At least one-third of the 100,000 genes in the mammalian genome are expressed only in the nervous system, but the diversity of types of neurons and the functional significance of this diversity is unknown. Given that much of what we know of the cell and molecular biology of neurons has been learned at the level of hundreds of thousands of cells, one essential tool for dissecting the diversity of neuronal types, and for understanding their function, is that of gene-expression profiling at the single-cell level. Therefore, we have used PCR-based methods for amplifying picogram quantities of complex RNA populations to synthesize and amplify cDNA from single, defined neurons from the mouse retina. To analyse these cDNA populations, we have used 1,100-element microarrays containing 130 known genes and 960 random clones from an adult mouse retina cDNA library. To test the ability of this system to distinguish between cell types and identify cell-specific genes, we are carrying out two types of experiments. To fingerprint cell types, we are comparing cDNA from each cell to a single reference sample of total brain cDNA and clustering cell types based on gene expression relative to entire brain. To directly isolate cell-type specific genes, we are also carrying out pairwise comparisons of gene expression between each of the cell types and identifying genes whose expression is restricted to each cell type. Our studies demonstrate the feasibility of carrying out gene expression profiling experiments at the level of single cells in the nervous system and that such methods can be used to distinguish cell types.