18.3. STOKES THEOREM 479

Definition 18.3.1 Denote as Λ(k) the set of finite sums of differential forms of or-der k. I will now describe the boundary ∂ : Λ(k)→ Λ(k−1) by first defining it on r andthen, if desired, one would know it on all of Λ(k). If k = 1, then ∂r ≡ r (b)− r (a) . Ingeneral, for r : [a,b]→ Rp and l ≤ k,

∂lr (u1, · · · , ûl , · · · ,uk)

≡ r (u1, · · · ,bl ,ul+1, · · · ,uk)−r (u1, · · · ,al ,ul+1, · · · ,uk)≡ rbl −ral .

Note that rbl ,ral are defined on

[a1,b1]×·· ·× [al−1,bl−1]× [al+1,bl+1]×·· ·× [ak,bk]≡ [a,b]l .

Specifically, if ω = ∑I aI (x)dxI where I denotes ordered lists of indices of length k− 1taken from {1, ..., p}∫

∂lrω ≡

∫[a,b]l

∑I

aI(rbl (u)

)det

(drI

bl(u)

du

)dmk−1

−∫[a,b]l

∑I

aI(rbl (u)

)det

(drI

al(u)

du

)dmk−1

where u comes from [a,b]l . Here rbl is r with bl in the lth position, similar for ral . Thus

det(

drIbl(u)

du

),det

(drI

al(u)

du

)are (−1)1+l AJ

1l where AJ1l is the (1, l)th cofactor of det

(DrJ

)where J has length k and the matrix DrJ is the matrix of 18.4, having the top row(

x j,u1 x j,u2 · · · x j,uk

)Then

∫∂r ω ≡ ∑l

∫∂lr

ω .

With this preparation, Stoke’s theorem follows from a computation.

dω ≡∑I∈J

p

∑j=1

∂α I

∂x j(x)dx j ∧dxi1 ∧·· ·∧dxik−1 , I = (i1, · · · , ik−1)

Definition 18.3.2 As discussed earlier, define∫r

dω ≡∑I∈J

p

∑j=1

∫[a,b]

∂α I

∂x j(r (u))

∂(x j,xi1 · · ·xik−1

)∂ (u1, · · · ,uk)

du (18.3)

By definition,∂

(x j ,xi1 ···xik−1

)∂ (u1,··· ,uk)

is the determinant ofx j,u1 x j,u2 · · · x j,ukxi1,u1 xi1,u2 · · · xi1,uk

......

...xik−1,u1 xik−1,u2 · · · xik−1,uk

 (18.4)

Note how this matrix is just the matrix of Df where

f (u) =(x j (u) ,xi1 (u) , ...,xik−1 (u)

)T.

18.3. STOKES THEOREM 479Definition 18.3.1 Denote as A (k) the set of finite sums of differential forms of or-der k. I will now describe the boundary 0 : A(k) + A(k—1) by first defining it on r andthen, if desired, one would know it on all of A(k). If k = 1, then dr =r(b)—r(a). Ingeneral, for r : |a,b] > R? and! <k,Or (uy,°** Uj, Uk)= T (ui,°°° br, Uj41,°°° Ug) — 7 (uy,°°° 34], Uj+1,°°° Uk) =Pp, —Ta;-Note that rp,,%q, are defined on[a1 bi] x +++ x [ay-1, bia] X any, diya] x +++ X (ax, bx] = [a,b],Specifically, if @ = ¥,a,(x) da! where I denotes ordered lists of indices of length k —1taken from {1,..., p}ark, (w)_ 1yn o= la, har (rp, (u)) det ( Ta dm_|dr! (u)Lar Tp, (u ya ( Ta Jam.where wu comes from ww, Here r,, is r with b; in the I" position, similar for rq, Thusdri, dr!det ( Py ‘) ,det ( i) are (-1)'ad, where At, is the (1, 1)"" cofactor of det (Dr’)where J has length k and the matrix Dr’ is the matrix of 18.4, having the top row( Niu Xjug 9° X jug )Then Jor O= Yi Jayr @With this preparation, Stoke’s theorem follows from a computation.0a . .Jo=S ye )dxj Ndxj, \-++\dxi_,, 1= (i1y++* sik-1)Tes j=l OX)Definition 18.3.2 As discussed earlier, defineO04 2 (xj,%i, “Xip_1)do= [, ) du (18.3)[woaEy Fu,b] Ox) ++ Ug)“ys A(x), “ig 1) : :By definition, Tuy IS the determinant ofUk)Xjuy X juz a X jugXj Xj ne1M i} U2 i Ug(18.4)Nip yup Nig_pg 0 Kip segNote how this matrix is just the matrix of Df whereFf (uw) = (xj(te) ,xi, (8) oxy, (w))”.