M = {
	 { -(n lambda), mu }
	,{ n lambda, -(mu + n-1 lambda) }
    }
VarM = {
	 { S + (n lambda), mu }
	,{ n lambda, S + (mu + n-1 lambda) }
    }
OneMatrix = {1,1}
VarOneMatrix = {{1},{1}}
Prob = {{1},{0}}
VarProb = {1,0}
SecondMoment = (Limit[(-2 VarProb . (D[(Inverse[VarM]),S])),S->0]) . VarOneMatrix
MTTDL = Simplify[- OneMatrix . Inverse[M] . Prob]
Variance = Simplify[SecondMoment - MTTDL^2]