We tested the precision and accuracy of common measurements of snakes by repeated measurement of the heads and trunks of 10 preserved snakes and 10 live snakes by two groups of five people over 10-wk periods. The measurements produced values with variances and ranges related to the nature of the variable, the measurer, and the snake, but accuracy could not be determined. Reporting sizes of snakes to high levels of accuracy is therefore unwarranted. Measurements of head variables on preserved and anesthetized live snakes had similar levels of variance that approximate half the variance of the same measures on live, unanesthetized snakes. Conversely, measurements of snout–vent length (SVL) on both preserved and unanesthetized live snakes had about twice the variance of the same measures made on anesthetized snakes. Measurers differed for all measurements of preserved snakes and for all head measurements of live, unanesthetized snakes, more experienced measurers generally yielding higher precision. Conversely, measurers did not differ for most measures of anesthetized snakes. Our data support suggestions that the most repeatable measures of SVL are made on anesthetized snakes. Lengths of the head and lower jaw can be measured with relative precision on a snake in any condition. Head width and supralabial length have both inter- and intrameasurer variances high enough to make them unreliable measures of head size. We conclude that features of live snakes most commonly measured vary because they have no exact size. We therefore suggest a new convention for reporting sizes of snakes.