Theorems of Phragmèn-Lindelöf type and other related results for solutions of elliptic-parabolic equations have been given by numerous authors in recent years. Many of these results are based upon the maximum principle and the use of auxiliary comparison functions which are constructed as supersolutions of the equations under various conditions on the coefficients. In this paper we present an axiomatized treatment of these topics, replacing specific hypotheses on the nature of the coefficients of the equations by a single assumption concerning the maximum principle and another concerning the existence of positive supersolutions, in terms of which the theorems are stated.

The aim of this paper is to characterize the finite simple group PSp4(3) by the structure of the centralizer of an involution contained in the centre of its Sylow 2-subgroup. More precisely, we shall prove the following result.

In der ebenen absoluten Geometrie, die hier stets in der Allgemeinheit meines Buches Aufbau der Geometrie aus dem Spiegelungsbegriff (AGS) verstanden sei, hat man guten Grund, die Geraden vor den Punkten zu bevorzugen. Grundlegende Relationen in der Menge der Geraden sind das Senkrechtstehen von zwei Geraden und das Im-Büschel-Liegen—die Abhängigkeit—von drei Geraden, das man nach Hjelmslev wie folgt definiert: Drei Geraden liegen im Büschel, wenn das Produkt der Spiegelungen an ihnen eine Geraden-Spiegelung ist.

The principal aim of this note is the proof of the following:

Main Theorem. The following properties of the group G are equivalent:

I. G is artinian and soluble.

II. (a) To every maximal subgroup S of G there exists a normal subgroup T of G with G = ST and S ∩ T = (S ∩ T)T.

Let P be a projective plane of finite order n and Γ a group of collineations of P. Gleason (6) and Wagner (10) have shown that if every point of P is the centre, and every line the axis, of a non-trivial perspectivity in Γ, then Γ contains a subgroup of order n2 which consists entirely of elations. It then follows that either P or its dual is a translation plane with respect to at least one line; in fact if Γ has no fixed elements, then P is desarguesian and Γ contains all elations of P. It was shown by Piper (7) and Cofman (4) that the hypotheses of Gleason and Wagner can be relaxed in certain cases, while the same conclusions hold.

Many years ago, I defined (8) three types of exceptional unode on an algebraic surface, which I called U*8, U*9, U*10, corresponding, on a non-singular model of the surface, to sets of six, seven, and eight rational curves, each of grade — 2, with the intersection patterns represented by the Coxeter-Dynkin graphs now usually known as E6, E7, E8:

where each dot represents a curve, and linked dots intersecting curves. In each case we shall denote the curves in the horizontal sequence by S1, s2, … from left to right, and the extra curve meeting s3 by s*.

Kummer's surface has the base surface F of a certain net of quadrics in [5] for a non-singular model. All the quadrics of have a common self-polar simplex ∑, and can, in a double-infinity of ways, be based on a quadric Ω1 and two quadrics that Ω1 reciprocates into each other. F is invariant under harmonic inversions in the vertices and opposite bounding primes of ∑ and (§2) contains 32 lines. In §3 it is shown, conversely, that those quadrics for which a given simplex is self-polar and which contain a line of general position constitute a net of this kind.

Let [Pn(k)] be the class of all subsets Pn(k) of the k-dimensional Euclidean space consisting of n distinct points and having diameter 1. Denote by dk(n, r) the maximum number of times a given distance r can occur among points of a set Pn(k).Put

In other words Dk(n) denotes the maximum number of times the same distance can occur between n suitably chosen points in k-dimensional space.

The figures studied in this paper are special convex bodies in Euclidean three-dimensional space which we shall call generalized convex bodies of revolution (GCBR). Such a set is obtained by the following procedure. Let K1 be a convex body of revolution and let x, y, z denote Cartesian coordinates in a system for which the z-axis is the axis of K1.

When a set of axioms is laid down as the basis of any mathematical doctrine, it must be proved that this set never leads to a contradiction. In this note we turn the question around. A set of axioms is given and we wish to adjoin an axiom of a specified type. How far does the demand of non-contradiction limit the choice of the new axiom?

The central problem in the present paper is the construction of symmetric and of skew-symmetric ( = skew) matrices C of order v, with diagonal elements 0 and other elements + 1 or — 1, satisfying

The following necessary conditions are known: v ≡ 2 (mod 4) and

a and b integers, for symmetric matrices C (Belevitch (1, 2), Raghavarao (14)), and v = 2 or v ≡ 0 (mod 4) for skew matrices C.

Two permutations (displayed in the two rows)

of the integers 1, 2, … , n are called discordant if ai ≠ bi, i = 1, 2, …, n. Let v(4, n), n ⩾ 4, be the number of permutations discordant with the three Permutations

Let a semigroup A be given by generators a1, a2, … , ad and defining relations u1 = v1, u2 = v2, … , ue = ve between these generators, the ui, vi being words in the generators. We then have a presentation of A, and write

The same generators with the same relations can also be interpreted as the presentation of a group, for which we write

It is well known that planes of characteristic 2 behave differently from other Pappian projective planes. For this reason their detailed properties are usually ignored in books on synthetic projective geometry, especially when conies are being discussed. This can give rise to the misleading impression that planes of characteristic 2 are more difficult to deal with, while a cursory introduction to conies in such planes (e.g. Theorem 4.3) may suggest that the notion of “pole and polar” no longer exists.

A closed curve Kn+1 of order n + 1 in real projective n-space Rn has a maximum number of n + 1 points in common with any (n — 1)-space. These curves are subjected to certain differentiability assumptions which make it possible to describe their singular points and to provide them with multiplicities in analogy with algebraic geometry.

One of the authors having just presented Cauchy's theorem on convex polyhedra to a class on “Convexity” at the University of Wisconsin, a discussion on Cauchy's lemma on convex polygons led to an exchange of letters which are here reproduced. The three letters are independently readable. Two new proofs of Cauchy's lemma are given, of which the second (§3) is very short.

In this issue honouring Professor Coxeter, I am pleased to present some results of an investigation that was prompted by questions which he himself raised over a decade ago.

With respect to a linear system of polarities in complex projective three-space, the polars of a fixed point Q form an axial pencil of planes. The axis of the pencil is called the axis of point Q with respect to the linear system of polarities. Since there are ∞3 axes and ∞4 lines in the space, not every line is an axis. The following discussion answers the questions of how many and which lines are axes with respect to the linear systems of polarities that have a fixed self-polar tetrahedron.

An incidence structure consists of two sets of elements, called points and blocks, together with a relation, called incidence, between elements of the two sets. Well-known examples are inversive planes, in which the blocks are circles, and projective and affine planes, in which the blocks are lines. Thus in various examples of incidence structures, the blocks may have various interpretations. Very shortly, however, we shall impose a condition (Axiom A) which ensures that the blocks behave like lines. In anticipation of this, we shall refer to the set of blocks as the set of lines. Also, we shall employ the usual terminology of incidence, such as “lies on,” “passes through,” “meet,” “join.” etc.

It is known that the four-colour problem for the faces of a map on a sphere is isomorphic with the four-colour problem for the vertices of its dual, and the problem is here discussed in the latter form. The isomorphs described below are concerned with codes for the four colours and for a change in colour as we pass along an edge from vertex to vertex. In the first (algebraic) isomorph, the coding involves the four fourth roots of unity, and leads to a graphical representation in the complex plane. In the second (arithmetical) isomorph, the coding involves the integers mod 4, and also the face-edge incidence matrix of the dual.

We use the term balls for congruent, closed spheres no two of which have interior points in common. In Euclidean n-space let Nn be the maximal number of balls which can touch a ball. Obviously, N2 = 6. R. Hoppe (see (1)) proved that N3 = 12, settling thereby a famous point of controversy between Newton and David Gregory, known as the problem of the thirteen spheres (see (3)). Simpler proofs were given by Günter (6), Schütte and van der Waerden (10), and Leech (7).