As part of the accreditation procedures, External Quality Control (EQC) must be performed for all biological determinations. In the exploration of male infertility, the spermocytogram is very important because it is often used as first line and an error of interpretation may have dramatic consequences. Alongside the EQC which usely consists of carrying out preparation slides (stained or not), we tested the use of a slide scanned from a stained specimen ("virtual slide"). All participants (n = 57) received a sample of the following supports: an unstained slide, a stained one and a virtual slide, all of them from a single human ejaculate. The required tests were the proportion of typical forms of spermatozoa and the degree of teratozoospermia using the Multiple Abnomalities Index (MAI) according to David's criteria. Results showed that for the two examinations, the dispersion of results remains similar regardless of the support. Furthermore, results seemed no to be influenced by the staining technique. This indicates that the discrepancy between results came from the quality of the observer. Moreover, the numerical values of half the participants were situated in the interval mean of ± 30% for the evaluation of typical forms and ± 15% for MAI. We recommend the virtual slide for EQC spermocytogram to evaluate and improve the reading ability. In addition, we propose to retain an interval of acceptability of ± 30% for the evaluation of typical forms and ± 15% for MAI.