Studies of tremor in the 19th century were based initially on simple observation and later on the use of crude graphic recording devices that had been modified from instruments developed for other purposes. Like several European contemporaries, American investigators studying tremor used and adapted various existing instruments, including tambours and sphygmographs. A tambour used a drum-shaped pneumatic mechanism to transmit movements to a recording instrument, whereas the sphygmograph was a nonpneumatic mechanical device initially used to record the pulse. 19th-century American neurologists who used such devices included Frederick Peterson, Hobart Amory Hare, Charles Loomis Dana, and Augustus A. Eshner. Their measurements of tremor frequency were generally consistent with modern estimates for various types of tremor. Eshner, in particular, was frustrated by the overlap of frequency domains for tremors in different diseases, because this precluded use of tremor frequency alone as a differentiating feature for diagnosis. Peterson and Dana recognized the variation in tremor frequency in different body parts resulting from different natural resonance frequencies of these parts as a function of weight and elastic properties. Peterson, Dana, and Eshner also recognized that tremor amplitude and frequency are inversely related. Finally, these 19th-century investigators recognized that the tremor of Parkinson's disease is a relatively low-frequency rest tremor, suppressed by action, and generally synchronous in symmetric body parts, but varying in amplitude and frequency in different body parts or over time.