Minor errors occurred in table 1 of the paper by Halasz et al (Gut1999;44:274–8) and hence appear throughout the text. The correct figures (given below) do not affect the significance levels quoted nor the interpretation of the data. The authors regret any inconvenience these errors may have caused.

• p274, abstract, left hand column, line 24: “14 (24)%” should read “13 (22%)”; line 27: “seven of ten” should read “six of nine”;

• p275, materials and methods, left hand column, lines 5 and 6: “. . .for 10 of the 58 patients that had GBV-C RNA or. . .” should read “. . .for nine of the 58 patients that had GBV-C RNA and/or. . .”; right hand column, line 35: “To simplify. . .SD.” should read “. . .SD. Only samples which were reactive in all three independent runs were considered to be positive.”; line 58: “14 (24)%” should read “13 (22)%”;

• p276, results, legend to fig 1, line 5: “The horizontal… (LTx).” should read “… (LTx). GBV-C E2 EIA data obtained from the first run have been plotted only as an example of an EIA outcome. However, all samples were run in triplicate and only those reactive in all three runs were regarded as true positive. Consequently some positive points in the figure were scored as negative in the final evaluation.”; right hand column, line 7: “(4/27; table 1)” should read “(3/27; table 1)”;

• p277, left hand column, line 30: “14” should read “13”; line 33: “Three” should read “Two”; discussion, right hand column, line 12: “36%” should read “34%”; line 52: “patient FH4” should read “patient FH7”.