6. J. Kochi, ibid., p. 1958.7. C. Walling and A. A. Zavitsas, ibid., p. 2084.8. J. Kochi and R. Subramanian, ibid., 87, 4855

(1965).9. 3. Kerr and A. Trotman-Dickenson, Progr.

Reaction Kinetics 1, 107 (1961).10. E. Collinson, F. Dainton, B. Mile, S. Tazuke,

D. Smith, Nature 198, 26 (1963); DiscussionsFaraday Soc. 29, 188 (1960).

11. G. Olah, Chem. Soc. London Spec. Pub. 19,21 (1965).

12. F. Lossing, J. deSousa, R. Pottie, R. Taubert,J. Amer. Chem. Soc. 81, 281 (1959); ibid. 83,4737 (1961); ibid. 84, 1521 (1962). The valuesfor oxidation-reduction potentials oan be in-fluenced greatly by solvent. Unfortunately,these values are klnown well only in aqueoussolutions. Estimates in nonaqueous solventsare qualitatively correct.

13. J. Halpern, R. Legare, R. Lumry, ibid. 85,680 (1964); P. Hurwitz and K. Kustin,Inorg. Chem. 3, 823 (1964).

14. R. Marcus, Trans. Faraday Soc. 56, 21 (1960),and earlier papers.

15. H. McConnell and H. Weaver, J. Chem.Phys. 25, 307 (1956).

16. J. Bjerrum, Kem. Maanelsblad 26, 24 (1945).

17. S. Manahan and R. Iwamoto, Inorg. Chem.4, 1409 (1965).

18. J. Kochi and R. Subramanian, Ibid., p. 1527.19. J. Kochi, J. Amer. Chem. Soc. 84, 2795

(1962); C. Walling and W. Thaler, Ibid. 83,3871 (1961).

20. J. Kochi and D. Mog, ibid. 87, 522 (1965).21. J. Kochi, ibid. 84, 774 (1962); Ibid., p. 3271.22. H. Taube, Chem. Rev. 50, 69 (1952).23. J. Kochi and D. Davis, J. Amer. Chem. Soc.

86, 5264 (1964).24. M. Szwarc, C. Leigh, A. Sehon, Proc. Roy.

Soc. London Ser. A 209, 97 (1951).25. A. Streitwieser and C. Perrin, J. Amer. Chem.

Soc. 86, 4938 (1964).26. R. Gilliom and B. Ward, ibid. 87, 3944

(1965).27. J. Kochi and D. Davis, Nature 202, 690

(1964).28. H. Taube and A. Ogard, J. Amer. Chem.

Soc. 80, 41084 (1958); "Mechanism of In-organic Reactions," Amer. Chem. Soc. Pub.No. 49 (1965).

29. W. Kray and C. Castro, J. Amer. Chem.Soc. 86, 4603 (1964).

30. C. Castro, private communication.31. L. Menapace and H. Kuivila, J. Amer.

Chem. Soc. 86, 3047 (1964); ibid. 88, 571(1966).

32. D. Benson, P. Proil, L. Sutcliffe, 3. Walkley,Discussions Faraday Soc. 29, 60 (1960);Trans. Faraday Soc. 56, 246 (1960).

33. J. Kochi, J. Amer. Chem. Soc. 87, 3609 (1965).34. , J. Org. Chem. 30, 3265 (1965).35. W. Lundberg, Autoxidation and Antioxidants

(Interscience, New York, 1962); Y. Kamiyaand K. Ingold, Can. J. Chem. 41, 2020 (1963),and later papers.

36. L. Boguslavskaya, Russ. Chem. Rev. (EnglishTransl.) 34, 503 (1965).

37. F. Minisci and R. Galli, Tetrahedron Letters1963, 357 (1963).

38. J. Kochi and H. Mains, J. Org. Chem. 30,1862 (1965).

39. M. Asscher and D. Vofsi, J. Chem. Soc.1963, 1887 (1963); ibid., p. 3921.

40. S. Dickerman, K. Weiss, A. Ingberman, J.Amer. Chem. Soc. 80, 1904 (1959); J. Kochi,ibid. 79, 2942 (1958).

41. W. Brackman and E. Havinga, Rec. Trav.Chim. 74, 937 (1955); ibid. 84, 579 (1965);and later papers.

42. Support from the National Science Founda-tion is gratefully acknowledged.

Nutrient Cycling

Small watersheds can provide invaluableinformation about terrestrial ecosystems.

F. H. Bormann and G. E. Likens

Life on our planet is dependent uponthe cycle of elements in the biosphere.Atmospheric carbon dioxide would beexhausted in a year or so by greenplants were not the atmosphere con-

tinually recharged by CO2 generatedby respiration and fire (1). Also, it iswell known that life requires a con-

stant cycling of nitrogen, oxygen, andwater. These cycles include a gaseousphase and have self-regulating feed-back mechanisms that make them rela-tively perfect (2). Any increase in move-

ment along one path is quickly com-

pensated for by adjustments along otherpaths. Recently, however, concern hasbeen expressed over the possible dis-ruption of the carbon cycle by theburning of fossil fuel (3) and of thenitrogen cycle by the thoughtless intro-duction of pesticides and other sub-stances into the biosphere (4).Of no less importance to life are

the elements with sedimentary cycles,such as phosphorus, calcium, and mag-

nesium. With these cycles, there is a

continual loss from biological systemsin response to erosion, with ultimatedeposition in the sea. Replacement or

return of an element with a sedimen-

424

tary cycle to terrestrial biological sys-tems is dependent upon such processesas weathering of rocks, additions fromvolcanic gases, or the biological move-

ment from the sea to the land. Sedi-mentary cycles are less perfect andmore easily disrupted by man than car-

bon and nitrogen cycles (2). Accelera-tion of losses or, more specifically, thedisruption of local cycling patterns bythe activities of man could reduce exist-ing "pools" of an element in local eco-

systems, restrict productivity, and con-

sequently limit human population. Forexample, many agriculturalists, food sci-entists, and ecologists believe that man

is accelerating losses of phosphorusand that this element will be a criti-cal limiting resource for the function-ing of the biosphere (1, 5).

Recognition of the importance ofthese biogeochemical processes to thewelfare of mankind has generated in-tensive study of such cycles. Amongecologists and foresters working withnatural terrestrial ecosystems, this in-terest has focused on those aspects ofbiogeochemical cycles that occur withinparticular ecosystems. Thus, informa-tion on the distribution of chemical

elements and on rates of uptake, re-tention, and release in various ecosys-tems has been accumulating (6). Littlehas been done to establish the role thatweathering and erosion play in thesesystems.

Yet, the rate of release of nutrientsfrom minerals by weathering, the addi-tion of nutrients by erosion, and theloss of nutrients by erosion are threeprimary determinants of structure andfunction in terrestrial ecosystems. Fur-ther, with this information it is possi-ble to develop total chemical budgetsfor ecosystems and to relate these datato the larger biogeochemical cycles.

It is largely because of the complexnatural interaction of the hydrologic cy-cle and nutrient cycles that it has notbeen possible to establish these relation-ships. In many ecosystems this inter-action almost hopelessly complicatesthe measurement of weathering or ero-sion. Under certain conditions, how-ever, these apparent hindrances can beturned to good advantage in an inte-grated study of biogeochemical cyclingin small watershed ecosystems.

It is the function of this article (i)to develop the idea that small water-sheds can be used to measure weather-ing and erosion, (ii) to describe theparameters of watersheds particularlysuited for this type of study, and (iii)to discuss the types of nutrient-cyclingproblems that this model renders sus-ceptible to attack. Finally (iv), the argu-ment is developed that the watershedecosystem provides an ideal setting forstudies of ecosystem dynamics in gen-eral.

Dr. Bormann is professof of forest ecology,Yale School of Forestry, Yale TJniversity, NewHaven, Connecticut; Dr. Likens is associate pro.fessor, department of biological sciences, Dart-mouth College, Hanover, New Hampshire.

SCIENCE, VOL. 155

on Septem

ber 6, 2020

http://science.sciencemag.org/

Dow

nloaded from

Ecosystem Defined

Communities such as fields and for-ests may be considered as ecologicalsystems (7) in which living organismsand their physical and biological en-vironments constitute a single inter-acting unit. These ecosystems occupyan arbitrarily defined volume of thebiosphere at the earth-atmosphere in-terface.

Lateral boundaries of an ecosystemmay be chosen to coincide with thoseof a biological community, such as theedges of a forest, or with the boundaryof some pronounced characteristic ofthe physical environment, such as theshoreline of a small island. Most often,however, the continuous nature of veg-etation and of the physical environ-ment makes it difficult to establishexact lateral boundaries on the basis of"community" or "environmental dis-continuity" (8). Often the investigatorarbitrarily selects an area that may beconveniently studied.From a functional point of view it

is meaningless to include within the ver-tical limits of an ecosystem all of thecolumn of air above and of soil androck below the laterally defined eco-system. For a working model of anecosystem, it seems reasonable to in-clude only that part of the columnwhere atoms and molecules may par-ticipate in the chemical cycling thatoccurs within the system (see the "in-trasystem cycle" of Fig. 1). When thebiological community is taken as adeterminant, the vertical extensions ofthe terrestrial ecosystem will be de-limited by the top of the vegetationand the depth to which roots and otherorganisms penetrate into the regolith(9). Vertical dimensions, defined in thismanner, can expand or contract de-pending on the growth potential ofpresent or succeeding communities.Thus, volumetric changes with time canbe considered-for example, those as-sociated with primary and secondarysuccession or with cliseral changes.

The Ecosystem and

Biogeochemical Cycling

The terrestrial ecosystem participatesin the various larger biogeochemicalcycles of the earth through a systemof inputs and outputs. Biogeochemicalinput in forest or field ecosystems maybe derived from three major sources:geologic, meteorologic, and biologic.Geologic input is here defined as dis-27 JANUARY 1967

Fig. 1. Nutrient relationships of a terrestrial ecosystem, showing sites of accumulationand major pathways. Input and output may be composed of geologic, meteorologic,and biologic components, as described in the text.

solved or particulate matter carriedinto the system by moving water or

colluvial action, or both. Depositionsof the products of erosion or mass

wasting and ions dissolved in incom-ing seepage water are examples of geo-logic input. Meteorologic input entersthe ecosystem through the atmosphereand is composed of additions of gase-ous materials and of dissolved or par-

ticulate matter in precipitation, dust,and other wind-borne materials.. Chem-icals in gaseous form fixed by biologicactivity within the ecosystem are con-

sidered to be meteorologic input. Bio-logical input results from animal activ-ity and is made up of depositions ofmaterials originally gathered elsewhere;examples are fecal material of animalswhose food was gathered outside thesystem, or fertilizers intentionally addedby man.

Chemicals may leave the ecosystemin the form of dissolved or particulatematter in moving water or colluvium,or both (geologic output); through thediffusion or transport of gases or par-ticulate matter by wind (meteorologicoutput); or as a result of the activityof animals, including man (biologic out-put).

Nutrients are found in four compart-ments within the terrestrial ecosystem:in the atmosphere, in the pool of avail-able nutrients in the soil, in organicmaterials (biota and organic debris),and in soil and rock minerals (Fig. 1).

The atmospheric compartment includesall atoms or molecules in gaseous formin both the below-ground and theabove-ground portions of the ecosys-tem. The pool of available nutrientsin the soil consists of all ions adsorbedon the clay-humus complex or dis-solved in the soil solution. The organiccompartment includes- all atoms incor-porated in living organisms and in theirdead remains. (The distinction betweenliving and dead is sometimes hard tomake, particularly in the case of woodyperennial plants.) The soil-rock com-partment is comprised of elements in-corporated in primary and secondaryminerals, including the more readily de-composable minerals that enter intoequilibrium reactions with the availablenutrients.The degree to which a nutrient cir-

culates within a terrestrial ecosystemis determined, in part, by its physicalstate. Gases are easily moved by ran-dom forces of diffusion and air circu-lation; consequently, nutrients with aprominent gaseous phase tend not tocycle within the boundaries of a par-ticular ecosystem but, rather, to be con-tinually lost a.nd replaced from outside.On the other hand, elements without aprominent gaseous phase may showconsiderable intrasystem cycling be-tween the available-nutrient, organic,and soil-rock compartments (Fig. 1).This internal cycle results from (i) theuptake of nutrients by plants, (ii) the

425

on Septem

ber 6, 2020

http://science.sciencemag.org/

Dow

nloaded from

release of nutrients from plants by di-rect leaching, (iii) the release of nu-trients from organic matter by biologi-cal decomposition, and (iv) equilibriumreactions that convert insoluble chemi-cal forms in the soil-rock compartmentto soluble forms in the available-nutri-ent compartment, and vice versa.

Available nutrients not only enterthe ecosystem from outside but areadded by the action of physical, chem-ical, and biological weathering of rockand soil minerals already within thesystem. Although some ions are con-

tinually withdrawn from the available-nutrient compartment, forming second-ary minerals in the soil and rocks,for most nutrient elements there is a

net movement out of the soil-rock com-partment. As the ecosystem is gradual-ly lowered in place by erosion or bythe downward growth of roots, new

supplies of residual rock or other par-ent material are included; in some sys-tems these materials may also be addedas geologic input.

Hydrologic-Cycle, Nutrient-CycleInteraction

At many points, nutrient cycles maybe strongly geared to the hydrologiccycle. Nutrient input and output are di-rectly related to the amounts of waterthat move into and out of an ecosystem,as emphasized by the "leaching" and"flushing" concepts of Pearsall (10),Dahl (11), and Ratcliffe (12), whiletemporal and absolute limits of biogeo-chemical activities within the systemare markedly influenced by the hydro-logic regime. Biologic uptake of nu-trients by plants and release of nu-

trients by biological decomposition are

closely related to the pattern of wateravailability. Potential levels of biomasswithin the system are determined inlarge measure by precipitation charac-teristics. Similarly, the nature and rateof weathering and soil formation areinfluenced by the hydrologic regime,since water is essential to the majorchemical weathering processes [ion ex-change, hydrolysis, solution, diffusion,oxidation-reduction, and adsorption andswelling (13)].

Biogeochemical Studies of Ecosystems

Although study of nutrient input,nutrient output, and weathering is nec-

essary for an understanding of fieldand forest ecosystems (6), ecologists,

426

foresters, and pedologists have general-ly focused attention on the internalcharacteristics alone. Thus, consider-able information has been accumulatedon uptake, retention, and release of nu-trients by the biota of ecosystems, andon soil-nutrient relationships (see, forexample, 6, 14). Rarely are these in-ternal characteristics of the ecosystemcorrelated with input and output data,yet all these parameters are necessaryfor the construction of nutrient bud-gets of particular ecosystems, and forestablishing the relationship of thesmaller system to the biosphere.

Quantitative data on input-output re-lationships are at best spotty. Thereare many data on nutrient output dueto harvesting of vegetation, but for par-ticular natural ecosystems there areonly sporadic data on nutrient inputin precipitation, or on nutrient outputin drainage waters (6). Recently, smalllysimeters have been used successfullyin the measurement of nutrient dynam-ics within the soil profile and in themeasurement of nutrient losses in drain-age water (15). The lysimeter tech-nique seems to be well suited for study-ing ecosystems characterized by coarse-textured soils with a relatively low fieldcapacity (16), high porosity, and nosurface runoff. For most ecosystems,however, lysimeters are probably oflimited value for measuring total nutri-ent output because (i) they are of ques-tionable accuracy when used in rockyor markedly uneven ground, (ii) theycannot evaluate nutrient losses in sur-face waters, and (iii) their installationrequires considerable disturbance of thesoil profile.The lack of information on the nu-

trient-input, nutrient-output relation-ships of ecosystems is apparently re-lated to two considerations: (i) inte-grated studies of ecosystems tend tofall into an intellectual "no man'sland" between traditional concepts ofecology, geology, -and pedology; (ii)more important, the measurement ofnutrient input and output requiresmeasurement of hydrologic input andoutput. Unquestionably this lack ofquantitative information is related tothe difficulties encountered in mea-suring nutrients entering or leavingan ecosystem in seepage water or insheet or rill flow, and to the high cost,in time and money, of obtaining con-tinuous measurements of the more con-ventional hydrologic parameters of pre-cipitation and stream flow. In manysystems the problem is further compli-cated by the fact that much water may

leave by way of deep seepage, eventu-ally appearing in another drainage sys-tem; direct measurement of loss ofwater and nutrients by this route isvirtually impossible.

Small-Watershed Approach to

Biogeochemical Cycling

In some ecosystems the nutrient-cycle, hydrologic-cycle interaction canbe turned to good advantage in thestudy of nutrient budgets, erosion, andweathering. This is particularly so ifan ecosystem meets two specifications:(i) if the ecosystem is a watershed,and (ii) if the watershed is underlainby a tight bedrock or other imperme-able base, such as permafrost. Giventhese conditions, for chemical elementswithout a gaseous form at biologicaltemperatures, it is possible to constructnutrient budgets showing input, output,and net loss or gain from the system.These data provide estimates of weath-ering and erosion.

If the ecosystem were a small water-shed, input would be limited to mete-orologic and biologic origins. Geologicinput, as defined above, need not beconsidered because there would be notransfer of alluvial or colluvial materialbetween adjacent watersheds. Althoughmaterials might be moved within theecosystem by alluvial or colluvial forces,these materials would originate withinthe ecosystem.When the input and output of dust

or windblown materials is negligible(this is certainly not the case irisome systems), meteorologic input canbe measured from a combination ofhydrologic and precipitation-chemistryparameters. From periodic measure-ments of the elements contained in pre-cipitation and from continuous measure-ments of precipitation entering a water-shed of known area, one may calcu-late the temporal input of an elementin terms of grams per hectare. Nonco-incidence of the topographic divide ofthe watershed and the phreatic dividemay introduce a small error (17).

Losses from this watershed ecosys-tem would be limited to geologic andbiologic output. Given an imperme-able base, geologic output (losses dueto erosion) would consist of dissolvedand particulate matter in either streamwater or seepage water moving down-hill above the impermeable base. Al-though downhill mass movement mayoccur within the system, the productsof this movement are delivered to the

SCIENCE, VOL. 155

on Septem

ber 6, 2020

http://science.sciencemag.org/

Dow

nloaded from

stream bed, whence they are removedby erosion and stream transportation.

Geologic output can be estimatedfrom hydrologic and chemical measure-ments. A weir, anchored to the bed-rock (Fig. 2), will force all drain-age water from the watershed to flowover the notch, where the volume andrate of flow can be measured. Thesedata, in combination with periodicmeasures of dissolved and particulatematter in the outflowing water, providean estimate of geologic output whichmay be expressed as grams of an ele-ment lost per hectare of watershed.The nutrient budget for a single ele-

ment in the watershed ecosystem maybe expressed as follows: (meteorologicinput + biologic input) - (geologic out-put + biologic output) = net loss orgain. This equation may be furthersimplified if the ecosystem meets athird specification-if it is part of amuch larger, more or less homogene-ous, vegetation unit. Biological outputwould tend to balance biological inputif the ecosystem contained no specialattraction or deterrent for animal popu-lations moving at random through thelarger vegetation system, randomly ac-quiring or discharging nutrients. Onthis assumption, the nutrient budget fora single system would become: (mete-orologic input per hectare) - (geologicoutput per hectare) = net gain or lossper hectare. This fundamental relation-ship provides basic data for an inte-grated study of ecosystem dynamics.

Small Watersheds for

Ecosystem Research

The relationship of the individual ter-restrial ecosystem to biogeochemicalcycles of the biosphere can be estab-lished by the small-watershed approach.Data on input and output of nutrientsprovide direct measurements of this re-lationship, while data on net loss pro-vide, as explained below, an indirectmeasure of weathering rates -for soiland rock minerals in relatively undis-turbed ecosystems.The small watershed may be used

for experiments at the ecosys;tem level.This has been shown by numerous ex-periments concerned with hydrologic re-lationships (see, for example, 18). Thus,it is possible to test the effects of vari-ous experimental treatments on the re-lationship of the individual ecosystemto the biospheric nutrient cycles. Ex-periments can be designed to determinewhether logging, burning, or use of27 JANUARY 1967

Fig. 2. A weir showing the v-notch, recording house, and ponding basin. [Courtesyof the Northeastern Forest Experiment Station]

pesticides or herbicides have an appre-ciable effect on net nutrient losses fromthe system. This information is notgenerally available at the ecosystemlevel.The small watershed, with its mea-

sured parameters of hydrologic andchemical input, output, and net change,is an excellent vehicle for the studyof interrelationships within a single eco-system. Nutrient output may be relatedto hydrologic parameters such as sea-sonal and diurnal variations in streamflow, seasonal patterns of precipitation,individual rainstorms, and variations inevapotranspiration. Characteristics ofthe nutrient cycle may also be relatedto phenological events occurring with-in the ecosystem, such as leaf develop-ment, initiation of root growth, leaffall, and litter turnover. In combinationwith current methods of biomass andnutrient analysis (see, for example, 6),the small-watershed approach providesa comprehensive view of the status andbehavior of individual elements withinan individual ecosystem.

Weathering, or the rate at which anelement bound in soil and rock min-erals is made available, can be esti-mated from net losses of that elementas calculated by the- nutrient-budgetmethod. Within the ecosystem (water-shed), atoms of an element (one thatlacks a gaseous form at ecosystemtemperatures) may be located in (i) soiland rock minerals, (ii) the biota and

organic debris, and (iii) the pool ofavailable nutrients (Fig. 1). There is anintense intrasystem cycling betweencategories (ii) and (iii) as large quanti-ties of ions are taken up by the vege-tation each year and released by directleaching or stepwise decomposition inthe food chain. Ions are continuallyreleased to the intrasystem cycle byweathering of soil and rock material.Some of these ions, however, are re-constituted as secondary minerals. Ifan ecosystem is in a state of dynamicequilibrium, as the presence of climaxforest would suggest (19), ionic levelsin the intrasystem cycle must remainabout the same for many years. Thus,in the climax ecosystem, net ion losses(output minus input) must be balancedby equivalent additions derived fromweathering of soil and rock materials.Thus, net ionic losses from an undis-turbed, relatively stable terrestrial eco-system are a measure of weatheringwithin the system. In a successionalecosystem (in which nutrients are ac-cumulating in biomass and organicdebris over the course of years), therate at which an ion is released, by-weathering must equal its rate of netloss from the ecosystem plus its rateof net accumulation in the biota andorganic debris (Fig. 1).The watershed model allows com-

parison in relative importance ofsolution and suspended bed load inremoving nutrients from an ecosystem.

427

on Septem

ber 6, 2020

http://science.sciencemag.org/

Dow

nloaded from

Table 1. Budgets for dissolved cations inwatershed No. 3 (42.4 hectares) for theperiod June 1963 to June 1964.

Input Output changeCation (kg/ (kg/ cag

hectare) hectare) hectgre)Calcium 3.0 7.7 -4.7Sodium 1.0 6.3 -5.3Magnesium 0.7 2.5 -1.8

Nutrient matter can be removed froman ecosystem by three forms of trans-portation in streams: in solution in thestream water, as inorganic and organicsuspended load kept in motion by tur-bulent flow, and as inorganic and or-

ganic bed load slid or rolled along thestream bottom (20). Solution lossesmay be measured, as described above,from stream-flow data and periodicmeasurements of dissolved substancesin the stream. Part of the losses ofsuspended matter may be estimatedfrom stream-flow data and periodicmeasurements of particulate matter ob-tained by straining or filtering streamwater as it comes over the weir. Therem'aining suspended matter and all ofthe bed load may be measured abovethe weir, where these materials collectin the ponding basin (Fig. 2). Thesecomparative measurements should be ofinterest not only to the ecologist con-

cerned with ecosystem dynamics butalso, since stream transportation is one

of the important aspects of fluvial de-nudation, to geologists.

NORTHERN HARDWOOD ECOSYSTEM

: 't .. 9 TREES Al; M ) -: / LITTER

The small-watershed approach pro-vides invaluable baseline informationfor the investigation of stream biology.Life-history studies of stream orga-nisms, population studies, and shifts incommunity structure and diversitymight be correlated with the measuredphysical and chemical parameters ofdrainage streams. Analyses of uptake,release, and transport of various nu-

trients by stream organisms could bemade. Moreover, the vegetation of a

watershed and the stream draining itare an inseparable unit functionally,and it would be of great interest to ob-tain information on the biological inter-action between them.

Sites for Watershed Studies

Small watersheds meeting the condi-tions outlined above are probably com-

mon. However, even if the desired con-

ditions are met, the investigator study-ing nutrient cycling is faced with thetask of initiating a hydrologic studybefore he can attack his major prob-lem. This is a time-consuming and ex-

pensive procedure, involving construc-tion and maintenance of weirs, estab-lishment of a precipitation network, andcontinuous collection of records, as

well as land rental fees and possibleroad construction costs. A practical so-

lution to this problem is inaugurationof nutrient cycling studies at establishedhydrologic laboratories, where the re-

Fig. 3. Estimated parameters for the calcium cycle in an undisturbed northernhardwood ecosystem in central New Hampshire. [Data on trees, litter, and exchangeablecalcium from Ovington (6)]

428

quired conditions exist and where hy-drologic parameters are being mea-sured and data are available.The feasibility of this approach is

demonstrated by our study at the Hub-bard Brook Experimental Forest inWest Thornton, New Hampshire.There, with the support of the Na-tional Science Foundation and the ex-cellent cooperation of the NortheasternForest Experiment Station, we arestudying nutrient cycling and ecosys-tem dynamics on six small monitoredwatersheds. We have accumulated dataon weathering rates, input, output, andthe annual budget of several ions inthis northern hardwood ecosystem.Also, studies of biomass, phenology,productivity, annual rates of nutrientturnover, and other factors are beingmade in one undisturbed watershedand in one in which conditions arebeing experimentally modified.

Preliminary data on input, output,and net change for three cations arepresented in Table 1. These results al-low us to add some numerical valuesto our ecosystem model (Fig. 3). Forthe calcium cycle, for example, inputwould be about 3 kilograms per hec-tare, while output (erosion) is esti-mated to be about 7.9 kilograms perhectare. Of this latter amount, 98 per-cent (7.7 kilograms per hectare) is lostin the form of dissolved substances inthe stream water, while first approxi-mations indicate that 2 percent is lostas calcium incorporated in organicmatter flushed out of the ecosystem.On the basis of assumptions discussedabove, it is estimated that the netamount of calcium lost, approximately5 kilograms per hectare, is replaced bycalcium released from soil and rockminerals by weathering. Hence 5 kilo-grams of calcium per hectare is addedto the system each year by weathering.

As yet we have not measured thecalcium content of the soil and vege-tation or the annual uptake and releaseof calcium by the biota. From Oving-ton's data (6). for a beeCh forest inWest England, which must be of aboutthe same magnitude as values for ourforest in New Hampshire, we see that203 kilograms of calcium per hectareare localized in the trees and litter,while 365 kilograms per hectare rep-resent exchangeable calcium in the soil.This gives a total of 568 kilogramsof calcium per hectare in organic mat-ter or as available nutrient. Assumingthat our forest (Fig. 3) contains a sim-ilar amount of calcium, we estimatethat a net annual loss of 5 kilograms

SCIENCE, VOL. 155

on Septem

ber 6, 2020

http://science.sciencemag.org/

Dow

nloaded from

per hectare would represent only nine-tenths of 1 percent of the total. Thissuggests a remarkable ability of theseundisturbed systems to entrap and holdnutrients. However, if these calcula-tions were based on actual amounts

of calcium circulated each year ratherthan on the total, the percentage losseswould be higher.On its completion, the Hubbard

Brook study will have yielded estimates,for individual elements, of many of theparameters and flux rates representedin the nutrient cycle shown in Fig. 1.These data will increase our under-standing of fundamental nutrient rela-tionships of undisturbed northern hard-wood forests, and they will providebaseline information from which we can

judge the effects on nutrient cyclingof such practices as cutting, burning,and the application of pesticides.

Studies similar to these at HubbardBrook could be established elsewherein the United States. There are thou-sands of gaged watersheds operated byprivate and public interests (17), andsome of these must meet the proposedrequirements. On selected watersheds,cooperative studies could be made bythe agencies or organizations control-ling the watershed and university-basedinvestigators interested in biogeochemi-cal cycling. Just such cooperation, be-tween federal agencies and universities,has been urged by the Task Group on

Coordinated Water Resources Research(21).

Cooperative studies of this type havethe advantage of providing a usefulexchange of ideas between scientists indiverse fields who are working on the

same ecosystem. The studies would pro-vide a larger yield of information on a

single system, the prospect of new con-

cepts arising from the available infor-mation, and a greater scientific yieldper dollar invested. Finally, coopera-tive studies would make available, forinterpretation from the standpoint ofnutrient cycling, an invaluable recordof past hydrologic performance and,in some cases, of the responses ofwatersheds to experimental manipula-tion.

Conclusion

The small-watershed approach toproblems of nutrient cycling has theseadvantages. (i) The small watershed isa natural unit of suitable size for in-tensive study of nutrient cycling at theecosystem level. (ii) It provides a means

of reducing to a minimum, or virtual-ly eliminating, the effect of the diffi-cult-to-measure variables of geologicinput and nutrient losses in deep seep-

age. Control of these variables makespossible accurate measurement of nu-

trient input and output (erosion) andtherefore establishes the relationship ofthe smaller ecosystem to the larger bios-pheric cycles. (iii) The small-watershedapproach provides a method wherebysuch important parameters as nutrientrelease from minerals (weathering) andannual nutrient budgets may be calcu-lated. (iv) It provides a means of study-ing the interrelationships between thebiota and the hydrologic cycle, variousnutrient cycles, and energy flow in a

single system. (v) Finally, with the

small-watershed system we can test theeffect of various land-management prac-tices or environmental pollutants on nu-

trient cycling in natural systems.References and Notes

1. L. C. Cole, Sci. Amer. 198, 83 (Apr. 1958).2. E. P. Odum, Ecology (Holt, Rinehart, and

Winston, New York, 1963).3. R. Revelle, W. Broecker, H. Craig, C. D.

Keeling, J. Smagorinsky, in Restoring theQuality of Our Environment (GovernmentPrinting Office, Washington, D.C., 1965), pp.111-133.

4. L. C. Cole, Saturday Rev. (7 May 1966).5. E. P. Odum, Fundamentals of Ecology

(Saunders, Philadelphia, 1959).6. J. D. Ovington, in Advances in Ecological

Research, J. B. Cragg, Ed. (Academic Press,New York, 1962), vol. 1.

7. A. G. Tansley, Ecology 16, 284 (1935).8. L. B. Slobodkin, Growth and Regulation of

Animal Populations (Holt, Rinehart, and Win-ston, New York, 1961).

9. H. 0. Buckman and N. C. Brady, The Natureand Properties of Soils (Macmillan, NewYork, 1960).

10. W. H. Pearsall, Mountains and Moorlands(Collins, London, 1950).

11. E. Dahl, "Rondane. Mountain vegetation inSouth Norway and its relation to the environ-ment," Skrifter Norske Videnskaps-Akad. Oslo1, Mat. Naturv. Kl. 1956 ( 1956).

12. D. A. Ratcliffe, J. Ecol. 47, 371 (1959).13. C. Bould, in Plant Physiology, A Treatise,

F. C. Steward, Ed. (Academic Press, NewYork, 1963), vol. 3.

14. J. R. Bray and E. Gorham, in Advances inEcological Research, J. B. Cragg, Ed. (Aca-demic Press, New York, 1964), vol. 2; K. J.Mustanoja and A. L. Leaf, Botan. Rev. 31,151 (1965).

15. D. W. Cole and S. P. Gessel, Soil Sci. Soc.Amer. Proc. 25, 321 (1961); , in Forest-Soil Relationships in North America, C. T.Youngberg, Ed. (Oregon State Univ. Press,Corvallis, 1965).

16. D. W. Cole, Soil Sci. 85, 293 (1958).17. C. 0. Wisler and E. F. Brater, Hydrology

(Wiley, New York, 1959).18. R. E. Dils, A Guide to the Coweeta Hy-

drologic Laboratory (Southeastern Forest Ex-perimental Station, Asheville, N.C., 1957).

19. R. H. Whittaker, Ecol. Monographs 23, 41(1953).

20. A. N. Strahler, The Earth Sciences (Harperand Row, New York, 1963).

21. R. Revelle, Science 142, 1027 (1963).22. Financial support was provided by National

Science Foundation grants GB 1144 and GB4169. We thank J. Cantlon, N. M. Johnson,R. C. Reynolds, R. H. Whittaker, and G. W.Woodwell for critical comments and sugges-tions during preparation of the manuscript.

Where Is Biology Taking Us?

Robert S. Morison

Others in this symposium (1) havedescribed the expected improvement inour knowledge of perception, cogni-tion, and learning, and have shownhow these improvements can be ex-

pected to facilitate the educationalprocess. It remains for me to try toidentify some of the long-term practicalconsequences of these trends, so thatwe can prepare ourselves to exploit27 JANUARY 1967

the advantages and minimize the dan-gers which accompany any advancein knowledge or technique. As formaleducation improves in effectiveness, itseems natural to suppose that its publicimage will continue to be enhanced.We in the United States have alwaysheld institutionalized education in highrespect and, second only to our Sovietfriends, have looked to it to solve all

manner of individual and social evils.As it becomes more and more capableof actually doing so, its prestige mustnecessarily continue to increase con-comitantly. What, then, are the prob-able consequences of the increasedprestige of institutionalized education?No doubt there will be a considerablenumber, but I should like to lookparticularly at its effect on more tra-ditional ways of transmitting accumu-lated experience to a new generationand to lay before you my reasons forbelieving that, as public recognitionof formal education continues to rise,

The author is director of the Division of Biolog-ical Sciences, Cornell University, Ithaca, NewYork. This article is adapted from a lecturepresented in October 1966 at the CaliforniaInstitute of Technology, Pasadena, on the occa-sion of the Institute's 75th Anniversary Convoca-tion.

429

on Septem

ber 6, 2020

http://science.sciencemag.org/

Dow

nloaded from

Nutrient CyclingF. H. Bormann and G. E. Likens

DOI: 10.1126/science.155.3761.424 (3761), 424-429.155Science 

ARTICLE TOOLS http://science.sciencemag.org/content/155/3761/424

REFERENCES

http://science.sciencemag.org/content/155/3761/424#BIBLThis article cites 8 articles, 1 of which you can access for free

PERMISSIONS http://www.sciencemag.org/help/reprints-and-permissions

Terms of ServiceUse of this article is subject to the

trademark of AAAS. is a registeredScienceAdvancement of Science, 1200 New York Avenue NW, Washington, DC 20005. The title

(print ISSN 0036-8075; online ISSN 1095-9203) is published by the American Association for theScience

of Science. No claim to original U.S. Government Works.Copyright © 1967 The Authors, some rights reserved; exclusive licensee American Association for the Advancement

on Septem

ber 6, 2020

http://science.sciencemag.org/

Dow

nloaded from