This paper presents the results of an investigation of voicing in utterance-initial and intervocalic stops in monolingual Russian speakers. Prevoicing was found in over 97% of the lenis stops; over 97% of the intervocalic stops were fully voiced. Utterance-initial fortis stops were pronounced as voiceless unaspirated and had short positive VOT. Intervocalic fortis stops were completely voiceless except for a short voicing tail into closure. These results are relevant for typological studies of voicing. Some studies of languages with a two-way contrast between initial stops with prevoicing and short lag VOT have reported that prevoicing is less robust than what might be expected. These findings have been attributed to influence from another language without prevoicing. Our results with monolingual speakers of Russian support these claims. Our results are also relevant for the debate about the laryngeal feature in aspirating languages, which often have some voicing of intervocalic lenis stops. Such voicing has been attributed to passive voicing, in contrast with active voicing that occurs in true voice languages such as Russian. We found that the voicing in Russian is much more robust than the intervocalic voicing in aspirating languages. This difference is explained if the features of contrast are different in the two types of languages: [voice] in the case of Russian and [spread glottis] in the case of aspirating languages.