College and Research Libraries To the Editor: The article by Joe A. Hewitt in the January 1972 issue of CRL describing an audit of a card catalog raised the question: "What is a realistic level of acceptable filing error in a catalog?" A sampling of our public catalog was carried out last summer to find the amount of filing error; the average error in both the author/title and the subject catalogs was estimated at 1.04 percent, at a confidence level of 95 percent. This result is strikingly close to the 1.1 percent reported by Hewitt, and was considered by our staff to be at an ac- ceptable level, although we, too, could find no comparative statistics on which to base a judgment. The sample was carried out by the chief cataloger, using a random sample of 94 sets of 100 cards each taken from the au- thor/title catalog and 62 sets from the sub- ject catalog. An average of 1.31 percent error was found in the former, and 0.63 percent error in the latter. It was noted, however, that many of the filing errors de- tected were only one card away from the correct location, and, therefore, the prob- ability of such cards being located during a search of the catalog seemed to be high. If we consider that the remaining cards represent a more serious problem for the user, the serious filing errors may be esti- mated at 0.86 percent in the author/ title catalog and 0.46 percent in the subject catalog, or an average of 0.7 percent in the catalog as a whole. (I would be glad to send more detailed information about our study to anyone who is interested.) Our original purpose was simply to find out what the error rate was, as we, too, are using student filers and clerical re- visers. However, we have since begun to consider alternatives to complete revision and plan to use the results obtained from the sample as a base figure against which 236 I Letters to measure the effects of such alternatives. To the Editor: (Mrs.) ]ill Le Croissette Assistant College Libmrian, Technical Services California State College Los Angeles, California The proposal suggested by Hans Wel- lisch in "Documentation-in-Source for Li- brary and Information Science" ( LRTS, Fall 1971) was read with much enthusi- asm. His recommendations deserve imple- mentation in all library service journals. If the discipline is committed to the training of indexers and abstracters (along with other information specialists) it should like- wise, it seems, be willing to provide great- er service in this respect for its own litera- ture. It should exert itself as a leader in the art of document retrieval. In essence, Wellisch was advocating not simply that abstracts should appear in the libra1y service journals with the articles- this is already a standard practice in many of them-but that the abstracts should ap- pear in the journals in a detachable form. This would be for the purpose of inter- filing into personalized information files. Wellisch also contends that these abstracts should carry concise subject headings, which in turn would facilitate retrospec- tive subject searches. He further proceeds to provide a recommendation for a stan- dardized format, and closes with the fol- lowing statement: "I hope that it will not take too long before the 'abstract frame' shown in Figure 3 will appear in this jour- nal which devotes itself to problems of bibliographical format and control, as an item in its regular feature of contents page-cum-abstracts, ready to be clipped and incorporated in your own personal in- formation file." I CRL is for me, as well as for many others, an important and heavily read jour- nal. Because I, also, often have occasion to refer back to earlier volumes, a file of abstract-index cards would prove to be of great assistance. I, therefore, urge the edi- torial board to consider the adoption of a plan to include in CRL detachable and indexed abstract cards. I would also hope that other readers would respond favor- ably to such a plan. Maybe a polling of the readership could be made on this is- Letters I 237 sue either through a questionnaire or an editorial remark. Glenn R. Wittig Reference Librarian Speer Library Princeton Theological Seminary Princeton, New Jersey Ed. note: Readers should write to the editor if they are interested in seeing the service implemented.