B. c. melanogaster
[Main Page]


Natural History



Back To The Boas


We wish to thank Jim Pomaville, Jr. for kindly providing this article.

A New Subspecies of
Boa Constrictor,

  Boa constrictor

  from Ecuador
(Serpentes: Boidae)

Detroit Zoological Park
P.O. Box 39, Royal Oak, Michigan 48068



The species Boa constrictor as currently recognized, with its many and varied mainland and insular races, has one of the largest distributions among land-dwelling vertebrates.  Almost central to the historically accepted range of the species and nestled against the eastern flank of the Andes in Ecuador is a little-known population that has the highest meristic count for circumferential body scales and a readily recognizable and distinctive color and pattern.  This subspecies is herein named Boa constrictor melanogaster and is compared to all its conspecifics recognized as valid taxa in the last century.

The boa constrictor of legend and of popular literary and cinematographic infamy bears little relation to the peaceful and handsome creature so familiar to hobbyists and to scientists who have knowledge of the herpetofauna of tropical America.  The species' formal description is attributed to Linnaeus as Boa constrictor in 1758, although a long list of literary references dating from 1648 is given by Bibron and Dumeril (1844).  Since 1758 the species has been juggled between the generic assignments of Boa and Constrictor on a frequent basis as the result of a probably irresolvable difference of opinion as to the systematic relationships of the American tree boas, Corallus caninus, C. enydris, and C. annulatus (Forcart, 1951).  For the moment I accept the present consensus that Boa is the best generic assignment for the boa constrictor.

Following Linnaeus's original description, a number of additional populations were given specific recognition but later were relegated to subspecific status when the mainland ranges were found to be continuous so that, with the exception of the isolated insular races, a large clinal entity seems to exist.  Boa constrictor is a remarkably adaptable species and ranges from sea level to elevations of at least 350 meters (Lazell, 1964; Duellman, 1978), with its southernmost range approximately 35degrees South latitude in Argentina and the northernmost range nearly 30degrees North latitude in western Mexico.  Cope (1877) records B. c. ortonii ( = B. c. imperator) from 3000 feet (923 meters), which, if correct, is probably far above the habitat of most populations of the species.  Although never really aquatic, the species inhabits wet tropical jungles, rain forests, xeric  (dry) jungles  (Yucatan), and even the deserts of Sonora (Mexico) and coastal Peru, where remarkable upper and lower extremes of temperature tolerance are required.

Repeated trips to Ecuador have drawn my attention to a distinctive population of Boa constrictor that needs formal recognition-not only to facilitate the scientific study of the species and its ecological associates, but also to provide a reference to a race that has assumed the dubious distinction of being the most desirable and valuable boa constrictor to reptile collectors and exhibitors. Hobbyists and zoo personnel have long been familiar with a rarely available yet beautiful and highly desirable phenotype of the boa constrictor usually referred to as the "red-tailed boa" in the trade.  This phenotype is sold at prices four to ten times higher than other Amazonian boa constrictors.  The first published photos of animals I would judge to be representative of this phenotype are in Ditmars (1931, upper photos in plates 6 and 7).

Prior to my trip to Ecuador in 1968, I'm not sure anyone realized that the so-called "red-tailed boas" were actually a stable population rather than exceptionally handsome individuals of an extremely variable species.  Since neither a formal description nor a valid scientific name has ever existed for this population, the trade has sold these animals under various names such as "red-tailed boas" or "Peruvian red-tailed boas." Erroneously, the specimens were labeled as Boa constrictor rubricauda (a name having no apparent validity) or as B. c. ortonii or B. c. amarali (both legitimate scientific names but for pallid races totally different from the strongly patterned red-tailed creature being offered for sale).

To further complicate matters, the boas offered for sale as "red-tailed boas" actually seem to represent two different stocks.  The one described herein as B. c. melanogaster has a deep mahogany-red coloration to the tail blotches, but the other has a bright orange-red color to the tail on a body typical in all other respects of the boas of the Amazon basin referred to Boa constrictor constrictor.  For this reason I have decided to avoid the confusion that continued use of the name "red-tailed" would be certain to engender among collectors and have named the Ecuadorian race the black-bellied boa, based upon a character quite reliably used to differentiate it from the eastern Amazonian population.  With a formal name available and access to the identifying characteristics provided herein, hobbyists should be able to be certain that they are getting what they are paying for, and both the hobby and the scientific communities will hopefully have a better understanding and appreciation of the entire species complex.

  Boa constrictor
melanogaster, SUBSP.  NOV.
(Figs. 1, 2 & 3)

  Holotype. -Adult female, University of Michigan Museum of Zoology (UMMZ) 172680, total length 233 cm.  Collected by Jivaro Indians in December, 1968, near the Rio Yaupi Catholic Mission upstream from the Rio Santiago in the Morona Santiago province of Ecuador.

  Topoparatypes. - UMMZ 172676-172679 and 172681; also University of Kansas Museum of Natural History (KU) 192081-192083.

  Etymology.-The subspecific name melanogaster literally translated means "black belly" and refers to the nearly uniform distribution of black pigmentation over the ventral surfaces of adults of this race.

  Diagnosis.-Boa constrictor melanogaster has the smallest and the most numerous dorsal body scales of all the conspecifics, ranging from 86-94 in circumferential count.  The dorsal pattern between head and anus consists of 20-21 dark blotches or saddles that at their narrowest (usually the tenth blotch behind the head) are typically only 4-7 scale lengths long at midbody (Fig. 3).  These dorsal body blotches are usually very dark blackish brown on a gray to pale yellow ground and typically show no light-colored centers in the area dorsal to the vertebral column.  Shortly before the anus, the dorsal blotches become elongate and are almost circular in outline on the tail.  These posterior dorsal blotches and their lateral counterparts are usually black-bordered with mahogany-red centers.  The cream-colored ventral surface in juveniles is uniformly black-spotted overall; however, the black spots expand and coalesce as individuals grow until the adult animal is characterized by an almost uniformly black venter (Fig. 2), especially posteriorly.  In very large specimens (3 meters and above), the dorsum also is invaded by black and the animal may appear melanistic overall.

Description of the holotype (Figs. 1, 2, and 3).-UMMZ 172680 is an adult female of total length 233 cm and has a complete tail.  Circumferentially there are 72 neck scales, 94 scales at the level of the tenth dark dorsal blotch behind the head, and 45 scales at the anus.  There are 244 ventral scutes and 49 subcaudals.  The tenth dark dorsal blotch is 5 scale lengths long at its narrowest transect.

  Variation.-Scutellation varies among the nine specimens (the holotype and eight topoparatypes) as follows: circumferential body scales at the neck 69-74 (avg. 71.1), at midbody (= tenth dorsal dark blotch) 86-94 (avg. 89.6), at the anus 43-49 (avg. 46); ventral scutes 237-252 (avg. 241.5); subcaudal scutes 45-54 (avg. 49.7). The tenth dark dorsal blotch varies in its narrowest transect from 4-7 scale lengths (avg. 5.7).

  Distribution.-Although all the type specimens came from the vicinity of the convergence of the Rio Yaupi and the Rio Santiago in the Morona Santiago province in Ecuador, the author has seen and examined identical material from Tena on the Rio Napo in the Napo province of Ecuador.  Also, the photograph on page 225 of Duellman (1978) would indicate that the range continues northward in the Napo province to Santa Cecilia on the Rio Aguarico.

The Amazonian jungles of eastern Ecuador are still relatively inaccessible, so the eastern limits of the range of B. c. melanogaster will remain unknown for some time.  It can be expected that the range will overlap into the adjacent Amazonian areas of southern Colombia and northern Peru.


  The relationships of the races or subspecies of Boa constrictor to one another can only be understood if one realizes that all the mainland races form a continuous clinal entity-each race merges genetically with one or more adjacent races.  In the traditional sense, only the insular races B. c. sigma, B. c. sabogae, B. c. orophias, and B. c. nebulosa are perhaps worthy of subspecific status, having gene pools isolated from continuous genetic drift.  Even so, these isolated gene pools are undoubtedly diluted and in all probability were established when tropical storms carried specimens of mainland Boa constrictor to these remote habitats.  The following remarks when used with the key and Table 1 should help considerably to clarify the relationships and physical characteristics of the published subspecies of Boa constrictor.

  (1) B. c. constrictor Linnaeus, 1758 is the only subspecies with which B. c. melanogaster is likely to be confused.  B. c. melanogaster differs from this subspecies by having more dorsal scale rows at midbody, 86-94 (avg. 89.6) versus 81-89; by having the tenth dark dorsal blotch only 4-7 (avg. 5.7) scale lengths long and uninterrupted middorsally by pale pattern elements (Fig. 3), versus 10 or more scale lengths long and with one or more pale pattern elements included on or near the vertebral midline within the dorsal blotches (Fig. 4); by having the ventral surface uniformly covered with black spots or even totally black in larger specimens (Fig. 2), versus the ventral surface being predominantly white or cream-yellow with only a few small black spots, except where the lowermost lateral dark blotches invade the ventrum and break up into clusters of relatively larger dark spots (Fig. 5).

  In most of the literature, the higher midbody scale counts for B. c. constrictor of up to 95 (Boulenger, 1893; Lazell, 1964) are suspect as being representative of specimens now being assigned to B c. melanogaster.  No specimens of B. c. constrictor known to this author exceed 88 dorsal scale rows, and Stull (1932) accepted 89 as the upper limit for this taxon without dealing with the higher counts of earlier authors.

Unquestionably, B. c. constrictor is the race to the east of B. c. melanogaster and probably to the north as well, separating it from B. c. imperator.

  (2)  B. c. occidentalis (Philippi, 1873) is a southern neighbor in Bolivia and Argentina of B. c. melanogaster and, due to the upper range of its dorsal scale counts (reported to be as high as 87 but unverified by me), populational intermediates may exist.  When compared to B. c. melanogaster, the more typical lower range of the extremely variable (65-87, per Boulenger, 1893) dorsal scale counts and the always higher (22-30) dorsal blotch count of B. c. occidentalis suggest to me that B. C. imperator may have in ancient times moved across the Andes from the west and become the phylogenetic source of B. c. occidentalis and B. c. amarali.  The geographical barrier of the present Andean range raises formidable doubts to such a hypothesis, and I am not sufficiently knowledgeable to defend the hypothesis myself.  Schmidt and Walker (1943) suggest that Peruvian faunal elements are known to have moved eastward from coastal Piura and Lambayeque and westward from Amazonian Cajamarca through Andean passes below 8,000 feet (=2,500 meters).  These particular passes are too far north to support the present hypothesis, which probably would require passage across the Andes farther to the south, perhaps in Chile (south of 35 degrees South latitude) where the Andean chain once again is crossed by passes at low elevations (Schmidt and Walker, 1943; Duellman, 1979).  This latter possibility means, of course, that historically the coastal population of B. c. imperator had to have ranged much farther to the south than its present terminus near Libertad, Peru, which does not seem implausible to me.

In B. c. occidentalis the dorsal blotches almost always include light middorsal pattern elements, and the blotches are interconnected with one another and with the lower lateral blotches to form a dark reticular pattern of great complexity (Fig. 6).  Occasionally in B. c. melanogaster the dorsal blotches coalesce dorsolaterally to form a simple chain-like pattern (Fig. 1).

  (3)  B. c. amarali (Stull, 1932) is known from southern Brazil, eastern Bolivia, and Paraguay and would not seem to be related to B. c. melanogaster, having only 71-79 dorsal scale rows versus 86-94 in the latter.  However, one of Stull's paratypes (UMMZ 68005) of B. c. amarali from Buena Vista, Department Santa Cruz, Bolivia, has the distinctive dorsal pattern of B. c. melanogaster, although in scale counts it conforms to its type series.  B. c. amarali typically also differs from B. c. melanogaster in its pallid coloration and in having more numerous (22 or more) dorsal blotches (Fig. 7) At middorsum the blotches are usually more than 10 scale lengths long, versus 4-7 in B. c. melanogaster.

  (4)  B. c. imperator (Daudin, 1803) probably has no direct genetic influence on  B. c. melanogaster, with which it shares no common borders to their ranges.  B. c. imperator inhabits the tropical Pacific coast of South America northward throughout most of Central America and Mexico.  B. c. imperator almost certainly is the original stock for the taxa that follow (# 5-13), with all of which it shares proximal ranges, 22 or more dorsal blotches, and fewer than 80 dorsal scale rows.  All are easily differentiated from B. c. melanogaster, which has 21 or fewer dorsal blotches and 86 or more dorsal scale rows.  See the accompanying key for the criteria to differentiate those latter taxa that are considered valid.

  (5)  B. c. orophias (Linnaeus, 1758) is an insular race restricted to Saint Lucia in the Caribbean.  Much confusion exists in the older literature regarding its distribution, but Lazell (1964) cleared this up to the satisfaction of most workers.  See the key and Table 1 for diagnostic characters.

  (6)  B. c. diviniloqua (Dumeril and Bibron, 1844) is considered by Lazell (1964) to be synonymous with B. c. orophias.

  (7)  B. c. nebulosa Lazell, 1964 is a distinctive insular race restricted to Dominica in the Caribbean.  See the key and Table I for diagnostic characters.

  (8)  B. c. sigma (Smith, 1943) is restricted to the Tres Marias Islands, Mexico.  Nine specimens were known in 1963 (Smith, personal correspondence), with ventral scale counts of 253 to 260, slightly in excess of the high for B. c. imperator of 253. (Stull, 1932).  Zweifel (1960) - synonymized it with B. c. imperator, but Stull still recognized it in 1964 (personal correspondence), as did Smith in 1963 (personal correspondence).

It would be wise to evaluate carefully the role of elevated ventral and subcaudal scale counts as the sole basis for taxonomically recognizing insular populations of snakes such as B. c. sigma.  Bogert and Oliver (1945, pp. 360-1) and Zweifel (1960) believe the character to be unreliable.  The embryogeny of many cold-blooded vertebrates is known to be affected by minor temperature differences between litters, with resultant differences of sex ratio, vertebral elements, and scale meristics.  Perhaps the differences in ventral scale counts seen in B. c. sigma, B. c. orophias, and B. c. nebulosa may have a basis in microclimatic conditions and might disappear if the different stocks were reproduced under similarly controlled conditions.  Although such elevated counts are not generally recognized in mainland populations of B. c. imperator, Smith (1943, 1964) mentions an odd specimen with 261 ventrals from Tehuantepec and thus undermines his own reluctance to merge B c. sigma into B. c. imperator.

  (9)  B. c. sabogae (Barbour, 1906) from Taboga Island, Panama, differs from the typical mainland B. c. imperator only in being dark reddish brown with the pattern largely obscured.  Stull (personal correspondence, 1964) found it indistinguishable from some aberrant mainland individuals and felt it should be synonymized with B. c. imperator.

  (10)  B. c. var. isthmica Garman, 1883 is considered a synonym of B. c. imperator by Peters and Orejas-Miranda (1970).  It was described from the Isthmus of Darien, Panama.

(11)  B. c. mexicana (Jan, 1863) was described from a single specimen from Mexico with 55 dorsal scale rows but which was otherwise like B. c. imperator.  Most authors have synonymized this taxon with B. c. imperator on the assumption that the type was an atypical specimen or was simply miscounted.  Smith (1963) states that no Mexican specimens of known provenance have ever shown 55 dorsal scale rows.  However, it appears that most workers have missed Andrew's report (1937) of four Mexican specimens (Field Museum 26989-92) from Chichen Itza having dorsals of 56 to 62.  These counts completely bridge the gap between B. c. mexicana (= 55) and B. c. imperator (= 61-79 in Boulenger, 1893) and should allay any remaining doubts that B. c. mexicana should be suppressed.

  (12)  B. c. ortonii Cope, 1877 is a desert population of pallid coloration found from Perico and the Upper Maranon valley to Piura and south to Libertad in northwestern Peru.  It was thought by Cope to be different from B. c. imperator in having a higher ventral count which Schmidt and Walker (1943) give as 246-252, well within the range of B. c. imperator's 225-252 (Boulenger, 1893) or 235-253 (Stull, 1932)-and in having small scales separating the suborbital and labial scales-which Schmidt and Walker found to be inconsistent in their study.  Stull (personal correspondence, 1964) considers it to be a synonym of B. c. imperator, which seems logical without additional data.

  (13)  B. c. eques (Eydoux and Souleyet, 1842) is based on a single specimen and is almost certainly an aberrant B. c. imperator, with which it has been synonymized by most workers.  The specimen came from Payta, Peru, and was characterized by one preorbital scale that was larger than all other scales in the orbital ring and equal in size to the eye itself.  Other scales of the head were mentioned as being unusually large also.  Since no further examples are known, it seems not to be a valid taxon.

  Before terminating this brief discussion of the subspecific taxa of Boa constrictor, I would like to suggest that Lazell (1964) was almost certainly in error in believing that an evolutionary continuity existed as a stepped-cline series from B. c. constrictor through  B. c. orophias to B. c. nebulosa.  The latter two insular populations are admittedly distinct subspecific taxa, but Lazell somehow failed to discuss their potential derivation from mainland populations of B. c. imperator to the west.  B. c. imperator is much more closely allied to the two Lesser Antillean taxa in Lazell's critical characters of dorsal scale rows, dorsal blotches, coloration, and patterning than is B. c. constrictor.  Furthermore, the insular B. c. sigma off the Pacific coast of Mexico shares the elevated ventral scale counts of B. c. orophias and B. c. nebulosa.  In my opinion the northward migration of B. c. constrictor ended with its colonization of Trinidad and Tobago.

Table 1. Comparison of subspecies and synonyms of Boa constrictor discussed in the text

Subspecies        Dorsal  Scales Ventral
No. of
 blotches between head  and anus 
Mid-dorsal length of blotches expressed as scale lengths Comments on color, pattern and taxonomic status
B. c. melanogaster 86-94 237-252 45-54 20-21 4-7 Dorsal blotches dark chocolate, much narrowed in length mid-dorsally, occasionally extended dorsolaterally and connected into a chain-like pattern; ground color pale gray or yellow; caudal blotches mahogany-red often blackening with age; ventrum uniformly black-spotted, usually becoming solid black posteriorly with age.
B. c. constrictor 81-87





10 or more Dorsal blotches brown, squarish, and usually including several pale pattern elements; ground color very variable—yellow, pale brown, gray, and even occasionally pinkish: caudal blotches
usually brown bordered with black but
occasionally orange-red; ventrum predominantly cream with only scattered black spots; highly variable subspecies, perhaps divisible, especially in northeastern South America. The higher dorsal scale counts are almost certainly specimens of B. c. melanogaster.
B. C. occidentalis 65-87 242-251 45 22 or more

Blothes rarely regular in outline

Dorsal pattern usually a blackish network or reticular pattern of great complexity on a gray or cream background; caudal blotches usually black; ventrum strongly invaded by black.
B. c. amarali 71-79 226-237 43-52 22 or more

10 or more

Pattern basically shades of gray—essentially no brown; black present usually only as spotting scattered on dorsum and ventrum; ventrum an unusual lichenose pattern of variably intense gray spots.
B. c. imperator 56-79






10 or more

Blotches usually brown to black-brown on a light brown to gray ground; caudal blotches usually brown; ventrum usually heavily black-spotted on a cream or occasionally pinkish orange ground.
B. c. orophias 65-75






  Subrectangular, dark dorsal saddles on a dorsal ground color of rich brown; venter is white with black or gray spotting pronounced.
B. c. diviniloqua           Synonymous with B. c. orophias
B. c. nebulosa 59-69 258-273   32-35   Dorsal blotches irregular in shape, scarcely darker than ground color anteriorly but becoming increasingly distinct in outline and color contrast posteriorly; ventrum is ash gray becoming slate gray to black posteriorly, black-spotted where contrast permits detection.
B. c. sigma 77 253-260 55-66 30(4)

10 or more

Although identical in almost all respects with B. c. imperator, the subspecies B. c. sigma is re­tained by most workers because of its insular isolation and higher ventral scale count. How­ever, I agree with Zweifel (1960) that B. c. sigma should be synonymized with B. c. imperator.
B. c. sabogae 65-67





    Synonymous with B. c. imperator.
B. c. var. lsthmica 78 246 57 25   Synonymous with B. c. imperator.
B. c. mexicana 55         Synonymous with B. c. imperator.
B. c. ortonii 57-72 248-252 46-59     Synonymous with B. c. imperator.
B. c. eques 65 240 55  

Synonymous with B. c. imperator.
(1)Boulenger, 1893; (2)Barbour, 1906; (3)Stull, 1932; (4)Zweifel, 1960 (plate 44); (5)Lazell, 1964.

[Back To Top] [Back To The Boas] [Main Page]



Bcmelanogaster1.jpg (26424 bytes)

Photo courtesy of Craig McSherry.  Taken in May of 1982, this 10 year old docile male was about 10 feet long.  It had a jet black belly with very little speckling.

Photo courtesy of Craig McSherry.  Taken in May of 1982, this 10 year old vicious female was 14+ feet long.  She had a black belly except in the chin area where the first 12 inches was 50% cream color.

This female (next four photos) was taken out of the wild in April, 1999 by John Johnson and his wife of TN.  She was found near the Aquaro River close to Santa Cecilia in the Napo Province of Ecuador.  She is 8 feet in lenght and feeding on rabbits. Her scale counts are:
Dorsal Rows:  89, Ventrals:  242, Subcaudals:  48, Saddles to Vent:   20.  J. Johnson is currently searching for a male to breed this female with. (J. Johnson, pers comm.)

JJMelanogaster.jpg (42912 bytes)

JJMelanogaster2.jpg (47372 bytes)

JJMelanogaster3.jpg (67367 bytes)

JJMelaonogaster4.jpg (42822 bytes)