The transgression form associated to p and the pair (ω0,ω1) is the (2n−1)-form Tp(ω0,ω1) on M uniquely characterized by the requirement that its pullback to P is
π∗Tp(ω0,ω1)=n∫01p(η∧Ωtn−1)dt,
where Ωtn−1 denotes the wedge product of (n−1) copies of Ωt and p(η∧Ωtn−1) means the R-valued form obtained by feeding the g-valued factors into p.
Because the integrand is basic (horizontal and G-invariant), Tp(ω0,ω1) is well-defined on the base. It is designed so that the difference of the corresponding Chern–Weil forms is exact:
Degree 1 (linear invariant polynomial). For n=1 and an Ad-invariant linear functional p:g→R, the formula reduces to
Tp(ω0,ω1)=p(ω1−ω0),dTp=p(Ω1)−p(Ω0).
Degree 2 on a trivial bundle (the usual Chern–Simons 3-form). On a trivial bundle and in a global gauge, a connection is represented by a g-valued 1-form A (see local connection 1-form
). For p(X,Y)=tr(XY) (degree 2), taking ω0=0 gives the standard 3-form
CS(A)=tr(A∧dA+32A∧A∧A),
with dCS(A)=tr(F∧F), where F=dA+A∧A is the local curvature
.
Abelian case. If G is abelian (e.g. U(1)), then Ad is trivial and A∧A=0. For a degree 1 invariant polynomial, the transgression between two U(1)-connections A0,A1 is simply T(A0,A1)=A1−A0, and dT=dA1−dA0.