Caecilia gracilis Shaw, 1802

Maciel, Adriano O. & Hoogmoed, Marinus S., 2011, Taxonomy and distribution of caecilian amphibians (Gymnophiona) of Brazilian Amazonia, with a key to their identification, Zootaxa 2984, pp. 1-53 : 6-9

Caecilia gracilis Shaw, 1802

Caecilia gracilis Shaw, 1802: 597 ; Dunn, 1942: 515; Taylor, 1968: 385; Hoogmoed, 1979: 273; Frost, 1985: 621; Azevedo- Ramos et al., 2004; Maciel et al., 2006; Frost, 2008.

Caecilia cf. gracilis Galatti et al., 2007:93 .

Diagnosis. Maximum known TL 735 mm (MPEG 1428), but most specimens less than 510 mm. PA 177–208, SG 11–52. Eyes in an open orbit or under bone. Dermal scales present; SI between annulus 41 and 141; at most two rows of scales per fold. Subdermal scales absent. Very slender Caecilia ; TL may reach 90 times BW. Terminal shield present.

Description. TL 30.9–90.6 times (63.6 ± 12.6; n = 95) BW. Head may be slightly narrower than body. Snout projecting distinctly beyond mouth (1.7 mm in MPEG 3898). Eyes generally visible, in an open orbit, but some specimens may have both (HUFMA 654; MPEG 1429) or one of the eyes covered by a fine bony layer (MPEG 1361 on the left side). Nuchal grooves may be distinct dorsally, laterally and ventrally, except third nuchal groove which ventrally is incomplete so that second collar ventrally is partially fused with the first primary annulus; a dorsal transverse groove is present on each collar, shorter and less distinct on the first; some specimens have indistinct or only partially distinct collars. Body subcylindrical, slightly wider than deep. Primary annular grooves incomplete, interrupted dorsally and ventrally, except for posteriormost primary annular grooves (at most 14, 9.7 ± 2.5; n = 12) which are complete; from general region where secondary grooves appear, to level of vent, primary annuli dorsally are complete. SG 11–52; at most nine secondary grooves complete (5.5 ± 2; n = 92). Vent transverse, may be slightly irregular; anal denticulations 9–16 (12.2 ± 5.2; n = 77); generally an approximately similar number in posterior and anterior edge of the vent (e.g. six anterior, six posterior in MPEG 16002). Paired anal papillae sometimes present; found only in two of the 43 males analyzed in a sample of 101 specimens. Unsegmented terminal shield posterior to the vent; slightly compressed in some specimens. SI between annulus 41 and 141 (89.5 ± 40.2; n = 4); at most two rows of dermal scales per fold. Dermal scales and number of rows per fold increasing, in size and in number, from anterior to posterior. Tongue anteriorly not completely attached to mandibular mucosa. PMT maximally 20, posterior maxillary teeth smaller, extending posteriorly of the level of the choanae, as PPT. PPT maximally 19, posterior ones smaller; teeth smaller than PMT. DT maximally 19, approximately same size as PMT; posterior ones smaller. ST maximally 6, approximately same size as PPT. Usually the highest number of ST in C. gracilis is four, except in one specimen (HUFMA 624) from a population in a transitional area between Amazonian forest and Cerrado in the state of Maranhão, Brazil, which has six.

Color. In life, specimens may be lavender, dark blue or grayish blue. The head is slightly paler than the rest of the body. The venter is slightly paler than the dorsum. Annular grooves in the posteriormost part of the body in a number of specimens with white lines. A narrow pale line links the eyes and the tentacles of some specimens.

Variation. No significant sexual dimorphism in morphometric and meristic characters (Maciel et al., in press). Comparisons of meristic and morphometric characters of populations reveal no great variations among sampled populations ( Table 1 ). Primary and secondary grooves counts vary slightly, with the most eastern population of C. gracilis (Urbano Santos, state of Maranhão) having a lower mean number of primary annuli than the other populations and even a lower minimum and maximum number for this character ( Table 1 ).

continued next page Amazonas Tocantins Amapá Maranhão Pará

TT 1.9 (1) 1.1–2.4 1.6 (1) 1.3–2.6 1.6–3.4

1.9±0.6 (11) 1.8±0.4 (65) 2.2±0.5 (17)

TN 0.6 (2) 0.4–0.9 0.6 (1) 0.4–0.9 0.6–1.3

0.7±0.2 (11) 0.6±0.1 (65) 0.8±0.2 (17)

TCM 4.1 (1) 2.7–5.4 3.6 (1) 2.3–5.3 3.1–6.7

3.9±1.8 (9) 3.6±0.8 (65) 4.4±1 (17)

TMM 0.6–0.7 (2) 0.3–0.9 0.6 (1) 0.4–1.1 0.6–1.8

0.7±0.2 (11) 0.7±0.1 (65) 0.9±0.2 (17)

NN 1.5 (1) 0.9–2 1.3 (1) 1–1.9 1.1–2.9

1.5±0.5 (11) 1.4±0.3 (65) 1.7±0.4 (17)

NMM 1.1–1.3 (2) 0.7–1.7 1.2 (1) 0.8–3.5 0.9–2.4

1.3±0.4 (11) 1.3±0.4 (65) 1.5±0.4 (17)

PA 182–193 (2) 191–206 208 (1) 177–197 180–205

199±74.3 (10) 187.3±23.1 (66) 192.7±7.4 (17)

SG 16–26 (2) 18–28 29 (1) 11–29 12–52

22.8±9.1 (10) 19.6±4.5 (66) 32.4±10.5 (17)

CSG 6 (1) 5–8 6 (1) 0–8 2–9

6.5±2.5 (10) 5.2±1.7 (58) 5.4±2.2 (17)

AD 13 (1) 10–13 10 (1) 10–14 9–16

11.7±5.5 (4) 12.3±4.4 (53) 11.7±1.5 (14)

AIV 2 (1) 2–4 2 (1) 1–3 1–3

2.8±1.1 (10) 2.3±0.5 (63) 2.1±0.6 (17)

PMT 15–17 (2) 14–18 16 (1) 9–19 12–18

15.4±4.3 (11) 14.3±3.3 (52) 15.3±1.4 (13)

PPT 13 (1) 14–18 16 (1) 10–19 12–14

15.8±6.9 (9) 14±4.3 (50) 13.3±1.1 (3)

DT 14 (1) 10–18 14 (1) 10–18 8–17

13.9±4.5 (11) 14.2±3.3 (52) 13.2±2.5 (9)

ST 4 (1) 0–5 4 (1) 1–6 3–4

3.1±1.8 (8) 3.2±1.1 (52) 3.8±0.4 (5) TL/BW 75.1–81 (2) 84.7–89.4 66.2 (1) 37.3–90.6 35.7–74.2

87±1.9 (3) 61.8±10.3 (50) 62.5±10 (13) SI 61–141 (2) 41–115 (2) The stepwise discriminant analysis of morphometric characters identified five variables (BW, BH, EN, ET, SP) that contribute most to differences between the OTUs ( Fig. 3 ; Table 2 ). This analysis revealed a significant difference among the geographic units in morphometric data (Wilks’s Lambda 0.600 p <0.001). Despite this significant difference, only a low separation was observed in the two canonical axes of the SDFA ( Fig. 3 ).

Axis 1 Axis 2

BW 0.192 0.757

The first discriminant function is responsible for 63.9% and the first and second together for 100% of the total variation. A total of 79% correct classifications were produced, 77% Jackknifed ( Table 3 ). The first axis of the SDFA is better represented by EN, body depth, ET and SP showing a weak separation of the MA (Maranhão) OTU from the TO (Tocantins) and PA (Belém) OTUs ( Fig. 3 ). The second axis is better represented by BW and EN, showing a partial separation of the TO OTU from the MA and PA OTUs.

MZUSP 57070 from Marabá, Pará, Brazil with TL 135 mm has approximately 60 “fetal” teeth in the lower jaw, arranged in numerous rows ( Fig. 4 ). Specimens of other populations, of about the same size, or even smaller have an adult dentition (e.g. MRT 6975 from Palmas, Tocantins with TL 139 mm; HUFMA 636 from Urbano Santos, Maranhão, with TL 112 mm), which may represent a geographic variation in the size when maturity is reached, or even individual variation because our sample of small specimens is small.

Distribution. Guyana, Suriname, French Guiana, Peru, Brazil (states of Amapá, Amazonas, Maranhão, Pará, Rondônia and Tocantins) ( Fig. 5 ).

Natural history. Taylor (1968) mentioned the hawk Leucopternis albicollis and the snake Anilius scytale as predators. Maschio et al. (2010) found a Caecilia cf. gracilis in the digestive tract of Anilius scytale from Brazilian Amazonia. Maciel et al. (in press) found earthworms in 92 % of the digestive tracts containing prey in a series of 59 dissected specimens of C. gracilis from Urbano Santos, Maranhão, Brazil.

Remarks. Nussbaum & Wilkinson (1989) recognized that the genus Caecilia might be non-monophyletic because some specimens of C. gracilis have eyes under bone like Oscaecilia . This character state is present in low frequency in different populations of C. gracilis , showing a fairly conservative pattern in this character across the wide distribution of this species.

Taylor (1972) stated that C. gracilis has a scale inception “near the end of the body, preceding the few secondaries a short distance, and also suggested a geographic variation in SI and “character of the squamulae between a specimen from Iquitos, Peru and specimens from French Guiana, although he did not clarify the nature of the geographic variation. Our data show that SI in C. gracilis is more variable than Taylor (1972) suggested.

Figures

FIGURE 3. Scatterplot graphs of scores of the first two axes of the Stepwise Discriminant Function Analysis employed for morphometric data of OTUs of Caecilia gracilis (left) and C. tentaculata (right).

FIGURE 5. Distribution of Caecilia gracilis. Open symbols represent data from literature and closed symbols specimens analyzed in this study. 1. Iquitos, Peru (Taylor, 1968); 2. Benjamin Constant, Amazonas, Brazil; 3. Urucú, Coari, Amazonas, Brazil; 4. Dunoon, Guyana (Taylor, 1968); 5. Wismar, Guyana (Taylor, 1968); 6. Orenoque Creek, Guyana (Nussbaum & Hoogmoed, 1979); 7. 30 km west of Paramaribo, Suriname (Nussbaum & Hoogmoed, 1979); 8. Onverwacht, Suriname (Nussbaum & Hoogmoed, 1979); 9. Sara Creek, Suriname (Nussbaum & Hoogmoed, 1979); 10 - 11. Localities in French Guiana based on the map on page 291 of Lescure & Marty (2000); 12. Igarapé Timbó, Rio Cupixi, Estrada para Serra do Navio, Amapá, Brazil; 13. Santarém, Pará, Brazil; 14. Melgaço, Pará, Brazil; 15. Belém, Pará, Brazil; 16. Santo Antônio do Tauá, Pará, Brazil; 17. Castanhal, Pará, Brazil; 18. Maracanã, Pará, Brazil; 19. Capanema, Pará, Brazil; 20. Bragança, Pará, Brazil; 21. Viseu, Pará, Brazil; 22. Tomé-Açú, Pará, Brazil; 23. Marabá, Pará, Brazil; 24. Palmas, Tocantins, Brazil; 25. Urbano Santos, Maranhão, Brazil; Jaci-Paraná, Rondônia, Brazil.

FIGURE 4. A juvenile of Caecilia gracilis (MZUSP 57070), showing “ fetal ” teeth in the lower jaw. Photo by Pedro Bernardo.

Tables

TABLE 1. Morphometric (in mm) and meristic data (range, mean ± standard deviation and sample size in parenthesis) of Caecilia gracilis. Data from populations sampled in this study.

	Amazonas	Tocantins	Amapá	Maranhão	Pará
TL	323–405 (2)	139–510 326.5±147.3 (11)	265 (1)	112–453 281.2±80.7 (66)	150–735 417±136.1 (17)
HW	4.3–4.8 (2)	2.8–5.6 4.2±0.8 (12)	3.7 (1)	2.8–5.3 3.9±0.8 (65)	3.5–7 5±1 (17)
HL	6.5–6.9 (2)	4.4–8.6 6.6±1.2 (12)	5.8 (1)	4.3–8.8 6.2±1.4 (65)	5.7–11.5 7.9±1.6 (17)
HH	3.2–3.4 (2)	1.9–4.8 3.1±1.1 (11)	2.3 (1)	2–4.5 3±1.5 (37)	2.5–5.9 3.7±0.9 (17)
BW	4.3–5 (2)	4.6–5.7 5.2±2.2 (3)	4 (1)	2.6–6.7 4.4±1.2 (51)	4.2–10.7 6.5±1.9 (13)
BH	3.9–4.4 (2)	5.1 (1)	3.1 (1)	2.3–5.9 3.6±0.9 (64)	3.6–9.2 5.5±1.6 (13)
WNC	4.3–4.8 (2)	2.8–6.2 4.5±2.1 (9)	3.8 (1)	2.7–5.9 4.3±1.6 (47)	3.8–10 5.9±1.8 (13)
WTR	3.9–5.3 (2)	2.7–5.5 4.4±1.8 (10)	3.9 (1)	2.5–6.3 4.2±1.1 (62)	4–9.8 6.1±1.7 (15)
WV	2.8–4 (2)	1.6–3.9 3.2±1.3 (10)	2.5 (1)	2.4–5 3.2±0.9 (31)	2.8–7.2 4±1 (16)
EE	2.9 (2)	1.9–3.5 2.8±0.9 (11)	2.2 (1)	1.9–5.2 2.5±0.6 (65)	2.3–5 3.2±0.7 (16)
EN	1.9 (1)	1.1–2.6 2±0.7 (11)	1.7 (1)	1.2–2.4 1.7±0.4 (65)	1.7–3 2.2±0.4 (16)
ET	1.8–1.9 (2)	1.2–2.2 1.8±0.6 (11)	1.6 (1)	1.2–2.4 1.7±0.4 (65)	1.6–3.2 2.2±0.4 (16)
EMM	0.8–0.9 (2)	0.5–1.1 0.8±0.2 (11)	0.6 (1)	0.4–1.1 0.7±0.2 (65)	0.5–1.4 0.9±0.2 (16)
SP	1.1–1.6 (2)	0.9–1.9 1.5±0.5 (11)	1.2 (1)	0.9–2.2 1.5±0.3 (65)	1.4–2.8 1.8±0.3 (17)

TABLE 2. Canonical Discriminant Functions: Standardized By Within Variances of the Stepwise Discriminant Analysis of Caecilia gracilis OTUs from east Brazilian Amazonia.

BH EN	0.521 1.012	0.329 - 0.225
ET	- 0.419	0.660
SP	- 0.374	0.370

TABLE 3. Classification Matrix from the Stepwise Discriminant Analysis of Caecilia gracilis OTUs from east Brazilian Amazonia.

MA	MA 63	PA 2	TO 0	% correct 97	Jackknifed % correct 95
PA	10	7	2	37	37
TO Total	4 80	2 8	5 5	45 79	36 77