We are concerned with the nth-order linear differential equation

1

where the coefficients are continuous. Aliev [1, 2] showed, in papers unavailable to the author that for n = 4

(see Definition 2). Theorems 1 and 5 give respectively nth-order generalizations of these two results.

Soit R un anneau primitif à droite et soit A un idéal à droite de R. Dans un récent article [2], Ménard a montré que A est minimal si et seulement si abA = baA ∀ a, b ∊ A. Le but de cette note est de montrer que A est minimal si et seulement si aA = a2A ∀ a ∊ A.

Lemme. Soit R un anneau semi-simple [1], soit A un idéal à droite de R et soit S = Rad(A) le radical de Vanneau A. Alors A est un idéal à droite minimal de R si et seulement si Ā = A/S est un corps.

Let In be the graph of the unit n-dimensional cube. Its 2n vertices are all the n-tuples of zeros and ones, two vertices being adjacent (joined by an edge) if and only if they differ in exactly one coordinate. A path P in In is a sequence x1, …, xm of distinct vertices in In where xi is adjacent to xi+1 for 1 ≤ i ≤ m-1; P is a circuit if it is also true that xm and x1 are adjacent. A path is Hamiltonian if it passes through all the vertices of In. Finally, for vertices x and y in In, we define d(x, y) to be the graph theorectic distance between x and y, i.e., the number of coordinates in which x and y differ.

Whitman [2] solved the word problem for lattices by giving an explicit construction of the free lattice, FL(X), on a given set of generators X.

The solution is the following:

For x, y ∊ X, and a, b, c, d ∊ FL(X),

(W1)

(W2)

(W3)

(W4)

where [p, q] = {x; p ≤ x ≤ q}.

The purpose of this note is to give a simple nonconstructive proof that the condition (W4) must hold in every projective (hence every free) lattice. Jonsson [1] has shown that in every equational class of lattices (Wl), (W2), and (W3) hold. Therefore the combination of these results gives a complete nonconstructive solution to the word problem for lattices.

Let Rm denote a m dimensional Euclidean space. When x ∊ Rm will write x = (x1, x2,..., xm). Let R+m ={x: x ∊ Rm, xi < 0 for all i} and R-m ={x: x ∊ Rm, xi < 0 for all i}. In this paper we consider a class of functions which consists of mappings, Er(K) and Hr(K) of Rm into R which are indexed by K ∊ R+m and K ∊ R-m respectively, and defined at any point α ∊ Rm by

1.1

Using the theory of algebraic numbers, Mordell [1] has shown that the Diophantine equation

1

possesses only two solutions in positive integers; these are given by n = 2, m = 1, and n = 14, m = 5. We are interested in positive integer solutions to the generalized equation

2

and in this paper we prove for several choices of k and l that (2) has no solutions, in other cases the only solutions are given, and numerical evidence for all values of k and l for which max (k, l) ≤ 15 is also exhibited.

Let G be a finite connected graph without loops or multiple edges. A maximal tree subgraph T of G is called a spanning tree of G. Denote by k(G) the number of all trees spanning the graph G. A. Rosa formulated the following problem (private communication): Let x(≠2) be a given positive integer and denote by α(x) the smallest positive integer y having the following property: There exists a graph G on y vertices with x spanning trees. Investigate the behavior of the function α(x).

The purpose of this note is to prove that an involutionf on a cyclic Peano space S leaves some simple closed curve in Ssetwise invariant.

We shall first define the required terms. A Peano space is a locally compact, connected and locally connected metric space. A connected space is called cyclic if it has no cut-point. An involution on a space is a periodic mapping whose period is 2; it is necessarily a homeomorphism. A mapping f: X → X is said to leave a subset E of S setwise invariant if f(E) = E. These definitions may be found, for example, in [2].

As has long been known, the irreducible tensor representations of GL(d) of rank n may be labeled by means of the irreducible representations of Sn, i.e., by means of the Young diagrams [λ], where λ1 + λ2 + … λr = n. We denote such a tensor representation by 〈λ〉. Using Young's raising operator Rij we can write [1, p. 42]

1.1

where the dot denotes the inducing process. For example, [3] . [2] is that representation of S5 induced by the identity representation of its subgroup S3 × S2.

Let Q denote the field of rational numbers. If m, n are distinct squarefree integers the field formed by adjoining √m and √n to Q is denoted by Q(√m, √n). Since Q(√m, √n) = Q(√m, √n) and √m + √n has for its unique minimal polynomial x4 —2(m + n)x2 + (m - n)2, Q(√m, √n) is a biquadratic field over Q. The elements of Q(√m, √n) are of the form a0 + a1√m + a2√n + a3√mn, where a1, a2, a3 ∊ Q. Any element of Q(√m, √n) which satisfies a monic equation of degree ≥ 1 with rational integral coefficients is called an integer of Q(√m, √n). The set of all these integers is an integral domain. In this paper we determine the explicit form of the integers of Q(√m, √n) (Theorem 1), an integral basis for Q(√m, √n) (Theorem 2), and the discriminant of Q(√m, √n) (Theorem 3).

In [3] D. B. Fuks defined a duality of functors in the category of weak homotopy types. In general this duality is more difficult to work with than the duality of functors of the category of pointed Kelley spaces [2]. It happens however that all so-called strong functors [2] F of induce functors of , and if we denote the duality operators of and by and D respectively, then there are many cases where .

Etant données deux variétés différentiables compactes M et N, nous avons la formule de Künneth

qui établit la relation entre la cohomologie réelle de la variété produit avec celle des facteurs. Cette formule est bien connue en topologie algébrique, et elle est valable dans le contexte plus général de la catégorie des espaces topologiques. Nous voulons donner ici une démonstration de cette formule par la théorie harmonique de Hodge-Kodaira [1]. Une telle démonstration n'est pas explicite, autant que je sache, dans la littérature; elle est désirable, par exemple, dans un cours de géométrie différentielle sur les formes harmoniques.

It is well known that the complex number field has infinitely many automorphisms. Moreover, it seems to be part of the folklore that the family of all automorphisms of the complex field has cardinality 2c, where c = 2ℵo. In this article the following generalization of this fact is proved: If k is any algebraically closed field then the family of all automorphisms of k has cardinality 2card k.

The complex field has infinite transcendency degree over its prime subfield. For fields of this type the proof is accomplished by essentially permuting the elements in a transcendency basis and extending each permutation to an automorphism of the field.

Lehmann [1] in his lecture notes on estimation shows that for estimating the unknown mean of a normal distribution, N(θ, 1), the usual estimator is neither minimax nor admissible if it is known that θ belongs to a finite closed interval [a, b] and the loss function is squared error. It is shown that , the maximum likelihood estimator (MLE) of θ, has uniformly smaller mean squared error (MSE) than that of . It is natural to ask the question whether the MLE of θ in N(θ, 1) is admissible or not if it is known that θ ∊ [a, b]. The answer turns out to be negative and the purpose of this note is to present this result in a slightly generalized form.

We consider an elliptic partial differential equation with constant coefficients and zero on the right hand side. It is well known [1] that every solution of such an equation can be approximated uniformly on each compact set by a sum of products of polynomials and exponential functions which satisfy the equation. Furthermore, if one assumes that the polynomial operator is homogeneous, then the approximation can be made with polynomials alone. It is our purpose to show, in the latter case, when the number of variables is two, that each solution can be written as an infinite series in certain specific polynomials. Our method is to factor the polynomial and build up the solution in terms of solutions of first degree equations.

Let (Ω, , P) be a probability space, and A1, A2… be a sequence of members of . The classical Borel-Cantelli problem is to determine the probability that infinitely many events Ak occur. The classical results may be found in Feller [2, p. 188]; while related work may be found in Spitzer [3, p. 317], and Dawson and Sankoff [1]. The latter works are generalizations of the Borel-Cantelli lemmas, taken in different directions.

In this paper, necessary and sufficient conditions will be given for infinitely many events Ak to occur, with probability 1. A lower bound for the probability that only finitely many Ak occur, is developed. In addition, necessary and sufficient conditions that only finitely many Ak occur, with probability 1, are given.