The application of microarray analysis to gene expression from limited tissue samples has not been very successful because of the poor signal qualityfrom the genes expressed at low levels. Here we discussed the use of catalyzed reporter deposition (CARD) technology to amplify signals from limited RNA samples on nylon membrane cDNA microarray. When the input RNA level was greater than 10 microg, the genes expressed at high levels did not amplify in proportion to those expressed at low levels. Compared to conventional colorimetric detection, the CARD method requires less than 10% of the total RNA used for amplification of signal displayed onto a nylon membrane cDNA microarray. Total RNA (5-10 microg), as one can extract from a limited amount of specimen, was determined to produce a linear correlation between the colorimetric detection and CARD methods. Beyond this range, it can cause a nonlinear amplification of highly expressed and low-abundance genes. These results suggest that when amplification is needed for any applications using the CARD method, including DNA microarray experiments, precaution has to be taken in the amount of RNA used to avoid skew amplification and thus misleading conclusions.