SUMMARY: Size of kidneys and 2nd lumbar vertebrae were calculated from radiographs and were measured directly at necropsy of 27 young adult male dogs of various sizes and unknown breeding. Direct measurement of kidney dimensions and weight at necropsy indicated that considerable variation occurred that was not due to differences in the sizes of the dogs. For estimation of kidney weight, kidney length was found to be superior to kidney area or to kidney length cubed. Kidney length was slightly better correlated with vertebra area than with vertebra length. However, the standard deviation for the ratio of kidney length to vertebra area was much greater than for the ratio of kidney length to vertebra length. The ratio of kidney length to vertebra length calculated from direct measurements at necropsy was less than the ratios obtained from measurements from radiographs, indicating that magnification of kidneys was greater than magnification of vertebrae. The relationship between magnification of kidneys and magnification of vertebrae was not constant; in some dogs, the ratio calculated from measurement of radiographs was nearly the same as the actual ratio, whereas in other dogs a considerable difference occurred. Comparison of 2 sets of ratios obtained from survey radiographs of the same dogs at different times suggested that the variation was more dependent on dog-to-dog variation than on radiographic technique. However, differences in ratios due to technique were also demonstrated in the study. Ratios calculated from intravenous nephrography were larger than those calculated from results of survey radiography and aortography. Survey radiographs rarely provided visualization of the right kidney adequate for measurements; adequate visualization of the left kidney was possible in only 26 of 52 radiographs. Visualization of both kidneys by intravenous nephrography was consistently adequate for measurements, making this technique the most feasible for determination of the dimensions of both kidneys. The numerous factors causing variation in ratios of kidney length to vertebra length determined from radiographs do not allow narrow numerical limits to be placed on this ratio for the normal dog. Although values for ratios of kidney length to vertebra length established in this study may serve as a general guide, the experimental dogs represented a group more homogenous with respect to age, sex, diet, and body condition than can be anticipated in a veterinary practice. Consequently, considerable deviation from mean values reported in this study should exist before kidney size is judged abnormal.