2001ApJ...560..719L

We present a study of 35 radio-selected supernova remnant (SNR) candidates in the nearby spiral galaxy NGC 6946. In this study, an optically selected sample of 27 SNRs from Matonick & Fesen is compared to our sample of radio-selected SNRs. The optically selected SNRs are found to have almost no overlap with the radio-selected sample. This dichotomy is further enhanced by the observation that the optically selected SNRs favor the interarm regions, while the radio-emitting SNRs lie predominately on the spiral arms in or near regions of high star formation. The separation of the two samples of SNRs is discussed in terms of selection effects and differences in cosmic-ray production. The optical sample of SNRs is probably biased toward those SNRs located in areas of NGC 6946 where the confusion of Hα emission by H II regions is relatively low and the SNRs are easier to identify by their emission-line signatures. The radio-selected sample is also subject to selection biases such that these SNRs favor the arms. However, the absence of radio emission from the optically selected, largely interarm SNRs and the relatively large number of radio-detected SNRs in the arms require additional explanation. The properties of the radio-selected SNRs are discussed in the context of diffusive shock acceleration theory. We find that the theory can account for the range in radio flux densities and the nondetection of the optically selected SNRs. The differences in the radio properties between the arm and interarm population of SNRs can be explained by differences in the average gas densities and magnetic fields since the latter affect both the cosmic-ray yield and the radio properties of the SNRs. The possibility that the arm and interarm SNRs arise from different stellar populations is also addressed.