A 2-step clustering procedure, using indexes derived from principal component analysis, was used to disclose sperm subpopulations within the canine ejaculate and its relationship to sperm cryoresistance. Semen from 4 dogs was frozen-thawed by a standard protocol: before freezing, computer-assisted sperm analysis of motility and morphometry were performed; after thawing, motility analysis was performed again; and cryoresistance was estimated as the percent changes in progressive motility and sperm velocities after thawing. We used indexes derived from principal component analysis (sperm velocity index [SVI] and sperm motility index [SMI]) and the SPSS 2-step cluster method to disclose sperm subpopulations. The 2-step clustering procedure revealed the existence of 6 subpopulations. Subpopulations 4 and 6 were characterized by high values of both SVI (>200 arbitrary units) and SMI (>90 arbitrary units), subpopulations 2 and 3 were characterized by medium values (SVI 100 to 130; and SMI 30 to 40), and subpopulations 1 and 5 were characterized by low values (SVI < 100; SMI < 30). The distribution of sperm subpopulations was completely different among dogs. Four sperm subpopulations based in morphometric parameters of the sperm head and midpiece were revealed. Models including SVI and SMI indexes explained curvilinear velocity (R(2) = 0.997; P < .001), straight-line velocity (R(2) = 0.98; P < .001), and average velocity (R(2) = 0.99; P < .0001) postthaw.