Siddle's (1979) assertion that the significance hypothesis is based on tautological argument regarding the effects of intensity decrease refer only to two papers which first suggested the need for research into determinants other than mismatch. Studies testing specific predictions followed in various laboratories, and it is their results that support the hypothesis. The hypothesis never depended on intensity decrease, and the charge of nonreplicability is incorrect. Siddle's argument concerning response stereotypy cannot account for the data showing response to significant 'novelty' coupled with nonresponse to nonsignificant 'novelty' in the same OR component, or for that showing consistent differences in various components between randomly selected groups. Arguments based on trait-lability are inadequate since tests of the lability hypothesis have confounded trait factors with the differences in stimulus assessment they hope to test. The significance hypothesis emphasizes stimulus information in the OR, significance serving to limit access to central processing channels to those signals whose information is relevant for the organism. OR components are seen as long-latency responses elicited after automatic early detection of uncertainty and significance, and reflect higher processing of significant information. A distinction between involuntary and voluntary phases of attention is acknowledged, but it is questioned whether OR components can be assigned to the former because of the frequent failure of 'automatic' ORs to innocuous stimuli, the long latency of most components, and failure to demonstrate structural differences between the classes of response.