This paper describes a four-year project undertaken to develop a set of performance tasks that could be used for assessing hands-on science in Irish primary schools. The project has evolved over five phases. In phase one, science assessments used in a variety of educational systems in Australia, Canada, New Zealand, the UK and the US were located and catalogued. In phase two, approximately 170 performance tasks were selected and adapted by the project team to suit the requirements of the Irish primary science curriculum. In phase three, a purposive convenience sample of 183 teachers evaluated the extent to which the tasks (a subset of 68) were suitable for use in Irish primary schools. Specifically, with respect to each task, teachers were asked to comment on and indicate their level of agreement with the following statements: The concept being assessed is stated clearly; The task focus is clear; The description of the prior knowledge required is clear; The skills being assessed are clearly identified; Relevant information about the task is provided for the teacher; Directions for how to do the task are clear; The pupil worksheet is clear and relevant. Overall teachers indicated that they were positively disposed to what the tasks were intended to achieve. A good deal of their feedback was extremely useful in amending and improving individual tasks and in improving the layout of tasks in general. In phase four, the researchers observed 11 different tasks being implemented in classes from infants to sixth. The 11 teachers involved were interviewed about their experiences immediately afterwards. Overall, teachers in this phase of the study like their counterparts who were surveyed were overwhelmingly positive about the tasks. They acknowledged their potential for helping them to integrate planning, teaching and assessment in science and in teaching science better. They liked them because they provide "something concrete to go on", to identify what "you're looking for before you prepare work" and to identify "who is getting it or not." They also indicated that they felt the tasks focused attention individual pupil learning and needs. Indeed, a number of them acknowledged that they were surprised at times about what they discovered about their pupils. Observational data indicated that the pupils enjoyed doing the tasks and that the tasks encouraged active involvement of pupils and teachers in the assessment process. However, observational and interview data also indicated that the teachers found the scoring guide difficult to use. A number of them acknowledged that they were not used to working in this way and were not confident with them. They were often unclear about whether they should be focusing on individuals, groups or the class as a whole. Another source of uncertainty arose from the lack of clarity about whether to use the rubric holistically or analytically i.e. to focus on a set of skills or individual skills. When asked to reflect on the challenges the task would pose for teachers in general, it was not surprising to hear them reflect on issues about having the time to undertake assessment with pupils, about the difficulty of recording assessment information and the need for "a lot more practice in doing that kind of work". Finally, one teacher's observation that "some teachers will need to adapt the tasks for different class situations" was insightful in so far as it pinpoints one of the challenges encountered in the experience of performance assessment in other systems. The fifth phase of the project, due to be completed in 2005, will involve dissemination to teachers via the Web and book/CD-ROM. Planning with respect to other elements of the project such as professional development courses and exemplars of performance has yet to be finalised.