Let a and b be in A and let x=(xn)n∈N be an element of W(A).
a) We have the formulas
τ(ab)=τ(a)×τ(b)
τ(a)×x=(apnxn)n∈N.
b) The series with general term Vn(τ(xn)) is convergent in W(A), with sum x. Let n be a positive integer. The polynomial Pn(X0,...,Xn;Y0,...,Yn) introduced in No. 3 is isobaric of weight pn in the family (X0,...,Xn) when Xi is assigned the weight pi. Hence
Pn(X0,0,...,0;Y0,...,Yn)=X0pnPn(1,0,...,0;Y0,...,Yn). Since
1=(1,0,0,...) is the unit element of the ring of Witt vectors with coefficients in
Z[(Xn)n∈N,(Yn)n∈N], we have
Pn(1,0,...,0;Y0,...,Yn)=Yn. By substituting
a for
X0 and
xi for
Yi, we deduce from (40) and (41) the relation:
Pn(a,0,...,0;x0,...,xn)=apnxn. According to the definition of the multiplication in
W(A), we have proved (39); formula (38) is a particular case of (39).
Let us prove b). By definition, Vn(τ(xn)) is the sequence all of whose components are zero, except the one with index n, which is equal to xn. It follows from Lemma 4, by induction on m, that we have
n=0∑mVn(τ(xn))=(x0,...,xm,0,0,...)
for every integer m≥0; we deduce b) by passing to the limit since the topology T on W(A) is the product of the discrete topologies of the factors A.
Bourbaki puts the proof directly after the statement, so what is above is the statement and its proof together.
Nothing.
Nothing.