The n! operations Ai of permutations upon n different ordered symbols correspond to n! matrices Ai of the nth order, which have in each row and in each column only one non-zero element, namely a unit. Such matrices Ai are called permutation matrices, since their effect in premultiplying an arbitrary column vector x = {x1x2….xn} is to impress the permutation Ai upon the elements xi. For example the six matrices of the third order

are permutation matrices. It is convenient to denote them by

where the bracketed indices refer to the permutations of natural order. Clearly the relation Ai Aj = Ak entails the matrix relation AiAj = Ak; in other words, the n! matrices Ai, give a matrix representation of the symmetric group of order n!.

It is well known that the canonical system of curves on an algebraic surface is only relatively invariant under birational transformations of the surface. That is, if we have a birational transformation T between two surfaces F and F′, and if K and K′ denote curves of the unreduced canonical systems on F and F′, then

where E and E′ denote the sets of curves, on F and F′ respectively which are transformed into the neighbourhoods of simple points on the other surface.

The following algebra possesses certain points of interest and is, I think, worth putting on record; it includes the algebra of matrices as a special case. Consider the algebra H over a ring F defined by

where hpq(p, q = 1, 2, …., n) are linearly independent over F. If are any elements of H, then from (1)

In the general theory of relativity, as in many other branches of theoretical physics, the material and energetical content of spacetime is considered, in the first instance, as an extended field, which is specified by means of field quantities (energy-momentum-stress tensor, charge-current density, electromagnetic field strength). From this point of view corpuscles (material particles, photons) are constructs obtained by first considering the field quantities as nonvanishing only within certain world tubes, and then passing by limiting processes to the idealisation in which these world tubes are shrunk into world lines. More precisely, this passage to the corpuscular description may be thought of as accomplished by replacing the original field by successive members of a sequence of field distributions, satisfying the same field laws, which cluster more and more in the neighbourhood of the world lines, and for which in some significant sense the total measure approaches a finite limit. Each such world line, together with the limiting measures of those portions of the field quantities associated therewith, is then a corpuscle; the form of the world line determines the motion of the corpuscle, and the associated “corpuscular quantities” its physical attributes (mass or energy, momentum, charge).

In this paper we consider the correspondence between tangential quadrics of [3] and points of [9]. Godeaux has considered this geometrically, with the object of obtaining a representation for a twisted cubic of three dimensions. We have considered it from a standpoint more algebraic than that of Godeaux, with particular reference to the types of pencils of quadrics that correspond to special lines of [9], and to the interpretation in [9] of the fact that the condition for a net of quadrics to be part of the polar system of a cubic surface is poristic.

If m, n are integers, m > 0, n > 1, the generalized Stirling numbers are defined by the identity in x,

We investigate in this paper a certain special family of quadric varieties, that is of in [R]. Now among the more important properties of a quadric in [R] is that it possesses a system or systems of “ generators,” i.e., the quadric may be taken as the locus of certain families of subspaces, the behaviour of these depending on the parity of R. If R is even, a quadric V2n–1 in [2n] contains a single family of [n—l]'s, so it seems likely that in discussing special families of quadrics in [2n] an important type will be obtained by constraining the quadric to pass through a number of [n — l]'s.

This paper deals with certain expansions of analytic functions in series of polynomials. Explicit forms for the polynomials are given in terms of the coefficients of the Taylor's expansion and of prescribed positive constants hk. Under suitable conditions, to be presently discussed, the series converge to the function in the Borel region.

It is proposed here to consider the sequence un determined by the relation

where, in particular,

and initially u1 = θ1. The following is the main result to be proved.

An important problem in finite-group theory is the determination of an abstract definition for a given group , that is, a set of relations

between k generating operations S1, …., Sk of , such that every other relation between S1, …., Sk is an algebraic consequence of (1).

The number of groups for which abstract definitions are actually known is relatively small, but a remarkable feature of the results already obtained is the extreme simplicity of the relations (1) in the case of several groups of quite high order. This fact constitutes an additional incentive to the search for abstract definitions, and many elegant results have doubtless yet to be discovered.

The infinite integrals

in which λ is positive and – π < θ < π were encountered by Kottler in a problem in the theory of diffraction. They have more recently been studied by Copson and Ferrar, who obtained the remarkably simple Fourier series

in which denotes a “cut Bessel function” of the third kind; this expansion is valid when the term has to be added to the expansion on the right.

It is well known that the real, skew-symmetric, non-singular, bilinear forms of n + n variables have no invariants. In fact, each of these forms may be transformed into one and the same form, for instance into the one which occurs in the usual representation of the complex group. The standard proofs of this theorem break down in case of infinite forms which are bounded in the sense of Hilbert, one of the impediments being the possibility of a continuous spectrum. The object of this note is to show that, while the usual proofs break down, the theorem itself is true in Hubert's case also.

Whittaker and Ruse have developed forms of Gauss's theorem in general relativity, their theorems connecting integrals of normal force taken over a closed 2-space V2 with integrals involving the distribution of matter taken over an open 3-space bounded by V2. The definition of force employed by them involves the introduction of a normal congruence (with unit tangent vector λi), the “force” relative to the congruence being the negative of the first curvature vector of the congruence (– δλi/δs). This appears at first sight a natural enough definition, because – δλi/δs at an event P represents the acceleration relative to the congruence of a free particle travelling along a geodesic tangent to the congruence at P. In order to give physical meaning to this definition of force it is necessary to specify the congruence λi physically, and it would seem most natural to choose the congruence of world-lines of flow of the medium. Supposing certain conditions satisfied by this congruence (cf. Ruse, loc. cit.), the theory of Ruse is applicable, and from this follows a form of Gauss's theorem.

The object of this paper is to investigate some properties of series which satisfy conditions of the form

where 0 < ρ ≦ p. denotes, as usual, the n-th Cesàro sum of order p for the series ∑an and the binomial coefficient . It is convenient to state here some properties of and to which we must constantly refer in the sequel.

It is well-known that if v≧–½ and

then ø (x) is said to be self-reciprocal in the Hankel transform and may be described as Rv. If v = ±½(1) reduces to the Fourier sine or cosine transform. Functions of these two classes may be described as Rg and Re.

In a list of pairs of reciprocal functions, G. A. Campbell gives the example that

is Rc, but there does not seem to be any explicit reference to this function in the literature. It suggests that there is a corresponding function which is Rv.

The circumstances explained in the footnote on p. 61 of the former paper with this title might well have necessitated a re-writing of the whole. Fortunately it appears that only a few short comments are required. For example it may be noted that the first question in §4 can be answered by counting the degrees of freedom in the two configurations. Eight points of a twisted cubic have freedom 20; four pairs of planes drawn at random through four lines of a regulus have freedom 21; therefore the eight planes of the second paragraph of §4 cannot always lead back to the eight points of §2. This is corroborated by the corresponding numbers in [4], which are 31 and 34.

In Kottler's theoretical discussion1 of the diffraction of a plane wave of monochromatic light of wave-length 2π/k by a black halfplane, the function

where (r, θ, z) are cylindrical coordinates, plays an important part. In particular it is necessary to have asymptotic formulae for f (r, θ), valid when r is either very large or very small compared with the wave-length.

An affine connection in an n-dimensional manifold Xn defines a system of paths, but conversely a connection is not defined uniquely by a system of paths. It was shown by H. Weyl that any two affine connections whose components are related by an equation of the form

where is the unit affinor, give the same system of paths. In the geometry of a system of paths, a particular parameter on the paths, called the projective normal parameter, plays an important part. This parameter, which is invariant under a transformation of connection (1), was introduced by J. H. .C. Whitehead. It can be defined by means of a Schwarzian differential equation and it is determined up to linear fractional transformations. In § 1 this method is briefly discussed.

Schouten and van Kampen (1) have studied the deformation of a . Applying the methods of that paper to the tangent vectors , which exist by hypothesis at all points of a certain region Vm′ (m′ > m) of Vn, we shall have

Whence we define the differentials

In the application of the to the lower index is treated as an ordinal index only. We shall not be concerned with any extension of the to indices other than those of the general Vb (see 1, equ. 3.24).

In an earlier “Note on Selection from a Normal Multivariate Population” the author considered the transformation of an n-variate normal population induced by “selection” operating to alter the parameters of frequency in a subset of p variates, while preserving normal frequency in the subset. The results were new in notation only; in substance they were first found by K. Pearson.