Energy Research
and Development
Administration
United States
Environmental Protection
Agency
Division of Biomedical
and Environmental Research
Germantown, Maryland 20767
Office of Research and Development
Office of Energy, Minerals and Industry
Washington, D. C. 20460
EPA-600/7-77-097
August 1977
PRODUCTION CYCLES IN
AQUATIC MICROCOSMS
Interagency
Energy-Environment
Research  and Development
Program Report

-------
                 RESEARCH REPORTING SERIES

 Research reports of the Office of Research and Development, U.S. Environmental
 Protection Agency, have been grouped into nine series. These nine broad cate-
 gories were established to facilitate further development and application of en-
 vironmental technology. Elimination  of traditional grouping  was consciously
 planned to foster technology transfer and a maximum interface in related fields.
 The nine series are:

       1.  Environmental Health Effects Research
      2.  Environmental Protection Technology
      3.  Ecological Research
      4.  Environmental Monitoring
      5.  Socioeconomic Environmental Studies
      6.  Scientific and Technical Assessment Reports (STAR)
      7   Interagency Energy-Environment Research and  Development
      8.  "Special" Reports
      9.  Miscellaneous Reports

 This report has been assigned to the  INTERAGENCY ENERGY-ENVIRONMENT
 RESEARCH AND DEVELOPMENT series. Reports in this  series result from the
 effort funded under the 17-agency Federal  Energy/Environment Research and
 Development Program. These studies relate to EPA's mission to protect the public
 health and welfare from adverse effects of pollutants associated  with energy sys-
 tems. The goal of the Program is to assure the rapid development of domestic
 energy supplies in an environmentally-compatible manner by providing the nec-
 essary environmental data and control technology. Investigations include analy-
 ses of the transport of energy-related pollutants and their  health and ecological
 effects; assessments of, and development of,  control technologies for energy
 systems; and integrated assessments of a wide range of energy-related environ-
 mental issues.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia  22161.

-------
                                                           LBL-5965
                  Production Cycles in Aquatic Microcosms
       A. Jassby, M. Dudzik, J. Rees, E. Lapan, D. Levy, and J. Harte
                       Energy and Environment Division
                        Lawrence Berkeley Laboratory
                          University of California
                         Berkeley, California 94720
Research supported by the U.S. Energy Research and Development Administration
and the Environmental Protection Agency D5-E681 through contract #77BCC

-------
                                 -1-
                              ABSTRACT




     Four 700-liter cylindrical containers were filled with demineralized




water, enriched with nutrients, and inoculated with 3.5-liter lakewater




samples.  The microcosms were maintained at a temperature of 18"°C under




a 12:12 L:D cycle for 6 months and several manipulations of their trophic




structure were carried out, including addition of snails (Physa sp.),




mosquitofish  (Gambusia affinis), and catfish  (Plaeostomas plaaostomas}.




Temporal variation of the phytoplankton resembled the bimodal patterns of




certain natural systems.  Further analysis demonstrated a close analogy




with the predator-prey oscillations of temperate marine waters:  an initial




bloom is terminated by zooplankton grazing; the resulting low phytoplankton




levels lead to gradual starvation of the zooplankton; and a second bloom




follows the final dieoff of zooplankton.  Both decreasing the concentra-




tion of initial nutrients or stocking the microcosms with Gambusia decreases




the time between the "spring" and "fall" blooms.   The problem of heavy




periphyton growth in microcosms was not solved with the introduction of




either Physa or Plaoostomas.   Possible solutions to this and to other




problems peculiar to microcosm research are discussed, and modifications




are suggested for  increasing the ability of microcosms to simulate




natural systems.

-------
                                -2-
                           INTRODUCTION



     The word "microcosm", when used in an ecological context,  refers




to a collection of chemicals and organisms within well-defined spatial




boundaries, generally under controlled physical conditions,  and in a




volume convenient for laboratory study,  i.e., much smaller than ecosys-




tems of interest in nature.  Unispecific cultures usually are excluded




from this definition.  A number of researchers view aquatic microcosms




as appropriate experimental objects for the investigation of systemic




properties of naturally occuring ecosystems and the delineation of



various details concerning trophic interactions, nutrient cycles, and




certain other topics (see reviews by Cooke, 1971, and Taub,  1974).




In particular, laboratory microcosms have the following desirable




properties:  (i) the small size permits replication; (ii) the chemical




composition of the medium and the trophic structure can be manipulated,




so that analogs of qualitatively different ecosystems can be created;




(iii) the lack of complicated spatial heterogeneity allows more complete




definition of physical, chemical, and biological characteristics;  (iv)




perturbations of different physical, chemical, or biological variables




can be carried out with little effort and expense; and (v) causal rela-




tionships often are more easy to deduce than in natural systems, where




uncontrolled environmental variability complicates interpretation.




    These advantages are not necessarily compelling.  A number  of draw-




backs inherent in the use of aquatic microcosms,  such as  the high surface-




to- volume ratio of the containing structure, can  be  pinpointed  on an

-------
                                 -3-
 a priopi basis.  Although  several  analogies  in biological patterns




 between  microcosms  and natural  systems  do exist  (e.g.,  diurnal  gas




 exchange,  Beyers,  1963; long-term community production, Cooke,  1967;




 nutrient cycling, Whittaker,  1961),  it  has  not yet been demonstrated




 that microcosms  can be made to  exhibit  most of the features essential




 to  seasonally- varying natural  systems  of interest, such as the typical




 succession of  phytoplankton and zooplankton,  the occurrence of both




 spring and fall  blooms of  primary producers,  etc.  Part of the  reason  for



 this lack of information has been  an undue  concern with creating




 systems  that exhibit steady-state  characteristics, a situation almost



 never observable in nature.




     Considering the potential  importance of microcosms in determining




 the macroscopic properties of natural ecosystems, as well  as the effect




 of toxic contaminants on these properties,  a need obviously exists for




 more detailed  investigation of the nature of small, synthetic aquatic




 ecosystems.  In this paper, results are presented from a study of 700-liter




 freshwater microcosms in an attempt to clarify some of the analogies




 between  laboratory microcosms and  natural water bodies, as well as to




 further define the major problems  associated with the use of micro-




 cosms in environmental research.   The microcosms differed in initial




 chemical composition and in certain features of trophic structure, and




 a variety of chemical and biological data were collected for periods of




up to 6 months.  Particular attention is paid  to those  factors motivating




the detailed design of the microcosms.

-------
                                -4-
                         MICROCOSM DESIGN




Size.  The size of freshwater microcosms employed in previous investi-




gations varies from less than 0.01 (Salt, 1971) to 200 liters (Whittaker,




1961), with the large majority less than 20 liters.  Size is critical




in at least 3 respects:  (i) smaller systems have larger container




surface-to-volume ratios, rendering surface effects more important




than in natural systems; (ii) smaller systems are disturbed to a greater




extent when samples are removed for analysis; and (iii) smaller systems



support a smaller number of trophic levels.  Accordingly, we found it




desirable to maximize the size of our microcosms within the boundaries




of the available temperature-controlled  space.  The 4 microcosms are




cylinders, 60.9 cm in radius and  75.8  cm in height, constructed of




fiberglass with a non-toxic seal.  When  filled to a depth of 60.1 cm,




the water volume is 700 liters in each tank.






Physical conditions.  The experiments  reported here were oriented toward




 an examination of microcosm behavior  in the absence of seasonal  changes




in temperature, light, and turbulence, i.e., toward patterns generated




solely as a result of the internal interactions between the various com-




ponents of the microcosms.  The systems  were maintained in a temperature-




controlled room at 19 ± 1 °C.  Each tank was illuminated by a bank of 8




4-ft high-output fluorescent lights on a 12:12 light:dark cycle.  The




water was agitated by air from a filtered  (Dayton Electric Speedaire




2Z435) laboratory supply, passing through a capillary tube 30 cm below




the water surface, at a rate of 1.2 liter rnin" .

-------
                                -5-
     No attempt was made to create a nonuniform vertical temperature




structure within the microcosms.  Creation of a thermocline would not




be a difficult matter.  However, we felt that the presence of the well-




illuminated hypolimnion of small volume that would result is so unrepre-




sentative of most naturally-occuring systems that the additional effort



was not merited.  This scaling problem undoubtedly is one of the outstan-




ding drawbacks of aquatic microcosms.  To a certain extent, the function




of the hypolimnion as a source of inorganic nutrients resulting from




decomposition processes, partially cut off from interactions with the




epilimnion, can be replaced by a porous benthic substrate that collects




sedimenting organic matter.




     There were no inflows or outflows of water during  this  experi-




ment.  Although the trophic state of an aquatic system is dependent




markedly on nutrient loading rates and hydraulic retention times




(Volenweider, 1975), the biological activity during summer stratifica-




tion of most temperate lakes appears to be determined by the concentra-




tion of dissolved nutrients already present at the onset of spring




overturn (Dillon and Rigler, 1974).  Because our first concern is with




the period of high productivity including and subsequent to the spring




bloom, justification therefore exists for setting flowthrough rates to




zero.  Water loss by evaporation  (about 1 cm each week) was compensated




for by weekly addition of demineralized water.






Benthio substrate.  The choice of substrate presented a difficult optimi-




zation problem.  If the sediments were too fine in texture, the total




particle surface area would support adsorption rates and bacterial

-------
                                 -6-
 activities that might be unnaturally high relative to the volume of




 overlying water.  On the other hand, sediments of too coarse a texture




 may preclude establishment of typical benthic macrofauna.  Because of



 the necessarily qualitative nature of these considerations,  a compro-




 mise between the 2 extremes had to be chosen without an explicit factual




 basis.  The benthic substrate consisted of a layer of river sand (sili-




 cates) 4 cm in thickness, and with a particle size ranging from 0.3 to




 3 mm.  The mean particle size was approximately 1.5 nan.




      Sediments  from natural  ecosystems  contain levels  of  organic matter




 and inorganic nutrients  partially determined by the  productivity and  depth




 of the overlying water.   When this material  is removed into  a  laboratory




 microcosm with  a smaller depth of overlying  water,  the danger  exists




 that  the   sediment  will  exert a long-term effect  on this  water surpassing




 its original effect on the parent system.  That is,  the material flowing




 from  sediments  to water  will  be diluted far  less  than  in  the parent




 system, and the resulting changes in water quality  may proceed over long




 time  periods and to an extent not found in natural  systems of  interest.




 Accordingly, we decided  to acid-wash the sand thoroughly  in  concentrated




 HC1 before use,  permitting the biological  and chemical characteristics




 of the microcosm water to determine the ultimate  organic  and inorganic




 chemical content of the  sediments.
             i





Miovooosm initiation.  Two extremes can be recognized in  the initiation




of aquatic microcosms in  the  laboratory.   The first appraoch is exempli-




fied by the systems  of Taub  (1971),  in  which defined inorganic chemical

-------
                                -7-
media under controlled physical conditions were inoculated with organisms


from pure culture  (an alga, a protozoan, and 2 bacteria).  Because the


community composition is known completely, these microcosms are referred


to as gnotobiotic  systems.  The second extreme is represented by the

                 i f
work of Beyers  (e.g., 1965), who removed portions of various naturally-


occurring aquatic ecosystems into the laboratory,  where they were  main-


tained under controlled conditions of light and temperature.  Each


approach was chosen to facilitate the attainment of different experimen-


tal goals.  The gnotobiotic approach was selected for analysis of physical


and chemical affects on steady-state community structure, and the confined


natural ecosystems for examination of diurnal community gas exchange.


     Although both designs were suited eminently for the problems under-


taken by the respective investigators, both have certain deficiencies


as general tools for the study of macroscopic properties of aquatic


ecosystesm.  The gnotobiotic systems differ from natural ones in that


the initial community lacks diversity and is synthesized by the investi-


gator, not by the natural selection of a diverse community from an even


more diverse initial assemblage.  Those properties that depend on the


existence of a large number of species with subtle properties suitable


for their coexistence thus will be lacking.


     The laboratory confinement of portions of naturally-occurring systems


entails a different set of difficulties.  The chemical and biological


structure of natural systems depends, in many ways, upon their geometry.


When samples of these ecosystems suddenly are confined, the chemical


and biological parameters inevitably change in a way not totally repre-


sentative of the parent system.  Although the qualitative behavior of the

-------
                                 -8-
 sample may resemble that of the parent for some time  after  confinement,


 providing valuable information on specific phenomena  such as  chemical


 transformations (e.g., Mortimer, 1941-42),  the  long-term behavior cannot


 be viewed with any confidence as analogous to the  parent unit or,  perhaps,


 to any natural system.  As an example, we  refer to the  discussion in the


 previous section on the possible consequences of decreasing the  depth  of


 water overlying sediments collected from some natural aquatic system.


      In parallel with certain other investigations (e.g., Maguire,  1971;


 Neill, 1975), the method of initiation that we  chose  represents  a compro-


 mise between the above two extremes.   The  4 tanks  (designated I,  II, III,


 and IV) each were filled with demineralized water  to  a  final  water volume


 of 700 liters.   The water then was enriched from stock  solutions of a


 modification of a common freshwater algae  medium (Woods Hole  MBL;


 Nichols,  1973).   Enrichment was identical  for each tank (Table 1),


 except for concentrations of inorganic phosphorous and  nitrogen.   Systems

                                                              2
 II,  III,  and IV were enriched with concentrations  of  3.0 x  10 ,  77, and


 19  ymol liter"   NaNO_, respectively.   System I  was  enriched identically to  III,


 and  all systems  had molar N:P = 16 in the  enrichment.


     Vitamins were  not added,  as we preferred that the  organisms in


each system  establish  levels  of vitamin activity reflecting their own


metabolic rates  and interactions.   Tris buffer  also was omitted.   The


levels  of tris normally used  for buffering activity (equivalent  to ca.10


mol liter    organic carbon)  far exceed the detrital carbon  concentrations


of any  freshwater system.   Most unpolluted inland  waters have organic


carbon  concentrations  of  0.2  to 3 mmol liter"  (Wetzel, ,1975).  The EDTA

-------
                                 -9-
 in our medium contributes 0.06 mmol liter"  to the detrital carbon pool


 and plays the essential role of a refractory humic-like chelating agent


 for trace elements.  The demineralized water contains an additional 0.10


 mmol liter   of organic carbon in unknown form.  Biological activity


 in the systems provides up to 0.32 mmol liter   (see below).  The organic

                                                                  _ j
 carbon in the microcosms thus ranges from 0.26 co 0.58 mmol liter  ,


 representative of many natural aquatic ecosystems.


      On the day following chemical enrichment of each microcosm, the


 systems each were inoculated with a 3,5-liter water sample collected


 from the littoral zone of eutrophic Lake Anza in the Tilden Park area


 of Berkeley, California.  Manipulation of the trophic structure was


 carried out at various times after inoculation.  These manipulations,


 summarized in Table 2, involved additions of juvenile mosquito fish


 (Gambusia affin-is}, South American catfish (Plaoostomas plaoostomas'),


 oligochaetes (Pptstina sp.), midge larvae (Tanytarsus sp.), and snails


 (Physa sp.), and were designed to fill niches not necessarily represented


 in the initial inoculum.




                         ANALYTICAL METHODS


      The following parameters were measured on a weekly basis:  tempera-


 ture,  0_, pH, inorganic carbon (1C), organic carbon (OC), NH4, NO, + N02,


 inorganic phosphorus (IP), total phosphorus (TP), phytoplankton species


 and number, and zooplankton species and numbers.  All chemical measure-


 ments  were duplicated.  Chlorophyll a_  (Chi a)  occasionally was  measured


in the microcosm water and in periphyton  scraped from the  tank  sides.

-------
                               -10-
Monitoring of phytoplankton and zooplankton continued until Day 198,




although all other analyses were terminated on Day 147.




     Methods and special instrumentation for each parameter monitored




in the microcosms are summarized in Table 3.   References to detailed




descriptions of the techniques are provided where necessary.  Vertically




integrated samples were collected for analysis at 1100 h, 3h after the




beginning of the light period, by immersing a hollow glass tube to a




depth of 5 cm above the bottom of each system.  Zooplankton were




collected by tows with a plankton bucket (Wildco) fitted with a 64 urn




straining net (Nitex).




     Total phytoplankton, protozoa, and rotifer volumes were estimated




from microscopic measurements on representatives of each species.  Crus-




tacean volumes were estimated from the numbers in various  length  classes




for each species, using the length-weight relationships of Pechen (1965)




for Daphni-a and Simooephalus and of Kelkowski and Shushkina (1966) for




copepods.  The Daphnia relationship also was  applied to Alona  and




Cypridopsis.

-------
                              -11-
                            RESULTS



     No additions of either Gambusia or Ptaoostomas were carried out



for system I, which thus contains the simplest structure of the 4



systems.  It is important to note that the biomass of Gambusia intro-



duced into systems II, III, and IV is much higher than the level that



could develop within any natural system.  The 5 fish in each microcosm



represent approximately 3 g or 4 mm  liter"  wet weight, compared to


                                            3      -1
a maximum crustacean level of less than 2 mm  liter  .  Although



crustacean production apparently was high enough to maintain the



Gambusia population (see below), the biomass of a planktivorous' trophic



level in natural systems normally is far less than the maximum biomass



of its food supply.  Any phenomena attributable to the presence of



Gambusia thus can be viewed only in a qualitative manner; the effects



in natural systems would be far less dramatic.  For this reason,



and in order to avoid unnecessary duplication in the presentation of



results for all 4 systems, we will concentrate on the details of system



I.  Only those phenomena in II, III, and TV that differ  significantly



from those in I are presented.



     On the basis of phytoplankton and crustacean biomass, the experi-



ment with system I can be divided into 3 periods of differing biological



activity (Fig. 2):(i) an initial bloom that terminates by Day 56,



(ii) a period of low phytoplankton biomass whose end is marked by the



disappearance of the cladocerans on Day 161, and (iii) a secondary



bloom beginning on Day 161.  A similar division may be applied to

-------
                               -12-




  systems  II  and  III  (Figs. 4,5), although the second interval terminates



  on  Day 140  and  198, respectively.  In IV, the distinction between an



  initial  bloom and a second interval of lower phytoplankton biomass cannot



  be  made  (Fig. 6), although the disappearance of crustaceans on Day 77



  may be taken as a natural division between the second and third period.



  Throughout  the  experiment, water temperatures remained at 18 ± 1  C and



  02  concentrations between 0.29 and 0.31 mmol liter



  (i) Initial bloom.   As exemplified by system  I,  a number of distinc-



 tive features characterize the chemical  data during the  first  interval,



 most of the parameters exhibiting a local extremum by Day 56(Fig.1) .  On



 Days 28 to 35,  pH and OC attained a maximum,  and 1C and  NO  +  N02



 a minimum for the period.   The increase  in OC of 0.18 ±  0.03 mmol



 liter   over the initial value corresponded in magnitude to the 1C



 decrease  of 0.21 ± 0.03 mmol  liter' .  The minimum NO, + NO  (64 ± 1
                                                      o     ^


 yiriol liter   below the initial level)  was accompanied by the end of a



 rapid decline in TP  and IP  from 3.8 to 0.5 ±  0.1 and 0.4 ± 0.1 ymol



 liter  ,  respectively.   By  the end  of the first  period,  NH. had



 risen to  a peak  of 23  ± 1 ymol liter  from previous values of less



 than 2 umol  liter"  .   Systems  II,  III, and IV qualitatively were



 similar in behavior, although  IV exhibited no significant rise in



 either NH or NO  + NO- after  the  initial decline in NO   + NO-.



      A well-defined sequence of plankton pulses  was observed during



 the  first  interval.  In all tanks,  the phytoplankton bloom attaining



 a maximum  on Days 28 to 35 was preceded  1 or  2 weeks by  a protozoa



peak, was  more or less  coincident with a rotifer peak, and was followed

-------
                                 -13-
 by a crustacean maximum within several weeks (Figs.  2, 4,  5,  6).

 For systems II, III, and IV, the size of the phytoplankton, protozoa,

 and rotifer maxima were in the same rank order as the initial N and P

 concentrations.  System I, which was treated in the same manner as III

 until the addition of Gambusia on Day 34, exhibited unexpectedly large


 deviations from III in levels of phytoplankton, protozoa, and rotifers


 before Day 34.


      The  initial bloom was dominated by diatoms and green algae.   In I,

 Cyclotella meneghin-iana and Oooystis sp. accounted for 55% of the

 total volume  on Day 28,  an unidentified LRGT and Oocyst-is for 84% on

 Day 35, and Synedva ulna and the LRGT for 76% on Day 42.   The first


 protozoa  peak was  dominated by Pseudomierothorax sp.  and  the initial

 rotifer peak  consisted almost solely of D-Leranophorus sp.   The crustacean

 community up  to Day 56 was formed primarily of the  cladocerans Alona guttata

 and Daphnia pulex,  although levels of the only copepod present,

 Cyclops vernalis,  also had begun to rise (Fig.  3).   As in I,  the

 local maximum of crustacean volume in II, III,  and  IV that occurred

 during the first interval  was dominated by Alona.

      Only qualitative  observations on periphytic growth and the associated


 fauna were recorded.   However,  certain phenomena observed on the  container
                                                                 L
 sides were of a dramatic nature,  occurring within a well-defined  time


interval,  and offer some insights  into the patterns developing in

the "pelagic" zone.  A light periphytic  covering in  system I  that


developed during the first 42 days evolved into a heavy growth by
                                     /'•'•,-'*-
Day 49, the same day the pH levels began  to rise a second  time (Fig.  la)

-------
                                -14-
 The growth consisted primarily of Oso-lViatovla and Cladophora sp.,



 the latter being covered by an epiphytic diatom community.   Similar



 growths appeared in II, III, and IV, although the taxonomic composition



 was highly variable; the qualitative denseness of the periphytic



 covering was in the same rank order as initial N and P concentrations.



 Tanytarsus larvae were observed on the container sides of all systems



 by Day 56.  The midge larvae fed on periphytic algae, as could be deduced



 from the presence of circular patches free of attached algal growth



 surrounding individual larvae.



 (ii) Period of low phytoplarikton biomass.   The second interval in



 system I was characterized by total phytoplankton volumes of less



 than 0.12 mm  liter" .   The NH. peak on Day 56 was followed 1 week



 later by a NO- + NO- maximum of 73 ± 4 umol liter" ,  close to the
              O     ^


 intial value of 77 umol liter"  (Fig 1).   The subsequent decline in



 NH.  and NO, + N02 was accompanied by an increase in pH to a smaller



 secondary maximum and a decrease of 1C to a smaller minimum.  No



 corresponding change in OC was observed.   Secondary peaks in NH. and



 N0_  + N02 occurred on Day 119 and 133, respectively.   Systematic



 problems  with the carbon analyzer after Day 105 prevented collection



 of accurate 1C and OC data for the remainder of the experiment.  Both



 IP and TP exhibited only erratic fluctuations at low levels after their




initial decline during  the  first interval.   The chemical behavior of



systems II and  III resembled that  of I, except that no secondary



maximum for NH4 or  NO-  +  N02 occurred in  III.   System IV exhibited



no significant  changes  in pH,  C, N, or P after the first interval.

-------
                               -15-
    Protozoa volume remained  low during the  second  interval  (fig.  2b);

3 small peaks dominated by Stpombidittm sp. were observed.  Rotifer
                                                4
levels also remained depressed, the  small community consisting of

Keratella cochlearis and  species of  Lecane,  Philodina, Tpiahotria, and
                                                             *
Voponkowia, until a final rise in  rotifers was detected  following

Day 140.  This secondary  rise consisted primarily of Polyarthra sp.,

accompanied by Keratella  quadrata  and Anupaeopsi-s sp.  Among the

crustaceans,  (Fig. 3), the major peaks for each group were segregated

in time in a clear manner.  The first peak of Alona guttata was followed

successively by Daphnia pulex, Cyclops vernal-is, Simoeephalus vetulus,

and the sole ostracod Cypvidopsis  sp.  Smaller peaks for Dftphnia occurred

before and after the maximum  for Simooephalus.  The disappearance

of cladocerans on Day 161 coincided  with the end of the  second interval.

Protozoa and rotifer volumes  in II,  III, and IV also remained at rela-

tively small values during this second interval, but only in II was

there a suggestion of a secondary  rotifer increase  (Figs. 4,5,6).  As

in I, the crustacean community was dominated by the cladocerans

Alona, Dapfmia and Simoeephalus (except for  IV, in  which Simoeephalus

was not detected).  Cyclops and Cyppidopsis  also were present, although

never in significant amounts.  The maximum levels attained for the

various crustacean species in II,  III, and IV were  in the same rank

order as the initial N and P  concentrations; the crustacean peaks  in

I exceeded those in III.

-------
                                -16-
      On Day 84,  when the pH in system I  had  ceased  to  increase  and


 NO  + NO  levels had stabilized (Fig.  1),  the  5 Physa  individuals,


 which had grown  to adult size,  were  observed for  the first  time to


 be exerting an effect on the periphytic  community:  snail movement


 on the container sides was  preserved clearly by a complex network of


 swaths free of algae.   On Day 105, a large number of tiny snails

                                                         _2
 (less than 4 mm  longest dimension, approximately  1  ind cm   of  side


 surface),  which  had hatched from egg capsules adhering to the container


 sides, were present on the  sides; on this  same day, NH.  levels  in the


 water column began a secondary increase  (Fig.  Ib).   By  Day 119, the


 sides essentially were free of algae,  aside  from  scattered patches of


 Cladophora,  and,  by Day 133,  the snail numbers had  dwindled and the


 Cladophora again were  present in dense amounts.   In system  II,


 population increases and the  subsequent  effects on  the periphyton


 community  resembled the corresponding phenomena in  I.  Although


 Physa egg  capsules  occurred on the sides of  III as  early as Day 105,


 no young snails  were observed until  Day  140  and the snail population


 always  remained  1 to 2 orders of magnitude less than in  I and II.  The


Plaeostomas  fed on the  side  growth and maintained  it at levels lower


than  in I  or  II,   although the periphyton levels remained high enough


to confer  a distinct green  color to  the  tank sides.  The introduction

-------
                              -17-
of Plaeostomas resulted  in a  shift  from unicellular to filamentous



periphytic forms.   In  IV, the Physa individuals added initially never




produced  egg capsules, but periphytic growth was removed completely



9 days  after the  introduction of Plaoostomas on Day 59 .  Side growth



never reappeared,  even when 3 of the Plaoostomas were transferred to



system  III.



     Two  Tanytarsus maxima were evident in  I, the first on Day 56



when water column concentrations of emerging larvae reached 2.0 ind



liter   ,  the second on Day 133 when levels  of 0.7 ind liter   were



recorded.  Only the first maximum occurred  in the remaining 3 systems.



In all  systems, ostracod abundance  always was far greater on and near the



sides and bottom  than  in the  water  column;  a similar comment applies



to adult  Cyclops  vernalis, but not  to the nauplius or early copepodid



stages.   Daphnia  pulex also occurred near the sides, but appeared



to prefer the open water.  Daphn-ia  ephippia consistently were present



on the sides near the water surface from Day 56 on.



(iii) Second bloom.  The disappearance of Daphnia pulex on Day 161



marked  the beginning of  a smaller secondary bloom in system I, a peak


          3        1
of 1.3  mm liter    appearing  on Day 175 (Figs. 2a,  3b).   The bloom was



due almost entirely to the cryptophyte Cpyptochpysis sp., which



accounted for more  than  85% of the  volume on Days 168 to 189.  Only



a small pulse of  protozoa was noted, mostly individuals of a Parameoium



sp., but  the rotifer community (mostly Polyarthva sp.) continued the



rise that began during the previous period.  The crustacean

-------
                               -18-
 community was  absent  completely  after  Cypvidopsis  finished  its decline




 on  Day 168.  As  in  I,  large  increases  in phytoplankton and  rotifer




 levels,  with only minor  changes  in protozoa concentrations, took




 place  after  the  beginning  of the third period in II and  III although




 the maximum  volumes attained even exceeded those of the  first period




 (Figs.  4,5].   System  IV  offers a. slightly more complicated  picture




 (Fig.  6).  Two distinct  blooms occurred during the third period,




 the beginning  of the  first on Day 77 coinciding with the disappearance




 of  the crustacean zooplankton, the second on Day 147 slightly preceding




 the disappearance of  rotifers.   The small quantities of  zooplankton




 sporadically appearing in  II, III, and IV during this third period




 all  were either  Cyclops  or Cypvidopsi,s individuals.




     No  dramatic neriphyton changes took place after the  second interval:




 I and  II retained their  heavy growths, III its lighter growth, and  IV




 remained free  of visible attachment.   All fish remained  active and healthy




 in appearance  throughout the  experiment.  For II,  III, and  IV, the




 Gombus-io. lengths increased from  11.8 ± 1.1 to 25.1 ± 4.8 mm.  There




were no  significant growth differences among systems.

-------
                              -19-
                          DISCUSSION






     The production cycle of system I  (Fig, 2a), less so that of II




and III, bears a striking resemblance to the bimodal patterns observed




in temperate marine waters  (Gushing, 1959) and in certain inland




water bodies, such as the Laurentian Great Lakes and productive




lakes of the English Lake District (Hutchinson, 1967).  Even the




interval length between the 2 peaks, approximately 5 months in the



case of I, is similar to the separation between peaks found in these




natural systems.  Despite the lack of temperature and irradiance



variations, the similarity between microcosm behavior and that of




certain naturally-occurring water bodies is marked, and the analogy




will be explored further in what follows.  As mentioned previously,




the presence of fish in II, III, and IV leads to a quantitatively




unrealistic trophic structure, so that our attention will be focused




primarily on system I.




(i) Initial period.  Conditions of Day 0 are similar to those at the




beginning of the spring bloom —  a large reserve of inorganic




nutrients and low levels of plankton biomass.  The resulting well-




defined sequence of plankton pulses is suited particularly well for




deducing the trophic relationships that characterize this initial




period.  In large natural systems, where a given trophic process




may start at slightly different times at different locations, hori-




zontal mixing tends to obscure the trophic relationships that can




be deduced from sampling at a fixed point.  It must be noted, of

-------
                              -20-
 course,  that this very mixing may influence the subsequent biological



 development of natural systems in a manner that cannot be captured



 with  the use of  isolated microcosms.



    The  initial  increase in phytoplankton productivity is retarded



 by protozoan grazing, as indicated by the larger standing crop of



 protozoa during  the first 2 or 3 weeks  (Figs. 2,4,5,6).  The subse-



 quent decline in protozoa appears to be due to grazing by rotifers



 and not  to a decrease in food supply, which continues to rise for 1



 or 2  weeks.  On  the other hand, the decline in rotifers clearly is



 related  to the decreasing phytoplankton levels, although it is not



 obvious  whether  the decline in the rotifer community occurs solely as



 a  result of overgrazing by rotifers or is aided by competition with



 the increasing crustacean community for a common food supply.  The



 fact  that the phytoplankton continue their rapid increase between



 Days  28  and 35,  when the rotifers already have attained maximum



 levels,  does suggest that the rotifer downturn is accelerated by



 competition with Alona.  Similar conclusions may be deduced for the



 remaining 3 microcosms.



      Whether by  rotifers or by crustaceans, grazing must be  the



 dominant factor  behind the termination  of the phytoplankton  bloom;



 inorganic N and  P levels observed on Day 35  (15 ±  1 and  0.4  ±0.1 umol



 liter" , respectively; Figs.  Ib, Ic) clearly are sufficient  to  support
                                          *


 further  growth (e.g., Carpenter and Guillard,1971; Perry,  1976),



while all other  required nutrients were added  in excess  (Table  1).



The fact that the second major bloom in II,  III, and  IV, when crustaceans



are absent, exceeds the first lends additional weight  tofthis viewpoint.

-------
                              -21-
    Although the "spring" bloom may have been halted by grazing,



the size of the bloom must be determined, at least in part, by the

                                                               <

quantity of nutrients initially present.  The peak phytoplankton



levels in II, III, and IV during the first period are 5.7, 0.76,



and 0.13 mm  liter  , respectively.  The protozoa and rotifer peaks



in this interval also fall in the appropriate rank order, reflecting



the relative levels of their food supply.



    The discrepancy: in magnitude between the phytoplankton peak of I



(2.4 mm  liter  ) and III is disconcerting, in view of the fact that



both systems were treated in an identical manner until the addition of



Gambusia on Day 34.  The protozoa attained a maximum of 2.0 mm  liter"


                    3      -1
in III, only 0.52 mm  liter   in I, suggesting that the lower phytoplank-



ton levels in III reflect a more intense grazing pressure from protozoa



during the first few weeks.  In turn, the discrepancy in protozoa may



result from a different protozoan composition in the initial inoculum;



although the inoculum volume was large  (3.5 liter), protozoa were



sparse and a significant random variation in the initial protozoa



levels may have resulted.  The rotifer peaks in I and III (0.12 and



0.082 mm  liter" , respectively) show less of a discrepancy than either



the phytoplankton or protozoa concentrations, probably a result of



the fact that a smaller amount of phytoplankton is compensated for by



a larger amount of protozoa in the food supply of the rotifers in III.

-------
                              -22-
      The correspondence between  the  magnitudes  of  the  1C decrease
 and the OC increase (Fig.  la)  must be  viewed  as somewhat coincidental.
 Although the 1C decrease should  equal  the  OC  increase  in a closed
 homogeneous system, aeration  and sinking of particulate organic
 matter will lower the absolute magnitude of each peak, respectively,
 and not necessarily by the same  amount.  Further evidence for  signi-
 ficant sinking losses (whether before  or after  being processed by
 grazers) is offered by the TP  (Fig.  Ic) and NO, +  NO-  concentrations
 (Fig.  Ib).
      In the case of TP, a drop of 4.3  umol liter    in  the water
 column takes place by Day 35,  although some of  these phosphorus
 losses may be due to adsorption  by the sand and tank sides and periphyton
 uptake, not necessarily to the  sinking  out  of  phytoplankton.  The rate
 constant for TP loss estimated for the first  35 days in system I is
     -1   2
 22  a   (r  = 0.81); using Vollenweider's  (1975) empirical approxima-
 tion of 10/z a   for the rate  constant, where z" =  mean lake' depth  (m),
                _ j
 a value of 18 a   is obtained  for the  microcosm.   The  agreement between
 the  2  numbers provides  some evidence that  the loss processes for TP
 in the microcosms resembles those of natural  systems.
     As  no  total  nitrogen  determinations were carried  out, the exact
 fate of the  NO,  + NO,  losses is  more difficult  to  pinpoint.  If all
of the NO, +  N09  decrease  up to  Day  35 (64 umol liter  ) represented
         •J      &
a transformation  to organic form remaining in the  water column, the
OC increase  of 0.18 mmol  liter   would imply  a  C:N molar ratio of  3
in this organic matter.  Although C:N  ratios  of this magnitude have

-------
                               -23-
been observed for marine deep water ditritus (e.g., Duursma, 1960"),



the detritus of lakes more generally is characterized by C:N ratios



exceeding 10 (Birge and Juday,  1934).  In fact, assuming a value of



 0.1 for the ratio of phytoplankton carbon to wet weight and a



 value of 3 to  6 for the molar  C:N  ratio in phytoplankton (e.g.,



 Antia et al.,  1963),  only 3 to 7 ymol liter'  N  is  required to



account for the nitrogen contained in the phytoplankton peak, a small



proportion of the actual decrease in NO  + NCL.  Accordingly, we can



conclude that a significant proportion of the NO, + N0_ decrease



results from sinking of biological material  (including that which may



have been grazed upon) onto the sediments or from uptake by periphyton.


It  is  interesting to note that  the molar N:P ratio for the lost



nutrients by Day 35 is  15, close to the mean Value traditionally



applied to phytoplankton  (e.g.,  Antia &t al.3 1963), and it is tempting



to  conclude that the decrease in NO  + N0? and TP is due primarily to



uptake by phytoplankton and periphyton, and subsequent sinking of



living and grazed phytoplankton. However, the lack of a correspondence



between N and P changes in the  intervals preceding Day 35 precludes



such a simple explanation.  Whatever the exact mechanism behind the



decrease in NO, + N0_, bacterial and zooplankton processing of nitro-
              O     «


genous organic matter leads to  a rapid building-up of NH. as the OC



decreases after Day 35; nitrification of the NH. results in the



reappearance of almost all the  NO, + N02 shortly after the start of



the second period.

-------
                             -24-
      The  large  loss in N and P and the OC increase of 0.18 mmol  liter"  ,



 when compared to the phytoplankton standing crop of approximately 20 ymol



 liter  on  Day  35  (using the C:wet weight ratio of  0.1), all suggest



 that the  maximum standing crop is only a small fraction  of the actual



 phytoplankton production during the first half of the bloom.  Assuming



 extracellular production of photosynthate is negligible  compared to



 particulate phytoplankton production, the increase in OC represents



 a minimum estimate of the quantity of phytoplankton produced up  to Day  35



 (uncorrected for respiration and sinking losses).   A minimum estimate



 for zooplankton grazing thus is obtained by subtracting  the standing



 stock of  phytoplankton, and it then appears that at least 90% of the



 particulate primary production was processed by zooplankton.



      A crude estimate of the loss rate in each interval  can be obtained



 by assuming that phytoplankton generation rates and loss rates are



 constant  within any given interval, and by choosing a reasonable value



 for generation  rates.  Because
     *p =  C°r - Cl)Xp  '                                       (1)



where x  = phytoplankton biomass (mm  liter  ),  c  = generation rate



(d  ) , and GI = loss rate (d  ) , c  can be determined from
                   i         rwi
      i = cr - ct, - 1 ) ln   xVr
               ^ 1    o      L P  ° -I
where t  and tj (d) mark the endpoints of the interval.   A reasonable



estimate of the generation rate cr can be determined from the maximum



rate of increase of phytoplankton standing crop,  usually at times when



the zooplankton are at lowest levels.   This maximum rate corresponds to a

-------
                              -25-
doubling time of  1.0 to  2.9 d  in  the  various  systems; accordingly,




we chose c^ = 0.7 d~ , equivalent to  a  doubling time of  1d .  The




resulting values  of cl for the first  period in system I  range from




0.1 to 1 d  .  The correlation between  c   and x , where  x   = zooplankton




volume   (mm  liter" ), is 0.75 (d.f.  =  8,  p < 0.05), suggesting that




the losses are due primarily to zooplankton grazing, although the




grazed material subsequently may  sink out.  Note that the correlation




is independent of the value chosen for  c  , providing that c  remains




constant for all  intervals; the fact  that  irradiance and temperature




is constant may be construed as partial evidence for this assumption.




(ii) Period of low phytoplankton  levels.   After the termination of




the spring bloom by zooplankton grazing, the  phytoplankton remain at




low levels due to continued grazing'pressure.  The continuing importance




of the crustaceans is illustrated in  Fig.  7,  which depicts the results



of removing part of the zooplankton community  with a 64-um mesh size




plankton bucket from a 4-liter sample of system I.  After a delay of




several days, a rapid increase in fluorescence takes place with respect




to a control beaker in which the  zooplankton  are left undisturbed.




When the zooplankton are removed  similarly from the control beaker




on the ninth day and transferred  to the beaker'containing no zooplank-




ton, fluorescence rises in the  control beaker and decreases in the




other.  Apparently, the generation rate of phytoplankton and the




grazing pressure of zooplankton remain more or less in balance during




the second period.

-------
                             -26-
     Further evidence that the production losses after the bloom



reasonably may be attributed to zooplankton grazing is obtained by


                                                -1-1
consideration of the quantity  c,/x'  (liter ind   d  ), where x^ =
                                _L  Z                            Z


zooplankton number (ind  liter~  ).   This  quantity is  the  filtering



rate that would be required to explain production losses in terms of



crustacean grazing pressure alone.  The mean value for Days 56 to 154,



that portion of the second period when crustaceans dominate the



zooplankton  (Fig. 2b), is 0.01 ± 0.01 liter ind~  d~ .  Lowest values



of 0.002 liter ind d~  are obtained when Cyclops vernalis is at its



maximum; intermediate values of  0.007 ± 0.004  liter ind   d   when



Alona guttata and Daphnia pulex are dominant; and highest values of



0.02 ± 0.02 liter ind   d   when Simooephalus vetulus is the dominant



species.  For the low food concentrations during this period (<10



cell liter  ), these values correspond to the results of various



zooplankton feeding studies (e.g., Wetzel, 1975, Table  16-8, for



copepods; Infante, 1973, for D. pulex; Sushtchenia,  1958, for  5.



vetulus}.



     Because the magnitude of the peaks for all 5 zooplankton  species



in systems II, III, and IV fall in the same rank order  as initial  N



and P levels, the peak sizes clearly  are dependent upon food supply



in the microcosms.  The limiting nature of the  food  supply  is  suggested



also by the presence of cladoceran ephippia in  all  4 microcosms after



the initial bloom.  The larger zooplankton peak sizes in  I  compared



with III imply that fish planktivory  also is a  factor in  determining

-------
                              -27-
the maximum development of zooplankton species in these systems.


Simooephalus, a large zooplankter and presumably one of the most

                                                                  f
susceptible to Gambusia feeding, was absent completely from IV, the


least productive microcosm.  In addition, a second Tanytapsus maximum


occurred only in I, where fish were absent.


     The clear temporal separation between the zooplankton peaks in


system I (Fig. 3) is probably not just a reflection of life cycle


timing.  In particular, the manner in which the Daphnia peaks are      ,


interspersed between those for the remaining species suggests signi- /


ficarit competition between members of the zooplankton community.


However, it is difficult to deduce the exact causal mechanisms that


lead to the temporal segregation of zooplankton peaks in Fig. 3.


Damped  oscillatory populations often are observed in unispecific


Daphnia cultures  (Pratt, 1943; Slobodkin, 1954), and frequently the


period of  oscillation is approximately 40 d.  One way to view the D.


pulex data is to hypothesize a series of 3 damped oscillations


generated  as a result of interactions between Daphnia and its food


supply, with peaks that would have occurred around Days 50, 90, and


130 in the absence of other crustaceans.  The presence of Alona


suppresses the first peak, and that of Simoaephalus the second; Cyclops


and Cyppidopsis then take advantage of those times when Alona and


Daphnia are suppressed.  The exact cause-effect relationships undoubtedly


are more complicated.  For example, the raptorial food habits of Cyclops


may initiate or accelerate the depression of Daphnia after Day  63.   In

-------
                               -28-
  any case, the value of this microcosm design for competition studies is


  clear; many of the details of competition phenomena probably would


  reveal themselves  by using an inoculum for  the microcosms  in  which  the



  crustacean composition was manipulated artifically.



      The rise in pH and the decrease  of 1C and NO,  +  NCL  during  the


 second period (Fig.  1)  appear  to  result from periphyton photosynthesis,


 as  evidenced by an  increase in periphyton density from Days  49 to


 84.   Periphyton scraped with a razor  blade off small  areas  of  the

                                                 _2
 tank sides revealed densities  as high  as 60 mg m   Chi a_ by Day 84,


 equivalent to 38 ug liter    Chi a_ if  dispersed throughout the  water.


 Chi  a_ measurements  occasionally collected in the water column  during


 the  second period never exceeded  4 ug liter" , so that the  phytoplankton


 biomass  was far less  than  the  periphyton biomass by Day 84.   It  is of


 some interest to estimate  the  maximum density that  can be attained


 by periphyton packed  on the tank  sides.  Assuming that the  thickness


 of the periphyton layer cannot exceed the distance  for which incident



 irradiance would be reduced to compensation  levels  of irradiance, the


 following  relationship  must be satisfied:
     Ie - V           ,                                         (3)


                                         _2
where  I  = compensation irradiance  (W m   PAR) ;   I   =  incident  irra-
        d-                                          O

           _2
diance (W m   PAR) ;  k  = specific extinction  coefficient  for Chi a
                      C                                            """"

  2-1                                               -3
(m  mg  ) ;  C = Chi a. concentration  on the  sides  (mg  m  ) ;  and  d = thick-



ness of periphyton layer (m) .  The mean  of   I   incident on the  sides is


                    -2                                     2   -1
approximately  7 W m   PAR.  Using a value  of   k   = 0.02 m  mg    and
                                                I*

          _2
I  = 1 W m   PAR (Platt and Jassby,  1976),  an  areal Chi a. concentration

-------
                               -29-
                  _2
(=Cd) of 100 mg m    is obtained, close to the value observed.  The


periphyton thus appear to have built up to their maximum light-limited


density.  The same argument does not apply to forms capable of developing


filaments, which avoid light limitation by extending into the water


column where individual filaments can avoid significant shading effects.


     The consequences for studies of nutrient cycles and trophic


dynamics in microcosms are overwhelming.  Once the periphyton are


established, temporal changes in nutrient concentrations (such as


the second NCL decrease illustrated in Fig. Ib), no longer can be


attributed to the behavior of planktonic organisms alone.  In addi-


tion, the explanation of population fluctuations in organisms that


partition their time between the "littoral" and "pelagic" zones  (such


as observed for Cypridosis,  adult Cyolops, and Daphnia) requires


more  complicated considerations of food supply.  Although the littoral


zone  fulfils a similar complicating role in natural systems, the produc-


tivity  of non-phytoplanktonic vegetation exceeds phytoplankton produc-


tivity  by an order of magnitude or more only in small, shallow lakes


 (e.g.,  Wetzel, 1975, Table 15-15).  The periphyton problem thus


limits  severely the use of microcosms as more general analogs of  inland


water bodies or for  investigation of purely planktonic relationships.


     The introduction of Physa and Plaaostomas into the microcosms


represents an attempt to decrease the periphyton abundance by herbivory.


In the  case of Physa, the initial  individuals  grew continuously  but


did not make their presence felt until Day 84, when visible effects of


their grazing coincided with the second decrease in pH  (Fig.  la).  The


resulting decrease in periphyton production apparently permitted aeration

-------
                              -30-
 to  return the pH to its initial level.  The appearance of large




 populations of young snails on Day 105 coincided with the second NH.




 increase, presumably because of snail excretion of grazed periphyton



 nitrogen.  The snails subsequently died from overgrazing after Day 119




 and the periphyton quickly reappeared.  The increase in pH accompanying




 the reappearance of periphyton is suppressed possibly because of CO™




 production from the decomposition of starved Physa.   The  lower concen-



 trations of young Physa in III, compared with I and II, can be




 attributed to Plaeostomas feeding on snail egg capsules attached  to



 the sides, and no secondary NH  peak from snail excretion was observed.




 In  I, productivity was not sufficient to support Physa reproduction.




 Because of the time lag between periphyton growth and snail reproduc-



 tion, and because of overgrazing, the snails generally proved incapable




 of  maintaining sides consistently free of periphyton.  Accordingly,




 we  do not feel that introduction of snails offers even a partial




 solution of the periphyton problem.  In fact, their short-term effects




 on  the pelagic N concentrations make interpretation of nutrient




 cycles even more complicated.




     The presence of Plaoostomas in IV proved to be completely




 effective in removing periphyton from the sides and preventing recoloni-




 zation.  However, system IV had the lowest  level of nutrients and the




 least dense periphyton growth.  In III, where initial N  and P  levels




were 4 times higher, the 3 Plaoostomas never succeeded in eliminating




periphyton growth, even with the aid of the snails present.   Although

-------
                             -31-
growth was always less dense than in I or II except when the snails

temporarily had eliminated the periphyton in these latter 2 systems,

the color suggested that Chi a_ densities were not an order-of-magnitude

lower than in I or II and probably still exceeded phytoplankton biomass.

The Placostomas appeared to be particularly ineffective against the

long Cladophora strands that developed in III during the third period.

     Accordingly, neither the snails nor catfish turned out to be an

adequate method for dealing with dense periphyton communities.  This

difficulty still remains unsolved and is the most serious hindrance

to long-term studies in microcosms.
                                       1
(iii) Second bloom.  In all 4 microcosms, the beginning of the second

bloom coincides precisely with the disappearance of crustaceans from

the water column.  The release from grazing pressure permits the phyto-

plankton community to take full advantage of the nutrients available

and increase to levels approaching or exceeding those of the initial

bloom.  The timing of the crustacean dieoff reflects the effects of

both food supply and predation.  In systems II, III, and IV, the

dieo'ff is postponed more when the initial N and P levels are increased;

in I, the lack of a planktivorous fish allows the intermediate

period between blooms to continue for the longest time of all 4 systems,

The bloom starting on Day 77 in IV appears to be prematurely destroyed

by the rotifer community, and it is only when the rotifers disappear

between Days 147 and 154 that a major phytoplankton rise takes place


(Fig.  6).

-------
                             -32-
    The termination of the "spring" bloom by grazing, and the onset




of the "fall" bloom after starvation of the zooplankton, both suggest




that the production cycles in these microcosms reflect a predator-




prey oscillation  (although not necessarily of the Lotka-Volterra




type).  The oscillations are forced by the initial conditions favoring



high phytoplankton growth rates with little interference from crustacean




grazers.  At a later time, the crustaceans in the inoculum spawn and




the grazing capacity of the zooplankton community increases.  Subsequent




grazing mortality among the phytoplankton results in primary produc-




tion levels too low to support the zooplankton, whose levels gradually




decrease through starvation.  Once the cladocerans, in particular, have




disappeared, conditions resemble those at the start and a new sequence




of production begins.




    The course of events is similar, in many ways, to that observed in




certain temperate marine waters where production cycles essentially




are a reflection of predator-prey relationships, and nutrient levels




respond to, rather than cause, the cycle (see discussion by Gushing,




1975).   As in the microcosms, the quantity of phytoplankton produced




is many times the maximum standing crop, and about 1 month is required




before crustacean grazing becomes effective enough to terminate algal




increases in the spring.  Although tropical and polar production cycles




also can be viewed as resulting from predator-prey interactions  (Gushing,




1959),  latitudinal differences in the annual temperature pattern result




in differences in the annual production pattern.  In arctic waters, the

-------
                             -33-




                                                                  \   ,   -



colder temperatures result in a longer delay time between the onset of
                                                                       , j


spring phytoplankton increases and the onset of effective grazing, and



only a single phytoplankton peak of high amplitude occurs before produc-



tion is halted by winter conditions.  In tropical waters, the short



delay time results in low amplitude phytoplankton changes without the



discontinuity characteristic of higher latitudes.  In general, the



longer the delay time, the greater the phytoplankton peak (up to the



onset of nutrient limitation) and the greater the proportion of phyto-



plankton biomass lost to sinking and respiration without being grazed.



Thus, the delay time between phytoplankton increases and the beginning



of effective grazing is an extremely important factor governing the



amplitude of production cycles and the ratio of secondary to primary



production.  The use of microcosms offers a unique opportunity for



exploring the precise conditions under which different production



patterns result; irradiance, temperature, initial nutrient levels,



and  initial  (i.e., "overwintering") levels of crustacean females or



resting eggs all can be manipulated to analyze effects on delay times



and  subsequent levels of primary and secondary production.



     The analogy between the microcosms and inland water bodies is less



general.  Evidence for predator-prey oscillations dominating production



cycles is clear only in the simplified associations of extreme habitats,



such as highly saline or alkaline lakes  (Anderson et at., 1955; Anderson,



1958).  In most well-studied cases, nutrient depletion is a more  signifi-



cant factor than grazing in causing the spring decline of phytoplankton;

-------
                             -34-
 for  example,  silicate depletion plays this role in parts of the English

 Lake District (Lund  et al-, 1963), and phosphate depletion in certain

 Laurentian  Great  Lakes (Schelske  et al. , 1972).  In addition, upwelling

 of subthermocline nutrients or algae  (Fee, 1976) during the fall overturn,

 rather  than zooplankton disappearance, appears to be the main cause of

 the  fall bloom.

     The possibility  of a more realistic simulation of inland waters

 with microcosms cannot be  ruled out.  As pointed out previously, protozoa

 grazing is  a major factor  in suppressing initial phytoplankton increases

 in the  microcosms.   In natural systems, protozoa usually constitute

 only a  minor portion of the zooplankton and achieve maximum levels

 after the spring bloom.   (e.g., Schonborn, 1962; Sorokin and Paveljeva,

 1972).  Initial microcosm  conditions  apparently allowed the inoculated

 protozoa to rapidly  outgrow predators and exert an undue effect on the

 spring  bloom.  It is possible that, if protozoa could be excluded from

 the  inoculum, phytoplankton increases would proceed at a rate sufficient

 to exhaust  nutrients before crustacean grazing became effective enough

 to terminate  the bloom.  In addition, certain aspects of the fall

 overturn perhaps could be  simulated by vigorous mixing of the benthic

 substrate into the water column for a short time, releasing the accumu-

 lation of nutrients  from decomposed organic matter and returning viable

algae that  have settled out.  Neither modification was attempted in  the

present set of experiments, but both  bear serious consideration for
            /
any future  work of this nature.

-------
                             -35-
                         CONCLUSIONS



    A number of important drawbacks deserving further discussion have



been made clear to us in the course of this work.  Some of these draw-



backs are inherent difficulties in size scaling.  For example, the



shallow depth results in higher sinking losses, smaller vertical



irradiance changes, smaller migration distances for zooplankton, etc.,



as compared to natural systems.  Little can be done about such scaling



problems and their possible interference must be interpreted within



the context of each individual experimental aim.



    Other major problems, however, do have potential solutions:



(i) It is not possible to stock planktivorous fish in realistic concentrations,



and their presence results in premature disappearance of the cladocerans.



The obvious answer to this difficulty is the complete exclusion of fish



from the microcosms.  Their exclusion also permits the use of smaller



systems, although proper zooplankton sampling sets a lower limit on


                                             2
microcosm volumes somewhere between 10 and 10  liters.



(ii) Poor replication will result unless all inoculated organisms that



can proliferate in the microcosms are present in sufficiently high



numbers in the inoculum; also, protozoa appear to play an unnaturally



important role.  Both of these problems suggest the use of gnotobiotic



systems, although experiments then are subject to the lack of realism



inherent in the gnotobiotic approach (see section on microcosm design).



Perhaps a compromise is suitable, in which the inoculum is prepared



from several unialgal (although not axenic) cultures, along with, say,



1 rotifer species and 1 crustacean species (with large numbers of indivi-



duals in each case).  Greater ability to exclude contamination by other

-------
                             -36-
 zooplankton will be obtained in the smaller systems without fish.


 (iii)  Periphyton growth is too high and can dominate events in the


 microcosms.  The periphyton problem can be dealt with in part by limiting


 experiments to  less than about 1.5 months, the time when the consequences


 of side  growth  first become apparent.  This limited time still permits


 a detailed examination of the circumstances surrounding the spring bloom.


 An alternative  method is mechanical removal of periphytic attachment,


 although daily  manual scraping of the tank sides is too tedious for


 routine  work and an automated method is desirable.  In any case, the


 solution of the periphyton problem deserves the most serious considera-

                                                      •
 tion in  any further work of this nature.  The dominance of biomass by


 side growth precludes the simplified study of nutrient cycles and


 energy flow that microcosms potentially can offer.


    Despite these drawbacks, this study has demonstrated certain resem-


 blances  between microcosms and natural water bodies where production


 cycles are determined by predator-prey relationships*  Also, modifica-


 tions have been suggested to increase the generality with which these


microcosms can  be made to simulate natural systems.  Because of the


resemblance to  certain natural production cycles, and because of the


clear sequence  of plankton pulses in the absense of spatial heterogeneity,


we feel  that the microcosm method described here offers a powerful


tool for assessing environmental effects  (whether of human origin or


not) on production patterns, trophic relationships, and competition


phenomena.

-------
                                  -37-
                              REFERENCES
Anderson, G.C. 1958,  Seasonal characteristics of two saline lakes
   in Washington. Limnol. Oceanogr.  3_:51-68.

Anderson, G.C., G.W. Comita, and V.  Engstrora - Heg.  1955.  A note on
   the phytoplankton-zooplankton relationship in two lakes in Washing-
   ton. Ecology 36_:757-759.

Antia,  N.J., C.D. McAllister, T.R.  Parsons, K. Stephens, and J.D.H.
   Strickland. 1963.  Further measurements of primary production using
   a large-volume plastic sphere. Limnol. Oceanogr. 8^:166-183.

APHA. 1971.  Standard methods for the examination of water and waste-
   water. American Public Health Assciation.

Beyers, R.J. 1963. A characteristic  diurnal metabolic pattern in
   balanced microcosms.  Publ. Inst. Mar. Sci., Texas 9^:19-27.

Beyers, R.J. 1965.  The pattern of photosynthesis and respiration
   in laboratory microecosystems.  Mem. 1st. Ital. Idrobiol. 18
   suppl.:61-74.

Birge, E.A., and C. Juday. 1934.  Particulate and dissolved organic
   matter in inland lakes.  Ecol. Monogr. 4_:440-474.

Carpenter, E.J., and R.R.L. Guillard.  1971.  Intraspecific differences
   in nitrate half-saturation constants for three species of marine
   phytoplankton. Ecology 52:185-185.

Cooke, G.D. 1967.  The pattern of autotrophic succession in labora-
   tory microcosms. Bioscience  17:717-721.

Cooke, G.D. 1971. Aquatic laboratory microsystems and communities,
   p.48-85. In J. Cairns (ed.), The  structure and function of freshwater
   microbial~cbmmunities. Virginia Polytechnic Institute and State
   University.

Gushing, D.H. 1959.  On the nature of production in the sea.  Fish.
   Invest. Lond. (Ser. 2) 22; 40 pp.

Gushing, D.H. 1975.  Marine ecology  and fisheries. Cambridge Univ.

Dillon, P.J., and F.H. Rigler. 1974.  The phosphorus-chlorophyll
   relationship in lakes. Limnol. Oceanogr. 19:767-773.

-------
                                  -38-
Duursma, E.K.  1960. Dissolved organic carbon, nitrogen, and phosphorus
    in the sea. J,B. Wolters, Groningen,

Fee, E.J. 1976. The vertical and seasonal distribution of chlorophyll
    in lakes of the Experimental Lakes Area, northwestern Ontario:
    implications for primary production estimates. Limnol. Oceanogr.
    2^:767-783.

Golterman, H.L. 1969. Methods for chemical analysis of fresh waters.
    Blackwell Scientific.

Guillard, R.R.L.  1973.  Division rates, p. 289-311. In J;R. Stein  (ed.)
    Phycological Methods. Cambridge Univ.

Hutchinson, G.E.  1967. A treatise on limnology, vol. 2. Wiley.

Infante, A. 1973. Untersuchungen uber die ausnutzbarkeit verschiedener
    Algen durch das Zooplankton. Arch. Hydrobiol. Suppl. 42:540-405.

Klekowski, R.Z.,  and E.A. Shushkina. 1966. Ernahrung, Atmung, Wachstrum
    und Energie-Umformung in Macrooyclops al'b'idus Jurine. Verh.  Int.
    Verein. Limnol. 16:399-418.

Lund, J.W.G.,  F.J.H. Mackereth, and C.H. Mortimer.  1963. Changes in
    depth and time of certain chemical and physical  conditions and  of
    the standing crop of Aster-ionella fovmosa Hass.  in the North Basin
    of Windermere  in 1947. Phil. Trans. Roy, Soc. London  (Ser. B)
    246:255-290.

Maguire, B. 1971. Community structure of protozoans and algae with
    particular  emphasis on recently colonized bodies of water, p.121-149-
    In J. Cairns (ed.), The structure and function of freshwater
    microbial communities. Virginia Polytechnic  Institute and State
    University.                            t

Mortimer, C.H. 1941-42. The exchange of dissolved substances between
    mud and water  in lakes. J. Ecol. 2£: 280-329; 30:147-201.

Neill, W.E. 1975. Experimental studies of microcrustacean competition,
    community composition and efficiency of resource utilization.
    Ecology 56_:809-826.

Nichols, H.W.  1973. Growth media-freshwater, p.7-24. In_ J.R. Stein (ed.).
    Phycological methods.  Cambridge University.

Pechen, G.A. 1965. Produktsiya vetvistousykh rakoobraznykh  ozernogo
    zooplankton. Gidrobiol. Zh. \\19-26

-------
                                  ^39-
Perry, M.J. 1976. Phosphate utilization by  an oceanic diatom in phos-
   phorus-limited chemostat culture and in  the  oligotrophic waters of
   the central North Pacific. Limnol. Oceanogr.  21:88-107.

Pratt, D.M. 1943. Analysis of population development in DapTmia at
   different temperatures. Biol.  Bull. Mar.  Biol.  Lab., Woods Hole,
   85:116-140.

Salt, G.W. 1971. The role of laboratory experimentation in ecological
   research, p.87-100.  In J. Cairns  (ed.),  The  structure and function
   of freshwater microbial communities. Virginia Polytechnic Institute
   and State University.

Schelske, C.L., and E.F. Stoermer. 1972. Phosphorus, silica, and
   eutrophication of Lake Michigan, p.157-176.  In_  G.E. Lidens (ed.),
   Nutrients and eutrophication.  Special Symposium, Amer. Soc.
   Limnol. Oceanogr.

Schonborn, W.  1962. Uber Planktismus und Zyklomorphose bei Difflugia
   lirmetiea (Levander) Penard.  Limnologies. 1^:21-34

Slobodkin, L.B. 1954. Population dynamics in Daphnia dbtusa Kurz.
   Ecol. Monogr. 24:69-88.

Solorzano, L.  1969. Determination of ammonia in natural waters by the
   phenol-hypochlorite  method,   Limnol. Oceanogr.  14:799-801.
Sorokin, Ju, I., and E.B. Paveljeva.  1972.  On the  quantitative
   characteristics of the pelagic ecosystem of  Dalnee Lake (Kamchatka).
   Hydrobiologia 40:519-552.

Strickland, J.D.H., and T.R. Parsons. 1968.  A practical handbook of
   seawater analysis. Fish. Res.  Bd. Canada, Bulletin 167.

Sushtchenia, L. 1958. Kolichestvennye dannye o_fil'tratsionnom
   pitaniiplanktonnykh  rachkov.  Nauch. Dokl. vyssh. shk,, Biol.
   Nauki 1:241-260.

Taub, F.B. 1971. A continuous gnotobiotic (species defined) ecosystem,
   p. 101-120.  In J. Cairns  (ed.),  The structure and function of
   freshwater infcrobial communities. Virginia Polytechnic Institute
   and State University.

Tuab, F.B. 1974. Closed ecological systems.  Ann. Rev. Ecol. System.  5_;139-160.

Vollenwieder, R.A. 1975. Input-output models, with special reference to
   the phosphorus loading concept in  limnology.  Schweiz. Z. Hydrol.  37:53-84.

Wetzel, R.G. 1975. Limnology. Saunders.

Whittaker, R.H. 1961. Experiments with radio-phosphorus tracer  in
   aquarium microcosms. Ecol. Monogr. 31:157-188.

-------
                               -40-
                           TABLE HEADINGS
Table 1.  Composition of the inorganic enrichment medium for the 4 700-
          liter microcosms.   The stock solution is diluted 1:700 in the
          microcosms.

Table 2.  Manipulation of the biological structure in the 4 700-liter
          microcosms.

Table 3.  Methods of analysis for the parameters monitored.

-------
                             -41-
Table I.
Nutrient
"'
CaCl.«2H_0
2 2
MgS04-7H20
NaHC03
Na0HPO -7H.01
242
NaN031
Na0SiO'9H00
232
KC1
Na2EDTA
FeSO.-7H00
4 2
CuSO.-SH-O
4 2
ZnSO.-7H00
4 2
CoCl0-6H00
2 2
MnCl -4H.O
2 2
Na0MoO.'2H00
242
Stock solution
(g liter'1)
37

37
37


--
30

10
1.5
1.0

0.010

0.020

0.010

0.20

0.010

Microcosm concentrations
(Vimol liter- 1)
3.6 x 102

2.1 x io2
6.3 x io2


—
1.4 x IO2

2.0 x IO2
5.7
5.3

0.057

0.10

0.060

1.4

0.060

see text

-------
                                       -42-
  Date      Day                           Manipulation
05-06-76    -  0     Inoculation of each of I,  II,  III,  IV'with  3.5-liter  sample  of  lakewater




05-21-76     15     Addition to each of I, II, III,  IV of 5 Pristina and  5 Tanytarsus larvae




06-09-76     34     Addition to each of II,  III, IV of  5 Gambusia of  length 1.2  cm




06-24-76     49     Addition to each of I, II, III,  IV of 5 Physa of length 2.5-5.Omm




07-04-76     59     Addition to IV of 4 Placostomas  of length 1.0 cm




08-15-76    101     Transfer of 3 Placostomas  from IV to III.

-------
Table  III.
  Parameter
          Method
         Special equipment
       Reference
°2   ..
pH
1C
oc
NH4
N03 + N02
IP
TP
Chi a
phytoplankton
zooplankton
polarography
electrometry
infrared absorbance
combustion to  1C
blue  indophenol reaction
reduction, diazotization
ascorbic acid  reduction
persulfate digestion  to  IP
fluorometry
Sedgewick-Rafter  cell
r64-urn tow
02 meter (YSI 57)
pH meter (Orion 601)
IR analyzer (Beckman 865)
TOC analyzer (Beckman 915A)
spectrophotometer (Zeiss PM2 DL)
fluorometer (Turner 111)
phase microscope (Reichert Zetopan)
dissecting microscope (AO 570)
Solorzano, 1969
Golterman, 1969
APHA, 1971
APHA, 1971
Strickland and Parsons, 1968
Guillard, 1973
                                                                     i
                                                                     -o
                                                                     u>

-------
                               -44-
                           FICURE LEGENDS
Figure 1.  Chemical measurements from System I.  Vertical lines indicate
           the standard error for duplicate measurements.  (a) pH, inor-
           ganic carbon (1C), and organic carbon (OC); (b) NH. and
           NOg + NC>2; (c) inorganic phosphorus (IP) and total phosphorus
           (TP).

Figure 2.  (a) Total volume of phytoplankton in system I; (b) total
           volume of protozoa, rotifera, and Crustacea in system I.

Figure 3.  (a) Resolution of the Crustacea in Fig.  2b into cladocera,
           copepoda, and ostracoda; (b) resolution of the cladocera in
           Fig. 3a into component species.

Figure 4.  (a) total volume of phytoplankton in system II; (b) total
           volume of protozoa, rotifera, and Crustacea in system II.

Figure 5.  (a) Total volume of phytoplankton in system III;  (b) total
           volume of protozoa, rotifera, and Crustacea in system III.

Figure 6.  (a) Total volume of phytoplankton in system IV; (b) total
           volume of protozoa, rotifera, and Crustacea in system IV.

Figure 7.  Importance of zooplankton grazing in 4-liter samples collected
           from system I on Day 63.  Removal of zooplankton with 64 um
           net (solid line) results in increased fluorescence with respect
           to a control beaker (dashed line).  When zooplankton are
           transferred from the control beaker to the beaker that is
           zooplankton-free, fluorescence trends reverse.

-------
                           Figure I

 o


 £
 c
 o
 o
 O
 o>
 o
 £
120



 80




 40
     0
    6.0
'    4.0
 i_
 (D


_  2.0
 o
 £
o.o
                                                     (b)
                                                     (c)
                   _L
                                                         9.0
                                                         8.0
                                                         7.0
      0     20   30    60    80    100   120    140    160

                        Time   (d)

                                               XBL77I - 120

-------
                          J*6-
                        Figure II
            Phytoplankton
              volume
0    20
60        100
      Time  ( d )
                                               XBL77I- 119

-------
  US EPA Headquarters Library
    '    Mail code 3404T
   1200 Pennsylvania Avenue NVV
                                   -1*7-
                                 Figure III

-------
                               Figure IV
     4.0
 ro
  £  3.0
  £
  CD  2.0

 ^
  o
 >  1.0
    0.0
    7.0
    6.0
    5.0
          Phytoplcmkton
            volume
       ~ I      I
       Protozoa
         *
 
-------
                                 -1*9-
                               Figure V
 o>
10
6.0
    5.0
    4.0
                     Phytoplankton  volume
 e  3.0
 E
 o>  2.0


 ~o
 >  1.0
   0.0
                                                               (a) _
            20    40    60     80   100    120   140   160    180  200

                               Time   (d )

                                                          XBL77I - 183

-------
                                   -50-
                                 Figure VI
    4.0
10
 6  iO
 e

 o>  2.0
 o
 >  1.0
    0.0


    50
   2.0
 E

"o
>  1.0 - i
   0.0
         Protozoa
                Rotifera x 10
                                     Phytoplankton volume
            20     40    60    80     100   120

                                   Time  (d)
               (b)  -
              I
140    160   180   200
                                                             XBL77I- 184

-------
  100


^-«»

en


"i so
   60
                        T      i       r

                             Zoop added
   40
o>
o



120
CD
u.
O
           1	^

        Zoop removed
                            Zoop  removed
    0
                        6      8

                        Time   (d)
10
12
14
                                          XBL77I-I82
                                                       H-




                                                       H I-1

                                                       (D I

-------