Once upon a time there was a Journal of Irreproducable Results, which published science humor. It started in 1955, says Wikipedia, and saw a schism in 1995 that led to the founding of the Annals of Improbable Research. AIR is noted for its awarding of the prestigous Ig Nobel awards, rewarding achievements that "first make people laugh, and then make them think".
Now the august journal Nature is getting into the fray, with an entire special issue on Challenges in Irreproducable Research. They describe it thusly:
No research paper can ever be considered to be the final word, and the replication and corroboration of research results is key to the scientific process. In studying complex entities, especially animals and human beings, the complexity of the system and of the techniques can all too easily lead to results that seem robust in the lab, and valid to editors and referees of journals, but which do not stand the test of further studies. Nature has published a series of articles about the worrying extent to which research results have been found wanting in this respect.
For more sources of irreproducable conclusions, a very good source is the Retraction Watch weblog which tracks scientific fraud through watching for retractions of published papers. Now not every irreproducable result is the result of fraud - sloppy mistakes and bad data handling are in part to blame, as are difficult to replicate experimental conditions. Retraction Watch also covers garden-variety scientific fraud like plagiarism and the sin of submitting a single paper to more than one journal.