Research Findings: This study reports initial descriptive and validity results of a new early childhood classroom observation measure, the Classroom Language Environment Observation Scales (CLEOS), designed to capture teachers' use of both implicit language supports (e.g., incidental scaffolding and shared reading) and more explicit language instruction (e.g., direct vocabulary instruction). Classrooms (n = 122) serving at-risk 3- to 5-year-old children and representing child care, Head Start, and public prekindergarten were observed; a subgroup was also observed with the Teacher Behavior Rating Scale (TBRS), a well-validated tool. Results indicated limited use of most language support strategies, particularly those that were more explicit. Concurrent validity for the CLEOS was supported via significant correlations with TBRS subscales. Greater use of higher quality linguistic input was significantly associated with teachers' years of experience but not with their education level. Findings supported the differential inclusion of linguistic input across settings, with large-group circle time being the most frequent setting for explicit instructional input and centers being the most frequent setting for incidental supports. Practice or Policy: Study results suggest a need to improve professional development and preservice training for preschool teachers related to supporting rich language interactions and explicit language and vocabulary instruction within classrooms. (author abstract)