30.4. STOKE’S THEOREM AND THE ORIENTATION OF ∂Ω 1069

30.4 Stoke’s Theorem And The Orientation Of ∂Ω

Here Ω will be an n dimensional orientable Lipschitz manifold with boundary in Rm. Letan oriented manifold for it be {Ui,Ri}p

i=1 and let a C∞ partition of unity be {ψ i}pi=1. Also

letω = ∑

IaI (x)dxi1 ∧·· ·∧dxin−1

be a differential form such that aI is C1(Ω). Since ∑ψ i (x) = 1 on Ω,

dω = ∑I

m

∑k=1

p

∑j=1

(ψ jaI

)∂xk

(x)dxk ∧dxi1 ∧·· ·∧dxin−1

It follows

∫dω = ∑

I

m

∑k=1

p

∑j=1

∫R j(U j)

(ψ jaI

)∂xk

(R−1

j (u))

∂(xk,xi1 · · ·xin−1

)∂ (u1, · · · ,un)

du

= ∑I

m

∑k=1

p

∑j=1

∫R j(U j)

(ψ jaI

)∂xk

(R−1

jε (u))

∂(xkε ,xi1ε · · ·xin−1ε

)∂ (u1, · · · ,un)

du+

∑I

m

∑k=1

p

∑j=1

∫R j(U j)

(ψ jaI

)∂xk

(R−1

j (u))

∂(xk,xi1 · · ·xin−1

)∂ (u1, · · · ,un)

du

−∑I

m

∑k=1

p

∑j=1

∫R j(U j)

(ψ jaI

)∂xk

(R−1

jε (u))

∂(xkε ,xi1ε · · ·xin−1ε

)∂ (u1, · · · ,un)

du (30.4.12)

where those last two expressions sum to e(ε) which converges to 0 as ε → 0 for a suitablesubsequence. Here is why.

(ψ jaI

)∂xk

(R−1

jε (u))→

(ψ jaI

)∂xk

(R−1

j (u))

because of the uniform convergence of R−1jε to R−1

j . In addition to this,

∂(xkε ,xi1ε · · ·xin−1ε

)∂ (u1, · · · ,un)

→∂(xk,xi1 · · ·xin−1

)∂ (u1, · · · ,un)

in Lr (R j (U j)) for any r > 0 and so a suitable subsequence converges pointwise. The inte-grands are also uniformly integrable. Thus the Vitali convergence theorem can be applied toeach of the integrals in the above sum and obtain that for a suitable subsequence, e(ε)→ 0.

Then 30.4.12 equals

= ∑I

m

∑k=1

p

∑j=1

∫R j(U j)

(ψ jaI

)∂xk

(R−1

jε (u)) m

∑l=1

∂xkε

∂ulA1ldu+ e(ε)

30.4. STOKE’S THEOREM AND THE ORIENTATION OF 0Q 106930.4 Stoke’s Theorem And The Orientation Of dQHere Q will be an n dimensional orientable Lipschitz manifold with boundary in R”. Letan oriented manifold for it be {U;,R;}/_, and let a C® partition of unity be {y;}?_,. AlsoletoO= Ya (x) dx;j, A-+: Adx;,_,1be a differential form such that a; is C! (Q). Since Y w; (x) = 1 on Q,3po (var)y on (x) dx, A dxj, Aves Adx;,_,1j=1 kdo=)T kIt followsm ?p O( Waa Xe Xi, Xj[ao p> y how) a) (R;! (u)) 7) ayI k=1j=1 “+ Un)Ox, JE O (uy,°+* ,Un)how a (war) (Rie (u)) O (Xe Xie **Xin_j€) ius. ,Un)hh y Jw) “ra (i'w) ao ah ay 04.19Ox,where those last two expressions sum to e(€) which converges to 0 as € — 0 for a suitablesubsequence. Here is why.2(va)(Rie (u)) + (vie) (R;! (u))because of the uniform convergence of Ri. to R;'. In addition to this,O (Xe sXipe***Xin_j€) 0 (X,Xi “Xin )O(uiy ln) (thy tn)in L’ (Rj (U;)) for any r > 0 and so a suitable subsequence converges pointwise. The inte-grands are also uniformly integrable. Thus the Vitali convergence theorem can be applied toeach of the integrals in the above sum and obtain that for a suitable subsequence, e(€) > 0.Then 30.4.12 equals“yyyT k=1j=1hh (var) (Riz (u)) YM A d+ e(€)j(Uj) OX re i Oui