In order to evaluate the morphometric differences of five populations of Mikrogeophagus ramirezi (Pisces: Cichlidae) from the lowlands of the Venezuelan Orinoquia,14 homologous landmarks from 83 specimens were taken and transformed to Procrustes variates. A MANOVA/CVA test was performed, however, this test did not perform well due to the low number of specimens used, and it was decided to increase the sample size. The probabilistic normal distribution for the variables was determined, and variances of these variables were similar. Subsequently, 100 values for each variable were generated, from the model Y= X + ε, being Y the variable value to estimate, X is the real mean value of that variable, and ? is the product between real standard deviate and random values of a normal distribution. This simulation was made for three types of standard deviates: the real variables SD, the prorated values from real variables, and mean bootstrap prorated values. The data generated values were validated with the real data, and some inner properties, both before and after applying the MANOVA/CVA test. In all tests, the values generated were no different from the real ones. Using the generated values as a valid surrogate, it was statistically determined that five populations were different morphometrically, and candidate species. On other hand, a regression between real data coordinates was performed, and compared with the generated data. Whilst the real data coordinates arranged on an almost perfect straight line, and were highly correlated, the generated data ones were dispersedly arranged. This fact, did permit infers that this line, the Stasis Centroid Line, is a phylogenetic signal emerging from the real data.

KEY WORDS: MANOVA/CVA; Mikrogeophagus ramirezi; Monte Carlo Techniques; Procrustes coordinates; Stasis centroid line.