Page 89 - Vector Analysis
P. 89

§3.4 The Fubini theorem                                                                                                                 85

Since ε ą 0 is given arbitrarily, we conclude that

                                      ż                        ż (ż                       )

                                                 f (z) dz ď              f (x, y)dx dy .

                                         AˆB                      BA

                ż (ż                   )ż

Similarly,               f (x, y)dy dx ď                    f (z) dz; thus (3.3) is concluded.                                          ˝

                AB                                     AˆB

Corollary 3.26. Let S Ď Rn be a closed and bounded set such that B S has volume zero,

φ1, φ2 : S Ñ R be continuous maps such that φ1(x) ď φ2(x) for all x P S, A = ␣(x, y) P

Rn  ˆ  R  ˇ  x  P  S, φ1(x)    ď   y  ď   φ2(x)(,      and     f     : A Ñ R be continuous.              Then f         is Riemann
          ˇ

integrable over A, and

                               ż                               ż ( ż φ2(x)                   )

                                      f (x, y) d(x, y) =                        f (x, y) dy dx .                                     (3.5)

                                   A S φ1(x)

Proof. To establish that f is Riemann integrable over A, by Theorem 3.23 it suffices to show

that B A has volume zero. Let m = min φ1(x) and M = max φ2(x). Since
                                                                      xPS xPS

                   B  A  Ď   ␣(x,  φ1(x))     ˇ  x  P  S(  Y   ␣(x,  φ2(x))  ˇ  x  P  S(  Y  (   S  ˆ  [m,  M ])  ,
                                              ˇ                              ˇ                B

to  see  BA     has   volume   zero       it  suffices      to  show  that      BS  ˆ     [m, M ],   ␣(x,  φ1(x))     ˇ  x     P  S(  and
                                                                                                                     ˇ

␣(x,  φ2(x))    ˇ  x  P  S(  have  volume        zero  because       of  Lemma        3.21   and  3.22.   Note    that        Theorem
                ˇ

3.23 implies that φ1 is Riemann integrable over S; thus for a given ε ą 0 there exists

partition P of S such that

                                                    U (φ1, P) ´ L(φ1, P) ă ε .

Let B =            Ť               []                                                 Then   C   ”  ␣(x,  φ1  (x))   ˇ  x     P  S(  Ď  B
                             ∆ ˆ infxP∆ φ1S(x), supxP∆ φ1S(x) .                                                      ˇ

             ∆PP ,∆XS ‰H

and

              żż                                                  ÿ      ()
         0 ď 1C(z) dz ď 1B(z) dz ď                                        supφ1S(x) ´ inf φ1S(x) ˆ νn(∆)
                                                                                                 xP∆
                C                     B                     ∆PP,∆XS‰H xP∆

          ď U (φ1, P) ´ L(φ1, P) ă ε .

Therefore,      C     =  ␣(x,  φ1(x))  ˇ  x   P  S(    has    volume     zero   and   similarly,    ␣(x,  φ2(x))        ˇ  x  P  S(  has
                                       ˇ                                                                                ˇ

volume zero.

    Now we show that B S ˆ [m, M ] has volume zero. Since B S has volume zero in Rn, for a

given ε ą 0 there exists a partition P of B S such that

                                                    U (1S, P)  ă     M     ε    +1    .
                                                                         ´m
   84   85   86   87   88   89   90   91   92   93   94