I have now read the Current Biology article, and some other papers referred to therein, and it seems we can put Rosinbio's fears of the authors being mistaken to rest. Apart from close observations, and "high definition videography", there is also extensive and convincing evidence from stable-isotope analysis of the ants, the plants, the spiders, and other spiders in the area:
They continue by noting that 89 +/- 13.2% of the C and N comes directly from the plants, and 8 +/- 7.9% comes indirectly from the plant-eating ants, giving the diet of the spider a heavy vegetarian slant. This was established by methods unfamiliar to me, but which can be found in the supplementary material, which I have not yet read. They continue:
But note that Costa Rican populations had a larger proportion of animal prey in their diet.
There is no end of interesting stuff in this article, though. They note, further, that the range of this species "coincides with that of ant-acacia systems throughout Mesoamerica", and that they are almost exclusively found on these plants. They speculate that, as the ants keep most other insects away from the acacias -- including, presumably, ones that might feed on the spiders -- the spiders may have "achieved herbivory by exploiting plant goods exchanged for animal services", which sort of makes sense. They can't eat too many of the ants, because then the acacia-ant system would be disrupted, and predators on the spiders would be introduced into the system. But they still need to feed on something, and the Beltian bodies are available.
Th article refers to two other cases of omnivory among spiders, though in neither case are these as elaborate as in Bagheera kiplingi. I have not yet read these articles (references below), but apparently some juvenile orb-weavers eat pollen which collect on their nets (Smith & Mommsen, 1984), though Meehan et al. suggests that this is more or less an accident, as they ingest them only when recycling their nets. Jckson et al. reported that 31 of 90 studied species of jumping spiders supplemented their diet with nectar, and apparently (from the abstract) did so to get sucrose, rather than water.
In all, a very interesting field, and the presence and wide spread of nectarvory in the Salticidae could certainly make the transition to herbivory more plausible, if there were already species that supplemented their diet with non-animalian food.
---
Jackson, Pollard, Nelson, Edwards & Barrion (2001): Jumping spiders (Araneae: Salticidae) that feed on nectar. Journal of Zoology 255, 25-29.
Meehan, Olson, Reudink, Kyser, Curry (2009): Herbivory in a spider through exploitation of an ant-plant mutualism. Current Biology 19, R892-R893.
Smith & Mommsen (1984): Pollen Feeding in an orb-weaving spider. Science 226, 1330-1332.
Thanks for the information!