History and Timing of Human Impact on Lake Victoria, East Africa Author(s): Dirk Verschuren, Thomas C.

Johnson, Hedy J. Kling, David N. Edgington, Peter R. Leavitt, Erik T. Brown, Michael R. Talbot and Robert E. Hecky Reviewed work(s): Source: Proceedings: Biological Sciences, Vol. 269, No. 1488 (Feb. 7, 2002), pp. 289-294 Published by: The Royal Society Stable URL: http://www.jstor.org/stable/3067918 . Accessed: 06/11/2012 20:20
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at . http://www.jstor.org/page/info/about/policies/terms.jsp

.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.

The Royal Society is collaborating with JSTOR to digitize, preserve and extend access to Proceedings: Biological Sciences.

http://www.jstor.org

THE ROYAL am SOCIETY

B4 -

Received 2 April2001 Accepted 5 September 2001 Published online 22 January 2002

Histowy and timing of human impact on Lake Victoria, East Africa

Dirk Verschurenl 2*, Thomas C. Johnsonl*, HedyJ. Kling3, David N. Edgengton4, PeterR. Leavitt5, ErikT. Brownl, MichaelR. Talbot6 and RobertE. Hecky7
lLargeLakes Observatory, University ofMinnesota, 10 University Drive,Duluth, MN 55812, USA 2Department ofBiologNy, Ghent University, Ledeganckstraat 35, B-9000 Gent, Belgium 3Department ofFisheries and Oceans,501 University Crescent, Winnipeg, Manitoba, Canada R3T 2N6 4Center forGreat LakesStudies, University of Wisconsin, Milwaukee, WI 53201, USA 5LimnologNy Laboratory, Department ofBiologNy, University ofRegina, Regina, Saskatchewan, Canada S4S OA2 6Geological Institute, University ofBergen, Allegt. 41, N-5007 Bergen, Nortway 7Department ofBiology, University of Waterloo, 200 University AvenueWest, Waterloo, Ontario, Canada N2L 3G1 Lake Victoria, thelargest tropical lake in theworld, suffers from severeeutrophication and theprobable extinction ofup to halfofits500+ speciesofendemiccichlid fishes. The continuing degradation ofLake Victoria's ecologicalfunctions has seriouslong-term consequences fortheecosystem services it provides, and may threaten social welfare in the countries bordering its shores.Evaluationof recentecological changesin the contextof aquatic food-web alterations, catchment disturbance and naturalecosystem variability has been hamperedby the scarcity of historical monitoring data. Here, we presenthighresolution palaeolimnological data, whichshow thatincreasesin phytoplankton production developed from the 1930s onwards, whichparallels human-population growth and agricultural activity in theLake Victoria drainage basin.Dominanceofbloom-forming cyanobacteria sincethelate 1980s coincided with a relative declinein diatomgrowth, whichcan be attributed to theseasonaldepletion of dissolved silica resulting from 50 years ofenhanced diatomgrowth and burial.Eutrophication-induced loss ofdeep-water oxygenstarted in the early1960s, and mayhave contributed to the 1980s collapse of indigenous fish stocksby eliminating suitable habitat forcertain deep-water cichlids. Conservation of Lake Victoria as a functioning ecosystem is contingent upon large-scale implementation of improved land-usepractices. Keywords:landscapedisturbance; eutrophication; fishintroduction; humanimpact;Lake Victoria; Nile perch 1. INTRODUCTION Lake Victoria has in recent decadesundergone a seriesof profound ecological changes, including strong increases in phytoplankton primary production(Hecky & Bugenyi 1992; Mugidde 1993), replacement ofdiatoms bycyanobacteria as thedominant groupofplanktonic algae (Kling et al. 2001), and theeradication ofseveral hundred species of endemic cichlid fishes afterthe 1980s population explosion ofNile perch an introduced piscivore (Barelet al. 1985; Ogutu-Ohwayo 1990; Witteet al. 1992). The observed changesin the phytoplankton community seem to be the logicalconsequenceof excessnutrient loading thatresults from deforestation and greatly intensified agriculturethroughout the drainagebasin of Lake Victoria, which comprisesportionsof five countries(Hecky & Bugenyi1992; Scherenet al. 2000). However,lack of long-term monitoring data and thecomplexity ofongoing ecosystem changes hamper fullappreciation oftheimpact of catchment disturbance on thelake'spresent condition. Apparent coincidencein the timing of the first massive cyanobacteria blooms offshore(Ochumba & Kibaara 1989) with the collapse of indigenousfish stockshas
*Authors for correspondence (dirk.verschuren(rug.ac.be, tcj(d.umn.edu) . Proc. R. Soc.Lond. B (2002) 269,289-294 DOI 10.1 098/rspb.200 1.1850

tempted thequestion: to whatextent does thedecimation of phytoplanktivorous haplochromine cichlids by Nile perch contribute to increased algal production? (Goldschmidt et al. 1993). Also, invigorated algal growth may have resulted in partfromthe documented rise in average surface-water temperature and water-column stability in the 1960s,a possibleregional manifestation of globalclimate warming (Heckyet al. 1994). In the absence of adequate historical monitoring data thatlinkobserved ecologicalchangeto possiblecausative factors, we used the palaeolimnological recordpreserved in offshore sediments to reconstruct the successionand timingof these changes,and to elucidatethe chain of cause and effect thatled to today'sconditions of severe eutrophication and restructured algal and fishcommunities.We describe the environmental history of Lake Victoriaoverthepast 180 years as recorded in coreV96-5MC from the deepestpartof the lake,whererapidsediment accumulation and the absenceofburrowing invertebrates combine to producea high-quality sediment archive. This core is one of six intactcoresrecovered to date from the mainoffshore depositional basinof Lake Victoria (Hecky 1993; Verschuren et al. 1998; figure 1), and representative ofthree records from 67-68 m waterdepth.

Z 289

2002 The Royal Society

EastAfrica on Lake Victoria, impact and others HistoCofhuman 290 D. Verschuren

Kenya, ofUganda, portions basincomprises The 263 000 km2 stations. basinand coring itsdrainage 1. LakeVictoria, Figure densities population Nations1995).Highest (United of27.7 million a 1995population with and Burundi, Rwanda Tanzania, ca. 90% of contribute together that rivers and Burundi Rwandan ofKenyan, occurin thedrainages activity and agricultural records ofsediment quality therelative reflect symbols 1988). Core-site & Bugenyi (Balirwa to LakeVictoria input total river open bioturbation; symbols, half-open sedimentation; undisturbed Closedsymbols, dynamics. bylocalbottom as determined at prepared bathymetry LakeVictoria arein metres. Depthcontours waveturbulence. and frequent bothbioturbation symbols, areas)and British offshore in 1995and 1996 (mainly echosounding basedon IDEAL project theLargeLakesObservatory, Beck(1998). datafrom Crul(1993); and population from drainages areas);river maps(nearshore Admiralty

AND METHODS 2. MATERIAL

bytheIDEAL was conducted that on Lake Victoria Fieldwork Lakes) Decade forthe East African (International programme sixoffcoresfrom surface-sediment in 1995 and 1996 recovered et 48 and 68 m waterdepth(Verschuren between shorestations V96-5MC, werecollected al. 1998). The 1996 cores,including (samplingarea 4x71 cm2), multi-corer witha Hedrick-Marrs after penetration dampenedsediment hydraulically thatprovides has come to reston thelake floor.Of thefour frame a support in the at each site,twowereextruded recovered core replicates near the top (0-10 cm), and 1 cm fieldin 0.5 cm increments was used forporeThe thirdreplicate down-core. increments the at 2 cm (0-10 cm) and 4 cm intervals; water extraction was capped and storedintact. replicate fourth ml-' wetmud) were and drymass (dryweight Watercontent age at depth for20 h at 105 C. Sediment bydrying detellllined and accumulationrates were determinedby 2'0Pb dating 2'0Pb (the fraction 1975). Unsupported (Robbins& Edgington by ill SitU decay of 226Ra)was not supported of 2'0Pb activity total2'0Pb and itsconstant between calculatedas thedifference
Proc.R. Soc. Lod. B (2002)

value of 1.74pCig-' below a core depth of 37cm (n=3). (Robbins regression to a nonlinear 2'0Pbwas fitted Unsupported at a core depth intoaccounta disconformity et al. 1978), taking of 25-26 cm (see 3). Solid-phasebiogenicsilica (Si) in the (DeMaster was analysedby wet alkalineextraction sediments solutionat 80 C for 1981), using a 0.5 m sodium hydroxide Dissolved Si in pore waterwas analysedby sample digestion. successivecore from on aliquotsextracted spectrophotometry, Fossil diatom assemupon recovery. immediately increments of in aqueous suspensions blages were analysedquantitatively forfossil (Kling 1998). Sampleprocessing sediment undigested midge analysis followed standard techniques (Walker & from the fossils sieveto extract 1985), usinga 100 ,um Patterson matrix. sediment

3. RESULTS

dynamics sedimentation (a) Offshore 2a) (figure in Lake Victoria sediments Deep-water mudscappedby a 1.5cm organic dark-brown are soft,

(a)

flocculent surface mud

and others 291 EastAfrica D. Verschuren onLake Victoria, impact ofhuman HistoCy

(b) _ mud organic soft (pCi g-') 2 10Pb unsupported

0.1 o
/ -

(c)

Sc (%) solid-phase 0 5 10 15 20 25

10

100

10 15 _
'e 'v

= r

20 :: 25 _ - 30 t 35_ 40 t
O 1 2 3 4 5 cumulative 2) (g cmmass dry

ax}-

o.s l.o l.s pore-waterSi (mM)

cumulative changein Lake Victoria.(a) Sedimentstratigraphy, evidenceforlake-wideecosystem Figure2. Palaeolimnological of core V96-5MC from68 m waterdepth.Black circles,data; solid line, model; dashed line, drymass and 210Pbactivity of biogenicSi in the solid phase (solid line) and dissolvedin pore water(black drymass (g cm-2). (b) Stratigraphy cumulative a hiatusof ca. 40 years(ca. 1885 to 1925). representing disconformity, to the inferred circles);the stippledzone corresponds reflecting acicularis, sp.,and Nitzschia spp., Cyclostephanos of the pelagic diatomsAulacoseira (c) Down-coredistribution surfacemuds are diatoms in flocculent changesin theirabsoluteabundance over the past 180 years;peak concentrations in and Chironomus burial. (d)(i) Abundanceratioof the midgetaxa Procladius beforepermanent dissolution subjectto further in the seasonal gradient the modern-day betweena waterdepthof 48 and 68 m, reflecting Lake Victoriasurfacesediments of the Procladiusl 10 months(r= 0.98, p < 0.01); (ii) evolution of bottomanoxia from0 to approximately persistence abundance ratioat 68 m waterdepthoverthe past 180 years. Chironomus

blanket mudand a thin offlocculent horizon surface thick of recentlysettled algae. Unsupported Pb activity drymass downwithcumulative decreasesexponentially at have accumulated thatsediments core,whichindicates constantrate (0.032 + 0.001 g cm-2yr-' or a relatively mixand biological '), and thatbothphysical 2.3 mmyr(Robbins et al. ing of recentdepositsare insignificant at a core depthof to lower2'0Pb activity 1978). A shift to that we interpret a disconformity 25-26 cm reflects an exceptionerosionduring sediment from haveresulted shaland relative The largewindfetch storm. allyviolent bottom offshore makeits entire lownessof Lake Victoria by highgenerated susceptibleto the wave turbulence storms(Johnson1980; Dearing 1997). Onlyin intensity a rare erosion thedeepestpartofthelakeis wave-induced event, and has the sediment record been preserved recent lake history intact to reconstruct sufficiently of wave turbulence et al. 1998). The effects (Verschuren recofthesediment quality on thearchival and bioturbation evidentat depthsof less than 64 m ord are increasingly than 50 m, areas shallower 1). In manyoffshore (figure depothe permanent prevents wave turbulence frequent 1991). al. et (Scholz muds organic sitionof fine-grained
91

2b) Si in coreV96-5MC (figure The recordofbiogenic et al. 1998) docucores(Verschuren deep-water and other in Lake of pelagic diatomproduction mentsthe history (21% ofdrymass) in the Peak Si concentrations Victoria. diatomssubject settled top 1.5 cm are recently flocculent burial(Conley& permanent before dissolution to further
Proc.R. Soc. Lod. B (2002)

diatomproduction (b) Historzcal

below thatlevelcorreSchelske1989). Si concentrations abundiatom oftotalfossil thelogarithm with latestrongly thatdiatoms dance (r= 0.89, p < 0.001, n = 20), implying Si in these biogenic sourceofsolid-phase are theprincipal sediments(Conley 1988). Low, constant Si concenoftheprofile half mass)in thelower (ca. 5% ofdry trations ofHolocene reto thosein theupperportions are similar the moderate, et al. 1998), and reflect cords (Johnson in Lake Victoria diatomproduction stableoffshore fairly before human impact became noticeable. Increasingly the 1930s to the from deposited Si concentrations higher 1980s reflectthe greaterdiatom productionthat midresultedfrom an increased supply of growth-limiting loss of this excess (Hecky 1993). Cumulative nutrients gradually sediments diatom Si to burial in deep-water Si. Concenofdissolved reservoir Lake Victoria's depleted in the upperwatercolumnduringstratification trations 70-80 ,uMin 1960 to less than8 ,uM have declinedfrom below in 1990 (Hecky 1993), and are now frequently limitdiato severely considered 1,uM a concentration lowerdia(Schelskeetal. 1986). The slightly tomgrowth deposited since the late 1980s tom-Si concentrations in whichdiatom situation the current 2b) reflect (figure concentration is limitedby the dissolved-Si production shallow-water ofSi from recycling through thatis realized through and new inputsfromthe watershed sediments soil runoff. Changes in the pelagic diatomfloraof Lake Victoria evidence further 2c) provide overthepast 60 years(figure loading.The changesin nutrient historical forsignificant 94% comprises acicularis speciesSNitzschia silicified thinly

292 D. Verschuren and others History ofhuman impact onLake Victoria, EastAfrica of totaldiatomabundancein recently settled algae, consistent withlive plankton data (Kling et al. 2001) that showthisspeciesto be themostcommonpelagicdiatom in Lake Victoriatoday.A 30-folddecreasein its fossil abundance immediately belowtheflocculent surface muds (as compared withca. 1.5 timesCyclostephanos sp. and ca. 1.2 times for Aulacoseira spp.) pointsto strong differential dissolution ofNitzschia frustules before permanent burial. This is supported by the depthprofile of dissolved Si in pore water(figure 2b), whichindicates significant Si dissolution and diffusion to theoverlying water columnfrom theuppermost fewcentimetres ofsediment only. Differentialdiatom dissolution implies that thespeciesabundances in fossilassemblagesare not quantitatively comparable withlivefloral composition (Flower1993); however, species abundancetrends in permanently buriedassemblages (i.e. thosebelowtheflocculent surface muds) can be considered to reflect realchangesin thepelagic-diatom community through time(Haworth1980; Leavitt etal. 1994). The fossil diatomdata (figure 2c) showthatthe pelagicdiatomcommunity of Lake Victoria was stablebetween about 1820 and 1940,withCyclostephanos and Aulacoseira co-dominant at ca. 80% and ca. 15%, and Nitzschia acicularzs absent,or too rareto be preserved. The abundance ofall three diatomtaxa started to increase between about 1940 and theearly1960s,with Nitzschia acicularzs achieving ca. 50% of fossilassemblages depositedin the late 1970s and early1980s; a levelofdominance notobserved at any previous time during the past 12 400 years (Stager& Johnson 2000). Consistent with thebiogenic-Si data, the fossil-diatom recordthenindicates a reduction of diatomproduction in the late 1980s, withdeclinesin all threemajordiatomtaxa. The 2l0Pb-inferred timing of thisreduction matches theincreased occurrence ofmassivecyanobacteria bloomsafter 1987 (Ochumba& Kibaara 1989). This indicates thatrestructuring ofthealgal community to cyanobacteria dominance in themid-1980s was influenced by Si limitation of diatomgrowth, due to seasonaldepletion ofa diminished dissolved-Si reservoir. Our fossildata also agree with long-term planktonsurveys (Kling et al. 2001) thatshow the virtual elimination of Aulacoseira spp. overthe past 20 years.In Lake Victoria todaythese diatomsare out-competed because of their high-Si needsand highsinking ratethrough thewater column (Talling 1966; Kilham 1990).
agricultural production (% of 1960value) 1996 100 150 200 250
l | l l l l l l l l l l l l l

1980 1960 . 19401920

1 900 | g |

/ r

<

pelagiccyanobacteriabloom Nile perch population surges, indigenous fish stocks collapse

/ /

lake-level risefloods riparian wetlands Nile perch introduced

railroad arrives inKampala, Uganda Ugandan cotton exports railroad arrives inKisumu, Kenya founding ofKampala,Buganda Arabtrade with Buganda
|
l

18801860 1840 1820 0

discovered byEuropean explorers Bugandakingdom

I
l
* Ww * 1 1

F I

* *

*,

10

20 30 catchment population (millions)

Figure3. Principaleventsin the recentenvironmental history of Lake Victoria, in relation to human-population growth and agricultural production in its drainagebasin.

ingseason (data from Heckyet al. 1994; Ochumba1996). Changesin theProcladiuslChironomus ratiooffossil midge faunas deposited through timeat V96-5MC (figure 2d(ii)) indicatethat,afterat least 140years of adequate yearroundbottom oxygenation, thedeep-water oxygen regime ofLake Victoria started to deteriorate in theearly1960s. Seasonallypersistent deep-water anoxia appearsto have reacheditscurrent spatialextent bythelate 1970s. These fossildata agree with fragmentary historical dissolvedoxygen data from the 1920s, 1960s and 1990s (Heckyet al. 1994), and implythat the intermittent deep-water anoxia first observedin 1960-1961 (Talling 1966) represented theearliest stageofeutrophication-induced deepwateroxygen loss in Lake Victoria. 4. DISCUSSION

The combinedfossilevidenceindicates thathistorical changes inphytoplankton productivity and composition of (c) Development f deep-water anoxia Lake Victoriahave been caused mainlyby bottom-up Fossil assemblages of midge larvae (Insecta: effects ofexcessnutrient loading, and less so byfood-web Chironomidae) in offshore after the 1980s upsurge ofNile perch.Timing surface sediments are a spati- alterations allyintegrated of the inferred productivity reflection of modern benthic communities and progress increasematch growth and agricultural across a large area of the offshore activity in the lake bottom (Frey human-population 1988). In Lake Victoriatheyare rather poor in species Lake Victoria basin (figure 3). Quantitative estimates of diversity, and reflecttoday's principal environmental totalhistorical and current nutrient exports to thelake do gradient in mud-bottom habitat from adequateyear-round not exist(Scherenet al. 2000), but strong correlation of size withagricultural oxygenation at shallow depths to persistent production overthepast seasonal population anoxia(up to 10 months) in deep water. Abundance ratios 40 years(1965-1991: r= 0.97, p < 0.001, n = 27) justifies population size as a proxy indicator of the anoxia-intolerant Procladius brevipetiolatus to the the use of historical soil disturbance and itseffect on nutrianoxia-tolerant Chironomus imicolaamong core stations foranthropogenic ent fluxesin catchment runoff (Meybeck1982; Caraco correlate strongly withwaterdepth(figure 2d(i) r= 0.98, 1995). The populationof the Lake Victoriabasin has p < 0.01, n = 5), and withtheintensity ofseasonalanoxia 4.6 millionin 1932 to 27.7 million (r= 0.91, p < 0.01, n= 5), whichis measured in 1995 as theratio grownfrom (UnitedNations1995). Between1900 and ca. 1930 ithad between theduration ofbottom anoxia(< 1 mg2 l-li in months) and deep-water oxygen content during themix- not increased appreciablybeyond pre-coloniallevels
Proc. R. Soc. Lond. B (2002)

History ofhuman impact on Lake Victoria, EastAfrica D. Verschuren and others 293 becauseofhighnatural mortality underearly colonialrule, exacerbatedby the toll of exotic diseases (Kuczynski 1949). Completion of the Uganda railroad by 1930 openedtheLake Victoria region to settlement bystimulatingplantation agriculture forthe exportof commercial crops (Maxon 1990). Rapid populationgrowth through immigration and improved healthconditions thenstarted the patternof large-scale deforestation and agricultural conversion thatcontinues to thepresent day. The occurrenceof seasonallypersistent deep-water anoxiasincethelate 1970s thatis inferred from our data supportsthe hypothesis (Kaufman & Ochumba 1993; Hecky er al. 1994) thatdeep-water oxygen loss in Lake Victoriamayhave facilitated the decimation of demersal haplochromine fishstocksby Nile perch,by eliminating the deep-water refugium thathad protected thesefishes fromsuch excessivepredationuntil then. This idea is attractive becauseitmight potentially explain why theNile perchpopulation explodedsuddenly, some 25 yearsafter its introduction fromlakes Albertand Turkana (LoweMcConnell 1987). However, theevolution offish species composition in experimental trawls sincethe early1970s (Witte er al. 1995) does notshowevidence thatexpansion of seasonaldeep-water anoxiacaused characteristic deepwater haplochromines to migrateto shallowerdepth rangesand augmentthe food base of Nile perch.Also before theNile perchupsurge, haplochromine densities at depthsof more than 30 m had been significantly lower thanin shallower areas (Kudhongania & Cordone 1974). Hence, although the development of deep-water anoxia musthavecontributed to thedemiseofcertain haplochromine fishtaxa thatwere dependenton the deep-water mudhabitat, itprobably did not'trigger' the 1980s populationexplosion ofNile perch. Our palaeolimnological data establish a strong chronological link betweenhistorical land use and algal productionin Lake Victoria, whichindicates thatlandscape disturbance rather than food-webalterations or climate changeis the dominant cause of the ongoingeutrophication.With current estimates projecting a doublingof the regional humanpopulation to 53 million by theyear 2020 (UnitedNations 1995), further degradation of the Lake Victoriaecosystem can be countered onlyif landmanagement strategies thatseverely restrict nutrient input to thelakeand itstributaries areimplemented on a multinational, basin-wide scale. The authors thank F. W. B. Bugenyi, Y. Chan,T. I=rdal, G. Ngobi,B. Odhiambo and M. Rosenmeier forassistance, theNational Agricultural Research Organisation ofUganda for research permission and access to the R. V. Ibis, and D. Conley, B. Cumming, C. Schelske, D. Schindler, J.F. Talling andF. Witte for discussion andcomments. Thiswork wassupported bytheUS National Science Foundation (T.C.J.),US National OceanicandAtmospheric Administration and Fund forScientific Research-Flanders (D.V.), Norwegian Research Council (M.R.T.) and NaturalSciencesand Engineering Research Council ofCanada (P.R.L.). Thisis IDEAL contribution no. 127. REFERENCES
Balirwa,J. S. & Bugenyi, F. W. B. 1988 An attempt to relate environmental factors to fishecologyin the lotichabitatsof lake Victoria.Verh. Int.Ver. Limnol. 23, 1756-1761.
Proc.R. Soc. Lond.B (2002)

Barel,C. D. N., Dorit,R., Greenwood, P. H., Fryer, G., Hughes,N., Jackson, P. B. N., Kanawabe,H., LoweMcConnell, R. H., Witte, F. & Yamaoka, K. 1985Destructionoffisheries in Africa's lakes. Nature 315, 19-20. Beck, J.1998In New Macmillan Social Studies Atlas for Uganda Kampala, Uganda:Macmillan. Caraco, N. F. 1995Influence ofhuman populations on phosphorus transfers to aquaticsystems: a regional scalestudy using large rivers. In Phosphorus inthe global environment (ed. H. Tiessen), pp. 235-244.NewYork:Wiley. Conley, D. J. 1988 Biogenic silicaas an estimate ofsiliceous microfossil abundance in Great Lakes sediments. BiogeochemistC 6, 161-179. Conley, D. J. & Schelske, C. L. 1989 Processes controlling thebenthic regeneration and sedimentary accumulation of biogenic silicain Lake Michigan. Arch. Hydrobiol. 116,2343. Crul,R. C. M. 1993 Limnology and hydrology ofLake Victoria.In UNESCOlInternational Hydrological Programme-IV Project M-5.1 'Comprehensive and comparative study of the GreatLakes '. Nijmegen, The Netherlands: Crul Consultancy. Dearing,J. A. 1997 Sedimentary indicators of lake-level changes in the humidtemperate zone: a critical review. . Paleolimnol. 18, 1-14. DeMaster, D. J. 1981The supply and accumulation ofsilica in themarine environment. Geochim. Cosmochim. Acta45, 1715-1732. Flower,R. J. 1993 Diatom preservation: experiments and observations on dissolution andbreakage inmodern andfossilmaterial. Hydrobiologia 269/270, 473484. Frey,D. G. 1988 Littoral and offshore communities of diatoms,cladocerans and dipterous larvae, and their interpretation inpaleolimnology. . Paleolimnol. 1, 179-191. Goldschmidt, T., Witte, F. & Wanink, J. 1993 Cascading effects of the introduced Nile perch on detritivorous/ phytoplanktivorous species in thesublittoral areasof Lake Victoria. Conserv. Biol.7, 686. Haworth, E. Y. 1980 Comparison ofcontinuous phytoplanktonrecords with thediatom stratigraphy in therecent sediments ofBlelham Tarn.Limnol. Oceanogr. 25, 1093-1103. Hecky, R. E. 1993The eutrophication ofLakeVictoria. Verh. Int.Ver. Limnol. 25, 39-48. Hecky, R. E. & Bugenyi, F. W. B. 1992Hydrology andchemistry oftheAfrican Great lakesand water issues: problems and solutions. Mitt. Int.Ver. Limnol. 23, 45-54. Hecky, R. E., Bugenyi, F. W. B., Ochumba, P., Talling, J.F., Mugidde, R., Gophen, M. & Kaufman, L. 1994Deoxygenation ofthedeepwater ofLakeVictoria, EastAfrica. Limnol. Oceanogr. 39, 1476-1481. Johnson, T. C.1980 Sediment redistribution bywaves inlakes, reservoirs andembayments. Proc. ASCE Symp. SuCface Water Impoundments 7-9, 13071317. Johnson, T. C., Chan, Y., Beuning, K., Kelts, K., Ngobi, G. & Verschuren, D. 1998 Biogenic silicaprofiles in Holocene coresfrom lakeVictoria: implications forlake-level history and initiation oftheVictoria Nile. In Environmental change and response in EastAfrican Lakes(ed. J.T. Lehman), pp. 75-88. Dordrecht, The Netherlands: Kluwer. Kaufman, L. & Ochumba, P. 1993Evolutionary and conservation biology of cichlid fishes as revealed by faunal remnants in northern LakeVictoria. Conserv. Biol.7, 719-730. Kilham, P.1990 Ecology ofMelosira species intheGreat Lakes ofAfrica. In Large lakes: ecological structure andfunction (ed. M. M. Tilzer & C. Serruya), pp. 414-427. Berlin: Springer. Kling, H. J. 1998 A summary ofpastand recent plankton of Lake Winnipeg,Canada using fossil algal remains. Kling, H. J., Mugidde, R. & Hecky, R. E.2001 Recent changes
. Paleolimnol. 19, 297-307.

294 D. Verschuren and others History ofhuman impact onLake Victoria, EastAfrica
in the phytoplankton communityof Lake Victoria in responseto eutrophication. In The greatlakes of thezvorld (GLOW): food-zveb, health and integrity (ed. M. Munawar& R. E. Hecky), pp. 47-65. Leiden, The Netherlands: Backhuys. Kuczynski, R. R. 1949 Demographic survey ofthe British colonial empire. New Jersey: OxfordUniversity Press. Kudhongania& Cordone 1974 Past trends, present stocksand possiblefuture stateof the fisheries of the Tanzania partof Lake Victoria. Afr.3t.Trop.Hydrobiol. Fish 3, 167-181. Leavitt,P. R., Hann, B. J.,Smol, J. P., Zeeb, B. A., Christie, C. E., Wolfe,B. & Hecky, R. E. 1994 Paleolimnological analysisof whole-lakeexperiments: an overview of results from Experimental Lake Area lake 227. Can. 3t.Fish.Aquat. Sci. 51, 2322-2332. Lowe-McConnell,R. H. 1987 Ecologzcal studies in tropicalfish communities. CambridgeUniversity Press. Maxon, R. M. 1990 Themes in Kenyan history (ed. W. R. Ochieng),pp. 29-35. Nairobi: Currey. Meybeck,M. 1982 Carbon, nitrogen and phosphorustransportby worldrivers. Am. 7. Sci. 282, 401-445. Mugidde,R. 1993 The increase in phytoplankton primary productivity and biomassin Lake Victoria(Uganda). Verh. Int. Ver.Limnol.25, 846-849. Ochumba,P. B. O. 1996 Measurement ofwatercurrents, temperature, dissolvedoxygenand winds on the KenyanLake Victoria.In Thelimnology, climatology and paleoclimatology of theEast Af7ican lakes(ed. T. C. Johnson & E. Odada), pp. 155-167. Newark:Gordon & Breach. Ochumba, P. B. O. & Kibaara, D. I. 1989 Observations on blue-green algal bloomsin theopen watersofLake Victoria, Kenya.Afr.. Ecol. 27, 23-24. Ogutu-Ohwayo, R. 1990 The decline of the nativefishesof Lake Victoriaand Kyoga (East Africa)and the impact of introduced species,especially the Nile perch,Lates niloticus and theNile tilapia,Oreochromis niloticus. Environ. Biol. Fish. 27, 81-86. Robbins,J. A. & Edgington, D. N. 1975 Determination of recentsedimentation rates in Lake Michigan using Pb-2l0 and Cs-137. Geochim. Cosmochim. Acta 39, 285-304. Robbins,J. A., Edgington, D. N. & Kemp, A. L. W. 1978 Comparative2l0Pb, l37Cs, and pollen geochronologies of sedimentsfrom lakes Ontario and Erie. Quat. Res. 10, 256-278. Schelske, C. L., Stoermer,E. F., Fahnenstiel,G. L. & Haibach, M. 1986 Phosphorus enrichment, silicautilization, and biogeochemical silicadepletion in theGreatLakes. Can. . Fish.Aquat. Sci. 43, 407-415. Scheren,P. A. G. M., Zanting, H. A. & Lemmens,A. M. C. 2000 Estimation ofwaterpollution sourcesin Lake Victoria, East Africa: application and elaboration of the rapidassessmentmethodology. . Environ. Mngmt 58, 235-248. Scholz, C. A., Rosendahl,B. R., Versfelt, J. W. & Rach, A. 1991 Resultsof high-resolution echosounding of Lake Victoria. 7. Afr.EarthSci. 11, 25-32. Stager, J. C. & Johnson, T. C. 2000 A 12,400 l4C yroffshore diatomrecordfrom east-central Lake Victoria, East Africa. T Paleolimnol. 23, 373-383. Talling,J.F.1966 The annualcycleofstratification and phytoplanktongrowthin Lake Victoria(East Africa).Int. Rev. Ges. Hydrobiol. 51, 545-621. United Nations 1995 Worldpopulation prospects: the 1994 revision. New York: United Nations. Verschuren, D., Edgington, D. N., Kling, H. J. & Johnson, T. C. 1998 Silica depletionin Lake Victoria:sedimentary signalsat offshore stations. 7. Great Lakes Res. 24, 118-130. Walker,I. R. & Patterson, C. G. 1985 Efficient separation of subfossil Chironomidae fromlake sediments. Hydrobiologia 122, 189-192. Witte, F., Goldschmidt,T., Wanink, J., Van Oijen, M., Goudswaard,K., Witte-Maas,E. & Bouton, N. 1992 The destruction ofan endemicspeciesflock: quantitative data on the declineof the haplochromine cichlidsof Lake Victoria. Environ. Biol. Fish. 29, 1-28. Witte, F., Goldschmidt, T. & Wanink, J. 1995 Dynamicsofthe haplochromine cichlidfaunaand other ecologicalchangesin the Mwanza Gulf of Lake Victoria.In The impact ofspecies changes in African lakes (ed. T. J. Pitcher& P. J. B. Hart), pp. 83-110. London: Chapman & Hall.

Proc. R. Soc. Lond. B (2002)