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 support@jstor.org.. The Johns Hopkins University Press is collaborating with JSTOR to digitize, preserve and extend access to American Journal of Mathematics. This paper is a discussion of some relations between the Hilbert theory and the existence problem for normal extensions of relatively complete fields. The Hilbert theory describes the subfields of a given normal field by means of the arithmetic inherent in the fields in question. The results of this theory may be interpreted as necessary conditions for the existence of normal extensions over a base field. We propose the question whether such conditions are sufficient for the solution of existence problems. To answer this query it is necessary to construct explicitly normal extensions taking advantage of information obtained by the Hilbert theory. A. A. Albert, 0. Ore and the author 1 have investigated independently such problems for the classical fields of p-adic numbers. In this article we take general relatively complete fields as reference fields. A field is termed relatively complete if Hensel's Lemma holds for the prime ideal of non-units with respect to the given valuation.2 The normal extensions involved are mostly of infinite degree over the ground field. This generality leads to an understanding of the mechanism underlying the structure of normal extensions. We have purposely selected relatively complete fields rather than complete fields as base fields since relative completeness can always be established for infinite extensions.3In the first part of the paper we develop the Hilbert theory for arbitrary normal extensions L over a relatively complete field F with value group r and perfect residcue class fields 5. We prove that L contains a unique maximal unramified field LI whose residue class field coincides with that of L. The field L contains, relative to LI, a radical extension LR, the ramification field, which is determined by the structure of the value group r (L) of L. As a matter of fact, the value group r (LB) of LR consists of all those elements of F(L) whose orders modulo r are prime to the characteristic X of F. Moreover, the field LI contains all roots of unity whose orders occur as orders of the elements in the factor group r (LR) /r. In case L contains only elements whose degrees over F are prime to X we find L = LR. Thus normal extensions * Presented to the Society, Dec. 29-31, 1941. Received October 29, 1941.