Equating multiple test forms is frequently desired. When multiple forms are linked in a chain of equating, error tends to build up in the process. This paper compares three procedures for equating multiple forms in a common-form design where each school administered, in a spiraled fashion, only a subset of multiple forms. Data used were from a tryout administration of a grade-11 proficiency examination for which 10 forms of reading, 10 of science, and 9 of mathematics examinations were administered. The results were evaluated with regard to the amount of equating error. The three approaches differed in terms of the presence or absence of "explicit" equating of equivalent groups based on the observed ability distributions and whether calibration samples were combined across common forms or not. The approach that involved explicit equating and combined-group calibrations resulted in less bias and standard error. (Contains two numbered tables, four numbered figures, five unnumbered tables, two unnumbered figures, and nine references.) (Author/SLD)