SYSTEMATICS OF SNAKES OF THE DIPSAS OREAS COMPLEX (COLUBRIDAE: DIPSADINAE) IN WESTERN ECUADOR AND PERU, WITH REVALIDATION OF D. ELEGANS (BOULENGER) AND D. ELLIPSIFERA (BOULENGER)

2005·51.1 MB·English

by  John E Cadle

SYSTEMATICS OF SNAKES OF THE DIPSAS OREAS COMPLEX
(COLUBRIDAE: DIPSADINAE) IN WESTERN ECUADOR AND PERU,
WITH REVALIDATION OF ELEGANS (BOULENGER) AND
D.
ELLIPSIFERA (BOULENGER)
D.
JOHN CADLE
E. 1
CONTENTS
Abstract 67 ABSTRACT. The systematics and biology of colubrid
Resumen 68 snakes from western Ecuador and northern Peru in
Introduction 69 the Dipsas oreas group, comprising the nominal taxa
Methods 70 D. oreas (Cope), D. elegans (Boulenger), and D. el-
Localities: The Rio Zana Study Site 70 lipsifera (Boulenger), are reviewed. The last two spe-
Systematic Characters, with Special cies are resurrected from the synonymy of D. oreas.
Reference to Dipsas 71 These species, especially D. elegans and D. ellipsi-
The Dipsas oreas Complex: Resurrection of fera, have been confused in previous literature be-
D. elegans and D. ellipsifera 74 cause of inadequate attention to patterns of sexual
Dipsas ellipsifera (Boulenger) 77 dimorphism and geographic variation. Dipsas elegans
Dipsas elegans (Boulenger) 88 and D. ellipsifera share a distinctive colorpattern that
Dipsas oreas (Cope) 97 is quite different from colorpatterns in D. oreas. Dip-
Notes on the Holotype 98 sas ellipsifera differs from both D. oreas and D. ele-
Diagnosis 98 gans in having lower ventral and subcaudal counts,
Description 100 but sexes must be analyzed separately to see the dis-
Hemipenis 108 tinctions clearly. Other subtle characters of scutella-
Geographic Variation in Dipsas oreas and tion, coloration, and dentition aid in distinguishing
the Identity of Peruvian Specimens 108 these species. Dipsas elegans is unusual in that males
Distribution and Type Locality 112 have significantly more ventral scutes than females,
Natural History: Habitats, Activity Patterns, the reverse of the more common colubrid pattern of
Eggs, and Hatchlings 116 sexual dimorphism, in which females have more ven-
Aggregation Behavior in Dipsas oreas 120 trals than males; neither D. oreas nor D. ellipsifera is
Taxonomic and Geographical Notes on sexually dimorphic for this character.
Dipsas gracilis, D. latifasciatus, and Dipsas ellipsifera is known only from the valley of
D. latifrontalis 122 the Rio Mira in extreme northwestern Ecuador (Im-
Dipsas gracilis 123 ababura Province). Dipsas elegans is known from the
Dipsas latifasciatus and D. latifrontalis in western versant of the Andes in Ecuador from just
Eastern Ecuador and Peru 126 north of the equator to about latitude 2°S; it is also
Key to Species of Dipsas in Western South found in the inter-Andean valley of the upper Rio
America 127 Guayllabamba east of Quito. Dipsas oreas is known
Acknowledgments 129 from southern Chimborazo and Guayas Provinces
Specimens Examined and Locality Records ____ 130 south to Loja Province in southern Ecuador, thence
Literature Cited 132 south along the western slopes of the Andes to at
least the Rio Zana (6°51'S) in northern Peru. The
occurrence and distribution of D. oreas in Peru is
detailed for the first time. Most localities for D. oreas
are on the Andean slopes, but the species is also re-
corded byspecimens from the lowlands in thevicinity
1 Department of Herpetology, Chicago Zoological of Guayaquil, Ecuador. The type locality of D. oreas
Society, 3300 Golf Road, Brookfield, Illinois 60513. and manyother South American amphibians andrep-
Associate, Department of Herpetology, Museum of tiles obtained by the naturalist James Orton is the
Comparative Zoology. "elevated Valley of Quito," which has been errone-
Bull. Mus. Comp. Zool., 158(3): 67-136, July, 2005 67
68 Bulletin Museum of Comparative Zoology, Vol. 158, No. 3
ously interpreted narrowly as the vicinity of Quito, serpientes de la familia Colubridae de Ecuador oc-
Ecuador. Orton's own writings show that a broader cidental y el norte del Peru en el grupo Dipsas oreas,
interpretation encompassing virtually all of the Ec- que se compone de los taxones nominales D. oreas
uadorian highlands was intended. The ranges of D. (Cope), D. elegans (Boulenger), y D. ellipsifera (Bou-
elegans, D. ellipsifera, and D. oreas are very likely lenger). Las ultimas dos especies se resucitan de la
extremely fragmented because ofsignificant destruc- sinonimia de D. oreas. Estas especies, especialmente
tion of forest habitats in western Ecuador and Peru. D. elegans y D. ellipsifera, se confusaron en la liter-
Hemipenes of Dipsas elegans, D. ellipsifera, and atura anterior debido a insuficiente atencion a los pa-
D. oreas are slightly bilobed, fully capitate, and have trones de dimorfismo sexual y la variation geografica.
ornamentation similar to other species of the tribe Dipsas elegans y D. ellipsifera comparten un patron
Dipsadini (Dipsas, Sibon, Sibijnomorphus, and Tro- de colores bien distinto y substantivamente diferente
pidodipsas). The capitulum is ornamented with pa- de los patrones en D. oreas. Dipsas ellipsifera es dis-
pillate calyces and the sulcus spermaticus bifurcates tinto de D. oreas y D. elegans al tener bajos cuentos
within the capitulum. Proximal to the capitulum a de ventrales y subcaudales, pero es necesario analizar
battery ofenlarged spines encircles the midsection of por separado los sexos para ver claramente las distin-
the organ. A large basal nude pocket is present on ciones. Otros caracteres mas sutil de la escamacion,
the hemipenis in all three species. la coloration, y la dentition ayudan al distinguir estas
Notes on coloration, natural history, and behavior especies. Dipsas elegans es extrafio ya que los machos
are reported for the three species ofthe oreas group, tienen considerablemente mas escamas ventrales que
although most observations are for Dipsas oreas. In las hembras, lo contrario del patron comun de di-
northern Peru, D. oreas shows extensive intrapopu- morfismo sexual en la familia Colubridae, en que las
lational variation in coloration, which could in part be hembras tienen mas ventrales que los machos; ni D.
sexually dimorphic. It is unclear whether this varia- oreas ni D. ellipsifera demuestra dimorfismo sexual
tion pertains to other parts ofthe distribution. Dipsas en esta caracter.
oreas is active nocturnally in low vegetation but seeks Dipsas ellipsifera se conoce solamente del valle del
seclusion in leaflitter or under surface objects on the Rio Mira al norte extremo del Ecuador (Provincia de
ground during the day. This diel behavior pattern is Imbabura). Dipsas elegans se conoce del lado occi-
amlisgohtrepboertceodmfmoronsevienratlheotgheenruss.pecIinesnoorftDhieprsnasPearnud, dental de los Andes en Ecuador desde justo al norte
del ecuador hasta aproximadamente la latitud 2°S; se
tchoienaccitdievsitwyiotfhatdhueltrDa.inyorseeaassoins.stAropnegcluylsieaarsoangaglreagnad- encuenta tambien en el valle interandino del alto Rio
Guayllabamba al este de Quito. Dipsas oreas se con-
tion of Dipsas oreas encountered at a locality in oce desde la parte sur de las provincias Chimborazo
northern Peru is described. Comprising one female y Guayas hastalaprovincia Lojaen el surde Ecuador,
and six males, this is the first reported case of "ag-
gregation behavior" in any snake of the tribe Dipsa- desde alii a lo largo de las vertientes occidentales de
dini and one ofvery few observations ofsuch behav- los Andes, por lo menos hasta el Rio Zana (6°51'S)
en el norte de Peru. La presencia y distribution de
ior in Neotropical colubrids. Details of the aggrega-
tion suggest an association with mating, and the ex- D. oreas en Peru se detallan para la primera vez. La
istence of communal nesting in this species suggests mayoria de las localidades para D. oreas se encuentra
that aggregation might occur for oviposition as well. en las vertientes Andinas, pero tambien es recordado
A key to all species ofDipsas reported or expected por ejemplares de tierras bajas vecino a Guayaquil,
in western South America (Colombia, Ecuador, and Ecuador. La localidad tipica de D. oreas y muchas
Peru) is provided. This includes six species known otras especies de anfibios y reptiles sudamericanos
from western Ecuador (andiana, elegans, ellipsifera, que fueron obtenido por la naturalista James Orton
gracilis, oreas, andtemporalis), two ofwhich (gracilis, es "el valle elevado de Quito", que se han interpre-
oreas) also occur in western Peru, and three addi- tado estrechamente y equivocadamente como los al-
tional species known or expected in Chocoan Colom- rededores de la ciudad de Quito, Ecuador. Las obras
bia (nicholsi, sanctijohannis, and viguieri). Dipsas propias de Orton demuestran que el quiso decir una
gracilis and D. viguieri are not distinguishable byany interpretation mas amplia, que abarco casi todas las
reported characteristics of the taxa. Brief notes are partes altoandinas Ecuadorianas. Los rangos de D.
provided on the occurrence of D. gracilis in north- elegans, D. ellipsifera, y D. oreas son probablemente
western Peru. Some taxonomic issues concerningtwo muy fragmentados debido a la destruction substan-
species ofthe Amazonian versant, D. latifasciata and tia de los bosques en Ecuador y Peru occidental.
D. latifrontalis, are outlined. These species are not Los hemipenes de Dipsas elegans, D. ellipsifera, y
clearly distinguishable on the basis ofcharacters dis- D. oreas son ligeramente bilobulados, completamente
cussed in the literature, and the assignment ofthese capitatos, y tienen ornamentation similar a otras es-
names to specimens from eastern Ecuador needs to pecies del tribo Dipsadini (Dipsas, Sibon, Sibyno-
be reviewed in conjunction with study of their holo- morphus, y Tropidodipsas). El capftulo es ornamen-
types. tado con calices que llevan papilas, y el surco esper-
matico se divide dentro del capftulo. Proximal al cap-
RKSUMEN. Se revisan la systematicay biologfa de las ftulo hay una serie de espinas agrandadas que rodean
Dipsas oreas Complex in Ecuador and Peru • Caclle 69
el organo. Un bolsillo desnudo y grande se encuentra eral species in western Ecuador and Peru
en la base del hemipene de las tres especies. have been poorly characterized in existing
Se reportan notas sobre la coloracion, la historia literature, making species determinations
natural, y el comportamiento para las tres especies
del grupo oreas, aunque la mayoria de las observa- ofcertain populations difficult. At the time
ciones son para Dipsas oreas. En el norte de Peru, that Peters (1960a) and Peters and Orejas-
D. oreas muestramuchavariacion intrapoblacionalen Miranda (1970) reviewed Dipsas, a few
colores. En parte, esta podria ser debido al dimorfis- species were known from western Ecua-
mo sexual. No es claro si esta variacion pertenece a
dor, but none was recorded from the west-
otras partes de su distribution. Dipsas oreas es activo
nocturnamente en vegetacion baja, pero por dia se ern slopes ofthe Andes or Pacific lowlands
encuentra en la hojarasca o debajo de objetos sobre in Peru. The availability of new material
la superficie de la tierra. Este patron de comporta- makes review of these species timely be-
miento diario se reporta para varias otras especies de
Dipsas y puede ser comun en el genero. En el norte cause of the confused state of the system-
del Peru, la actividad de adultos de D. oreas es muy atics of several species in western Ecuador
estacional y corresponde fuertamente con la epocade and misapplied names to many specimens
lluvias. Se describe una agregacion extrafia de Dipsas already in collections. A subsequent paper
oreas que se hallo en una localidad de campo en el (Cadle, unpublished data) will review new
norte del Peru. Compuesto de una hembray seis ma-
material of the related genus Sibynomor-
chos, esto es el primero reporte de "comportamiento
de agregacion" en una especie del tribo Dipsadini y phus.
es uno de muy pocas observaciones de tal compor- Peters (1960a) and Peters and Orejas-
tamiento en colubrideos Neotropicales. Los detalles Miranda (1970) thought that three species
sugeren que la agregacion fue asociada con el aco-
groups of Dipsas were represented in
plamiento y la existencia de nidos comunales en esta
especie sugere que agregacion puede suceder tam- western Ecuador: (1) the variegata group,
bien por poner huevos. represented in western Ecuador by D.
Se provee una clave para todas las especies de Dip- variegata variegata and D. variegata ni-
sas conocidas o esperadas en Sudamerica occidental cholsi; (2) the oreas group, represented by
(Colombia, Ecuador, Peru). Estas incluyen seis es-
pecies conocidas del Ecuador occidental (andiana, D. oreas, D. ellipsifera, and, as clarified lat-
elegans, ellipsifera, gracilis, oreas, temporalis), de las er by Kofron (1982), D. elegans; and (3)
cuales dos (gracilis, oreas) tambien se encuentran en the articulata group, represented in west-
el Peru occidental; se conocen o se esperan tres es- ern Ecuador by D. gracilis and D. tem-
pecies adicionales en el Choco de Colombia (nicholsi,
sanctijohannis, viguieri). No se pueden distinguir D. poralis. Subsequent to Kofrons (1982) re-
gracilis y D. viguieri por las caracteristicas reportadas view, D. elegans and D. ellipsifera were
en la literatura. Se proveen notas breves sobre lapre- recognized as subspecies of D. oreas (Or-
sencia de D. gracilis en el noroeste del Peru. Se re- ces and Almendariz, 1987). Until recently,
sumen algunos problemas acerca de la taxonomia de none of these species was known from
dos especies de lavertiente amazonica, D. latifasciata
y D. latifrontalis. No se pueden distinguir estas es- Peru, but Cadle and Chuna (1995) and
pecies por las caracteristicas en la literatura, y es ne- Tello (1998) reported D. oreas and D.
cesario revisar el uso de estos nombres para especi- gracilis, respectively, from northern Peru.
menes desde Ecuador oriental conjunto con estudio Species in the variegata and oreas groups,
de los holotipos.
in particular, have presented complex sys-
tematic problems, resulting in numerous
INTRODUCTION
misidentified specimens (e.g., several un-
Neotropical snakes of the genus Dipsas related species confused with D. variegata;
(Colubridae: Dipsadinae: Dipsadini) were Cadle and Myers, 2003). Consequently,
last comprehensively reviewed by Peters several species within these groups from
(1960a). Many South American species re- Ecuador and Peru have routinely been
main poorly known because they are rep- confused, prompting me to review these
resented by few specimens and/or they ex- groups while identifying collections result-
hibit complex variation in characters, such ing from biological inventories in northern
as color pattern and scutellation, typically Peru.
used to infer species limits in snakes. Sev- Cadle and Myers (2003) reviewed the
70 Bulletin Museum of Comparative Zoology, Vol. 158, No. 3
"Dipsas variegata" group (sensu Peters, and 1989 (early rainy season: January) for
1960a) in Panama and western South a combined total of 11 weeks, I made her-
America. Their principal conclusions rel- petological collections in the Monte Seco
evant to the identity of species of Dipsas region; some results from those surveys
in western Ecuador were (1) specimens were reported previously (Cadle, 1989,
from Ecuador previously referred to Dip- 1991; Cadle and Chuna, 1995; Cadle and
sas v. variegata were misidentified speci- McDiarmid, 1990). In the late 1980s and
mens of D. oreas, D. andiana, or Sibijno- early 1990s, humid montane forest cov-
morphus petersi, all of which are well ered the Andean slopes above Monte Seco
known from western Ecuador, and (2) Ec- from approximately 1,500 to 2,500 m, with
uadorian specimens identified as D. v. ni- an areal extent estimated to be 2,500 ha
cholsi by Peters (1960a) are a distinct spe- (Sagastegui et al., "2003" [2004]). Above
cies, Dipsas andiana (Boulenger), a name 2,500 m in the immediate vicinity were ta-
resurrected from the synonymy of D. or- blelands and slopes that had been largely
eas, where it had been placed by Peters cleared of forests, although forest frag-
(1960a). Thus, no definitive records of ments still existed up to nearly 3,000 m.
Dipsas variegata are known from Panama, Specific survey sites at Monte Seco varied,
Colombia, Ecuador, or Peru, and the close and fieldwork sampled a wide elevation
relationship of D. andiana to D. variegata range concentrated between 1,200 and
seems highly questionable. See Cadle and 2,500 m; less intense sampling was con-
Myers (2003) for further details. ducted down to 1,000 m and as high as
Cadle and Myers (2003) briefly charac- 3,000 m. The primary focus of the field
terized Dipsas oreas in diagnosing D. an- work was dense humid montane forest
diana. They examined specimens of the covering the mountain slopes, but other
nominal taxa elegans and ellipsifera but habitats, including agricultural land (coffee
left their relationship to D. oreas as an plantations, secondary forests) and open
"open question" (Cadle and Myers, 2003: highland grasslands and bushlands, were
footnote 13). This paper extends their dis- also sampled. The known flora of Bosque
cussion of Dipsas oreas and evaluates the Monte Seco comprises approximately 380
status of D. elegans and D. ellipsifera. In species of flowering plants and 40 species
addition, I summarize natural history and of pteridophytes (Sagastegui et al., 2004).
behavioral observations for these species Wet and dry seasons at the Rio Zana
and provide a key for the identification of Study Site are very pronounced, with a
all species of Dipsas known or expected distinct rainy season occurring from ap-
from west of the Andes in Colombia, Ec- proximately December to April. Temper-
uador, and Peru. atures recorded using a maximum/mini-
mum
METHODS mercury thermometer at the base
camp (1,800 m) were as follows:
Localities: The Rio Zana Study Site.
Most of my observations on the natural 5 May-26 June 1987 (50 days, early diy
history ofDipsas oreas in Peru are derived season)
from field work conducted in the vicinity Maximum daily average, 28.5° C
of Monte Seco (also known as Monteseco), (range 21.5°-35.5° C)
a small town on the north side of the Rio Minimum daily average, 9.2° C
Zana in western Cajamarca department, (range 5.5°-12° C)
Peru (6°51'S, 79°6'W). Thus, it seems per-
tinent to make a few brief comments on 14-29 January 1991 (12 days, early rainy
this locality and others in northern Peru season)
where D. oreas is known to occur. Maximum daily average, 19.5° C
In 1987 (early dry season: May—June) (range 16.5°-21° C)
Dipsas oreas Complex in Ecuador and Peru • Cadle 71
Minimum daily average, 10.7° C calities, unless given by the collectors or
(range 10°-12° C) otherwise stated, were derived from orni-
thological gazetteers of the Neotropics
The difference between the dry and rainy (Paynter, 1993, 1997; Stephens and Tray-
season maximum daily temperatures at the lor, 1983); from Peruvian departmental
study site is because many days during the maps or 1:50,000 topographic maps pro-
rainy season are prone to dense fog or duced by the Instituto Geografico Nacion-
cloud cover, whereas fog is rare during the al, Lima; or from the online versions ofthe
dry season. These records show that the gazetteers ofthe U.S. Board on Geograph-
minimum daily temperature range was ic Names at the GEOnet© names server
much greater during the dry season (char- (http://earth-info.nga.mil/gns/html/index.
acterized by clear, cool evenings) than dur- html; the Web address seems to change
ing the rainy season (foggy/cloudy, warmer with great frequency but should be search-
evenings), even though average minimum able with the name GEONET). Bracketed
temperatures are similar. data in localities are inferences from these
Cadle (1989, 1991), Cadle and Chuna sources. Abbreviations for museums in
(1995), and Cadle and McDiarmid (1990) which cited specimens are housed are giv-
give additional details on the geography en in the list of specimens examined.
and habitats of the Monte Seco region. Systematic Characters, with Special
Dillon et al. (1995) included the humid Preference to Dipsas. I followed standard
forest of Monte Seco in a floristic study of methods used in previous studies of snake
western Peru and Ecuador, Sagastegui and systematics (e.g., Cadle, 1996; Cadle and
Dillon (1991) provided a checklist of its Myers, 2003; Myers, 1974). Because some
flora, and Sagastegui et al. (2004) included scutellation characters are highly variable
the site in a comparative floral analysis of intraspecifically within Dipsas, I here com-
humid montane forests of northern Peru. ment on the nature of the variation and
For simplicity, in this paper I refer to all ways in which some of these characters
specific localities within the Monte Seco were recorded.
area as the Rio Zana Study Site and give Peters (1960a: 25) noted the extreme
elevation and habitat detail where perti- variability of head scale patterns within
nent. Descriptions or photographs of the Dipsas: "Division, extra suturing, or fusion
Rio Zana Study Site and some other lo- takes place in nearly every head scale in
calities cited herein are given by Cadle this genus." Scale patterns ofthe temporal
(1991), Cadle and Chuna (1995), and Sa- region in Dipsas are especially prone to
gastegui et al. (2004). The herpetological intraspecific variation (Peters, 1960a: 26;
collections made at the Rio Zana Study Cadle and Myers, 2003: table 1), and the
Site in 1987 and 1990 were timely, inas- patterns of fragmentation and fusion of
much as a considerable portion of the hu- temporal scales make concise character
mid forest was cut and burned in the mid- definitions of species difficult. Generally, I
1990s to provide pasturage for cattle and attempted only to score primary and sec-
land for coffee production (see Sagastegui ondary temporals, and even this proved
et al., 2004: figs. 6, 7). The site should be difficult in some cases. A few other head
resurveyed to determine how much of its scale patterns seem useful in distinguish-
unique herpetofauna survives and to pro- ing some species pairs, especiallywhen the
vide a baseline for its future recovery or frequencies of alternative states are con-
extinction. sidered. In this category are scales in the
Elevations for my field sites were deter- loreal region involving the loreal, postnas-
mined with a Thommen© altimeter, usu- als, preoculars, and prefrontals; a few com-
ally in conjunction with topographic maps. ments concerning scale patterns in this
Coordinates and elevations for other lo- area are warranted because they are help-
72 Bulletin Museum of Comparative Zoology, Vol. 158, No. 3
ful in distinguishing species considered unusually long preocular above the loreal
herein. (loreal pattern 6, Fig. 1).
In those cases in which only a single Loreal patterns 1 and 2 are identical ex-
scale is present between the postnasal cept for the presence or absence of a dis-
scale and the eye, I follow Peters (1960a: crete suture separating the preocular scale
7—8) in considering this scale to be a lo- from the prefrontal scale. Patterns 3 and
real. A preocular scale, when present in 4 are related in a similar way. Usually a
Dipsas, is small, usually located superior to single pattern overwhelmingly predomi-
the loreal, and bordered by the loreal, pre- nated for each species (see Tables 1 and 3
frontal, supraocular, and eye (Fig. 1). This and the discussion of Dipsas oreas for ex-
terminology is also consistent with defini- ceptions). In all cases of intraspecific var-
tions given by Savage (1973), but Dipsas iability in patterns 1—4, alternative states
differs from the more common colubrid within a species were always either Pat-
conditions commented upon by Savage. terns 1 and 2 or patterns 3 and 4 (and
Savage (1973: fig. 1) considered the com- never, e.g., pattern 1 with pattern 4). For
mon condition in which the preocular example, in D. oreas, in which pattern 2
scale separates the loreal scale from the was overwhelmingly frequent, alternate
anterior border of the eye (i.e., the preo- states involved pattern 1 or other rare pat-
cular is intercalated entirely between the terns, but never patterns 3 or 4. Patterns
eye and loreal); Savage also considered 5 and 6 appeared with less frequency than
any other patterns, except for a high fre-
patterns of scale fusions involving the lo-
quency of pattern 6 within one population
real + prefrontal. In contrast, in Dipsas
of D. oreas (see below).
the preocular typically lies superior to the
One reviewer ofthis paper suggested in-
loreal (Fig. 1), and the loreal contacts the cluding total segmental counts (ventrals +
anterior border of the eye; scale fusions
usually involve the preocular + prefrontal. subcaudals) as part of the diagnoses and
Differences among the species of Dip- descriptions of these species, and as a
summary to aid in sexing ofspecimens (the
sas considered herein involved primarily
idea being that total segmental counts
two characters of the loreal region: (1) the would add together whatever smaller sex-
shape of the loreal scale itself, which was
ual dimorphism might exist within ventral
either square/polygonal or much longer
or subcaudal counts considered separately,
than tall (rectangular), and (2) whether a thus enhancing the distinction between
preocular appeared as a distinct scale (usu-
sexes). This would perhaps permit easier
ally superior to the loreal) or was fused sexing of specimens without the need for
with the prefrontal scale (in which case the dissection. Although this approach seems
fused prefrontal—preocular touches the an- useful for some groups of snakes, several
terior border of the eye). Four patterns in aspects of variation within the snakes con-
scales of the loreal region were most fre- sidered herein make the use of total seg-
quent, loreal patterns 1—4 (Fig. 1), distin- mental counts less useful, and exclusive
guished by the two characters ofthe loreal consideration of total counts actually ob-
and preocular scales just mentioned. In scures some patterns apparent from con-
addition to these common patterns, a few sideration of ventral and subcaudal counts
rarer conditions were observed, which in- separately.
volved either the presence of a subpreo- Two of the species considered herein
cular (formed by a suture segregating the (Dipsas oreas and D. ellipsifera) are not at
posteroventral corner of the loreal from all sexually dimorphic in ventral counts,
the main part of that scale; loreal pattern but strongly dimorphic in subcaudal
5, Fig. 1) or horizontal partition of the lo- counts (male counts higher than female
real, nasals, and/or prefrontals to form an counts with virtually no overlap in counts
Dipsas oreas Complex in Ecuador and Peru • Cadle 73
Loreal Pattern 1 Loreal Pattern 2
Loreal Pattern 3 Loreal Pattern 4
Loreal Pattern 5 Loreal Pattern 6
Figure 1. Variation in scale patterns in the loreal region of species of Dipsas. Patterns 1 and 2: loreal squarish orpolygonal;
separate preocularpresentabove loreal (pattern 1) orfusedwith prefrontal scale (pattern2). Patterns3and4: loreal rectangular,
much longer than tall; separate preocular present above loreal (pattern 3) orfused with prefrontal scale (pattern 4). Patterns5
and 6: rare patterns; pattern 5 is similar to pattern 2 but has a small triangular preocular separated from the posteroventral
corner of the loreal; pattern 6 has an elongate preocular above the loreal and extending from the eyeto the internasal. Patterns
1 and 2 are the common patterns in D. oreas, with patterns 5 and 6 less frequent. Patterns 3 and 4 are characteristic of D.
elegansand D. ellipsifera. See text for discussion.
between the sexes; see Tables 1—3). In between the sexes. As in D. oreas and D.
these species, use of total segmental ellipsifera, the use of total segmental
counts yields no insights that are not ap- counts in D. elegans reveals nothing that
parent from separate consideration ofven- would not be seen by separate consider-
tral and subcaudal numbers, specifically ation of ventrals and subcaudals, each of
because the difference between the sexes which is strongly sexually dimorphic for
with regard to total segmental counts is this species. Moreover, it obscures the
contributed entirely by the subcaudals; highly unusual pattern of sexual dimor-
this fact is obscured when the subcaudal phism in ventral counts in D. elegans (dis-
count is evaluated solely as a component cussed later herein).
of the total segmental count. On the other Thus, the use of total segmental counts
hand, in D. elegans, males have signifi- elucidates no patterns of sexual dimor-
cantly greater numbers of ventrals and phism or interspecific differences within
subcaudals than females, with virtually no the Dipsas oreas group that would not be
overlap in the counts for either character evident from separate analysis of ventral
74 Bulletin Museum of Comparative Zoology, Vol. 158, No. 3
and subcaudal numbers. It also potentially fera (Boulenger, 1898) with type locality
obscures other features, such as the rela- "Ibarra" [Ecuador]. These nominal taxa
tive contribution of ventrals or subcaudals have remained poorly known since their
to any revealed pattern, and atypical pat- descriptions. Peters (1960a) placed D. ele-
terns of sexual dimorphism (as for D. ele- gans and D. ellipsifera in his oreas group,
gans). Nonetheless, I report total segmen- but he had insufficient data to infer the
tal counts in the descriptions and summary relationships ofLeptognathus andrei. Kof-
data tables that follow so that others can ron (1982) provided photographs of the
evaluate their utility for themselves, holotypes or syntypes of these forms and
Whether the approach of using total seg- presented evidence that Leptognathus an-
mental counts yields generally useful in- drei Sauvage is a synonym of nominotypi-
sights into the systematics of Dipsas is not cal D. oreas, sl conclusion with which I
clear to me. A casual inspection ofdata for concur on the basis of the photograph of
several other South American species (Ta- the type and information provided by Kof-
ble 4; unpublished data) indicated no par- ron (see Table 3).3
ticular utility over and above separate con- The status of the nominal forms Dipsas
sideration ofventral and subcaudal counts elegans and D. ellipsifera and their rela-
for this small sampling of taxa. tionship (ifany) to D. oreas require further
consideration. Contrary to reports in the
THE DIPSAS OREAS COMPLEX:
literature, these three taxa are diagnosable
RESURRECTION OF DIPSAS ELEGANS
from one another. Peters (1960a) recog-
AND DIPSAS ELLIPSIFERA
nized the similarity of the color patterns of
D
Cope (1868) described Leptognathus elegans and D. ellipsifera but thought,
oreas from a specimen (holotype: ANSP erroneously, that D. elegans was a Mexican
10115) obtained by James Orton with the species because the original description
type locality "elevated Valley of Quito" (Boulenger, 1896: 452) stated that the ho-
[Ecuador]. The "Valley of Quito" is the lotyPe was from "Tehuantepec." Conse-
type locality for many amphibians and rep- quently, Peters assigned all Ecuadorian
tiles from Ortons large South American specimens with the distinctive elegans-el-
collections, but it has been misinterpreted Upsifera pattern to D. ellipsifera (Peters,
by most authors dealing with his collec- 1960a: 92)- Peters' concept ofD. ellipsifera
tions. It is clarified herein (see the account therefore included specimens of D. ele-
for Dipsas oreas). gans, even though he recognized the latter
Subsequent to Cope's description of as a distinct species from Mexico. Subse-
Dipsas oreas, several taxa now perceived quent discussions of these snakes (Kofron,
°
as related to it were described: Leptogna- 1982; rces and Almendariz, 1987) have
thus andrei Sauvage (1884) with type lo- been misled by Peters' confusion of the
^°
cality "Loja, Nouvelle-Grenade"2 D. ele- species in western Ecuador and by in-
gans (Boulenger, 1896) with type; locality adequate attention to sexual dimorphism
"Tehuantepec" [Mexico];, and D. ellipsi- and geographic origin of samples. This
confusion, combined with tacit acceptance
by most recent workers of specimen iden-
2After 1830, New Granada comprised Panama +
Colombia, a result of the secession of Ecuador and
Venezuela from Gran Colombia, which had existed 3 Dipsas oreas was, for many years, also confused
for a short time subsequent to colonial independence with Sibynomoiyhus mikanii (e.g., Sibynomorphus
from Spain in the early 19th century. Sauvage (1884) mikanii oreas or Dipsas mikanii oreas reported from
reported "Nouvelle-Grenade" as the type locality, but Ecuador by Amaral 1929a, "1929"b [1930] and Park-
the data recorded in the MNIIN catalogues for the er [1934, 1938]). This resulted from synonymy ofthe
holotype (MNIIN 6285) explicitly include "Loja" ac- two species by Giinther (1872) and Boulenger (1896:
cording to R. Roux-Esteve as reported by Kofron 453). The confusion was discussed by Peters (1960a:
(1982: 48-49). 93-94).
Dipsas oreas Complex in Ecuador and Peru • Cadle 75
tifications in some older literature (e.g., vicinity of the Isthmus of Tehuantepec,
Boulenger, 1896), has led to considerable with the type of D. elegans. The error was
uncertainty concerning differential char- promulgated by Gunther (1885—1902:
acteristics of these nominal taxa and their 141) and continued in checklists and fau-
respective distributions. nal works (e.g., Peters, 1960a: 86; Smith
Peters' (1960a) confusion of two Ecua- and Taylor, 1945) until Kofron discovered
dorian species under the name Dipsas el- the mistake. The unfortunate association
lipsifera affected his assessment of varia- of Sumichrast's name with the specimen
tion within this species. For example, his that ultimately became the holotype of D.
ventral and subcaudal counts for males elegans resulted in it being considered a
and females of "Dipsas ellipsifera" (Peters, Mexican species by Peters (1960a) and
1960a: 87) are exceptionally broad for a others (e.g., Smith and Taylor, 1945).
species of snake having a small range in Kofrons (1982) conclusion that the ho-
western Ecuador. Nonetheless, Peters lotype of Dipsas elegans most likely came
(1960a: 91) recognized a distinctive geo- from Ecuador was a major step toward re-
graphic pattern to the variation within his solving the systematics of these snakes.
concept of D. ellipsifera: Nonetheless, because the type specimens
ofD. ellipsifera and D. elegans have nearly
The material available can be divided into two identical color patterns, Kofron (1982: 48)
groups as to provenance, the first coming from the synonymized D. ellipsifera with D. elegans,
pRoipoulMaitrioan,d"raainndagteheanodthererfferrroemdtthoe wasesttheern"tsylpoipceasl which is the earlier name. He did not thor-
of the Andes in Ecuador, to the south of the Rio oughly consider other character variation
Mira. Although the indicated differences be- in available samples of these snakes, par-
. . .
tween these two populations suggest subspecific ticularly in the context ofgeographic origin
status, I am not assigning them a name, because and sex, even though some of these pat-
no satisfactory holotype is available. The most
striking differences between the two. .p.opulations terns had already been elucidated by Pe-
are in counts involving body segments. ters (1960a; see previous quotation).
Despite having examined much of the
Peters goes on to detail differences be- same material as Peters, Kofron (1982)
tween the Rio Mira specimens and the used Peters' (1960a) data when comparing
others when samples are segregated by the scale counts of the holotype of Dipsas
sex. However, he did not make the con- elegans with D. ellipsifera, apparently fail-
ceptual leap to recognizing distinct species ing to realize that Peters had confused the
on the basis of these populational differ- two species under the name "Dipsas ellip-
ences. Nor did he recognize that the "oth- sifera" and overlooking Peters' more de-
er" population from the western slopes of tailed discussion of geographic variation
the Andes represented Dipsas elegans, be- within his concept of D. ellipsifera: "The
lieving as he did that D. elegans was a ventral and subcaudal counts of [the ho-
Mexican species. lotype of] D. elegans (183, 95) are within
Kofron (1982) clarified much of the the ranges of 155-187 ventrals and 72—105
confusion engendered by the uncertain or- subcaudals reported for D. ellipsifera by
igin of the holotype of Dipsas elegans. He Peters (I960)" (Kofron, 1982: 47). Thus,
summarized evidence that the holotype of Kofron (1982) accepted Peters' (1960a)
D. elegans came from western Ecuador on characterization of "Dipsas ellipsifera,"
the basis ofthe known travels and contacts which included specimens of D. elegans;
of Adolphe Boucard, who sold the speci- not surprisingly, characteristics of the ho-
men to the British Museum (Kofron, 1982: lotype ofD. elegans conformed well to this
47). Kofron also reviewed the erroneous composite taxon. Kofron did not reexam-
association of the name of the collector, ine the meristic data of specimens segre-
Francois Sumichrast, who collected in the gated by sex and locality and failed to no-
76 Bulletin Museum of Comparative Zoology, Vol. 158, No. 3
tice a few other subtle differences be- tunate because color patterns are highly
tween the two (see below). Peters (1960a), variable within some species of Dipsas.
although he elucidated the pattern of geo- Nonetheless, Peters' oreas group included
graphic variation, did not differentiate the D. elegans, D. ellipsifera, and D. oreas by
population samples as distinct taxonomic virtue of having a color pattern in which
entities or associate the name "Dipsas ele- "the blotches are wider than the inter-
gans" with one of the population samples. spaces, with little contrast in color be-
Kofron (1982: 50) further suggested that tween them (pi. IVa) [and] the centers of
"A comprehensive analysis ofD. oreas and the blotches are considerably lightened,
D. [elegans] may show the two to be sub- which often gives the species an appear-
specifically related, as I can distinguish the ance of having paired ellipses on the
specimens only by pattern." Kofron's sides. The interspaces are rather heavi-
. . . . . .
(1982: 50) use of the name D. ellipsifera ly streaked and spotted with dark colors"
throughout the last paragraph ofhis text is (Peters, 1960a: 31).
an apparent lapsus for D. elegans in view A glance at the photograph that Peters
of his synonymizing the former name with chose to illustrate color patterns charac-
the latter earlier in the paper (Kofron, teristic of the oreas group (Peters, 1960a:
1982: 48). Plate IVa) shows that even this example
Orces and Almendariz (1987) acted on does not fit the definition well (e.g., some
Peters' observations of geographic varia- of the interspaces are wider than the
tion and examined additional specimens of blotches and the the color pattern exhibits
both Dipsas elegans and D. ellipsifera. Un- great contrast). Examination of only a few
fortunately, in their reconsideration of the specimens will quickly erode confidence in
geographic pattern, they again lumped Peters' definition of the oreas group. In all
specimens by locality without regard to sex three species ofthe oreas group, the width
and thus failed to elucidate the correlation of the blotches (or bands) varies along the
between sexual dimorphism in meristic length of the body; anteriorly, they are vir-
data and geography that had been outlined tually always wider than the interspaces,
by Peters (1960a). On the basis of average but posteriorly, they are narrower than the
differences in meristic characters between interspaces. Most specimens of all three
population samples and little or no overlap species have highly contrasting patterns
in segmental counts between them, Orces (see later discussion of D. oreas for some
and Almendariz (1987) maintained D. ele- exceptions, Cadle and Myers [2003: 21—
gans and D. ellipsifera as distinct taxonom- 25], the photograph of the holotype in
ic entities but, following Kofron's (1982: Kofron [1982], and later discussion and il-
50) suggestion, considered them subspe- lustrations herein).
cies of D. oreas Cope. Orces and Almen- The similarity and uniqueness of the
dariz provided the new combinations Dip- color patterns of Dipsas elegans and D. el-
sas oreas elegans and Dipsas oreas ellipsi- lipsifera, and their narrowly allopatric dis-
fera. The taxonomic justification was based tributions (documented herein), suggest a
on the broad overlap in ventral and sub- close relationship between the two species.
caudal counts between nominotypical D. However, the similarity in segmental
oreas and D. elegans (Orces and Almen- counts between Dipsas elegans and D. or-
dariz, 1987: 138) and on the similar color eas, particularly among females (see Table
patterns of D. elegans and D. ellipsifera. 1), is of minimal consequence in assessing
The relationship, if any, of Dipsas ele- a possible relationship or conspecificity be-
gans and D. ellipsifera to D. oreas warrants tween them because many species ofDip-
renewed scrutiny. Peters (1960a: 30-31) sas have broadly overlapping scale counts
defined species groups on the basis of col- (Cadle and Myers [2003] and subsequent
or pattern characteristics. This was unfor- discussions herein of D. gracilis). Thus,