These notes are an expanded and updated version of a course of lectures which I gave at King's College London during the summer term 1979. The main topic is the Hermitian classgroup of orders, and in particular of group rings. Most of this work is published here for the first time. The primary motivation came from the connection with the Galois module structure of rings of algebraic integers. The principal aim was to lay the theoretical basis for attacking what may be called the "converse problem" of Galois module structure theory: to express the symplectic local and global root numbers and conductors as algebraic invariants. A previous edition of these notes was circulated privately among a few collaborators. Based on this, and following a partial solution of the problem by the author, Ph. Cassou-Nogues and M. Taylor succeeded in obtaining a complete solution. In a different direction J. Ritter published a paper, answering certain character theoretic questions raised in the earlier version. I myself disapprove of "secret circulation," but the pressure of other work led to a delay in publication; I hope this volume will make amends. One advantage of the delay is that the relevant recent work can be included. In a sense this is a companion volume to my recent Springer-Ergebnisse-Bericht, where the Hermitian theory was not dealt with. Our approach is via "Hom-groups," analogous to that followed in recent work on locally free classgroups.

We begin by making clear the meaning of the term "tame." The higher ramifi cation groups, on the one hand, and the one-units of chain groups, on the other, are to lie in the kernels of the respective representations considered. We shall establish a very natural and very well behaved relationship between representa tions of the two groups mentioned in the title, with all the right properties, and in particular functorial under base change and essentially preserving root numbers. All this will be done in full generality for all principal orders. The formal setup for this also throws new light on the nature of Gauss sums and in particular leads to a canonical closed formula for tame Galois Gauss sums. In many ways the "tame" and the "wild" theory have distinct features and distinct points of interest. The "wild" theory is much harder and - as far as it goes at present - technically rather complicated. On the "tame" side, once we have developed a number of new ideas, we get a complete comprehensive theory, from which technical difficulties have disappeared, and which has a naturality, and perhaps elegance, which seems rather rare in this gen, eral area. Among the principal new concepts we are introducing are those of "similarity" of represen tations in both contexts and that of the Galois algebra of a principalorder., One might expect that this Galois algebra will, also be of importance in the wild situation."