Natural History Articles
THE
DESCRIPTION OF A NEW SUBSPECIES OF BOA
CONSTRICTOR FROM PERU *
Present Address: 1450 Blue Spruce Lane, Wantagh, N.Y. 11793 USA Abstract:
The taxonomy of the genus Boa is poorly elucidated as a result of the
spotty systematic fieldwork on this wide-ranging genus. Given
current suggested modifications of the species concept, the insular taxa, nebulosa
and orophias, warrant specific recognition. Evidence for the
occurrence of Boa constrictor
imperator in southern Ecuador and Peru is lacking, and its retention
on checklists of Peru is a literature artifact. A new Boa,
restricted to Tumbes, Peru, merits at least subspecific recognition, as it
differs from all conspecifics in length of tail and hemipenis, color
pattern, and is geographically isolated. The
systematics of the species Boa constrictor
is in a lamentably poor state considering that it has been known to
the scientific community since at least the time of Linnaeus, is one of
the most widely distributed terrestrial vertebrates known, and is arguably
the- most widely recognized snake in the world. Boa
constrictor is polytypic on two continents comprising a geographic
range of over 6 million square miles, including five known distinguishable
insular populations (at least one of which has not been formally
recognized) and at least two mainland forms apparently allopatric with
either of the major continental subspecies. Our incomplete
knowledge of this species is likely a result poor sampling over the
extensive distributional range which has dissuaded investigators from
pursuing any comprehensive study. Recently
dissatisfaction among herpetological taxonomists with the old Biological
Species Concept and suggestions for new specific and subspecific criteria
does allow for more latitude in analysis of the populations and demes of Boa constrictor, but will,
of course, fall short of satisfying the more extreme interpretations.
We have chosen to base our determinations on the more conservative,
traditional rule that two populations are subspecifically different if
they different in a suite of characters to the degree of at least 75%
(full species differing at the 100% level). In many interpretations
of this rule, an area of parapatry and intergradation must occur for-
subspecies recognition. We emphasize that time frame may create
operational difficulties in this theory, as a current observation may not
provide adequate data as to long term characteristics and fates of
evolving populations. Nonetheless, this method of subspecific
determination (with minor variations) is supported by Smith (1990), who
emphasizes that recognition of dichopatric populations as different
species is subjective and subject to opinion, Janis Roze (pers. comm.),
and Herndon Dowling (pers. comm.).
If one accepts the Phylogenetic Species Concept of Rosen (1978, 1979), the
smallest diagnosable populations of Boa constrictor with a
derived character difference would be full species, including the
undescribed pallid population from Islas de Bahia and Cayos Cochinos,
Honduras, the new Peruvian population described herein, the isolated
subspecies ortonii, and the
insular Caribbean orophias and
nebulosa. The somewhat more conservative, cladistically
influenced Evolutionary Species Concept (ESC) of Frost and Hillis (1990)
would also elevate orophias,
nebulosa, and likely the new Peruvian population to full species, as
they are allopatric and appear not be reproductively compatible with other
Boa. Lazell (1964), in describing nebulosus
(which should have been named nebulosa,
as Boa is a feminine genus), expressed some misgivings about not
according full species status to orophias
and nebulosa, especially as he was concerned about their ability to
interbreed with mainland forms. He stated that he would have
described orophias as a full
species had nebulosa not existed
to form a cline from the mainland form. Proponents of the ESC
consider allopatric clinal subspecies to be full species, particularly if
they will be subsumed into the larger species in future. We see no
reason to infer that the islands of St. Lucia and Dominica (as a single
unit) will afford boas opportunity for genetic exchange with the mainland
barring major tectonic catastrophe. Stull (1935) placed these Lesser
Antillean forms of Boa in the
species orophias, as a
consequence of their insular occurrence on St. Lucia and Dominica and
their substantial phenotypic differences from mainland Boa. We
concur, these snakes being 100% genetically isolated from the mainland,
although it is uncertain whether they comprise one species or two.
Insufficient data is available on the Honduran insular populations to make
a determination of their proper taxonomic rank. Peters and Orejas-Miranda
(1970), and Vanzolini (1986) consider the insular Panamanian sabogae a subspecies based on traditional methodology. The
distance from Saboga Island to Panama is not sufficient for a
determination of reproductive isolation to have been conclusively
made. Zweifel (1960) has placed sigma
in the synonymy of imperator based on traditional criteria.
Again, the distance from the Tres Marias Islands to Mexico proper is
insufficient for isolation or important character variation. The
Peruvian Boa Fauna
Despite the facts that Boa
constrictor has never been reviewed in its entirety and the
relationships between the two major contiguous subspecies, constrictor
and imperator, are poorly
elucidated, there exist several insular and/or relictual subspecies
recognized by nearly all workers (vide
supra). Several American workers (Wilson and Meyer, 1985)
have argued against new or additional Boa subspecies and have, in
fact, called a taxonomic system archaic. While the authors stand apposed
to the naming of gratuitous or dubiously documented taxa at any level, as
has recently been the case with Lampropeltis Triangulum, and
possibly with Boa constrictor Melanogastor, (Lang hammer, 1983), we
are compelled to formally report as new such taxon as merits hierarchical
level. Having found a population of Boa which has
evolved sufficiently to appear unique, we herein describe a new subspecies
from a relictual area in Tumbes, Peru.
Part of problem in assessing subspeciation in Peruvian Boa constrictor
is the lack of certain geographic ranges for the known taxa. Stimson
(1969) and Peters and Orejas-Miranda (1970) both list the southernmost
range B. c. imperator and the only range of B. c. ortonii as
northwestern Peru. Schmidt and Walker (1943), list only B. c.
ortonii as occurring from Piura south to Libertad. Assuming
acceptance of the validity of ortonii,
only the province of Tumbes lies outside its range on the northwest
coastline of Peru, and the snake inhabiting this province is unique.
We would accord it specific status, were we absolutely certain that no
intergradation with B. c. imperator occurs, although none has been
documented.
With regard to Boa constrictor melanogastor, Langhammer’s (1983)
poor taxonomic procedure renders this name a nomen dubium, although it may
well be a recognizable taxon. His most egregious error is his
blatant disregard for the high dorsal scale counts of B. c. constrictor
documented by such respected workers as Lazell (1964), Dixon and Soini
(1977, x=929.2 for seven specimens) and Vanzolini, Ramos-Costa and Vitt
(1980). These counts were later supported by Chippaux (1986) who
examined 15 specimens from Guyana with 91 to 95 mid-dorsal scales.
Langhammer stated that snakes with dorsal counts in excess of 88 were most
likely misidentified melanogastor. The inference that melanogastor
inhabits Guyana and Caatingas, Brazil as well as Ecuador is untenable.
Also worthy of note is Langhammer’s (1983) contention that melanogastor
has mid-dorsal saddles only 4-7 scales in length; his photograph of the
holotype shows that while this is true for saddle number 10, saddles
number 6 and 7 cover 10 scales, and saddle number twelve covers 8 scales.
Langhammer admits to being unsure of the range of melanogastor; he
states (1983) that the range should overlap into the Amazonian areas of
northern Peru, where Dixon and Soini (1977) have, in fact, documented the
occurrence of subspecies of constrictor. Finally, we are left
with no criterion to distinguish this population from constrictor
except the dark ventral surface, whose status at the area of
intergradation with constrictor remains uncertain. Nonetheless,
Langhammer's tables of meristic counts are fairly complete and well
documented. His assertion that imperator has 22 or more dorsal saddles is supported
by hundreds of observations, however, his support for the traditional
assumption that B. c. constrictor
has 21 or fewer saddles is not entirely correct, as we have seen 2
specimens from Colombia with 22. This being the case, ortonii,
which have 15-19 body saddles and the Tumbes "black boas" with
19-21 cannot be imperator sensu
Langhammer (1983). The inference that imperator occurs in
Peru at all thus lacks support. Its retention as a Peruvian taxon in
recent accounts (Langhammer, 1983; Peters and Orejas-Miranda 1970) is thus
either artifactual or suppositional, as the Tumbes population was not well
known until 1987. We thus propose a restriction of the range of imperator
to southern coastal Ecuador, an area poorly represented in
museum collections. With
regard to the "black boas" of Tumbes Province, Peru, the data
for taxonomic recognition is much more convincing. These snakes
differ in numerous respects from both subspecies imperator and
constrictor, and are isolated from the latter by mountains almost 3000
meters feet in elevation. We thus describe them below as: Boa
constrictor longicauda subsp.
nov. Holotype.
--Museum of Comparative Zoology (MCZ) 176002, a subadult male, was
collected on June 14, 1988, east of Tumbes, Tumbes Province, Peru, by a
native collector. Paratypes.--MCZ
176003, an adult male (illustrated on cover plate), collection data same
as for holotype, is represented by a complete shed skin, the specimen
still extant (RMP 551, collection of the senior author) to be donated to
MCZ upon its demise. RMP 553, a subadult male, was collected
in 1989 in Tumbes Province, Peru, obtained from importer, no further data
available. Diagnosis.
--Boa constrictor longicauda
may be distinguished from all conspecifics by the much longer tail (in the
male) and hemipenis. Additionally, it may be distinguished from B.
c. imperator by fewer dorsal body blotches (20-21 in longicauda,
22 or more in imperator), the
longitudinal mid-dorsal band without projections to the eyes, and the
generally darker head and body coloration, without tan or red color on the
tail in adults; from B. c.
constrictor by lack of red coloration on the tail, fewer midbody scale
rows (a maximum of 76, usually fewer in longicauda,
a minimum of 81, usually several more in constrictor),
the darker body pattern without tan color, and the grey, black-spotted
head in adults; from B. c. ortonii by
the much darker overall coloration, lack of red tail color and lower
ventral count (246-252 in ortonii); from B. c. melanogastor (sensu Langhammer, 1983) by the much longer tail,
lack of melanistic ventral coloration, and fewer middorsal scale rows (86
or more in melanogastor); from B. c. amarali by the overall darker coloration, greater number of
subcaudals (43-52 in amarali), and fewer dorsal body saddles (22 or
more in amarali); and from B.
c. occidentalis by the
greater number of subcaudals (45-46 in occidentalis),
and the very different color patterns (a blackish network or reticulum
on a cream ground color in occidentalis).
Description of Holotype. --The following description was made postmortem
on the as yet unpreserved specimen. The holotype is a small adult
male 159.5 cm in length with a 23.6 cm tail. The tail/total length
ratio of 0.148 is the longest documented for any specimen of Boa. The head is grey with a few black-specks and a mid-dorsal
longitudinal spear-shaped black band which bulges laterally at the level
of the eyes. It has thick black pre-ocular stripe from the nostril
narrowing at the eye, and a post ocular stripe underlined with a narrow
white band extending past the angle of the jaw. The tip of the snout
is slightly abraded. It has 19 mid-dorsal body saddles, 7-10 scales
in length at midbody. The first 4 saddles are connected by lateral
bands. These saddles are almost entirely black, numbers 4 to 14
having elongate white patches in their lateral expansions. There are
23 lateral body blotches, charcoal gray and 5 scales long anteriorly,
originating on the third or fourth scale row, becoming progressively
darker and longer posteriorly, reaching a maximum of 18 scales in length,
originating on scale row 1 posteriorly. Several of the
posterior body blotches have white centers. The flanks are grey, the
interspaces between the saddles dark brown. The ventral ground color
is cream and almost without spotting on the anterior 1/3 of the body,
laterally spotted and flecked with black on the middle 1/3, and heavily
spotted with black on the posterior 1/3 of the body. 'There are 4
black dorsal tall blotches with a black tail spine. The underside of
the tall is cream, with 4 large black spots, each 4 to 4 1/2 subcaudals
long. There is no red coloration on the tail. The ground color
of the tail is yellow. There
are 247 ventrals and 62 subcaudals. The dorsal scutellation is 54 at
the neck, 66 at midbody, and 38 at the tail. There are 20
supralabials and 22 infralabials. The infralabials are lightly
flecked with black. The hemipenis probed 40 subcaudal scales while
the animal was alive, and the right organ subtended 29 subcaudals when
everted (but not fully turgid, retractor muscle not cut). The left
hemipenis is partially everted. Remarks.
--The holotype is one of only 2 specimens with 19 dorsal saddles; it
appears to be congenitally missing the first dorsal body saddle. Variation.
-A total of IO males and IO females were included in this study.
Males mature sexually at about 150 cm, and both sexes attain lengths in
excess of 280 cm. There are 19 to 21 black dorsal body saddles, 7-10
scales in length at midbody, often joined anteriorly by lateral stripes,
but separate posteriorly. There are 4 to 6 black or
brown-black dorsal tall blotches. The head surface ranges from gray
to charcoal with a coal black longitudinal middorsal spear shaped band
which may send small lateral projections toward, but not reaching the
eyes, and there may be a sub ocular black projection not reaching the
upper labials. Suffusion of black on the head, snout, and
infralabials is highly variable, the paratype having numerous dark flecks
and spots. There is pronounced ontogenetic variation in the color of
the flanks; juveniles having brown flanks which change to gray or black in
a period of less than two years as maturity is reached. Adults may
develop extensive white patches on the flanks and in the lateral aspects
of the saddles. The interspaces between the saddles are brown in
younger specimens, but become strongly suffused with black in larger
snakes (see illustration of paratype). There are 50 to 58 (x=54.0)
dorsal scales around the neck, 60 to 76 (x=67.7) at midbody, and 32 to 40
(x=37.9) posteriorly. Ventrals range from 223 to 247 (x=239.2), and
subcaudals from 60 to 67 (x=62.8) in males and 50 to 54 (x=52.7) in
females. The tall is long, comprising 11.1 to 12.0 percent of the
total body length in females, and 12.8 to 14.8 percent in males.
There is significant sexual dimorphism in tail length. The hemipenis
probes 36 to 40 subcaudals.
Distribution. --B. c. longicauda
is known only from Tumbes Province, Peru, the only coastal tropical wet
forest in Peru.
Etymology. --The name longicauda is
Latin for "long-tailed."
Discussion
Boa constrictor longicauda has a longer tail than other subspecies,
and there is significant sexual dimorphism at the .02 level by Student's t
test, the males' tails averaging 14.1% of total body length, the females
11.6%. Although the literature contains inadequate analyses of tail
length to total length ratio in other populations to make a statistically
valid statement, it is worthy of note that there are no records for other
males with tails as long as B. c. longicauda. Dixon and Soini (1977) list figures of 11.9
and 11.4% for males and females of B.
c. constrictor from Iquitos, Peru, although their samples were quite
small. We have found similar measurements in Colombian constrictor.
Our ratios for 3 male specimens of Colombian imperator and 2 ortonii
fall between 10.6 and 11.0%.
In the past many workers considered Boa to possess a proportionately small hemipenis. This may be
a result of' Cope's (1900) often cited 85 mm hemipenial illustration of a
specimen from Brazil. Cope died before the plate section of
this text was finished, however, and there is no further data or scale
included with the drawing. The authors have observed an everted
hemipenis estimated to be in excess of 220 mm on the specimen herein
illustrated. Inverted, the same organ subtended 38 subcaudals on the
221 cm specimen. The maximum hemipenial probes we have recorded for
other subspecies are: constrictor, 27
subcaudals; imperator, 23 subcaudals; melanogastor, 28 subcaudals; ortonii,
14 subcaudals. Given the differences, it is questionable whether
large male longicauda could successfully copulate with other
subspecies. The single male observed in copula shows no interest in
females of other subspecies.
Boa constrictor longicauda is
almost certainly confined to Tumbes Province, a wet tropical refugium
surrounded to the north, east, and south by cordillera rising to at least
3000 km. The only possible access for other subspecies to Tumbes
would be by Pacific coastal drift from Ecuador. The large Peruvian
collection at the Museum of Comparative Zoology at Harvard University
includes no Boa specimens from Tumbes or adjacent southern
provinces of Ecuador (Jose Rosado, pers. comm.).
While the phylogeny of the subtaxa of Boa remains speculative,
Stull (1932) implies the existence of a constrictor group within the genus, with amarali intermediate
between constrictor and occidentalis
geographically and with respect to the mental shield. Melanogastor sensu Langhammer (1983), must be added to this group
based on meristic and mensural characters. An imperator
group is also implied by Schmidt and Walker (1943) and Zweifel (1960)
consisting of at least subspecies imperator, sigma,
and ortonii. The derivation of the insular orophias
and nebulosa is problematical; Lazell (1964) correctly points out that
their geographically closest neighbor is constrictor
of Trinidad and Tobago, some 350 kilometers to the south, and that a
stepped cline series exists with respect to meristic and mensural
characters. It must be bourne in mind that the Isthmus of Panama has
only existed for 3-4 million years in recent times, therefore imperator
may be a relative newcomer to Central America and Mexico. The
likelihood that it crossed over 2000 kilometers of the Caribbean to reach
St. Lucia and Dominica in such a short time is slim. The likelihood
of a recent relationship between imperator
and longicauda is greater, as their ranges approach parapatry and the
scale meristics are close. Ultimately, many of these relationships
may only be resolved through biochemical analyses, if at all. Specimens Examined.--With the exception of the holotype, all 20 specimens are currently in private collections, including 7 in the collections of the authors. All counts and measurements were done at least two times on living snakes except for the holotype and shed skin of the paratype. [Back To Top] [Back To The Boas] [Main Page]
1991 Male - Photo courtesy of Tim Meade. 1991 Female - Photo courtesy of Tim Meade. 1992 Female - Photo courtesy of Tim Meade. Wild caught male - Photo courtesy of Tim Meade. [Back To Top] [Back To The Boas] [Main Page] |