Small-animal PET is becoming an increasingly common tool to explore biological function noninvasively, particularly in the neurosciences. However, it is a relatively new phenomenon with its own special challenges. Although small-animal PET scans suffer less from scatter and attenuation than their human counterparts due to the smaller subject size, background rates can be significant. And while the spatial resolution of ∼2 mm was much better than for human scanners, the improvement did not scale with the size of the regions of interest in rats and mice, making the partial volume effect relatively more problematic. The first commercial systems were the micro-PET scanners from Concorde Microsystems (Knoxville, TN), which appeared around the year 2000. Concorde (now part of Siemens) produced the rodent micro-PET systems R4 (Knoess et al., 2003) and Focus 120 (Laforest et al., 2004), and the primate systems P4 and Focus 220 (Tai et al., 2001). The micro-PET R4 scanner was the first system and has perhaps the largest installed base and thus will serve as a specific context for the issues discussed later. Recently several other commercial small-animal PET systems have been offered, including the HIDAC (Oxford Positron Systems, Oxfordshire, UK) (Jeavons et al., 1999; Schafers et al., 2005), MOSAIC (Philips Medical Systems, Philadelphia, PA) (Surti et al., 2003), X-PET (Gamma Medica, Northridge, CA), and YAP-PET (ISE, Pisa, Italy) (Del Guerra et al., 1998). Many of the concepts discussed herein can be applied to these scanners as well. The initial micro-PET systems were robust in terms of hardware, but the processing software had few of the quantitative corrections that researchers had come to expect of state-of-the-art human systems. In the last few years the software has improved significantly, but careful attention to quantitative issues remains important to maximize its power to resolve subtle biological effects in small regions.