Harttia villasboas, Oyakawa & Fichberg & Py-Daniel, 2018

Oyakawa, Osvaldo T., Fichberg, Ilana & Py-Daniel, Lucia Rapp, 2018, Three new species of Harttia (Loricariidae: Loricariinae) from Serra do Cachimbo, Rio Xingu basin, Pará, Northern Brazil, Zootaxa 4387 (1), pp. 75-90: 81-88

publication ID


publication LSID


persistent identifier


treatment provided by


scientific name

Harttia villasboas

new species

Harttia villasboas   , new species

( Figs. 3 View FIGURE 3 , 4 View FIGURE 4 , 5 View FIGURE 5 , 6 View FIGURE 6 ; Tables 1, 2)

Harttia duriventris Buckup et al., 2011   , In: Castilhos & Buckup, 2011.

Harttia   sp. Xingu3 — Covain et al., 2016 (reference, distribution, molecular phylogeny).

Holotype. MZUSP 97078 View Materials , 126.0 mm SL, Brazil, Pará, Rio Xingu Basin, Rio Curuá, tributary of Rio Iriri , above two waterfalls, near the Churrascaria Cachoeira do Curuá at highway Cuiabá-Santarém (BR-163), mun. Novo Progresso , 08°44'09"S 54°57'46"W, 20 October 2007, J.L. Birindelli, L.M. Sousa, A.L. Netto-Ferreira, M.H. Sabaj & N.K. Lujan. GoogleMaps  

Paratypes. Brazil, Pará, Rio Xingu Basin : ANSP 200508, 5, 80.3–104.8 mm SL   , INPA-ICT 0 53237, 20, 46.4– 115.9 mm SL, MCP 50553, 5, 69.4–98.3 mm SL, MZUSP 115485, 64, (9 c&s), 36.5–126.0 mm SL, MZUSP 96937, 4, 40.2–49.3 SL, INPA-ICT 0 53238, 4, 55.0–80.0 mm SL, collected with holotype GoogleMaps   . MZUSP 101304, 3 View Materials , 81.1 View Materials –114.0 mm SL, Rio Curuá, tributary of Rio Iriri , just above the dam, mun. Novo Progresso, 0 8º44'30 S 54°57'35" W GoogleMaps   , 23 January 2009, A.L. Netto-Ferreira, J.L. Birindelli, L.M. Sousa & P.H. Hollanda-Carvalho. MZUSP 106473 View Materials , 1 View Materials , 44.9 View Materials mm SL, Rio Curuá, tributary of Rio Iriri, Serra do Cachimbo , near the highway Santarém-Cuiabá (BR-163), mun. Novo Progresso , 8˚44’9”S 54˚57’46”W, 15 August 1984, M. Goulding   . MZUSP 118550, 17, 31.2–109.9 mm SL, Rio Curuá, tributary of Rio Iriri at the Cachoeira do Curuá waterfall, near the village of PCH Curuá and Churrascaria Cachoeira do Curuá , at highway Cuiabá-Santarém (BR-163), between the 60 m and 40 m falls, mun. Novo Progresso, 8˚44’2.3”S 54˚57’42.1”W, 0 6 August 2015, O.T. Oyakawa, W.M. Ohara & M. Pastana   . MZUSP 120907, 4 View Materials , 64.0– 85.4 mm SL, Rio Curuá, tributary of Rio Iriri at the Cachoeira do Curuá waterfall, near the village of PCH Curuá and Churrascaria Cachoeira do Curuá, between the 40 m fall and the three small falls of approximately 1 to 5 m, mun. Novo Progresso, 8˚44’9.5”S 54˚57’46.5”W, 0 6 August 2015, O.T. Oyakawa, W.M. Ohara & M. Pastana   . MNRJ 35553 View Materials , 36 View Materials , 20.0– 72.1 mm SL, Rio Curuá waterfall, Rio Iriri drainage, between two falls, ca. 54 km from Castelo dos Sonhos, mun. Novo Progresso , 8˚44’6”S 54˚57’43”W, 1 October 2008, M.R. Britto, J. Gomes, F. Carvalho & L. Fries.  

Diagnosis. Presence of abdomen completely covered by plates readily discriminates H. villasboas   from members of the H. loricariformis   group (naked abdomen) and from members of H. rhombocephala   group (abdomen partially covered). Harttia villasboas   can be distinguished from H. surinamensis   , H. fowleri   , and H. duriventris   by having caudal peduncle slightly compressed laterally after confluence of lateral keels vs. caudal peduncle strongly compressed laterally after confluence of lateral keels. In addition, H. villasboas   can be distinguished from all other species of the H. fowleri   group by having greater head length 24.0–29.9% [26.3%] of SL vs. 22.4–24.8% [23.5%] of SL in H. surinamensis   , 22.7–24.5% [23.6%] of SL in H. fowleri   , 21.0–24.3% [22.3%] of SL in H. duriventris   , 21.1–23.9% [22.7%] of SL in H. dissidens   , 22.8–24.8% [23.4%] of SL in H. panara   and 21.0–24.7% [22.8%] of SL in H. absaberi   . H. villasboas   differs from all other species of H. fowleri   group by presenting, on snout, a conspicuous median longitudinal ridge, bordered by two slightly lateral ridge from the supraorbital rim to the snout tip, vs. absence.

Description. Measurements and counts in the Table 2. Member of H. fowleri   group. Body dorsoventrally depressed and elongated, widest at cleithrum. Dorsal profile of body straight and gently ascending from tip of snout to anterior region of orbit, and horizontally straight from this point to dorsal-fin origin and gently descending to end of caudal peduncle. Ventral profile of body straight from tip of snout to origin of anal fin and gently ascending from this point to caudal peduncle.

Anterior profile of the head elliptical in dorsal view. Eye roughly round, inferior margin of orbit slightly straight. Dorsal flap of iris present. Interorbital and parieto-supraoccipital flat. Tip of snout with small, oval and naked area circumscribed by rostral plates; naked area sometimes extending backward and joined with naked region of upper lip. Snout with conspicuous median longitudinal ridge, bordered by two slightly lateral ridges from the supraorbital rim to the snout tip. Lips with numerous marginal short fringes. Maxillary barbel short, reaching half of length of lower lips and ornamented with small papillae; barbel joined to lip by small flap of tissue. Upper lip covered by numerous series of papillae; papillae of outer series larger than inner series. Lower lip well developed and covered by numerous series of papillae; papillae of outer series slightly smaller than inner series.

Posterior border of lower lip not reaching anterior margin of pectoral girdle. Premaxilla with 32–53 [45] and dentary with 33–48 [41] bicuspid teeth; all cusps with the same length. Presence of a conspicuous spherical papilla on roof of mouth anterior to oral valve. Infraorbital series with five plates; infraorbital 5 contacting inferior branch of sphenotic. Inferior region of orbit delimited by infraorbitals 3 to 5. Canal plate exposed, roughly triangular.

Abdominal region completely covered by minute plates. Gular area naked. Nine to 13 [10] lateral abdominal plates, rectangular and elongate. Preanal plates 2, roughly hexagonal and well developed, bordered anteriorly by a line of irregular plates. Five longitudinal series of plates on trunk. Median series with 29–32 [30] perforated plates. Two weakly developed, parallel and longitudinal odontodes keels coalesced at 20th –21 st plates   .

Dorsal fin II,7; its origin on vertical through pelvic-fin origin. Spinelet, or first spine, roughly oval and small, its width approximately half of width of base of second dorsal-fin spine. Chain-link structure, which articulates dorsal-fin spine with second dorsal-fin pterygiophore, incomplete. Tip of last rays of dorsal fin, when adpressed, reaching vertical through of origin of last anal-fin ray. Pectoral fin I,6; tip of pectoral-fin spine and first two branched rays surpassing insertion of pelvic-fin spine. Mature males with dorsal region of pectoral-fin spine covered by well-developed odontodes strongly bent and turned backwards. Pelvic fin i,5; tip of pelvic-fin spine reaching insertion of anal-fin spine. Anal fin i,5; external portion of first anal-fin pterygiophore roughly rectangular shape and covered by skin. Caudal fin forked, i,12,i rays, with 5 supracaudal plates on its base, median supracaudal plate bearing lateral line canal. Two procurrent rays on base of upper and lower caudal-fin rays. Caudal fin slightly emarginate. Caudal peduncle slightly compressed laterally after confluence of lateral keels.

Color in alcohol. Dorsal region of body dark to light brown. Six, inconspicuous, transverse dark brown marks, first near the origin of dorsal fin, second starting at end of last rays of dorsal fin, followed by third to fifth in middle of caudal peduncle, and sixth at origin of caudal-fin rays. Ventral region yellow pale. All fins with four transverse dark brown marks. Anal-fin base surrounded by a small blotch of dark brown chromatophores.

Etymology. The specific name, villasboas   , is a patronym that honors the Orlando, Claudio and Leonardo Villas Boas brothers, three frontiersmen that led the Expedição Roncador-Xingu during the years of 1943 to 1949, with the mission to explore a wide and unknown territory of the Amazonian regions of Brazil. In 1949, the expedition reached the tablelands of the Serra do Cachimbo. One of the most important results of this expedition was the establishment of the Parque Indígena do Xingu in 1961, the first huge indigenous area in all South America. The area of the park is 26.420 km 2 in the northern of Mato Grosso State and nowadays protects 14 indigenous ethnicities. A noun in apposition.

Distribution. Harttia villasboas   is, so far, known only from above the two great falls of Rio Curuá, a tributary of Rio Iriri, in the area of Serra do Cachimbo. Collections made bellow the two falls of Rio Curuá have failed in capture the species, suggesting that the new species might be restricted to the portion of the river above the two great falls ( Figs. 4 View FIGURE 4 , 5 View FIGURE 5 ).

Discussion. Although the presence of plates on the abdomen of species of Harttia   was discussed in Oyakawa et al. (2013), some details deserve additional comments. As already noted by Rapp Py-Daniel & Oliveira (2001) and Provenzano et al. (2005), three conditions of abdominal plating, or abdominal covering, can be found in the species of the genus. Abdomen without plates, abdomen partly covered or completely covered by plates. The abdomen is the area delimited by the pectoral girdle, lateral abdominal plates and the preanal plates, or by the urogenital pore in case that the preanal plates are absent. The abdominal region of Harttia rondoni   is partially covered by minute irregular plates usually concentrated on posterior portion of the abdomen and contacting the preanal plates, and the larger specimens usually bear a narrow median longitudinal row of plates. Besides of H. rondoni   , H. longipinna   , H. rhombocephala   , H. trombetensis   and H. punctata   have the abdomen partially covered by plates. Harttia rhombocephala   was described by Miranda-Ribeiro (1939) based on only one specimen with 120.5 mm of SL, collected in the Rio de Janeiro State, and no additional specimens were collected since. Even though Miranda-Ribeiro (1939), Oyakawa (1993) and Provenzano et al. (2005) had stated that the abdomen of H. rhombocephala   is covered by plates, the careful examination of the holotype in this work clearly shows two naked areas on the abdomen, delimited by the median row and the lateral abdominal plates. The implications of this alternate pattern of abdominal covering deserve future examination, as the variation seems to be unrelated to environment or locality.

Covain et al. (2016) produced a molecular phylogeny of the subfamily Loricariinae   and concluded that the only supported group within Harttia   is the clade including Amazonian representatives with the following interrelationships ((((( Harttia   s p. Xingu 2, Harttia   sp. Xingu3) Harttia   sp. Xingu1) H. duriventris   ) ( H. dissidens   , Harttia   sp. Tapajós)) (( H. punctata   , Harttia   sp. Tocantins) H. aff. punctata   )). All Harttia   species from Xingu correspond to the taxa described herein ( Harttia   sp. Xingu1 = Harttia rondoni   , Harttia   s p. Xingu 2 = H. panara   , Harttia   sp. Xingu3 = H. villasboas   ).

The Serra do Cachimbo, located at the southwest of Pará State and incrusted in the north border of the Brazilian Shield, is an ancient eroded plateau oriented in the SE/NW direction and encompassing highlands that may reach 740 m of altitude. These highlands are the divide between the Xingu and Tapajós basins and congregate part of the headwaters of the Rio Xingu (Curuá and its tributaries, the Rio Treze de Maio, Rio Escorpião and others), which runs from south to north direction, and Rio Tapajós (São Benedito, Teles Pires and its tributaries, the Rio Braço Norte, Rio Cristalino and others as well), which runs from east to west direction.

Harttia The   Rio Curuá departs Serra do Cachimbo through two major waterfalls approximately 40 m (upstream) and 60 m (downstream) ( Fig. 5 View FIGURE 5 ) in height, separated from each other by about 50 m of river channel. Above the 40 m waterfall, there is a sequence of three smaller waterfalls of approximately 10 m each ( Varella & Sabaj Pérez 2014). In the same way, the Rio Treze de Maio leaves the Serra do Cachimbo through a large waterfall with the same name, and then runs for about 20 km, discharging into the Rio Curuá near the village of Cachoeira da Serra downstream to the two major waterfalls of Rio Curuá. These waterfalls clearly isolate the headwaters of the Rio Curuá and its tributary, the Rio Treze de Maio, from the middle and lower part of the Rio Curuá. The most impressive feature of the hydrography of the region is the close proximity between the headwaters of the Rio Curuá (Xingu basin) and the Rio Teles Pires (Tapajós basin). This proximity is so small that, in certain points, these headwaters are separated by less than one kilometer of relatively flat terrain.

Harttia rondoni   is known from the Rio Curuá and Treze de Maio only bellow the waterfalls. So far no records of this species are known upstream the waterfalls of these two rivers. On the other hand, there are two records of the species in the Rio Fresco, a tributary of the middle Rio Xingu ( Fig. 4 View FIGURE 4 ). Judging for this pattern of distribution, it is quite reasonable to assume that H. rondoni   is more widely distributed in Xingu basin than suggested by its current distribution, which only may reflect an incomplete sampling.

Harttia panara   is known from only one locality of Rio Curuá upstream of the 40 m waterfall; and H. villasboas   is known from two localities, syntopically with H. panara   , upstream of the 40 m waterfall and in a small stretch of the 50 m of river channel of the Rio Curuá between the 60 m and 40 m waterfalls. Contrary to Harttia rondoni   , H. panara   and H. villasboas   apparently have a very restricted distribution, at least based on their current distribution, because no other records of these two species are known in the upper Rio Xingu above the waterfalls.

In the last ten years several species, said to be endemic to the Serra do Cachimbo, were described. Leporinus guttatus Birindelli & Britski, 2009   ; Moenkhausia petymbuaba Lima & Birindelli, 2006   ; M. clorophthalma Sousa et al., 2010   ; M. plumbea Sousa et al., 2010   ; Jupiaba kurua Birindelli et al., 2009   ; Lebiasina marilynae   Netto- Ferreira, 2012; L. melonoguttata Netto-Ferreira, 2012   ; L. minuta Netto-Ferreira, 2012   ; Apistogramma kullanderi Varella & Sabaj Pérez, 2014   ; A. eleutheria Varella & Britzke, 2016   ; A. flavipedunculata Varella & Britzke, 2016   ; Erythrocharax altipinnis Netto-Ferreira et al., 2013   and Ituglanis apteryx Datovo, 2014   were described from upstream of waterfalls of Rio Curuá or Rio Treze de Maio, with the exception of M. plumbea   , which was described from a tributary of Rio Braço Norte, Rio Tapajós drainage. All the aforementioned species were described based on material collected at only one locality or within a very constrained area. More recently, based on collection efforts made in 2015 by a team of MZUSP supported by the South American Characiformes Inventory   project (Projeto SACI), Lebiasina melonoguttata   was recorded from the headwaters of Rio Teles Pires, Tapajós basin.

The presence of Harttia panara   and H. villasboas in Xingu   headwaters (above 300 m) and their absence from lower regions is puzzling and properly also raises some question about the apparent scarcity of these fish from the Xingu, as Harttia   representatives tend to be abundant in their rock-covered environments. Rio Xingu fishes have been in the center of the international tropical aquarium trade for decades, and although Harttia   has never been a high-rated ornamental fish in this specific market, anecdotal sampling could have occurred and noticed. No records of Harttia   have been registered from the Xingu whatsoever ( Camargo et al. 2014). This might be explained by a non-Loricariinae oriented professional diving collection in the Xingu or a real scarcity of Harttia   . Recently, owing to the necessity of evaluating the impact of damming the Rio Xingu (UHE Belo Monte), several ichthyological surveys were conducted in the area between São Felix do Xingu down to the mouth of Rio Xingu in the Amazon, and from middle portion of Rio Iriri down to its mouth in Xingu (2012–2016 —D. Bastos, A. Gonçalves, pers. comm.) without any record of Harttia   representatives. This way, we consider that more studies and samplings efforts are necessary to comprehend the Harttia   diversity of the Xingu basin.


Museu de Zoologia da Universidade de Sao Paulo


Prestwich and Pilkington Botanical Society














Harttia villasboas

Oyakawa, Osvaldo T., Fichberg, Ilana & Py-Daniel, Lucia Rapp 2018

Harttia duriventris

Buckup et al. 2011