J. H. HUBER*
* Laboratoire d'Ichtyologie, Muséum national d'Histoire naturelle, 43 rue Cuvier, 75231 Paris Cedex 05, France.
Abstract:
The author studies the Cyprinodonts of the genus Rivulus that have been obtained as the result of four collecting trips in French and Dutch Guyana. Six species are disclosed, from which two are new, Rivulus xiphidius and Rivulus amphoreus, two are positively identified, Rivulus geayi and its yellow counter part Rivulus agilae, two bear some uncertainty, Rivulus cf.urophthalmus and Rivulus aff. holmiae where comparison with living material from the type locality is strongly needed. Another species, Rivulus manaensis could be a simple synonym of agilae. Rivulus xiphidius is thought to be very isolated in the genus. Finally primary information about biology and aquarium maintenance are given.
The genus Rivulus Poey, 1861 contains about sixty taxa, of which more than 80% were described before 1945 and 50% before 1916. The types are very often in bad condition, the descriptions inadequate or the type location is not precise.
Regan (1912) lists 22 species; Myers (1927) creates the genera Rivulichthys and Rachovia and Hoedeman (1959) proposes a method of determining species based on frontal scalation and morphology.
In fact my experience tends to show that the first criterion, although probably of use in defining superspecies, proves to be particularly variable and cannot confidently be considered as a means of determining a species of fish. As to the second, it does not allow us to separate cryptic species which are isomorphs. It seems that our present knowledge of the systematics can be compared to that of the genus Aphyosemion 25 years ago, with the additional difficulty of the variability of the colour pattern.
In that sense, this paper cannot be regarded as definitive, but rather as a preliminary approach. My warm thanks are due to the collectors and breeders who kindly provided me with their material for study: P. Blanc, Surennes; V. Montiel, Vierzon; G. Oelker, Kourou; D. Poliak, Chateau-Landon and F. Vermeulen, Lent.
1. Rivulus xiphidius n.sp.
Holotype: (MNHN 1979-284), female, 22.3mm standard length, 26.9mm; total length, from location "A" up stream from Saint Georges, set back from the Maripa Falls a few hundred metres from Oyapock river, French Guyana (fig.1B). P. Blanc, coll. 22.11.1978.
Paratypes: (MNHN 1979-285),1 male and four females from location "B", situated 14kms. from Kourou near the inland road, a few hundred metres away from the road, "on a hunters path towards the hill". G. Oelker, coll. end of 1977.
Colouration of Live Fish:
Characterised by a black band which covers the lower third of the body in both sexes. In the male this band extends into the caudal fin; below and above it is edged by a bluish-white border, especially in the rear part of the body. The rest of the body is bright orange and shows bluish and purple in reflected light. The anal and ventral fins also have a black band near the body, which is made more prominent by a bluish border; the outer part of the fins is orange. In the males from location "B" the black band on the body merges with that on the anal fin. The top half of the caudal is orange and the lower half contains the extended black band on the body with its bluish border. Below this there is another bluish band, sometimes white, and finally there is an orange marginal area. This caudal pattern resembles a sword, hence the specific name xiphidius, from the Greek "xiphos" a sword.
The female has a brown body with numerous irregular spots; the black band turns dark grey when the fish is frightened and a few dark spots may appear at the top of the caudal or dorsal fins.
Colouration of Preserved Fish:
All that remains after the fish have been preserved are the black band and the melanistic pigments scattered irregularly over the rest of the body and the fins of the male. It should also
be added that in the male there is a whitish lower band on the caudal and anal fins.
Morphology:
One of the smallest species of the genus. Meristic details and proportions of the type fish see Table 1.
The sensory system is composed of a complex group of frontal pores, fig. 2, a small number of pores along the lateral line, two barely visible facial pores, lying about a third of the
way between the eye and the operculum. The frontal scalation is of the E type (fig. 2)
Fig.2. Frontal scalation (E-type) and frontal neuromasts of Rivulus xiphidius n.sp.
The slender cylindrical shape of this fish and its unique colour pattern set it apart from the rest of the genus and in the future it may be given separate status.
2. Rivulus amphoreus n.sp.
Holotype: (MNHN Paris 1979-286), male 57.9mm (SL), 64.7mm (TL), from location "H" situate near the Tafelberg, an area of high hills reaching more than 1000 metres, some 120kms. to the south-south west of Paramaribo. R.P. Solanus-Essers coll.
Paratypes: (MNHN 1979-287) a male and a female from the same place.
Supplementary material (ZMA116136 and 116137) 5 specimens from the same location, deposited with the Instituut voor Taxonomische Zoologie, Zoologisch Museum Universiteit van Amsterdam.
Frans Vermeulen disclosed these fish at a Nijmegen dealer's where their drab colouration caused them to be unnoticed.
Colour Pattern of Live Fish:
The body of both sexes are the same; a rather irregular chequered pattern of brownish green spots on a bright yellow background; the fins are spotted with grey. The female is more drab, with a black ocellus on the top half of the caudal peduncle. The male's caudal has a black margin with a yellow edge and towards the centre, concentric circles of dark spots.
Colour Pattern of Preserved Fish:
No changes from live pattern.
Morphology:
Typical of the genus, the fish is relatively large, rather massive and with rounded fins. The unpaired fins have an average number of rays: D = 11-12, A = 16-17 and there are many small scales: L.L. = 47-49. For meristic characteristics and proportions see Table 1.
Nevertheless the fish has one characteristic which has never been reported in the other species. Rivulus amphoreus has a pair of facial neuromasts ,situated one third of the way between the eye and the operculum and clearly exposed inside an open vertical canal. Although these neuromasts have not been mentioned in literature, they exist in some preserved specimens at the MNHN and also in Rivulus xiphidius. With Rivulus amphoreus these organs are not in the form of pores, but of clearly visible neuromasts. It is not known yet if this characteristic has a specific value. Besides the facial neuromasts, there is a complex frontal sensory structure and an incomplete lateral line with a sensory pore every two or three scales. The frontal scalation is indistinct and probably of the E type.
The specific name "amphoreus" comes from "amphora", an allusion to the general morphology and fine features (facial neuromasts) as well as to the colouration of the fish.
Discussion:
Rivulus amphoreus can be separated from the small species of the Guyanas (geayi group) by its large size. It differs from the large species by its colour pattern, by its large dorsal fin (D= 11-12, instead of D = 8-10 as in waimacui and others) and by its geographic origin. It is in fact the only species from the Guyanas living at a high altitude. Hoedeman (1959: 54) reported the presence of a Rivulus in the Tafelberg, but rather curiously he suggests in this case, as in others, the existence of natural hybrids.
3- Rivulus geayi Vaillant 1899
Although the original description is old, it was not difficult to identify the specimens from locations "B" and "A" (near the type location). In fact the original collecting place is described in detail: "finally, beyond the source of the River Carnot, M. Geay passed the watershed which separates the basin of the Carsevenne river from that of the Cachipour, where this latter river. begins".
The same is true of the colour pattern: "dark head; there are four or five vertical black markings on the caudal peduncle, the ones nearest the caudal being more or less chevron shaped. There is vertical band on the caudal fin". The fish under discussion are the first caught since the original description and we can now see the colouration of live fish.
D = 8-9, A = 11-12, L.L.30-32.
The number of haploid chromosomes is 21: 12 metacentric or submetacentric and 9 telocentric. Scheel (1972) published 10 karyotypes of Rivulus varying from n = 20 to n = 23.
4- Rivulus agilae Hoedemann 1954.
A small Rivulus - 40 to 50mm TL - was caught on the outskirts of Cayenne, below Mont Mahuri (loc."E"). The same species was caught 19km from Kourou towards the interior (loc."D"), together with Rivulus cf. urophthalmus.
We have identified these specimens as Rivulus agilae even though the type locality (Agila) is in Dutch Guyana and despite the fact that between the two locations Hoedeman (1961:61) described an isomorphic species, manaensis (river Mana in French Guyana). Rivulus manaensis is distinguished from agilae mainly by colour pattern of the male: rear rays of the anal and lower part of the caudal black (one fifth). To me, these differences seem minor and, in addition, comparison of the colour pattern of the form from location "D" with that of the form from Zanderij, "45kms. from the south of Paramaribo, near Agila" Hoedeman (1961) is convincing: manaensis could only be a colour morph of agilae.
D = 7-9, A = 11-12, L.L.= 29-33.
Rivulus agilaeArial; is also very close to Rivulus geayi: they are isomorphs having a similar body colour pattern on the body comprising irregular blue 'V' shapes. The main difference is in the colour pattern of the caudal fin: a black lower marginal band in agilae and red vertical bars in geayi. Maybe the former represents a yellow phase and the latter a blue phase, as is often the case with Aphyosemion (Huber,1978). Moreover, there remains the possibility that they are sympatric, locations "B" and "D" being only 5km apart.
5- Rivulus cf. urophthalmus Günther.1866
Some Rivulus, with bodies striated with reddish brown dots. and of average size, -60 to 70 mm TL- were- discovered in several places in French Guyana, near the coast: loc. "D" mentioned above, loc "C" 18km north-west of Kourou on the national road RN1; loc. "F" on the island of Cayenne about 20km from the city and loc. "G" at Montjoly, about 6 km south-east of Cayenne. The male has a whitish yellow marginal band at the top and bottom of the caudal fin. The female, less colourful, has the usual black ocellus surrounded by bright yellow. Variability between the populations is slight in the males and of some importance in the females. It tends to disappear in the aquarium. whilst a temporary fasciated pattern appears when the female is frightened.
For the time being the fish has been placed together with Rivulus urophthalmus, one of the most abundant species, with an enormous distribution range and which probably covers several different species.
Hoedeman (1961, op. cit.) reports a related fish at Zanderij (with agilae) in Dutch Guyana: the colour pattern of the body is the same but the caudal fin is edged in its entirety by a black marginal band. The identification of this fish will remain in doubt until specimens of the typical Rivulus urophthalmus have been collected again at Para (Belem), in northern Brazil about 800km south-east of the Guyana locations.
D = 8, A = 12, L.L.= 33.
6- Rivulus aff. holmiae Eigenmann 1901.
A very large fish. up to 150 mm TL was caught at location "D" with Rivulus xiphidius and agilae. Its morphology is similar to that of the fish described tram Holmia, British Guyana. However the colour pattern is different. Here the male has a black marginal band on the lower part of the caudal fin, whereas Rivulus holmiae has font-family: a whitish yellow double band at the top and bottom of the caudal fin. Here the adult female does not seem to have an ocellus whilst in Dutch Guyana it does.
0ne cannot but point out the parallel with Rivulus urophthalmus, in which the two colour patterns also exist. But. the two bands of urophthalmus are associated in French Guyana with the black band of holmiae (and conversely in Dutch Guyana).
Until now we have had little information on the biotope of Rivulus species. That is why I have decided to describe them in detail.
A small stream with a current, shallow (20cms in February,1978). in the forest of 'Varzea' which is persistently flooded. Black water over sand. No aquatic plants and weak light. Rivulus xiphidius: a trio, at the edge in a semi-stagnant spot with dead leaves. Rivulus geayi, numerous in flowing water.
Water: pH 13.0pt; = 5; TH 1; 23°C. (P. Blanc, pers. comm.).
Permanent water course even at the end of the dry season. Weak light: The bottom is of sand and stones. Rivulus xiphidius very rare, with a predominance of females, and geayi. In a darker deeper part, Rivulus aff. holmiae (G. Oelker, pers. comm.).
A pond rapidly drying up. Grassy area, very sunny. This water hole is formed in the rainy season and it is there connected with a few watercourses. Rivulus cf. urophthalmus. Other fish: Cichlids, Hemigrammus sp., Nannostomids, and Hoplosternum (very numerous). Temperature about 35°C. (G. Oelker pers. comm.).
A small very slowly flowing river. Very sunny with considerable of water in the dry season, but it never dries out completely. Rivulus13.0pt; agilae and rarer Rivulus cf. urophthalmus. Other fish as in location "C" but the Hoplosternum do not seem to be present, whereas some Copella sp. were caught. Water temperature: up to 30°C. (G. Oelker. pers. comm. ).
Two biotopes, very close to each other were examined.
1. Stagnant ditch, in a sunny place near the roadside, 1.5 m wide, 30 cm deep, clayey soil, muddy, the edges densely covered with grass, a few Nymphea sp. Rivulus cf. urophthalmus only. Percentage of males. about 50%. Physico-chemical conditions of the water on 27th February 1978, at 3 pm: pH= 6.6, TH= 1, water temperature: 28°C.
2. Stagnant pool in a mangrove (Rhizophora present) of flooded savannah, muddy bottom, depth 70cm, well lit. Vegetation: Cyperaces. Rivulus cf. urophthalmus abundant, with one dominant male. Monopoecilus melanozonus, large Cichlids (G. Oelker, pers. comm.).
Descriptions of locations "E", "F" and "H" have not been reported.
The beginner can breed most Rivulus species. They are hardy fish and so their maintenance is no problem.
This is true of Rivulus amphoreus. Frans Vermeulen states (20 Nov. 1977): "A very easy fish, but it is an excellent jumper and it will get through the smallest gap. It prefers a well planted aquarium. If several pairs are kept together, one male always becomes dominant and the edge of its eyes becomes bright yellow green. The species is prolific and the large eggs (1.9 - 2 mm in diameter) are laid near to the surface and hatch in two weeks. Growth is rapid with the result that after four months the young begin spawning".
Here too are the observations of V. Montiel (Dec.1978).
1. Rivulus agilae and geayi.
Not a very prolific fish (4 to 5 eggs daily); average size of the eggs (1 to 1.2mm in diameter) which hatch in 18 days at 23°C. Month old fry Arial; are 10mm long and they grow rapidly until they are 30mm long, when the growth rate slows down. But, strangely enough, the fry of agilae can be sexed from the first week, because by then the ocellus appears at the top of the caudal peduncle. Adult size is reached at 5 1 - 6 months and the fish begin to breed at about 8 months.
2. Rivulus cf. urophthalmus.
This fish is, in nearly every respect, similar to the typical Rivulus amphoreus. However the eggs are smaller, with a diameter of 1.2 mm. Both sexes show the ocellus until the age of four months, when it disappears in the males.
3. Rivulus aff. holmiae.
This is a large fish and a big eater, taking chopped earthworms. It spawns happily at the bottom and at the surface. The very large eggs, 2.2 - 2.8 mm, hatch after wet or dry storage of about a month at 24°C. All the fry have an ocellus until the age of about 4 months, when it fades and disappears. Growth is rapid until about 4 months, when the fish should be about 80mm, then it slows down until the fish reaches adult size after 12 5.0pt;font-family: - 13 months.
4. Finally some species, for example Rivulus xiphidius, are very difficult. It is a very small fish in which the females seem to outnumber the males. The eggs are large, 1.6mm in diameter, and very few in number. The first hatching took place after nine days in difficult conditions. The first fry took two days to get out of its shell completely and then eight days to absorb its yolk sac. Subsequently incubation lasted 18 days. Growth is extremely slow and adult size is reached in approximately six months.
Sexual development comes later and the first spawning has not been achieved until the fish are a year old.
Before preserving the specimens, I was able to observe the six Rivulus species in a single aquarium. Actually their relative behaviour is clearly different. Rivulus amphoreus is active, living in open waters. Rivulus cf. urophthalmus is quite passive, often near the surface, in the so called "lazy" position, well Arial; known with Rivulus, which consists of keeping the head horizontal and the rest of the body hanging down in a curve. Rivulus agilae and geayi swim peacefully in open water and seem more shy than the others. With Rivulus aff. holmiae the male is active and often stays near the surface, whereas the female prefers the areas near the bottom of the tank. Lastly, Rivulus xiphidius is completely different: usually it stays motionless in open water and behaves remarkably like Characoids of the nannostomus group, a characteristic which again gives it an isolated position within the genus Rivulus.
These are a few preliminary observations which need following up with closer study of these fish. With the morphological characteristics, the detailed descriptions of the fishes pattern and colouration and the study of the biotopes, this is a first step in our getting to know about the genus Rivulus..
TABLE 1 Proportions and meristic characteristics of the 2 new species.
Species | Sex | SLmm | TLmm | TL | PD | PA | PV | Hd | Ht | D | A | D/A | LL |
Riv. xiphidius, holotype | F | 22.3 | 26.9 | 121 | 71 | 62 | 51 | 25 | 20 | 9 | 11 | +8 | 32+1 |
paratypes | M | 22.1 | 27.1 | 122 | 71 | 66 | 52 | 24 | 20 | 9 | 11 | +7 | 32 |
F | 21.8 | 27.1 | 124 | 72 | 65 | 54 | 25 | 20 | 9 | 11 | +8 | 31 | |
F | 20.5 | 25.2 | 123 | 69 | 62 | 52 | 25 | 18 | 9 | 12 | +8 | 31+1 | |
F | 22.0 | 27.0 | 123 | 69 | 62 | 52 | 26 | 18 | 9 | 12 | +8 | 31+1 | |
F | 17.5 | 21.4 | 122 | 70 | 64 | 52 | 26 | 18 | 9 | 12 | +7 | 30 | |
Riv. amphoreus, holotype | M | 57.9 | 64.7 | 112 | 67 | 55 | 48 | 21 | 23 | 12 | 17 | +8 | 47+2 |
paratypes | F | 51.4 | 58.5 | 114 | 67 | 55 | 48 | 20 | 22 | 11 | 16 | +9 | 48 |
M | 53.2 | 57.5 | 108 | 68 | 57 | 47 | 19 | 20 | 11 | 17 | +9 | 48+2 |
SL= Standard Length ( mm ). TL= Total length ( mm). TL%= total length, as a % of standard length. PD%= pre-dorsal length (%). PA%= pre-anal length (%). PV%= pre-ventral length(%). Hd = length of head (%) Ht= depth of head (%). D= number of dorsal rays. A= number of anal rays. D/A= Deviation of dorsal fin insertion vs. anal fin. LL= number of scales in a longitudinal series.
Literature.
Hoedeman (J.J.) 1959.- Rivulid Fishes of Suriname and other Guyanas. Studies Fauna Suriname, 3(7): 44-98, figs, 8 tables 5 pls.
Hoedeman (J.J.),1961.- Notes on the Ichthyology of Surinam and other Guianas, Bulletin of Aquatic Biology, 2(17): 61- 64.
Hoedeman (J.J.),1961.- Preliminary key to the species and subspecies of the genus Rivulus. Bull.of Aquatic Biology, 2(18): 65-74.
Huber (J.H.),1978.- Caractères taxinomiques et tentative de regroupement du genre Aphyosemion, Rev. fr. Aquariol. 5(1): 1-32.
Myers (G.S.),1927.- An analysis of the Genera of Neotropical Killifishes allied to Rivulus. Ann. Mag. Nat. Hist., 9(19): 115-129.
Translation by Peter Watkins for the BKA.
Reviewed by J. Huber (February 2000, for Internet)
Original publication in French: Huber, J.H. 1979. A propos de quatre nouvelles Collections de Rivulus des Guyanes, avec Description de Rivulus xiphidius n. sp. et Rivulus amphoreus n. sp. Rev. fr. Aquariol., 6 (3): 65-72, 11 figs., 2 maps, tab.
also published in English in Killi News (BKA, 1982), 206: 1-12
Subsequent notes:
Rivulus cf. urophthalmus is described as Rivulus lungi Berkenkamp, 1984.
Rivulus aff. holmiae is described as Rivulus igneus Huber, 1991.
Rivulus geayi from locality "B" is disclosed to be distinct from the standard geayi (locality "A") and described Rivulus cladophorus Huber, 1991.
Additional reading of this author's publications ? Select any ONLINE article.