A statistical approach using sequentially principal component analysis (PCA)clustering and discriminant analysis was developed to disclose morphometricsperm subpopulations. In addition, we used a similar approach to disclosesubpopulations of spermatozoa with different degrees of DNA fragmentation. It iswidely accepted that sperm morphology is a strong indicator of semen quality andsince the sperm head mainly comprises the sperm DNA, it has been proposed thatsubtle changes in sperm head morphology may be related to abnormal DNA content.Semen from four mongrel dogs (five replicates per dog) were used to investigateDNA quality by means of the sperm chromatin structure assay (SCSA), and forcomputerized sperm morphometry (ASMA). Each sperm head was measured for nineprimary parameters: head area (A), head perimeter (P), head length (L), headwidth (W), acrosome area (%), midpiece width (w), midpiece area (a), distance(d) between the major axes of the head and midpiece, angle (θ) ofdivergence of the midpiece from the head axis; and four parameters of headshape: FUN1 (L/W), FUN2 (4π A/P2),FUN3 ((L – W)/(L + W)) and FUN 4 (πLW/4A). The data matrix consisted of 2361 observations, (morphometricanalysis on individual spermatozoa) and 63 815 observations for the DNAintegrity. The PCA analysis revealed five variables with Eigen values over 1,representing more than 79% of the cumulative variance. The morphometric datarevealed five sperm subpopulations, while the DNA data gave six subpopulationsof spermatozoa with different DNA integrity. Significant differences were foundin the percentage of spermatozoa falling in each cluster among dogs (p < 0.05). Linear regression modelsincluding sperm head shape factors 2, 3 and 4 predicted the amount of denaturedDNA within each individual spermatozoon (p< 0.001). We conclude that the ASMA analysis can be considered apowerful tool to improve the spermiogram.