astro-ph0002242/ms.tex
1: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
2: % AAS style double spaced
3: \documentstyle[12pt,aaspp4]{article}
4: 
5: % AAS style single spaced
6: %\documentstyle[12pt,aasms4]{article}
7: 
8: 
9: \pagestyle{headings}
10: %\pagestyle{empty}
11: %+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
12: %                       abbreviations
13: %+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
14: \newcommand{\uc}{\rm U Com~}
15: \newcommand{\cdo}{$^{12}C(\alpha ,\gamma)^{16}O\;$}
16: \newcommand{\dydz}{$\Delta Y / \Delta Z\;$}
17: \newcommand{\kms}{$\,km\,s^{-1}$}
18: \newcommand{\lsun}{log $L/L_{\odot}\,$}
19: \newcommand{\msun}{$M/M_{\odot}\,$}
20: 
21: %====================================================================
22: 
23: \begin{document}
24: 
25: \title{The RR Lyrae {\rm U Com} as a test for nonlinear pulsation models}
26: 
27: 
28: \author{Giuseppe Bono\altaffilmark{1}, Vittorio Castellani\altaffilmark{2}, 
29: Marcella Marconi\altaffilmark{3}}
30: 
31: \affil{1. Osservatorio Astronomico di Roma, Via Frascati 33,
32: 00040 Monte Porzio Catone, Italy; bono@coma.mporzio.astro.it}
33: \affil{2. Dipartimento di Fisica Universit\`a di Pisa, Piazza Torricelli 2,
34: 56100 Pisa, Italy; vittorio@astr18pi.difi.unipi.it}
35: \affil{3. Osservatorio Astronomico di Capodimonte, Via Moiariello 16,
36: 80131 Napoli, Italy; marcella@na.astro.it}
37: 
38: \begin{abstract}
39: We use high precision multiband photometric data of the first
40: overtone RR Lyrae \uc to investigate the predictive capability 
41: of full amplitude, nonlinear, convective hydrodynamical models.  
42: The main outcome of this investigation is that theoretical predictions 
43: properly account for the luminosity variations along a full pulsation 
44: cycle. Moreover, we find that this approach, due to the strong dependence 
45: of this observable and of the pulsation period on stellar parameters, 
46: supply tight constraints on stellar mass, effective temperature, and 
47: distance modulus. Pulsational estimates of these parameters appear 
48: in good agreement with empirical ones.
49: Finally, the occurrence of a well-defined bump just before the luminosity 
50: maximum gave the unique opportunity to calibrate the turbulent convection 
51: model adopted for handling the coupling between pulsation and convection.  
52: \end{abstract}
53: 
54: \keywords{stars: distances -- stars: evolution -- stars: horizontal branch -- 
55: stars: individual (\uc) -- stars: oscillations -- stars: variables: RR Lyrae}  
56: 
57: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
58: \pagebreak 
59: \section{Introduction} 
60:  
61: Variable stars play a key role in many astrophysical problems, 
62: since their pulsation properties do depend on stellar parameters, 
63: and therefore they can supply valuable and independent constraints 
64: on a large amount of current evolutionary predictions. 
65: In particular, the empirical evidence found long time ago in 
66: Magellanic Cepheids of the correlation between period and 
67: luminosity was the initial step for a paramount theoretical and 
68: observational effort aimed at using variable stars as standard 
69: candles to estimate cosmic distances. The current literature is still 
70: hosting a vivid debate on the intrinsic accuracy of the Cepheid distance 
71: scale (Bono et al. 1999; Laney 2000) and on the use of RR Lyrae 
72: variables to evaluate the distance -and the age- of Galactic globulars  
73: (Caputo 1998; Gratton 1998, G98). 
74: 
75: Theoretical insights into the problem of radial stellar pulsations 
76: came from the linearization of local conservation equations governing 
77: the dynamical instability of stellar envelopes. Linear, nonadiabatic 
78: models typically supply accurate pulsation periods and plausible 
79: estimates (necessary conditions) on the modal stability of the 
80: lowest radial modes. 
81: However, a proper treatment of radial pulsations does require the 
82: solution of the full system of hydrodynamic equations, including 
83: a nonlocal and time-dependent treatment of turbulent convection 
84: (TC) to account for the coupling between radial and convective 
85: motions (Castor 1968; Stellingwerf 1982, S82). 
86:  
87: The development of nonlinear, convective hydrocodes (S82; Gehmeyr 1992; 
88: Bono \& Stellingwerf 1994, BS; Wuchterl \& Feuchtinger 1998) 
89: gave the opportunity to provide plausible predictions on the properties 
90: of radial variables, and in particular on the topology of the instability 
91: strip, as well as on the time behavior of both light and radial velocity 
92: curves. This new theoretical scenario allowed to investigate, for the 
93: first time, the dependence of pulsation amplitudes and Fourier 
94: parameters on stellar mass, luminosity and effective temperature 
95: (see e.g. Kovacs \& Kanbur 1997; Brocato et al. 1996; Feuchtinger 2000, F20).
96: However, all these investigations dealt with parameters related to 
97: the light curve, whereas nonlinear computations supply much more 
98: information, as given by the detailed predictions of the light 
99: variation along a full pulsation cycle. Therefore, the direct 
100: comparison between observed and predicted light curves appears as 
101: a key test only partially exploited in the current literature 
102: (Wood, Arnold, \& Sebo 1997, WAS). 
103: 
104: In order to perform a detailed test of the predictive capability of our 
105: nonlinear, convective models we focused our attention on the photometric  
106: data collected by Heiser (1996, H96) for the field, first 
107: overtone -$RR_c$- variable \uc. The reason for this choice relies 
108: on the detailed coverage of the U, B, and V light curves, as well as   
109: on the characteristic shape of the light curve, with a well-defined 
110: bump close to the luminosity maximum. This secondary feature provides 
111: a tight observational constraint to be nailed down by theory.  
112: Since the period of the variable strongly depends on the structural 
113: parameters (mass, luminosity, and radius) of the pulsator, 
114: the problem arises whether or not nonlinear pulsation models 
115: account for the occurrence of similar pulsators, and in affirmative 
116: how precisely the observed light curves can be reproduced by 
117: theoretical predictions.
118: 
119: In \S 2 we present the comparison between theory and observations, 
120: while in \S 3 we discuss the calibration of the TC model. Finally,
121: in \S 4 we briefly outline the observables which can further 
122: validate this theoretical scenario. 
123:   
124: 
125: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
126: 
127: \section{Comparison between theory and observations}
128: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
129: 
130: On the basis of spectroscopic measurements Fernley \& Barnes (1997, FB97) 
131: estimated for \uc a metallicity $[Fe/H]=-1.25\pm0.20$, 
132: while Fernley et al. (1998, FB98) found a negligible interstellar 
133: extinction ($E(B-V)=0.015\pm0.015$).  
134: According to these empirical evidence and to a well-established 
135: evolutionary scenario, we expect for a metal-poor RR Lyrae a 
136: stellar mass of the order of 0.6 $M_\odot$ and a luminosity 
137: ranging from \lsun=1.6 to 1.7. At the same time, pulsation 
138: predictions on double-mode pulsators suggest similar mass values 
139: (Cox 1991; Bono et al. 1996). 
140: As a consequence, we computed a sequence of nonlinear models at fixed 
141: chemical composition (Y=0.24, Z=0.001) and pulsation period 
142: $P_{FO}\approx0.29$ d. 
143: Along such an iso-period sequence the individual models were constructed 
144: at fixed mass value (\msun = 0.60), while both the luminosity and the 
145: effective temperature were changed according to the pulsation relation 
146: given by Bono et al. (1997, BCCM). 
147: Both linear and nonlinear models were computed by adopting the input
148: physics, and physical assumptions already discussed in BS, BCCM, 
149: and in Bono, Marconi \& Stellingwerf (1999). According to S82, 
150: current models were computed by assuming a vanishing efficiency 
151: of turbulent overshooting in the region where the superadiabatic 
152: gradient attains negative values. This means that the convective flux 
153: can only attain positive or vanishing values ($F_c \ge 0$). 
154: In the next section we show that the assumption adopted 
155: in our previous investigations -i.e. $F_c$ can attain both 
156: positive and negative values- marginally affects the topology 
157: of the instability strip, but the predicted light curves are 
158: somewhat at variance with empirical ones.    
159: 
160: The top panels of Fig. 1 show that at fixed stellar mass (\msun=0.60), 
161: the double peak feature appears in models characterized by 
162: luminosities approximately equal to \lsun$\approx$1.61, and effective 
163: temperatures ranging from 6950 to 7150 K. These models also present 
164: {\rm B} amplitudes in reasonable agreement with empirical estimates 
165: ($A_B=0.64$ mag). The bottom panels of Fig. 1 display the light 
166: curves of models 
167: along the iso-period sequence constructed by adopting a fixed effective 
168: temperature ($T_e=7100$ K) but different assumptions on stellar mass 
169: and luminosity. A glance at these curves shows that the luminosity 
170: amplitude is mainly governed by the stellar mass, whereas the shape 
171: of the light curve is only marginally dependent on this parameter. 
172: 
173: We find that the best fit to the observed B light curve is obtained 
174: for \msun=0.6, \lsun=1.607, $T_e=7100$ K, $P_{FO}=0.290$ d, together 
175: with a distance modulus $(m_B-M_B) = 11.01$ mag. The fit -though not perfect- 
176: appears rather satisfactory, thus suppling a substantial support 
177: to the predictive impact of the adopted theoretical scenario.
178: On the basis of this finding, we are now interested in testing the 
179: accuracy of theoretical predictions in different photometric bands. 
180: Fig. 2 shows from left to right the comparison between predicted 
181: light curves (solid lines) and empirical data (open circles) in the 
182: U, B, V, and K band respectively. The comparison was performed by 
183: adopting the same distance modulus, i.e. by neglecting 
184: the interstellar extinction, and the agreement between theory and 
185: observations seems even better than for the B light curve.  
186: Thus suggesting that nonlinear models account for luminosity 
187: amplitudes which are a long-standing problem of pulsation theory.   
188: 
189: 
190: Not surprisingly, we also find that the time average colors predicted 
191: by our model appear, within current uncertainty on both  reddening and 
192:  photometry, in very good agreement with empirical estimates 
193: (see Table 1). This result supports the evidence that nonlinear models, 
194: at least in this case, can constrain stellar colors by best fitting 
195: the light curve in a single photometric band. At the same time,
196: this agreement suggests that the pulsational constraints on the 
197: temperature of the pulsator, as derived by the B light curve, are
198: consistent with the theoretical light curves in the other photometric 
199: bands.
200: 
201: \small 
202: \begin{center}
203: \begin{tabular}{cccc}
204: \tablewidth{0pt}\\  
205: \multicolumn{4}{c}{TABLE 1. {\rm U Com}: theoretical and empirical
206: colors$^a$}\\
207: \hline 
208: Color    & Theory        & Observ.      & Observ.          \\   
209: % mag      &              & $E_{(B-V)}=0$ & $E_{(B-V)}=0.015$ \\   
210:  mag      &              & $E(B-V)=0$ & $E(B-V)=0.015$ \\   
211: \hline 
212: $<U>-<B>$ &$0.06\pm0.01$ & $0.11\pm0.02$    & $0.09\pm0.02$  \\ 
213: $<B>-<V>$ &$0.23\pm0.01$ & $0.21\pm0.02$    & $0.19\pm0.02$  \\ 
214: $<V>-<K>$ &$0.77\pm0.01$ & $0.81\pm0.07$    & $0.77\pm0.07$  \\ 
215: \hline 
216: \end{tabular}
217: \end{center}
218: \normalsize 
219: \begin{minipage}{1.00\linewidth}  
220: \noindent $^a$ Empirical estimates are based on photometric 
221: data collected by H96 and by FB97. Theoretical colors refer to the best 
222: fit model, and the errors were estimated by assuming an uncertainty of 
223: 50 K in the temperature of this model.  
224: \end{minipage} 
225: 
226: However, we note that on the basis of both the period and the 
227: shape of the {\rm B} light curve we predicted the effective temperature, 
228: the intrinsic luminosity, and in turn the distance modulus of this object. 
229: The plausibility of the theoretical constraints can be further tested
230: by comparing them with independent evaluations available in the literature. 
231: We find that the effective temperature predicted by nonlinear models 
232: -$T_e=7100\pm50$ K- is in remarkable agreement with the empirical 
233: temperature -$T_e=7100\pm150$ K- derived by adopting the true intensity 
234: mean color $(<V>-<K>)_0=0.77\pm0.07$ provided by FB97 and the 
235: CT relation by Fernley (1989). The same outcome applies by assuming 
236: E(B-V)=0, and indeed 
237: $(<V>-<K>)=0.81\pm0.07\,\rightarrow\,T_e=7050\pm150$ K, while the 
238: semi-empirical estimate provided by H96 suggests $T_e=7250\pm150$ K. 
239: We also note that the effective gravity of the best fit model 
240: (log $g\approx3.0$) is also in very good agreement with both the 
241: photometric estimate obtained by H96 (log $g=3.1\pm0.2$) and the 
242: spectroscopic measurements for field RR Lyrae variables provided 
243: by Clementini et al. (1995) and by Lambert et al. (1996). 
244: 
245: As far as the distance modulus is concerned, Fig. 3 shows the comparison 
246: of our pulsational estimates (filled circles) with empirical and 
247: theoretical \uc absolute magnitudes obtained by adopting different 
248: methods. The top and the bottom panel refer to absolute magnitudes  
249: based on visual and NIR magnitudes respectively. Data plotted in 
250: this figure show that our estimates appear in satisfactory  
251: agreement with distances based on the Baade-Wesselink method,  
252: on the statistical parallax method, on Hipparcos trigonometric 
253: parallaxes and proper motions, and on RR Lyrae {\rm K} band 
254: PL relation. 
255: 
256: However, one also finds that the distance determination based on 
257: HB evolutionary models constructed by including the most recent
258: input physics (Cassisi et al. 1999, C99) seems to overestimate 
259: the distance 
260: modulus by approximately 0.18 mag when compared with the current 
261: pulsation determination. A disagreement between evolutionary and 
262: pulsation predictions concerning the luminosities of RR Lyrae 
263: stars was brought out by Caputo et al. (1999), and more recently by 
264: Castellani et al. (2000) who found that up-to-date
265: He-burning models seem too bright when compared with Hipparcos 
266: absolute magnitudes. Data plotted in Fig. 3 confirm this discrepancy
267: between evolutionary and pulsational predictions, possibly due to 
268: an overluminosity of HB models. Part of this discrepancy might be 
269: due to the higher temperature of U Com when compared with the mean 
270: temperature of RR Lyrae gap ($T_e=6800$ K) adopted in evolutionary 
271: estimates.  
272: 
273: Finally, we also constructed several sequences of models by increasing
274: or decreasing the metal abundance by a factor of two. We find that the 
275: bump becomes more (less) evident at lower (higher) metal contents, and 
276: that this change does not allow us to obtain a good fit between 
277: theory and observations. This result is in satisfactory  agreement with 
278: the spectroscopic estimate by FB97 and supports the evidence that the 
279: shape of the light curve can also be used to constrain the $RR_c$ 
280: metallicity (Bono, Incerpi, \& Marconi 1996).  
281: 
282: 
283: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
284: \section{Calibration of the TC model}
285: 
286: The first set of models we constructed for fitting the empirical 
287: light curves was characterized by an unpleasant feature: 
288: the peak of the bump was, in contrast with empirical evidence, brighter
289: than the "true" luminosity maximum. According to Bono \& Stellingwerf
290: (1993) the bump along the rising branch presents a strong dependence 
291: on the free parameters adopted in the TC model. However, in the 
292: calibration of the TC model suggested by BS both the eddy viscosity 
293: and the diffusion scale lengths (see their equ. 4 and 9) were scaled 
294: to the value of the mixing length parameter. We performed several 
295: numerical experiments by 
296: changing along each sequence only one of the three free parameters. 
297: As a result, we find that full amplitudes models constructed by adopting 
298: plausible changes of the free parameters do not simultaneously account 
299: for the pulsation amplitude, the shape of the light curve, and the 
300: temperature width of first overtone instability region. 
301: 
302: Due to the lack of a self-consistent theory of time-dependent, nonlocal, 
303: convective transport, current investigations were mainly aimed at 
304: calibrating the free parameters adopted for treating the coupling 
305: between pulsation and convection (Yecko et al. 1998; F20). This is not 
306: a trivial effort, since the observables and the comparison between theory 
307: and observations are affected by the thorny problem of the transformation 
308: into the observational plane and/or by systematic deceptive errors 
309: such as reddening and distance estimates. 
310: In order to overcome some of these difficulties, F20 calibrated  
311: the TC model by performing a detailed comparison between theoretical 
312: and observed luminosity amplitudes of field RR Lyrae variables. On the basis 
313: of the fine tuning of both mixing length and turbulent viscosity length, 
314: F20 found that the Fourier parameters of fundamental light curves agree 
315: with observational data. The same outcome did not apply to $RR_c$ 
316: variables, and indeed predicted values appear, at fixed period, 
317: smaller than the empirical ones.  
318: A detailed comparison with the convective 
319: structure of RR Lyrae models constructed by F20 is not possible because 
320: he adopted a convective flux limiter in the turbulent source function 
321: and in the convective flux enthalpy, and neglected both the turbulent 
322: pressure and the turbulent overshooting. As a consequence, we decided 
323: to test the dependence of full amplitude models on the last two 
324: ingredients. 
325: 
326: Interestingly enough, we find that the models constructed
327: by assuming a vanishing overshooting efficiency satisfy empirical 
328: constraints i.e. the bump is dimmer than the luminosity maximum, the 
329: luminosity amplitudes attain values similar to the observed ones 
330: and the temperature width of the region in which the first overtone 
331: is unstable agrees with BCCM findings for cluster $RR_c$ variables. 
332: Fig. 4 shows the 
333: {\rm B} light curve of two models constructed by adopting 
334: the same input parameters, but different assumptions on the efficiency 
335: of turbulent overshooting. It is noteworthy the simultaneous change in 
336: the pulsation amplitude and in the shape of the light curve. 
337: The dependence of the convective structure on both overshooting 
338: and turbulent pressure will be investigated in a forthcoming paper.  
339: 
340: 
341: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
342: \section{Discussion and Conclusions} 
343: 
344: The comparison between theory and observations, namely the period and 
345: the shape of the {\rm B} light curve of U Com, allowed us to supply 
346: tight constraints on the structural parameters such as stellar mass, 
347: effective temperature, and gravity, as well as on the distance of this 
348: variable. We also found that the occurrence of a well-defined bump 
349: close to the luminosity maximum can be safely adopted for 
350: constraining the metallicity of this object and for calibrating 
351: the TC model adopted for handling the coupling between convection 
352: and pulsation. 
353: 
354: The approach adopted in this investigation seems quite promising since 
355: it only relies on nonlinear, convective models and on stellar atmosphere 
356: models. In fact, the best fit model to the empirical data was found by 
357: constructing sequences of iso-period models in which the stellar mass, 
358: the luminosity and the effective temperatures were not changed according 
359: to HB models but to the pulsation relation (BCCM).  
360: The comparison between theory and observations  
361: shows that both the structural parameters and the distance are 
362: in very good agreement with estimates available in the literature.  
363: No evidence for a systematic discrepancy was found in pulsation estimates, 
364: thus supporting the evidence that the individual fit to light curves  
365: can supply independent and firm constraints on the actual parameters 
366: and distances of variable stars. This finding confirms the results 
367: of a similar analysis on a LMC Bump Cepheid by WAS.  
368: 
369: Accurate radial velocity data for \uc are not available in the 
370: literature and therefore we could not constrain the accuracy of the 
371: velocity variation along the pulsation cycle. The three radial velocity 
372: points collected by FB98 agree quite well with the 
373: predicted curve. However, the radial velocity curve is a key 
374: observable for constraining the consistency of the adopted TC 
375: model (F20), and therefore new spectroscopic measurements 
376: of \uc would be of great relevance for assessing the 
377: predictive impact of nonlinear, convective models. 
378: Theoretical observables of the best fit model discussed in this 
379: paper, as well as both radius and radial velocity variations 
380: are available upon request to the authors. 
381: 
382: 
383: %\acknowledgments
384: It is a pleasure to thank M. Groenewegen for providing us the
385: \uc distances based on the reduced parallax method and on 
386: the modified Lutz-Kelker correction, as well as for insightful 
387: discussions on their accuracy. We are indebted to R. Garrido 
388: for sending us radial velocity data and to T. Barnes for useful 
389: suggestions on current data. We also acknowledge an anonymous 
390: referee for some useful suggestions that improved the readability 
391: of the paper. This work was supported by MURST -Cofin98- under the 
392: project "Stellar Evolution". Partial support by ASI and CNAA is 
393: also acknowledged.  
394: 
395: 
396: \begin{references}
397: \reference{} Barnes, T. G., \&  Hawley, S. L. 1986, \apj, 307, L9 (BH)  
398: \reference{} Bono, G., Caputo, F., Castellani, V., \& Marconi, M. 1996, \apj, 471, L33 
399: \reference{} Bono, G., Caputo, F., Castellani, V., \& Marconi, M. 1997,
400: \aaps, 121, 327 (BCCM)  
401: \reference{} Bono, G., Caputo, F., Castellani, V., \& Marconi, M. 1999, \apj, 512, 711  
402: \reference{} Bono, G., Incerpi, R., \& Marconi, M. 1996, \apj, 467, L97 
403: \reference{} Bono, G., Marconi, M., \& Stellingwerf, R. F. 1999, \apjs, 22, 167
404: \reference{} Bono, G., \& Stellingwerf, R. F. 1993, Soc. Astr. Italiana Mem., 64, 559
405: \reference{} Bono, G., \& Stellingwerf, R. F. 1994, ApJS, 93, 233 (BS) 
406: \reference{} Brocato, E., Castellani, V. \& Ripepi, V. 1996, \aap, 111, 809
407: \reference{} Caputo, F. 1998, \aapr, 9, 33   
408: \reference{} Caputo, F., Castellani, V., Marconi, M., \& Ripepi, V. 1999,
409: \mnras, 306, 815    
410: \reference{} Carney, B. W., Fulbright, J. P., Terndrup, D. M., Suntzeff, N. B.,
411:  \& Walker, A. R. 1995, \aj, 110, 1674 (C95)  
412: \reference{} Cassisi, S., Castellani, V., Degl'Innocenti, S., Salaris, M. \& Weiss, A. 1999, \aaps, 134, 103 (C99)
413: \reference{} Castellani, V., Degl'Innocenti, S., Girardi, L., Marconi, M.,
414:  Prada Moroni, P. G., \& Weiss, A. 2000 \aap, in press 
415: \reference{} Castelli, F., Gratton, R. G., \& Kurucz, R. L. 1997, \aap, 324, 432
416: \reference{} Castor, J. I. 1968,  unpublished  
417: \reference{} Clementini, G., Carretta, E., Gratton, R., Merighi, R. 1995, \aj, 110, 2319
418: \reference{} Cox, A. N. 1991, \apj, 381, L71
419: \reference{} Fernley, J. 1989, MNRAS, 239, 905
420: \reference{} Fernley, J., \& Barnes, T. G. 1997, \aaps, 125, 313 (FB97)  
421: \reference{} Fernley, J., et al. 1998, \aap, 330, 515 (FB98)  
422: \reference{} Fernley, J., Carney, B. W., Skillen, I., Cacciari, C., \&
423: Janes, K. 1998, \mnras, 293, L61 (FC98)  
424: \reference{} Fernley, J., Skillen, I., \& Burki, G. 1993, \aaps, 97, 815  
425: \reference{} Feuchtinger, M. U. 2000, \aap, in press (F20)  
426: \reference{} Gehmeyr, M. 1992, \apj, 399, 272
427: \reference{} Gould, A. \& Popowski, P. 1998, \apj, 508, 844 (GP)
428: \reference{} Gratton, R. G. 1998, \mnras, 296, 739 (G98)    
429: \reference{} Groenewegen, M. A. T., \& Salaris, M. 1999, \aap, 348, L33 (GS)  
430: \reference{} Heiser, A. M. 1996, \aj, 112, 2142 (H96)  
431: \reference{} Jones, R. V., Carney, B. W., Storm, J., \& Latham, D. W.
432: 1992,  \apj, 386, 646 (J92)  
433: \reference{} Kovacs, G., \& Kanbur, S. 1997, \mnras, 295, 834  
434: \reference{} Lambert, D. L., Heath, J. E., Lemke, M., Drake, J. 1996 \apjs, 103, 183
435: \reference{} Laney C. D. 2000, in IAU Colloq. 176, The Impact of Large-Scale 
436: Surveys on Pulsating Star Research, ed. L. Szabados, \& D. Kurtz 
437: (San Francisco: ASP), in press  
438: \reference{} Layden, A. 1999 in Post-Hipparcos Cosmic Candles, ed. A. Heck,
439: \& F. Caputo (Dordrecht: Kluwer), 37 (L99)  
440: \reference{} Layden, A., Hanson, R. B., Hawley, S. L., Klemola, A. R. \&
441: Hanley, C. J. 1996, \aj, 112, 2110 (L96)  
442: \reference{} Longmore, A. J., Dixon, R., Skillen, I., Jameson, R. F., \&
443: Fernley, J. 1990, \mnras, 247, 684 (L90)  
444: \reference{} Stellingwerf, R. F. 1982, \apj, 262, 330 (S82)  
445: \reference{} Tsujimoto, T., Miyamoto, M., \& Yoshii, Y. 1998, \apj,
446: 492, L79  (TMY)  
447: \reference{} Wood, P. R., Arnold, A., \& Sebo, K. M. 1997, ApJ, 485, L25 (WAS) 
448: \reference{} Wuchterl, G., \& Feuchtinger, M. U. 1998, 340, 419   
449: \reference{} Yecko, P. A., Kollath, Z., Buchler, J.R. 1998, A\&A, 336, 553  
450: \end{references}
451: 
452: 
453: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
454: %                           Figure Captions
455: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
456: \pagebreak
457: % 1
458: \figcaption{Top panels: blue light curves of iso-period -$P=0.29$ d- RR Lyrae 
459: models constructed by adopting the same chemical composition (Y=0.24, 
460: Z=0.001) and stellar mass (\msun=0.60), but different assumptions on 
461: effective temperatures and luminosities (see labeled values). 
462: Bolometric curves where transformed into {\rm B} magnitudes by 
463: adopting bolometric corrections and color-temperature (CT) relations 
464: by Castelli, Gratton \& Kurucz (1997).  
465: Bottom panels: similar to top panels, but for models constructed at 
466: fixed effective temperature ($T_e=7100$ K), and different assumptions 
467: on stellar masses and luminosities (see labeled values).}
468: 
469: % 2
470: \figcaption{Comparison between theory (solid lines) and observations 
471: (open circles). From left to right the panels refer to photometric 
472: data in U, B, V (Heiser 1996), and K (Fernley, Skillen \& Burki 1993).  
473: Empirical data were plotted by assuming $E(B-V)=0$. Photometric errors 
474: in optical bands are equal to the symbol size.}  
475: 
476: % 3 
477: \figcaption{\uc absolute magnitudes vs. time. Top and bottom panels 
478: show {\rm V} and {\rm K} estimates respectively. The acronyms refer 
479: to different methods: 
480: BW: Baade-Wesselink (Jones et al. 1992); HB models: Horizontal Branch
481: models (Cassisi et al. 1999, C99); SP: statistical parallax
482: (Barnes \& Hawley 1986, BH); TP: trigonometric parallax (Gratton 1998, G98);
483: BW + PLK: zero point from BW and slope from {\rm K} band Period-Luminosity
484: relation (Longmore et al. 1990, L90; J92; Carney et al. 1995, C95);
485: LK + BW: zero point from modified Lutz-Kelker correction (Groenewegen \&
486: Salaris 1999, GS) and slope from BW (Fernley et al. 1998, FC98);
487: RP + BW: zero point from reduced parallax (GS) and slope from BW (FC98);
488: SP + BW: zero point from SP (Layden et al. 1996, L96; Fernley et al. 1998,
489: FB98; Gould \& Popowski 1998, GP; Tsujimoto, Miyamoto, \& Yoshii 1998, TMY;
490: Layden 1999, L99) and slope from BW (FC98).} 
491: 
492: % 4 
493: \figcaption{ {\rm B} light curves of RR Lyrae models vs. phase. 
494: Solid lines refer to models constructed by adopting the calibration 
495: of the TC model suggested by BS, while dashed lines to models 
496: constructed by assuming that the convective flux is vanishing in the 
497: regions in which the superadiabatic gradient is negative 
498: (see equ. 7 and \S 3 in BS).}  
499: \end{document}
500: