Dear Dr. Belsole,

first of all, we like to thank the referee for the positive assessment and the
useful remarks.

We like to address the issues raised in the report:
1. Different decay times in the coincidence detections shown in both parts of
figure 3:
Two parts of the figure correspond to different decay paths X and Y as
indicated in figure 1. The strength of the collective effect is directly
proportional to the dipole moment associated with the transition under
consideration. Another way to look at this is to consider the difference in
optical density (OD) for the different transitions. The OD of 32 we quote is
always referred to the lower part of the X branch, or the 5P3/2, F = 3 to
5S1/2, F = 2 transition. We do state this right at the beginning of section 5
("The transition 5P3/2, F = 3 to 5S1/2, F = 2 is 2.8 times stronger than from
5P3/2, F = 2"). 
The measured coherence times for decay paths X and Y are respectively
τX=5.6±0.1ns and  τY = 13.1±0.2ns, are mentioned right after equation 4, and
yield a ratio of 2.3, comparable to the ratio of the interaction strength,
but not compatible within the uncertainties out of the fit parameters. We can
not fully explain this difference, but a likely cause could be a change
in the optical density between the two measurements. To point out the at least
qualitative agreement between the two ratios, we added a sentence near the end
of section 5:
"The extracted ratio $\tau_Y/\tau_X=2.34\pm0.05$ deviates by about 20\% from
the ratio of the transition strengths, possibly due to a change in the atomic
density between the two measurements."
We also added a sentence to the caption of figure 3 to highlight this ("The
coincidence events show an exponential rise due to the response of the etalon,
and an exponential decay with different time constants for the two paths due to different transition strengths.")

2. FFT analysis of the quantum beat data
We assume the referee was commenting on data shown in figure 4, not figure
2. Although presenting a FFT seems tempting to support the claim that the
center frequency is at the 266MHz, the evidence is relatively weak. Reason for
this is the small number of beats, leading to a relatively large uncertainty
in the center frequency estimation. We attach an FFT obtained from the data,
but feel that it adds very little to better understand the beats, mostly
because the uncertainty in extracting a center frequency is limited by the
relatively fast exponential decay time of the beat to perhaps 50MHz. We
therefore would opt to not show it, as it would not add more than seen already
by the presence of the oscillation.

3. Spelling issues on names in references.
We thank the referee for pointing out the spelling inconsistencies in the
bibliography. We fixed references 7,20,24,33 and checked that the others are
consistent with the respective spellings in the journal sources. 

4. Lastly, we compactified the author names/affiliations in the manuscript
according to the IOP style guide. 

With this, we hope to have addressed the issues raised by the referee.

Thanks, and Best Regards on behalf of all authors,

Christian Kurtsiefer