On behalf of the Canadian Mathematical Congress, I wish to thank the University of Toronto for its hospitality, the members of the Local Arrangements Committee (especially Chandler Davis) for the many comforts and pleasures they have provided, and our nine distinguished visitors for the courses of lectures they gave at our Seminar.

The definition of injectivity, and the proof that every module has an injective extension which is a subextension of every other injective extension, are due to R. Baer [B]. An independent proof using the notion of essential extension was given by Eckmann-Schopf [ES]. Both proofs require the p reliminary construction of some injective overmodule. In [F] I showed how the latter proof could be freed from this requirement by exhibiting a set F in which every essential extension could be embedded. Subsequently J. M. Maranda pointed out that F has minimal cardinality. It follows that F is equipotent with the injective hull. Below Icon struct the injective hull by equipping Fit self with a module strucure.

In [2], a construction for the lower radical class R∘ (η) with respect to a class η of rings was given as the union of an inductively defined ascending transfinite chain of classes of rings. It was shown there that this construction terminates, for associative rings, at ω∘, the first infinite ordinal, in the sense that if {ηα: α an ordinal} is the chain, then R∘ (η) =ηω∘. Also, examples of classes η for which R∘ (η) = η1, η2, η3 were given.

Let Mm(F) be the vector space of m - square matrices over a field F. If X belongs to Mm(F), then xij will denote the element occuring in row i and column j of X.

Let Sm be the symmetric group of degree m and ε: Sm → F the alternating character on Sm (i.e. ε(σ) = 1 or - 1 according as σ is an even or odd permutation).

Blackett [4] introduced the concepts of near-ring homomorphism and near-ring ideal. Beidleman [1] established the fundamental homomorphism theorem and the isomorphism theorems for (left) near - rings obeying the condition that 0.a = 0 for every a in the near-ring. Several others, for example [3], [5], and [7], have taken up the study of ideals. This paper takes up the study of homomorphisms of (left) near-rings not subject to the condition 0.a = 0. It is shown that such homomorphisms can be decomposed into homomorphisms of two special sub-near-rings. Conversely, conditions are sought under which homomorphisms of the two sub-near-rings may be mated to produce a homomorphism of the sub-near-ring.

In [1] the Principle of Duality of mathematical logic (which states that the duals of equivalent well-formed formulas are also equivalent) is established by first showing that certain transforms related to the duality transform, also possess this property. The transforms involved are ,, , and which are defined as follows.

Let F[a, b] be the linear space of all real valued functions defined on [a, b]. A linear operator L: C[a, b] → F[a, b] is called positive (and hence monotone) on C[a, b] if L(f)≥0 whenever f≥0. There has been a considerable amount of research concerned with the convergence of sequences of the form {Ln(f)} to f where {Ln} is a sequence of positive linear operators on C[a, b].

Mikusinski [1] has extended the operational calculus by methods which are essentially algebraic. He considers the family C of continuous complex valued functions on the half-line [0,∞). Under addition and convolution C becomes a commutative ring. Titchmarsh's theorem [2] shows that the ring has no divisors of zero and, hence, that it may be imbedded in its quotient field Q whose elements are then called operators. Included in the field are the integral, differential and translational operators of analysis as well as certain generalized functions, such as the Dirac delta function. An alternate approach [3] yields a rather interesting result which we shall now describe briefly.

Consider the subset containing those functions for which

One never attempts to visualize ; it is just a compact blur in the infinite-dimensional space . Nevertheless, we want to establish that it shares with several other sets an odd but rather remarkable "geometric" property: it is overwhelmingly concentrated around a single element. This element we call the nucleus of .

Let be a non-decreasing sequence of non-negative numbers, and let U∘=0. Then we have

Yang [3] proved the following integral inequality:

If y(x) is absolutely continuous on a≤x≤X, with y(a) = 0, then

Let f(x) be a polynomial of degree d ≥ 3 with integral coefficients, 'say,

In a previous paper [6] I deduced, from a deep result of Lang and Weil [2], that there is a constant k1(d), depending only on d, such that for all primes p ≥ k1(d), p ⫮ ad, f(x) has a pair of consecutive residues (mod p), that i s, there exists an integer r(0 ≤r ≤ p-1) with the property that

In graph - theoretic terms a homogeneous p-dendrite, p ≥ 2, is defined as a finite singly-rooted tree in which the root has valency 1 while every other vertex has valency 1 or p. More descriptively, a homogeneous p-dendrite may be imagined to start from its root as the main, or 0th order, branch which proceeds to the first - order branch point where it gives rise top first - order branches. Each of these either terminates at its other end (which is a second-order branch point) or it splits there again into p branches (which are of third order), and so on. The order of the dendrite is the highest order of a branch present in it. For completeness, a 0-th order dendrite is also allowed, this consists of the 0-th order branch alone.

In [S], H. Sharp characterizes each topology on a finite set S = {s1, s2,…sn} with a n×n zero-one matrix T = (tij) where tij=1 if and only if . In this paper we seek matrix characterizations of certain topological properties of finite spaces. Such characterizations will provide purely mechanical ways of determining if a space has a certain topological property.

Consider a two associate class partially balanced incomplete block (PBIB) design [2] with parameters of the first kind t, b, r, λ,1 λ,2 n1, n2, and parameters of the second kind . Let the letters m, p, l, s represent treatments and define associates of treatment m, λmps = number of times the treatments m, p and s all occur together in the same block, λmpls = number of times the treatments m, p, l and s all occur together in the same block.

Some of the most important G-structures of the first kind (1) are those defined by linear operators satisfying algebraic relations. Let J be a linear operator acting on the complexified space of a differentiable manifold V, and satisfying a relation of the form

where λ is a complex constant and I is the identity operator. In the case λ ≠ 0 the manifold has an almost product structure (2) which in the case λ = i reduces to an almost complex structure (3). In the remaining case, λ = 0, the manifold has an almost tangent structure (4).

Let X be a random variable whose density (or distribution if discrete) f(x; θ) depends on an unknown parameter θ, real or vector-valued. By making observations on X we want to know whether there exist estimates of prescribed accuracy for the real-valued parametric function g(θ). By an estimate of prescribed accuracy for g(θ) we mean a confidence interval of prescribed length and confidence coefficient or a point estimate with prescribed expected loss W. In the following our loss functions W will always satisfy the requirement that W(δ, θ) = V(|δ - θ|), where V is a strictly increasing function of its argument. The class of such loss functions includes among others the squared error loss.

In this paper we give necessary and sufficient conditions that a continuous transformation f: A→A of a metric space A with the metric r should be a contraction with respect to an equivalent metric s. This is the solution of a problem stated by J. S. W. Wong [2].