A new fish-scale gecko from Madagascar 1

A paper by my colleagues and me appears today (7 February 2017) in PeerJ, describing a new species of gecko from northern Madagascar with large scales that come off with extreme ease. This is actually a characteristic of the whole genus Geckolepis, to which the new species belongs, and has been known about for at least 150 years. But what is really remarkable about the new species, Geckolepis megalepis Scherz, Daza, Köhler, Vences & Glaw, 2017, is that the scales are quite a bit larger than any other species in the genus (Scherz et al. 2017), and are indeed the largest scales of any known gecko, both relative to body size and in absolute millimetres.

Geckolepis megalepis in life. Look at those huge scales! Photo by Frank Glaw.

Click this image to be taken to the paper in PeerJ.

If you want to read the press coverage on the new species, you can do so at any of the following links (I will dynamically update this list as press stories emerge):

Last updated: 20 February 2017, 12h39 CET.

The reason I direct you away from my blog, is that I want to talk about this story from a slightly different perspective, namely my own part in the research, and the significance of the new species “straight from the author’s mouth,” as it were. This is quite different from what the press is talking about, which doesn’t so much tell the story of the paper as the findings.

Geckolepis: a few cursory facts.

First, a brief overview of the genus Geckolepis: This genus was first described by Alfred Grandidier in 1867. Today it contains five recognised species, including our new one: G. typica, G. polylepis, G. maculata, G. humbloti, and G. megalepis. The genus is endemic to Madagascar and the Comoro islands. They are best known for their large scales that come off easily. These scales regenerate very quickly. They are all nocturnal, arboreal lizards, with a mostly insectivorous diet. Their taxonomy needs a lot more work. That’s where I come in.

Scale loss in Geckolepis megalepis. Photo by Frank Glaw.

The story of the paper

In 2014 I started the second semester of my Master’s degree at the Ludwig-Maximilian Universität München. In this semester, I was to work on a research project that was to culminate in a poster. So I went to Dr. Frank Glaw, one of the leading experts on the reptiles and amphibians of Madagascar, who had been my supervisor for the frog side-project I had conducted in the first semester, and together we developed a project that would not only satisfy the requirements of the programme, but that would ideally also result in a paper. I had suggested to work on a reptile, because I had already begun gaining experience with frogs. Ideally it would be a gecko because… I like geckos.

Geckolepis typica, from southern Madagascar. Photo by Mark D. Scherz

From among the dozens of potential projects (given the many undescribed reptile species already known from DNA barcoding studies on Madagascar’s reptiles), we selected Geckolepis. The decision was made for several reasons. First and foremost, the species in question had been relatively well circumscribed already in a previous study by Lemme et al. (2013). In this study, they had characterised the genetic variation of the genus Geckolepis, and found there to be numerous cryptic lineages, meaning an estimated eight to twelve species, of which at least four were undescribed. They then briefly circumscribed the morphological variation in each of their Operational Taxonomic Units (OTUs). One of these, OTU ‘D’, was particularly different; they characterised it as follows:

OTU D. ― This OTU occurring at Ankarana differs from all other OTUs by the lower number of scale rows around the midbody (17–20 vs. 22–38 in all other OTUs) and a lower number of ventral scale rows (27–32 vs. 33–55 in all other OTUs). It furthermore differs from OTU EFG, G. typica and G. polylepis by a larger body size (SVL <67 vs. <54 mm).

This made OTU D the most morphologically distinct of the genetic lineages—a prime candidate with which to begin to resolve the taxonomy of the genus. It is important to remark at this point that Lemme et al. (2013) also assigned a different OTU, OTU ‘AB’, to the historical name Geckolepis maculata, the identity of which has been unclear since it was realised that the diversity of Geckolepis has been so underestimated.

Another reason this project was selected was that we hoped that micro-CT would prove invaluable in their resolution. With micro-CT it might be possible to verify G. maculata as a member of the OTU AB and thus put that question to rest once and for all, for the lack of genetics from this 120 year-old specimen made its attribution uncertain at best. At the same time, we hoped to identify characters of these geckos that are invariable and not damaged by collection—after all, most specimens in natural history museums lack at least part of their integument, as a result of the quite incredible scale-loss of these geckos. Having osteological characters to provide an unambiguous identification would greatly ease the study of these geckos.

The first step we took was to obtain micro-CT scans of the specimens. We scanned a selection of the animals available to us, including the holotype of G. maculata, two specimens of the new species, and a few specimens of other species/lineages. I should note also that shortly after this project started, we began work on another group from the genus, which was completed more quickly, wherein we revalidated the name Geckolepis humbloti for a species mostly restricted to the Comoro Islands (Hawlitschek et al. 2016). These first scans covered the whole bodies of the animals, and as such were at a relatively low resolution. As a result, the sutures between bones were difficult to see—this is an important character in lizard taxonomy, so we later had to go back and re-scan the skulls of those where the sutures were not shown.

The poster I presented at the IRT2 conference of my Master’s programme. Click to see at full resolution.

After this, I began to look in close detail at the morphology and scalation of the geckos. The scheme would follow Köhler et al. (2009), because theirs was the most comprehensive morphological analysis of the genus to date. Immediately it was clear that Lemme et al. (2013) had been quite right: OTU D does indeed have the largest scales of any member of the genus, and there are proportionally fewer of them as a result. With this solid distinction in hand, I went to look at the osteology of the specimens I had scanned. It soon became clear that the bones of these geckos are not much less variable than their scalation, though obviously for different reasons. Certainly a few differences were apparent between the new species and the G. maculata holotype, but in the context of the rest of the scans, it was not clear to me how great these were.

Nevertheless, the data added up to a clear meaning: this was indeed a distinct and unique species, and a name would be necessary. After toying with a few, Frank and I went with ‘megalepis’, due in part to the pleasant sound it has in apposition to the genus name (say ‘Geckolepis megalepis‘ out loud a few times to understand what I mean), but also because it accurately reflects what is really special about this species: its huge (‘megas’ in Greek) scales (‘lepis’ in Greek). I prepared the poster, and at the same time assembled a manuscript to describe the species.

What followed was a break of almost a full year on this project; in 2015, I worked all of six hours on the description paper. This is not because I lost interest, but rather because I was focussing on my thesis and a few other, smaller papers. For this year also we placed priority on G. humboti, as this paper was slightly more straightforward and simple.

The story of the paper describing G. megalepis might be shorter if the work involved in the poster would have sufficed—indeed, the paper would have appeared far sooner. But I was determined to use this opportunity to describe the skeleton of Geckolepis in extreme detail—far greater detail than I had achieved in Hawlitschek et al. (2016). These osteological descriptions are tedious, and if done wrong can spread a lot of misinformation. Having had little previous experience with reptile osteology, I was keen to bring on someone with a real expertise on geckos. So I invited Juan D. Daza to join the paper. Juan’s rather incredible publication record speaks for itself in terms of his extensive knowledge of gecko bones. He accepted the invitation, and we set to work.

A rotational gif of the skull of Geckolepis megalepis. Gif by Mark D. Scherz and Juan D. Daza.

Juan’s experience proved invaluable, for not only was he able to identify the small errors in my own work, but he could also delve deeper into the description of each element, and the significance of certain characters, than I—something that, by myself, I could only have done after months spent buried in the literature. He was able to highlight, for instance, the fact that the skull of Geckolepis has two very unusual characters for geckos, namely unfused frontal bones and subolfactory processes, the former of which is merely uncommon among geckos, and the latter of which is known only from a single, fossil gecko. But Juan and I also found evidence that the skull of the specimen of the AB OTU from Lemme et al. (2013) that we had scanned was not only different from G. maculata, but was in fact the most different of all the skulls examined. So we were able to reject the hypothesis of Lemme et al. (2013) that G. maculata and OTU AB are conspecific, and suggest that this OTU AB might have some skull characters that will make it relatively easy to describe, all things considered.

Quite quickly, a much improved manuscript came together. We started also to piece together more interesting information from the literature, which stretches back further on this genus than you might expect. We found for instance an old and partly forgotten previous study that had done an excellent job of characterising the integument of these geckos (Schmidt 1911), which gave us room for evolutionary and functional interpretations of the larger scales of G. megalepis. After a bit of to-and-fro among the coauthors, we finally submitted to PeerJ on the 27th of October 2016. On the 2nd of December we got reviewer feedback—which was far more positive than I could have hoped for—and 22 days later we resubmitted it. Eight days after that, it was accepted! This is certainly the fastest turnaround I have ever experienced, and the quality of the operation was remarkable. But I am not here to sing PeerJ’s praises; to hear my feelings about the whole experience, read my post on their blog.

Stereo 3D (magic eye) image of the full skeleton of Geckolepis megalepis. To get the 3D effect, position your face >30 cm/>12 in from your screen, relax your eyes, and cross them.

Geckolepis megalepis

In many ways, the new species is like any other member of the genus Geckolepis. Just like all other species, it has large, fish-like scales that come off with ease. It lives a nocturnal, arboreal lifestyle, and probably eats a wide range of invertebrates. It is not especially big. So why do we care?

A close-up of the boundary between lost and intact scales in Geckolepis megalepis. Note the clear attachment zone of each scale.

Well, firstly, the new species represents the height of gecko scale size achievement. Geckolepis as a whole have always been among the top competitors for this title, but G. megalepis is far and away the winner, with scales that are around 8% of the body length of the gecko in diameter. These scales, moreover, are ‘bony’ everywhere except on the lips. Only after the publication of the paper did I realise why this might be the case: The lip scales lie flat on the body of the animal, and are therefore not folded over against themselves like the scales that cover the rest of the body (Schmidt 1911). As a result they are thinner and not compacted, and probably do not have the space occupied by the bony particles found throughout body scales (Schmidt 1911). In the body scales, Schmidt (1911) showed that these bony formations lack bone cells—meaning that they are not technically ‘osteoderm’, despite their chemically similar composition.

Schmidt (1911) also described what he called the ‘Anwachsfläche’, German for ‘attachment zone’, at the base of scales, which is a region that was later characterised by Schubert & Christophers (1985) and Schubert et al. (1990) as possessing cells that make something of a pre-formed zone along which the scales, and their underlying skin, tear away. These studies showed that this is an active process, induced by one layer of cells in the skin that allows the scales and the skin to tear free. This is an incredibly sophisticated adaptation to avoid predation, with few parallels among terrestrial amniotes. In 2016, Frank Glaw returned to Ankarana and collected one additional specimen of G. megalepis to include in our study, but he actually found several individuals, and he photographed the skin of one after it had lost its scales. The photograph shows remarkably clearly the attachment zones of the scales that had not yet come away.

Secondly, G. megalepis represents yet another instance of microendemism to a single reserve of 182 km2 in northern Madagascar: Ankarana National Park (formerly a Special Reserve). I will avoid listing them exhaustively here, but recently described species endemic to this park include two frogs, Stumpffia be and Tsingymantis antitra (T. antitra is the only member of its genus and represents a very deep lineage within Madagascar’s endemic frog radiation Mantellidae); a snake, Madagascarophis lolo; and a day gecko, Phelsuma roesleri, not to mention other groups of vertebrates and invertebrates. This incredible diversity can be put down to the bizarre habitat and the isolation that happens almost directly as a consequence of colonisation of it; the limestone karst of Ankarana is riddled with canyons and gullies. Populations in two different canyons can be separated quite effectively, lending selection and drift strong substrate for change over evolutionary time. The microclimate of these habitats is also different from the surrounding forest, which might also provide a source for different selective clines inside and outside of the karst. Due to the difficulty getting in between them to study their biodiversity, many new species doubtless still persist in these areas undetected.

Finally, the study itself is something of a showcase of the direction taxonomy is heading. Calls for ‘integrative taxonomy’ have been taken seriously. Our study is based on the integration of data from pholidosis (scale counting), morphometrics, osteology, and genetics, which together have helped us to begin to tease apart this nightmarish genus, and I expect that these data sources will be invaluable in the complete resolution of the genus. It is something of a triumph to finally be making progress on this genus. It has, after all, been 75 years since the last species was described from this genus, and the youngest valid species—Geckolepis polylepis Boettger, 1893—is 124 years old! It is time for some progress!

I am very pleased to have this paper finally published, two and a half years after starting to work on it. I am also pleasantly surprised at the press reception of the article. We had expected there to be some interest, but I never anticipated that I would be doing interviews with the likes of National Geographic, Discovery, and Science Magazine. I am so happy that these secretive little geckos are getting a bit of exposure—they are strange and interesting creatures that deserve their time in the spotlight.


Hawlitschek, O., Scherz, M.D., Straube, N. & Glaw, F. (2016) Resurrection of the Comoran fish scale gecko Geckolepis humbloti Vaillant, 1887 reveals a disjointed distribution caused by natural overseas dispersal. Organisms Diversity & Evolution, 16, 289–298. 10.1007/s13127-015-0255-1

Köhler, G., Diethert, H.-H., Nussbaum, R.A. & Raxworthy, C.J. (2009) A revision of the fish scale geckos, genus Geckolepis Grandidier (Squamata, Gekkonidae) from Madagascar and the Comoros. Herpetologica, 65, 419–435.

Lemme, I., Erbacher, M., Kaffenberger, N., Vences, M. & Köhler, J. (2013) Molecules and morphology suggest cryptic species diversity and an overall complex taxonomy of fish scale geckos, genus Geckolepis. Organisms Diversity & Evolution, 13, 87–95. 10.1007/s13127-012-0098-y

Scherz, M.D., Daza, J.D., Köhler, J., Vences, M. & Glaw, F. (2017) Off the scale: a new species of fish-scale gecko (Squamata: Gekkonidae: Geckolepis) with exceptionally large scales. PeerJ, 5, e2955. 10.7717/peerj.e2955

Schmidt, W.J. (1911) Beobachtungen an der Haut von Geckolepis und einigen anderen Geckoniden. In: Voeltzkow, A. (Ed.) Reise in Ostafrika in den Jahren 1903-1905 mit Mitteln der Hermann und Elise geb. Hickman Wentzel-Stiftung ausgeführt: Wissenschaftliche Ergebniss von Alfred Voeltzkow. Schweizerbart’sche Verlagsbuchhandlung, Stuttgart, pp. 331–352.

Schubert, C. & Christophers, E. (1985) “Dermolytische Schreckhäutung” – ein besonderes Autotomieverhalten von Geckolepis typica (Reptilia, Gekkonidae). Zoologischer Anzeiger, 214, 129–141.

Schubert, C., Steffen, T. & Christophers, E. (1990) Weitere Beobachtungen zur “dermolytischen Schreckhäutung” bei Geckolepis typica (Reptilia, Gekkonidae). Zoologischer Anzeiger, 224, 175–192.

Leave a comment

Your email address will not be published. Required fields are marked *

One thought on “A new fish-scale gecko from Madagascar

  • Anonym

    Really nice blog post, much better than the newspaper articles! 🙂 Thanks for sharing your personal as well as scientific experiences, Mark.