The Death Goby and Her Kin


The day Hurricane Isaac rolled through Baton Rouge (Aug 28th) saw the publication of what I consider the best project that I’ve ever been a part of: a paper showing the sister relationship between Milyeringa, a genus of cavefishes from Australia and Typhleotris, a genus of cavefishes from Madagascar. [Get the original open access paper by clicking here: Chakrabarty et al. 2012] I’d like to tell the tale of the Malagasy trip that was the inception of the project and to tell a bit about the story behind the paper before diving into the science.

When I was a young boy dreaming of being a zoologist, I dreamt of going to far off countries and discovering wildly amazing new creatures in places no one has ever been. In my childhood nightmares I feared the dark and being trapped in small spaces. My trips to Madagascar and Australia fulfilled both the dreams and nightmares of my youth.

Sinkhole fever and the Death Goby.
Fig1. The dreaded sinkhole.
I went to Madagascar shortly before my postdoc ended at the AMNH (American Museum of Natural History) and after I had been hired at LSU (Louisiana State University) in the summer of 2008. It was a weird period in my career when I was probably focusing more on the future than the present. My wife and I were in the process of buying our first house; in fact I was signing documents (including my LSU contract) up to the moment I was getting into the cab headed to the airport. The trip to Madagascar came at the perfect time for me; I was done with the stresses of having to find a job, and was ready for something to highlight the end of my old position. As luck would have it the Constantine S. Niarchos Expedition Fund provided my postdoctoral mentor, John Sparks (Curator, AMNH), with a grant to travel to Madagascar and collect blind gobies. I actually didn’t pay much attention to why we were going to Madagascar, I just knew I wanted to go: I’ve always wanted to go. My first work in science, as an undergrad with Melanie Stiassny, was on Malagasy cichlids (Stiassny et al. 2001) and I always saw Madagascar as one of the last remaining wild places on Earth. I was so busy buying a house in Baton Rouge, negotiating my start-up, and packing for the trip that I really didn’t know much about the actual taxa we were after in Madagascar. I had been working on bioluminescent fishes my entire postdoc and never worked with gobies or caves before. I figured I would catch-up on these subjects during the plane ride over to Antananarivo (Tana). As it happened I did more catching up with Scott “Scuppy” Holtz (now a grad student at Cal State Fullerton) and Phil Willink (then at the Field Museum) who were joining John and I in the field. By the time we got around to our first field site (after much stamping of documents, preparation of materials, shaking of hands, and days of impossibly slow and shaky driving across terrible Malagasy “roads”) I sort of had an idea that we would be collecting some rarely seen and poorly understood freshwater cavefishes, namely two species of Typhleotris. Most of what we knew about these species was from their original descriptions (in the 1930s and 50s) and that others had noted that they share a striking resemblance to the cavefishes of Western Australia. John had done a ton of research on the caves of the region, and with the help of the wonderful Steve Goodman, discovered a large sinkhole that was mentioned in a French geologist’s dissertation that possibly contained blind cavefishes. Sinkholes are not caves, they are exposed to the open air and sunlight; what is visable is just part of a larger subterranean habitat (i.e., a karst window). I was horrified and impressed when I was first confronted by this football-field sized sunken lake (Fig 1). The vintany, as the locals called it, made my heart sink into my stomach. ‘What could possibly live down there?,’ I thought to myself. I was trying not to let my nervousness show when Scuppy secured a flimsy chain ladder to a dangling tree at the edge of the sinkhole. As luck would have it I went into the water first while the others secured more equipment. The first part of the climb down over the lip of the sinkhole was particularly nerve-racking because you couldn’t see over the edge to your next foothold. The local guides sat a safe distance away with no interest in getting closer to the insides of this strange pit.
As I snorkeled I started thinking of what I would do if I encountered one of the huge Nile crocodiles this region is famous for. I felt a bit like a miner’s canary, but I didn’t see anything: no fish, no crocs, no beluga whales. All I encountered were some spiders and insects and a beautiful view of the deep dark waters below. As I climbed back up the ladder I saw John and Phil were getting ready to head down. They swam for a long while and after some 25 minutes John yelled out, “I got one.”
‘Got one, what?’ I thought.
Fig 2. A new species of Typhleotris.
He passed the specimen to Scup who climbed up the chain ladder with an ease I wish I possessed. I looked at this little beast, no longer than my pinky, and as dark as a Hershey’s chocolate. How strange this creature was: it lacked eyes and had a smooth bony head (Fig 2). Why would there be an eyeless fish here? Why is it so darkly colored?  The water is clear and sunlight penetrates as far as you can see; the sinkhole appeared more than 50 meters deep, perhaps more. What lies beneath, the French geologist’s dissertation told us, was a subterranean groundwater connection to caves several kilometers away. After some time John yelled out again: he had gotten another one. Around this time I started to get nauseous, I had ingested some of the water while snorkeling and it wasn’t sitting well. I tried not to think about it while I dutifully took photos, GPS coordinates, and tissue samples. The locals had warned me that this was a revered site and that I should not urinate near the mouth of the sinkhole. I didn’t need to pee but I did need to vomit, although the water was clear it was full of loose vegetation and was effectively a pit trap for anything unlucky enough to fall in.
I walked out of sight of everyone and vomited the sinkhole water I had accidentally ingested. I came back to my notebook, camera and specimen just in time for the second specimen to be brought up to me. John and Phil who had been swimming for more than an hour finally came up. This second specimen was dark and eyeless like the first one, and a similar size. John and Phil told me they had seen others but those individuals slowly sank away into the deep just out of reach. The dark coloration is a great camouflage and the fish are certainly aware of their surroundings, enough so to move away from larger moving objects splashing around. Were this species white, as most subterranean species are, it would have been easy pickings for birds that could strike from above and perhaps even to swimming snakes or other creatures that could handle the initial 30ft descent.
We were feeling pretty good about ourselves having collected something amazing from the very first site. We were in good spirits when we retreated back to our camp. A couple days later when we started collecting at a new location a little further north it became clear that all was not well. In deference to my colleagues I won’t replay how sick the team got, but it was pretty scary stuff - especially given how far we were from civilization. I think I was spared major illness because I expelled most of the water I had ingested while snorkeling. We jokingly named the mysterious illness “sinkhole fever,” but seeing your mentor and friend so ill that you consider using the satellite phone to call in a helicopter is no joke. The next day we sent Scuppy and John back to Tana. Scuppy went to a hotel to rest up while John flew back to New York.
            We did quite a bit more collection after that. Phil and I went to a dozen or so more caves over the next week and had quite the amazing experience in the southeastern part of Madagascar. Scuppy, who had recovered well in Tana, was well enough to join me for a trip to the northern tip of Madagascar (Ankarana) after Phil departed. In the north, Scuppy and I collected both blind and sighted members of Glossogobius, sometimes right next to each other.  Those caves in the north remain the most spectacular places I’ve ever been, but it will always be the vintany, home of the darkly pigmented “Death Goby,” that will be the defining site of the trip. [The description of this darkly pigmented form is in press and should come out later this year.]

Back Home.
            On returning from Madagascar I effectively was done at AMNH and was now a new faculty member at LSU. All the sequencing and analysis was done at LSU for the project that would eventually become our PLoS paper. The addition of Matt Davis as my postdoc, the first person to join my lab in 2010, really transformed how we viewed our data. I am not a goby biologist, and although we had some gobies collected we didn’t have enough for a wide-scale global phylogeny. Luckily there was already a well-sampled phylogeny completed using mitochondrial genes (Thacker 2009). It was our intention from the start to stick our newly collected samples in with the previous data available on GenBank. The other piece of the puzzle was the Milyeringa from Australia. Several of the genes for this species had been sampled already for a few specimens loaned to Chris Thacker, but not a large sampling. In my 2009 trip to Australia I collected some additional specimens, including what turned out to be a new species of Milyeringa (described in Chakrabarty, 2010). With both pieces of the puzzle sequenced we discovered what we and others had suspected, these two cave lineages, now separated by more than 6,000km of Indian Ocean, are sister taxa.

The Paper.
Fig 3: Summary of the relationships of Gobiiformes. See original PLoS paper for details.
            The cavefish paper describing the relationship between Milyeringa and Typhleotris published in PLoS One includes John Sparks and Matt Davis (now at the Field Museum in Chicago) as co-authors (Chakrabarty et al. 2012); everyone contributed equally. The dated phylogeny we present for Gobiiformes is the only time-calibrated phylogeny of this extremely diverse group and uses mostly taxa that had been sampled by others for molecular data (again mostly from Thacker, 2009). Our main concern for this paper was the relationships among blind and cave dwelling forms. For further discussion of goby relationships see the recently published Biology of Gobies (Patzner et al., 2011; or papers like Hoese and Gill, 1993; Larson, 2009; or Thacker, 2009). Besides the five species of Milyeringa and Typhleotris we also included blind members of Glossogobius (also collected in the 2008 trip) and Typhlogobius in our phylogeny. Less than 1% of described fish species are blind, and a stygobitic (aquatic cave dwelling) lifestyle has only evolved in 20 families of the 500 or so families of bony fishes. Stygobitic forms are especially poorly known among gobies. Besides those mentioned above there is Caecogobius (from Philippine caves), Oxyeleotris (from Papua, and the only blind eleotrid we don’t have samples for) and a few species of Luciogobius and Typhlogobius, found in seaside caves in marine or brackish water). Given the limited number of blind gobies species known it probably isn’t a surprise that our optimization recovered the common ancestor of Milyeringa and Typhleotris also as a blind taxon, with independent origins of blindness recovered for both Glossogobius and Typhlogobius (Fig 3). When collecting Glossogobius in Madagascar, I couldn’t help but notice how different they were from the Typhleotris we collected in the first part of the trip. We could easily scoop up most individuals of Typhleotris with a simple swipe of a dipnet (or even our hands) while individuals of Glossogobius ankaranensis were much more skittish and swam away at almost the same speed as the sighted G. callidus that we collected in the same caves. Individuals of G. ankaranensis that we collected were depigmented and blind but you could still see the remnants of a dark bead-sized lens right where the full eyes would be, implying a more recent loss of sight. The Typhleotris species had bone covering the region of the orbit and no eye to speak of. The specimens of Milyeringa I sampled in Australia look very much like their cousins in Madagascar. I couldn’t help but note the red dirt and baobab trees that are also common to the regions where these fishes are found; I like to image that this must have been what parts of Gondwana looked like.
            Using what limited data there is on goby fossils, and using a number of outgroups that did have a good fossil record, we were able to date the divergences across the tree and as it is related to our node of interest. Incorporating the error bars around the divergence estimates, we recovered ages that were congruent with the break up of eastern Gondwana. A couple of caveats come with this: (1) our phylogeny uses four mitochondrial genes, so it is essentially dated on a single locus, and, (2) Australia and Madagascar were never directly abutting in any current Gondwanan model, India and Antarctica were always in between.
            Part of the early criticism we have received for our paper is that we don’t discuss the basal relationships of the other gobies in our dated phylogeny (see FB thread below). We didn’t discuss the other non-blind taxa much because we are fully cognizant of the fact that outside of the subterranean taxa of interest we did not collect most of the other gobies in our tree. Although certainly our dated phylogeny is unique, we do think a discussion of gobiiform relationships is better suited for a paper by those with vested interests in those relationships. Our interests centered around the relationships among the blind members. Our tree was the first to incorporate several of these lineages but we didn’t add many new taxa and no new genes from previous work on Gobiiformes. Our approach of using Bayesian inference to estimate the phylogeny and divergence times simultaneously while using a more expansive outgroup sampling may be novel, but the real work on gobies will have to come from the addition of morphological and nuclear characters by folks who know these taxa better than us (I’m looking at you Chris Thacker, Luke Tornabene and others). My co-authors and I were trying to get a better understanding of the relationships of the subterranean lineages, so forgive us if we did not mention the rest of the tree much, or cite particular papers discussing those relationships.
As for caveat #2: are we suggesting that there are likely cavefishes in India that may be related to Milyeringa and Typhleotris? Or that they were on Antarctica? Possibly, one can only speculate about “missing” lineages. We continue to discover that our knowledge of major geological events are not quite as stable as we once thought. (Remember when the Closure of the Isthmus of Panama was only 3.5mya?) I think assuming that the current timing and scenario of the break-up of Gondwana is the last word on the subject would be a mistake. One intriguing possibility is of a possible direct connection between Madagascar and Australia in a land connection called Pandora (Parenti and Ebach, 2010). The cavefishes in their current incarnation as Milyeringa and Typhletoris are probably not capable of dispersing much beyond their isolated aquatic cave systems but that doesn’t mean that the cave systems themselves haven’t moved around and evolved. More than likely they did, just look at the karst window that gave rise to the Death Goby. The scenario I picture in my head is of a widespread lineage probably composed of both cave dwelling and non-subterranean species on eastern Gondwana; the drifting continents then separated this lineage. Most of the members of this lineage slowly went extinct with time (while adding a few new ones now and again) and all that remains today are a handful of species in isolated caves in Madagascar and Australia. What better way to escape extinction that to hide out in a cave? (Just ask Al Qaeda.) Of course there are other potential scenarios, as we discuss in our paper, this one is just the most palatable for me. However, without evidence of sighted members of the Milyeringa + Typhleotris clade, the simplest conclusion is that the ancestor was a blind cavefish too (as we report in the paper).
Any way you slice it, the best explanation we currently have for this lineage of cave dwellers remains that they are a Gondwanan relict. Given their very limited dispersal ability (blind, restricted to their cave habitats) and the ancient age we recover for them, it would seem that this group might be one of the best cases for a group that shows a vicariant Gondwanan pattern.

References
Chakrabarty, P., (2010) Status and phylogeny of Milyeringidae, with the description of a new blind cave fish from Australia, Milyeringa brooksi, n. sp. Zootaxa 2557:19-28.

Chakrabarty, P., Davis, M.P., Sparks, J.S. (2012) First reported case of a trans-oceanic sister-group relationship between vertebrate troglobites. PLoS One 7:e44083 (pg.1-8).

Hoese, D.F. and Gill, A.C. (1993) Phylogenetic relationships of eleotridae fishes (Perciformes: Gobioidei) Bulletin of Marine Science 52: 415-440.

Larson, H.K. (2009) Review of the gobiid fish genera, Eugnathogobius and Pseudogobiopsis (Gobioidei: Gobiidae: Gobionellinae), with descriptions of three new species. The Raffles Bulletin of Zoology 57: 127-181.

Parenti, L.R. and Ebach, M.C. (2010) Wallacea deconstructed. In: Beyond Cladistics: The Branching of a Paradigm (D.M. Williams and S. Knapp, eds.) University of California, pp. 303-318.

Stiassny, M.L.J., Chakrabarty, P. , Loiselle, P. (2001) Relationships of the Madagascan cichlid genus Paretroplus, with a description of a new species from the Betsiboka River drainage of northwestern Madagascar. Ichthyological Explorations of Freshwaters 12: 29 – 40.

Thacker C.E. (2009) Phylogeny of Gobioidei and placement within Acanthomorpha, with a new classification and investigation of diversification and character evolution. Copeia 2009: 93­–104.