Gymnotus ucamara, William G. R. Crampton, Nathan R. Lovejoy & James S. Albert, 2003

William G. R. Crampton, Nathan R. Lovejoy & James S. Albert, 2003, Gymnotus ucamara: a new species of Neotropical electric fish from the Peruvian Amazon (Ostariophysi: Gymnotidae), with notes on ecology and electric organ discharges., Zootaxa 277, pp. 1-18 : 5-13

Gymnotus ucamara n. sp.

(Fig. 3)

Holotype. - UF 126182 (156.0 mm TL, WGRC 09.200902), Peru, Loreto, Rio Ucayali, Rio Pacaya, Cocha Zapote, in Pacaya-Samiria National Reserve, 05°20.03'S, 74°29.08'W, collected by J. Albert, W. Crampton, N. Lovejoy, H. Ortega and R. Reis, 9 September 2002.

Paratypes. UF 126121 (1, 174 mm TL, WGRC 02.200902) ; UF 126183 (1, 172 mm TL, EOD not recorded) ; UF 126184 (2, 190 mm TL, WGRC 12.200902; 146 mm TL, WGRC 13.200902). All collected by J. Albert et al. with holotype .

Nontypes. MUSM 9274 (1, 134 mm TL), Loreto, Contamana, Aguas Calientes, approx. 07°02'S, 74°14'W, 3 June 1996 ; MUSM 10184 (7, 117-156 mm TL, 2 cleared and stained), Loreto, Contamana, Rio Ucayali, approx. 07°02'S, 74°14'W, 31 May 1996. All collected in Peru by H. Ortega et al.

Diagnosis. Gymnotus ucamara differs from other species in the Gymnotus carapo species-group in possessing the following unique combination of characters: a coloration pattern with 18-24 dark brown bands separated by narrow pale interbands which are less than one-third the width of the dark bands (a pattern that is readily distinguishable from all other described species of the G. carapo species-group except G. mamiraua ); a long head (12.2-13.4 % total length vs. 7.9-11.8 % in all other species except G. carapo and G. arapaima ); many (10-11) scales over anal fin pterygiophores (vs. 4-9 in all other species except G. arapaima ); few (38-43 [median 42]) pored lateral-line scales to first ramus (vs. 32-38 [median 37] in G. mamiraua , vs. 42-52 [median 48] in G. carapo , and vs. 50-64 in all other species); a low (75-91 [median 82]) total number of pored lateral-line scales (vs. 93-108 [median 98] in G. carapo , and vs. 106-140 in all other species except G. mamiraua with 75-79 [median 78]), and a relatively large eye (orbital diameter 0.09-0.10 % HL vs. 0.06-0.07 % in G. carapo ). G. ucamara is superficially most similar to G. mamiraua from which it can be readily distinguished on the basis of the following characters; a long head (12.2-13.4 % total length vs. 9.7-10.7 % total length in G. mamiraua ) and relatively more pored lateral line scales to first ramus (38-43 [median 42] vs. 32-38 [median 37] in G. mamiraua ).

Description. Body shape and pigment patterns illustrated in Fig. 3. Morphological and meristic data presented in Tables 2 and 3. Cephalic sensory canal pore configurations summarized in Fig. 4. Size up to 190 mm TL. Size at maturity unknown. Sexual dimorphism unknown. Scales cycloid, ovoid, present on entire post-cranial portion of body from nape to tip of caudal appendage. Scales on dorsal surface relatively large at mid-body, six rows from lateral-line to dorsal midline. Scales small over pterygiophores, 10-11 scale rows. Lateral-line scales (in holotype) approximately 1.3 mm high by 1.5 mm long in humeral region, 2.3 mm high by 2.6 mm long at midbody, 1.4 mm high by 2 mm long dorsal to anterior margin of clear patch on anal fin. Gape size in mature specimens large, extending beyond posterior nares. Mouth position superior, lower jaw longer than upper, rictus decurved. Chin fleshy and bulbous with thick pad of electroreceptor organs and support tissues overlying tip of snout and oral jaws. Anterior narial pore included within gape in large narial fold. Anterior nares large, subequal to diameter of eye. Branchial opening moderate (32-40 % in HL). Circumorbital series ovoid. Ethmoid region between anterior nares moderate, its anterior margin rounded. Eye position lateral, lower margin of eye slightly ventral to rictus. Eye relatively large, orbital diameter 0.09-1.0 % HL. Premaxilla and dentary with one or two rows of large, slightly recurved, conical teeth. Premaxilla with 11-12 (mode 12, n = 3) teeth disposed in single row along outer margin. Dentary with 16 (n = 2) teeth disposed in single row along outer margin.

Rib 5 approximately same width as rib 6. Body cavity of moderate length with 32-34 (mode 33) precaudal vertebrae. Hemal spines present. Gas bladder not extending beyond first hemal spine. Displaced hemal spines absent. Multiple anal-fin ray branching posterior to rays 10-17. Variable number (11-19) of asymmetrically arranged lateral-line rami extending posteroventrally at posterior end of lateral line. Dorsal lateral-line rami absent in all specimens examined. Anal-fin pterygiophores at posterior portion of body cavity shorter than first hemal spine. Caudal appendage short, less than half pectoral-fin length in undamaged and unregenerated specimens. Single hypaxial electric organ, extending along entire ventral margin of body. Three or four (mode 4) rows of electroplates at one HL distance from end of caudal appendage.

Color in life. 18-24 (median 21, n = 13) broad, dark, chocolate-colored bands, separated by narrow, pale interbands (sensu Albert et al. 1999) less than one-third the width of the dark bands. Slight countershading in specimens longer than 150 mm TL, more pronounced anteriorly. Pale interbands extend from pectoral-fin base to tip of caudal appendage and oriented obliquely (anterior-ventral to posterior-dorsal). Dark bands occur singly (never divided into band pairs sensu Albert et al. 1999) and are occasionally divided into Y or inverted-Y shapes but never divided into X shapes on anterior two-thirds of body. Pale interbands sometimes interrupted. Pigment density slightly greater at the dark band margins than in the middle at mid-body. Band-interband margins wavy and highly contrasted with one another. Pale interbands irregular in shape and width, narrower and more regularly shaped anteriorly. Pale interbands usually do not extend to mid-dorsum along anterior 2/3 of body. Incomplete pale interband present in middle of some dark bands. Three pale interbands from either side terminate near ventral midline, often not meeting, between the anus and anal-fin origin. One or two bands lie posterior to last anal-fin ray.

Head never banded, spotted or blotched, uniformly dark brown but slightly paler in gular region. Numerous minute chromatophores speckled over branchiostegal membranes and ventral surface of head. Pectoral-fin rays dark brown or black, interradial membranes hyaline. Anal fin never blotched, spotted or marked. Anal fin rays and membrane dark gray or black on anterior 80% of fin length, translucent on posterior 20%. Color variation is not known to be correlated with size, sex or EOD waveform. Specimens fixed in 10% formalin and preserved for 1-6 years in 70% ethanol maintain approximate colors of life, although the darker pigments sometimes pale with time.

Electric organ discharge. G. ucamara n. sp. generates EODs as discrete pulses of 1.25 -1.73 ms duration (mean 1.51, n = 5). These comprise four or five phases (Fig. 5). The waveform comprises a dominant tri-phasic component (P0, P1, P2) with a duration of approximately 0.78-0.92 ms (mean 0.87, n = 5) followed by a low-voltage positive final phase (P4). A very low-voltage initial positive phase (P-1) precedes P0 in most specimens. The Peak Power Frequency (PPF) (Fig. 4) of the Fourier Transform of EODs of G. ucamara n. sp. ranges from 1.71-1.95 kHz (mean 1.78, n = 5).

The EOD pulse repetition rate of G. ucamara n. sp. is relatively low and less variable during the day when this species lodges itself into the submerged root mats of floating plants (range 44.5-45.9 Hz, mean 45.3, standard deviation [SD] 0.3, n = 4). G. ucamara n. sp. can instantaneously increase its EOD pulse repetition rate to around 90-110 Hz in response to a sudden stimulus such as a loud underwater noise or a light prod with a glass rod. Following such a ‘fright response’ the EOD repetition rate returns to close to the normal resting rate within a few milliseconds and back to the normal resting rate within a few seconds. The EOD pulse repetition rate is usually higher and more variable at night. The highest pulse rates occur during swimming (range 62.5-76.9 Hz, mean 69.4, SD 6.2, n = 4) and the lowest rates occur when a specimen stops swimming, usually by resting its body against a submerged structure, or wedging itself between submerged roots (range 50.0- 66.7 Hz, mean 59.7, SD 8.7, n = 4).

Distribution. Known only from the lowland Rio Ucayali basin in Peru, at sites near Contamana and in the Rio Pacaya-Samiria National Reserve near the confluence of the Rio Ucayali and Rio Marañon.

Habitat and Ecology. All the type specimens were captured from static, vegetationchoked water in a shallow (maximum depth 0.5 m) channel connecting a floodplain lake (Cocha Zapote) to the Rio Pacaya (Figs. 1,2). The habitat structure, vegetation and water quality of the locality are typical of whitewater floodplains of the Central and Upper Amazon(Ayres 1993; Junk 1997). The type specimens were captured during the low-water month of September. Aquatic vegetation at the site consisted of plants typical of floating meadows, an important substrate for floodplain gymnotiform fish communities throughout the year (Crampton 1996; Albert & Crampton 2001). The dominant species were Cyperus sp., Eichhornia crassipes Solms. , Pistia stratiotes (L.) , Ludwigia sp., Salvinia spp., and Utricularia sp. Water quality at the locality was recorded as: dissolved oxygen, 2.4 mg/l; electrical conductivity, 240 µScm-1 at 25°C; transparency with Sechhi disc 0.6 m, surface temperature 31.2 °C. Other gymnotiform electric fish collected from the same locality and habitat were Gymnotus carapo (L.) , Eigenmannia sp., and three undescribed species of Brachyhypopomus . One of the two non-type lots from Contamana (MUSM 10184) was collected in rafts of floating vegetation along the edge of the Rio Ucayali. These rafts may have been swept out of nearby floodplain lakes (H. Ortega pers. comm.). The other nontype lot (MUSM 9274) was collected from floating vegetation along the edge of a sediment-laden ‘whitewater’ stream near its confluence with the Rio Ucayali. Water temperature was reported as 28°C. This stream flows out of hot springs from nearby mountains, but gymnotiforms were only found well downstream of the influence of unusually warm waters (H. Ortega pers. comm.). G. ucamara n. sp. feeds on aquatic invertebrates. Stomach content analyses of the paratype and non-type series are summarized in Table 4.

Etymology. Named for the geological term “Ucamara Depression” describing the lowlying region between the lower reaches of the Ucayali and Marañon Rivers caused by subsidence in the Upper Amazon foreland basin.