THE EXTENSION AGENT CONDUCTING THE FARM AND HOME MANAGEMENT PROGRAM FOR FARMERS AND THEIR WIVES IN A NEW YORK COUNTY BEGAN USING SMALL EXPERIMENTAL STUDY GROUPS IN THE SECOND YEAR OF THE PROGRAM. USE OF SOCIOMETRIC TECHNIQUES FOR FORMING THESE GROUPS RESULTED IN REDISTRIBUTION OF THE 15 MEMBERS OF TWO ORIGINAL GROUPS TO FIVE DIFFERENT GROUPS AND IN THE RECRUITMENT OF TEN NEW MEMBERS FOR THE FIVE GROUPS. LITTLE CHANGE OCCURRED IN THE MEMBERSHIP OF THE SMALLER ORIGINAL GROUP, WHOSE FIVE MEMBERS CAME FROM A SMALL SECTOR OF THE TOWNSHIP. TEN MEMBERS OF THE LARGER ORIGINAL GROUP, PLACED BY USE OF SOCIOGRAMED CHOICES, WERE REGROUPED IN FOUR DIFFERENT GROUPS. COMPOSITION OF THE NEW GROUPS CORRELATED CLOSELY WITH NEIGHBORHOOD AND AREA LEADERSHIP PATTERNS. ONE YEAR LATER, NO SIGNIFICANT BEHAVIOR CHANGE WAS NOTED IN THE SMALL GROUP, WHICH FUNCTIONED SATISFACTORILY BOTH BEFORE AND AFTER THE SURVEY, BUT MEMBERS OF THE LARGER GROUP SHOWED GREATLY IMPROVED COMMUNICATION AND COHESIVENESS. ALTHOUGH OBJECTIVE MEASURES OF SUBJECT MATTER LEARNING WERE NOT PROVIDED, EXTENSION AGENTS GAVE A HIGHLY FAVORABLE EVALUATION OF BEHAVIOR CHANGES PRODUCED BY THE REDISTRIBUTION OF STUDY GROUPS. THE DOCUMENT INCLUDES ONE CHART, ONE TABLE, AND FOOTNOTES. (LY)