The EU Water Framework Directive requires assessment of the ecological quality of running waters using macroinvertebrates. One of the problems of obtaining representative samples of organisms from streams is the choice of sampling date, as the scores obtained from macroinvertebrate indices vary naturally between seasons, confounding the detection of anthropogenic environmental change. We investigated this problem in a 4th order calcareous stream in the western Carpathian Mountains of central Europe, the Stupavský potok brook. We divided our 100 m study site into two stretches and took two replicate samples every other month alternately from each stretch for a period of 1 year, sampling in the months of February, April, June, August, October and December. Multivariate analysis of the macroinvertebrate communities (PCA) clearly separated the samples into three groups: (1) April samples (2) June and August samples (3) October, December and February samples. Metric scores were classified into two groups those that were stable with respect to sampling month, and those that varied. Of the metrics whose values increase with amount of allochthonous organic material (ALPHA_MESO, hyporhithral, littoral, PASF, GSI new, DSI, CSI), the highest scores occurred in February, April, October and December, while for metrics whose values decrease with content of organic material (DSII, DIS, GFI D05, PORI, RETI, hypocrenal, metarhithral, RP, AKA, LITHAL, SHRED, HAI) the highest values occurred in February, April, June and December. We conclude that sampling twice a year, in early spring and late autumn, is appropriate for this type of metarhithral mountain stream. Sampling in summer is less reliable due to strong seasonal influences on many of the metrics examined while sampling in winter is inappropriate for logistical reasons.