about a line which is oriented at an angle $i$ to our
94: line-of-sight with jet velocity $\beta$ in units of $c$ ($\gamma = (1
95: - \beta^2)^{-1/2}$), the redshifts measured from the west jet
96: ($z_{+}$) and the east jet ($z_{-}$) are given, if the jets are
97: symmetric, by:
98: \begin{equation}
99: \label{eq:redshift}
100:   z_{\pm} = -1 + \gamma[1 \pm \beta\sin{\theta}\sin{i}\cos{\phi} \pm
101:   \beta\cos{\theta}\cos{i}],
102: \end{equation}
103: where $\phi$ is the phase of the precession cycle (see
104: http://www-astro.physics.ox.ac.uk/$\sim$kmb/ss433/).  Addition of
105: $z_{+}$ and $z_{-}$ in Eqn\,\ref{eq:redshift} gives an expression
106: relating the observed redshifts to the jet speed independently of any
107: angular variation.  Re-arrangement gives
108: \begin{equation}
109: \label{eq:sum}
110:   \beta = \left[ 1 - \left[1 + \frac{z_{+} +  z_{-}}{2}\right]^{-2}  \right] ^{1/2}.
111: \end{equation}
112: The quantity $z_{+} + z_{-}$ fluctuates very
113: substantially (see Fig\,\ref{fig:resids_v_time}). 
114: Subtraction of the expressions for $z_{+}$ and $z_{-}$ gives the angular properties $a$ of the orientation of the
115: jet axis, with the speed divided out using
116: Eqn\,\ref{eq:sum}:
117: \begin{equation}
118: \label{eq:ang}
119: a =   \frac{z_{+} -  z_{-}}{2\beta\gamma} = \sin{\theta}\sin{i}\cos{\phi}
120:   + \cos{\theta}\cos{i}, 
121: \end{equation}
122:   
123: Fluctuations in $\beta$, $\theta$ and $\phi$ are predominantly
124: symmetric, as described by Eqns 1--3, when fluctuations in $z_{+} +
125: z_{-}$ represent symmetric fluctuations in speed.  (The velocity
126: variations in the two radio jets \citep{Blu04} are highly
127: symmetric; the standard deviation on the difference in the speeds is
128: less than $0.004\,c$ and on the common velocity $0.014\,c$.)
129: 
130: The disadvantages of the variables $s = z_{+} + z_{-}$ and $a$
131: (Eqn\,\ref{eq:ang}) are that their interpretation is simple only for
132: perfect symmetry and that they may only be used for the 395 out of 486
133: observations which record a simultaneous pair.  Their merits are
134: exemplified by Fourier analyses of the time distributions.  We used
135: the algorithm of \cite{Rob87} which accounts for the uneven
136: time-sampling of the data.  The angular data $a$ clearly revealed
137: periodicities corresponding to the nodding of the precession axis
138: \citep{Kat82,New82,Col02} and the 162-day precession period, clearly
139: seen in Fig\,\ref{fig:ft}a.  There is no periodicity in the speed data
140: $s$ (Fig\,\ref{fig:ft}b) common to the angular data,
141: consistent with perfect symmetry.  The speed data also indicate a
142: periodicity at 13.08\,days (the periodicity at 12.58\,days matches a
143: beat with Earth's orbital period 365\,days); to investigate this, we
144: folded the data over 13.08\,days in 20 phase bins, and in each bin the
145: mean speed ($\beta$, from Eqn\,\ref{eq:sum}) was
146: derived. Fig\,\ref{fig:fold} shows a clear sinusoidal oscillation with
147: orbital phase.  The rms variation in speed which oscillates with
148: orbital phase is smaller by a factor of three than the overall speed
149: dispersion.  This oscillation with amplitude $2000\,{\rm km\,s^{-1}}$
150: may be because the speed with which the bolides are ejected is a
151: function of orbital phase, but the excursions in Fig\,\ref{fig:fold}
152: could also be interpreted as due to orbital motion; in that case
153: SS\,433's orbital speed is $\sim 400\, {\rm km\, s}^{-1}$.
154: 
155: \section{Anti-correlated deviations in jet speed and $\theta$}
156: \label{sec:correlations}
157: 
158: We fitted Collins' dataset with the kinematic model, including
159: nodding.  From our fit, we derived model redshift pairs and hence the
160: variables $s$ and $a$.  Subtraction of these model variables from
161: those constructed from the data gave residuals $\Delta s$ in $s$ and
162: $\Delta a$ in $a$.  The $\beta$ variation is shown in
163: Fig\,\ref{fig:resids_v_time}c.  The standard deviation of this
164: histogram is 0.013, in excellent agreement with the result from our
165: radio image (0.014).  Examples of the residuals in $s$ and in the
166: angular variable $a$ are plotted in Fig\,\ref{fig:resids_v_time}; the
167: variations in speed and angular residuals are anti-correlated.  From
168: Eqns\,\ref{eq:redshift}--\ref{eq:ang}, maintaining the assumption of
169: symmetry:
170: 
171: \begin{equation}
172: \label{eq:s_2}
173:    \Delta s^2  = 4 \beta^2 \gamma^6  \Delta
174:   \beta^2 ,
175: \end{equation}
176: \begin{eqnarray}
177: \label{eq:a_2}
178:    \Delta a^2  &=& (\cos\theta \sin{i} \cos\phi -
179:   \sin\theta \cos{i})^2   \Delta  \theta^2  +
180:   (\sin\theta \sin{i} \sin\phi)^2  \Delta\phi^2   \\ \nonumber
181:   &-&  2 \sin\theta \sin{i} \sin\phi (\cos\theta \sin{i} \cos\phi -
182:   \sin\theta \cos{i})   \Delta  \theta \Delta \phi ,
183: \end{eqnarray}
184: \begin{eqnarray}
185: \label{eq:as}
186:    \Delta a\,\Delta s  &=& 2 \beta \gamma^3 
187: [(\cos\theta \sin{i} \cos\phi - \sin\theta \cos{i}) 
188:   \Delta\beta\,\Delta\theta  \\ \nonumber
189:  &-& \sin\theta \sin{i} \sin\phi  \Delta\beta \Delta\phi],
190: \end{eqnarray}
191: where $\Delta \beta$, $\Delta \theta$ and $\Delta \phi$ represent the
192: variations in $\beta$, $\theta$ and $\phi$ respectively. Averaging
193: over many cycles for any given value of the 
194: phase $\phi$ yields the averages $\langle \Delta a^2 \rangle$,
195: $\langle \Delta s^2 \rangle$, $\langle \Delta a \Delta s \rangle$ as
196: functions of $\phi$, in terms of the parameters $\langle \Delta
197: \beta^2 \rangle$, $\langle \Delta \theta^2 \rangle$, $\langle
198: \Delta\beta\,\Delta\theta \rangle$ and so on.  The fit to $\langle
199: \Delta\beta\,\Delta\theta \rangle$ is shown in Fig\,\ref{fig:a_s}.
200: The parameters are given in Table\,\ref{tab:fits}; the global $\chi^2
201: / NDF$ is 34.8/24, where $NDF$ is the number of degrees of freedom.  
202: 
203: Fig\,\ref{fig:a_s} shows that the quantity $\langle \Delta a\,\Delta s
204: \rangle$ has an almost pure cosinusoidal variation with phase, as
205: given by the first term on the right hand side of Eqn\,\ref{eq:as}.
206: This shape is the unique signature of a correlation between variations
207: in $\beta$ and $\theta$.  $\langle \Delta \beta^2 \rangle$ shows no
208: correlation with $\phi$ and $\langle \Delta a^2 \rangle$ very little.
209: The latter requires fluctuations in both $\theta$ and in $\phi$. We
210: remark that $\langle \Delta (z_{+} - z_{-})\,\Delta s \rangle $ does
211: not show a strong correlation with $\phi$; nor should it, using the
212: parameters from Table 1.  Removing the varying speed from $z_{+} -
213: z_{-}$ was crucial in revealing this correlation in $\langle \Delta
214: a\,\Delta s \rangle$.
215: 
216: 
217: \section{The redshift residual plot}
218: \label{sec:eikfigfive}
219: 
220: Consider the plane of redshift residuals, as in figure~5 of
221: \cite{Eik01}, with symmetric excursions from the kinematic model in
222: $\beta$, $\theta$ and $\phi$.  Comparison of their figure~5 (similar
223: to our Fig\,\ref{fig:eikfigfive}f) with our
224: Figs\,\ref{fig:eikfigfive}a and \ref{fig:eikfigfive}b requires the
225: presence of both angular variations (to spread the points along the
226: line $y = -x$) and velocity variations (to spread the points
227: perpendicular to this line --- note that even their quoted redshift
228: measurement error of 0.003, likely an over-estimate, will not account
229: for this breadth).  Inclusion of all these variations, correlated as
230: in Table\,\ref{tab:fits}, gives Fig\,\ref{fig:eikfigfive}d which
231: resembles that from the data (Fig\,\ref{fig:eikfigfive}f).  The
232: simulations in Fig\,\ref{fig:eikfigfive}d take no account of the
233: (stochastic) duration of the excursions; the duration of variations in
234: Fig\,\ref{fig:eikfigfive}e were drawn from a gaussian with half-width
235: 2\,days.
236: 
237: Thus Figs\,\ref{fig:varywithphase} and \ref{fig:eikfigfive} establish
238: the consistency of the optical data with perfect symmetry and with
239: speed fluctuations whose magnitude is in excellent agreement with
240: those found in the radio image.  In addition, the slope of
241: Fig\,\ref{fig:eikfigfive}e is $-0.765 \pm 0.034$, and that for
242: Fig\,\ref{fig:eikfigfive}f from the Collins data set is $-0.786$ and
243: not $-1$; this long standing curiosity is explained by the physics
244: from the \S\,\ref{sec:correlations} fit.
245: 
246: \section{Other assumptions}
247: \label{sec:other}
248: 
249: If the jet speed were constant, the residuals to $z_{+}$ and $z_{-}$
250: would, for the case of strictly antiparallel jets, be correlated as in
251: Fig\,\ref{fig:eikfigfive}a; to spread the distribution of points
252: perpendicular to this line it is necessary to allow some independence
253: in the pointing of the two jets.  The variation with precessional
254: phase of the quantities $\langle \Delta s^2\rangle$, $\langle \Delta
255: a^2\rangle$ and $\langle \Delta a \Delta s\rangle$ is then almost as
256: well described ($\chi^2 / NDF = 36.3/24$) as by our fit of
257: \S\,\ref{sec:correlations}.  Such a model would not be able to explain
258: the radio jet morphology \citep{Blu04} and this fit required angular
259: fluctuations breaking symmetry to have rms values $\sim 1/2$ of those
260: preserving symmetry.  Angular jitter alone cannot account for the
261: slope of $\Delta z_{-}$ versus $\Delta z_{+}$
262: (Fig\,\ref{fig:eikfigfive}) differing from $-1$, unless the jitter in
263: the East jet is systematically smaller than in the West jet.
264: Velocity variation breaks symmetry in the Doppler shifts and accounts
265: naturally for this observation.  A better fit than either
266: ($\chi^2/NDF$ = 25.5/21) was achieved by allowing some symmetry
267: breaking angular fluctuations in addition to symmetric velocity
268: fluctuations: in this case the rms symmetry breaking fluctuations were
269: $\sim 1/4$ of those preserving symmetry.  In all cases the rms
270: fluctuations in $\theta$ were $\sim 1/3$ of the rms fluctuations in
271: $\phi$, as would be expected for pointing angle fluctuations described
272: by \cite{Kat82a} as ``isotropic''.
273: 
274: \section{Concluding remarks}
275: 
276: Archival optical spectroscopic data on SS\,433 reveal variations in
277: jet speed, in cone opening angle, and in the phase of the precession.
278: These appear in the plane of redshift residuals
279: (Fig\,\ref{fig:eikfigfive}) and through the new (symmetry dependent)
280: technique of combining simultaneously observed redshift pairs for the
281: speed-only ($s$) and angular-only ($a$) characteristics
282: (Fig\,\ref{fig:resids_v_time}).  The velocity variations $\sim
283: 0.014\,c$ are strongly anticorrelated with cone angle $\theta$, in the
284: sense that when faster bolides are ejected the cone angle is smaller.
285: We also found smaller amplitude sinusoidal oscillations in speed as a
286: function of orbital phase.  If this is due to ejection speed, perhaps
287: the orbit of the binary is eccentric.  If these 13.08-day oscillations
288: are orbital Doppler shifts the orbital velocity is $\sim 400\,{\rm
289: km\, s}^{-1}$ (twice that inferred by \cite{Cra81,Fab90}). If this
290: were the case, the mass of the companion to SS\,433 would be $> 86\,
291: {\rm M}_\odot$ and if the mass fraction were 0.1, then the mass of the
292: companion would be \gtsim\ $100\, {\rm M}_\odot$.  Such masses would
293: be hardly consistent with an A-type companion \citep{Gei02,Cha04}.
294: 
295: \acknowledgments
296: 
297: K.M.B.\ thanks the Royal Society for a University Research Fellowship.
298:   It is a pleasure to thank
299: Avinash Deshpande, James Binney \& Philipp Podsiadlowski for helpful
300: discussions.
301: 
302: \begin{thebibliography}{}
303: 
304: \bibitem[Blundell \& Bowler(2004)]{Blu04}
305: Blundell, K.~M., \& Bowler, M.~G.\ 2004, \apjl, 616, L159 
306: 
307: \bibitem[Charles et al(2004)]{Cha04}
308: Charles, P.A.\ et al, 2004, RevMexAA, 20, 50
309: 
310: \bibitem[Crampton \& Hutchings(1981)]{Cra81} 
311: Crampton, D.~\&  Hutchings, J.B.\ 1981, \apj, 251, 604 
312: 
313: \bibitem[Collins \& Scher(2002)]{Col02}
314: Collins, G.W.\ \& Scher, R.W., 2002, \mnras, 336, 1011
315: 
316: \bibitem[Collins \& Scher(2000)]{Col00}
317: Collins, G.W.\ \& Scher, R.W., 2000, in ``the Kth Reunion'', ed
318: A.G.D.\ Philip, L.\ Davis Press, Schenectady, NY, pg105
319: 
320: \bibitem[Eikenberry et al(2001)]{Eik01} Eikenberry, S.S. et al, 
321: 2001, \apj, 561, 1027 
322: 
323: \bibitem[Fabrika \& Bychkova(1990)]{Fab90} 
324: Fabrika, S.~N.~\& Bychkova, L.~V.\ 1990, \aap, 240, L5 
325: 
326: \bibitem[Gies et al(2002)]{Gei02}
327: Gies, D.R., Huang, W.\ \& McSwain, M.V., 2002, \apj, 578, {\sc l}67 
328: 
329: \bibitem[Katz \& Piran(1982)]{Kat82a} 
330: Katz, J.~I.~\& Piran, T.\ 1982, \aplett, 23, 11 
331: 
332: \bibitem[Katz et al(1982)]{Kat82} 
333: Katz, J.~I., Anderson, S.~F., Grandi, S.~A., \& Margon, B.\ 1982, \apj, 
334: 260, 780 
335: 
336: \bibitem[Margon(1984)]{Mar84}
337: Margon, B.\ 1984, \araa, 22, 507
338: 
339: \bibitem[Newsom \& Collins(1982)]{New82}
340: Newsom, G.H.\ \& Collins, G.W., 1982, \apj, 262, 714 
341: 
342: \bibitem[Roberts et al(1987)]{Rob87}
343: Roberts, D.H., Lehar, J.\ \& Dreher, J.W., 1987, \aj, 93, 968
344: \end{thebibliography}
345: 
346: \clearpage
347: 
348: \begin{deluxetable}{llll}
349: \tablecaption{\label{tab:fits} 
350: Fits to excursions from the standard kinematic model (incorporating
351: nodding).
352: $\beta$ is in units of $c$, and $\theta$ and $\phi$ are in radians.}
353: \tablehead{\colhead{Quantity} & \colhead{fitted value} 
354: & \colhead{quantity} & \colhead{derived value}  \\ }
355: \startdata
356: $\langle \Delta \beta^2 \rangle     $  
357: & $\phantom{-}1.67 \pm 0.18 \times 10^{-4}$
358: & rms speed variation 
359: & 0.0129$c$
360: \\
361: $\langle \Delta \theta^2 \rangle    $  
362: & $\phantom{-}2.24 \pm 0.35 \times 10^{-3}$
363: & rms $\theta$ variation
364: & 2.71\,deg
365: \\
366: $\langle \Delta \phi^2 \rangle      $  
367: & $\phantom{-}1.31 \pm 0.26 \times 10^{-2}$
368: & rms $\phi$ variation
369: & 2.96\,days
370: \\
371: $\langle \Delta \beta\Delta\theta \rangle $  
372: &$-3.81 \pm 0.52 \times 10^{-4}$ &  & \\
373: $\langle \Delta \beta\Delta\phi \rangle   $  
374: & $\phantom{-}1.70 \pm 0.80 \times 10^{-4}$ &  & \\
375: $\langle \Delta \theta\Delta\phi \rangle$
376: & indistinguishable from zero & & \\
377: \enddata
378: \end{deluxetable}
379: 
380: \clearpage
381: 
382: \begin{figure}
383: \epsscale{0.5}
384: \plotone{f1.eps}
385: \figcaption{\label{fig:varywithphase} Demonstration that the
386:   dispersion in redshift with phase is unaffected by velocity
387:   variations as long as they are accompanied by pointing variations:
388:    {\bf (a)} random variations in velocity drawn from
389:   a gaussian of half-width 0.013$c$, {\bf (b)} random variations in
390:   $\theta$ and in phase drawn from gaussian distributions of 2.7\,deg
391:   and 2.95 days respectively and {\bf (c)} uncorrelated variations in
392:   all three of the quantities above.
393:  } 
394: \end{figure}
395: 
396: \clearpage
397: 
398: \begin{figure}
399: \epsscale{1.0}
400: \plotone{f2.eps}
401: \figcaption{\label{fig:resids_v_time} {\bf (a)} and {\bf (b)} examples
402: of angular and speed residuals versus time.  The variations suggest
403: anti-correlation, \S\,\ref{sec:correlations}.  {\bf (c)} Distribution
404: of speeds from the data compiled by Collins. The mean is indicated by
405: the vertical red line, and the standard deviation in $\beta$ is 0.013. }
406: \end{figure}
407: 
408: \clearpage
409: 
410: \begin{figure}
411: \epsscale{1.0}
412: \plotone{f3.eps}
413: \figcaption{\label{fig:ft} Fourier transform of data as described in
414:   \S\,\ref{sec:variations}.  {\bf (a)} The differences of all redshift
415:   pairs, divided by $2\beta \gamma$ (see Eqn\,\ref{eq:ang}) which, if
416:   the jets are symmetric, depend only on the angular properties of the
417:   jet.   {\bf (b)} The sums of all
418:   redshift pairs which, if the jets are symmetric, only depend on the
419:   speed of the jet.  }
420: \end{figure}
421: 
422: \clearpage
423: 
424: \begin{figure}
425: \epsscale{0.7}
426: \plotone{f4.eps}
427: \figcaption{\label{fig:fold} Speed data (Eqn\,\ref{eq:sum}) folded
428: over the orbital period of 13.08 days showing a clear sinusoid, mean
429: $0.2581 \pm 0.0005$, amplitude $0.0066 \pm 0.0007$, phase offset with
430: respect to optical ephemeris $2.17 \pm 0.11$\,rad.  }
431: \end{figure}
432: 
433: \clearpage
434: 
435: \begin{figure}
436: \epsscale{0.7}
437: \plotone{f5.eps}
438: \figcaption{\label{fig:a_s} The mean value of $\langle \Delta a \times
439:   \Delta s \rangle$ in 162-day period phase bins, showing the best fit
440:   to Eqn\,\ref{eq:as} averaged.  There are two free parameters in the
441:   fit shown as a solid line, $\langle \Delta\beta\,\Delta\theta
442:   \rangle$ and $\langle \Delta\beta\,\Delta\phi \rangle$. The dotted
443:   line shows the fit with fixed velocity but angular asymmetry.  }
444: \end{figure}
445: 
446: \clearpage
447: 
448: \begin{figure}
449: \epsscale{1.0}
450: \plotone{f6.eps}
451: \figcaption{\label{fig:eikfigfive} The effect on redshift residuals in
452:   each jet for deviations from the kinematic model in $\theta$,
453:   $\beta$ and $\phi$, for each date on which there is a redshift pair
454:   (replicating the sampling function).  {\bf (a)} Variations only in
455:   $\theta$, drawn from a gaussian distribution with half-width 2.7
456:   degrees.  The form of this plot, points lying only on the $y = -x$
457:   line, is identical for variations only in $\phi$, or for variations
458:   in both $\phi$ and $\theta$.  {\bf (b)} Speed-only variations (on a
459:   given day $\beta$ is drawn from a gaussian of half-width 0.013 ---
460:   Fig\,\ref{fig:resids_v_time}c). {\bf (c)} Variations in $\beta$ {\em
461:   anti-correlated} with those in $\theta$: on a given synthesized
462:   observation date the same randomly-drawn number from a gaussian is
463:   scaled by 0.013 for the speed variation and by $-2.7$ degrees for
464:   the $\theta$ variation.  {\bf (d)} As (c), but with $\theta$
465:   partially anti-correlated with $\beta$ and uncorrelated $\phi$
466:   fluctuations (as Table\,\ref{tab:fits}). {\bf (e)} As (d), but the
467:   duration of the $\beta$-variations, and that of the correlated
468:   $\theta$-component, is drawn from a gaussian of half-width
469:   2\,days. {\bf (f)} Observed residuals. }
470: \end{figure}
471: \end{document}
472: