Almost necessary and sufficient conditions for survival of species (Q2386001)

The paper investigates a three-species system (LV-3) \[ u_k^{\prime }=u_k[a_k(t)-\sum_{l=1}^3b_{kl}u_l] \] by making use of an auxiliary two-species system. Here, \(u_k(t)\) is the density of the \(k\)th species at time \(t\), \(a_k(t)\) represents the growth rate of population \(k\) at time \(t\), and \(-b_{kl}\) stands for the measure of the inhibiting effect of population \(l\) on population \(k\). The key result for the two-species system is given in Theorem 2.1 (a persistence result that collects and improves some previous results in the literature). The main results of the paper are Theorems 2.2 and 3.1. They read as follows. The first one: The system (LV-3) is persistent if for any permutation \(i,j,k\) of \(1,2,3\), \(M[a_i]=\lim \frac 1T\int_{t_0}^{t_0+T}a_i(t)dt\) (a quantity assumed independent of \(t_0\)) \(>b_{ij}M[a_j]/b_{jj}\), \(M[a_j]>b_{ji}M[a_i]/b_{ii}\) and \(M[a_k]>b_{ki}x_i^{*}+b_{kj}x_j^{*}\), where \(x_i=x_i^{*}\), \(x_j=x_j^{*}\), is the unique solution of \(M[a_i]=b_{ii}x_i+b_{ij}x_j,M[a_j]=b_{ji}x_i+b_{jj}x_j\). The second one regards the autonomous case: Let \(a_k(t)=a_k\). Assume that \(a_1>b_{12}\frac{a_2}{b_{22}}\), \(a_2>b_{21}\frac{a_1}{b_{11}}\) and \(a_3<b_{31}x_1^{*}+b_{32}x_2^{*}\) hold, where \((x_1^{*},x_2^{*})\) is the intersection of the lines \(a_1=b_{11}x_1+b_{12}x_2,a_2=b_{21}x_1+b_{22}x_2\). If \((u_1(t),u_2(t),u_3(t))\) is any solution of (LV-3) such that \((u_1(t_0),u_2(t_0))\) is close to \((x_1^{*},x_2^{*})\) and \(u_3(t_0)\) is sufficiently small and positive, then \(u_1(t)\rightarrow x_1^{*},u_2(t)\rightarrow x_2^{*}\), and \(u_3(t)\rightarrow 0\) as \(t\rightarrow +\infty \). The proof of the latter result relies on the standard linearized stability theory. The paper is almost self-content, highly interesting and beautifully written.