The main purpose of this study was to formulate a classificatory system for dances on the basis of how these are perceived by the spectators, using factor analytic techniques. Forty dances on 16mm achromatic films were serially presented on screen to 48 female students in a four-year university and a junior college. Each subject rated each dance on a set of 40 bipolar adjectival seven-point scales. A 40×40 between-dance correlation (over scale means) matrix was generated, and then subjected to analysis by the principal factor method. Five factors were extracted, with the total eigenvalue of 33.87 that accounts for 84.68% of the total variances. Varimax rotation, applied to the factor loading matrix derived by the principal factor method above, yielded four meaningful orthogonal factors. These four factors represent four groups into which the 40 dances are classified. They were named (1)lyric, (2)epicformal, (3)dramatic, and (4)fanatic, and this system constitute the proposed Factor Analytic Classification (FAC) system. Secondly, a three dimensional Euclidean space was constructed with three primary axes obtained after the Varimax rotation, and the 40 dances were allocated into the resultant Octants. This octant classification was compared with Laban's eight Effort Elements, namely, Float, Glide, Press, Wring, Glick, Dab, Punch and Slash, called Effort Octant Classification (EOC). Thirdly, the 40 dances were classified into comparable octants on the basis of their SD-scores obtained along the three semantic dimensions of (1)dynamics, (2)evaluation, and (3)clarity. This classification is referred to as Semantic Octant Classification (SOC). By combining FAC and SOC, it was possible to derive a quasi family tree of dances. Finaly, an attempt was made to construct Semantic Effort Octant Classification (SEOC) system by combining SOC and EOC, without a marked success. The failure may be attributed to the insensitivity of the SD scales to the Effort differences.