1: Dr. Chris Sneden, Letters Editor
2: THE ASTROPHYSICAL JOURNAL LETTERS
3: apjletters@letters.as.utexas.edu
4:
5: Dear Chris,
6:
7: Many thanks for your efforts in finding a referee and for the report on
8: our paper "A Spectroscopic Orbit for Regulus" by Gies et al. (ApJL #22909).
9: We appreciate the referee's speedy and careful review, and below we
10: outline the changes we made in response to the report.
11:
12: > This manuscript presents the discovery that Regulus is a spectroscopic
13: > binary with a period of about 40 days. Normally a simple spectroscopic
14: > orbital solution would not warrant publication as an ApJ Letter, but
15: > in this case the star is of special interest in that its very rapid
16: > rotation is difficult to explain given its present evolutionary
17: > status. The authors' results and their plausible conclusion that the
18: > companion is a white dwarf provide a natural explanation through a
19: > previous episode of mass and angular momentum transfer. Indirect
20: > evidence for such a white dwarf has been reported previously in the
21: > form of ultraviolet excess. In my opinion this justifies publication
22: > in ApJL. The paper is fairly straightforward, concise, and very well
23: > written, and I have only a few corrections, comments, and
24: > suggestions the authors may wish to consider:
25:
26: > * The correct HIP number is 49669.
27:
28: Corrected.
29:
30: > * My own orbital solution based on the authors' data gives slightly
31: > different results for the orbital elements and derived quantities.
32: > Although this difference is not terribly important for this paper, it
33: > was a bit unexpected for such a straightforward problem. Perhaps the
34: > authors could double-check their results. Fixing the orbital period to
35: > 40.11 days and assuming a circular orbit, as the authors have done, I
36: > get the following using my codes:
37: >
38: > Gamma = +4.90 +/- 0.20 km/s
39: > K = 7.03 +/- 0.29 km/s
40: > T = 2,444,527.02 +/- 0.27 (HJD)
41: > asini = 5.58 +/- 0.23 R_Sun
42: > f(M) = 0.00144 +/- 0.00018 M_Sun
43: >
44: > in which I have weighted the observations as the authors claim to have
45: > done. Perhaps the disagreement has to do with differences in the
46: > methodology (I used Levenberg-Marquardt optimization)...
47:
48: We are grateful that the referee made an independent assessment of
49: the orbital solution. The difference in the results is a consequence
50: of a slightly different weighting scheme that was not fully given
51: in the first version of the paper. We revised Section 3, paragraph 1,
52: line 10 as follows:
53: "Each measurement was assigned a weight proportional to the
54: inverse square of the larger of the measurement error
55: or 1~km~s$^{-1}$ (to account for possible systematic
56: errors between results from different groups of observations).
57: Trials with other weighting schemes gave similar results."
58:
59: > * The authors have fixed the period in their solution to the value
60: > determined with the DFT and CLEAN methods. Given that the observations
61: > cover more than 250 cycles, I tried leaving P as a free parameter, and
62: > got a seemingly more precise value of 40.1160 +/- 0.0096 days (a
63: > factor of two better than reported in the manuscript). The change in
64: > the other elements is insignificant, except that the nodal passage
65: > becomes formally more poorly determined (2,444,525.9 +/- 1.9 days),
66: > although this is only because of the very strong correlation between P
67: > and T (with a correlation coefficient of -0.9899). Holding the period
68: > fixed, as the authors have done, artificially suppresses this
69: > correlation so that the nodal passage looks more precise but doesn't
70: > account for the error in P.
71:
72: Again, the different (lower) error estimate for the period is due
73: to differences in the weighting scheme. The referee used the inverse
74: square of the quoted errors while we adopted the nominal equal weights
75: used in the DFT/CLEAN method. We prefer to adopt the more conservative
76: errors associated with the standard time series analysis.
77:
78: We agree with the referee that the formal error for epoch does not
79: account for the error in period (set to a constant in the orbital fit).
80: We added to Section 3, paragraph 1, line 13 the following:
81: "Note that the error in $T_0$ increases to $\pm 3.9$~d
82: when the full range in acceptable period is considered."
83:
84: > * In Sect. 4 the authors comment on the lack of evidence for the
85: > companion in observations that should have resolved it. It would be
86: > useful to include a few references here for those negative attempts.
87:
88: We added citations to the interferometric and lunar occultation
89: work on Regulus that failed to find evidence of a companion.
90:
91: > * In Table 3, it would be useful to include the estimate of the
92: > coefficient in the expression for M*sini, along with its error.
93:
94: We added a new row to Table 3 with this functional parameter.
95:
96:
97: We hope you will now find the paper ready for publication in the
98: Astrophysical Journal Letters.
99:
100: Sincerely yours,
101:
102: Douglas Gies
103: