A new approach is described to screen for protein nanocrystals based on the reversibility of crystallization. Methods to characterize nanocrystals are in strong need to facilitate sample preparation for serial femtosecond X-ray nanocrystallography (SFX). SFX enables protein structure determination by collecting X-ray diffraction from nano- and microcrystals using a free electron laser. This technique is especially valuable for challenging proteins as for example membrane proteins and is in general a powerful method to overcome the radiation damage problem and to perform time-resolved structure analysis. Nanocrystal growth cannot be monitored with common methods used in protein crystallography, as the resolution of bright field microscopy is not sufficient. A high-performance method to screen for nanocrystals is second order nonlinear imaging of chiral crystals (SONICC). However, the high cost prevents its use in every laboratory, and some protein nanocrystals may be "invisible" to SONICC. In this work using a crystallization robot and a common imaging system precipitation comprised of nanocrystals and precipitation caused by aggregated protein can be distinguished.