Wolbachia pipientis is a rather peculiar bacterium. It's an intracellular organism, and is found in a wide variety of tax, including nematodes, crustacea, and arachnids. About 20% of insect species are thought to have Wolbachia. Wolbachia has evolved a number of mechanisms to ensure transmission, which is generally maternal. There are a number of consequences of infection with these intracellular bacteria observed in different species - including reproductive isolation between infected and uninfected strains of the mosquito Culex pipiens. I also recall seeing a old paper (though I've long-since lost the citation) which claimes a growth advantage of infected Drosophila larvae in crowded culture, though I think this must have predated the identification of Wolbachia.
So some laboratory Drosophila strains are infected with Wolbachia, while some are not. This can lead to artefacts in many experiments. For example, in my own lab, we had some perplexing results concerning the effects of particular transgene on viability of some combinations - the involvement of Wolbachia became clear firstly when we noticed the effects depended on which which strain was paternal and which maternal in the cross, and finally the observation that this effect vanished after treating the transgenic stocks with antibiotic. These kind of artefacts are particularly important to consider when working with complex characteristics such as lifespan (as we do in my lab).
An interesting question is why these bacteria persist in insect populations - is there a positive selection? In this brief paper published in Science, Hedges et al show that Wolbachia infection does influence one important factor in Drosophila: resistance to viral infection.
Viral infections aren't something I think the average Drosophila researcher worries about - we're more likely to worry about infestation with mites, or bacterial infection on the culture medium! Much of the research on Drosophila immune systems concerns bacterial or fungal infection. Nonetheles, there are viruses out there that do infect Drosophila, and some with quite odd characteristics, such as sigma virus, which I believe makes flies die on exposure to high levels of carbon dioxide (a problem since nowadays we use carbon dioxide instead of ether to anaesthetise our flies!). In this paper, Hedges et al look at outcomes of infection with three viruses - Drosophila C virus (DCV), cricket paralysis virus (CrPV) and Flock House virus (FHV). The first two are natural pathogens: all three are lethal to Drosophila when inoculated by injection.
This is a brief paper, with a single figure (there is further supplementary material at the Science website). Essentially, what they show is that Wolbachia-infected flies survive for longer following injection with DCV than do uninfected flies. This experiment was done with two Drosophila strains: Oregon RC (a standard laboratory wild type strain - figure 1A) and w1118(a standard white eyed mutant strain - figure 1B), and interestingly the experiment used "cured" flies as the non-Wolbachia-infected control flies. this point is important, as it means that genetically, the two strains are the same. The same effect can be seen with CrPV (figure 1C), and even more markedly with FHV (figure 1D) infections. [Click on figure for larger version -may require subscription]
So, does this research suggest a positive selection for Wolbachia infection? Quite possibly. I don't know very much about natural viral infection in Drosophila, but I suppose that the injection oute of infection used experimentally might be rather more severe than a naturally acquired infection, and while the Wolbachia infected flies do still die in this experiment, perhaps Wolbachia infection leads to significantly greater survival of natural viral infection.