1: \begin{subequations}

2: \label{ERE}

3: \begin{align}

4: \frac{\mathrm{d}\bar{n}}{\mathrm{d}t}& =-A\left( \bar{n}+1\right) -2BW_{21}%

5: \bar{n}, \\

6: B& =\frac{\pi \mu ^{2}}{3\hbar ^{2}\varepsilon _{0}},  \label{B}

7: \end{align}%

8: which coincides with ERE (\ref{Einstein}) and gives the correct result for $B

9: $ (see Ref.\thinspace 5).

10: 

11: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

12: % SECTION: Light incoherence vs atomic incoherence. From ERE to laser rate...

13: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

14: 

15: \section{Light incoherence vs atomic incoherence. From ERE to laser rate

16: equations}

17: 

18: In our previous derivation of ERE we have assumed that the light spectrum

19: was sufficiently broad as to bring $\bar{n}$ outside the integral in (\ref%

20: {dn1})---the Markov approximation---and to make the replacement $e^{-\gamma

21: _{\bot }\left( t-t^{\prime }\right) }\rightarrow 1$. Nevertheless one can

22: still adopt the Markov approximation without imposing $e^{-\gamma _{\bot

23: }\left( t-t^{\prime }\right) }\rightarrow 1$, and this is what we face in

24: this Section. We are hence considering the possibility that, either because

25: the light spectrum is broad or the atomic line is, or both, (i.e., $\max

26: \left( \gamma _{\bot },\Delta \right) \gg A$), the function $I\left(

27: t-t^{\prime }\right) e^{-\gamma _{\bot }\left( t-t^{\prime }\right) }$ under

28: the integral in (\ref{dn1}) is strongly peaked around $t^{\prime }=t$,

29: allowing again the substitution $\bar{n}\left( t^{\prime }\right)

30: \rightarrow \bar{n}\left( t\right) $ under that integral. Hence, under the

31: Markov approximation Eqs.\thinspace (\ref{npunt_I}) become, after performing

32: the time integration, 

33: \end{subequations}

34: