Saturday, April 7, 2012

Eel mocassins


In all likelihood, you'll never find me.
If you held a gun to my head (or anyone's) and told me to find you a rainbow snake, even just a single one, I guarantee I wouldn't be able to do it without extraordinary luck. And yet, there are tens, possibly hundreds of thousands of rainbow snakes out there in the rivers and wetlands of the southeast. So it's not a problem of rarity per se. Why is it, then, that these snakes are so seldom seen, even by avid snake hunters?

Rainbow snakes epitomize several of the qualities that make studying snakes challenging. For this reason and others, they have long been among my favorite snakes, a Holy Grail of sorts. They really are worthy of further study, because we know so little about their natural history.

You might ask me: have you ever found a rainbow snake? And I don't mean to brag, but yes, I have. Three times (two were the same snake), I have been smiled upon by the herping deities. At first glance, it doesn't seem that difficult to find one: they are bright pink, yellow, and blue violet. They reach lengths of over six feet long. A snake like that should stick out like a sore thumb. Additionally, they occur throughout the southeastern Atlantic and Gulf coastal plains, areas well-populated by snake hunters and herpetologists.

Range of the rainbow snake - SEGAP 

But in reality, very few people are lucky enough to see even one, and finding a rainbow snake is a watershed event in the life of a herper. What, then, causes the apparent rarity of this species?

Let's start by reviewing what we know about the natural history of this snake. It is one of two species in the genus Farancia, the other being Farancia abacura, the mud snake. While the common names of the two species conjure quite different images, the two snakes are actually quite similar in appearance and morphology.

Farancia abacura (left) and F. erytrogramma (right) 

The mud snake is a bit better known and a bit more commonly encountered, but finding one is still an event, and as little is known about mud snake natural history as about rainbow snakes. Although the two species were once considered to be in different genera (Abastor for erytrogramma), they are clearly sister taxa. The ease of classifying them stops there, however. In order to understand the broader phylogenetic position of Farancia, we need to look at the historical biogeography of North America. Although the dates in question are a matter of some debate, the order of the events I am about to chronicle is fairly solid. The colubroid snake fauna that inhabited the continent 16 million years ago were called the xenodontine snakes. Before the more recent invasion from Asia by and radiation of many of the familiar North American snakes (natricines like garter and watersnakes, colubrines like rat and kingsnakes, elapids such as coral snakes, and vipers), xenodontines had the continent mostly to themselves, save for a few boas and blindsnakes. There is no question that there are fewer xenodontine species alive today than there were at that time. Whether the xenodontines were never very abundant, or whether they have been pushed to marginal niches by more competitive colubroids, their ancestors' common descendants survive in South America, in the hyper-diverse radiation known as dipsadines (including many unfamiliar snakes, but a few semi-familiar ones, such as the goo-eating Dipsas, coral-snake mimicking Eyrthrolamprus, and pencil-thin Imantodes). The reason the xenodontines and dipsadines have separate names despite sharing a common ancestry is that they developed in isolation for several million years, when North and South America were separated by ocean (though the separate names , like much else in herpetology, aren't universally accepted, applied, or used). A few dipsadines have re-invaded North America following the recent reconnection of NA and SA (~3.5 million years ago), including the desert nightsnake, Hypsiglena torquata, in the west, and the pine woods snake, Rhadinaea flavilata, in the southeast. In a few places they have met the descendants of their great-great-grandparents, the relict North American xenodontines. It is to this group that the two Farancia belong, together with a few other interesting North American snakes, including the hog-nosed snakes (Heterodon), ring-necked snake (Diadophis), worm snakes (Carphophis), and sharp-tailed snakes (Contia). The convergence between these xenodontines and some dipsadines is remarkable, but that's a story for a different article.

Farancia erytrogramma
Pseudoeryx plicatilis









So just what is the niche of the rainbow snake? Any field guide will tell you that they are completely aquatic, in part due to their highly specialized diet of American eels (Anguilla rostrata), which they forage for at night (see this video). But a 1964 review by Wilfred T. Neill, herpetologist extraordinaire of Florida and the southeastern US, provides more insight and detail than is typically found in a field guide. Neill had a keen eye for natural history, and it shows in his publication on the rainbow snake.

In his article, Neill states that alkaline (basic) aquatic habitats are inhabited by rainbow snakes, whereas mud snakes tend to inhabit more acidic streams, bogs, and ponds. This is not strictly true - but in general, it serves as a good rule for ecologically separating the two species. Another consideration is that the prey of the mud snake generally inhabit isolated wetlands, which tend to be more acidic, whereas the prey of the rainbow snake can only be found in watercourses with a permanent connection to the ocean. This is because American eels are catadromous fishes, meaning that they spawn in the ocean but live much of their lives in fresh water, sort of the inverse of a salmon, lamprey, or sturgeon. Rainbow snakes are so particular about their diet as adults that they are often called the 'eel moccasin' by "rural residents of the Georgia Coastal Plain" because "ever' time you see one, he got a eel tail a'hangin' out his mouth" (quotations from Neill).

A video still of a rainbow snake eating an eel
Now, you may be familiar with the problems inherent when migrating fishes, like salmon or eels, encounter man-made impoundments along river courses. In the Pacific northwest, anadromous salmon are unable to move upstream past the dam to spawn, which has led fisheries managers to construct 'fish ladders' to help salmon bypass dams. Baby salmon can then move downstream through the dam, although many are certainly obstructed on their journey to the ocean. The situation for the eel is a little different, because they begin life in the ocean instead of at the headwaters of a stream, but the challenges faced during the life of the fish are the same. Upstream of dams along southeastern rivers, eel populations have dwindled to almost nothing, and it's likely that rainbow snake populations in those areas have been affected. In addition to being largely unable to move past dams in either the up or downstream direction themselves, rainbow snakes living upstream of a dam surely have to contend with slowly dwindling eel populations. However, data on rainbow snake population declines are nonexistent. Why? 

Who wouldn't want to study a snake like that? 

One big challenge is the inherent uncertainty in estimating what a species' population size is, given incomplete detection of all individuals. The stochasticity associated with these kinds of estimates is controllable to an extent dependent on the detectability of the species in question (i.e., you can get more accurate population size estimates for highly detectable, recapturable species like watersnakes than you can for undetectable species like rainbow snakes). Because we rely on recapture rate to inform our estimates of population size, animals that are never recaptured don't do much good, which is all of them when detectability is low enough (advances in models using presence/absence data have somewhat mitigated this problem). What's more, detectability varies with species life history, with time of year, with survey method, with spatial scale, with observer experience and bias, and with environmental variables such as temperature, moon phase, rainfall, other species interactions, basically anything that affects animal activity is going to impact detectability. So it's really hard to separate environmental effects that extirpate or reduce the size of a species population (i.e., affect species occupancy) from those that influence their activity (i.e., affect species detectability). Also, anthropogenic environmental effects are often spatially and temporally autocorrelated, so it's hard to assign a decline, if you can find one, to just one cause. These problems are particularly bad for animals whose natural history we don't know that much about, like rainbow snakes.

Some colleagues and I did a study to measure just how hard it was to catch a rainbow snake. We compared the rate of capture to that of six other aquatic snake species that also inhabit isolated wetlands in the South Carolina coastal plain. We used robust presence-absence based models that estimate the probability of detecting a species rather than the population size of that species, so that recapturing the same individual animals was unnecessary. We set minnow traps overnight in different wetlands for the entire summer, and recorded all the snakes that got caught in them. We also recorded information on environmental variables, such as the prey community at each wetland, the wetland hydroperiod, and the wetland's connectivity to other wetlands. We put this data into the model, and asked it to tell us which site features best predicted a high probability of capturing snakes of different species.

Figure from Durso et al. 2011 

As you can see, the likelihood of catching a rainbow snake can be quite low, but it improves in wetlands located close to the nearby Savannah River floodplain. This actually isn't too surprising, considering that their prey live only in rivers. The Savannah has relatively few dams along it, because lowering the water level in certain spots would expose Cesium-contaminated sediments that resulted from the nuclear reactors on the Savannah River Site, where the research was conducted. So, in an indirect way, the radioactive waste might actually be helping the rainbow snake, although how it impacts them directly is difficult to say.

One more enigma about rainbow snakes - it turns out that the conventional wisdom that they are entirely aquatic isn't entirely accurate. You see, unlike most of the aquatic snakes in North America, Farancia lay eggs (remember, they are xenodontines, which are oviparous, as opposed to natricines, like garter and watersnakes, which are viviparous). To do this, they must go onto land, much like a sea turtle must also return to land to lay her eggs. Actually, I think turtles are a good ecological analogy for Farancia. The male Farancia may never leave the water, but the females, when they reproduce, must. Rainbow snakes have sometimes been found in dry, sandy fields, and observations from 1940s Virginia assert that they are "the most abundant snake" and "more often seen on land, [but] equally at home in the water". Many were turned up by plowing fields, but some were found crawling on the surface as well. Nests have been found in the same dry, sandy fields, among mud turtle, slider, and cooter nests. At the aforementioned Virginia site, young rainbow snakes were found under boards in spring and fall, suggesting that some may migrate to the water soon after hatching, whereas others may overwinter in the nest, again in the manner of many turtles.

The mud snake has the largest recorded clutch size of any North American snake, but the rainbow snake is no slouch in the egg department, laying clutches up to 52 eggs in size. Some females may remain with the eggs until they hatch, but this phenomenon is poorly studied.

Juvenile rainbow snake captured in a drift fence 
Wilfred Neill, indefatigable field naturalist and publisher that he was, described two populations of distinctively colored rainbow snakes from brackish water habitats in Florida, with very deep red coloration replacing the yellow on the chin and throat. Neill also described the South Florida Rainbow Snake, Farancia erytrogramma seminola, a supposedly now-extinct subspecies, from three specimens he collected and placed in the collection of Ross Allen (of Reptile Institute fame). These specimens were collected in Fisheating Creek in Glades County near Lake Okeechobee, and they differed from other rainbow snakes in having predominantly black ventral scales. Neill considered this, but not the saltwater populations, distinct because the biogeographic history of south Florida, which was a large island in an archipelago during the Pleistocene, would have allowed some degree of divergence and concomitant isolation, whereas the saltwater populations probably were not isolated from other populations. Argue all you want about what a subspecies is supposed to be (I think they're often relatively meaningless), but recently several conservation groups seized upon the South Florida Rainbow Snake as a posterchild for snake conservation. Now, the rainbow snake can use all the publicity it can get among lawmakers and the general public, and I would eventually like to see a concerted effort to survey and monitor all rainbow snakes rather than just this single population. However, one has to start somewhere, and when it's rainbow snake surveying you're talking about, anywhere is about as good as anywhere else. I spoke with Cameron Young, the director of the Center for Snake Conservation, about his recent expedition to Fisheating creek to look for F. e. seminola. He reported excellent habitat, a creek full of eels, and reliable stories from local folks of rainbow snakes seen recently - so hopes are high for a rediscovery, and an opportunity to learn more about the biology of this remarkable snake.

If you have found a rainbow snake, leave a comment below! You're a member of an elite club, and I'd love to know about it. 

ACKNOWLEDGMENTS
Thanks to photographers John White, Pierson Hill, JD Willson, Todd Pierson, Nathanael Herrera, Cameron Young, and Antoine Baglan, and to Cameron Young for an advance synopsis of his results from Fisheating Creek in March 2012.

REFERENCES
Durso AM, Willson JD, Winne CT, 2011. Needles in haystacks: estimating detection probability and occupancy of rare and cryptic snakes. Biological Conservation 144:1506-1513. <link>

Gibbons, J. W., J. W. Coker, and T. M. Murphy, Jr. 1977. Selected aspects of the life history of the rainbow snake (Farancia erytrogramma). Herpetologica 33:276-281.

Haro A, Richkus W, Wahlen K, Hoar A, Busch W, Lary S, Brush T, Dixon D, 2000. Population decline of the American Eel: implications for research and management. Fisheries Management 25:7-15.

Neill, W. T. 1964. Taxonomy, natural history, and zoogeography of the rainbow snake, Farancia erytrogramma (Palisot de Beauvois). American Midland Naturalist 71:257-295.

Palisot de Beauvois, A. 1801. Mémoire sur les Serpens.in C. S. Sonnini and P. A. Latreille, editors. Histoire naturelle des reptiles. Imprimerie de Crapelet : Chez Deterville, Paris.

Powell C, Stevenson DJ, Smith M, Jensen JB, 2010. A new clutch size record for the Mud Snake (Farancia abacura). Southeastern Naturalist 9:177-178.

Richmond ND, 1945. The habits of the rainbow snake in Virginia. Copeia 1945:28-30.

Semlitsch RD, Pechmann JHK, Gibbons JW, 1988. Annual emergence of juvenile mud snakes (Farancia abacura) at aquatic habitats. Copeia 1988:243-245.

Steen, D., C. Guyer, and L. Smith. 2012. Box 9: Relative abundance in snakes: a case study. Pages 287-294 in R. McDiarmid, M. Foster, C. Guyer, J. Gibbons, and N. Chernoff, editors. Reptile Biodiversity: Standard Methods for Inventory and Monitoring. University of California Press, Berkeley, CA.

APPENDIX
The taxonomic history of Farancia erytrogramma is not overwhelmingly complex, but it is somewhat interesting, at least to me (which is why this part is in the appendix). The species was originally described as Coluber erotrogrammus, placed in the catch-all non-venomous snake genus Coluber as most snakes were in the early 1800s, only a few short decades after Linnaeus. The description was published in Sonnini & Latreille's "Histoire Naturelle des Reptiles", but the actual author was Palisot de Beauvois, whose "Memoire sue les Serpents" constituted the snakes chapter of HNdR. None of these guys were specifically herpetologists, but the distinction between herpetology and entomology was still pretty hazy back in those days, with most people tending to lump all the creeping things together. They were all three explorers, naturalists, and probably really fun guys. Supposedly Latreille, an orphan who discovered isopods, created the concept of a type species, and was generally considered the foremost entomologist of his time, escaped prison and execution by identifying a rare species of beetle in his cell and so impressing the prison doctor that he secured Latreille's release.

So Palisot de Beauvois gets the naming credit (called the authority) for F. erytrogramma - but he kind of misspelled it. The Latin prefix erythro- means red, but PdB left out the h for some reason, which Neill derided as "etymologically poor" but not an "inadvertent error". Many authors inadvertently gave credit to Sonnini & Latreille or to Latreille alone, and the date of the work was frequently miscited as 1802 (the correct year is 1801). Neill cleared this up in his 1964 review, and synonymized Farancia and Abastor, both named by J.E. Gray of the British Museum, who was "well known for inventing many apparently meaningless scientific names." The only other change in need of explaining is the gender change of erytrogrammus to erytrogramma, to agree with the feminine Farancia instead of the masculine Coluber or Abastor.

In my research for this article, I came across the German words for these species, which I think are just wonderful: schlammnatter for mudsnake and regenbogen-schlammnatter for rainbow snake. Natter means snake, schlamm means mud, and regenbogen means rainbow (from regen: rain, and bogen: arc).

1 comment:

city said...

thanks for sharing.