The authors evaluated the accuracy of EMG needle placement using the methods of two common EMG reference atlases. Fourteen cadavers were used to conduct needle placement by two board-certified electromyographers with a blinded anatomist completing the needle dissection. The two methods were found to be similarly accurate. Certain muscles were particularly difficult to sample by both methods. Certain muscles were uniquely difficult to sample by one or the other method. Many inaccuracies were not significant in that needle placement occurred in muscles of similar root and nerve innervation. Generally, lower extremity, superficial, and larger muscles were easier to sample than smaller or deeper muscles. Typical diagnostic difficulties are discussed with means to improve needle sampling accuracy with present technology and future study. Muscle needle placement, as would be conducted for electromyography, shows differential accuracy in needle placement depending on muscle location, size, depth, and local anatomy. Understanding these limitations allows the electromyographer to recognize patterns of inaccuracy and make corrections for a more accurate diagnosis.