I thank the two reviewers for their comments. I have taken them all into
account. The corresponding changes to the manuscript are marked in blue.
Below my detailed response.

> Reviewer: 1
> The basic Hosking integral is given in (2.1), but the formula is not
> clear. In particular, the right side contains a volume integral d^3r,
> then also two different position vectors x and r. So first, which of
> these is the one that goes along with the volume element d^3r? Second,
> once one of these has been integrated out, what happens to the other
> one? Why is the left side then, once the integral has been evaluated,
> not a function of either x or r? At the moment anyway it is not clear
> what this integral actually means. Presumably this is all properly
> defined in the previously existing literature on this Hosking integral,
> but readers should not have to go there to look up and understand the
> basic definition of the quantity itself...

To clarify this better, I have now emphasized in two places (pages 1
and 2) that angular brackets denote averaging over x. Thus, it should
now be clear that <h(x)h(x+r)> is a function of r and is equal to <h^2>
for r=0. I have now added some extra text after Eq.(2.1) on page 2 to
explain that the r integral is over volumes V_R=4pi R^3/3, and that
script I_H increases like R^3 for small R. The actual Hosking integral
is computed for values of R, for which script I_H levels off.

> Just a few lines later it says "the definitions of I_H and I_H are",
> where one of the I's is \mathcal script, and the other one is just
> regular italic script. Are there actually two different I_H quantities??

Yes, at this point it may almost look like a misprint. To emphasize this
difference better, I have now displayed the relevant part of the first
sentence of Section 2 in italics (but not in blue).

> Another few lines down is the magnetic correlation length xi(t),
> which seems to play a very important role in all of the analysis,
> but readers are never given any formal definition of what it actually
> is. Many experts would likely know at least roughly what is meant here,
> but again, readers should not have to refer back to previous literature
> to look up precise formulas for quantities that turn out to be important
> here, so a proper definition of xi should be given here.

Yes, I agree, so I have now added this in the second sentence of the
paragraph just before Eq.(2.5).

> The calculations are described as being from the previous paper called
> B20 here, but no further details are given. Please include at least a
> paragraph or so reminding readers of what the calculations actually were,
> that is, what initial conditions, etc.

Yes, good point! I have now added some 5 lines after Eq.(2.5) to explain
the basics about this run.

> I do not understand the inset in figure 4. By definition, shouldn't
> all of these quantities start out at 1 for very small t? They do in the
> main figure anyway, so what is different about the inset? And how does
> the inset actually relate to the main figure? Shouldn't one of the lines
> in the inset correspond to one of the lines in the main figure? At the
> moment anyway I'm not seeing any connection at all.

Oops, my mistake! The y title was correct (script I_H as opposed to roman
I_H in the main figure), but the x title was wrong. Is it now corrected
and should have read R. I have also indicated the sense of time with an
arrow (see Fig.4 on page 6). The related text in the last 2 sentences
of Sect.3 (also page 6) did already explain this correctly.

> Reviewer: 2
> 1. How does the author view the role of magnetic reconnection in
> the Hall decay? Magnetic helicity is conserved by electron MHD because
> magnetic flux is frozen into the electron flow; are helicity-conserving
> but topology-violating events, i.e., reconnections, required for energy
> decay, in the author's opinion?

I have now picked up this point in the last paragraph of the conclusions
(page 7), along with some relevant references. At the moment, based on
the presented pq diagrams, there is no strong case in favor of invoking
reconnection.

> 2. An equivalent, but perhaps conceptually simpler, route to get at
> the t^{-10/13} decay law would be to use (1.1) to argue that the timescale
> tau for magnetic decay is proportional to L^2/B, with L and B the integral
> length and magnetic-field scales respectively. Since tau is proportional
> to t for self-similar decay, we get B^2 proportional to t^{-10/13} if I_H
> ~ B^4 L^5 ~ const. The difference between MHD and electron MHD is that,
> in the former, tau is proportional to L/B, not L^2/B.

I have now added this discussion with reference to the anonymous referee
in the footnote of page 4.