Brown 1935
Brown 1935
Brown 1935
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp
.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact [email protected].
American Mathematical Society is collaborating with JSTOR to digitize, preserve and extend access to
Transactions of the American Mathematical Society.
http://www.jstor.org
I[x
Ov av\ [ 7 vi v
71 ( /v) (vi /v\ ) 11
X 9x2 OxI Ox2 LOx, axr+1 ax, Oxr+iJ
a(vl, V2, * .
. Vr+i)
kl 2 k!)
Therefore
thetransforms ofthepointsofq lie in a cube ofedge <t ; k, hence
ofvolume_ mE km. Sincethevolumeof theoriginalcube q is en theratioof
thevolumesis ;mekm-n. Now kmr-n0, by (3.1), since k= km;and r and e
oneon L2,whoserespective
coordinates
x2, * *, xndiffer
byat mostE. Hence
theircoordinatesxi differby at mostME1> Ei. In applyingTaylor's theorem,
all derivativesof ordersless than k are takenat the firstpoint,and the re-
maindertermis a sum of derivativesof orderk. Therefore,by (3.6), the
corresponding twopointsin (u)-spacehave coordinatesrespectively differing
by less than
nk(Me,) HI
1
k! 2
Therefore ifwe covera portionof (x2, , xn)-spacecontaining
12by a net-
workof (n- 1)-cubes,of edge e1,each of whichhas at least one pointin 12,
each of thesecubes is mappedby (3.5) and (3.4) on a subsetof a cube in
(u)-space of edge H(El) k. The ratio of the volumesof two such cubes is
n
(Hi) m(El) km-n+l. Since k = k km-n+ 1 >0. As Ei can be made as small as
we like,we concludethattheimageofL2 has Jordanmeasurezero,henceis
nowheredense.
If k=2, Lemma 3.I is now proved.If n? 2m thenk mustequal 1 or 2,
and we see thatwe have nowprovedthefollowing:
LEMMA 3.II. If n?2m, Lemma 3.I is true.
If k> 2, we continuewiththeproofofLemma3.I. Againby use ofLemma
2.I we findthatwe can nextturnour attentionto a closed subsetL3 of L,
neighboringa point of L, with all partial derivativesof vi, ,)vm,of orders
1 to k-2 inclusive, zeroat eachpointofL3,butwithsomeparticular(k- 1)st-
orderpartialderivative,say ak-lVl/dxlk-l, not zero in L3. Applyingthe im-
plicitfunctiontheoremto the locus a k-2v1/xI k-2= 0, we obtain xl =2(X2,
* Xn), 42 of class C(t-k+2). We substitute
thisfunction
intotheequations
(3.4), obtaining
(3.7) ui = (X2, Xn) ( = 1, ,m)
Let (cf.(3.1))
(3.13) j =
?j(Ul, x2, . , xn) (j = 2,*. , m), ifn > 1,
= fj(l)
?>j (j = 2, , m), if n = 1,
withs of class C(8). (Note that m>1 sincer>0 and m>r.) We treatthe
case n >1, as it is thenobvioushow to treatthe case n=1. We take A so
small thatit is part of thelocus forwhich(3.12) is valid. We now let X be
* If n < m, theonlycondition
is thattheu's be ofclassC', whichis theresultofK andS.
the2 withthefirsttermoftheright
From(3.2), t(m,2m) = 2, and combining
hand memberof (3.17) we obtain
2
(3.18) t(m, n) = km + km + km + ** + km ifn > 2m + 1.
Case I. n> 2m+2. Then n-2m+r>I and (3.8) holds.Sincein thiscase
n > m, kmn-r>-k_Ra relationbetweenthe firsttermsof the sumsin (3.8) and
(3.18) respectively.The second termin (3.8) is obviously(from(3.1)) at
leastas greatas thesecondtermin (3.18), etc. As thereare at least as many
lowing properties.
(i) F(u, y) is definedin all of real (u, y)-space and has continuouspartial
derivativesof thefirstorderthere.
(ii) For each (y0), F(u, y0) 0 in each regionof (u) -space.
(iii) F[vi(x, y), * v**m(X, y), yi . , yp](f(x, y) =0 at each point ofB.
DEFINITION 2. Functions vZ(xi, m, , de-
Xnx Yi, . .* yP), j=1,
fined in a domain D of (x, y)-space, are said to be dependentin x1, X, *
where the A's are analytic near (0, 0, 0), and A (O, 0, 0) =0, j=1,I , m.
In (6.1) the firstidentitysimply defines 4. Next we note that since
f(xl, 0, *
... *, O)0, (xi, 0, * * , O)O0, so that we can apply the Weier-
0,
stressPreparationTheorem:to 4, givingus thesecondidentityin (6.1), to-
getherwith(6.2).
Since 4. = -1I 50, kis irreducible? at (0, 0, 0), hencewe see from(6.1)
thatH is irreducible at (0, 0, 0). Therefore the discriminant|l/A(x2,* * *, Xn,
-= - fz8/fXz1
ax8