Dark origins: the mysterious evolution of blood-feeding in bats

Continuing the vampire theme, I here want to discuss another of those really, really interesting things about vampire bats: namely, how did their blood-feeding behaviour evolve in the first place?

First off, a big thank you to everyone who's been visiting, commenting and generally saying nice things. The Speculative Zoology post in particular generated a lot of recent interest, and as of yesterday Tetrapod Zoology was the third most-hit site in the scienceblogs community (though still way way behind Pharyngula of course). For some reason I'm unable to post new comments on the Speculative Zoology article which is annoying as I wanted to respond to Sordes' comments on the life appearance of the earliest whales. Oh well. You'll note that I finally worked out how to add banners and a profile photo to the site. I aim to keep changing banners as and when I feel like it.

Despite requests, I regret that I will not be producing a post on Chthulhu, sorry Steve, as I am not convinced that this and other Lovecraftian entities are of tetrapod affinities. Funnily enough though, I did have a run-in with one tentacled beast recently: while wandering around the Southampton docks (really) I discovered, there on the side of the road, a dead little baby cuttlefish. This was several hundred metres away from the water, so either cuttlefish have evolved teleportation, jet propulsion is far more effective than generally thought or.. well, I'm fresh out of ideas! (just kidding, it was surely dropped by a gull).

Enough nonsense, on with the bats... In previous posts we've looked quickly at the biology and behaviour of vampire bats, at the vampire fossil record, and at the interesting fact that the extinct vampire species were somewhat more diverse in terms of ecology and behaviour than the three living species. As covered here, there is also the fascinating possibility that some of the supposedly extinct species inspired certain mythological stories, or might linger on and await discovery, though the evidence for this is presently circumstantial. While vampire bats have a reasonably good fossil record extending back to the Pliocene, we unfortunately lack fossils of a proto-vampire: of an ancestral species that, judging from its morphology, was just beginning to make the first forays into the bizarrely specialised lifestyle of the obligate sanguivore*. Nor are there any non-sanguivorous bats that possess a vampire-like dentition, but use their teeth for another function. These shortcomings are particularly frustrating because we really want to know how sanguivory evolved. As it happens this is a much-discussed question within the bat literature, and many bat experts have had a stab at proposing an answer. And that wasn't a pun, because vampire bats don't stab, they slice [image from here].

* Note that the habit of blood-feeding is termed sanguivory, and not 'sanguinivory' as you might expect.

To date five different theories have been proposed to explain the origins of sanguivory, though some aren't well known as they were published in relatively obscure sources. Reviews of this issue are provided by Fenton (1992) and Wetterer et al. (2000).

-- Hypothesis 1. Vampires evolved from frugivores (fruit eaters) that evolved enlarged, sharp teeth to pierce fruit rind. These teeth then became an exaptation: a structure that later proved useful for another purpose; namely, the piercing of animal skin in order to feed on blood. This hypothesis rests on the assumption that vampire ancestors were part-time or dedicated frugivores.

-- Hypothesis 2. Vampires evolved from bats that fed on the ectoparasites of large animals, gradually shifting to blood-feeding and eventually to dedicated sanguivory. The assumption here is that vampire ancestors were insectivores, and highly specialised ones at that.

-- Hypothesis 3. Vampires evolved from insectivorous bats that started to take advantage of the insects that collected around wounds on the bodies of large animals. Over time the bats started to not only feed on the congregating insects, but on the blood and tissue of the animal that was causing the insects to congregate. Eventually the full transition to sanguivory occurred. Again, this hypothesis rests on the assumption that vampire ancestors were insectivores.

-- Hypothesis 4. Vampires evolved from carnivorous bats that preyed on small arboreal tetrapods. The predatory habits and specialised teeth of these bats were later exapted for use in biting large animals. This hypothesis is the most vague of the five. It requires the assumption that vampires evolved from carnivorous bats, similar to the vampyrines (which are so named as they include Vampyrum, the false vampire: image at left from here).

-- Hypothesis 5. Vampires evolved from arboreal omnivorous ancestors that learnt to take chunks of flesh from wounds on large animals. Over time they became increasingly specialised for sanguivory, eventually losing the flesh-eating habit. Again, this hypothesis assumes that vampire ancestors were part-time or dedicated carnivores.

The first way to find support for either one of these hypotheses is to look at diet records for vampires. Are they really sanguivores 100% of the time, or might they sometimes eat other foods? If they do then we might get a hint as to their ancestral diet. The good news is that some vampire specimens have been discovered with other objects in their stomachs, including insects, fruit and pieces of flesh. The bad news is that these objects were either probably swallowed accidentally (this goes for the insects and flesh fragments), or are so rare (fruit have only ever been reported in one study) that their occurrence is regarded as accidental. Arguably, the diet of living vampires therefore tells us nothing about the evolution of sanguivory.

A huge amount of morphological and genetic data shows that vampires, the desmodontines, are part of the New World leaf-nosed bat family Phyllostomidae. In the absence of fossils that appear intermediate between vampires and other phyllostomids, and of informative dietary data from living vampire bats, the next best thing we can do is look at those phyllostomids closest to vampires, in order to see if they might give us any clues to the origins of sanguivory. Phyllostomids are one of the most diverse bat clades of them all (there are over 140 species), and include insectivores, carnivores, frugivores, folivores and nectarivores, as well as many species that combine some or even most of these dietary habits. It would be really informative in view of the hypotheses listed above if we found that, for example, vampires belonged to a mostly carnivorous phyllostomid clade, or to a, say, mostly frugivorous clade. Unfortunately phylogenetic comparisons of this sort aren't as informative as they might be, because vampires actually appear to be the earliest branch of Phyllostomidae to diverge, meaning that they don't group closely with other phyllostomid species (Wetterer et al. 2000, Jones et al. 2002) [adjacent image, of the nectarivorous phyllostomid Glossophaga, from here].

We might learn something about the primitive condition for vampires by looking outside of Phyllostomidae and at the most closely related bats, the mormoopids (moustached or spectacled bats) and noctilionids (bulldog, fisherman or hare-lipped bats). Phylogenetic work on these groups indicates that their primitive feeding habit was insectivory, and this - together with the fact that part-time insectivory is widespread throughout Phyllostomidae - suggests that the same feeding style was the case for the ancestral phyllostomid, and hence for the ancestral vampire. Hypotheses 2 and 3 therefore have some phylogenetic support.

How do the other hypotheses hold up? Hypothesis 1 (proposing that sanguivory might have evolved from fruit-eating) lacks support given that vampires are not part of a frugivorous phyllostomid clade, nor are the phyllostomid outgroups frugivorous. Furthermore, Fenton (1992) was suspicious of the fact that other bats - notably fruit bats - haven't ever switched from frugivory to sanguivory if this transition was so easy. Hypothesis 4 (proposing that vampires switched from eating small vertebrates to sanguivory) doesn't make much sense, and again lacks phylogenetic support given that none of the bats closely related to vampires are true carnivores. Hypothesis 5 (positing that vampires took chunks of flesh from animals before switching to sanguivory) also requires that vampires evolved from carnivores and again lacks support.

Of the two hypotheses that hold up best, hypothesis 2 (the proposal that vampires first fed on ectoparasites) is particularly intriguing given that elsewhere among tetrapods - among oxpeckers - we seem to be seeing the same thing. Namely, a possible transition from the eating of ectoparasites to the eating of body tissues. It's been argued that a diet of ectoparasites is unlikely to have led to sanguivory in bats however, partly because the nocturnality of bats would make them poor at finding ectoparasites (I'm not sure I accept that, given the sensory abilities bats have evolved to locate favoured food sources), and also because ectoparasites are present worldwide and yet vampires only evolved in the Americas (Fenton 1992, Timson 1993) [oxpecker image from Turnikator's flickr site].

Fenton's (1992) favoured hypothesis - that wound-feeding led to sanguivory - is in agreement with the phylogenetic evidence for an insectivorous origin of vampire bats. However, like the ectoparasite-feeding hypothesis it fails to explain why sanguivory didn't evolve among other insectivorous bats. Furthermore, Fenton (1992) argued that the evolution of vampires was fuelled by the fact that the Americas had more megafaunal mammals that other continents, and that wounds may have been more frequent in these taxa than in those elsewhere in the world. However, the phylogenetic data strongly indicates that vampires did not evolve to drink the blood of mammals on the ground, but of birds in the trees (among vampires only Desmodus feeds on mammals - the other species predominantly feed on birds). It's true that, during the Cenozoic, the Americas were home to an awful lot of birds, but the same could be said of other continents. The wound-feeding hypothesis remains unsatisfactory, therefore.

So what are we gonna do? Well, I have a new hypothesis, with a really catchy title. Check back soon: I will be posting it within the next few hours (life permitting).

Refs - -

Fenton, M. B. 1992. Wounds and the origin of blood-feeding in bats. Biological Journal of the Linnean Society 47, 161-171.

Ferrarezzi, H. & Gimenez, E. d. A. 1996. Systematic patterns and the evolution of feeding habits in Chiroptera (Archonta: Mammalia). Journal of Comparative Biology 1, 75-94.

Jones, K. E., Purvis, A., MacLarnon, A., Bininda-Emonds, O. R. P. & Simmons, N. B. 2002. A phylogenetic supertree of the bats (Mammalia: Chiroptera). Biological Reviews 77, 223-259.

Timson, J. 1993. How vampire bats acquired a taste for blood. New Scientist 137 (1855), 18.

Wetterer, A. L., Rockman, M. V. & Simmons, N. B. 2000. Phylogeny of phyllostomid bats (Mammalia: Chiroptera): data from diverse morphological systems, sex chromosomes, and restriction sites. Bulletin of the American Museum of Natural History 248, 1-200.