Hydrobiologia 458: 141–149, 2001. © 2001 Kluwer Academic Publishers. Printed in the Netherlands.

141

Litter input from riparian vegetation to streams: a case study of the Njoro River, Kenya Adiel E. M. Magana Department of Zoology, Egerton University, P.O. Box 536, Njoro, Kenya Tel: 254-37-61284, Ext. 3628. Fax: +254-37-61213. E-mail: [email protected] Key words: organic matter, CPOM, tropical rivers, riparian vegetation, Lake Nakuru

Abstract Allochthonous coarse particulate organic matter (CPOM) input into the Njoro River was measured between January and June 1998 at two contrasting sites: open-canopy and closed-canopy sites. Bank runoff and aerial drift traps were used for collecting CPOM inputs over periods of two weeks. Collected litter was sorted into four categories: leaves, fruits, wood and plant fragments. Monthly input ranged from 77 to 228 g ash free dry weight m−1 for bank runoff input and from 64 to 129 g ash free dry weight m−2 for aerial input. The highest input of 228 g ash free dry weight m−1 was recorded in May at the closed-canopy site. Wood, fruits and plant fragments of particle size >100 µm contributed a mean ± SE of 60±9% of the total inputs with the rest from leaf litter. The closed-canopy site had higher inputs (P<0.05) of bank and aerial input than the open canopy site. There was no relationship between total bank runoff input and rainfall (rs = 0.08), however, total aerial input increased with decrease in rainfall (rs = – 0.59). There were differences between inputs from different plant species (P<0.05) that ranked in the following order: Syzygium cordatum > Rhus natalensis > Pittosporum viridiflorum > Vangueria madagascariensis. Removal of riparian vegetation from the banks of the Njoro River would alter the quantity and quality of the litter and reduce CPOM inputs to the river and to a downstream lake with attendant consequences to the energy budget of biocoenoses in the two ecosystems. Introduction The importance of terrestrial leaf litter as the major source of energy in small woodland low order streams has been stressed by many authors (Cummins, 1974; Anderson & Sedell, 1979; Vannote et al., 1980, Iversen et al., 1982; Bretschko, 1990; Bretschko & Moser, 1993). Inputs of organic matter in form of fragments of riparian vegetation in rivers and streams form an important link between terrestrial and aquatic productivity. Organic matter enters the stream by aerial drift or throughfall, bank runoff and subsurface pathways (Dudgeon, 1992; Dudgeon & Bretschko, 1996). Events in the riparian ecosystem largely determine the quantity, quality, timing and retention of allochthonous organic matter received by streams. Moreover, the riparian vegetation is the source of coarse particulate organic matter (CPOM) and the leaf phenology of the vegetation determines the amount of CPOM input to stream in time and space (Conners & Naiman, 1984;

Bretschko, 1990). Furthermore, climate seasonality and land–water interactions to a large extent influence allochthonous organic matter inputs in streams of a given region (Dudgeon & Bretschko, 1996). Particulate allochthonous inputs may contribute 81–92% of the organic carbon supplied to first and second order streams (Conners & Naiman, 1984). Formal study of tropical streams and rivers has a long history beginning with early explorers and naturalists who visited these regions and collected aquatic specimens (Jackson & Sweeney, 1995). However, most long term studies on inputs and retention of organic matter in streams and rivers have been done in temperate regions. Reports on tropical streams and rivers are rare (Mathooko, 1995). Comprehensive studies of tropical lotic ecosystems are restricted to Amazonia (Jackson & Sweeney, 1995) and tropical Asia, particularly the Hong Kong rivers (Dudgeon, 1992).

142

Figure 1. Location of the Njoro River, Kenya, showing the study sites.

Study on the dynamics of CPOM in rivers and streams of Kenya is at its infancy. The two sections of the Njoro River, Nakuru District, which were the focus of the present study have a belt of natural riparian vegetation approximately 20 m wide on either side of the river channel. Mathooko (1995) observed that although this riparian vegetation does not exhibit a distinct phenology or defoliation timing, the CPOM retained in the dry-store and wet-store zones on Njoro River fluctuated in time and space. The dry-store zones retain a larger amount of CPOM and they serve as storage or pre-processing areas for allochthonous material before its ultimate input into stream. The Njoro River experiences at least two periods of high discharge per year (Karanja et al., 1986) with a profound effect on CPOM standing stock on both dry-store and wet-store zones as the flood carries away the retained CPOM. This study assumes that various meteorological factors as well as density and composition of riparian vegetation influences the inputs of particulate organic matter into the Njoro River in time and space. The objectives of the present study were to determine: (i) The spatio-temporal pattern of plant litter inputs to the Njoro River. (ii) The influence of factors such as wind, bank slope, rainfall, density and composition of riparian vegetation on plant litter inputs to the Njoro River.

The Njoro River and the study sites The study sites were situated on two reaches of the

Njoro River within the Egerton University campus (2240 m.a.s.l.) about 26 km SW of Nakuru Town. The Njoro River (Fig. 1) in Nakuru District, Rift Valley Province, Kenya is a low gradient river originating from the Eastern Mau Hills (2700 m.a.s.l.). The catchment area (0◦ 15 S, 0◦ 25 S; 35◦ 50 E, 36◦ 05 E) is 200 km2 . The length of the river is about 50 km and it discharges into Lake Nakuru (1700 m.a.s.l.), a soda lake in the Rift Valley. Its main tributary is the Little Shuru. The soils are mainly volcanic clay loam except near the Lake, were silt clay is predominant. Average annual rainfall is 1200 mm, distributed trimodally with peaks in April, August and November (Karanja et al., 1986). The River itself is structured in a typical poolriffle sequence with predominantly soft substrata in pool area and bedrock in the rifle sections (Bretschko, 1995). The upper study site designated ‘Open Canopy Site’ is devoid of canopy over the river channel. However, vegetation is dense on both sides of the river channel with a 60% canopy cover. The most dominant riparian vegetation is comprised of Syzygium cordatum (Myrtaceae), Rhus natalensis (Anacardiaceae), Pittosporum viridiflorum (Pittosporaceae) and Dovyalis caffra (Flacourtiaceae). The lower study site designated ‘Closed Canopy Site’ is about 1 km below of the upper study site. It has a shaded (closed) canopy covering 90% of the channel area, comprising Syzygium cordatum (Myrtaceae), Acacia abyssinica (Mimosaceae), Rhus natalensis (Anacardiaceae), Pittosporum viridiflorum (Pittosporaceae) and Vangueria

143

Figure 2. Climatic variables measured at the study sites at the Njoro River, Kenya, from January to June 1998 (a) rainfall and temperatures (b) humidity and wind speed.

madagascariensis (Rubiaceae). Both stream sections are approximately 20 m wide and 40 m long. The width of the river channel ranged from 4 m to 7 m.

Materials and methods During the study, the quantity and composition of the bank runoff (lateral) and aerial (throughfall) inputs of organic matter from the riparian vegetation of the River Njoro was determined at the closed and open canopy sites. As possible influencing factors, the temperature, wind speed, rainfall and relative humidity were measured daily for entire 6-month duration of the study. The bank runoff CPOM inputs to Njoro River were measured at the closed and open canopy sites from January to June 1998 using specially designed bank runoff traps. The traps consisted of a steel plate (length 50 cm and width 50 cm) placed on the bank. The steel

Figure 3. Plant litter input from bankside runoff into the Njoro River, Kenya, at the closed canopy from January to June 1998 (a) Syzygium cordatum and Pittosporum viridiflorum (b) Rhus natalensis and Vangueria madagascariensis (c) fruits, wood and plant fragments.

plate led material rolling down the slope into the horizontal slot of a wooden box (length 50 cm, width 30 cm, height 10 cm) with its lower side covered by a nylon net (mesh size 100 µm) to allow water to pass through. Five traps were distributed randomly at each study site.

144 Aerial CPOM inputs were also measured over the same period at both sampling sites. Five coneshaped traps made of a fine nylon material (mesh size 500 µm) with a diameter of 50 cm and 60 cm high were positioned along a transect across the river, above the water surface with their openings directed upwards at each site to collect the aerial drifting litter. Both bank runoff and aerial traps were emptied every two weeks. Samples were oven-dried at 60 ◦ C for at least 48 h to a constant weight, then sorted into leaves, fruits, wood and unidentifiable plant fragments. The leaf component was separated further into different plant species and weighed to the nearest 0.001 g. Sub samples were ashed at 500 ◦ C in a muffle furnace for 4 h, re-weighed and their ash free dry weight (AFDW) calculated (Benfield, 1996). Aerial and bank runoff inputs were expressed as g AFDW m−2 and g AFDW m−1 , respectively (Moser, 1991; Weigelhofer & Waringer, 1994). Kruskal–Wallis one-way ANOVA by ranks test, Spearman’s rank correlation coefficient (rs ) and Ttests (Zar, 1996) were used to determine the significance of any differences between data sets.

Results Table 1 summarises the total bank runoff input recorded from January to June 1998 at both the closed and the open canopy sites. Total bank runoff inputs were higher at the closed canopy site than at the open canopy site (P<0.05). Monthly inputs mean ± SE was 175±22 and 86±19 g AFDW m−1 at the closed and open canopy sites, respectively. There was no distinct temporal pattern of bank runoff input of the various components of plant litter at the closed canopy site. The endemic riparian plant species such as Syzygium cordatum and Pittosporum viridiflorum had a continuous litter input which increased slightly in March and January, respectively (Figs 2a, b and 3a). Furthermore, their inputs were positively wind speed and negatively correlated to in rainfall (Table 2). There was a general pattern of decrease of inputs from terrestrial species growing in the riparian zone e.g. Rhus natalensis and Vangueria madagascariensis with cessation of inputs in June and April respectively (Fig. 3b). Fruits and wood (Fig. 3c) had an initial decrease of input in February followed by a gradual increase reaching peak inputs in May. Wood input showed a positive correlation with rain-

Figure 4. Plant litter input from bankside runoff into the Njoro River, Kenya, at the open canopy site from January to June 1998 (a) Syzygium cordatum and Pittosporum viridiflorum (b) Dovyalis caffra and Rhus natalensis (c) fruits, wood and plant fragments.

fall (rs = 0.44, P > 0.05), and a negative correlation with wind speed (P < 0.01). Input of plant fragments ranged from 24 to 74 g AFDW m−1 , with peak input in April (Fig. 3c). Except for S. cordatum and P. viridiflorum at the open canopy site (Fig. 4a) which showed a similar pattern of input to that of the same species at the closed

145 Table 1. Total particulate organic matter bank runoff input to the Njoro River, (g ash free dry weight (AFDW) m−1 ) January to June 1998. Mean ± SE of 5 replicate bank runoff traps Closed canopy site (g AFDW m-1 ) SE

Species Syzygium cordatum Rhus natalensis Pittosporum viridiflorum Vangueria madagascariensis Dovyalis caffra Fruits wood Large plant fragments (>1 mm) Small plant fragments (>1 mm) Total

195 14 8 4

3.5 0.5 0.1 0.3

106 274 327 78 1006

7.9 13 7.2 2.88

Open canopy site (g AFDW m −1 ) SE 222 21 11

7.3 1.6 0.6

20 68 96 95 13 546

1.0 3.0 2.6 3.8 0.9

Table 2. Correlation of particulate organic matter bank runoff input with climatic variables measured at the Njoro River, Kenya, from January to June 1998. Spearman’s rank correlation coefficient was used for determining the correlations. Significance levels indicated as ∗ P < 0.05 and ∗∗ P < 0.01 Species

Syzygium cordatum Rhus natalensis Pittosporum viridiflorum Vangueria madagascariensis Dovyalis caffra Fruits wood Large plant fragments (>1 mm) Small plant fragments (<1 mm)

Closed canopy site RainTemp. Humid fall -ity

Wind speed

−0.71 −0.20 −0.30 −0.33

0.74 −0.29 0.09 −0.03

−0.81∗ −0.03 −0.17 −0.49

0.58 0.64 0.35 0.83∗

0.43 0.44 −0.03 −0.71

−0.71 −0.37 0.54 0.37

0.99∗∗ 0.84∗ −0.38 −0.46

−0.93∗∗ −0.99∗∗ −0.06 −0.17

canopy site described above, other fractions had no consistent pattern of input (Fig. 4 b, c). The leaf litter comprised 35.5% of the total bank input at the two sites. Table 3 shows that input by aerial drift (throughfall) of 576 g AFDW m−2 recorded at the closed canopy site was twice that in the open canopy site. Monthly aerial inputs ranged from 8 to 75 and 49 to 129 g AFDW m−2 at the open and closed canopy sites, respectively. Leaves were the most dominant component of aerial input comprising 40.0% of the total aerial input, mostly originating from the adjacent riparian vegetation. All components of the aerial CPOM input at the two sites showed a more or less similar temporal pattern with the highest values recorded in March and

Open canopy site Rain Temp. Humid -fall ity −0.03 0.23 −0.66

0.89 0.64 0.49

0.60 −0.37 −0.77 0.31 0.38 −0.77 −0.14 0.31 −0.43 −0.03

Wind speed

−0.72 −0.72 −0.58

0.35 0.62 0.46

0.55 −0.81∗ 0.84∗ −0.06 0.23

−0.32 0.93∗∗ −0.064 −0.38 −0.46

April (Figs 5 a, b and 6 a, b), except for the plant fragments, whose input increased considerably in June at the closed canopy site (Fig. 5b). The relationship between aerial input and the prevailing weather conditions varied with components and site. Peak leaf litter inputs were, however, associated with relatively high wind speed and low rainfall while peak wood inputs were associated with low wind speed and high rainfall (Table 4) The total CPOM inputs (bank runoff and aerial) from January to June 1998 at the closed and open canopy sites ranged between 804 and 1581 g AFDW m−2 . Wood, fruits and plant fragments of the particle size >100 µm contributed a mean ± SE of 60 ± 9% of the total inputs, the rest was from leaf litter. There were differences of CPOM inputs

146 Table 3. Particulate organic matter aerial input to the Njoro River, (g ash free dry weight (AFDW) m−2 ) from January to June 1998. Mean ± SE of 5 replicate throughfall traps Species

Closed canopy site (g AFDW m 2 ) SE

Syzygium cordatum Rhus natalensis Mytenus senegalensis Fruits wood Large plant fragments (>1 mm) Small plant fragments (>1 mm)

144 13 2.8 36 97 193 90

Total

576

Open canopy site (g AFDW m 2 ) SE

4.7 0.7 0.2 1.4 4.1 5.7 2.9

118 2

7.1 0.2

91 0.5 41 3

5.7 0.1 1.2 0.1

256

Table 4. Correlation of particulate organic matter aerial input with climatic variables measured at the Njoro River, Kenya, from January to June 1998. Spearman’s rank correlation coefficient was used for determining the correlation. Significance levels indicated as ∗ P < 0.05 and ∗∗ P < 0.01 Species

Syzygium cordatum Rhus natalensis Mytenus senegalensis Fruits wood Large plant fragments (>1 mm) Small plant fragments (<1 mm)

Closed canopy site RainTemp. Humid fall -ity −0.54 −0.20 0.38 0.14 0.09 −0.26 −0.26

0.77 0.89∗ 0.38 0.94∗∗ −0.26 0.37 0.37

−0.99∗∗ −0.81 0.62 −0.64 0.20 −0.29 −2.9

by plant species (Kruskal–Wallis one-way ANOVA by ranks test: P<0.05) ranked in the following order: Syzygium cordatum > Rhus natalensis > Pittosporum viridiflorum > Vangueria madagascariensis.

Discussion The world average litter fall of temperate forest is about 350 g m−2 , mean annual litter fall to upland forest streams of the Eastern U.S.A. is 414 g m−2 . Southern Australia Eucalyptus forest streams have annual litter fall values ranging between 250 and 1000 g m−2 (Campbell & Fuchshuber, 1994). The total amount of litter inputs recorded for six months at the Njoro River in the present study ranged from 804 to 1581 g AFDW m−2 , which is higher than litter fall reported from elsewhere. This is consistent with observations made by Bray & Gorham (1964) that litter fall

Wind speed 0.81 0.70 −0.82 0.29 −0.32 −0.11 −0.12

Open canopy site RainTemp. Humid fall -ity

Wind speed

−0.03 −0.04

0.71 −0.03

−0.81∗ −0.19

0.64 0.57

−0.46 −0.39 0.09 −0.03

0.52 −0.38 0.14 0.54

−0.87∗ 0.27 0.23 −0.38

0.97∗∗ −0.40 −0.64 −0.60

tends to increase with increase in rainfall, but decrease with higher latitude and declining temperature. The amount of rainfall recorded at the study sites during the present study was abnormally high due to El Niño weather and temperatures ranged from 18.8 to 21.3 ◦ C advantaging leafing and vegetative growth of riparian plants and so contributing large inputs to the river. In the present study, no distinct temporal pattern of leaf litter input was observed. However, it was noted that there was a slight increase in both the throughfall and bank runoff leaf litter inputs during a relatively dry spell between February and April. Since temperature was conducive for plant growth throughout the study duration, shortage of rainfall could have limited biological processes thereby facilitating abscission of leaves and flowers. The highest bank runoff CPOM input was dominated by fruits of S. cordatum and woody material of Acacia abyssinica and occurred during the wettest

147

Figure 5. Plant litter input from aerial drift (throughfall) into the Njoro River, Kenya, at the closed canopy site from January to June 1998. (a) Syzygium cordatum and Rhus natalensis. (b) Fruits, wood and plant fragments.

Figure 6. Plant litter input from aerial drift (throughfall) into the Njoro River, Kenya, at the open canopy site from January to June 1998. (a) Syzygium cordatum and Rhus natalensis. (b) Fruits, wood and plant fragments.

period, i.e. May and June at the closed-canopy site. It is presumed that rainfall enhanced rotting of wood, maturation and ripening of fruits, hence shedding of these components. However, it is estimated that only a small percentage ranging between 30 and 50% of the plant litter on the banks enters the streams depending on topography and seasonality of the area under investigation (Britton, 1990; Benfield, 1997). Although wood contributed a large biomass of particulate organic matter into the stream only a few species of aquatic organisms use wood directly. The refractory nature of wood renders it a much less important source of food than leaf litter (Anderson & Cummins, 1979; Dudgeon, 1992). Similar studies in some tropical Asian streams of Hong Kong show rather small temporal variations in the amount of inputs to streams due to lack of pronounced seasonal variation and no consistent pattern of defoliation and leafing. However, minor increases in the standing stocks of CPOM from December to February reflecting litter inputs by deciduous plants

such as Liquidambar formorsana have been reported in some parts of tropical Asia which have distinct dry and wet seasonality (Dudgeon, 1992). In contrast to the present study, studies on inputs of CPOM in Rold Kindle in Denmark (Iversen et al., 1982) & Oberer Seebach in Austria (Moser, 1991; Bretschko & Moser, 1993) both low order temperate streams show a distinct temporal pattern with peak inputs in autumn. Whilst climatic seasonality determines timing of biological events such as synchronised leaf fall and litter input to temperate streams, it seems that the dry / wet seasons were the main determinants of timing events like litter fall in the present study. Studies done in Tai Po Kau tropical forest streams in Hong Kong, show that there is a close and strong correlation between input of organic matter and both prevailing temperature and rainfall. Strong winds such as typhoons are also known to have an important effect on the magnitude and timing of litter production (Dudgeon, 1992). The amount of input recorded at the closed canopy site was nearly twice that of the open canopy. A

148 similar spatial pattern has been reported by many authors (Gurtz et al., 1988; Moser, 1991; Weigelhofer & Waringer, 1994). Since the bank slope at the two sites was similar (24.6◦ and 23.3◦ at the closed and open canopy site, respectively), it is therefore presumed that the density and composition of riparian vegetation was crucial in determining the quantity of inputs at the two sites. The litter collected in the bank runoff and aerial traps came from vegetation within the immediate vicinity. This implies that wind had a limited and local effect on transport of plant litter. It was noted that the dominance in canopy cover by a particular plant species did not necessarily result in a corresponding change in amount of input. For example, Pittosporum viridiflorum was more dominant than Rhus natalensis at the open canopy site but it contributed less of leaf litter than the R. natalensis, probably because of its low turnover in leafing and defoliation. The CPOM inputs varied from species to species of the riparian plants. Inputs were dominated by native or endemic species such as Syzygium cordatum whose litter accounted for 32.5% of the total input. Contrary to Stout (1980), not all riparian plants in tropical woodland streams release their litter continuously into streams. In the present study, while endemic riparian plants such as Syzygium cordatum and Pittosporum viridiflorum, shed their litter continuously with minor variations over a period of time, terrestrial plants growing in the riparian zone, e.g. Vangueria madagascariensis and Rhus natalensis shed all their litter at a particular time. This is of great significance to stream zoocoenoses such as shredders which use litter from riparian vegetation for food. This, therefore, implies that if endemic plant species are replaced with ‘exotic’ species in the riparian zone, litter input could become intermittent and ecological dependants at the site and further downstream could be disadvantaged. In conclusion, it is emphasised that fragments of riparian vegetation entering rivers and streams form an important link between terrestrial and aquatic ecosystems. It is, therefore, desirable to maintain buffer strips of native riparian vegetation to maintain quantity and timing of allochthonous inputs (Dudgeon & Bretschko, 1996; Pozo et al., 1997). The findings of the present study provide useful information on the dynamics of litter fall and the role of local riparian vegetation in allochthonous inputs to the Njoro River. The study also highlights the allochthonous input contributions of the Njoro River into Lake Nakuru because the allochthonous detritus, which is not pro-

cessed in the river channel is carried to the lake, and hence is important to the energy budget of the lake.

Acknowledgements I thank Prof. Dr Gernot Bretschko (Biologische Station Lunz, Austria) for advice during this work, Dr J. M. Mathooko (Department of Zoology, Egerton University) who read and commented on various aspects of this manuscript, Mr S. T. Kariuki (Department of Botany, Egerton University) for assisting in identification of riparian plants, Mr C. T. Meme for assisting in collection and processing of samples and the Egerton University for providing working space. I appreciate useful comments by an anonymous referee. This study was supported financially by ÖAD (Austria) through the North–South Dialogue Programme (EH-Project 894/97).

References Anderson, N. H. & K. W. Cummins, 1979. Influence of diet on the life histories of aquatic insects. J. Fish Res. Bd Can. 36: 335–342. Anderson, N. H. & J. R. Sedell, 1979. Detritus processing by macroinvertebrates in stream ecosystems. Ann. Rev. Entomol. 24: 351–377. Benfield, E. F., 1996. Leaf breakdown in stream ecosystem. In: Haver, R. & G. A. Lambert (eds), Methods in Stream Ecology, 579–589. Benfield, E. F., 1997. Comparison of litter fall into streams. In: Webster, J. R. & J. L. Meyer (eds), Stream Organic Matter Budgets. J. n. am. Benthol. Soc. 16: 3–161. Bray, J. R. & E. Gorham, 1964. Litter production in forest of world. Adv Ecol Res 2: 101–157. Bretschko, G., 1990. The dynamic aspect of coarse particulate organic matter (CPOM) on the sediment surface of a second order stream free of debris dams (Ritrodat–Lunz study area). Hydrobiologia 203: 15–28. Bretschko, G., 1995. Tropical River Ecology Initiative. Proceedings of the first workshop, Egerton University, Njoro, Kenya: 92 pp. Bretschko, G. & H. Moser, 1993. Transport and retention of organic matter in riparian ecotones. Hydrobiologia 251: 95–101. Britton, D. L., 1990. Fire and dynamic of allochthonous detritus in a South African mountain stream. Freshwat. Biol. 24: 347–360. Campbell, I. C. & L. Fuchshuber, 1994. Amount, composition and seasonality of terrestrial litter accession to an Australian cool temperate stream. Arch. Hydrobiol. 130: 499–512. Conners, E. M. & J. R. Naiman, 1984. Particulate allochthonous inputs: Relationship with stream size in an undisturbed watershed. Can. J. Fish aquat. Sci. 41: 1473–1484. Cummins, K. W., 1974. Structure and function of stream ecosystems. Bio. Science 24: 631–641. Dudgeon, D., 1992. Patterns and processes in stream ecology: A synoptic review of Hong Kong running waters. Schweizerbart’sche Verlagsbuchhandlung, Stuttgart: 147 pp. Dudgeon, D. & G. Bretschko, 1996. Allochthonous inputs and land–water interactions in seasonal streams: tropical Asia and

149 temperate Europe. In Schiemer F. & K. T. Boland (eds), Perspectives in Tropical Limnology. SPB Academic Publishing, Amsterdam, The Netherlands: 161–179. Gurtz, M. E., G. R. Marzolf, K. T. Killingbeck, D. L. Smith & J. V. Mcarthur, 1988. Hydrologic and riparian influences on the import and storage of coarse particulate organic matter in a Prairie stream. Can. J. Fish. aquat. Sci. 45: 655–665. Iversen, M. T., J. Thorup & J. Skriver, 1982. Inputs and transformation of allochthonous particulate organic matter in a headwater stream. Holarct. Ecol. 5: 10–19. Jackson, K. J. & W. B. Sweeney, 1995. Present status and future directions of tropical stream research. J. n. am. Benthol. Soc. 14: 5–11. Karanja, A. K., S. S. China & P. Kundu, 1986. The influence of land use on Njoro River catchment between 1975 and 1985. In Proceeding of the Third National workshop on soil and water conservation in Kenya, Nairobi. Mathooko, J. M., 1995. Retention of plant coarse particulate organic

matter (CPOM) at the surface of the wet-store and dry-store zones of the Njoro River, Kenya. Afr. J. Ecol. 33: 151–159. Moser, H., 1991. Input of Organic Matter (OM) in a low order stream (Ritrodat–Lunz study area, Austria). Verh. int. Ver. Limnol. 24: 1913–1916. Pozo, J., E. Gonzalez, J. R. Diez, J. Molinero & A. Elosegui, 1997. Input of particulate organic matter to streams with different riparian vegetation. J. n. am. Benthol. Soc. 16: 602–611. Stout, J., 1980. Leaf decomposition in Costa Rica lowland tropical rain forest. Biotropica 12: 264–272. Vannote R. L., G. W. Minshall, K. W. Cummins, J. R. Sedell & C. E. Cushing, 1980. The river continuum concept. Can. J. Fish. aquat. Sci 37: 130–137. Weigelhofer, G. & J. A. Waringer, 1994. Allochthonous input of coarse particulate organic matter (CPOM) in a first to fourth order Austrian forest stream. Int. Rev ges. Hydrobiol. 79: 461–471. Zar, J. H., 1996. Biostastistical Analysis, Prentice-Hall Inc. New Jersey: 662 pp.