“…Assuming r individuals have recovered after infectious periods w 1 < · · · < w r < T , we have ℓRfalse(w1,…,wrfalsefalse|θ,kfalse)=false(k−rfalse)logHγfalse(Tfalse)+∑i=1rloghγfalse(wifalse),where Hγfalse(⋅false) and hγfalse(⋅false) are, respectively, the survival function and the probability density function of the exponential distribution with rate γ . Averaging the infectious periods used in the previous analysis [38,39], we assume here that the recovery times have an exponential distribution with mean γ −1 = 5.5 days (see also [40,41]), so γ was not estimated. The complete log-likelihood conditional on the population size n , the parameters and observables is then ℓ0(t1…,tk,w1,…,wr|θ,n)=ℓI(t1,…,tk|θ,n)+ℓ…”