I have started publishing papers this year - 3 so far, of which 2 are in an international journal, the 3rd in a local journal.
My question is, is there any particular disadvantage with publishing the majority of papers in one or a very limited array of journals? Or is it a case that it makes no difference either way?
I am sure opinions and traditions vary concerning this question. For me the quality of the journal (say, impact factor) is the most important, you try to get published in good journals. In some cases, the field may be so narrow that only one or a very few journals may be good. In such cases the selection may result in a single journal.
To publish in a variety of journals seems to be a means by itself for many but I think the reasons still vary. There may be advantages in getting published in many journals to show that the research has wider applications and is accepted more generally. It may be a way to avoid the suspicion that you have a back way into a single journal (not that that would be true). In some cases, you may select different journals because you know your ideas are not favoured by someone in a specific journal's editorial board (strategic reason).
So reasons for spreading the publications may vary quite substantially. The main point for me is, however, still that primarily the quality of the journal will decide. I would not chose a different journal just to get a new journal mentioned in my list of publications.
In my field (mathematics), the quality of journals is a quite fuzzy notion; we do not rely very much on metrics like the impact factor, but have a subjective and qualitative sense of prestige of journals. Inevitably, the way a given journal is regarded changes from one person to another. As a consequence, publishing in a variety of journals increases the odds that someone looking at your publication list will think "whao, she published in X!".
Another point is that if one publishes a large fraction of its research in one journal, people can wonder whether she has a friend in the editorial board that help her getting accepted there.
All in all, it seems preferable (and possible, given the large amount of journal in the field) to publish in a variety of journal for hiring, promotion and more generally evaluation matters. Of course, the picture is certainly different in other fields.
I agree with @PeterJansson that impact factor is certainly an important factor in selecting a journal. But I would also think about who you want to reach with your paper. Although Google Scholar and similar search engines now ensure a much higher probability of finding your paper based on its title, abstract and keywords, it will only be found by people actively searching for terms associated with your work.
Most researchers still subscribe to specific journals in their field, even if only through e-mail alerts, so I target a journal also for its audience (while trying to select the highest impact factor from this subset). This is especially important if your work bridges several fields.
For example, my research is about human-robot interaction, so I have to decide whether to publish an article in a journal that is read primarily by roboticists or social scientists/psychologists.
In this case, it is also vitally important to adjust the paper to its target demographic. Psychologists don't know (and likely don't care) about my implementation details while roboticists are more likely to be interested in nitty gritty details about the code and less about the nuances of the social psychology theory behind it.
