The effect of the strict classification of spermatozoal morphology on the outcome of in-vitro fertilization and embryo transfer treatment cycles has been assessed in a retrospective analysis of 2144 consecutive cycles. The patients all had a standardized long protocol gonadotrophin-releasing hormone analogue cycle, with luteal phase start, to achieve down-regulation. All treatment cycles where the sperm density was abnormal (< 20 x 10(6)/ml), or where progressive motility was abnormal (< 40%), were excluded. The study excluded treatment cycles where the oocytes inseminated did not include at least one grade 1 or grade 2 oocyte. The percentage of couples achieving the normalized results of the clinic, including median fertilization rate per patient, insemination rates, numbers of embryos transferred, rates of 'spare embryo' blastocyst formation, cumulative pregnancy rates and pregnancy outcome, were calculated. No statistically significant difference arose between the two groups of patients with regard to the percentage of patients achieving the normalized median fertilization rate or higher (group 1 with > 14% normal forms, and group 2 with > or = 4%, < or = 14% normal forms). There was a statistically significantly lower chance of achieving this rate in patients of group 3 (< 4% normal forms) (P < 0.005), but 68.6% did achieve that fertilization rate or higher. There was no statistically significant difference in any of the other end points. In conclusion, a morphological classification may be appropriate as an indicator for counselling patients with regard to treatment expectations, but its use would be seem inappropriate as an index of fertilizing potential in clinical management.(ABSTRACT TRUNCATED AT 250 WORDS)