"Our objective was the analysis of the process of social and environmental adaptation of sojourners to two isolated oceanic sites far from the Northeastern coast of Brazil. Both the Biological Reserve of Rocas Atoll and the Saint Peter and Saint Paul Archipelago (an Environmental Protection Area) offer harsh conditions for human survival: no fresh water, humid tropical heat, sunny outdoors, remote oceanic location with very limited help for any sort of emergency, among others. A four people simple wood shelter (~ 45 m2) was built in the Atoll and a scientific station in the Archipelago to offer as much comfort as possible under design principles based on similar buildings used in the Brazilian Antarctic Station, adapted to tropical oceanic conditions. The buildings were planned as refuges amidst the mildly hostile conditions for humans, and had to comply with the demands of minimal ecological interference with the many species of oceanic life, depicting a reassuring silhouette within the otherwise desolate landscape. For periods of 20-30 days teams of 4 persons are sent to those places by boat, being replaced by their substitutes at the end of that time. They are scientific researchers, graduate and post-graduate students typically of areas such as Marine Biology, Oceanography and the like. Our preliminary efforts of studying the adaptation of the sojourners to those places had to deal with the lack of literature about that type of isolated and confined environment, distinct from spacecraft simulators, polar stations or nuclear submarines. After overcoming effects of the social inflation of risk in our own work, we have been able to recognise: 1) the importance of the social atmosphere of the group of sojourners, and not only of the physical environment by itself; 2) the need of knowing specific determinants of each team context, in personal, professional and social terms; 3) the operational difficulties to interview all participants; 4) the limitations of a questionnaire as data collecting instrument employed in the first stages of the project; and 5) our expectations about information that could express visitors' life as it was lived in those places, and not as remembered/reported in later interviews performed after their return to the continent. Video tape recording and/or live interviews through telephone or internet would have allowed for interesting possibilities of data collection in those circumstances. Material and financial constraints, however, led us to employ personal diaries, a paper-and-pencil simple yet insightful alternative. The diary kit personally supplied by our research team to the groups of travellers hours before their departure on the boat trip contained: a 15x25 cm notebook specifically prepared with project identification and instructions about the type of information sought for; 12 colour pencils, a ball pen, a pencil-with-eraser, and a return self-addressed envelope; everything packed within a plastic re-sealable bag. Some sojourners were reluctant about registering information on their diaries; we received several back entirely blank, others with just a few lines written. During the post-trip interviewing, some of these told us about their self-censoring, for "not knowing what to write about", or for fear of furtive glances by group members. Others adopted a colloquial writing style, producing detailed texts, typically more about their peers' conduct than themselves, with more references to behavioural consequences than to behavior, with few accounts of feelings or emotions explicitly expressed. Also mentioned were personal transformations felt as consequence of experiences lived there, and special events leading to accounts such as "I don't like to write, but today something happened that you guys ought to know". Drawings and poetry were not rare and some such instances provided valuable data difficult to be conveyed through verbal report. This illustrates the unstructured nature of the technique, which allows for an open participation of the researched and low control of the researcher. An analogy seems appropriate here. When evaluating environments, researchers usually start with less structured techniques, such as trace measures or behavioural mapping, progressing later on towards more sophisticated instruments, planned in accordance with results provided by the early ones. Such heuristic value also holds true for our experience with data from the diaries. The content analysis of personal journals entries has provided important clues for later stages of our work. Although in general referring to professional or scientific objectives of people's mission there, such data included accounts on administrative and ideological profile of the two organisations managing the areas, functional virtues/deficiencies of the shelter and the scientific station, hard/good times in work or in relation to group members, group dynamics and social climate, social and ecological adequacy of personal manners, maintenance activities and domestic tasks division, preservation vs. conservation, territoriality and space appropriation and many other relevant themes. AcknowledgementThe author acknowledges the partial support from Conselho Nacional de Desenvolvimento Científico e Tecnológico - CNPq (Proc. 350501/2000-9)"