Keel-bellied mud snake, Bitia hydroides. Gray, 1842

Bitia hydroides Gray, 1842: 64.  Holotype: BMNH 1946.1.2.38.  Type locality: Unknown.
Homalopsis hydrina Cantor, 1847: 104.  Holotype: possibly in the ZSI collection, Calcutta, India. Type locality: sea off Kedah Coast, peninsular Malaysia.
Hipistes fasciatus Gray, 1849a: 78. Holotype: BMNH 1946.1.2.50. Type locality: from East Indies.
Hipistes hydrinus – Günther, 1864: 287.
Distribution: Coastal waters of southern Myanmar, Thailand, peninsular Malaysia, Singapore, and Borneo. Possibly in the Greater Sundas (Murphy 2007).
Diagnosis: Scale rows at mid body 37–43; ventral scales keeled; parietal scales fragmented; single internasal; dorsal scales small, triangular, exposed skin between them; loreal in contact with first 3 or 4 upper labials; ocular ring scales large, plate-like; upper labials 3–4 or 4–5 below the orbit; tail short 8–11% of SVL.

Bitia hydroides. Photo credit: Bruce Jayne

Diagnosis.  Bitia is one of the most distinctive homalopsid snakes. At midbody, it has 37 – 43 rows of triangular scales with exposed skin between them; it has keeled ventral scales, and an exceptionally long and narrow anal plate. No other snake has this combination of character states. Overall the gestalt and banded color pattern suggests a true sea snake (Figure 15), but Bitia lacks the expanded paddle tail of true sea snakes. The only other homalopsid snakes with this number of dorsal scale rows are members of the genus Homalopsis, but it has strongly keeled, striated dorsal scales that overlap and has no exposed skin between them. Homalopsis has a tail that is more than 30% of the SVL, while B. hydroides has a tail that is 8 – 13% of the SVL.

Size. The largest specimen measured for this study was a female with a total length of 718 mm, and a 54 mm tail. The largest male measured had a total length of 641 mm, and a 55 mm tail.  Female SVL averaged 500 mm. Male SVL averaged 338 mm. The smallest specimen measured had a total length of 153 mm, and a 19 mm tail. In a series of 33 small specimens, presumably neonates, the SVL ranged from 134 – 172 mm (X¯ = 153.9 mm). Jayne et al. (1995) reported females were more massive than males of equal length, and their analysis suggested that a 40 cm SVL male would weigh 26 g while a 40 cm female would weigh 31.3 g.  A 15 cm neonate was predicted to weigh 2 grams. 

External Morphology. The body is cylindrical to slightly compressed; the tail is slightly flattened. Despite the fact that Gyi (1970) and Smith (1943) have stated that the head is not wider than the neck, Jayne et al. (1995) found this true only for smaller specimens and demonstrated that larger specimens do have the head distinctly enlarged from the neck. The eyes are round and set dorsally on top of the head. They are also relatively small; having an orbital diameter that is about one third of the length of the supraocular scale. The lower jaw is countersunk. On the head the rostral scale is about as broad as it is tall; it is almost round (most homalopsids have a pentagonal rostral); and it has two notches on the ventral edge. The mental scale opposite the rostral scale’s ventral edge has a small pad of tissue with folds that fit the rostral’s groove. This is probably an adaptation for complete mouth closure and may allow the tongue tips to be protruded slightly without opening the mouth. The nasal scales are separated by an internasal. Each nasal is divided unequally by the nare with the posterior portion being much smaller than the anterior; the nasal cleft touches the loreal and the internasal. The single internasal is about three times longer than it is broad and penetrates between the prefrontal scales but does not completely separate them. The prefrontal scales are in contact. The frontal is elongated; in some specimens, it is semi-divided, and in outline may take on a distinctive hour-glass shape. The parietal scales are fragmented into several plates or smaller scales, see Figure 15. The loreal is single and is in contact with the first three or four labials.  The scales of the ocular ring are exceptionally large and plate like, there is one supraocular, one preocular, one postocular, and one subocular. One specimen had a divided postocular; another had a divided subocular. The temporal scale formula was 1 + 2 + 3. However, because the parietals are fragmented distinguishing between the one primary temporal and an enlarged fragment of the parietal is difficult. The upper labials usually number seven, occasionally eight. Labials 3 – 4 or 4 – 5 can be below the orbit, but not entering the orbit because of the presence of the subocular. On the chin the lower labials, number 8 – 12. The first pair of chin shields contacts the first 4 – 6 labials. There are two pairs of chin shields; the first pair is larger than the second. The gulars number 5 – 7, usually six. Teeth. Jayne et al. (1995) found enlarged anterior palatine teeth to be a unique, derived character state in this genus. The fourth palatine tooth was 3.1 mm and exceeded the length of the posterior maxillary fang (2.1 mm). Also, the fourth and longest tooth in the lower jaw exceeds the length of the posterior maxillary fang [maxillary teeth = 13 – 14; palatine teeth = 7 – 8; pterygoid teeth = 19 – 23; dentary teeth = 18 – 23].
     On the body the dorsal scales on the neck are in 32 – 39 rows, they are smooth, lanceolate, and have gaps between the base of the scale and the apex of the preceding scale; dorsal scales at midbody are in 37 – 43 rows, and at posterior body are in 31 – 35 rows. The first dorsal rows are approximately the same size, but scales become enlarged toward the midline. The ventral scales are about twice the length of a dorsal scale, and they are keeled at the outer margins (Plate 1, Figure E). However, these keels most likely play a role in allowing the snake to laterally compress its body when swimming. Ventral scales number 151 – 161 (X¯ = 158) in males, and 157 – 164 (X¯ = 161) in females. On the tail the dorsal scales are more ovate than those on the body, and are also smooth. The subcaudal scales are divided, and number 33 – 40 in males, and 24 – 29 in females and thus the counts are sexually dimorphic. The anal plate is divided, narrow, and at least twice as long as a normal ventral. Males had tails that averaged 11.1% of the total length, while female tails averaged 8.0% of total length (Jayne et al., 1995). At the base of the tail, the width is 81% of the height.
     Color and pattern on the head are uniform gray. The body has 37 – 46 dark transverse blotches that are 2 – 3 scales wide. The ventral surface is  cream color. The tail’s dorsum has a pattern similar to the body; the tail’s ventral surface is uniform cream. 
    The morphological convergence of Bitia with other marine snakes has been commented upon by several other authors. Boettger (1898) considered Bitia involved in a mimicry complex with Acrochordus granulatus. Cantor (1847) wrote, “In general appearance and colours the present is more closely allied to the pelagic serpents than any other known species.” And, Gyi (1970) noted, “The small head; narrow anterior third of the body; narrow ventrals and subcaudals; and compressed tail resemble the characteristics of members of the family Hydrophidae.” 

    Bitia is usually collected in the intertidal environment including mud flats, estuaries, and possibly mangroves. However, it is seldom encountered, and the microhabitat used for most of its life history is unknown. Jayne et al. (1995) collected this snake in the intertidal zone of the Muar estuary and found it to be the second most common snake; only one specimen was found in the channel of the Muar River. Released specimens were observed burrowing into the mud, and they note that the small eyes, smooth scales, and short tail are adaptations seen in terrestrial burrowing snakes. This species was also observed to use a sidewinding locomotion on mud surfaces, as do some other homalopsids that forage on mudflats. More specimens were found active at night than during the day. It should be pointed out that this species showed up at this location for a few nights and that it was not common before or after the few days of its abundance.  Theobald (1868) wrote, “This species has quite the coloration of a Hydrophis through probably an estuary rather than an open sea species.  It is, however, largely captured in company with great numbers of Hydrophis in the sluice nets (or creeks) in the Bassein River below Gnaputau.” This may be an aquatic-burrowing homalopsids, and it is possible that it can spend long periods of time submerged, obtaining oxygen through cutaneous respiration via the exposed skin between the scales. However, mangrove mud is often anoxic, and how this environment influences (if it does) snake behavior is unknown.
Diet and Feeding Behavior
    Boulenger (1890) wrote, “It lives almost entirely upon fishes.” And, Cantor (1847) commented, “In one examined the stomach contained remains of two small pelagic fishes.” The most detailed account of food habits was provided by Jayne et al. (1995); they examined 117 field captured individuals and found one prey item in each of six individuals. Three items could only be identified as the tail fin rays of fish. Two individuals contained the goby Oxuderces dentatus, and for one of these snakes (SVL = 16.5 cm, mass = 3.0 g), the prey had a total length of 6.6 cm and a mass of 0.4 g (13.3% of the predator’s mass). The other prey item found was the goby Trypauchen raha, which was the largest prey item consumed in terms of both absolute size (total length = 14 cm; mass = 9.1 g) and mass, it was 38% of the predator’s mass (snake SVL = 39 cm, mass = 26 g). Captive snakes were offered a variety of sympatric species of oxydercine gobies (Periopthalmus chrysospilos, Scartelaous pectinirostris, and Oxuderces dentatus), but they were unsuccessful in eliciting feeding behavior. The presence of Trypauchen raha in the diet suggests that Bitia may feed in deeper water than is usually available in the intertidal, mudflat habitat.
    There are no observations of this snake feeding, and Jayne et al. (1995) noted that the enlarged palatine teeth do not seem suited for capturing the prey they found in the specimens examined, and they wrote, “Whether or not Bitia catches and consumes prey common to other marine snakes by using specialized behaviors remains an open question.”

    Jayne et al. (1995) found the smallest gravid female had an SVL of 36.5 cm and mass of 26 g (excluding the embryos). Of 28 females captured with SVLs greater than 36.5 cm, 14 were gravid. Gravid females were captured between 20 November and 18 December, and this correlates with Smith’s (1943) report of two females obtained in Myanmar (northern edge of the range) in September containing three and four fully formed embryos. Litter size varied from 1 – 10 (X¯ = 4.2) and increased significantly with female SVL. Clutch mass averaged 14.5 g (r = 2.3-52 g) and increased significantly with SVL. The mean RCM was 0.22 (range = 0.08 – 0.35), the RCM was not significantly correlated with maternal SVL or mass. The mean RCM is low for a viviparous colubrid, and it is exceeded by nine of 11 North American species of natricines. The mean mass of embryos per litter ranged from 2.3 – 5.2 g (X¯ = 3.2 g), and this quantity was not significantly correlated with maternal size, litter size, or RCM. Three litters had an average SVL of 16.1 cm (r = 15.5 – 16.8) and the average mass was 2.5 g. Heatwole (1999) stated that Bitia matures early. And, at least some female Bitia reproduce in the first year of life despite the fact that they are very small at birth. Jayne et al. (1995) found parturition  to be highly seasonal and does not overlap with any of the other marine snakes studied in the Muar area. The smallest gravid female Jayne et al. (1995) report had an SVL of 365 mm, this investigation of additional museum material produced a 293 mm female (UMMZ 61291) containing three embryos and one infertile egg. Jayne et al. (1995) estimated the growth rate for neonates at 0.06 cm/day.