The In x Ga 1− x N alloy system is used as an example to describe achievements and limitations that are given by noise levels in lattice images from transmission electron microscopes. Unlike many other experimental techniques, noise limits must be determined afresh from image to image since they vary with microscope performance, sample preparation, radiation damage if present, and operator skills. Both, the determination of the indium distribution in In x Ga 1− x N alloys by strain mapping and the detection of indium clusters in InN by imaging, are largely affected by noise limitations. As a result, it is challenging to probe for the distribution of indium atoms in alloys with x<0.2 or x>0.8, which is of importance if one aims at understanding the nucleation of indium atom clusters in GaN or InN. We present results that point towards cluster formation in quantum wells with x<0.2 and show that photoluminescence at 0.5–0.7 eV relate to the presence of indium clusters in InN. A reliable detection of single indium atoms, however, will require further improvement of detection limits, which one can expect from the next generation of electron microscopes that are developed in DoE's TEAM Project.