This list of books and articles in the field of numerical analysis is intended to supplement and bring up to date one published by the present writer as an appendix to his Principles of Numerical Analysis (McGraw-Hill, 1953; this will be designated PNA). In the two years since its preparation a number of omissions have been noted. More to the point, however, has been the rapid growth of interest in the subject and the concomitant proliferation of the literature. As a byproduct and aid to one person's endeavor to keep up with the field, the supplementary list has grown to a length comparable to that of the original.Admittedly it is incomplete, and doubtless always will be. Nevertheless, with the encouragement of some, it is offered for such help as it may provide to others who are trying to find their way about. The incompleteness is partly deliberate, in that the list is intended to cover only that area that was delimited in the preface to PNA. In general, the somewhat self-contradictory designation \"algebraic numerical analysis\" might describe the area in question. Thus, papers on functional equations (integral, differential, etc.) are included only when they seem to contribute to the algebraic phase of the subject. A few references on continued fractions are listed even though continued fractions were nowhere discussed in PNA. On the other hand, the noninclusion of papers on linear inequalities has no such rational justification. They have been left out quite simply because they have not been included. In general, while a few excisions have been made, the omission of a paper is by no means necessarily a reflection upon its quality, even when the author knows that an actual reprint is in my file. It may signify careless book-keeping. But, in particular, papers which appear in collections have not been listed separately. Hence a number of important papers are included implicitly but not explicitly, papers by Fan, Todd, Fox, Geiringer, and numerous others. There may appear to be a disproportionate number of papers dealing with matrix algebra, and this may well be so. The representation may or may not reflect accurately either the relative importance or the relative degree of activity in the field. Moreover, it may be entirely fortuitous that in the day-today work of one particular computing laboratory (here at ORNL) it has become necessary , over the past two or three years, to draw …

P. G. Guest

Papers

The Estimation of Standard Error from Successive Finite Differences