Solution:
Let b2=ca. The conditions of the problem give b2=ca∣a4 and a+1∣ca+1, and this is equivalent to
c∣a3anda+1∣c−1.
Let c=d(a+1)+1, d∈N0. Since a3≡−1(moda+1), we have ca3≡−1(moda+1), i.e. ca3=e(a+1)−1 for some e∈N. It follows that a3=(d(a+1)+1)(e(a+1)−1), which after multiplying out and cancelling a+1 becomes a2−a+1=de(a+1)+(e−d). From here we have e−d≡a2−a+1≡3(moda+1), hence
e−d=k(a+1)+3andde=a−2−k(k∈Z).
We distinguish the following cases:
(1) k∈/{−1,0}. In this case (*) implies de<∣e−d∣−1, which is possible only for d=0. Now c=1 and b2=a, hence (a,b)=(t2,t).
(2) k=−1. From (*) we obtain a=d+1. Now c=a2 and b2=a3, hence (a,b)=(t2,t3).
(3) k=0. From (*) we obtain a=d2+3d+2. Now c=d(a+1)+1=(d+1)3 and b2=ca=(d+1)4(d+2). It follows that d+2=t2 for some t∈N, which gives (a,b)=(t2(t2−1),t(t2−1)2), t≥2.
Answer: These are the pairs (a,b) of the form (t2,t), (t2,t3) and (t2(t2−1),t(t2−1)2), where t∈N.