Mata and Rosengart1 believe that interobserver variability in interpreting radiographs may play a part in the disparate incidence of necrotizing enterocolitis (NEC) reported from various centers. Underlying their contention is the implication that the diagnosis of NEC rests upon abnormal radiographic findings. I do not dispute their observations, but let us examine the underlying premise. Leonidas et al2 and Daneman et al3 analyzed the radiographic signs of NEC. Both groups found bowel dilation to correlate well with severity of the disease, but dilation is a highly subjective sign and is entirely nonspecific.4