- Open Access
Food habit of the endangered yellow-spotted newt Neurergus microspilotus (Caudata, Salamandridae) in Kavat Stream, western Iran
© Farasat and Sharifi; licensee Springer. 2014
Received: 15 September 2013
Accepted: 27 August 2014
Published: 10 September 2014
Diversity and abundance of macroinvertebrate fauna were simultaneously determined in selected benthic samples and in regurgitated stomach contents in Neurergus microspilotus in Kavat Stream (western Iran) during April and May 2012. The aim of this study was to determine the degree of reliance of this species to benthic macroinvertebrates during their reproductive season in aquatic habitat.
Twenty-one taxa of macroinvertebrates were identified in the benthic samples while 19 taxa were presented in the regurgitated stomach contents. Data obtained from benthic samples showed that the most abundant macroinvertebrate groups included Lumbricidae (27.2%), Mycetophilidae (20.06%), Gammaridae (12.19%), and Plananariidae (9.3%). Data obtained from 45 stomach contents indicated that on average the highest importance values combining number, frequency, and volume for prey categories consumed included Mycetophilidae (14.03%), Baetidae (13.68%), Corbiculidae (12.57%), Gammaridae (10.8%), and Lumbricidae (9.34%). N. microspilotus also consumed small stones, plant materials, and their own eggs (0.91%). The analysis of selectivity in feeding using Ivlev’s index showed that the prey taxa that appeared to be preferred (E i > 0.5) were generally rare in the environment.
Comparison between benthic macroinvertebrates and those taken by the newt demonstrates that although high similarity (Sorenson index of 78.94%) exists between the two communities, the dominance of the items taken by N. microspilotus (Simpson index = 0.32) is higher than that of the benthic community (Simpson index = 0.20) indicating that the newts rely on fewer number of species with higher proportion of individual prey items. Feeding habits of 45 N. microspilotus have shown that the newts rely extensively on Mycetophilidae, Baetidae, Corbiculidae, Gammaridae, and Lumbricidae as important food items for N. microspilotus.
Relatively few Caudata occur in Iran. These include seven species of the genera Triturus, Paradactylodon, Neurergus, and Salamandra (Baloutch and Kami ). Newts of the genus Neurergus have a relatively wide geographic distribution, ranging from western Iran (Zagros Mountains) and extending to Iraq and southern Turkey (Baloutch and Kami ). The yellow-spotted newt occupies an assortment of aquatic microhabitats during the breeding season. Visual determination of substrate texture in Kavat Stream indicated that this newt tends to occupy substrates containing gravels and pebbles (Sharifi and Assadian ). Two species of the genus Neurergus (N. kaiseri and N. microspilotus) are listed as critically endangered by the IUCN criteria (IUCN Red List of Threatened Species. Available from http://www.iucnredlist.org). This species is listed as critically endangered because its area of occupancy is less than 10 km2, and there is a continuing decline in the extent and quality of its stream habitat and in the number of subpopulations and individuals because of habitat degradation, drought, and over collection of animals for both national and international pet trade. Habitat loss through divergence of streams for irrigation is probably the most important factor that threatens the species in its Iranian range (Sharifi and Assadian ). Investigation made by Sharifi and Assadian () on N. microspilotus have confirmed that this newt occurs in several highland streams in the mid-Zagros mountains but is highly vulnerable to the rapid changes occurring in their aquatic and terrestrial habitats.
There is no available information regarding feeding habits of N. microspilotus in Iran, Iraq, or possibly in southern Turkey. Among closely related species in southern Turkey, it is evident that the Lycian salamander (Mertensiella luschani) chiefly preys upon aquatic insects. Serdar and Rizvan () have shown that this species preys mostly on aquatic Coleoptera. In a similar study based on the stomach contents in Salamander leurognathus (Martof and Scott ), it was shown that more that 70% of food items in this species consist of Ephemeroptera and Trichoptera. Maerz et al. () have shown that although feeding habits of salamanders differ at different times of the year and in various habitat types, however, Oligochaeta, Coleoptera, and Isopoda are the most important food items of this animal. Feeding habits have also been reported in different species of amphibian including Bufo melanostictus by Sreelatha et al. () and Pleurodema diplolistris by Santos et al. (). These studies have shown that although frog diet consisted of a wide variety of arthropods including Diptera and Coleoptera, the aquatic forms did not contribute much to their diet. Other studies have demonstrated that the prey items identified in the diets of different species of Anura shows that these species are generalist and opportunistic predators whose diet is most strongly influenced by prey availability (Toshiaki ; Kerim and Ahmet ; Caldart et al. .
The present investigation aims to determine variation in diversity and abundance of benthic macroinvertebrates. This study also intended to show food preference and feeding habits of N. microspilotus which may reflect the availability of prey items in Kavat Stream.
2.1 The species
Three species of the genus Neurergus have been reported to occur on the Iranian plateau, in Northern, Central, and southern parts of the Zagros Mountains. These include N. crocatus Cope  from Northwestern Iran, Northeastern Iraq, and Southeastern Turkey; N. microspilotus (Nesterov ) from Western Iran and Iraq; and N. kaiseri Schmidt  from the Southern Zagros Mountains in Lorestan and Khusistan provinces in southern Iran. Previous studies on this genus are scant and mainly limited to original descriptions and anecdotal explanations. However, Schmidtler and Schmidtler () studied different populations of Neurergus and confirmed the presence of three allopathic species belonging to this genus from Iran and a fourth species in Turkey.
2.2 Study area
2.3 Benthic macroinvertebrate fauna
Sampling from benthic macroinvertebrates was performed on 20 April and 15 May 2012. Benthic macroinvertebrate fauna were collected by kick sampling in the fast-flowing waters in Kavat Stream. This sampling involved shuffling through the substrate within a quadrate (0.4 × 0.4 m) against another quadrate with the same size which had a mesh bag and located perpendicular to the benthos in the opposite direction to the flow of the stream about 1 m from the water’s edge. All invertebrates were killed in the field using small quantities of 40% formaldehyde and later preserved in 96% ethanol for further examination. Further analyses carried out in the laboratory included counting and sorting the specimens under suitable magnifications (×7 to 40). The benthic macroinvertebrates were identified using manuals of Bouchard (), Thyssen (), and Parker and Consulting (). Shannon-Wiener index of general diversity (H = −∑ (n i /N) log (n i /N)) was used to express the diversity of benthic macroinvertebrates. Simpson index of dominance (c = ∑ (n i /N)2) was used to determine how relative importance of different species is distributed within the community. Sorenson index of similarity (I S = 2C/A + B) was also used in order to compare degrees of similarity between benthic macroinvertebrates and those taken by the yellow-spotted newt.
2.4 Stomach content
N. microspilotus used in the present study (45 individuals, 22 in 20 April and 23 in 15 May 2012) were all caught at daytime. The newts were captured by hand. Gastric lavage was used to extract the stomach contents of the live animals. A tube was inserted through the newt’s mouth into its stomach, and the stomach was pumped by a 60-ml syringe full of water until the newt regurgitates the stomach contents. These food items were filtered from the water and preserved in 96% ethanol solution for later identification and quantification. Although it may not be a pleasant experience for the newts, the survival rate of N. microspilotus that has gone through the gastric lavage was 100% in the present study. The permit for collecting N. microspilotus for the present study was issued by the Kermanshah Department of Environment. In extracting the stomach contents using gastric lavage, we firmly determined to avoid any casualty to the specimens. The newts were kept in small (30 × 30 cm) pools by putting up several stones at the sampling site for approximately 2 hours to see if this experiment causes any visible side effect and then released. This experiment showed no mortality.
Prey categories consumed by N. microspilotus were analyzed in terms of the number, occurrence, and volume of each prey category. The volume of each prey was estimated by the formula of an ovoid spheroid, proposed by Dunham (), V = 4/3 π (L/2) (W/2), where L corresponds to the greatest length and W to the largest width of the prey. An index of importance (I x) was calculated for each prey category by the formula proposed by Caldart et al. () by summing the percentage of occurrence and the numeric and volumetric percentages of each prey in the diet and dividing it by 3. The sufficiency of the sample to assess in feeding habit was evaluated by an accumulation curve of prey categories, using EstimateS 9.1 software with 1,000 random additions (Colwell ).
Ivlev’s E i index (Ivlev ), E i = (n i − r i) / (n i + r i ), was used to estimate selectivity in feeding behavior in N. microspilotus. In this equation, n i represents the proportion of prey taxa i in the stomach contents and r i represents the proportion in the benthic macroinvertebrate community. E i can vary between −1 and 1. In this study, the thresholds of E i = 0.5 (Cogalniceanu et al. ) are used to determine the selectivity in feeding behavior. The thresholds of E i > 0.5 are considered preferred, and those with E i < 0.5 are considered avoided food items i. To evaluate the correlations between relative abundance of benthic macroinvertebrate in Kavat Stream and regurgitated stomach contents, Pearson’s correlations were performed.
2.5 Statistical analysis
In order to determine correlation and frequency of species in sampled quadrates and the stomach contents, correlation coefficients were determined between relative abundance of various taxa in sampled quadrats and the newt regurgitates using Microsoft Office Excel 2007 and SPSS statistical package (version 15, SPSS Inc., Chicago, IL, USA).
A checklist of the benthic invertebrates sampled
Number of classes, orders, families, and species in each phylum of the benthic organisms
Electivity values of prey categories consumed by N. microspilotus
Electivity (E i)
Eggs (of themselves)
Frequency of empty stomachs and group characteristics of preys in the diet of N. microspilotus
Empty stomachs (%)
Maximum no. of prey/individual
Average no. of prey
9.9 ± 8.4
Average number of prey species
3.2 ± 1.6
Aquatic prey (%)
Terrestrial prey (%)
Prey categories in the diet of N. microspilotus
Amphipoda-Gammaridae (aq- ad)
Coleoptera-Dytiscidae (aq-l and ad)
Ephemeroptera- Heptageniidae (aq-l)
Eggs (of themselves)
Unidentified arthropod remains
The feeding selectivity in N. microspilotus as expressed by the Ivlev’s selectivity index (E i) was computed (Table 3). The analysis of selectivity in feeding using Ivlev’s index (E i) showed that most of the prey taxa that appeared to be preferred were generally rare in the environment. The highest electivity was found for Caelifera and eggs (E i = 1), followed by Corbiculidae (E i = 0.88), Bithyniidae (E i = 0.87), and Baetidae (E i = 0.52). The lowest values were obtained for Planorbidae, Tipulidae, Chamaemyiidae, and Ephemeridae (E i = − 1), followed by Lumbricidae (E i = − 0.72), Planariidae (E i = − 0.70), and Heptageniidae (E i = − 0.58) (Table 3).
There is no available information concerning freshwater macroinvertebrate fauna in highland streams in western Iran. Relatively low diversity of benthic macrofauna in this study is not unusual and may be the consequence of low order of the stream. A wide range of α-diversity has been reported for low order streams. Hawkeswood () has reported only 14 species for the Murrumbidgee River, near Wagga Wagga, New South Wales, Australia. George et al. () have reported 19 species in the Okpoka Creek in the Niger Delta. Also, Kazanc et al. () has reported 21 species belonging to five classes from the channel entrance of Lake Koycegiz to the Mediterranean Sea, Turkey. Whereas macroinvertebrates are indicators of the water quality, the absence of polychaetes in the macroinvertebrate fauna from Kavat Stream may be attributed to the high level of water quality and lack of organic pollutants in this stream. This assertion is in agreement with the observation of many researchers (e.g., Mendez et al. ; Harlan ; Musale and Dattesh ; Omena et al. ) who reported that polychaetes were found in association with sites grossly polluted with organic matter, heavy metals, and petroleum hydrocarbons.
Benthic macroinvertebrate diversity in the two sampling occasions in April and May varies. In April, when water discharge was considerably high, 15 taxa were presented in the sampled quadrats. This increased to 17 taxa in May. The relative abundance of different taxa as expressed by the percentage of number also changed. In April, Lumbricidae (Aporrectodea rosea and Eiseniella tetraedra) and Mycetophilidae (Rhymosia sp.) comprised 75.15% of the number of benthic macroinvertebrates, whereas in May, Gammaridae (Gammarus daiberi), Planariidae (Polycelis feline), Heptageniidae (Maccaffertium sp.), Hydropsychidae (Cheumatopsyche sp.), Baetidae (Baetis sp.), and Mycetophilidae (Rhymosia sp.) cover over 84.3% of the number of individuals. From April to May, α-diversity as measured by Shannon-Wiener index of diversity of the benthic macroinvertebrates increased (0.69 to 0.90) but dominance as measured by the Simpson index reduced from 0.26 to 0.14.
Analysis of regurgitated stomach contents including the empty stomachs (11.81%) showed that on average 9.9 ± 8.4 benthic macroinvertebrate items are consumed by N. microspilotus. Compared to similar values for prey diversity in other Caudata, Covaciu-Marcov et al. () have demonstrated that the Carpathian newt (Lissotriton montandoni) feeds on only 2.72 items. However, similar value for the great crested newt (Triturus cristatus) is 9.8 (David et al. ). The feeding intensity of the yellow-spotted newts in the first sampling occasion is lower (10.44) than the second sampling occasion (13.14). A low rate of the feeding intensity at the beginning of activity period has been reported in other species of amphibians (Hirai and Matsui ; Kovacs et al. ). Such an increase in feeding intensity is believed to result from unfavorable weather conditions, mainly low temperatures, which affect both the predators and the prey (Guidali et al. ; Covaciu-Marcov et al. ). Similar to other species of newts such as Triturus cristatus (Kutrup et al.  and Dobre et al. ), the yellow-spotted mountain newts mainly feed on aquatic prey. In the present study, the average terrestrial item found in regurgitated stomach contents of the yellow-spotted newts comprise only 0.6% of the total prey items. The consumption of these terrestrial preys indicates that some individuals of N. microspilotus may leave the water and forage in the terrestrial environment; it is also possible that the terrestrial insects have drifted into the water by wind.
In the second sampling occasion in May, some newts consumed (1.81%) their own eggs. N. microspilotus also consumed a high proportion of cobbles and plant materials. Although some stones may have been swallowed accidentally, it is possible that most of them were similar to soil-associated species such as caddis flies (Trichoptera) that were taken as food. The presence of caddis fly and stone fragments has been reported from the stomachs of another similar-sized salamander species, Salamander leurognathus (Martof and Scott ). Prey categories contained in the stomachs of 45 individuals indicate that these mountain newts feed heavily on aquatic arthropods. However, in early spring when the water discharge is high, the newts avoid the main stream and occur in subterranean seepages and parallel shallow streams. At this time, they feed mainly on Corbiculidae (31.6%), Lumbricidae (24.89%), Cecidomyiidae (14.38%), and Gammaridae (14.1%). Later, when the water discharge is reduced, the newts enter into the main stream and feed mainly on Mycetophilidae (30.77%), Baetidae (30.55%), and Gammaridae (19.17%).
The feeding selectivity in N. microspilotus is expressed by Ivlev’s selectivity index (E i) indicating that there is an inconsistency among the abundance of benthic macroinvertebrates and the feeding items taken by N. microspilotus. For example, the most abundant prey taxa in the benthic community (Lumbricidae (27.20%), Plananariidae (9.3%), Hydropsychidae (6.38%) show very low and negative Ivlev’s index values of −0.72, −0.70, and −0.58, respectively (Table 3). The analysis of selectivity in feeding using Ivlev’s index showed that the prey taxa that appeared to be preferred (E i > 0.5) were generally rare in the environment. The low selectivity values obtained in the present study may be due to the lack of closely related species of Caudata in the study area (Sharifi and Assadian ). Competition among such species has been proposed as one of the principal mechanisms that can eventually lead to resource partitioning and species coexistence.
Comparison between benthic macroinvertebrates and those taken by the newt demonstrates that although high similarity (Sorenson index of 78.94%) exists between the two communities, dominance of the items taken by the yellow-spotted newts as expressed by the Simpson index (0.32) is higher than that of the benthic community (0.20). This indicates that the newt relies on fewer numbers of species with more balanced numbers of individuals of different species. Considering that N. microspilotus consumes the majority of the benthic macroinvertebrates (17 out of 21) reported in this study, it should be considered a nonspecialist or generalist predator. However, comparison between the relative abundance of the benthic macroinvertebrates and those taken by the newt (Figure 3) shows a coefficient of determination r 2 = 0.17. Feeding habits of the 45 N. microspilotus have shown that the newts rely extensively on Mycetophilidae, Baetidae, Corbiculidae, Gammaridae, and Lumbricidae and other important food items for N. microspilotus.
Hossein Farasat (first author) is a PhD student at the Department of Biology, Razi University, Kermanshah, Iran. He is studying various aspects of ecology and taxonomy of the yellow-spotted mountain newt under supervision of Mozafar Sharifi (second author). The present study is a part of the first author’s PhD course.
We thank the organizations that supported this study, in particular, the Iran National Science Foundation (Contract code: 91057377) and Razi University that financially supported this study as a part of a PhD research project.
- Baloutch M, Kami G: Amphibian of Iran. Tehran University Publications, Tehran; 1995.Google Scholar
- Bouchard RW Jr: Guide to Aquatic Macroinvertebrates of the Upper Midwest. Water Resources Center University of Minnesota, St. Paul, MN; 2004.Google Scholar
- Caldart VM, Iop S, Bertaso TRN, Cechin SZ: Feeding ecology of Crossodactylus schmidti (Anura: Hylodidae) in Southern Brazil. Zool Stud 2012, 51: 484–493.Google Scholar
- Cogalniceanu D, Aioanei F, Ciubuc C, Vadineanu A: Food and feeding habits in a population of common spadefoot toads ( Pelobates fuscus ) from an island in the lower Danube floodplain. Alytes 1998, 15: 145–157.Google Scholar
- Cope ED: On Neurergus crocatus from Iran and Iraq. Proc Acad Nat Sci Phil 1862, 1862: 343.Google Scholar
- Covaciu-Marcov SD, Cupsa D, Cicort A, Naghi N, Vesea L: Date despre spectrul trofic al unor populatii de Triturus alpestris din zona Muntelui Ses (jud. Bihor, România). Oltenia, Studii si Comunicari. Stiint Nat 2003, 19: 171–176.Google Scholar
- Covaciu-Marcov SD, Cicort-Lucaciu AS, Diana CIS, ÉVA HK, Ferenti S: Food composition of some low altitude Lissotriton montandoni (Amphibia, Caudata) populations from North-Western Romania. Arch Biol Sci 2010, 62: 479–488. 10.2298/ABS1002479CView ArticleGoogle Scholar
- Colwell RK: EstimateS: Statistical Estimation of Species Richness and Shared Species from Samples., Vers. 9. 2013.Google Scholar
- David A, Cicort-Lucaciu AS, Roxin M, Pal A, Nagy-Zachari A: Comparative trophic spectrum of two newt species, Triturus cristatus and Lissotriton vulgaris from Mehedinţi County, Romania. Biol 2009, 3: 133–137.Google Scholar
- Dobre F, Bucur DM, Mihut R, Birceanu M, Gale O: Date asupra compozitiei hranei a unei populatii de Triturus cristatus (Laur. 1768) din Parcul National “Defileul Jiulul”. România Biol 2007, 1: 23–28.Google Scholar
- Dunham AE: Realized niche overlap, resource abundance and intensity of interspecific competition. In Lizard Ecology: Studies of a Model Organism. Edited by: Huey RB, Er P, Schoener TW. Harvard Univ, Cambridge, MA; 1983:261–280.Google Scholar
- George ADI, Abowei JFN, Daka ER: Benthic macroinvertebrate fauna and physico-chemical parameters in Okpoka creek sediments, Niger Delta, Nigeria. Int J Anim Vet Advs 2009, 1: 59–65.Google Scholar
- Guidali F, Scali S, Carettoni A, Fontaneto D: Feeding habits, niche breadth and seasonal dietary shift of Rana dalmatina in northern Italy. Current studies in herpetology. Soc Eur Herpetol 1999, ᅟ: 161–166.Google Scholar
- Harlan KD: The use of polychaetes (Annelida) as indicator species of marine pollution: a review. Rev Biol Trop 2008, 56: 11–38.Google Scholar
- Hawkeswood TJ: A brief investigation into the benthic macroinvertebrate fauna of one section of the Murrumbidgee River, New South Wales, Australia, using kick, net and surber sampling methods. Calodema 2004, 2: 1–5.Google Scholar
- Hirai T, Matsui M: Feeding habits of the Japanese tree frog, Hyla japonica , in the reproductive season. Zool Sci 2000, 17: 977–982. 10.2108/zsj.17.977View ArticleGoogle Scholar
- Ivlev VS: Experimental ecology of the feeding of fishes. Yale, New Haven, Connecticut; 1961.Google Scholar
- Kerim ÜEK, Ahmet M: Food composition of the marsh frog, Rana ridibunda Pallas, 1771, in Thrace. Turk J Zool 2007, 31: 83–90.Google Scholar
- Kazanc N, Oguzkurt D, Girgin S, Dugel M: Distribution of benthic macroinvertebrates in relation to physic-chemical properties in the Koycegiz-Dalyan estuarine channel system (Mediterranean Sea, Turkey). Indian J Mar Sci 2003, 32: 141–146.Google Scholar
- Kovacs EH, Sas I, Covaciu-Marcov SD, Hartel T, Cupsa D, Groza M: Seasonal variation in the diet of a population of Hyla arborea from Romania. Amphibia Reptilia 2007, 28: 485–491. 10.1163/156853807782152534View ArticleGoogle Scholar
- Kutrup B, Cakir E, Yilmaz N: Food of the banded newt, Triturus vittatus ophryticus (Berthold, 1846) at different sites in Trabzon. Turk J Zool 2005, 29: 83–89.Google Scholar
- Maerz JC, Myers EM, Adams DC: Trophic polymorphism in a terrestrial salamander. Evol Ecol Res 2006, 8: 23–35.Google Scholar
- Martof B, Scott D: The food of the Salamander Leurognathus. Ecology 1957, 38: 494–501. 10.2307/1929894View ArticleGoogle Scholar
- Mendez N, Flos J, Romero J: Littoral soft-bottom polychaete communities in pollution gradient in front of Barcelona (Western Mediterranean, Spain). Bull Mar Sci 1998, 63: 167–178.Google Scholar
- Musale AS, Dattesh VD: Distribution and abundance of macrobenthic polychaetes along the South Indian coast. Environ Monit Assess 2011, 178: 423–436. 10.1007/s10661-010-1701-3View ArticlePubMedGoogle Scholar
- Nesterov PV: Trois formes nouvelles d’Amphibiens (Urodela) du Kurdistan. Annu Mus Zool Acad 1916, 21: 1–30. [in Russian]Google Scholar
- Omena EP, Lavrado HP, Paranhos R, Silva TA: Spatial distribution of intertidal sandy beach polychaeta along an estuarine and morphodynamic gradient in a eutrophic tropical bay. Mar Pollut Bull 2012,64(9):861–873.View ArticleGoogle Scholar
- Parker D, Consulting AT: Identification Key to the Orders of Saskatchewan Aquatic Insect Larvae and Adults. 2012.Google Scholar
- Santos JWA, Damasceno RP, Rocha PL: Feeding habits of the frog Pleurodema diplolistris (Anura, Leptodactylidae) in Quaternary sand dunes of the middle Rio Sao Francisco, Bahia, Brazil. Phyllomedusa 2003, 2: 83–92. 10.11606/issn.2316-9079.v2i2p83-92View ArticleGoogle Scholar
- Schmidtler JJ, Schmidtler JF: Untersuchujngen an westpersischen Bergbachmolchen der Gattung Neurergus (Caudata, Salamandridae). Salamandra 1975, 1: 84–98.Google Scholar
- Schmidt KP: Diagnoses of new amphibians and reptiles from Iran. Nat Hist Miscellanea 1952, 93: 1–2.Google Scholar
- Serdar DM, Rizvan TM: Analysis of the stomach contents of the Mertensiella luschani (Steindachner, 1891). Asiat Herpetol Res 2004, 10: 164–167.Google Scholar
- Sharifi M, Assadian S: Distribution and conservation status of Neurergus microspilotus (Caudata: Salamandridae) in western Iran. Asiat Herpetol Res 2004, 10: 224–229.Google Scholar
- Sharifi M, Vaissi S: Captive breeding and trial reintroduction of the endangered yellow-spotted mountain newt Neurergus microspilotus in Western Iran. ᅟ 2014, 23: 159–166. Doi: 10.3354/esr00552Google Scholar
- Sreelatha KS, Natarajan P, Ritakumari SD: Studies on the food and feeding behaviour of Bufo melanostictus . J Eco Biol 1990, 2: 232–233.Google Scholar
- Thyssen PJ: Keys for Identification of Immature Insects. 2010.Google Scholar
- Toshiaki H: Ontogenetic change in the diet of the pond frog, Rana nigromaculata . Ecol Res 2002, 17: 639–644. 10.1046/j.1440-1703.2002.00521.xView ArticleGoogle Scholar
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly credited.