Consider a commutative diagram of commutative groups:
Aa↓A′uu′Bb↓B′
Then there exists one and only one homomorphism u1:Ker(a)→Ker(b) and one and only one homomorphism u2:Coker(a)→Coker(b) such that the diagrams
Ker(a)i↓Au1Ker(b)j↓B
and
A′p↓Coker(a)u′u2B′q↓Coker(b)
are commutative, i and j denoting the canonical injections and p and q the canonical surjections.
If x∈Ker(a), then a(x)=0 and b(u(x))=u′(a(x))=0, hence u(x)∈Ker(b), and the existence and uniqueness of u1 are then immediate. Similarly, we have u′(a(A))=b(u(A))⊂b(B), then by taking quotients u′ gives a homomorphism u2:Coker(a)+Coker(b), which is the only homomorphism for which (15) is commutative.
Let us now start with the commutative diagram (10) of commutative groups; there corresponds to it by Lemma 1 a diagram
where i,j,k are the canonical injections, p,q,r the canonical surjections and u1,u2 (resp. v1,v2) the homomorphisms canonically associated with u,u′ (resp. v,u′) by Lemma 1. It is immediately verified that this diagram is commutative.
Bourbaki puts the proof directly after the statement, so what is above is the statement and its proof together.