One major obstacle to progress in neuroscience is the lack of knowledge about the neuron classes that make up the CNS. How many discrete cell types exist, and what is the "morphospace" of each of these types? In order to tackle this problem, we propose a combined genetic and anatomical approach to neuron classification at a large scale using larval zebrafish (Danio rerio). Zebrafish have ideal properties for this approach since they are translucent and easily amenable to genetic modification. Their brains are relatively small (ca. 90,000 neurons at 6 days post-fertilization) but display the basic architecture that is conserved across vertebrates. The small size permits this approach to reach full brain coverage in the near future. We are using a genetic tool that stochastically labels individual neurons with membrane-targeted fluorophores. Optical access and visualization are further facilitated by prior tissue clearing with a streamlined Clarity protocol. Labeled neurons are imaged by confocal microscopy at high resolution and reconstructed for quantitative morphological analysis. Because the method is based on the Gal4/UAS system, it will be possible, in the future, to target defined neuronal populations in order to combine genetic identity with morphological analysis. A variation of this tool tags pre-synaptic sites with a second, genetically encoded fluorophore, which gives information about the direction of information flow between neighboring cells and between interconnected brain regions. Currently, we have reconstructed >1,000 neurons from a pan-neuronal expression pattern and recorded their morphologies. Future efforts in our laboratory will be directed at automating image acquisition and establishing an image-processing pipeline, which will greatly accelerate our throughput. In order to compare neurons between individual fish, we are aligning them to a high-resolution, age-matched standard brain. This reference brain will consist of several different markers giving other researchers the flexibility of using a reference stain that best suits their experimental paradigm. Furthermore, combining all these markers into one standard brain will give us further insight into the functional modules that make up the larval zebrafish brain. This new database of zebrafish neurons will initially be populated by their morphological properties, anatomical location and transmitter type, but can be expanded to include functional, developmental and gene expression data. With support from a local supercomputing center, it is planned to make the database a web-accessible resource for the community and link it to ZFIN. This will give zebrafish neuroscientists the ability to browse the stored information, upload data from their own experiments and carry out similarity searches, analogous to BLAST or PDB.