Con base en descriptores para la caracterizacion In Situ de Passifloraceae desarrollados por el IPGRI, se caracterizaron morfologicamente 189 genotipos del genero Passiflora colectados en el municipio de Pasto. El Analisis de Componentes Principales (ACP) y el Analisis de Correspondencias Multiples (ACM) permitieron caracterizar las accesiones y formar grupos mediante el metodo de clasificacion jerarquica. En el ACP, los tres primeros componentes explicaron el 75.63% de la variacion total; se formaron 5 grupos. El primer grupo se destaco por la longitud del fruto (106mm) frente al promedio general (86mm) y peso del fruto (97.03g) superior al promedio general (77.59g); estuvo conformado por P. tarminiana Coppens & Barney, P. tripartita var mollissima (Kunth) Bailey y una aseccion de P. mixta. El grupo dos lo conformaron P. Mixta, p. Cumbalensis y una accesion de P. Tripartita var. Mollissima; en el grupo tres esta P.manicata, en el grupo cuatro P. ligularis y en el grupo cinco P. gracilima, P. alnifolia, P. andreana y P. edulis f. edulis. El ACM expreso un 41.81% de la variacion en los tres primeros factores; el analisis clasificatorio permitio conformar 6 grupos, el subgenero Tacsonia (P: tripartita var mollissima, P. tarminiana, P. mixta y P. cumbalensis) se ubico en los grupos uno, dos, tres y cuatro presentando poca divergencia entre si; ademas presento diferencias en la forma, sabor y color del fruto y dureza de la cascara, con los grupos cinco (P. ligularis, P. edulis f. edulis y P. manicata) y seis (P. andreana y P. alnifolia) existiendo la posibilidad de realizar hibridaciones para el mejoramiento de las especies comerciales. ABSTRACT A series of phenotypic descriptors developed by The IPGRI were used to characterize species in situ of Passifloraceae. The present research characterized and evaluated morphologically 189 specimens of the genus Passiflora collected in the Pasto municipality. The results, using Principal Component Analysis and Multiple Classification Analysis, indicated significant differences among the data collected. As a consequence of these differences evident groups of data were showed (hierarchical classification). Using the PCA method, 75.63% of the total variation was explained by the first three components. As a result, five groups were formed. The first group is visibly separated from the rest because of the following reasons: fruit size and fruit length (Lfr) (106mm), compared with the average of 86 mm, and fruit weight (97.03gr) compared with the general average (77.59gr). This first group was composed by P. tarminiana Coppens & Barney, P. tripartita var mollissima (Kunth) Bailey and an aseccion of P. mixta. As a second group, species like P. Mixta, p. Cumbalensis and an accesion of P. Tripartita var. Mollissima, also P. Manicata. The fourth group was represented by P. ligularis and the fifth group by P. gracilima, P. alnifolia, P. andreana and P. edulis f. edulis. The results of the MCA showed a 41.81% of variation within the first three factors. The method of classification allowed clustering 6 groups. The subgenus Tacsonia (P. tripartite var mollissima, P. tarminiana, P. mixta y P. cumbalensis) was found into the groups one, two, three and four; presented little divergence between thern, but high variability of size, taste, color and hardiness of their seed coat when it is compared to the fifth (P. ligularis, P. edulis f. edulis y P. manicata) and sixth group (P. andreana y P. alnifolia); results that give the opportunity to improvement hybrid commercial species.