Agama picta Wied, 1823

Vanzolini, Paulo E. & Myers, Charles W., 2015, The Herpetological Collection Of Maximilian, Prince Of Wied (1782 - 1867), With Special Reference To Brazilian Materials, Bulletin of the American Museum of Natural History 2015 (395), pp. 1-155 : 31-37

Agama picta Wied, 1823

Plate 10 and figures 5 – 8B

1823 Abbildungen: Lief. 3.

1824 Isis : 663 (listed, with reference to the Abbildungen).

1825 Beitra¨ge: 125, 604.

PRESENT STATUS: Enyalis pictus (Schinz, 1822) .

Agama picta Wied, 1823 , was placed as a species of Enyalius by Etheridge (1969: 240), and changed to subspecific status by Jackson (1978: 21). Because Wied’s name subsequently was shown by Myers et al. (2011) to be an objective junior synonym of Agama picta Schinz, 1822 , the current name of this taxon becomes Enyalius catenatus pictus (Schinz, 1822) . Only the senior authorship is changed; ‘‘the original concept of the taxon is that of Wied (1823, Lief. 3; 1825a: 125, 604)’’ fide Myers et al. (2011: 7).

REMARKS: Agama picta was in recent years dated from the 1825 Beitra¨ge ( Etheridge, 1969: 240; Jackson, 1978: 21), but the 1823 plate and accompanying text diagnosis and description clearly have precedence; Etheridge (1970: 118) later correctly gave the Abbildungen as source but incorrectly retained the 1825 date.

The Maximilian collection at the American Museum contains but a single specimen of Enyalius, AMNH R-108, from which all color pattern has been bleached after nearly two centuries in alcohol. More than a little confusion surrounds this poor specimen, as outlined below, following a brief description. It should become evident to the reader that intraspecific variation and species limits in Enyalius are still not well understood.

DESCRIPTION OF AMNH R-108: It is a female, with convoluted oviducts but lacking enlarged ova; the body has been previously opened. The specimen is very soft and has lost all vestiges of color pattern. Size: 90 mm SVL + 197 mm tail (including brokenoff piece of 145 mm) 5 287 mm total length (tail 2.19 X SVL); tibia length 23 mm, tibia/ SVL 5 0.26. Canthal ridge virtually straight, barely curved toward midline anteriorly; anteriormost canthal scale somewhat approaching upper edge of nasal scale. Supraoculars smooth, some moderately enlarged, approaching size of circumorbital scales. No enlarged suboculars. Midbody scales cannot be accurately counted owing to condition of specimen. A low vertebral crest starting at rear of head and continuing weakly onto base of tail: 54 enlarged vertebral crest scales (or 65 scales including small ones) to anterior edge of thigh held at right angle to body, or 74 scales (large and small) to level of rear edge of thigh. Small conical scales on sides of body, becoming distinctly larger, flattened, and virtually smooth dorsally on either side of vertebral crest. Ventral scales very weakly keeled on chest, becoming smooth over most of venter. Scales on underside of shank (infratibials) keeled, about as wide as long. Subdigital lamellae smooth, distal several divided. Caudal scales arranged in segments, with about 4–5 dorsal and 3 ventral scales per segment.

Because the specimen has lost all vestiges of color pattern, Maximilian’s lovely plates of picta and catenata (see pls. 10–11 herein) cannot be used to assign it, and other avenues must be explored.

ASSIGNMENT OF AMNH R-108: Although only this specimen is known to have reached the American Museum, Maximilian had collected at least four and probably at least five specimens of two species of ‘‘ Agama ,’’ which he named catenata and picta in 1821 and 1823, respectively. In 1860, Maximilian brought their identifications and then-current synonymies up to date in his manuscript catalog, as follows:

1.

2. Enyalius Wagl. rhombifer D. et B. f[ide] Wagl. ( Agama catenata Wied ). Lophyrus rhombifer Spix ). ([species no.] 380).

Uperanodon D.B. pictum D.B. ( Agama picta Wied ) ([species no.] 335).

AMNH R-108 was identified after (1) above, with the name Enyalius rhombifer entered in the AMNH book catalog for reptiles; of two ‘‘original’’ numbers shown, one (380) was Maximilian’s species number and the other (152) probably an older AMNH card catalog number preceding the first (1920) book catalog (Myers, 2000: 100). Myers examined the catalog entry in 1976, finding that someone had decided that AMNH R-108 was the type of Maximilian’s catenatus and that the catalog had been so marked. The original book entry had consequently been emended by attempting to erase ‘‘ rhombifer’’ (still legible under magnification) and superimposing the name ‘‘ catenatus ’’; the same hand added ‘‘Bahía’’ to the original entry ‘‘ Brazil.’’ 21

Etheridge (1969: 244, 246) examined AMNH R-108 for his revision of Enyalius and accepted the designation of holotype for E. catenatus . In his account for E. pictus, Etheridge (1969: 240) designated a neotype of pictus under the assumption that the holo-

21 These catalog emendations were not necessarily incorrect if the original entry rhombifer had been taken from the bottle containing Wied’s specimen, but that cannot be determined. It is not known when the emendations were made, but they were entered prior to about 1969, after which changes or data interpolations were made in pencil and usually dated and initialed (Myers, 2000: 101).

type of that species ‘‘apparently was in the American Museum...and is now lost.’’

In a subsequent revision of the genus, Jackson (1978: 19–20) also examined AMNH R-108, which he thought did not agree with his concept of catenatus . On the basis of the aforesaid emended catalog entry (Myers in letter to J.F. Jackson, Jan. 13, 1976), Jackson concluded that the specimen did not have type status and that the holotype of catenatus is apparently lost. Jackson (1978: 21) accepted also Etheridge’s conclusion that the holotype of E. pictus was likewise lost and stated that the neotype designated by Etheridge corresponds well with Wied’s illustration of pictus .

Jackson (1978: 19) further concluded that AMNH R-108 ‘‘is clearly a member of an unnamed taxon of which all other specimens have been collected in southeastern Brazil.’’ Jackson (1978: 26) assigned AMNH R-108 to his new Enyalius perditus .

Unfortunately, in accepting that AMNH R-108 either was, or was not, the holotype of A. catenatus , neither Etheridge nor Jackson seems to have compared it with specimens of the one other species described by Wied , namely A. picta . Jackson furthermore did not compare the specimen critically with specimens of his Enyalius perditus .

The specimen comes out readily to Enyalius pictus in Etheridge’s (1969: 255–256) generic key, or to E. catenatus if the faintly keeled chest scales count as ‘‘ventral scales keeled.’’ However, because of the relatively close approach of the anterior canthal ridge to the nasal scale, the specimen keys out to the later-described E. perditus of Jackson (1978: 27–28). Nonetheless, AMNH R-108 does not appear to be conspecific with Enyalius perditus Jackson , despite Jackson’s (1978: 26) flat-out claim that it is a specimen of that southern species, which claim was given as ‘‘further evidence that AMNH R-108 is not Wied’s type of [the more northern] catenatus since the type locality is specified as the interior of Bahia.’’ Jackson asserted that AMNH R-108 was a specimen of perditus but did not include it as a paratype or discuss the characters that led him to his conclusion. He may have been impressed mainly by the condition of the anterior part of the canthal ridge, inasmuch as several other characters are inconsistent with his description of perditus , which has, for example, keeled supraorbitals and keeled belly scales. We have compared AMNH R-108 with four specimens of E. perditus collected at the type locality Boracéia. 22 These specimens are immediately distinguishable from AMNH R- 108 by a character not used by Jackson—the sides of the body in perditus are covered with tiny conical granules that become slightly larger dorsally but remain conical or strongly keeled in the paravertebral region. In contrast, the dorsal scales on either side of the median crest in AMNH R-108 are larger, flattened, and virtually smooth—much as in a few available specimens of pictus from Bahia (AMNH R-131859–131860).

Wied illustrated, described, and measured one specimen of Agama picta , although there is heretofore unpublished indication that at least one other specimen might have been

22 AMNH R-120470 (ex MZUSP 38381), a paratype collected by B. Faria in 1975, and AMNH R-119750– 119752 collected by Myers in 1979.

preserved (see below). Wied’s published measurements (in Abbildungen and Beitra¨ge) convert to 86 + 189 5 275 mm, compared with our recent measurements of the nowflaccid specimen of 90 + 197 5 287 mm. The Abbildungen plate (see pl. 10) shows the lizard with a small erect gular sac, 23 described in the Beitra¨ge as loose skin under the throat that puffs up when the lizard is excited; the skin under the throat of AMNH R-108 seems loose enough to form a small fold in profile ( fig. 8 ). Wied implied that the belly scales are smooth when he mentioned that those on the chest are somewhat keeled, which is exactly the case in AMNH R-108. (In the Beitra¨ge,

23 Enyalius is supposed to lack a gular pouch according to Etheridge (1969: 256), but there is no mistaking the structure in plate 10.

Wied noted that the scales of Agama catenata are mainly as in A. picta , one of the few differences being that they are keeled on the under parts of catenata .)

We think that AMNH R-108 represents Wied’s Agama picta and, although we cannot say with absolute assurance that it is the same specimen illustrated and described, we provisionally regard it as the holotype of that species, thereby setting aside Etheridge’s (1969: 240) designation of a neotype (ICZN, 1999: article 75.8). This view could be disproved were it shown that the illustrated color pattern is actually confined to males as implied by Jackson (1978: 21–22), who described variant patterns in the few females available to him, in contrast to Etheridge (1969: 241) who stated that ‘‘The pattern is equally vivid in both sexes.’’

The plate is very colorful (see pl. 10), although the colors may now differ a bit from Wied’s field sketch painted at Morro d’Arara in March 1816, which is reproduced here (from Bosch, 1991: 239) as figure 8 . Wied’s descriptive field identification on the sketch reads ‘‘ Lacerta fasciata (m).’’ Wied’s de- termination of the specimen as male almost certainly was due to the small extended gular pouch, which seems about as large in AMNH R-108, a female. The gular pouch also is extended in one of two quick sketches of the head, which show that the tongue was orange like the gular pouch ( fig. 8 ). Wied’s field painting and the specimen were used by an illustrator (H.J. Beckers) to prepare the published plate (compare fig. 8 with pl. 10); the work was critiqued by Wied (in a back section of the Beitra¨ge 604), who commented that while the color and habitus were correct and the scales of the head were fairly accurate, the body scales were less accurately copied by the illustrator.

The above field sketch is accompanied in Bosch (1991: 239) by a separate study of a tail, which must have been drawn sometime later, after Wied’s return home to Neuwied since it is labeled Agama picta (vs. Lacerta fasciata on the field sketch). The tail sketch emphasizes an ‘‘error’’ in scutellation (a regeneration event). This cannot be matched on AMNH R-108, indicating that Wied had at least two specimens of A. picta . Possibly he interpreted this as a character and instructed his illustrator to include a drawing of the tail in the published plate. Although it is not a very good representation, the illustrator obviously saw nothing unusual, as would be expected if he were drawing from the present AMNH R-108.

TYPE LOCALITY: Wied had only a single locality for Agama picta , which therefore is the type locality. In the Abbildungen it was given as ‘‘ Morro d’Arara am Flusse Mucurí,’’ and in the Beitra¨ge as ‘‘ Lagoa d’Arara am Mucurí.’’ Etheridge (1969: 240) seemed to interpret ‘‘ Lagoa d’Arara am Mucurí’’ as two localities, and, on that basis, he incorrectly restricted the type locality to the town of Mucuri at the mouth of the Rio Mucuri (18 ° 059S, 39 ° 349W), some distance from the actual locality.

Morro d’Arara was on the Lagoa d’Arara —the names are interchangeable in Wied’s publications—above the lake’s outlet to the Rio Mucuri. At the time of Wied’s visit, Morro d’Arara was a new fazenda whose name meant ‘‘Araraberg’’ 24 (Reise 1: 249–250). According to Wied (Reise 1: 250, 252), Morro d’Arara was reached in a day and a half up the Rio Mucuri. Entrance to the Lagoa d’Arara was through a narrow, shady channel on the north side of the

24 Araraberg —hill of the araras, common name for the Scarlet Macaw ( Ara macao ). Wied had a long encampment at Morro d’Arara on Lagoa d’Arara (Feb. 5–July 23, 1816), which is the type locality of his Anolis viridis , Agama picta , Coluber formosus , and Rana macrocephala .

Mucuri, with Morro d’Arara another quarter hour up the Lagoa d’Arara —described as a fine lake of fair size, encompassed by forested mountains. Bokermann (1957: 224) placed Morro da Arara on the margin of the Rio Mucuri, about 30 km from its mouth, and Vanzolini (1992: 24) provided the approximate coordinates 18 ° 069S, 39 ° 489W.

The above account and the one following were prepared several years ago. Since then the phylogeny has been partially elucidated by Rodrigues et al., 2014. Subspecies of Enyalis are no longer recognized.

Figures

Fig. 5. Holotype of Agama picta Wied in dorsal view (AMNH R-108); 90 mm SVL. The once-vivid color pattern has completely faded.

Fig. 6. Holotype Agama picta Wied. Details of head in dorsal, lateral, and ventral view (AMNH R-108); 5 mm scale line.

Fig. 7. Holotype of Agama picta Wied. Details of hindquarters in ventral view (AMNH R-108); 5 mm scale line.

Fig. 8. Agama picta Wied. Inset (A) showing orange-pigmented tongue and small dewlap. Pen-andwatercolor illustration (B) made by Prince Maximilian at Morro d’Arara in March 1816, with descriptive field identification ‘‘Lacerta fasciata (m).’’ The Abbildungen plate (see pl. 10) was prepared from this sketch. Detailed study (C) of a tail, from a different specimen than shown in the life drawing. Wied’s handwritten explanation to artist (from typescript in Bosch, 1991: 238): Schwanz der Agama picta. um etwa die Vertheilung der Schuppen fu¨r den Kupferstecher genauer zu haben. In a. und b.[regenerating tail at break] sind die Schuppen aus Versehen verschieden gezeichnet, sie sind aber am ganzen Schwanze u¨berall gleich ross, klein und 4-eckig in Reihen gestellt. (‘‘Tail of Agama picta. In order to prepare for the copperplate engraving [see pl. 10], the scales are drawn differently in a and b; along the entire tail, however, the small square scales are equally disposed in rows.’’) From Bosch (1991: 238–239), courtesy of the Brazilian Library of Robert Bosch GmbH, Stuttgart.