Accepted by H. Zaher: 16 Sept. 2013; published: 17 Oct. 2013
ISSN 1175-5326 (print edition)
1175-5334 (online edition)
Copyright © 2013 Magnolia Press
Zootaxa 3721 (5): 455–474
A new giant Atractus (Serpentes: Dipsadidae) from Ecuador, with notes on some
other large Amazonian congeners
WALTER E. SCHARGEL
, WILLIAM W. LAMAR
, PAULO PASSOS
, JORGE H. VALENCIA
DIEGO F. CISNEROS-HEREDIA
& JONATHAN A. CAMPBELL
Department of Biology, The University of Texas at Arlington, Arlington, TX 76019, USA
Department of Biology, The University of Texas at Tyler, 3900 University Blvd., Tyler, TX 75799, USA
Departamento de Vertebrados, Museu Nacional, Universidade Federal do Rio de Janeiro, Quinta da Boa Vista, Rio de Janeiro, RJ,
Fundación Herpetológica Gustavo Orces, Av. Amazonas 3008 y Rumipamba, Casilla 17 03 448, Quito, Ecuador
Museo de Zoología, Pontificia Universidad Católica del Ecuador, Av. 12 de Octubre y Roca, Quito, Ecuador
Colegio de Ciencias Biológicas y Ambientales, Universidad San Francisco de Quito, Calle Diego de Robles y Vía Interoceánica,
Corresponding author: firstname.lastname@example.org
We describe a new species of Atractus from Cordillera de los Guacamayos in the Andes of Ecuador. This new species is
the largest known species of Atractus, reaching almost 120 cm in total length with a robust habitus. We also use multivar-
iate statistical analyses of morphometric data to look into the taxonomic confusion involving other large, banded/blotched,
species of Atractus in Western Amazonia. We show that A. snethlageae has a widespread distribution in Amazonia and
has been repeatedly confused with A. major in Ecuador owing to its color polymorphism. Our multivariate statistical anal-
yses support previous suggestions to recognize A. snethlageae as a distinct species relative to A. flammigerus. Taxonomic
accounts are provided for both A. major and A. snethlageae including detailed color pattern descriptions. We also find that
there are no valid morphological differences to support recognizing A. arangoi as a separate species from A. major; con-
sequently we synonymize the former name with the latter.
Key words: Atractus major, Atractus snethlageae, Atractus touzeti, Cordillera de Los Guacamayos, rainforest, reptiles,
The genus Atractus consists of generally small, semi-fossorial or cryptozoic snakes that feed mostly on earthworms
and other small invertebrates (Martins & Oliveira 1993, 1999). The genus is distributed from central Panama
(Myers 2003) to northeastern Argentina (Giraudo & Scrocchi 2000); ranging across a wide variety of habitats from
lowland rainforest and Neotropical savannas to cloud forest and páramos in the Andes and even inhabiting the
summits of some tepuis (the table-top mountains in the Guayana Region). Currently more than 130 species are
recognized in Atractus (Passos et al. 2012), which makes it the most species-rich genus of snakes in the world. The
taxonomy of this genus, however, remains in a confused state especially in regard to species boundaries. Several
factors have made taxonomic studies of Atractus a difficult task. The main problem seems to be the scarcity of
specimens in collections, which may reflect the difficulty of collecting these snakes because of their secretive
habits (Downs 1967; Myers 2003). Consequently, many species of Atractus are known from only a few specimens.
To complicate matters, the original descriptions of many of the species in this genus are basic and lack a proper
diagnosis. Regrettably, most of the work done so far to expand our knowledge on the variation of these species
covers only a fraction of the genus.
The first modern taxonomic review of Atractus was undertaken by Savage (1960) for Ecuador. This work
SCHARGEL ET AL.
456 · Zootaxa 3721 (5) © 2013 Magnolia Press
remains one of the most important contributions to the taxonomy of the genus. Savage (1960) described the
variation for the 16 species that were known in Ecuador; he also proposed three putatively monophyletic groups
within the genus and provided an account of all the names (including synonyms) that were available at the time.
Subsequently, other reviews have been published for Venezuela (Roze 1961), Suriname (Hoogmoed 1980),
Panama (Myers 2003), northeastern Argentina (Giraudo & Scrocchi 2000) and particular regions of Brazil (Cunha
& Nascimento 1983; Martins & Oliveira 1993; Passos et al. 2010c) and Colombia (Silva Haad 2004; Passos et al.
2009b; Passos & Lynch 2011). The fact that many of these reviews have been restricted to politically defined areas
(e.g. country, state), which in some cases are relatively small, has posed limitations to the definition of species
boundaries, especially in widespread species complexes (e.g. Amazon species of Atractus). Therefore, a complete
taxonomic review of Atractus is still warranted.
Although the great majority of species of Atractus are quite small, rarely exceeding 40 cm, a few have been
found that approximate or even exceed one meter in total length (e.g. A. gigas, A. obesus and A. trihedrurus). These
so called “giant” Atractus tend to also have a robust habitus and in some cases are known from only a handful of
specimens. In this study we report on yet another giant species of Atractus known from specimens collected in the
Cordillera de los Guacamayos in Ecuador. This species, although extremely large in size relative to other Atractus,
has a color pattern (brown background with transverse bands) and scalation similar to other relatively large species
of Atractus that have been found in cis-Andean Ecuador as well as other regions of Amazonia. Several species
names have been used to refer to these species in Western Amazonia, including A. badius, A. flammigerus, A.
major and A. snethlageae. While examining specimens and the literature concerning these other species we
realized that much confusion still persists about their taxonomy. In this study we use statistical analyses of
morphological data to clarify the confusion regarding these species, as well as to provide support for the new giant
species of Atractus.
Material and methods
Although we have collectively examined a vast number of specimens of Atractus, belonging to many different
species, we list in the appendix only those specimens that were used for the purpose of statistical analyses and/or
for documenting morphological variation reported in the species accounts. These specimens are deposited in the
following collections: UNITED STATES: American Museum of Natural History (AMNH); California Academy of
Sciences (CAS); Carnegie Museum of Natural History (CM); Field Museum of Natural History (FMNH); Florida
Museum of Natural History (UF); Louisiana Museum of Natural History (LSUMZ); Museum of Comparative
Zoology (MCZ), Museum of Vertebrate Zoology (MVZ); Natural History Museum of Los Angeles County
(LACM); National Museum of Natural History (USNM); Museum of Zoology University of Michigan (UMMZ);
University of Kansas (KU); The University of Texas at Arlington (UTA); Texas Cooperative Wildlife Collection
(TCWC). COLOMBIA: Museo de Historia Natural, Universidad de La Salle (MLS), Bogotá D.C. VENEZUELA:
Colección de Vertebrados de la Universidad de Los Andes (CVULA), Mérida; Museo de Ciencias Naturales de
Guanare (MCNG), Guanare. ECUADOR: Fundación Herpetológica Gustavo Orcés (FHGO); Museo de Zoología,
Pontificia Universidad Católica del Ecuador (QCAZ), Museo Ecuatoriano de Ciencias Naturales (MECN).
NETHERLANDS: National Natuurhistorisch Museum in Leiden (RMNH).
For the purpose of statistical analysis we recorded for each specimen data for 18 quantitative external
morphology variables, which included the following body and head plate measurements (abbreviations in
parenthesis): Total length (TL), head length (HL), head width (HW), frontal length (FL), frontal width (FW),
prefrontal suture length (PREFS), loreal length (LORL), loreal height (LORH), parietal length (PARL), parietal
width (PARW), rostral height (ROSTH), rostral width (ROSTW), Internasal width (INW), internasal suture (INS),
chinshield length (CHINL), eye diameter (EYED), eye-nostril distance (EYENOS) and supraocular length
(SUPRAOC). Total length was measured to the nearest millimeter by stretching the specimens along a metric ruler.
All other measurements were taken with a dial caliper to the nearest 0.1 millimeter and with the aid of a dissecting
scope. Paired head plates were all measured only on the right side. Morphological differentiation between
operational taxonomic units was examined by conducting principal component analyses (PCA) using the
correlation matrix of the data. PCA has the advantage over other statistical techniques in that it condenses the
information contained in a large number of original variables into a smaller set of new composite dimensions,
Zootaxa 3721 (5) © 2013 Magnolia Press · 457
A NEW GIANT ATRACTUS FROM ECUADOR
while making no a priori assumptions about groupings in the data (McGarigal et al. 2000). The results of the PCA
were examined by making scatterplots using the principal components (PC) recovered and grouping the
specimens’ values by other putatively diagnostic characters (e.g. coloration) not included in the data matrix. Many,
possibly all, species of Atractus are sexually dimorphic in relative tail size and segmental counts (ventral and
subcaudal scales) examined separately. However, there seems to be little or no sexual dimorphism in the total
segmental count (the sum of ventral and subcaudal scales), which might indicate that the sexual difference in
relative tail size is mostly accounted by the position of the vent on the body instead of a difference in tail size
relative to all other measurements. Based on this observation we decided to use TL instead of having snout-vent
length (SVL) and tail length as two separate measurements in the analyses. The purpose of doing so was to reduce
the effect of sexual dimorphism while maximizing sample size by combining males and females together in the
statistical analyses. To assess whether this was a valid decision we examined grouping specimens within species by
sex when plotting the principal components in analyses in which TL was used versus SVL and tail length. As
suspected, gender did not have an obvious effect in the analyses (i.e. males and females did not separate in
multivariate space) whenever TL was used, whereas separation by gender was observed whenever SVL and tail
length were included together. All statistical analyses were performed on SYSTAT 11 (SPSS Inc.).
The banded Atractus species confusion in the Amazonian Andean slopes. The first analysis was to evaluate the
identity of the species involved in what previous authors have referred to A. badius, A. flammigerus, A. major and
A. snethlageae in the upper Amazon Basin. Three distinctive color pattern types (Fig. 1) were observed on
specimens for which these four names have been applied in the literature. The typical A. major color pattern type
(I) is a pale to medium brown dorsum with dark brown blotches or bands edged by pale coloration, and agrees with
the pattern stated by Boulenger (1894) when he described the species, as well as with patterns A and B and perhaps
also C defined by Savage (1960) for A. major. Specimens with this color pattern type have also been called A.
badius at least one time relatively recently in the literature (see remarks under A. major). In the second color
pattern type (II) the dorsum is dark brown or dark grey with pale (cream or buff in preservative) bands usually
edged by black, and it agrees with pattern D described by Savage (1960) for A. major. The names A. badius and A.
flammigerus have also been applied to specimens with this color pattern in Peru (see species account for A.
snethlageae). Passos et al. (2010b) also indicated that Savage (1960) included specimens of A. gigas under the
pattern D of A. major. In the third color pattern (III) the dorsum is pale brown to buff with dark brown blotches not
edged by pale coloration, and it seems to agree with pattern E, and maybe also C, described by Savage (1960) for
A. major as well as with the picture provided by Duellman (1978) for A. major. A PCA of specimens with these
three color patterns shows that the first two principal components explained 81.9% and 6.1% of the variance in the
sample, respectively. Because all other components explained independently very little of the variance (<3%) and
did not seem to discriminate between groups, we present results only on the first two components. A scatterplot
with the factor scores for the first two PCs is shown in Figure 2 and the component loadings are shown on Table 1.
The scatterplot shows that specimens with color patterns II and III widely overlap in multivariate space but
specimens with color pattern I form a well-defined cluster that barely overlaps with specimens of the other two
color patterns. Separation of specimens with color pattern I from those with color patterns II and III occurs mostly
along the second principal component axis, with EYED and SUPRAOC being the two variables that had the higher
loadings on PC2. The first principal component does not seem to contribute to the separation observed. However,
when using morphometric data in a PCA the first principal component is generally considered a “size factor”
(Humphries et al. 1981). A plot of TL against the first principal component scores shows that there is indeed a
strong linear relationship between size and the first principal component. Therefore, the differences observed in the
confidence ellipses of color pattern B and C can be mostly explained by the fact that the three largest specimens
between these two groups have color pattern C. In addition to the results of the PCA it was noted that specimens
with color pattern I always have three infralabials in contact with the chinshields on each side whereas specimens
with color pattern II and III always have four. These observations suggest that two species are involved, one of
which is A. major and clearly corresponds with color pattern I.
SCHARGEL ET AL.
458 · Zootaxa 3721 (5) © 2013 Magnolia Press
FIGURE 1. Specimens of large banded/blotched Atractus from the Amazon Basin grouped into three color pattern types (I: top
row, II: middle, III: bottom). Extremes of variation within each color pattern type are shown.
Zootaxa 3721 (5) © 2013 Magnolia Press · 459
A NEW GIANT ATRACTUS FROM ECUADOR
TABLE 1. Component loadings for the first two principal components of a multivariate analysis that included specimens
of Atractus grouped into three different color patterns.
TABLE 2. Component loadings for the first two principal components of a multivariate analysis that included specimens
of Atractus badius, A. flammigerus and A. snethlageae.
The species represented by color patterns II and III is assigned in this work to A. snethlageae because it agrees
well with this species in aspects of coloration (pattern II agrees with the description of the holotype of A.
snethlageae) and meristics. However, the status of A. snethlageae itself has to be reassessed owing to its
nomenclatural history. This taxon was first proposed by Cunha & Nascimento (1983) as a subspecies of A.
flammigerus, in a description based on specimens from Pará, northeastern Brazil. Only three years before the
description of A. snethlageae, Hoogmoed (1980) resurrected A. flammigerus, which at the time was considered a
synonym of A. badius. Hoogmoed’s definition of A. flammigerus included specimens from the Guiana Shield (A.
flammigerus sensu stricto) as well as from Peru (herein A. snethlageae). Vanzolini (1986) elevated A. snethlageae
to full species status but without providing a justification for this decision. Subsequent to these works, the names A.
badius and A. flammigerus have been applied (see species account for A. snethlageae) in western Amazonia to
Variable PC1 PC2
TL 0.946 -0.064
HL 0.983 -0.096
HW 0.919 0.001
FL 0.903 -0.282
FW 0.947 0.174
PREFS 0.963 0.010
PARL 0.965 0.021
PARW 0.944 0.018
LORL 0.965 0.109
LORH 0.769 0.528
ROSTH 0.901 0.063
ROSTW 0.935 0.067
INS 0.780 0.129
INW 0.914 0.011
CHINL 0.892 0.269
EYED 0.736 -0.528
EYENOS 0.978 0.040
SUPRAOC 0.790 -0.548
Variable PC1 PC2
TL 0.946 0.141
FL 0.927 -0.105
FW 0.923 0.309
PARL 0.955 0.206
LORH 0.908 0.188
ROSTW 0.933 0.135
INS 0.888 -0.240
CHINL 0.964 0.170
EYED 0.855 -0.351
SUPRAOC 0.726 -0.639
SCHARGEL ET AL.
460 · Zootaxa 3721 (5) © 2013 Magnolia Press
what we herein consider to be A. snethlageae. In light of this confusion we decided to conduct a PCA (Table 2) of
the morphometric dataset comparing the specimens we refer to A. snethlageae with specimens of both A. badius
and A. flammigerus (sensu stricto). Because of the concern of having a small number of specimens relative to the
number of variables included, we reduced to original dataset to only ten variables following the same procedure as
in the previous analysis. A scatterplot (Fig. 2) of the first two PC (which explain 81.9% and 8.4% of the variance,
respectively) shows that the specimens of A. snethlageae do not overlap in multivariate space with either A. badius
or the two specimens examined of A. flammigerus sensu stricto. These results further support Vanzolini’s (1986)
decision to elevate A. snethlageae to full species status. Other aspects of morphology that distinguish A.
snethlageae from A. flammigerus are discussed under the account for the former species.
FIGURE 2. Results of the multivariate statistical analyses (ellipses always represent 95% confidence intervals). Top left:
Scatterplot of the first two PC of an analysis grouping specimens into color patterns types I (solid dots), II (Xs), and III (open
dots). Top right: Scatterplot of the first two PC of an analysis including specimens of A. snethlageae (open dots) and A. touzeti
(solid dots). Bottom left: scatterplot of TL vs. PC2 in specimens of A. snethlageae (open dots) and A. touzeti (solid dots); fitted
straight line and squared correlation coefficient are for the first species. Bottom right: Scatterplot of the first two PC of an
analysis including specimens of A. badius (solid dots), A. flammigerus (Xs), and A. snethlageae (open dots).
Zootaxa 3721 (5) © 2013 Magnolia Press · 461
A NEW GIANT ATRACTUS FROM ECUADOR
Taxonomic evaluation of a new giant Atractus. The new giant species of Atractus is known from only three
large (>1000 mm in TL) and robust specimens collected from cloud forests (>2000 m) in the Cordillera de los
Guacamayos in Ecuador. The largest specimen in this small series is 1195 mm in TL and represents a size record
for Atractus. The size and robustness of this species is only paralleled by Atractus serranus and A. trihedrurus from
Atlantic Rainforest of southeastern Brazil (see Passos et al. 2010c), and Atractus gigas (Myers & Schargel 2006;
Passos et al. 2010b) from western Ecuador, which was until now the largest species of Atractus, exceeding 1000
mm in SVL.Because the new giant species is most similar in coloration to A. snethlageae, which occurs in close
geographic proximity but at lower elevations, we made an exhaustive comparison between these two species. A
PCA (Table 3) of the new giant species and A. snethlageae shows that the first two PC explain 95.9% and 2.0% of
the variance, respectively. A scatterplot of the first two PC (Fig. 2) shows that the two species included in the
analysis have a large separation in the first PC. However, this was expected because, as mentioned before, the first
PC is highly correlated with size. The difference in size between the two species is quite obvious considering that
the largest specimen of A. snethlageae is 718 mm in TL, but only four of all the specimens examined and reported
in the literature (n= 85; Cunha & Nascimento 1983; Martins & Oliveira, 1993) exceed 500 mm, whereas all three
specimens of the new species exceed 1000 mm in TL. In terms of the second PC there is overlap in factor scores
between the two species yet there seem to be allometric differences between them. For A. snethlageae there is a
strong inverse relationship between the first (a size factor) and second PC, that is, factor scores in the second PC
have the tendency to become smaller as individuals get larger. Although it is not known whether this allometric
relationship is also characteristic of the new species, the three specimens in the analysis have factor scores on the
second PC that are comparable to small specimens of A. snethlageae (Fig. 2). We believe this is strong evidence
supporting the distinctiveness of the new species relative to A. snethlageae; that is, if the specimens of the new
species were just very large individuals of A. snethlageae we would have expected them to have small factor scores
on the second PC consistent with the allometric relationship found in A. snethlageae.
TABLE 3. Component loadings for the first two principal components of a multivariate analysis that included specimens
of A. snethlageae and A. touzeti sp. nov.
The two species also differ in other morphological aspects. The new species has a proportionally much wider
head (HW/HL= 97–112%) than A. snethlageae (HW/HL= 55–89%). In the new species the frontal scale is always
longer than wide (FL/FW = 1.01–1.17) whereas in A. snethlageae the frontal scale is usually wider than long and
rarely as wide as long (FL/FW = 0.75–1.0). All three specimens of the new species have 8/8 supralabials whereas
only two out of 85 specimens (including data from Cunha & Nascimento 1983, and Martins & Oliveira 1993) of A.
snethlageae have 8/8 supralabials, and an additional five specimens have 7/8 or 8/7 supralabials. To assess the
significance of the observed difference in the frequency of supralabials we used the binomial equation to obtain the
probability that three specimens randomly drawn from pooling the data of the two species in question would all
have 8/8 supralabials. Even being conservative (specimens with asymmetry in number of supralabials counted as
having 8/8) the probability of obtaining three specimens with 8/8 supralabials is so low (p<0.001) that we are
confident our data reflect taxon-specific differences in supralabials.
Variable PC1 PC2
TL 0.990 0.008
HW 0.979 0.132
FL 0.975 0.176
PREFS 0.964 -0.233
PARW 0.987 0.110
LORL 0.980 -0.165
LORH 0.970 -0.105
INW 0.989 0.065
EYENOS 0.984 -0.139
SUPRAOC 0.970 0.148
SCHARGEL ET AL.
462 · Zootaxa 3721 (5) © 2013 Magnolia Press
Atractus major Boul
Figs. 1, 3 and 4
Rhabdosoma maculatum—Günther, 1859: 411 (in part).
Rhabdosoma badium—Jan 1862: 13 (in part); Jan & Sordelli 1865: Plate 1, Fig. 1
Atractus major Boulenger 1894: 307; Amaral 1930b: 61; Savage 1960: 47 (in part, see remarks. Lectotype designated); Roze
1961: 114, 1966: 84; Duellman 1978: 229 (in part, see remarks); Dixon & Soini 1986: 95; Pérez-Santos & Moreno 1991:
94 (in part, see remarks); Martins & Oliveira 1993: 28, 1999: 96; Silva Haad 2004: 435; Duellman 2005: 367; Esqueda &
La Marca 2005: 7; Passos et al. 2010b: 76.
Atractus arangoi Prado 1939: 1. Holotype: female, “Colombia” (MLS 136). Daniel 1949: 314 (distribution restricted to Puerto
Asís). [new synonymy, see remarks]
Atractus badius (F. Boie)—Peréz-Santos & Moreno 1988: 68 (in part, pictures 17 and 18 on plate 10).
Lectotype and type locality. Juvenile male, BMNH 19220.127.116.11 (formerly BMNH 18.104.22.168; Fig. 4), collected
by Buckley at locality of Canelos (01°36’S, 77°48’W; ca. 520 m), Province of Pastaza, Ecuador. The lectotype was
designated by Savage (1960), restricting the type locality from eastern Ecuador to “Canelos, Ecuador.”
Diagnosis. A species of Atractus with 17 midbody dorsal scales rows, differing from all congeners by the
following combination of characters: (1) large size, adults reaching almost 800 mm in TL; (2) loreal long (about
three times longer than high); (3) generally seven (rarely six) supralabials with third and fourth in contact with eye;
(4) generally seven (rarely six) infralabials with first three in contact with chinshields; (5) five to seven maxillary
teeth, usually with a single postdiastemal teeth; (6) 163–185 ventral scales in females and 150–173 in males; (7)
27–37 subcaudal scales in females and 29–53 in males; (8) tail of moderate length: Tail/TL 9.4–12.5% in females
and 11.6–18.8% in males; (9) dorsal color pattern consisting of dark brown blotches or bands edged by yellow or
cream on a pale to medium brown ground color.
Comparisons. Atractus major is widely sympatric with A. snethlageae and these two species have been
repeatedly confused with each other. In terms of color pattern A. major has brown crossbands which are darker than
the background coloration and are also edged by yellow or cream. In A. snethlageae the crossbands are typically
paler than the background coloration and are edged by black; less frequently the background color is pale brown
with dark crossbands/blotches, but these are never edged by a different color as in A. major. When confusion
persists, the best way to distinguish between the two species is examining the number of infralabials on each side in
contact with the chinshields and the number of postdiastemal teeth (three infralabials contacting chinshields and a
single postdiastemal tooth in A. major vs. four infralabials and two postsiastemal teeth in A. snethlageae). With
respect to other Amazonian species of Atractus having 17 dorsal scale rows, A. major is similar in aspects of color
pattern and size to A. schach, and A. torquatus. Atractus major differs from A. torquatus by having blotches/
crossbands edged by a pale color (no pale edges on the bloches/crossbands in A. torquatus) and having two
postoculars (as opposed to one postocular). Atractus schach is sympatric with A. major in northcentral Amazonia
but this species lacks pale edges on the dark dorsal blotches/crossbands and it also has four infralabials in contact
with the chinshields and two postdiastemal teeth. Atractus natans is sympatric with A. major in western Amazonia
and specimens of this species oftentimes have pale-edged dark blotches; however, A. natans is a much smaller
species with adults rarely exceeding 400 mm in TL, which, as opposed to A. major, has a mostly dark (dark grayish
brown to black) venter (mostly cream in
jor) and four infralabials in contact with the chinshields.
Color pattern. The dorsal ground color is pale to medium brown or grayish brown, with irregular or
ellipsoidal dark brown, primary blotches or crossbands narrowly edged by pale coloration, which can be pale
brown, cream or yellow. The first dorsal blotch is usually elongated forming a short middorsal stripe on the neck,
with separate, irregular blotches on the sides. The size, shape, extent on the lateral regions, and distance between
the blotches/crossbands is variable but they generally expand longitudinally on the middorsal region. Typically the
blotches/crossbands occupy a smaller area than the interspace between them, but the two specimens examined from
Bolívar State, Guayana region of Venezuela, are notable exceptions in which the bands are at least twice as large as
the interspace between them. In some specimens the blotches are in series of two and do not contact each other
either because they do not reach the vertebral scale row (typical condition in specimens from Táchira, Venezuela)
or, if they do, they do not coincide with each other. Most individuals also have smaller, irregular secondary
blotches on the lower half of the sides. These secondary blotches are the same color as the primary blotches and
Zootaxa 3721 (5) © 2013 Magnolia Press · 463
A NEW GIANT ATRACTUS FROM ECUADOR
might alternate with them or they might appear as broken continuations of the primary blotches on the sides. The
first two scales rows are typically paler than the dorsal background coloration but they are often also overlaid with
small, irregular dark spots and mottling. The head is usually darker than the dorsal background color. The venter is
also variable in color pattern, but generally the background color is cream with brown spots and/or blotches. In
some specimens the blotches are located contiguously across scales forming broken ventral stripes, in which case
the most common pattern is a single midventral stripe.
FIGURE 3. Photographs of individuals in life of A. major (top left: Amazonas, Venezuela; top right: Bolívar, Venezuela;
bottom left: Ecuador) and A. snethlageae (bottom right: Ecuador).
FIGURE 4. Lectotype (BMNH 1922.214.171.124) of A. major.
SCHARGEL ET AL.
464 · Zootaxa 3721 (5) © 2013 Magnolia Press
Distribution. Amazon Rainforests of Ecuador, Colombia, Peru, Venezuela, and Brazil.
Remarks. Savage (1960) redescribed A. major based on a significant number of specimens from Ecuador;
however, he erroneously included many specimens of A. snethlageae as A. major in his work. Color pattern D, and
most likely color pattern E, as well as some of the specimens allocated to color pattern B, described by Savage
(1960) for A. major, represent specimens of A. snethlageae. Duellman (1978) & Pérez-Santos & Moreno (1991)
committed the same mistake apparently following Savage (1960); that is, their accounts for A. major include
specimens of A. snethlageae. Moreover, the pictures referred as A. major in Duellman (1978), figure 138, page
229, and in Pérez-Santos & Moreno (1991), Photo 59, page 484, are indeed A. snethlageae. Duellman (2005) later
realized his mistake and indicated that he had included specimens of A. flammigerus (= A. snethlageae) in his
earlier account of A. major in Amazonian Ecuador (Duellman 1978).
Prado (1939) described A. arangoi from Colombia without specifying the type locality. Daniel (1949)
mentioned that this species is known from Puerto Asís, Department of Putumayo, southeastern Colombia. Prado
(1939) only compared A. arangoi with A. major, which he indicated was related, but in his view A. arangoi differed
in color pattern, having a smaller size, and fewer ventrals and subcaudals. However, all the putative diagnostic
characters for A. arangoi fall within the variation in A. major as herein defined. The examination of the holotype of
A. arangoi (Fig. 5) by one of us (PP) has confirmed an agreement with A. major in all other examined
morphological features that were not reported in the original description of the species. The holotype of A. arangoi
is a female (MLS 136), 373 mm in TL, tail is 15.5% of TL. The variation is standard meristic characters is: 161
ventrals, two preventrals, three gulars separating chinshields from first preventral, 32 subcaudals, 6/7 supralabials,
seven infralabials (first three in contact with chinshields), 6/5 maxillary teeth. The dorsal color pattern consists of
rsal, pale bordered, dark brown blotches on the body and 10 on the tail. Based on all the evidence at hand we
consider A. arangoi to be a junior synonym of A. major.
FIGURE 5. Holotype (MLS 136) of A. arangoi.
Zootaxa 3721 (5) © 2013 Magnolia Press · 465
A NEW GIANT ATRACTUS FROM ECUADOR
Atractus snethlageae Cunha & Nascimento 1983
Figs. 1 and 3
Atractus badius (F. Boie)—Amaral 1930a: 93 (in part), 1930b: 59 (in part); Carrillo & Icochea 1995: 13; Doan & Arizábal
Atractus major Boulenger—Savage 1960: 47 (in part, see remarks); Duellman 1978: 229 (in part, see remarks); Pérez-Santos &
Moreno 1991: 94 (in part, see remarks)
Atractus sp. B—Dixon & Soini 1977: 37.
Atractus flammigerus (F. Boie)—Hoogmoed 1980: 20 (in part, specimen from Peru); Dixon & Soini 1986: 93; Duellman &
Salas 1991: 9; Duellman 2005: 366.
Atractus flammigerus snethlageae Cunha & Nascimento 1983: 19.
Atractus snethlageae—Vanzolini 1986: 25; Martins & Oliveira 1993: 34; Martins & Oliveira 1999: 97; Giraudo & Scrocchi
Holotype and type locality. Adult male, MPEG 10131, collected in October 03, 1976, at locality of Colônia Nova
(01º26’S, 47º32’W), Rio Gurupi, Rodovia BR-316, State of Pará, Brazil.
Diagnosis. A species of Atractus with 17 dorsal scales rows, differing from all congeners by the following
combination of characters: (1) large size, adults reaching about 700 mm in TL; (2) loreal long (about three times
longer than high); (3) generally seven (rarely eight) supralabials with third and fourth (fourth and fifth whenever
eight supralabials are present) in contact with eye; (4) generally eight (rarely seven) infralabials with first four in
contact with chinshields; (5) six to seven maxillary teeth; (6) 151–180 ventral scales in females and 137–165 in
males; (7) 19–28 subcaudal scales in females and 27–45 in females; (8) tail of moderate length: Tail/TL 6.9–
17.6%; (9) dorsal color pattern of dark gray/brown ground coloration with pale crossbands or blotches usually
edged in black, or dorsal ground color pale brown with dark brown or black blotches.
Comparisons. The comparisons and confusions in the literature involving A. snethlageae and A. major are
discussed in the previous species account and will not be repeated here. With respect to Amazonian species with 17
dorsal scales, the frequent color morph of A. snethlageae is rather unique in having well-defined pale crossbands,
typically edged in black, on a dark background. Some specimens of A. latifrons are somewhat similar in color
pattern, but they can be easily distinguished from A. snethlageae in having one postocular (as opposed to two), a
loreal that is less than twice as long as wide (as opposed to twice as long as wide), and typically having six
supralabials (as opposed to typically having seven). Juvenile specimens of A. gigas have a color pattern similar to
A. snethlageae but the two species differ in the number of infralabials in contact with the chinshields (three in A.
gigas, four in A. snethlageae). The less common color pattern, a pale brown background with dark blotches, is
similar to A. schach (see remarks) and A. torquatus. Atractus torquatus has one postocular (as opposed to two in A.
snethlageae) without reported exceptions. Atractus snethlageae is most similar to A. flammigerus (character states
in parenthesis) but differs from this species in lacking keels on dorsal scales (adults have keels on dorsal scales at
the level of the vent), typically seven supralabials (typically eight), and a venter with tiny dark spots forming
irregular diffuse markings (spots are solid, regular, about the size of a dorsal scale and which tend to form
longitudinal stripes). According to the statistical analysis A. snethlageae and A. flammigerus also differ
morphometrically (Fig. 2) with a large separation in PC2. Because SUPRAOC had the highest component loadings
on PC2 we looked into this variable to see if it could be helpful for discriminating the two species. Indeed A.
flammigerus has longer supraoculars than A. snethlageae and this character, for the purpose of being used as
diagnostic, can be presented as SUPRAOC/HL ratio. In A. flammi
gerus this ratio is 0.182-0.188 whereas in A.
snethlageae it is 0.093-0.152.
Color pattern. Two different color pattern types occur in this species with considerable variation within them.
The most common pattern (pattern II in statistical analyses) consists of a dark brown or dark grayish brown ground
coloration with cream or buff yellow dorsal crossbands that extend down to the second or third dorsal scale row on
the sides. The dorsal bands typically expand longitudinally into one or two dorsal scales and are edged by black,
but in some specimens the bands are irregular and form blotches that may expand longitudinally more than two
dorsal scales. Rarely the black edges are inconspicuous. The crossbands may be complete across the dorsum,
interrupted in the vertebral region or alternating on the sides. The interspace between crossbands is longer than the
crossbands themselves, encompassing two to seven dorsal scales but with variation within individuals. The first
two dorsal scales have an irregular pattern of dark and pale mottling and solid cream spots might also be present at
that level. A large white or cream nuchal band is present in juveniles and small adults but gradually darkens and
SCHARGEL ET AL.
466 · Zootaxa 3721 (5) © 2013 Magnolia Press
disappears as individuals get larger. In adults the top of the head is either the same color as the background color of
the body or slightly paler. Irregular and inconspicous dark spots or mottling are frequently observed on the head. In
the other, less common color pattern type (pattern III in statistical analyses) the dorsal background coloration is
pale brown or creamy brown (in preservative) with a series of irregular paired dark brown blotches that may or may
not contact each other middorsally. A dark brown nuchal band is usually present with the head usually darker than
the dorsal background. The ventral coloration in both color pattern types is creamy with varying levels of dark
brown pigmentation in the form of tiny dark spots that may form diffuse blotches. In some specimens the dark
markings can form a broken midventral stripe. Rarely the venter is so heavily pigmented that it is almost uniform
Distribution. Atractus snethlageae as currently defined is a widespread species occurring in northern Brazil,
and the most part of the upper Amazon Basin in Colombia, Ecuador, Peru, Bolivia and Gran Chaco in northern
Argentina (Giraudo and Scrocchi, 2000; Passos 2008). This species inhabits lowland and lower montane
rainforests, from sea level up to 1800 m.
Remarks. As stated above, A. snethlageae has been repeatedly confused with A. major in Ecuador (e.g.
Savage 1960; Duellman 1978; Pérez-Santos & Moreno 1991). We think that part of this confusion stems from the
great amount of intraspecific color pattern variation in A. snethlageae, especially the fact that two distinct color
morphs are found in this species, one of which (pattern III in this study) is similar in general aspects to the color
pattern observed in A. major. Discrete polymorphism in color pattern as documented here for A. snethlageae,
although rare in snakes, has been found in other species of Atractus (Fajardo 2000; Passos & Prudente 2012;
Schargel, unpublished observations) and will likely become a more common observation as we obtain larger
samples of poorly known species and a better understanding of the taxonomy of the genus.
The name A. badius has also been used for specimens of A. snethlageae from Argentina, Bolivia, Brazil,
Ecuador and Peru in the literature or on a significant number of museum specimens that we have examined. This
has been the case in the literature (e.g. Carrillo & Icochea 1995; Doan & Arizábal 2000) even after Hoogmoed
(1980) had redescribed A. badius and resurrected A. flammigerus (which at the time included A. snethlageae) from
its synonymy. As a matter of fact, the name A. badius has historically been a “dumping ground” for several
different species of Atractus having crossbands. Our current understanding of the species together with the
examination of material and records misidentified as A. badius seems to indicate that this species is endemic to the
Guiana Region as delimited by Hoogmoed (1979). Although A. badius likely occurs in southeastern Venezuela,
records of this species for this country in the Coastal and the Mérida Mountain Ranges represent misidentifications
of A. lancini, A. univittatus, and
Comparisons between A. snethlageae and A. schach are currently problematic. The color pattern III of A.
snethlageae is similar to the color pattern that has been reported in A. schach. Atractus schach was originally
described from Suriname and, just like A. flammigerus, had long been considered a synonym of A. badius (see
Hoogmoed 1980). The few works (Cunha & Nascimento 1983, 1993; Martins & Oliveira 1993, 1999) that have
provided taxonomic accounts for both A. schach and A. snethlageae have separated these two species solely on
color pattern. As a matter of fact, before Hoogmoed (1980) resurrected A. schach, Cunha and Nascimeno (1978)
had considered specimens that they examined and which they later referred to A. schach and A. flammigerus
snethlageae (Cunha and Nascimento 1983) as conspecific and under the name A. badius. At this point we have no
evidence from our meristic or morphometric data that would support recognizing the two color patterns that we
have assigned to A. snethlageae in western Amazonia as two different species. However, without a more
comprehensive sampling we cannot establish whether what has been referred to A. snethlageae and A. schach from
Brazil are conspecific or not. If specimens referred to A. schach in Brazil prove to be a different species from A.
snethlageae it would still be necessary to determine whether those specimens are indeed conspecific with topotypic
material of A. schach. As such, we are aware that Atractus snethlageae as herein defined might represent a species
complex and a more in depth taxonomic examination of this species is in progress (Passos in prep.)
Atractus touzeti sp. nov.
Figs. 6 and 7
dult female, FHGO 517, collected dead on road on 16 August 1992 by Peter Pearman at Cosanga–
Archidona road (00°37’S, 77°48’W; 2200 m) in the Cordillera de Los Guacamayos, Province of Napo, Ecuador.
Zootaxa 3721 (5) © 2013 Magnolia Press · 467
A NEW GIANT ATRACTUS FROM ECUADOR
Paratypes. Two adult females, FHGO 2035 and FHGO 2036, collected dead on road on 5 April 1998 by
Felipe Campos at La Virgen (ca. 2000 m) in the Cordillera de Los Guacamayos, Province of Napo, Ecuador.
Diagnosis. A species of Atractus with 17 dorsal scales rows differing from all other members of the genus by
the following combination of characters: (1) large size and thick body, adults reaching more than 1000 mm in TL;
(2) loreal long (about three times longer than high); (3) eight supralabials, four and five in contact with eye; (4)
seven or eight infralabials; (5) eight maxillary teeth; (6) 167–170 ventral scales in females; (7) Tail of moderate
length Tail/TL 9.5–11.4%; (8) 31 subcaudal scales in females; (9) dorsal color pattern of short pale crossbands
edged by black borders that separate the pale color from the brown ground color.
Comparisons. Regarding all known species of Atractus with 17 dorsal scales, only A. gigas, A. latifrons, A.
major, A. obesus, A. sanctamartae, A. serranus, A. snethlageae, A. titanicus, A. torquatus, A. trihedrurus, and A.
zebrinus attain a TL close to or exceeding 600 mm. Adults of A. serranus and A. trihedrurus have a uniform
dorsum, but they also differ from A. touzeti in having fewer ventrals (<164) and subcaudals (<24). In both A.
latifrons and A. torquatus there is a single postocular whereas there are two in A. touzeti. Atractus torquatus further
differs from the new species in having a cream venter with well-spaced small spots or nearly immaculate whereas
in A. touzeti the venter is heavily pigmented with large, roughly rectangular or irregular blotches. Atractus latifrons
further differs from A. touzeti in having a color pattern of rings that encircle the whole body. Atractus major, A.
obesus, A. sanctamartae, A. titanicus, and A. zebrinus all have seven supralabials. Atractus sanctamartae, A.
titanicus and A. zebrinus further differ from A. touzeti in having fewer subcaudals (>28). Atractus major and A.
s further differ from A. touzeti in aspects of dorsal coloration with A. major having a pattern of pale-edged
dark brown blotches or bands, while A. obesus has black rings that encircle the body.
Among the large species of the genus Atractus touzeti is most similar to A. gigas and A. snethlageae. Atractus
gigas is the only species in the genus that is really comparable in size and habitus to A. touzeti, but adults of the
former species are either uniform or with inconspicuous pale bands dorsally whereas the dorsal crossbands in A.
touzeti are conspicuous and well-defined. All three specimens of A. touzeti have eight supralabials whereas
specimens of A. gigas typically have six or seven supralabials (see Passos et al. 2010b). The number of infralabials
that contact the chinshields on each side is four in the two paratypes of A. touzeti but the holotype is asymmetric
and has four on one side and three on the other; all specimens of A. gigas have three infralabials in contact with the
chinshields. Atractus snethlageae is a smaller species that is not known to reach a TL of over 718 mm, it has six to
seven maxillary teeth (eight in A. touzeti), and typically seven infralabials (eight in A. touzeti). The maximum
number of subcaudals observed (including data from Cunha and Nascimento, 1983) in females of A. snethlageae is
28, whereas all three specimens of A. touzeti have 31 subcaudals. Although it is possible that the range of
subcaudals of these two species will overlap when more specimens of A. touzeti are obtained, it is obvious that they
differ in typical values for this character. These two species also differ in other aspects of morphometrics as
explained in the results section; specifically the HW/HL ratio seems to be a useful character for separating them.
Among small species of Ecuadorian Atractus (<50 cm TL), all of them except A. multicinctus have dorsal color
patterns other than crossbands, usually being striped, spotted or uniformly colored. Atractus multicinctus further
differs from A. touzeti by having fewer maxillary teeth (5 or 6) and supralabials (7).
tion of the holotype. An adult female, 1195 mm in TL, tail length 114 mm (9.5% of TL); head slightly
distinct from neck, as wide as long (HW: 28.8 mm; HL: 28.6 mm). Snout truncated in dorsal view, rounded in
lateral view; eye moderately small (2.9 mm), about same size as the upper postocular, pupil circular; eye-nostril
distance 0.29 of HL; 2.8 times eye diameter; rostral bell-shaped, barely visible in dorsal view, 1.6 times broader
than high, lingual groove reduced; internasals small, as long as wide, laterally contacting anterior and posterior
nasals; prefrontals large, 1.2 times longer than wide, in contact with eye; frontal roughly triangular, 1.2 longer than
wide, 1.5 times longer than median suture of prefrontals; parietals 1.6 times longer than wide, median suture of
parietals 0.8 times length of frontal. Nasal divided, posterior scale twice the size of the anterior; preocular absent;
loreal 2.5 times longer than high, narrowly touching the eye, anterior edge 2.2 times higher than posterior; two
postoculars, lower very small, upper 6 times larger; temporals 1 + 2, upper posterior temporals elongated, as large
as the frontal; supralabials eight on both sides, first contacting nasals, second contacting posterior nasal and loreal,
third and fourth contacting loreal, fourth, fifth and sixth in contact with eye on the right side, fourth and fifth in
contact with eye on left side, sixth contacting postocular and anterior temporal, seventh contacting anterior and
posterior temporals, eighth contacting posterior temporal and dorsals; infralabials eight on the right side seven on
the left, four contacting the chinshields on right side, three on left side, first pair short, meeting at ventral midline,
SCHARGEL ET AL.
468 · Zootaxa 3721 (5) © 2013 Magnolia Press
separating mental from chinshields; one pair of chinshields, each 2.3 times longer than wide. Maxillary teeth eight,
most teeth bearing a prominent longitudinal ridge on the labial side; palatine teeth large; dorsal scales in 17–17–17
rows, smooth without apical pits; ventrals 170, preventrals four; anal plate single; subcaudals 31, paired.
FIGURE 6. Holotype (FHGO 517) of A. touzeti.
Color pattern of the holotype. In preservative (alcohol 70% after formalin), the dorsal ground color is hair
brown with dark and cream mottling on dorsal scale rows 1–3; there are 34 dorsal crossbands along the body, seven
on the tail, extending laterally to the first dorsal scale row except for the first two which extend to dorsal scale row
six (blotches). The crossbands are tawny olive, one dorsal scale long, edged by black borders of about the same
length, separated from each other by two to three dorsal scale lengths, they become indistinct and darkly mottled on
the first and/or second dorsal scale rows; they alternate with small lateral blotches of the same color that cover
partially the first and second dorsal scale rows on each side. The dorsum of the head is olive brown, slightly paler
than the dorsum of the body, with some inconspicuous dark brown spots; an incomplete short dark band is present
in the nuchal region. Supralabials have cream spots with dark mottling in the lower portion. Infralabials, mental
and chinshields are mostly dark brownish olive with cream spots. The venter is cream heavily pigmented with large
rectangular and irregular dark brownish olive blotches.
Variation. The two paratypes are adult females with TL of 1050 and 1030 mm, and tail lengths of 120 and 115
mm, respectively comprising 11.4 and 11.2% of TL. The cephalic index (width/length x 100) is 101–120 (mean =
114, n = 3). There are eight supralabials (both sides) with the fourth and fifth entering the orbit in both specimens;
infralabials are eight in FHGO 2035 and seven (right side) and eight (left side) in FHGO 2036, with four in contact
with the chinshields. In the paratypes the upper posterior temporal is not as large as in the holotype, and is about the
same size as the first temporal. The upper postocular is about twice the size of the lower postocular in both
specimens. There are 169 and 167 ventrals and 31 subcaudals in the paratypes. Other aspects of lepidosis agree
well with the holotype. Both paratypes have eight maxillary teeth on the right side, reducing in size posteriorly.
Palatine and pterygoid teeth in FHGO 2035 are seven and five, respectively, on each side. We could not count teeth
in the other specimens because of damage to dentigerous bones (FHGO 2036) or the specimen was sufficiently
brittle as to preclude opening the mouth without risking breaking the jaw (holotype). The numbers of dorsal
crossbands in the paratypes are 34 and 32 along the body, and seven and eight on the tail, respectively. In the
holotype the percentage of the venter covered with dark blotches is almost 50%; in the paratypes this percentage is
lower, about 35–40%, and the color of the blotches is dark neutral gray.
Color in life. Although we do not have notes about the color in life of the type series of A. touzeti, we have
received a picture taken in situ of a live individual found in the same locality from which the paratypes were
obtained. The specimen was found crossing a road and was not collected but its large size and robust habitus are
apparent in the picture. The background color of the dorsum is dark brown (darker tone than what was observed in
the preserved specimens), becoming almost black towards the middorsal region. The crossbands are conspicuous
Zootaxa 3721 (5) © 2013 Magnolia Press · 469
A NEW GIANT ATRACTUS FROM ECUADOR
and bright yellow on the side but become suffused with brown color towards the middorsal region. The black
borders of the crossbands are not distinct in this specimen. There are about 42 crossbands total from neck to tail.
FIGURE 7. Dorsal (top) and lateral (bottom) view illustrations of the head of the holotype (FHGO 517) of A. touzeti.
Etymology. We take great pleasure in naming the new species after Jean-Marc Touzet, an enthusiastic
promoter of herpetology in Ecuador and the founder of the Fundación Herpetológica Gustavo Orcés.
Distribution. Known only from the Cordillera de Los Guacamayos in the Eastern Andean range of Ecuador
(Fig. 8). According to the recent classification of ecosystems in Ecuador (Ministerio del Ambiente de Ecuador,
2013) the area where the type locality is located is part of the “evergreen montane forest of the north and central
Cordillera Oriental of the Andes.” The vegetation in this ecosystem is constituted mostly by evergreen forests
reaching 20-25 m in height, dominated by Andean species in the families Melastomataceae (Miconia), Solanaceae,
Myrsinaceae, Aquifoliaceae, Araliaceae, Rubiaceae and several families of ferns. Valencia (1995) studied the
composition and structure of a forest near Baeza (2000 m), Napo, Ecuador, which is about 20 km north of the type
locality of A. touzeti and found that the vegetation was composed of a combination of typical Amazonian and
Andean species, with trees reaching 30 m tall and the understory being dominated by the palm tree Geonoma
SCHARGEL ET AL.
470 · Zootaxa 3721 (5) © 2013 Magnolia Press
FIGURE 8. Map of Ecuador indicating type locality (solid circle) of A. touzeti. The two localities in which the species has been
collected are nearby and cannot be separated on the map.
Taxonomic research of the genus Atractus has been traditionally a difficult task because of the plethora of species
described and the fact that many species remain poorly characterized and are known from only a few specimens. To
further complicate matters, it is now clear that intraspecific color pattern variation within Atractus can be complex
and include discrete polymorphism (Fajardo 2000; this work), sexual dimorphism (Passos et al. 2009a) and drastic
ontogenetic changes (Passos et al. 2010c; Passos & Prudente 2012). These aspects of color pattern variation have
confused taxonomists working with Atractus even when relatively large series have been available (e.g. Savage
1960; Bernal-Carlo & Roze 1997). Although color pattern is still an important and useful character system for
Zootaxa 3721 (5) © 2013 Magnolia Press · 471
A NEW GIANT ATRACTUS FROM ECUADOR
taxonomic research in Atractus, we stress that it should not be used in isolation to support taxonomic decisions.
Because of the generally conservative nature and overlap in traditional meristic and scalation characters within
Atractus it is important to include alternative morphological characters in taxonomic studies. For example,
hemipenial morphology has been shown to be quite variable in the genus (Schargel & Castoe 2003) and
informative to establish relatedness and to separate species (Passos et al. 2010a, 2010c; Prudente & Passos 2010).
Herein we also show that morphometric data, mostly from the head, used in the context of multivariate statistical
analyses (e.g. PCA; see also Schargel 2003) can significantly help support taxonomic decisions. Even if species are
closely related, as seems to be the case for A. flammigerus, A. snethlageae and A. touzeti, morphometric data can
still be quite informative. The only important limitation that we see for using multivariate statistical techniques on
morphometric data to delimit species in Atractus is the relatively large sample size required to be able to include
several variables in the analyses and to understand intraspecific variation. As mentioned before, many species of
Atractus remain known from very few specimens.
Atractus touzeti is now the largest known species of Atractus and it is paralleled in size only by A. gigas from
western Ecuador. These two species are so large relative to other species in the genus that they almost double the
size of species that are considered large (e.g. A. major, A. torquatus) and they quadruple the size of the smaller
species in the genus. Interestingly, neither one of these two giant species is atypical in other aspects of external
morphology. Not even segmental counts (e.g. ventral scales), which tend to be correlated with size in snakes
(Lindell 1994; Head & Polly 2007), are noticeably higher relative to other species in the genus. This observation
further supports the notion that gigantism in snakes is achieved by modification of post-somitogenic somatic
growth rather than with an increase in somite numbers (i.e. pleomerism; Head & Polly 2007).
The discovery of such a large species of Atractus underlines the fact that we are still in a phase of big
discoveries with respect to the diversity of this genus. The amount of new species discoveries in Atractus as well as
the number of more revisionary-oriented taxonomic work has been unprecedented in the last decade. Regardless, it
is still clear that a whole lot more work is required before we can obtain a solid taxonomic framework for
understating other aspects of the diversification of this genus. Even if several recent efforts have focused on
resolving the taxonomic chaos involving some regions (e.g. Andes of Colombia) from which several species were
poorly known, our taxonomic approaches are still limited by the fact that pretty much nothing concrete is known
about the phylogenetic relationships within Atractus. This limitation has prevented those of us working on the
taxonomy of this complicated genus from framing our research projects around hypotheses of species relatedness,
as opposed to emphasizing our taxonomic revisions on politically defined areas or, at best, biogeographical
provinces. The large number of species is certainly intimidating to anyone up to the task of looking into the
phylogenetic relationships of Atractus, which will likely have to rely largely on molecular data. However, we are
reaching a point at which any phylogenetic information of Atractus, even if preliminary, would certainly be
welcomed and might even have a major impact on taxonomic research if at least major clades can be identified and
We are most grateful to all the museum curators and collection managers that allowed us to examine specimens
under their care: Darrel Frost and Charles Myers (AMNH), Robert Drewes and Jens Vindum (CAS), Amelia Díaz
(CVULA), Harold Voris and Alan Resetar (FMNH), Katty Garzón (FHGO), Linda Trueb (KU), David Kizirian
(LACM), Juan García-Pérez (MCNG), James Hanken and José Rosado (MCZ), Mario Yánez-Muñoz (MECN),
David Wake (MVZ), Laurie Vitt (OMNH), Omar Torres-Carvajal (QCAZ), Marinus Hoogmoed and Pim Arntzen
(RMNH), Lee Fitzgerald and James Dixon (TCWC), Ronald Nussbaum and Greg Schneider (UMMZ), Roy
McDiarmid (USNM), Brother Roque Casallas and Arturo Rodriguez (MLS), Ana Prudente (MPEG), and Collin
McCarthy (BMNH). Cesar Barrio allowed us to examine specimens he collected before they were incorporated
into CVULA. For letting us use their pictures we are grateful to Cesar Barrio, Luis Rodriguez and Eric Smith. For
help with statistical analyses we are grateful to Jesse Meik. Special thanks are due to Miguel Castañel and Pablo
Tomás Caiza for sharing with us a picture of a live individual of A. touzeti.
SCHARGEL ET AL.
472 · Zootaxa 3721 (5) © 2013 Magnolia Press
Amaral, A. (1930a) “1929” Contribuição ao conhecimento dos ophidios do Brasil IV. Lista remissiva dos ophidios do Brasil.
Memorias do Instituto Butantan, 4, 69–126. [The year printed on the article is added in quotes if it is different from the
actual publication year]
Amaral, A. (1930b) “1929” Estudos sobre ophidios neotropicos XVIII. Lista remissiva dos ophidios da Região Neotropica.
Memorias do Instituto Butantan, 4, 127–271.
Bernal-Carlo, A. & Roze, J.A. (1997) Snake genus Atractus (Colubridae) from Sierra Nevada de Santa Marta, Colombia, with
a description of a new species. Bulletin of the Maryland Herpetological Society, 33 (4), 165–170.
Boulenger, G.A. (1894) Catalogue of the snakes in the British Museum (Natural History), vol. 2. Trustees of the British
Museum, London, 382 pp.
Carrillo, N. & Icochea, J. (1995) Lista taxonómica preliminar de los reptiles vivientes del Perú. Publicaciones del Museo de
Historia Natural, Universidad Nacional Mayor de San Marcos, 47, 1–27.
Cunha, O.R. & Nascimento, F.P. (1978) Ofídios da Amazonia. X. As cobras da região lest do Pará. Publicações Avulsas do
Museu Paraense Emílio Goeldi, 31, 1–218.
Cunha, O.R. & Nascimento, F.P. (1983) Ofídios da Amazonia. XX. As espécies de Atractus Wagler, 1928 na Amazônia oriental
e Maranhão (Ophidia: Colubridae). Boletim do Museu Paraense Emílio Goeldi, Nova série, Zoologia, 128, 1–39.
Cunha, O.R. & Nascimento, F.P. (1993) Ofídios da Amazonia. As cobras da região leste do Pará. Boletim do Museu Paraense
Emílio Goeldi, Série Zoologia, 9, 1–39.
Daniel, H. (1949) Las serpientes en Colombia. Revista Facultad Nacional de Agronomía de Medellín, 9, 301–333.
Dixon, J.R. & Soini, P. (1977) The reptiles of the upper Amazon basin, Iquitos region, Peru. II. Crocodilians, turtles and snakes.
Milwaukee Public Museum Contributions in Biology and Geology, 12, 1–91.
Dixon, J.R. & Soini, P. (1986) The reptiles of the upper Amazon basin, Iquitos region, Peru. Milwaukee Public Museum,
Milwaukee, 154 pp.
Doan, T.M. & Arizabal, W. (2000) New herpetological records for the Tambopata Province, Department of Madre de Dios,
Peru. Herpetological Review, 31 (3), 188–189.
Downs, F.L. (1967) Intrageneric relationships among Colubrid snakes of the genus Geophis Wagler. Miscellaneous
Publications Museum of Zoology, University of Michigan, 131, 1–193.
Duellman, W.E. (1978) The biology of an equatorial herpetofauna in Amazonian Ecuador. Miscellaneous Publication,
University of Kansas Museum of Natural History, 65, 1–352.
Duellman, W.E. (2005) Cusco Amazónico: the lives of amphibians and reptiles in an Amazonian rainforest. Comstock Books in
Herpetology, Cornell University Press, Ithaca, 433 pp.
Duellman, W.E. & Salas, A.W. (1991) Annotated checklist of the amphibians and reptiles of Cuzco Amazónico, Peru.
Occasional Papers of the Museum of Natural History University of Kansas, 143, 1–13.
Esqueda, L.F. & La Marca, E. (2005) Revisión taxonómica y biogeográfica (con descripción de cinco nuevas especies) de
serpientes del género Atractus (Colubridae: Dipsadinae) en los Andes de Venezuela. Herpetotropicos, 2 (1), 1–32
Fajardo, A.M. (2000) Revisión de las serpientes del género Atractus, (Serpientes-Colubridae) en la sabana de Bogotá.
Unpublished thesis. Pontificia Universidad Javeriana, Facultad de Ciencias, Bogotá, 69 pp.
Giraudo, A.R. & Scrocchi, G.J. (2000) The genus Atractus (Serpentes: Colubridae) in northeastern Argentina. Herpetological
Journal, 10 (3), 81–90.
Günther, A.C.L.G. (1859) Second list of cold-blooded Vertebrata collected by Mr. Fraser in the Andes of western Ecuador.
Proceedings of the Zoological Society of London, 1859, 402–427.
Head, J.J. & Polly, P.D. (2007) Dissociation of somatic growth from segmentation drives gigantism in snakes. Biology Letters,
3 (3), 296–298.
Hoogmoed, M.S. (1979) The herpetofauna of the Guianan Region. In: Duellman, W.E. (Ed.), The South American
herpetofauna: Its origins, evolution, and dispersal. University of Kansas Museum of Natural History Monograph 7,
Lawrence, pp. 241–279.
Hoogmoed, M.S. (1980) Revision of the genus Atractus in Surinam, with the resurrection of two species (Colubridae, Reptilia).
n the herpetofauna of Surinam VII. Zoologische Verhandelingen, 175, 1–47
Humphries, J.M., Bookstein, F.L., Chernoff, B., Smith, G.R., Elder, R.L. & Poss, S.G. (1981) Multivariate discrimination by
shape in relation to size. Systematic Zoology, 30 (3), 291–308.
Jan, G. (1862) Prodromo dell’iconografia generale degli ofidi. II. Parte V. Gruppo: Calamaridae. Archivio Zoologia, Genoa,
Jan, G. & Sordelli, F. (1865) Iconographie generale des ophidiens, livraison 11. Libraires de L’Académie Impériale de
Médicine, Milan, 6 pp.
Lindell, L.E. (1994) The evolution of vertebral number and body size in snakes. Functional Ecology, 1994 (8), 708–719.
Martins, M. & Oliveira, M.E. (1993) The snakes of the genus Atractus Wagler (Reptilia: Squamata: Colubridae) from the
Manaus region, central Amazonia, Brazil. Zoologische Mededelingen, 67 (2), 21–40.
Martins, M. & Oliveira, M.E. (1999) Natural history of snakes in forests of the Manaus region, Central Amazonia, Brazil.
Herpetological Natural History, 6 (2), 78–150.
Zootaxa 3721 (5) © 2013 Magnolia Press · 473
A NEW GIANT ATRACTUS FROM ECUADOR
Ministerio del Ambiente de Ecuador. (2013) Sistema de Clasificación de los Ecosistemas del Ecuador Continental. Secretaria
de Patrimonio Natural, Quito, 205 pp.
McGarigal, K., Cushman, S. & Stafford, S.G. (2000) Multivariate statistics for wildlife and ecology research. Springer, New
York, 283 pp.
Myers, C.W. (2003) Rare snakes—Five new species from eastern Panama: Reviews of northern Atractus and Southern Geophis
(Colubridae: Dipsadinae). American Museum Novitates, 3391, 1–47.
Myers, C.W. & Schargel, W.E. (2006) Morphological extremes—two new snakes of the genus Atractus from northwestern
South America (Colubridae: Dipsadinae). American Museum Novitates, 3532, 1–13.
Passos, P. (2008) Revisão taxonômica do gênero Atractus Wagler, 1828 (Serpentes: Colubridae: Dipsadinae). Unpublished
Ph.D. Thesis, Museu Nacional, Universidade Federal do Rio de Janeiro, Rio de Janeiro, 671 pp.
Passos, P. & Lynch, J.D. (2011) “2010” Revision of Atractus from upper and middle Magdalena drainage of Colombia.
Herpetological Monographs, 24, 149–173.
Passos, P. & Prudente, A.L.C. (2012) Morphological variation, polymorphism and taxonomy of the Atractus torquatus complex
(Serpentes: Dipsadidae). Zootaxa, 3407, 1–21.
Passos, P., Cisneros-Heredia, D.F., Rivera, D., Aguilar, C. & Schargel, W.E. (2012) Rediscovery of Atractus microrhynchus and
reappraisal of the taxonomic status of A. emersoni and A. natans (Serpentes: Dipsadidae). Herpetologica, 68 (3), 375–392.
Passos, P., Chiesse, A., Torres-Carvajal, O., & Savage, J.M. (2010a) “2009” Testing species boundaries within Atractus
occipitoalbus complex (Serpentes: Dipsadidae). Herpetologica, 65 (4), 284–403.
Passos, P., Dobiey, M. & Venegas, P.J. (2010b) Variation and natural history notes on giant groundsnake, Atractus gigas
(Serpentes; Dipsadidae). South American Journal of Herpetology, 5 (2), 73–82.
Passos, P., Fernandes, R., Bérnils, R.S. & Moura-Leite, J.C. (2010c) Taxonomic revision of the Brazilian Atlantic Forest
Atractus (Reptilia: Serpentes: Dipsadidae). Zootaxa, 2364, 1–63.
Passos, P., Lynch, J.D. & Fernandes, R. (2009a) “2008” Taxonomic status of Atractus sanctaemartae and A. nebularis, and
description of a new species of Atractus from Atlantic coast of Colombia. Herpetological Journal, 18 (4), 175–186.
Passos, P., Mueses-Cisneros, J.J., Lynch, J.D. & Fernandes, R. (2009b) Pacific lowland snakes of the genus Atractus
(Serpentes: Dipsadidae), with description of three new species. Zootaxa, 2293, 1–34.
Pérez-Santos, C. & Moreno, A.G. (1988) Ofidios de Colombia. Museo Regionale di Scienze Naturali di Torino, Monografia VI,
Turin, 520 pp.
Pérez-Santos, C. & Moreno, A.G. (1991) Serpientes de Ecuador. Museo Regionale di Scienze Naturali Torino, Monografia XI,
Turin, 540 pp.
Prado, A. (1939) Notas ofiologicas 4. Cinco especies novas de serpentes colombianas do genero Atractus Wagler. Memorias do
nstituto Butantan, 13, 1–9.
Prudente, A.L.C. & Passos, P. (2010) New cryptic species of Atractus (Serpentes: Dipsadidae) from Brazilian Amazonia.
Copeia, 2010 (3), 397–404.
Roze, J.A. (1961) El género Atractus (Serpentes: Colubridae) en Venezuela. Acta Biologica Venezuelica, 3 (7), 103–119.
Roze, J.A. (1966) La Taxonomia y Zoogeografia de los Ofidios en Venezuela. Ediciones de la Biblioteca UCV, Caracas, 362 pp.
Savage, J.M. (1960) A revision of the Ecuadorian snakes of the colubrid genus Atractus. Miscellaneous Publications of the
Museum of Zoology, University of Michigan, 112, 1–86.
Schargel, W.E. (2003) Taxonomy of certain related and problematic snakes in the genus Atractus (colubridae: dipsadinae) with
comments on hemipenial morphology in the genus. Unpublished M.Sc. Thesis, The University of Texas at Arlington,
Arlington, 90 pp.
Schargel, W.E. & Castoe, T.A. (2003) The hemipenes of some snakes of the semifossorial genus Atractus with comments on
variation within the genus. Journal of Herpetology, 37 (4), 718–721.
Silva Haad, J.J. (2004) Las serpientes colombianas del género Atractus Wagler, 1828 (Colubridae, Xenodontinae) en la
Amazonia Colombiana. Revista de la Academia Colombiana de Ciencias Físicas, Exactas y Naturales, 28 (108), 409–446.
Valencia, R. (1995) Composition and structure of an Andean forest fragment in eastern Ecuador. In: Churchill, S.P., Balslev, H.,
Forero, E. & Luteyn, J.L. (Eds.), Biodiversity and conservation of neotropical montane forests. The New York Botanical
Garden, New York, pp. 239–249.
Vanzolini, P.E. (1986) Levantamento herpetológico da area do estado de Rondônia sob a influência da rodovia BR 364.
Programa Polonoroeste, CNPq, Brasília, Brazil, 50 pp.
SCHARGEL ET AL.
474 · Zootaxa 3721 (5) © 2013 Magnolia Press
APPENDIX. Specimens examined.
Atractus badius: BRAZIL: Amapá: Serra do Navio forest: LACM 44686. FRENCH GUIANA: Cayenne: RMNH 12; road
between Cayenne and Tonate: CAS 146918. SURINAME: Afobaka: RMNH 13574; Powakka: KU 221545; Brownsberg:
UTA 60138; Brownsberg Nature Reserve plateau: MCZ 152691; Djai Creek: RMNH 12860; Nassau Mountains: RMNH
12862; Republiek: RMNH 18677; Paramaribo: RMNH 13773.
Atractus flammigerus.—SURINAME: Nassau Mountains: RMNH 13571, 13572.
Atractus major.—BRAZIL: Rondônia
: Rio Formosa: OMNH 37533. COLOMBIA: Amazonas: Ginogoje, Río Apaporis:
MCZ 53220; Caquetá: Florencia: MCZ 65384; Putumayo
: Puerto Asis: AMNH 53446. ECUADOR: Morona Santiago:
Taisha, Makuma: FHGO 788, 815, 890, 959, 2370, 3435, 3814, 4042, 4043, 4046, 4050, 4496, 4517, 4761, 5717; Taisha,
Amazonas: FHGO 4762; Taisha, Kiim: FHGO 3963; Taisha Paantim: FHGO 4476; road 45 between Puyo to Macas:
QCAZ 10146. Napo
: Alto Sindi: QCAZ 3689; Estación Biológica Jatun Sacha: MCZ 178334, 178335, 173886-173889;
Tena: (FHGO 7627); Jatun Sacha: FHGO 3495. Orellana
: Parque Nacional Yasuní: QCAZ 3079, 6272. Pastaza: -1.12, -
76.85: USNM 321108; Arajuno, Villano: QCAZ 8157, 8265, 8379; Charapacocha: FHGO 3813; Mera: KU 98625,
121307, 133527; Mera, Shell, Fábrica Te Zulay: FHGO 4496. Sucumbios
: Lago Agrio: KU 125991; Pañacocha: QCAZ
7881, Putumayo, Santa Elena, Bohórquez: MECN 8343; Rio Cotopino: UMMZ 89005, 89006, 92032, 92036, 94068.
Santa Cecilia: KU 109830, 109831, 112250, 112251, 121841, 175348, 175399. Zamora Chinchipe
: Zamora: FHGO 4496.
: Pagaat: USNM 316576; Cuzco: San Martin, 5 km N of Camisea River: KU 538472-538475. Madre de
Dios: 15 km E of Puerto Maldonado: KU 207772, 214838-214842; Tambopata Reserve KU 343043. Loreto: Centro
Unión: TCWC 42807; Monte Carmelo, Requena: AMNH 55638, 55640; Monte Alegre, Pachitea: AMNH 52780; Pampa
Hermosa, Cushabatay: AMNH 55709; Quebrada Grande, near junction between Sucusari River and Napo River: KU
: Alto Utuquinia: AMNH 53474; Chiyacu: AMNH 52095; Ollanta: AMNH 52105; Zona Reservada
Pucacuro: FHGO 7552. VENEZUELA: Amazonas
: road Puerto Ayacucho—Gavilan: CVULA 6560-6566, Bolívar: El
Triunfo: CVULA 6567, 6568. Táchira
: La Trampa, Uribante-Caparo: CVULA 4317, 4454, 4727, 4280; Presa La Honda,
Campamento Siberia: MCNG (1 uncatalogued specimen).
Atractus snethlageae.—BRAZIL: Amazonas:
Cuieira River, Between Manaus and Rio Branco: CAS 101607; Manaus: CAS
49797; Presidente Figueiredo: RMNH 26020. BOLIVIA: unknown locality: AMNH 2987. Pando
: Ivón: AMNH 28838;
Senna, near junction with Río Madre de Dios: UMMZ 59778, 59779. ECUADOR: Morona-Santiago
: Chiguaza: USNM
232690, 232691; Méndez: USNM 232692, 232693, 283945, Taisha, Makuma, Tumpaim: FHGO 5614. Napo
: 6.5 k, SE of
Misahualli: QCAZ 3476, 3477; Archidona, Cotundo, Los Cocodrilos: QCAZ 11075; Cascada de San Rafael: QCAZ 3256,
3257; El Chaco, road to Linares: QCAZ 4047; Estación Biológica Jatun Sacha: MCZ 173870, 173871; San Francisco de
Borja, Sardinas: QCAZ 1494, MECN 80; Tena, Sumac Shagcha: FHGO 2178 Orellana
: km 10 Maxus road: QCAZ 10614.
: Abitagua: UMMZ 92033, 94067; Arajuno, Kurintza: QCAZ 8287, Arajuno, Villano: QCAZ 8367; Río Pastaza:
UMMZ 89026; Sarayaku: UMMZ 89024. Sucumbios
: Lago Agrio: KU 125988-125990, 125992-125998; El Reventador:
QCAZ 444; Río Salado, 1 km upstream from Río Coca: KU 164206; Santa Ana: MECN 7873; Santa Cecilia: KU 112249,
125985, 125986, 175397, 175400, MCZ 96674. Tungurahua
: San Francisco de Mapoto UMMZ 89023, 89025. Zamora
Chinchipe: Yantzaza, Los Encuentros: MECN 8437. PERU: Amazonas: Huampami, Río Cenepa: MVZ 163247. Cuzco
40 km E of Quincemil, along road to Puerto Maldonado: LSUMZ 48718; Cashiriari, S of Río Camisea: USNM 538466;
Pagorini, on Río Camisea: USNM 538467; San Martin, 5 km N of Río Camisea: USNM 538464. Loreto
: Paraiso: TCWC
42107; Río Itaya, Iquitos: AMNH 54085, 54257, 54268, 54279, 54368, 54643, 54645, 54667, 54953, 54958; Upper
Amazon: FMNH 11180 Madre de Dios
: Cuzco Amazónico, 15 km E of Puerto Maldonado: KU 214843, 214905, 220191;
Puerto Maldonando: MVZ 247494 Ucayali
: Balta, Río Curanja: LSUMZ 14582.