Rcv.Diol. Trop.46. Supl. 6: 185·199. 1998 WWW.UCr.Bc.er
The benthic macro invertebrate community in Caribbean Costa Rican streams and the effect of two sampling methods
Pia Paabyl.l, Alonso Ramirez] and Cathy M. Pringle] 'Centro de Investigaciones en Ciencias del Mar y Limnologia (ClMAR). Univcl'>idad de Costa Rica. 2060 San JosC. Costa Rica. Fax: (S06) 253·566\ 'MeRida S.A. Consuitores Ambientales. [email protected] l\nsiitute of Ecology, University of Georgia. Athens. Georgia 30602. USA.
(Rec. 20-V-\997. Rev. 13-!V-I998. Acep. 14-VJ.l998)
Ab!tract: The benthic maeroinvertebrate community of streams draining Ihe south-Caribbean coasl of Costa Rica. was studied to detcnnine community stnlcture and composition through time (September 199� - May 1996), the c ITect of vegelation cover on drainage area (forest vs. disturbed sites) and to compare two sampling mcthods; Surber and kitchen sieve. Five sites, predominantly with a forest cover drainagc area, show abundances (rncan� 157.60 in divs.lm1) and taxa richness (mean><9.65 spp.lm') significantly higher thWlthose found in the disturbed draina�e a reas (rneans--82.08and 4.62 respectively). All sites have high community hcterogeneity indices (> O.80) which. to gether with the low similarity indices « 0.40) between sampling times indicate a macroinvenebrate community subject to frequent disturbance evenlS that diminish species pennanence in the aquatic syslem. The kitchen sieve sampling method with len replicates (4 000 em') is significantly more effective in sampling the community in tcrms of abundance (mean"'165.\0) than the traditional Surber sampler (mean=83.03) using three replicates (3 33J em'). Both are equally effective in the number of taxa collected, although each device captures different tan. The species similarity between the Surber and the sieve method ranges between 0.1379 to 0.7619 when combining all sampling times.
Key words: Streams. community structure, tropical Caribbean. sampling method. macroinvertebrates. Costa Rica. richness.
Freshwater ecosystems can be divided into McConnell el at. 1992) contain such taxa two main categories; systems with longitudi richness that attracts the attention of numerous nal current movements (streams and rivers), biologists and human communities, mainly and those with circular or no movement at all due to their icthyc production. Central Ameri (ponds, lagoons, swamps, lakes). Most can systems are, comparatively, poor in fish commonly, these systems obtain the attention species and the opinion within the human of the surrounding human population because population about streams and rivers is that of human dependence on water to survive and they only serve as water conduits. develop (Middleton 1995). Only rarely are The South Caribbean coastal area of Costa these water ecosystems analyzed as something Rica is currently experiencing an increase in beyond water conduits; for example, as eco the rate of development, -especially as a result systems containing a unique biodiversity in of tourism. In consequence, there is a nonnal teracting with the terrestrial and marine eco augmentation in the need for water, as well as systems (Abramovitz 1996). By contrast, the the production of garbage disposed into the Amazon river (Lowe-McConnell 1987, 1994), various surrounding water systems. In respon the rivers in Asia and Australia (Dudgeon and se to these problems, the local community has Lam 1994) and the deep African lakes (Lowe- initiated various efforts towards the unders- REVISTA 1)1'. 111()I.O(jIA '1 RI )['KAI.
I 700, XII II 0001 LOWEST / " \-,,:1 / -__ Average 400 HIGHEST 79·94 � < X' _ . '" - ·",·�,Low I 300 I' ..--- ,/ • f.-- ,/ -- High , �oj.,--,.(I � - Average "-. 95·96 x N --Low , - - High ,x / VlII v, Fig. I. Average monllly rainfall (mm) recorded between (1979-[994) and during tllis stud)' 1995-1996 from tile meteoro logical station SixlOla in Tulamanca. Costa Rica. Source: Hydrology Depanment, InstitUlo Costarriccnse de Electricidad. Roman numbers Kpl'CSC'nleach month with the scate K�sent;ng the rainfall. Aquatic ecosystems ufCosta Ri�a '" sites were located in the Rio Carb6n. Black two unfiltered samples were taken in acid and Sands Creek. Rio Cocles. Rio Cai"io Negro. destilled water washed polyethylene 60 ml Hone Wark stream and Rio Gandoca. Sites bottles to measure total phosphorus and am one. two. five, six and eight are those which monia. as well as two 0.45 �m Millipore fil primarily have a forested drainage area. Sites tered samples for the measurement of nitrate three. four, seven and nine have a disturbed and soluble phosphorus. These samples were (i.e. original vegetalion is CUI) drainage area frozen in the field and taken to the laboratory (Table 1, Fig. 2). of the CIMAR (Centro de Investigaciones en Ciencias del Mar y Limnologia. University of Physical and chemical water analyses: In Costa Rica) to be analyzed within a period of each site and visit, data were gathered water three months. TSS and alkaliniry were deter and air temperalUre with a hand thennometer, mined within a period of seven days. channel width with a 30 m measuring tape, The total suspended solids were detennined depth with a em calibrated stick, and water after filtering a known volume through pre velocity with a Gurly meter at various points dried and pre-weighed Whatman GF/C fillers, across the channel. In addition, dissolved oxy dried at 80°C for 24 hr and weighing the dif gen was measured with an oxygen electrode, ference in my!. The alkalinity detennination pH and conductiviry with a field electrode. was done titrating with a weak H�S04 acid For the analysis of total suspended solids (Wetzel and Likens 1979). (TSS) a I 000 ml water sample was taken, The soluble reactive phosphorus (P04-P) together with a separate 500 ml sample for the and the total P (TP) were done utilizing the detennination of total alkaliniry. Similarly, molybdenum blue colorimetric method (AP- Fig. 2. Samplingsites (numbers) in thearea ofPucrto Viejo Manzanillo and Gandoca, Talamanca,Costa Rica. II. REVISTA DE BIOLOGIA TROPICAL TABLE I Physical and �"'U:ufclKlrucleris/ics oflilt Jlrtii'" ,ftQches visiled in &plemh.u 199J '''rough/996. May P�rto V�jo de TaJomanca, Man::aniflo and Gandoco, Cos/a Rica, G: grazu. B: bano/'ItJplon/s, J: Inga. £: Erylhrin.. Co' COCQ(}, Cu: Cassia. DRF: de"n 'Dlrt/oresl, P: Pel1lachethra matroloba. Cr: Cl)'SOphyllumcaimito . • TSS: la/a/$Uspended soUth. (-) no lfMasurtmt!n/ ... • • • US. ll' tl"S!J')t,' ll".n r F 0 f r o f 0 0 12' �s II 12 II> 11 , • G.8 G.8 O.I:I.I,O,E G.o.ce. 0II.f 0" O".C.E..[ I.J.B.C o.'.Cr ' " " " , , " " " '" " " '�I " n , " " " " " " w n " " I'.J " " " " " " " " " ... ". ... � ,� '" . � •• 'M' •• '" •• •• • • " ... •• ".d. •• ". .., •• •• '" ." ...... OlJJ ...... � .. " hip '" •• 1l-:7 0117 .... ,� '" '" '" '" '" ,. '" " • " , " • " " • " " TSS· " ., ••• " " " " ,."" " " " " ., '" .., .. 9,2 ,�, �oo., .. :.. 2�.' 14.1 '" , ,� .. .. " " " " " • ,,� " .. ,,,,"I n •• • " ,� '" '" " no � " " 'v� • • " '" " " " �-, • " " •• n • " • , ' .... 1 , " • • •• , " " • ...... " " • • .. .. " • • • • •• .. " " """""' ..... " " '" " .. " " I�"'I " " " .. " " � " " " � " " " .. " " � " " " TOIaI·' ,� " '" " '" " " '..,1 " • " " u " ., • " " ., " " • " " " '" � '" � HA 1985), as well as the TP following a con through the indophenol method (Sol6rzano centrated sulfuric acid digestion and a 25' au \969). toclaving. The nitnste (NO]'N) was determined through the hydrazine reduction method Macroinyertebrale eommunily: The aquatic (Kamphake et aI. 1961) and the ammonia macroinvertebrales were sampled at each one Aquatic ccosystem$of Costa Rka 189 of the sites four times through the year; in general tenns, these streams can be defined as September, December, February and May. having wann waters (23-29 "C) during the [ow The most common method to capture the or precipitation months and cool waters du-ring ganisms living in between the substrate parti the rainy months (17-27 "c). These waters are cles of the stream bottom is simple: a net. In influenced by the magnitude of the rains, the present study two slightly different meth showing changes of several orders of magni ods were utilized: tude in total suspended solids and ammonia-N I) the Surber method (I mm1 mesh size), (Table I). which basically encloses a flat area of I ft2 The nutrient concentrations found in the (1111.11 cml). Manually this limited area is streams draining this coastal area are similar to disturbed into approx. I em depth. This those found in other areas of the Costa Rican method was replicated at each site three times for a total area of3333 .33 cm2) (Surber 1937 ). 2) "kitchen sieve" method (2 1 cm diameter, round, lmml mesh size). With this method a limited flatarea in the bottom of the stream of 1 20 x20 cm (400 cml) was manually disturbed. 1 , -and replicated ten times in each site for a total area of 4000 cm1 and a mesh size of I mm2. These replicates were taken choosing a ran dom point to start and then moving upstream in a zig-zag direction, taking a sample every B. 80 -100 em including the center and edges of . the stream. All samples were fixed in the field with 70 i • % ethanol and brought to the laboratory to , separate the organisms from the sediment and , organic matter. All organisms were identified to the minimum taxonomic level possible and counted to detennine the abundance per unit are •. Aquatic macroinvertebrates can be catego rized by their various feeding habits and dif ferential abilities to withstand stream current and turbulence as well as different refuge needs against predators. To detennine if pres ence and abundance of the various types of substratepresent in the streameffects diversity and biomass of the macro invertebrate groups, we sampled: leaves, sediment, and pebbles in rimes and pools. Macroinvertebrates were fixed in the field with 70 % ethanol and taken to the laboratory to be identified, counted and measured for their body biomass detennina tion. The results of this part of the research are presented in Ramfrez e/ af. (1998). RESULTS Fig. 3. Abundanceof the benthic macroinvertebrate com Physical and chemical characteristics: The munity. A. Forest (n-IO) vs. Disturbed (n-8) sites. B. Overall sieve vs. physical and chemical characteristics of the (n-9) Surver (n-9) sampling. C. Distri bution of organisms per forest site and method (n-4).D. sampled sites in the area between Puerto Viejo Distribution of organi� per disturbed site and method. and Gandoca are summarized in Table I. In Bars::I: I standard error. ". REVISTA DE BIOLOOJA TROPICAL F.' D-3 1 � i 1 , I s..... i 1 I 0-7 F-8 i m IXI95 1 o XII/95 , o 11/96 am V/96 ...... Fig. 4. Total nwnbcrof indivklualslm' through time:. Sieve method: 4 000 em'; Surbermethod: 3 333 em'. September 1995-May 1996.Note thai vertical scale is not tmsame for all graphs. Aquatic ccosystems ofCosta Rica 191 Caribbean (Paaby and Goldman 1992, Sanford R._ el al. 1994). In addition, and similar to other A. Costa Rican systems, these stream waters contain occasionally sufficiently high concen· t trations of nitrate-N, ammonia-N and phos , phorus-P as to define them as eutrophic (Bis 1 hop 1973, Jordan 1985. MargalefT 1983). Abundance: The benthic macroinvenebrate F""... ed ""'""" community is differentially represented be Mdhod tween the sites grouped as forested and dis B. turbed according to their drainage vegetation cover, with average abundance/m2 signi 1 ficantly higher in the forested sites (Two-Way , ANOVA F=5.49, P=0.0344. Fig. 3a). Addi 1- tionally, significantly higher abundanceslm2 of macroinvenebrates were captured using the "sieve method" (Two-Way ANOVA F=5.82, S;.,n .... P=0.0302, Fig. 3b-d). ,.� The abundance of macroinvenebrates os • C. cillates temporally for both methods and sites as.... " types (Fig. 4). In December the abundance . " - was the highest. It is wonh mentioning that � " ; " this month was a dry month (Fig. I). In the 1 , month of February, the absence of organisms , in all sites may evidence the effect of flooding (Fig. I). Samples from May, at all sites, except " " " R " 9, show no recovery in the numbers of ma • croinvenebrates back to values similar to [);st>.,rbod " D. those found in December (Fig. 4). " [] So<". " Taxa richness (taxa): A total of 82 taxa were " c_ , " collected from the six sampled streams (Table , -1. 2). However, only a few species constilUted the communities of the sampled sites (Fig. 5a· d, Table 2). Each site, combining both sam pling methods) contains from 16�3 taxa in the forested sites and from 13-41 in the dis turbed ones, with statistical evidence sup Fig. 5. Species richness of thc benthic macroinvcrtcbratc poning a higher taxa richness in the forested community. A. Forest (n�IO) vs. Disturbed (n�8l sitcs_ B. Overall sieve (n-9) vs. Survcr (noo9) sampling. C. drainage area sites (Two-Way ANOVA Distribution of organisms per for est sitc and mc:lhod F=6,J1, P=O.0249). Temporal effects on the (n�). D. Distribution of organisms per disturbed site and taxa richness is evident, showing a larger mcthod. Bars: ± I standard efTOr. number of taxa during the less rainy months of the year, September and December (Fig. 6). from 0.237 - 0.667, with 30 % of the compari Composition: The total number of taxa pres sons> 0.45. The similarity index between the ent in each sampled site (Table 2) is a panial disturbed sites (3,4,7,9) are almost the same indicator of community structure as this ab with values ranging between 0.286 - 0.667 but solute number hides imponant composition with only 16.6 % of the comparisons> 0.45. changes throughout the year and between When comparing the "forested" sites and the sites. The species composition similarity be "disturbed" sites the similarity ranges between tween the five forested sites (1,2,5,6,8) ranges 0.280·0.788, with 35 % of the comparisons> 192 REVISTA DE BIOLOGIA TROPICAL F-' "� 11 ., � ., � s""'" D-4 H � I I Sieve F-' D-7 Sieve Sieve F"; � � ., illl 1X/95 �, ., o XII/95 � o 11196 IDJl V/96 Sieve Fig. 6. Total numberormorphospecieslm1 through time. Sieve method: 4 000 em'; Surber method: 3 333 em', September I99S-May 1996. F - Forestsites, D - disturbed sites. Aquatic ecosystems of Costa Rica 193 0.45. There was not a clear trend showing a 60 % of the temporal comparisons and < 0.40 higher similarity between sites under the same in 81 % of them (Table 3b). These values are degree of disturbance in their drainage area low and indicate continuous changes in the are absent (One-Way ANOVA F=O.09, species composition within 2-3 month inler P=0.7683, Table 3a). vals. Similarly, the morphospecies composition The sampling methods, Surber and "Kit of the sampled community through time varies chen sieve" do not differ significantly (Two considerably, with similarity indexes < 0.25 in Way ANOVA F= 1.88, P=0.1914) in the TABLE 2 Taxa prtstnt (+) persilt and sampling fIItthod combining all dates throllgh September 1995 and May 1996. Puerto Viejo de Sarapiquf. Manzanilla and Gandoca de Talamanca. Cruta Rica Forested Disturbed 3 4 7 9 I 2 568 Si � 51 Su Si 5u 5i h Si h 5i 5u Si Su Si 5u 5i Su + Epbemcropterl + + Baetidac: Camt/abatlidus + + + Batlis + Baelodts + Cloeodts + + + + + + + sp.1 + + + sp.2 Leptophlebiidae Thrau/odts + + + + + + + + + + Travarella + + + + Fam)(us + + + + + + Hydrrumi/OOarf + + + + + + + sp,l Leptohyphidae LeptahypMs + + + + + + TriearylhOOts + + + + Hap/ohypMS + + + + + + + + + + + + sp.1 + + sp.2 + spJ Caenidae Caenis + + Ephemeridae Hexngenla + + + Palymitarcyidae + + + + + Trichoptcrl + + Helicopsychidae Pkilopotamidae Chimarra + + sp.1 Hydroptilidae Oclvotrlehla + Mayalriehla + + + sp.l Hydropsychidae LeptGntmtJ + + Sfllicri�a + + + + + + + + + ".1 + + + + + + + + + + Glossosomatidae Mae,�ma + + + Leptoceridae + Odontoccridae Calamoceratidae Phyllaicw + + + Coleopterl Gyrinidae DlneutUJ + '" REVISTA DE BIOLOGIAlltOPICAL ConI. TABLE 2 Forested Distu!bed 2 , , • ) 6 7 , Tm SI S, SI S, SI S, SI S, SI S, SI S, SI S, SI S, SI S, Coleoptera Elmidae Jlexacylfotpus + Phunocerus + MicrocylfotplLS + $p.' + + + + + + + + + Psc:phenidae + + + Scil1idat Cyphorn + Plitodaclylidac flcmipltra Dclostomatidac BelosfOma + + A"duJ + + + Vcliid¥ RhagovrfiD + Micro�fia + + + + + :r;,'Olll'a Zygopltnl + + + + + Lcstidac ArclrihWt! + + + Anisoptcra Libcllulidac Perif�mi3 + + + Gomphidac A8riogOltlplrUJ + + + + Ptrlgomphl4 + + + + C'pdogOl1lphus + + Progomphus + + £pigQmphus + 1\If1!lloplUI Coryd�hdac COIJ'liolus + + + + PlctoptcrII J'erlidae AnacrOfH"mia + Dipltrl + Chirooomidae sp.l + + + + + + + + + + + + + + + + + + ¥:.2 + + + + ipulidac + + + Jlexaloma + + + 51'.1 + + Limoniinac + Simuliidae + Ccratopogonidac + Oolichopodidac + Ephidridae + lAp'doplcra Pymlidac (Crambidac) Ptlrophllo sp.1 + + + Ptlrophlla 51'.2 + Cruslun Macrobrachiam + + + + + + + sp. L + + + + + + + + + + + + + + Alydac ' .. + + + + + + + Dcupoda + + + + + + + + + 1\101111$(. GUlropoda + + + + + + + + 5p.1 + 51'.2 + + + Aqualic ecosyslems ofCosla Rica '9' COIIf,TABLE 2 FOrf:sled Dislurbed , 2 • , , ) 6 7 9 Tm s; s, s; s, Si s, s; s, s; s, Si S" s; s, Si Su S; s, 8i.... t�il + Ollgochlrll + + TOlltlul 32 " 36 21 12 6 13 12 16 , " , 12 , 20 , " 7 number of taxa captured. The species simila rity between both sampling methods do not TABLE] differ between forested and disturbed sites, SorrnMn Simifariry indu Clmon8 situ (oj. comparing with a maximum similarity not > 76 % in site sampling methodsper site (b). comparing sampling datu 3 (One-Way ANOVA F=O.OO. P=0.9945. Ta (&pt. Drc /995, FtbruClry. May/996) (cj. Purrto Virjo. ble 3c). Man:aniflo and Gandocu de Tu/umunco, Costa RicCI. F-Jorested. D- disturbed Species diversity (H') (Whitaker 1965): Be •• cause the diversity of a community is directly Fores!ed Di,rurbed dependent upon the total number of taxa pres 2 , , • 7 • , ent, the species diversity (H') was shown as Ihe diversity relative 10 the site's maximum 0.67 possible diversity (H'/H ....,). (Fig. 7). The 0.24 0,27 relative diversity is high (88.6 % > 0.5 HfHm. . for forested; and 59 % > 0.8 H/Hm.. for dis 6 0,4) 0.48 O.H turbed) with a weak trend towards showing & 0,3S 0.31 0.33 OJ7 higher values during the rainy months (e.g. 0.32 0.4S 0.40 0050 0.35 May, Fig. 7). 4 0.34 0,37 O.H 0,47 0.33 0.67 The distribution of individuals among the 7 0.38 0.35 0.77 0.79 0,49 0,36 0.34 taxa is a common indicator of community 9 0.32 0.21 O.B O.l I 0,59 0.29 0.l3 0." I evenness vs. dominance. In all of the sampled sites, except 4, the homogeneity (J) of the community is high (> 60%) with temporal •• differences thai bring the evenness down 10 only 20 % (Fig. 7). A highly homogeneous Si!c 'Vcln. lXJ95 1XI'J5 X1W5 XIIl9' community structure is an indicator of liule lion ..,. YO. ..,. ..,. species dominance. type' xn,l9s V/96 lt196 V/% To explain diversity trends it is common to invoke physical (e.g. substrate) and temporal F 0,00 OJ7 0.00 F 0.66 0,22 0, 1 2 (e.g. intermediate disturbance) heterogeneity. , F 0.33 In the studied systems there was not a signifi , F 0." cant evidence for a positive correlation be • F o.n 0.11S 0.21 0. .'16 0.00 tween substrate heterogeneity (Table 4) and J 0.S9 0.00 0.00 D the total abundance of macroinvertebrales pre • D 0." 0.33 OJ75 7 D 0.32 0.22 0.32 sent in a site (r.,.....- 0.44; r_= 0.23), or the , D 0.00 0.40 0.00 0.00 0.00 taxa richness (r...... "" 0.27; r.....,.,.= 0.23) or bio diversity (r,,-- 0.00; r..-,= - 0.10). <. Diarurocd DISCUSSION 8 3 .. , 6 • State of the freshwater ecosystems draining Sieve �I. the coastal area between Puerto Viejo and ,.. .. 0.S1 0.'9 0.22 0.S6 0.33 0.76 0 .... 0.14 0.44 Gandoca de Talamanca: The benthic mac roinvertebrate communities were highly 196 REVISTA DE BIOLOGIATROPICAL , . F-I D-) .. "' ,. J " J .. .. .. .. , Sieve S""", , Sieve S""", F-2 : D-4 .. I J : J �: ! I I 01 I Sieve S",boo- IJSieve S""'" F-5 D-7 .. J .. J .. Sieve S""'" Sieve S""'" F-<; 0-, .. " .. J J .. .. .. "' Sieve S",boo- Sieve m IXl95 0 XII/95 ['] 11/96 OlD V/96 Sieve 8""'" Fig. 7. Evenness Index (J) through lime. Sieve mCIIIod: 4 000 em'; Surber mClhod: 3 333 em', September I99S-M,y 1996. F - Forest siles. D - disturbed sites. diverse at all the sampled sites in both the for· These high values are accompanied by high esledand disturbed immediate drainage areas. total abundance in the sites of less disturbance Aquatic c�osystems of Costa Ri�a 197 suggesting higher stability and permanence. the average age of the resident species. There However, there are very low similarity values is an obvious absence of dominant species and between sites and through time suggesting a a low species similarity between the sampling community in constant change. In addition, intervals (2-3 months). there is little evidence supporting the theory In conclusion, the benthic macroinverte that physical heterogeneity (substrate) is cur brate communities of the sampled region evi rently playing a determining factor in the denced , in their structure and function. cons maintenance of diversity. tant changes and under great influence by the The Puerto Viejo, Manzanillo and Gandoca rains. During the study period, the frequency areas are characterized by a very variable rain and intensity of Ihe rains were such that the pattern between years. The sampling year in benthic species were unable to maintain a sta cluded months with precipitation levels' out ble structure through time. Disturbanc..e in side any levels reported in the previous 19 biological systems constitutes a mechanism years, either with too much rain (e.g. 200 mm that maintains diversity; however, ils pQsitive distributed in only IWO days) or 100 little « impact is reversed when the magnitude and 100 mm/month of December) (Fig. 1). The the frequency surpasses an intennediate state temporal variance in rain patterns generates (Connell 1978; Ward and Stanford 1983). The direct effects on the riverine systems. Most implications of this in watersht;d management obviously, the disturbance caused by the policy are important because the changes in heavy rains may result in a reduction or total land use through construction, road opening, removal of the benthic community. Conse forestry plantations. and urbanization, among quently, future changes in the structure of the many. may result in a homogenization as well community and subsequent pattern of organic as in an augmentation of the normal distur matter transformation will be the result of the bance pattern affecting the aquatic systems. pattern of recolonization of individuals and Any of these extremes can have effects on the species. The drainage vegetation cover may presence and abundance of the species. the diminish the effect of heavy rains; however, rate and pattern of recolonization, as well as during this climatic event no differences be on the species interactions and functioning in tween forested and disturbed drainages were the organic matter processing within the sys evident. tem. The composition of the macro invertebrate community in the Gandoca-Manzanillo study Sampling or the benthic community: The area is very heterogeneous among sites and kitchen-sieve constitutes a good method to sampling times. Thisheterogeneity means that acquire a good representation of the composi there is a high species turnover rate within the tion of the benthic community, it is accessible community, which in tum suggests that the to most people, it is easy to handle and trans development of the community reinitializes port. and it permits a sampling of more repli frequently « 2 months). When disturbances cates and microsites per site. In the current occur within the margin of predictability of study, there were \0 replicates per site; how magnitude and frequency, the result is a ever, it is recommended to duplicate this up 10 unique successional sequence generated from at least 20, to increase the probability of cap the initial group of species present. Hence, turing rare species in the area as a result of a two sites may share a high proportion of spe high disturbance rate. cies but differ in the relative abundance of each. However, when the magnitude and fre ACKNOWLEDGEMENTS quency of disturbance surpasses the "normal" behavior, the result is the disappearance of This study was done as a result of the great species and of communities like was encoun efforts made by Jenny Myers: her constant tered with inthe GandocalManzanillo area and perseverance and patience in the long and te found elsewhere by Reice (1994) and Connell dious road to acquire funding for the project. (1978). Additionally, the high evenness of the The funding support comes from The Nature communities observed supports a pattern of Conservancy's Ecosystem Research Program disturbance whose turnoverrate is shorter than in the project "Value of Wetland Transition ,,, REVISTA DE BIOLOGIA TROPICAL Zones in ProtCi:ting the Nutrient Balance of REFERENCES Coaslal Aquatic Ecosystems. Talamanca-Cari Abramovitz. J. N. 1996. Imperiled Waters, Impoverished bbean Biological Corridor, Costa Rica". In Future: The Ikeline of Freshwiler Eoosystems. addition. we profoundly thank the support Worldwatch Paper 128. Worldwalch Institute. Wash received by Ihe CIMAR personnel, especially ington. D.C. 18 p. Carmen Durin. Of course, work like ours al ways requires additional help in the field and Bishop. J. E. 1973. Limnology of a Small MaI.yan River. Sungai. Gombak. Dr. W. Junl:. TheH�uc. 485 p. laboratory which we received from Grettel AgUero, Andrea Borel and Olman Alfaro. The Connell. J. II. 1978. Divenity in tropical n1infOfeSllI and local work was done with the cooperation of coralreefs . Science 199: 1302-\310. ANAl Associalion who provided the local Dudgeon, O. &. P. K. S. Lam. 1994. Inland .....ters of facilities. ANAl's good and posilive altitude tropical Asia and AUSlJIIia: Conservalion and man always made our stay comfortable. To all agement. Mitt. Internal. Verein.limnol. 24: 1-4. many Ihanks. Jordan. C. F. 1985. Nutrient Cycling in Tropical Forest RESUMEN Ecosystems. Principles andTheir Application in Man· agement and Conservation. Wiley. New York. EI presentc estudio fue realizado en los rlos Kamphake. L J.. S. A. Ilannah &. J. M. Cohen. 1967. y quebradas de la zona eostern sur-caribel'la de Automated analysis for nitrate by hydrazine reduction. Costa Rica. invtsligandose la comunidad de Wat. Res.l: 205·216. macroinvertebrados benlicos con los objetivos Lowe.McConnel. R. 1987. Ecological Studies in Tropical de delerminar el estado de la estructura de la fish Communities. Cambridge University. Cambridge. comunidad. la composici6n a lraves del liem 382 p. po. el efecto de la cobertura de vegetaci6n en lowe-McConnel. R.• C. M. Crul &. F. C. Roes!. 1992. el Mea de drenaje y la comparaci6n de dos R. Symposium on Resource Usc and Conservation of the metodos de muestreo; el "surber" y un cola African Great Lakes, Bujumbura. 1989. Mill. Imerna\. dor. EI estudio fue realizado entre setiembre Verein. Limno1. 23. 128 p. de 1995 y mayo de 1996. Cinco sitios predo minantemenle con una cuenca forestada Lowe-McConncl1, R. 1994. Threats to, and conservation of. tropical fresllwater fishes. Mill. Internal. Vcrein. muestran una abundancia (promedio" 157.60 LimnoL 24: 47·n. Im1) y riqueza de taxa (promedio=9.65/m1) significativamenle mayor que la encontrada en MargalelT. R. 1983. Limnologfa. Omega. Bartclona, Esplfl•. 1010 p. los 511105 con una cuenca perturbada (promedio - 82.08 y 4.62 respectivamente). Middleton. R. 1995. EI AgWl Limpia: Un Recurso Frigil. Todos los sitios muestran un indice de helero Documc:nto Verde If 4. Scrvicio Cultural e Infonna geneidad alto (> 0.80), el cual junto con los tivo de los Estldos Unidos de Am6ica. Washington, bajos valores de similitud « 0.40) entre D.C.80p. epocas de muestreo indican que la comunidad Paaby. P. & C. R. Goldman. 1992. Chorophyl1. priml\tY de macroinvertebrados se encuentra sujeta a productivity. and respiration in a lowland Costa Rican eventos de perturbaci6n frecuentes que dis stream. Rev. BioI.Trop. 40: 185-198. minuyen la pennanencia de especies en el sistema. EI metodo de muestreo utilizando el Ramirez. A.. P. Puby. C. M. Pringle &. G. AgOero. 1998. ElTect of habitl1 type on benthic maeroillvenebrites in colndor es significativamenle mas efectivo en two lowland tropi(;al streams. Costa Rica. Rey. Bioi. muestrear la comunidad de macroinverte Trop. 46. Sup!. 6: 185·199. brados (promedio = 165.lO/m� que el Surber (promedio - 8J.OJ/m2. Ambos metodos de Reice, S. R. 1994. Noncquilibrium dctenninanlS of bio muestreo. Surber y colador. son igualmente logical community strudure.Am. Sci.. 82: 424-4]5. erectivos en el numero capturado de taxa, Resh. V. II.. M. J. Myers&. M. J. Hannaford. 1996. Mae· aunque la similitud en la composici6n de roillvertebratcs IS biotic indicators of environmental captura oscila entre 0.1379 y 0.7619). quality. p. 647-667. /" R. H.lICr &. G. A. Lamberti (Eds.). Methods in Stream Ecology. Academic. San Oie80. California. Aquatic ecosystems ofCosta Rica 199 Sanford, R, L, P. Paaby, J. C. Luvall &. E. Phillips. 1994. production in one nIile of stream. Trans. Am. Fisheries Climate. geolnOl"phology and lIquatic ecosystems. p. Soc. 66: 193-101. 19-33. In L. A. McDade, K. S. Bawl, H. A. Hespen heide &. G. S. Hartshorn (Eds.). La Selva: Ecology Ward, J. V. & J. A. Stunford 1983. The intemlediPle and Naturo[ HislOry of 0 NeOlropical Rain Forest. The disturbance hypothesis: An e,xplamion for biotic di· University of Chicago, Chicago. Illinois. versity pallcrns in lotic ecosystcms. p. 347-356. In T. D. Fontaine &. S. M llar1ell (Eds.). Dynamics of Lotic Sol6rzano. L 1969. Detennination of ammonia in natural Ecosystems. Ann, Arbor Science. Ann Arbor. Michi waters by the phenolhypochlorite method, Limnol. gw. Oceanogr. 14: 799-80 1. Wetzel. R. G. Y E. Likens. 1979. Limnological Analyses. Surber, E. W. 1937. Rainbow trout and bottom fauna W. D. Saundcrs. l'ttiiadclphia, I'cnnsylvannia.