changes following review comments

article/content.tex

@@ -99,13 +99,13 @@ \section{Methods}
 global dynamical behaviour of the system.
 We used \textit{Julia v1.3.1} \citep{bezanson2017}, along with packages \verb| DifferentialEquations.jl|
 \citep{rackauckas2017} to compute the numerical integrations and \verb|Plots.jl| to represent
-our results. 
-Not knowing the exact algorithm used in the original article, we let the \verb|solve()| function 
+our results.
+Not knowing the exact algorithm used in the original article, we let the \verb|solve()| function
 select the appropriate algorithm to solve our differential equations.
 In our implementation, it selected a composite algorithm combining, amongst others,
 algorithms Tsit5 and Rosenbrock23.
-We did note, however, that setting a specific algorithm such as RK4 sometimes resulted in a similar 
-behaviour for the system. 
+We did note, however, that setting a specific algorithm such as RK4 sometimes resulted in a similar
+behaviour for the system.
 
 To fully replicate key findings of the original paper, we focused on replicating
 original figures.
@@ -145,8 +145,8 @@ \section{Methods}
 dynamics. Moreover, \citeauthor{hastings1991} mention removing points resulting from
 secondary local maxima, but do not provide details on how to identify these points.
 Hence, we adopted the following method:
-1) we selected the 1000 last solutions for our system between time steps 1 and 10 000, in
-order to eliminate transient behaviour; %Reviewer comment: This is unclear to me. Do you mean time steps 9000-10000? Please clarify.
+1) we selected the solutions for the last 1000 evaluated time steps of our system, in
+order to eliminate transient behaviour;
 2) we selected the values that were greater than both their preceding and following
 values, which identified local maxima only;
 and 3) we only kept values that were greater than a given threshold of the cycle's maximal
@@ -235,8 +235,8 @@ \section{Results}
 confirm that chaos occurs for values other than $b_2$ = 2.0. Chaos is apparent for both
 smaller or greater values.
 However, while \citeauthor{hastings1991} reported that chaos was more likely for greater values of
-$b_2$, our results highlight that $z$ instead converges to a single value and starts to
-crash past $b_2$ = 2.35. %Reviewer comment: Please clarify what you mean by "crash" here. e.g. "starts to rapidly approach zero"
+$b_2$, our results highlight that $z$ instead converges to a single value and
+starts to rapidly approach zero past $b_2$ = 2.35.
 
 Lastly, although \citeauthor{hastings1991} did not specify the equation of the plane that
 crosses the trajectories of the phase plot at its "handle", we were able to accurately
@@ -293,8 +293,7 @@ \section{Discussion}
 
 While we also found chaos for values of $b_2$ other than the default one of 2.0, both
 smaller or greater, we do not totally agree with \citeauthor{hastings1991} that "chaos is more
-likely for larger values of $b_2$". As \autoref{fig:figS1} shows %Reviewer comment: missing verb?
-, chaos can be quite likely for both
+likely for larger values of $b_2$". As \autoref{fig:figS1} shows, chaos can be quite likely for both
 smaller or larger values.
 We find important to note, however, that at a certain value of $b_2$, $z$ converges and
 starts to crash, thus exhibiting non chaotic behaviour within a given range of $b_1$
@@ -320,7 +319,7 @@ \section{Discussion}
 In general, the model, including its equations and parameters, was well described by the
 authors. The most significant obstacles to reproducibility in \citeauthor{hastings1991}'s paper
 were the absence of the values of the initial conditions, which have a huge impact on a
-chaotic system, and the insufficient description of certain methods. %Reviewer comment: Be explicit here about what was insufficiently described. E.g. the numerical integration methods.
+chaotic system, and the insufficient description of the numerical integration methods.
 Consequently, there are slight differences between our results and theirs.
 Furthermore, since we tried to keep our implementation as close as possible to the
 original one, some steps did rely on arbitrary thresholds (for instance for the primary