Practical work is a characteristic element in science teaching at all levels. Intuitively, it appears justified to most science teachers, because observation and manipulation of isolated parts of "the physical world" is part of the very nature of science. The reasons for having practical work in school science teaching have been discussed recently . European university teacher's objectives for labwork have also been studied. Whatever the reasons, at tertiary level it is essential both to learn science, to learn about science, and to do science (a useful way to think about three perspectives in science teaching ). To include practical work or not in a modern curriculum leading to a science degree is a matter beyond any reasonable discussion. Since practical work, however, is not a single, well-defined category it looks differently and plays different roles in different science disciplines. Many universities use curricula with "practicals" or "exercises" dominating in the earlier parts, while "experimental work" or "investigations" may be included in thesis-work. In new science curricula problem-oriented project work may be introduced from day one, and experimental work will be included when relevant. The Natural Sciences Basic Programme (NSBP) of Roskilde University includes such projects occupying half of the study load. This two years programme serves as the first part of a number of degree-programmes.

From a chemistry point of view it is obviously important that the two first years (the NSBP) include introductory chemistry and give qualifications relevant to chemistry. Among these is genuine experience with experimentation being essential to chemistry and so important to most chemistry jobs. It is believed that the general experience with experimentation obtained through the students' research-like project work (on problems involving any natural science subject) actually gives qualifications of importance in chemistry or any other subject with a characteristic experimental dimension. But is it true? To address this general question is therefor relevant for the evaluation of the NSBP as the first part of a chemistry degree.

The first issue is to describe the characteristic elements of experimental work. The actual formulation could be different and more or less detailed^,. It appears that in the first place a moderate degree of specification is more useful, when experimental work from very different parts of science is to be described and compared. The next problem is to assess to which extent the students get adequate experience with experimentation (experimental investigation) during their projects. One method may involve the knowledge of what they actually did. A source for such information is the project reports. Such reports may reveal whether the students during the project have experienced a specific element of experimental work. A question is, of course, if the reports themselves give sufficient information to assess this. If, however, the foregoing questions can be answered acceptably, the top question could be elucidated: does experimental work of any science tradition contain the same elements as chemical experimentation or to which extent is this true?

The present work deals with one of the steps in this chain of possible investigations: the analysis of the experimental work of the students' projects as brought to light by their reports. 25 second-term project reports produced by the latest total population of first year students were preliminarily studied, using the five fairly general categories of most scientific experimental work (Objectives, Design, Experimental, Results and Evaluation) as a template. The analysis focuses on which of the five categories that seemed to have been subject to intellectual challenge to the students and which were mostly due to the supervising teacher and taken as granted by the students and not discussed further. All project reports reflected the students' engagement in some practical work, and some elements of experimental work were represented in all the reports. The students' arguments in the reports for performing the experiments seemed to be reasonable in almost all cases. Design and choice of methods seemed to be one of the weaker parts of several projects, and not all reports witnessed enough attention on standardisation, controls etc. Most of the projects were "single-subject" projects, where the problem studied matched the expertise of the supervising teacher. The "chemistry projects" seemed in the first place not to reveal better experimental training in the above sense than the average. Further details of the investigation will be

Key words: experimental work , project, freshman, labwork