BENTHIC AND SUBTERRANEAN AQUATIC OLIGOCHAETE FAUNA (ANNELIDA, OLIGOCHAETA) FROM COIBA ISLAND (PANAMÁ) AND CUBA

A total of 15 taxa of aquatic oligochaetes have been identified in two collections, one of which was from Coiba island, at the National Museum of Natural Sciences (MNCN, Madrid), and the other from Cuba (leg. Dr. Ana Camacho). Morphological and systematic remarks are provided for Limnodrilus variesetosus, Pristina aequiseta forms evelinae and foreli, P. breviseta, P. osborni, and P. sima. Pristina cabacuensis Botea, 1983 and P. decui Botea, 1987 are proposed as junior synonyms of Pristina aequiseta Bourne. The taxonomy of Pristina osborni and P. sima is discussed and their independent specific status supported. The alluroidid Brinkhurstia americana is reported from Coiba island widening the known geographical distribution of the species to South and Central America.


Introduction
In the present work, new data from two collections of oligochaete worms from both hyporheic and running waters in Coiba island and from subterranean waters in Gibara, Cuba, are provided.Located in the Pacific Ocean, close to the west shore of Panamá, Coiba island is exceptionally unaltered because it has been a prison since 1920 (García-Valdecasas et al., 1997).The National Museum of Natural Sciences (MNCN) in Madrid, Spain, is currently studying the freshwater fauna of Coiba as an example of unaltered, neotropical, small-sized rivers and they have generously made the aquatic oligochaete collection available for identification.This paper also includes descriptions of oligochaete worms from subterranean waters in Cuba, sampled by Dr. Ana Camacho who has also kindly sent me the oligochaete worms for determination.

Study sites and methods
The studied localities in Coiba island and Cuba, with sampling data and the reference sampling codes in the collections at the MNCN (Museo Nacional de Ciencias Naturales) are listed below.See Fig. 1 for the geographical location of the two study areas.
Collection from Coiba island (Panamá).Except for restricted agricultural activity, the land of Coiba island has not been influenced by human activity (García-Valdecasas et al., 1997).The island presents a well-preserved tropical forest with a large number of small rivers and streams, in which no evidence of organic pollution has been detected.Environmental variables and sampling procedures are described in detail by García-Valdecasas et al. (1997).The studied sites show conductivity values within the range of 125-450 µS cm -1 , water temperatures between 24-26°C and dissolved oxygen over 6.4 mg l -1 .The worms were fixed in 4% formaldehide or 70% ethanol and conserved in 70% ethanol.The collection is maintained in the National Museum of Natural Sciences (MNCN) in Madrid and the Department of Zoology of the University of the Basque Country (UPV/EHU).Rodriguez, 1999 RECORDS: site 30 (10 ind.), type locality of the species (Rodriguez, 1999) Limnodrilus variesetosus Brinkhurst, 1979 (Fig. 2) RECORDS: site 32 (10 ind.)DESCRIPTION: Immature worms.Triangular prostomium.Bifid chaetae.Dorsal bundles 2-4 chaetae, usually 4; ventral bundles 3-6 chaetae.In anterior segments, ventral chaetae thicker and longer than dorsal of the same segment.Anterior dorsal chaetae from II to V with upper teeth about 1.2 to 1.6 times longer than lower, being of about the same length from VI or VII, backwards.Ventral chaetae of these same segments with upper teeth about 2 to 3.5 times longer than lower, gradually reducing the difference from segment VI backwards.Posterior segments have thinner chaetae, with the upper teeth equal or shorter, and thinner than the lower.The position of the nodulus in both dorsal and ventral chaetae goes from median (distal portion/total length: 0.48-0.50) in the anterior segments to distal (up to 0.40) in the most posterior segments.

Monopylephorus camachoi
REMARKS: Among the American Limnodrilus species, both L. variesetosus Brinkhurst, 1979 andL. rubripenis Loden, 1977 show enlarged ventral chaetae with very long upper teeth in some segments of the anterior part of the body.The examined material has been identified as L. variesetosus because of the presence of the enlarged chaetae in the first segments II (III) to V, instead of IV (V) to IX, characteristic of the L. rubripenis.Chaetae in the new material from Cuba and in the Jamaican population of L. variesetosus are basically similar (Fig. 2A, B and C).Brinkhurst and Wetzel (1984) considered this species as a possible variant of L. udekemianus.Brinkhurst and Marchese (1989) also discussed the possible synonymy of both species.Compared with   L. variesetosus, L. udekemianus (studied in Spanish specimens) can be distinguished by the teeth proportions in chaetae which always have the upper tooth longer than the lower, in both anterior and posterior bundles (Table 1).In both species, the upper tooth is thicker than the lower in the anterior segments (Fig. 2 B).However, in L. udekemianus posterior chaetae have upper and lower teeth of the same width (Fig. 2 D), whereas in L. variesetosus the upper tooth is thinner than lower (see Fig. 2 C and Brinkhurst, 1979).The scales in the original drawing (Fig. 1, in Brinkhurst, 1979) are necessarily wrong for the chaetae.
The existence of variants in the size of chaetae has been demonstrated in other tubificid species related to ionic characteristics of water (form grandiseta of Tubifex tubifex, see Rodríguez and Armas, 1983).Environmental data from the site (García-Valdecasas et al., 1997) do not suggest any apparent characteristic different from a typical freshwater site.Since the worms of our collection are immature, the contribution to clarify the taxonomic status of this species is limited.

NAIDIDAE
Stephensoniana trivandrana (Aiyer, 1926) (Fig. 3) RECORDS: site 1 (1 ind.), 26 (1 ind.)DESCRIPTION: 1 to 1.6 mm long.Diameter of the body 142 µm in IV and 164 µm in IX.Number of segments 11 and 18. Prostomium very short, 40 µm long (Fig. 3A).The body wall is 5-8 µm thick and it is incrusted by foreign material, not densely covered in the anterior part and very dense in the posterior part of the body.Dorsal chaetae start in segment II.Dorsal bundles consist of 1-3 smooth hair chaetae and 1-3 thin needles, suddenly tapering towards the distal end.Two hair chaetae per bundle in the first segments (115-128 µm long), 3 in the following V to VII segments (144-150 µm long), decreasing to only 1 in the last segments of the body (120-152 µm long).Hair chaetae are about 2 µm thick at their base, close to the tegument.Needles, 40 to 43 µm long (ca. 1 µm thick), simple-pointed, with a distal, inconspicuous nodulus at about 1/4 of the distal apex.
Ventral bundles with 3 bifid chaetae, decreasing to 1-2 in most posterior segments, with the nodulus proximal, at about one third of the base.Ventral chaetae 62-64 µm long in anterior segments, up to 85 µm in middle and posterior segments.Upper tooth grows backwards, being 1.6 to 2 times longer  than lower in anterior segments and up to 3 times longer in the posterior segments.
REMARKS: The chaetae measurements are within the range reported in the literature (Table 2).In America, the species has been observed in several localities (USA: see Brinkhurst 1986; Mexico: see Brinkhurst and Marchese 1989), Surinam (Harman, 1974), Haiti (Dumnicka, 1986) and Argentina (Di Persia, 1980).The finding in Panamá confirms the extensive presence of this species in American freshwater habitats.The species has not been reported in Europe and Australia, so far.Dorsal bundles formed by one "hairy" hair chaeta, increasing in length in the anterior segments of the body (from 85 to 126 µm) and decreasing to 111 µm, backwards.One needle per dorsal bundle, 28-31 µm long and about 1.5 µm thick, bifid, with equal tiny teeth, of about 1 µm long.

Pristina aequiseta
Ventral chaetae, 3-4 per bundle in segments II to IV, one giant chaetae in V, 5-6 in the following segments, and (2) 3-4 per bundle in the posterior part of the body.Ventral chaetae of segment II are longer than the following ones, except for the giant chaeta (see Table 3).Chaetae of segment II are thinner (1.5 µm) than the following ones (2 µm); the giant chaetae reach a thickness of 4.8 µm.The nodulus of ventral chaetae in segment II is proximal (distal portion /total length = 0.67), median in the next anterior segments, with a tendency to a more distal position in the posterior segments (up to 0.41).Upper tooth (U) longer than lower (L) in anterior segments (U/L = 2.5-3.0 / 2.0 µm), changing to an equal length backwards (2.2 / 2.2 µm).
Ventral chaetae 3-4 ( 5) per bundle in anterior segments, and occasionally 1 or 2 per bundle, backwards.Anterior chaetae with proximal nodulus (distal portion /total length = 0.51-0.57) in segment II and gradually more distal in the following segments.Upper (U) tooth longer than lower (L) in anterior segments (U/L = 3.7 / 2.5 µm), changing to an equal length in the median part of the body (2.5-3.1 µm), and shorter than the lower in the most posterior segments (U/L = 2.4 / 3.0 µm).REMARKS: The "hairy" appearance of hair chaetae is not an easy character to observe, even under a magnification of x1250.Harman (1966) did not see serration in the Mississippi material, and he expressed some doubts about its reliance (Harman, 1974).Ventral chaetae in segment II are not, or just slightly, differentiated in length and thickness from the other ventral chaetae.This fact has previously been reported in populations examined by Dumnicka (1986) in the Caribbean Islands.
Pristina aequiseta, forms foreli and aequiseta have been reported in several localities in the Caribbean region: the West Indies and Haiti (Dumnicka, 1986), Surinam (Harman,1974), Venezuela (Botea, 1987), as well as in other parts of America (Costa Rica and Nicaragua: Harman, Colombia is also reported in Brinkhurst and Marchese, 1989).
In Cuba and Venezuela, Botea described 5 new Pristina species, and among them Pristina cabacuensis Botea, 1983 andP. decui Botea, 1987 are probably synonyms of P. aequiseta.P. cabacuensis was described from specimens sampled from 2 localities in Cuba (one specimen in brackish waters).The author discussed their resemblance with P. foreli and P. aequiseta.We have incorporated the information of chaetotaxy of the species in Table 4, and their range of values is within those reported for P. aequiseta.The author stated that the most significant difference of the species is the length of the needle teeth which is described as long; however, their length is only 0.5 µm, thus  they are in fact very short.Other supposed differences relate to the number of ventral chaetae per bundle, 2-4 forwards and 2-3 (5) backwards, but this is within the range of variation known for P. aequiseta (form foreli) (see Sperber, 1948).In my opinion, the only significant difference with P. aequiseta is that the teeth of ventral chaetae in anterior segments are of equal length.
Pristina decui was described from Venezuela hyporheic fauna by Botea (1987).The author compares the new species with P. aequiseta.The description of the species is very poor and does not provide information regarding measurements of needles and ventral chaetae.The absence of serration in hair chaetae of P. decui could be due to both the difficulty to see this character depending on the orientation, and also to its variability (Harman et al., 1988).The position of the stomach dilatation in IV or V is consistent with the possibility of an individual regenerating from fission, since the stomach dilatation in the Pristina genus is usually found in VI to VIII.Again in this case, the only significant difference with P. aequiseta (form foreli) is in my opinion the length of the teeth in the anterior ventral chaetae.In this respect, it is interesting that the specimens described by Dumnicka (1986) as P. foreli from the West Indian Islands in the Caribbean region also showed a very slight difference in the length of the teeth of anterior ventral chaetae.
One smooth hair chaeta, thinner in segment II (ca.1.0 µm) than in the following segments (1.5-2 µm).The length of hair chaetae increases at the anterior region of the body (76 to 221 µm), and in the posterior region they gradually shorten, down to 48 µm.One needle per dorsal bundle (30-48 µm long, 1.2 µm thick), bifid, with equal teeth, 3.0-5.0µm long in segments II to VII (Fig. 4B), 2.5-3.0 µm in segments VIII to X, and 2.0-2.2µm in the posterior segments of the body.Posterior needles show unequal short teeth, with upper teeth slightly thinner and shorter than lower (Fig. 4D).
Ventral chaetae 3 per bundle in the anterior and median segments, decreasing to 1-2 in the posterior part of the body.In the anterior segments, ventral chaetae gradually increase in length backwards (48-58 µm), and gradually shorten towards the posterior part of the body (48-53 µm).The thickness of ventral chaetae is similar all along the body (2.5 µm), although slightly thinner in the most posterior bundles.The nodulus of ventral chaetae in segment II is nearly median, with a progressive displacement towards a distal position in the following segments (distal portion / total length=0.47 in segment II, and 0.37 in the most posterior segments).Upper tooth (U) longer than lower (L) in segment II (U/L = 4.3 / 2.5 µm), gradually becoming of equal length in the median region, and the upper smaller than the lower in the posterior segments (U/L = 2.5 /3.0 µm) (Fig. 4A).
REMARKS: The species P. breviseta may or may not present a proboscis, as is clearly illustrated by Marcus (1943, Fig. 101).The presence/absence of a proboscis is now recognised as a character that may be variable at an infrageneric or even infraspecific level within the Pristina genus, since Collado and Schmelz (2000) have included the Pristinella genus in synonymy with Pristina.
The question of long vs. short needle teeth can be a confusing character for identification purposes, since the length may be variable, as is shown in the studied individual (Fig. 4).The ratio of needle tooth length to the distal part of the needle (from the nodulus to the tip) varies between 0.25 and 0.1,  which is shorter than the relation of one third reported by Marcus (1943, Fig. 102C).A similar relationship (ca.0.20) is present in the African species P. rotundirostris Grimm, 1985, which also has smooth hair chaetae and parallel needle teeth.However, in this species the ventral chaetae are more numerous per bundle and the length of teeth of ventral chaetae is described as equal.
Measurements of the chaetae reported here for P. breviseta are within the range or are smaller than those reported by other authors (Table 5).The length of ventral chaetae increases in the anterior segments of the studied specimen of P. breviseta, a fact that was also reported by Marcus (1943), although it is a common characteristic in many species of the genus Pristina.Immature specimens of P. breviseta can be distinguished from P. synclites by the length of the teeth in ventral chaetae, described as equal in P. synclites, and the upper longer than the lower in P. breviseta.
Ventral chaetae 3-4(5) per bundle in anterior segments (28-37 µm long), increasing in length in first segments of the body, and (2)3-4 in the posterior segments (32-41 µm long).Chaetae of segment II thinner than the following ones.Nodulus of ventral chaetae median in segment II, being gradually more distal towards the middle and posterior part of the body (distal portion/total length = 0.44-0.5 in segment II up to 0.33-0.37 in the most posterior segments).Upper (U) tooth longer than lower (L) in ventral chaetae of segment II (U/L = 2.8-3.2/1.8-2.2 µm), with a tendency towards an equal length in the following segments, and the upper shorter than the lower in the posterior ones (U/L = 1.8-2.0/2.0-2.2 µm).
Clitellum not clearly differentiated.One pair of spermathecal pores in segment VII (Fig. 6A).One  (c) Sociedad de Amigos del Museo Nacional de Ciencias Naturales y Consejo Superior de Investigaciones Científicas Licencia Creative Commons 3.0 España (by-nc) http://graellsia.revistas.csic.espair of male pores in segment VIII.Tubular atria.Spermathecae are located in the most anterior part of the segment, close to the anterior septum.They are formed of a spherical ampulla (15 µm diameter) and a wide duct (23 µm long, 15 µm max. width).One pair of seminal funnels in septum VII/VIII which connect with atria.No modified genital chaetae.
REMARKS: Sperber (1948) stated the probable synonymy of P. minuta (Stephenson) and P. osborni (Walton), an opinion shared by Brinkhurst and Kathman (1983).As P. minuta, the species has been reported in Texas (Harman, 1973) and Brazil (Marcus, 1943;Righi, 1973), and it has also been reported in Africa and Asia.The range of variability of the chaetae measurements is very similar for both species (Table 6).
There has been a considerable taxonomical confusion between P. sima and P. osborni, and a part of the reports of P. sima are probably due to the presence of P. osborni with pectinated needles.The presence or absence of intermediate teeth in the same individual in this species was first reported by Loden and Harman (1980) who pointed out the possibility of misidentification of P. osborni as P. sima because of the occasional presence of pectinated needles.The original description of P. sima by Marcus (1944, Fig. 59D) shows the typical needle bent in its distal third, with relatively long teeth, distal thinner and shorter than lower, while needles are straight with short and about equal teeth in P. osborni.One of the needles drawn by Grimm (1986) for African P. osborni populations could be of the sima-type.The needles drawn for P. sima by Varela (1990) in individuals from Argentina are of the osborni-type.Harman et al. (1988) and Brinkhurst and Marchese (1989) regarded P. osborni as a probable synonym of P. sima and P. minuta.On the other hand, it is my opinion that the form and size of the needles support the independent status of P. sima (see remarks of this species below).
(c) Sociedad de Amigos del Museo Nacional de Ciencias Naturales y Consejo Superior de Investigaciones Científicas Licencia Creative Commons 3.0 España (by-nc) http://graellsia.revistas.csic.esdiverging, upper tooth shorter and thinner than lower (U/L: 2.2/2.8-3.9 µm), and all pectinated (3 intermediate teeth).Ventral chaetae 4-6 per bundle in anterior segments, and 4 (rarely 3) in the posterior segments (34 to 45 µm long); nodulus located between one third and one half of the distal end of the chaetae (distal portion /total length=0.36-0.43)and a clear tendency in the size, from anterior to posterior chaetae is not observed.Upper tooth longer than lower in ventral chaetae of segment II, and smaller in posterior ventral bundles.
REMARKS: The presence of pectination in the needles is a constant in the studied collection.The number of the intermediate teeth in the needles is variable in some studied populations (e.g. 1 to 3: Rodriguez, 1984;2 or 3: Martínez-Ansemil and Giani, 1980).Ercolini´s description (1969) includes the presence of denticulate hair chaetae, in contrast to Marcus´s description (1944), Martinez-Ansemil and Giani (1980), Grimm (1990) and this study.
The specimen reported by Davis (1982) is probably a tubificid as can be judged from the shape of the needles (sigmoids and with longer distal teeth in needles) and their size (approx.71 µm long).Varela (1990) reported P. sima from Argentina; however, this material is probably P. osborni.Loden and Harman (1980) proposed a potential relation of P. sima with P. rosea-group species, due to the difference in length of needle teeth.Brinkhurst and Marchese (1989) suggested the synonymy of P. sima and P. minuta with P. osborni.However, P. sima can be well separated from these species by the characteristic bayonet-like form of the needles, with teeth longer and markedly unequal compared with those in P. osborni/minuta (compare data in Tables 6 and 7).
Pristina sima has been reported by several authors in Africa, Europe and America.It has been found in North America (N.York: Strayer and O'Donnell, 1988) and Central and South America (Martinique: Chagné and Giani, 1989;Guyana: Stacey and Coates, 1996;Brasil: Marcus, 1944;Argentina: in Brinkhurst and Marchese, 1989;Chile: Gluzman, 1990).Interestingly, the species has been reported in both freshwater and estuarine brackish water (Martínez-Ansemil andGiani, 1980 and1982).This is the first report of the species in Cuba.Ventral chaetae very variable in number, differing in number also in the bundles of the same segment.Ventral chaetae of segment II longer than the following ones, and thickness is ca. 2 µm, all along the body.Tooth length is variable within the chaetae of the same segment, the upper tooth being longer, equal or shorter than the lower in anterior segments, and shorter in posterior ones.Nodulus in ventral chaetae median or slightly distal in segments II and III (distal portion / total length = 0.43-0.50)and distal in the following segments (0.34-0.41).
Septal glands in segments V to IX. Intestine from VIII.One pair of testes in the anterior part of segment X, one pair of ovaries in the anterior part of segment XIII.Over one of the testes, a sperm sac could be seen.Egg sac extends backwards to XVII.Sperm funnels open down septum X/XI (Fig. 7A).Male duct, "atria" according to Brinkhurst (1964) and Righi et al. (1978) descriptions, and"prostate" after Omodeo (2001), long and tubular, convoluted in segment XIII (Fig. 7B).Three different sections can be distinguished in the male duct: a first section formed by finely granulated cells, about 170 µm long, 30-62 µm diameter; a second non-granulated section, about 700 µm long, 25-45 µm diameter, formed by very dense cells; and a third narrow tubular section, about 550 µm long, 18 µm diameter.The second section and part of the first are covered by a diffuse glandular layer (6-16 µm high), made of cells containing large granules, which form some masses between adjacent coils.The duct which connects the sperm funnel with the ectal section of the male duct was not observed, probably broken during dissection.
One pair of penial chaetae (380-400 µm long, 13 µm max.width), simple-pointed, associated with the male pores, into a muscular sheath (36 µm max.width) in segments XII and XIII.A thick retractor muscle (up to 27 µm thick) joins the longitudinal musculature of the tegument in the anterior part of segment XII to the proximal part of the muscular sheath which contains the penial chaeta.
One single spermatheca in the dorsal part of segment IX, formed by a compressed spherical ampulla (236 µm max.diameter) which opens in the mid-dorsal line through a short vestibulum (about 90 µm long), with thick muscular layer (max.25 µm thick).REMARKS: There are few differences with previous descriptions of the species.The penial chaetae are longer than those described by Brinkhurst and Jamieson (1971) (115 µm long, 18 µm max.diameter), but shorter than the description by Omodeo and Coates (2001) from Guyana (>560 µm by 12.2 µm).Compared with the original description (Brinkhurst, 1964), the position of the spermathecal pore is median rather than anterior in segment IX.The form of the spermatheca resembles that described by Omodeo and Coates (2001) more than that by Brinkhurst (1964).Sperm sacs have not been reported in previous descriptions.
The type specimens studied by Brinkhurst (1964) were from Argentina.The species has also been reported in Brazil (Righi et al., 1978) and Guyana (Stacey and Coates, 1996;Omodeo and Coates, 2001).Therefore, the new finding of the species in Coiba island, reflects an extensive distribution of the species Brinkhurstia americana in South and Central America.

Marionina gr. subterranea
RECORDS: site 1 (15 ind.), 28 (2 ind.)REMARKS: the worms lack lateral bundles of chaetae all along the body and they have 2 chaetae per ventral bundle, simple, straight and bent in their proximal end.

Discussion
The freshwater oligochaete fauna of Central America and the Caribbean region has been previously studied by several authors.The oligochaete families Naididae, Opistocystidae and the aphanoneuran Aeolosomatidae have received special attention through contributions by Michaelsen (1933) and Harman (1965 and1982), who reported twenty one species from Central America (Salvador, Costa Rica, Guatemala, Nicaragua and Honduras).More recently, Dumnicka (1983 and1986) studied a collection from subterranean habitats in several Caribbean islands and North Venezuela, reporting a total of four tubificids and twelve naidids.Botea (1983) studied three collections of subterranean oligochaetes and aelosomatids from Cuba and provided a list of 34 species, of which 17 were described as new to Science, although descriptions are quite poor and incomplete.A posterior study by the same author reported 15 species of aquatic oligochaetes and aelosomatids from subterranean waters in Venezuela, of which 3 were new to Science (Botea, 1987).Medeiros and Neves (1982), Coates and Stacey (1994), and Stacey and Coates (1996) reported about 50 taxa from Bonhaire island and Guyana, respectively.More recently, Chagné and Giani (1998) have reported 35 species from Martinique.The diversity of the fauna in the region is relatively large taking into account the limited number of studies, and more than one hundred taxa have been reported so far.In this context, the taxa identified from Coiba island in the present study should be regarded as a preliminary list of the oligochaete species, which will undoubtedly be amplified with further studies.
The absence of a cosmopolitan fauna in the samples is in concordance with the general characteristics of the sites in Coiba island, which is practically undisturbed and unpolluted.It will be also interesting to determine in future studies if any important differences in the distribution of oligochaete fauna exists between the Pacific islands and the Atlantic Caribbean islands.The present study would seem to suggest that such differences are minimal.

Fig. 1 .
Fig. 1.-Location of the two study regions in Central America (Coiba island and Cuba).

Table 5 .
-Several morphological features of Pristina breviseta from different sources.Abbreviations: s, number of segments.