A ring R is called left QF-3 if it has a minimal faithful left R-module. The endomorphism ring of such a module has been recently studied in [7], where conditions are given for it to be a left PF ring or a QF ring. The object of the present paper is to study, more generally, when the endomorphism ring of a Σ-quasi-projective module over any ring R is left QF-3. Necessary and sufficient conditions for this to happen are given in Theorem 2. An useful concept in this investigation is that of a QF-3 module which has been introduced in [11]. If M is a finitely generated quasi-projective module and σ[M] denotes the category of all modules isomorphic to submodules of modules generated by M, then we show that End(RM) is a left QF-3 ring if and only if the quotient module of M modulo its torsion submodule (in the torsion theory of σ[M] canonically defined by M) is a QF-3 module (Corollary 4). Finally, we apply these results to the study of the endomorphism ring of a minimal faithful R-module over a left QF-3 ring, extending some of the results of [7].

There are a number of theorems which bound d.l.(G), the derived length of a group G, in terms of the size of the set c.d.(G) of irreducible character degrees of G assuming that G is in some particular class of solvable groups ([1], [3], [4], [7]). For instance, Gluck [4] shows that d.l.(G)≤2 |c.d.(G)| for any solvable group, whereas Berger [1] shows that d.l.(G)≤|c.d.(G)| if G has odd order. One of the oldest (and smallest) such bounds is a theorem of Taketa [7] which says that d.l.(G)≤|c.d.(G)| if G is an M-group. Most of the existing theorems are an attempt to extend Taketa's bound to all solvable groups. However, it is not even known for M-groups whether or not this is the best possible bound. This suggests that given a class of solvable groups one might try to find the maximum derived length of a group with n character degrees (i.e. the best possible bound).

In [4], R. P. Dilworth introduced the concept of a Noether lattice as an abstraction of the lattice of ideals of a Noetherian ring and he showed that many important properties of Noetherian rings, such as the Noether decomposition theorems, also hold for Noether lattices. It was later shown, in [1], that every Noether lattice is not the lattice of ideals of any Noetherian ring, yet many studies have successfully been undertaken to relate other concepts between Noetherian rings and Noether lattices as had been begun by Dilworth. (See [3], [5], and [6].) In this paper we undertake such a study and show that some results of M. Brodmann in [2] and L. Ratliff in [7] concerning prime divisors of large powers of a fixed element of a commutative Noetherian ring may be generalized and extended to the setting of a Noether lattice. It is shown (Theorem 2.8) that if A is an element of a Noether lattice then all large powers of A have the same prime divisors and (Corollary 3.8) included among this fixed set of primes are those primes that are prime divisors of the integral closure of Ak for some k≧l. We note that the ring proof of this latter result does not generalize directly since it uses the notion of transcendence degree which to our knowledge has no analogue in multiplicative lattices.

Let K be a totally real algebraic number field of degree n > 1, and let OK be the maximal order. We denote by гk, the Hilbert modular group SL2(OK) associated with K. On the extent of the weight of an automorphy factor for гK, some restrictions are imposed, not as in the elliptic modular case. Maass [5] showed that the weight is integral for K = ℚ(√5). It was shown by Christian [1] that for any Hilbert modular group it is a rational number with the bounded denominator depending on the group.

The first paragraph of Section 3 (p. 89) of the paper, published in the Glasgow Mathematical Journal 30 (1988), 87–96, should have read as follows.

One of the fundamental results of representation theory is the identification of the irreducible representations of a semisimple group by their dominant weights [3]. The purpose of this paper is to establish similar results for a class of reductive algebraic monoids.

Let k be an algebraically closed field. An algebraic monoid is an affine algebraic variety M defined over k, together with an associative morphism m:M × M → M and a two-sided unit 1 ∈ M for m.

Let X and Y be normed spaces and let L(X, Y) denote the set of linear transformations (henceforth called “operators”) T with domain a linear subspace D(T) of X and range R(T) contained in Y. The restriction of T to a subspace E is denoted by T/E; by the usual convention T|E = T|E∩ D(T). For a given linear subspace E the family of infinite dimensional ssubspaces of E is denoted by (E). An operator Tis said to have a certain property ℙ ubiquitously if every E ∈ (X) contains an F ∈(E) for which T|F has property ℙ For example, T is ubiquitously continuous if each E ∈(X) contains an F∈ (E) for which T|F is continuous. In the present note we shall characterize ubiquitous continuity, isomorphy, precompactness and smallness. A subspace of X is called a principal subspace if it is closed and of finite codimension in X. The restriction of an operator to a principal subspace will be called a principal restriction. The symbol T will always denote an arbitrary operator in L(X, Y).