Anaerobic digestion of organic matter

Download Anaerobic digestion of organic matter

Post on 16-Mar-2017

216 views

Category:

Documents

1 download

TRANSCRIPT

  • This article was downloaded by: [Cornell University Library]On: 13 November 2014, At: 22:20Publisher: Taylor & FrancisInforma Ltd Registered in England and Wales Registered Number: 1072954Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

    C R C Critical Reviews inEnvironmental ControlPublication details, including instructions for authors andsubscription information:http://www.tandfonline.com/loi/best18

    Anaerobic digestion of organicmatterP. N. Hobson a , S. Bousfield a , R. Summers a & E. J. Kirsch ba Rowett Research Institute , Bucksburn, Aberdeen, Scotlandb Purdue University , West Lafayette, IndianaPublished online: 09 Jan 2009.

    To cite this article: P. N. Hobson , S. Bousfield , R. Summers & E. J. Kirsch (1974) Anaerobicdigestion of organic matter, C R C Critical Reviews in Environmental Control, 4:1-4, 131-191,DOI: 10.1080/10643387409381614

    To link to this article: http://dx.doi.org/10.1080/10643387409381614

    PLEASE SCROLL DOWN FOR ARTICLE

    Taylor & Francis makes every effort to ensure the accuracy of all the information(the Content) contained in the publications on our platform. However, Taylor& Francis, our agents, and our licensors make no representations or warrantieswhatsoever as to the accuracy, completeness, or suitability for any purpose of theContent. Any opinions and views expressed in this publication are the opinions andviews of the authors, and are not the views of or endorsed by Taylor & Francis. Theaccuracy of the Content should not be relied upon and should be independentlyverified with primary sources of information. Taylor and Francis shall not be liablefor any losses, actions, claims, proceedings, demands, costs, expenses, damages,and other liabilities whatsoever or howsoever caused arising directly or indirectly inconnection with, in relation to or arising out of the use of the Content.

    This article may be used for research, teaching, and private study purposes. Anysubstantial or systematic reproduction, redistribution, reselling, loan, sub-licensing,systematic supply, or distribution in any form to anyone is expressly forbidden.Terms & Conditions of access and use can be found at http://www.tandfonline.com/page/terms-and-conditions

    http://www.tandfonline.com/loi/best18http://www.tandfonline.com/action/showCitFormats?doi=10.1080/10643387409381614http://dx.doi.org/10.1080/10643387409381614http://www.tandfonline.com/page/terms-and-conditionshttp://www.tandfonline.com/page/terms-and-conditions

  • ANAEROBIC DIGESTION OF ORGANIC MATTER

    Authors: P. N. HobsonS. BousfieldR. SummersRowett Research InstituteBucksburn, AberdeenScotland

    Referee: E. J. KirschPurdue UniversityWest Lafayette, Indiana

    I. INTRODUCTION

    The treatment of waste products so as toreduce them in bulk and render them less obnox-ious is becoming more and more necessary inpresent-day society. Some wastes pose almostinsurmountable problems of disposal; radioactivewastes may remain dangerous for thousands ofyears. Other dry wastes, such as power station ash,may be relatively easily disposed of as land-fill orbuilding materials. Still other dry wastes may bereconstituted, as are paper, rags, and glass, intofeedstock for producing more of the same articles.Chemical wastes may be reacted to render theminnocuous, or burnt along with other wastes. Insome cases proper design of furnaces has led to areturn from burning waste in the form of heatwhich can be usefully employed. But there stillremains probably the largest fraction of the totalwaste output which is difficult or impossible totreat by physical methods. This is organic materialof animal or vegetable origin. The high watercontent of most of this material makes incinera-tion impossible or uneconomic; separation of theliquid and solid fractions can be difficult and thesefractions in themselves are still polluting. The

    water content makes transport uneconomical, andthis, and the fact that the waste is subject toalmost immediate attack by microorganisms, makeon-site treatment necessary. Since the pollutionalproperties of these wastes are due in large measureto their being good substrates for microbialgrowth, they can be treated by controlled actionof microorganisms.

    The growth of microorganisms can take placeeither aerobically or anaerobically. A larger rangeof substrates may be degraded aerobically thananaerobically and even many chemical wasteswhich in concentrated form are corrosive orbactericidal may, when in dilute solution, betreated microbiologically by aerobic methods.Complete action of microorganisms in the pres-ence of excess oxygen, and on a substrate balancedin carbon, nitrogen, and minerals theoreticallyresults in carbon dioxide, water, and microbialcells. Because of imbalances in the waste substratessuch complete conversion is seldom attained, butin a properly designed plant the pollutional load ofwaste waters may be reduced to acceptable levels.However, of more importance than imbalance ofsubstrates is the question of excess oxygen. Thedifficulties of supply of oxygen to the microbes

    June 1974 131

    Dow

    nloa

    ded

    by [

    Cor

    nell

    Uni

    vers

    ity L

    ibra

    ry]

    at 2

    2:20

    13

    Nov

    embe

    r 20

    14

  • make aerobic waste-water treatment systems ofuse only for the treatment of dilute, largelysolubilized, materials. Some solids may be present,but these should be in the form of fine, suspendedparticles and their removal from the waste-water iscaused as much by entrapment and sedimentationin zoogleal floes of microorganisms as by actualdegradation. Aerobic microorganisms can, ofcourse, break down solid organic materials giventime and, again, a sufficient supply of air. Thisability is made use of in composting systems whererelatively dry, solid, organic wastes are degraded.But this process depends on the dryness of thesolids to provide a loose matrix through which aircan diffuse. If the solids are wet and compact,oxygen becomes limiting and anaerobic metabo-lism results. The other difficulty with aerobictreatment is that as respiration gives a high energyyield a large proportion of the substrate carbonand nitrogen is converted to microbial cells. In thecase of composting of solid wastes the productionof microbial cells does not matter. Indeed it is adesirable function of the process as the endproduct is generally used as fertilizer. But in thecase of treatments designed to produce a clearwater the mass of microbial cells is an embarrass-ment. These cells have to be removed from thetreated liquid. They can only be removed as a verywet sludge and, as they themselves can be subjectto microbial attack, and so are potentially pollu-tional, their disposal presents a further problem.

    Anaerobic metabolism on the other hand pres-ents no problems of oxygen supply and is suitablefor breakdown of concentrated and insoluble, wet,organic wastes. This is exemplified by the fact thatfor millions of years anaerobic microbial metabo-lism has been used in the alimentary tracts ofherbivorous animals for digesting vegetable food-stuffs. "Organic" is used here in the sense ofanimal or microbial bodies or products of theirmetabolism, native vegetable materials or manu-factured products such as paper, but it does notinclude many synthetic organic chemicals whichare degraded only slowly, if at all by anaerobicprocesses. Its disadvantage from the point of viewof waste treatment is that although a mixedanaerobic flora will under the correct conditionsconvert a large amount of the carbon in an organicwaste into methane and carbon dioxide, which arenonpolluting, the process is never entirely com-plete, and some of the acidic intermediary pro-ducts remain to pollute the residual liquid, and,

    except in the largely theoretical case where nitro-gen is growth-limiting, ammonia and possiblyother nitrogenous compounds may also remain.Thus in most cases the liquid from an anaerobictreatment system may not be suitable for uncon-trolled discharge. On the other hand anaerobicmetabolism is of low energy yield and the conver-sion of substrate to microbial cells is small, so thevolume of residual microbes to be disposed of isless than in an aerobic system. The residualmaterial from an anaerobic system is also stableand odorless and so presents less problems per seof disposal.

    Anaerobic metabolism can then, by itself, notoften be a complete waste disposal system inreducing a liquid organic waste to present-day riverboard standards. It can be a system for reducingthe volume of, and stabilizing the solids in a waste,and for greatly reducing the pollutional load of thewaste liquid. This, in itself, may be sufficient, butgenerally some further aerobic or other treatmentof the liquid will be desirable before its finaldischarge.

    Anaerobic microbial metabolism may takeplace whenever the ingress of oxygen is stoppedcompletely or limited to such an extent thataerobic microbial metabolism will quickly removethe oxygen. Thus it will take place beneath thesurface of still waters where rate of diffusion ofoxygen from the air is slow and the oxygen can beused up by microbial metabolism in the surfacelayers. This then is the basis for the simplest typeof anaerobic, waste-treatment system; the lagoon.A lagoon consists of a large tank, of a few feetminimum depth, open to the atmosphere. Wastematerial, such as farm animal excreta, is slowly runinto one end and partially purified liquid runs outat the other. Because there is little turbulence inthe lagoon the waste solids settle out. An aerobicmetabolism takes place in the top layer of theliquid, but beneath this, and in the settled solids,anaerobic metabolism slowly breaks down andstabilizes the organic matter. The principal dis-advantage of a lagoon is its size. Since themicrobial action takes place at an ambient temper-ature it is slow, and to obtain sufficient break-down of the waste its detention time must be long,a matter of months or a year or so. Thus if theoutput of waste to be treated is large, a very largelagoon is needed. This in turn brings otherdisadvantages. Unless the waste is very liquid thelagoon may have to be filled with water initially to

    132 CRC Critical Reviews in Environmental Control

    Dow

    nloa

    ded

    by [

    Cor

    nell

    Uni

    vers

    ity L

    ibra

    ry]

    at 2

    2:20

    13

    Nov

    embe

    r 20

    14

  • obtain dispersion and settling of the solids. Thelarge surface area makes evaporation in dry clim-atic conditions and flooding in wet ones a prob-lem. While the large surface allows release ofcarbon dioxide and methane from the anaerobicmetabolism, it also allows release of sulphurcompounds and other microbial metabolic pro-ducts which can create a nuisance.

    Under some conditions a lagoon can alsobecome a breeding ground for flies and otherpests. The lagoon has applications on certain sizesof farms under certain conditions, or for sometypes of factory waste, but its use is limited.

    The "septic tank" in its various forms is anenclosed anaerobic system where ii) a speciallydesigned chamber or chambers a settlement anddigestion of solid waste takes place and a partiallypurified liquid is run off while gaseous endproducts are vented to the atmosphere. Feed is bygravity and is intermittent and controlled only bythe rate of flow in the sewers feeding the tank.The septic tank principle is largely used fordomestic wastes and is now generally limited tosingle house use, although some undergroundstorage pits for farm wastes may, by accident ordesign, obtain a similar process. Since, like thelagoon, a septic tank is at ambient temperature itis again a slow process and the tank must bedesigned for its particular purpose. Besides havingadequate design, both the lagoon and the septictank must be run under controlled conditions orbreakdown of the system and creation of a pooreffluent and a public nuisance may result. Also inboth cases there is a gradual accumulation of solidsand these have to be removed at intervals anddisposed of.

    The third system of anaerobic waste treatmentin general use is that referred to as "anaerobicdigestion." This is anaerobic metabolism takingplace at a temperature usually above ambient andin a closed system which can be of simple orcomplex design. When it is run at elevated temp-erature the microbial metabolism is quicker thanin the lagbon or septic tank and the detention timeof the waste is less. Also the rate of charge (anddischarge) is subject to greater control than theseptic tank and the gaseous end products can becollected to do useful work.

    The most common use of the anaerobic digesteris as a treatment for the solids settled off fromdomestic sewage before the aerobic treatment,combined with treatment of the microbial sludge

    from the aerobic plant. This utilizes the previouslymentioned ability of the anaerobic microorganismsto degrade solid organic material in very thicksuspension. This property also makes the process apossible one for primary treatment of thickslurries such as those from farms, and of the solidresidues from, for instance, vegetable or meatprocessing factories, slaughter houses, or the fer-mentation industries. Since, of course, anaerobicmetabolism of completely dissolved compounds isalso possible, anaerobic digestion can also be usedfor soluble wastes. However, a complex flora, thecomponents of which have different growth rates,is necessary in practice, if not in theory, foranaerobic digestion and anaerobic bacteria aregenerally of lower maximum growth rate thanaerobes, so to stabilize this flora for treatment ofsoluble wastes, especially of high flow rate, asystem different from that used for high solidswastes is needed. There is also some indicationthat certain components of the anaerobic florarequire a solid surface for optimum growth andsolid components in the waste can supply this.

    Since anaerobic digestion at higher than ambi-ent temperature seems the process which is con-tributing, and is most likely to contribute most tosystems of waste treatment, and is the mostamenable to control, some theoretical and prac-tical considerations of the process will be thepurpose of this review, although the materials andmethods of construction of digesters are outsidethe scope of this article. Lagoons and septic tankswill not further be considered. Also, in the presentclimate of decrease in conventional energy suppliesmicrobial production of useable gas would seem,at least on a limited scale, to offer possibilities.

    Since anaerobic digesters have been part ofdomestic sewage plants for at least 60 years manypapers and reports, sometimes conflicting, on theoperation of digesters have been published. Butdetailed investigation of the microbiology andbiochemistry is comparatively recent, as it hasdepended on methods developed for use in otherfields. There is also difficulty in comparing reportson digester operation because the substrates fordigestion are heterogeneous and the methods ofanalysis commonly used are not sufficiently exactto show up differences in detailed composition ofthe wastes which may affect digestion. In effortsto correlate data on anaerobic digestion, themicrobiology and biochemistry and the theoreticaland practical aspects of digester design have been

    June 1974 133

    Dow

    nloa

    ded

    by [

    Cor

    nell

    Uni

    vers

    ity L

    ibra

    ry]

    at 2

    2:20

    13

    Nov

    embe

    r 20

    14

  • the subjects of some reviews and symposia. Somemore recent ones are listed under references.1"4

    In this article the authors have not attemptedto review every paper on anaerobic digestion orrelated subjects; this would be impossible, as wouldindeed including every aspect of the subject. Whatthey have tried to do is to give an indication of thepresent status of digester technology and to try torelate the biochemistry, microbiology, and theoryof digester operation to the often more detailedwork in related fields.

    II. THE BIOCHEMISTRY OFANAEROBIC DIGESTION

    Anaerobic Metabolism in the Rumen and inDigesters

    Although anaerobic metabolism was probablythe mode of life of the most primitive microorgan-isms it was only with the advent of large amountsof organic material that metabolism of the typeconcerned in anaerobic digestion could develop.The production by microorganisms of energy forgrowth by partial breakdown of organic materialshas been going on in suitable habitats for millionsof years. It occurs at low temperatures in under-water muds, it occurs at higher temperatures in theintestinal tract of herbivores and presumablyprehistoric herbivores had a similar mode ofdigestion. The multicompartment stomach andcaecum digestion of herbivores is most akin toanaerobic sewage digestion, and from the well-documented bacteria and reactions of the rumencan come information of value in the study ofanaerobic digestion. However, there is one majordifference between the functions of a rumen andan anaerobic digester. To optimize rumen condi-tions means that production of the volatile fattyacid products of the primary fermentation offeedstuffs and of microbial mass, which form theanimal's nutrients, should be maximized, and thesecondary production of methane, which is largelya waste to the animal, should be minimized. Onthe other hand, in an anaerobic digester produc-tion of acids and microbial cells should be aminimum, as these pollute the final effluent,whereas production of gaseous end products,which are nonpolluting and can be a useful sourceof energy, should be a maximum.

    Since anaerobic metabolism of organic com-pounds is so widespread the bacteria responsibleare widely distributed and both the herbivore's

    intestine and the anaerobic digester rely on anatural inoculation of the bacteria to develop amixed flora, which is basically similar in all cases,but is modified by the conditions of availablesubstrates and mode of "feeding" of the sub-strates. In each case a very complex flora develops,some constituents of which are vital to thereactions, but others of which may be onlygrowing fortuitously and playing no real part inthe reactions.

    The basic reactions in anaerobic digestion areshown in Figure la and compared with those ofthe rumen in Figure lb. But whereas the digesterreactions are shown only in broad outline and aremainly deductions, all the pathways shown in therumen metabolism have been demonstrated in vivoor in vitro. Although the degradable material canbe classified into proteins, nonprotein nitrogenouscompounds, carbohydrates, lipids, and salts thematerials making up these basic constituents mayvary considerably. In the rumen the variation isbetween the insoluble carbohydrate (mainly cellu-lose) of the natural plant diet and the soluble(molasses or starch) carbohydrate of the intensive-rearing diet, and between natural proteins ofplants and "manufactured" proteins (e.g., soya-bean or fishmeal) or nonprotein nitrogenous com-pounds (e.g., urea or ammonium salts) of theintensive feeding. In the rumen these variationslead to well-documented changes in predominantmicrobial types and fermentation products.

    In the domestic digester the principal consti-tuents of the primary sludge screened off from theincoming sewage will be very similar in all westerncountries and so will the constituents of aerobicsludge. The differences in feed to the digesters willlie in the proportion in which these two wastes aremixed. However, in largely vegetarian countriesthe composition of the primary sludge will bedifferent from that of western countries in that itwill contain more fibrous plant residues. Indeed, itwas suggested many years ago that the gas produc-tion in anaerobic digesters in India was due to thefaecal residues from a vegetarian diet. Thesematerials will then be more akin to the feed to adigester processing farm wastes, where partlydegraded cellulosic plant residues form the maincarbohydrate constituents of the waste. Wastefrom vegetable processing plants will be evenhigher in plant residues, either undegraded ordegraded to varying extents by the washing andpeeling processes. On the other hand meat factory

    134 CRC Critical Reviews in Environmental Control

    Dow

    nloa

    ded

    by [

    Cor

    nell

    Uni

    vers

    ity L

    ibra

    ry]

    at 2

    2:20

    13

    Nov

    embe

    r 20

    14

  • Proteins nonprotein N Carbohydrates Lipids Salts

    A Ammonia Simple sugars Glycerol/ plus

    long chain acids

    Hydrognationegradation (?)

    Waxes,hydrocargonoils, plastics,

    etc.

    Bacterial cells Volatile fatty acids, H2, CO2(which?) ^ lactic,

    CH4 CO2

    Bacteria, saltsindigestible residues

    SO4NO;

    SHNH,

    Bacteria, bufferingof medium

    A "liquefaction"B "methanogenesis"

    FIGURE 1A. General scheme of the reactions occurring in anaerobic digestion. The question marks indicate reactionsabout which there is doubt as to the extent or nature of the reaction or its products. Other reactions have been shownexperimentally or can reasonably be inferred.

    or slaughter house wastes will be largely protein-aceous and lipid material, together with someplant residues from gut contents of the animals.Waste from distilleries or similar factories couldcontain relatively large amounts of soluble carbo-hydrates as well as cellulosic plant residues.

    As with the rumen, variation of the feed toanaerobic digesters should result in variations incomposition of the microbial flora, variations inrates of digestion, and variation in some of thepathways of digestion. However, compared withthe rumen there is little fundamental knowledge ofthe microbiology and biochemistry of anaerobic

    digestion. Some of the variations in digesterfunction are known, others can only be deducedand must be later subjected to experimental proof.

    The Composition of Digester FeedstocksExcept in the case of some factory wastes, such

    as have been mentioned, the materials fed todigesters will have been subject to the actions ofintestinal enzymes and microbial action, either inthe animal body or in sewers or collection tanks,or in the aerobic sewage system. Very little easilydegraded or soluble carbohydrate or proteinaceousmaterial will be present. Although in the liquid of

    June 1974 135

    Dow

    nloa

    ded

    by [

    Cor

    nell

    Uni

    vers

    ity L

    ibra

    ry]

    at 2

    2:20

    13

    Nov

    embe

    r 20

    14

  • Food

    Proteinnonprotein N

    Amino acids

    Carbohydrates

    Simple sugars

    Ammonia

    Lipids

    GlycerolGalactoselong-chainFatty acids

    Salts

    Saliva

    Bicarbonate Urea

    Bufferingrumen fluid

    Ammonia

    Formic propionic butyrict 4-

    lactic

    4-succinic

    J

    COj

    Hydrognation

    Ingested byprotozoa

    Saturated acids

    *Some straight chain also

    Passing on to intestines.

    Absorbed from rumen.

    Excreted by mouth.

    Bacteria, protozoa, vitamins, long chain acids, unabsorbedproducts of microbial metabolism, food residues, salts.

    VFA, NH3

    CH4 CO2

    FIGURE IB. Miciobial metabolism in the rumen.

    farm wastes which have been standing underanimal houses or in collecting tanks, quite largeamounts of acid fermentation products may bepresent. In our own work we have found up toabout 8,000 ppm volatile fatty acids in waste fromunder piggery floors. However, before becomingamenable to fermentation most of the constituentsof the waste must be hydrolyzed to monomeric ordimeric compounds. Since little, if any, of theseprimary hydrolysis products are found in digesterfluids, and as discussed later they may be inhibi-tory to the primary hydrolysis, this primaryhydrolysis could be the rate-limiting step inanaerobic digestion.

    As was mentioned before, sewage sludges fed todigesters might vary in composition and this seemsto be borne out by the published analyses ofsludges. However, it is difficult to compare suchanalyses as techniques of analysis vary and thefractions, although perhaps given the same name,are in fact crude materials defined only by themethod of analysis. Also the terminologies"BOD," "COD," "permanganate value," etc., usedto define sewage treatment, while valuable incontext have no real meaning in terms of chemicalentities. Some analyses are shown in Table 1. Itwill be seen that the composition of the domesticsludge varies considerably, especially in the con-

    136 CRC Critical R eviews in Environmental Control

    Dow

    nloa

    ded

    by [

    Cor

    nell

    Uni

    vers

    ity L

    ibra

    ry]

    at 2

    2:20

    13

    Nov

    embe

    r 20

    14

  • TABLE 1

    Some Analyses of the Solids of Sewage Sludges Used asDigester Feedstock

    a. Ether soluble 34.40%Soluble in coldand hot water 9.52%

    Alcohol soluble* 2.49%Hemicellulose 3.20%Cellulose 3.7Lignin 5.7Crude protein 27.12%Ash 24.13%

    b. Protein (N X 6.25) 19.4%Fat (ether extract) 25.2%Crude fiber 10.8%Cellulose 1.4%Humic acid 4.0%

    (pyridine soluble)

    Total volatile matter 60.8%

    *Waxes, resins, alkaloids, choline

    c. Hemicllulose 6.15%Cellulose 34.48%Lipids 14.01%Protein 18.98%Ash 34.88%

    d. Carbohydrate 24.2%Protein 20.6%Lipid 20.4%Ash 27.8%

    e. Protein 20.9%Lipid 7.7%Neutraldetergentfiber* 53.8%

    Aciddetergentfibert 33.0%

    Lignin 10.1%Ash 17.6%

    *Cellulose, hemicellulose, lignintCellulose, lignin

    Terminology as used by investigators:a., b., c. domestic sewage sludges3 >s >6

    d. domestic sewage sludge, from average figures quoted inReference 7e. piggery waste (authors)

    tent of fatty and cellulosic materials. The fattymaterials may consist of glycerides and other lipidsfrom natural sources and hydrocarbon oils, andthe apparent content of the former will vary with

    the solvents used for extraction. The content ofcellulosic material may be a reflection of theproportions of aerobic sludge added to the crudesewage, or of other things such as the screeningmethods used and whether screenings are added tothe sludge. It is also difficult to say how thecomplex polysaccharides of bacterial residues willappear in the different methods of analysis.Analysis c. approaches the analysis e. which is ofwaste from pigs fed on a barley ration where muchof the cellulose of the barley husks and sawdustfrom the piggery appears in the waste.

    Digestion of CelluloseTaking the analyses as a whole, it is apparent

    that the three main digestible components ofsludges fed to digesters are the polymers desig-nated as cellulose and hemicellulose, protein, andfatty materials. Lignin, although an ill-definedmaterial in most analyses, is regarded as virtuallyundegradable by anaerobic processes and so willnot contribute to the biochemical processes ofdigestion. However, it does play a part in digestionof the first constituent of sludge to be considered cellulose.

    Cellulose, although basically a (3-linked glucosepolymer, can be found in many forms which differin biodegradability. The microbiology and enzy-mology of cellulose degradations have been con-sidered in a number of reviews (e.g., see Reference8) and will not be dealt with at length here. Butsome points are relevant to a discussion of anaerobicdigestion. In its native form in plant structure, orspecialized constituents such as cotton fibers,cellulose is highly polymerized and the polymersare highly orientated. In addition the fibers maybe coated with wax, or interlocked with lignin andpectic materials or other polysaccharides. Thismakes the native cellulose resistant to attack bymicroorganisms. Pectin and "hemicellulose" (thisis a rather ill-defined term, but is usually taken tomean pentosans) are relatively easily attacked bymicroorganisms such as occur in the animal gut,and so much of this will have been removed fromfaecal material in the digestive processes, and anyremaining will be amenable to degradation in thedigester. Domestic cooking of vegetables couldalso remove much of these polymers. However,lignin, as shown by studies on the digestion in therumen of plant fibers of different ages can renderplant cellulose almost undegradable. Althoughthere is little direct evidence, there is likely to be

    June 1974 137

    Dow

    nloa

    ded

    by [

    Cor

    nell

    Uni

    vers

    ity L

    ibra

    ry]

    at 2

    2:20

    13

    Nov

    embe

    r 20

    14

  • some cellulolytic activity in the human intestines.Lignification of the vegetable fibers and the lowcellulolytic activity and rapid passage of foodstuffsthrough the intestine will tend to result in littledegradation of cellulose, but this will be offset tosome extent by the usual cooking of vegetableswhich will make the fibers more readily degrad-able. However, in the case of wastes from herbivor-ous animals such as cattle and pigs the food isuntreated, or relatively so, and only the moreeasily degradable cellulose will be attacked in thedigestive tract. The residue, which could be afeedstock for anaerobic digestion, will consist ofthe more lignified and difficultly degradable fibers.The digestibility of wastes from vegetable proces-sing plants will depend upon the extent to whichthe native plant fibers have been degraded by thewashing and peeling processes of the factory. Forinstance, alkali treatment will render the fibersmore easily degradable than the native material.Papers in domestic sewage will have been treated,mechanically and chemically, in manufacture andthe fibers in these will be very easily degraded.

    Thus the chemically- or physically-defined ma-terial known as "cellulose" consists of materials ofvery different biodegradabilities. Present studiessuggest that more than one enzyme is concerned inthe hydrolysis of cellulose and that while anenzyme (or enzymes) which will hydrolyze treatedcellulose is widespread among cellulolytic bacteria,the presence of another enzyme needed to degradenative cellulose is not so common. Owing to thefact that paper either in powder or strip form, orchemically and mechanically disintegrated cotton,are the substrates used as a matter of conveniencefor isolation of "cellulolytic" bacteria, it is notpossible to say how many anaerobic "cellulolytic"bacteria produce the enzymes needed for hydro-lysis of native cellulose. But in some studies ofrumen bacteria only one out of five strains ofthree species of "cellulolytic" bacteria was asactive in degrading native cellulose (cotton fibers)as the ground filter paper on which it had beenisolated. The other strains varied from 0 to about60% dissolution of the cotton fibers.9 The rate ofdegradation of native or treated cellulose is likeany microbial attack on a solid substrate, slow andmany pure cultures need some days at least, andoften weeks, before complete solubilization ofcellulose is achieved. But here again bacteria vary.Van Gylswyk and Labuschagne1 found thatwhile four strains of Ruminococcus albus solubil-

    ized 400 to 495 mg of ground filter paper per 100ml medium in 24 hr and six strains of R.flavifaciens solubilized 250 to 310 mg, five strainsof Butyrivibrio fibrisolvens solubilized only 30 to70 mg. Three strains of Cillobacterium cellulo-solvens were variable in solubilizing cellulose andgradually lost their cellulolytic activity on keepingin stock culture on cellobiose agar.

    While some cellulolytic bacteria can also de-grade pectins and pentosans, some cannot, and socomplete attack on a plant structure may need theassociation of other bacteria with the cellulolyticones. However, degradation of pectins or pento-sans is usually quicker than that of cellulose, sothe ultimate degradation of the plant structure willdepend on the activity of the cellulolytic bacteria.

    As will be seen in Section IV. the majority ofbacteria isolated from anaerobic digesters utilizecarbohydrate as a source of energy. This may be tosome extent an artifact of the isolation proceduresso far used, but so far as can be seen from analysis,cellulose and associated hemicellulose are themajor, or only, carbohydrates present in sewagesludges. The position of bacterial carbohydrates isunclear, but these are likely to form only a minorproportion of the total carbohydrate, and willusually yield, even if they can be hydrolysed,amino-sugars or deoxypentoses or other sugars ofless general utility to bacteria than the glucose,cellobiose, xylose, and xylobiose of cellulose orhemicellulose breakdown. It would seem thatthese latter sugars are the major source of energyfor a large proportion of the nonmethanogenicbacteria, and so it might be suggested, as wasoriginally done by Maki6 that cellulose hydrolysisis one of the, if not the, rate-limiting steps inanaerobic digestion. Cellulose fibers remain in therumen for up to two or three times the residencetime of the liquid fraction, but the cellulose-hydrolysis products, cellobiose or glucose, areimmediately fermented and the concentration offree sugars is extremely low. There are fewer dataon anaerobic digesters, but as with the rumenbacteria, if cellulolysis were proceeding faster thanthe rate of utilization of the product, cellobiose,then any accumulation of cellobiose would de-crease the rate of cellulolysis. At a concentrationof 0.1%, cellobiose in a medium will inhibitcellulolysis by the most actively cellulolytic rumenbacteria, the ruminococci.1 '

    Apart from having a possible controlling role inthe overall rate of the digester reaction, cellulolysis

    138 CRC Critical Reviews in Environmental Control

    Dow

    nloa

    ded

    by [

    Cor

    nell

    Uni

    vers

    ity L

    ibra

    ry]

    at 2

    2:20

    13

    Nov

    embe

    r 20

    14

  • can play an important part in reduction of thetotal solids content of the sewage as it is, in thecase of some sewages at least, the only insolublematerial that can be significantly reduced in mass(see below).

    Digestion of Nitrogenous MaterialsProteolytic activity has been demonstrated in

    laboratory and domestic digesters12 and proteo-lytic and deaminative bacteria have been isolated(Section IV.)- There seems no reason to doubt thatthe degradation of proteins in digesters follows thesame pathways as that in the anaerobic rumenwhere proteins are hydrolyzed to peptides andamino acids and some of these are then deamina-ted with the formation of ammonia, carbondioxide, and volatile fatty acids. Deamination ofamino acids and not carbohydrate fermentation isthe source of branched chain volatile fatty acidsfound in small amounts in the rumen or digestersystems. Peptides, amino acids, and ammonia canall be reconverted to microbial protein so that thenet loss in proteinaceous material in digestion maynot be very great. If the deamination reactionsproceed to too great an extent then ammonialevels in the digester can become unacceptablyhigh.

    Nonprotein nitrogen compounds may also bedegraded in the digester. Bacterial hydrolysis ofurea to ammonia takes place rapidly in the rumenalthough it is difficult to isolate ureolytic bacteriafrom this organ.1 3 Urea could be a constituent offresh farm wastes. A number of species of bacteriawill degrade purines and pyrimidines under an-aerobic conditions. The mixed rumen bacteria willslowly ferment in vitro xanthine, uric acid, andguanine to carbon dioxide, ammonia, and aceticacid.14 Micrococcus aerogenes will fermentadenine and guanine to lactic acid, mainly, withammonia, carbon dioxide, and some hydrogen.15

    Micrococcus lactilyticus ferments hypoxanthineand xanthine to hydrogen, carbon dioxide,ammonia, urea, and propionic and acetic acids.16

    While Clostridium acidi urici and Cl. cylindro-sporum ferment uric acid to ammonia, carbondioxide, acetic acid, and formic acid, and Cl. acidiurici also ferments xanthine and hypoxanthine tocarbon dioxide, acetate, and ammonia.17 Micro-coccus aerogenes will slowly ferment the pyrimi-dines uracil, cytosine and thymine to carbondioxide, hydrogen, ammonia, lactic acid and aceticacid, while Cl. uracilicum grown in the presence of

    uracil will degrade this compound to /3-alanine,carbon dioxide and ammonia.18 Experiments withM. lactilyticus16 showed that the degradation ofpurines by cell suspensions could be affected bythe presence or absence of hydrogen in the gasphase.

    Purines and pyrimidines will be present insludge fed to digesters in the degradation productsof the dead bacteria which form a large part of theaerobic sludge and the faecal material. Degradationof these compounds has been shown to be causedby bacteria which could be present in digestingsludge. Under the conditions of rapid uptake ofhydrogen by the methanogenic bacteria of thedigester one might expect that the rate or extentof degradations involving the production of hydro-gen might be increased, so that purine andpyrimidine decomposition would be affected asdiscussed below.

    Nitrate will be reduced to nitrite and then toammonia in anaerobic systems so nitrate could bea source of nitrogen for digester bacteria. How-ever, Gasser and Jeris19 testing nitrates (sodium orammonium) as nitrogen sources in laboratorydigesters treating a corn starch syrup trade wastefound that a large proportion (some 50%) of thenitrate nitrogen was given off as nitrogen gas andwasted. This may have been due to the design oftheir digesters or the way in which the digesterswere loaded, or to the particular bacterial popula-tion that built up in the digesters, as loss ofnitrogen as gas would not be expected to be largein an anaerobic system. Possibly a large populationof bacteria depending on nitrate respiration forremoval of electrons from carbohydrate fermenta-tions built up under these conditions. The pointneeds further investigation.

    Digestion of Lipid MaterialThe question of lipid metabolism in the

    anaerobic medium of the digester appears to posesome problems, and the metabolism differs fromlipid metabolism in the rumen.

    The lipids of ruminant feedstuffs contain a highproportion of unsaturated fatty acids, some 60%of pasture acids being linolenic acid, with about10% of linoleic and oleic acids and approximately5% of palmitoleic acid. Most of these fatty acidsare present as glycerides, mainly galactoglycerides,but there are also sterols, waxes, and phospho-lipids, as well as some free fatty acids (for reviewssee References 20 to 22). In the rumen, galactogly-

    June 1974 139

    Dow

    nloa

    ded

    by [

    Cor

    nell

    Uni

    vers

    ity L

    ibra

    ry]

    at 2

    2:20

    13

    Nov

    embe

    r 20

    14

  • cerides, triglycrides, and phospholipids arerapidly hydrolysed by bacterial Upases.23 Theglycerol or galactose liberated in the hydrolysis isfermented to volatile fatty acids. The liberatedlong chain fatty acids are, if unsaturated, rapidlyhydrogenated by certain bacteria in the rumencontents to the corresponding saturated acids (forreview see Reference 24). However, no appreciabledegradation of the free fatty acids can be demon-strated in in vitro incubation experiments withrumen contents, and experiments in vivo with14C-linoleic acid showed only 0.6 to 1.0% of theradioactivity appearing as steam volatile degrada-tion products.25

    Some of the long chain fatty acids may also beabsorbed on to, and possibly integrated into thestructure of rumen bacteria,26 and the presence oflong branched-chain fatty acids in some species ofrumen bacteria is apparently due to their synthesisby chain elongation from the short branched-chainacids isobutyric, isovaleric, and 2-methyl-butyricacids.27 The rumen protozoa can also assimilatelong chain fatty acids and can convert some acidsinto others.

    Although there may be some microbial degrada-tion of existing long chain fatty acids andsynthesis of new ones there is no evidence forsignificant degradation of long chain acids tovolatile acids. Some straight chain hydrocarbonsare present in plant lipids, and although oxidationof these to the corresponding fatty acids wasshown to occur in the rumen epithelium, microbialoxidation did not occur.28 Degradation of hydro-carbon oils would not be expected in anaerobicdigesters. Lipid metabolism in anaerobic digestershas been much less intensively investigated thanthat in the rumen, but there is evidence that theprocesses differ in one important aspect. A generalsurvey of the published analyses of raw anddigested sewage sludge may be difficult to inter-pret, as different methods of extraction of "fat"have been used. Thus, the "fat" fraction maycontain, with glycerides, phospholipids and freefatty acids, waxes and hydrocarbon grease andoils, so a change in one component on digestionmay be difficult to quantify. For instance, asimple ether extraction will not extract structurallipids of bacteria so any originally free lipidcombined in this way by growing bacteria will beclassified as "digested." Nevertheless, there isundoubted evidence that lipid material does dis-appear during digestion. In the case of piggerywaste where the lipid is only a few per cent of the

    total solids (Table 1), this disappearance, while alarge fraction of the lipid (some 64%), was a muchsmaller fraction of the total solids digestion (17%).But in the case of some sludges of high lipidcontent, digestion of the lipid can be a majorcomponent of the total solids degradation.7 Theknown pathways of metabolism of lipids in sewagedigestion are summarized in the next paragraphs.

    The lipid fraction of domestic sewage fed toanaerobic digesters appears to contain a lowerproportion of unsaturated acids than do the lipidsof vegetation fed to ruminants, and the analyses ofHeukelekian and Mueller29 also showed the un-saturated acids to be largely monoenoic. There isalso a higher proportion of free long chain fattyacids in the domestic sewage. These differencesmight be expected. The sewage lipids are subjectto the same reactions as occur in the rumen, that ishydrolysis of glycerides and hydrognation ofunsaturated fatty acids.29 However, there is onemajor difference between rumen and digestermetabolism of lipids. In the rumen the free fattyacids remain unchanged and pass on to theabomasum and intestines. In the digester the freefatty acids are degraded. Analytical proceduresshow a disappearance of lipid material and sincelipid can form a large proportion of domesticsewage solids (Table 1), the inference is that lipidsmust be one of the substrates for methanogenesis.Chynoweth and Mah7 incubated digester sludgefor short periods with or without the addition ofthe lipid fraction of raw sludge or palmitic acidand found that acetic acid was a product of fattyacid degradation. The results of Heukelekian andMueller29 suggest that this acetic acid is formedby stepwise degradation of the long chain fattyacids by a process akin to the -oxidation degrad-ation of acids found in tissues and aerobicbacteria. They measured the changes in concentra-tion of stearic, palmitic, and myristic acids duringa batch digestion of sludge and found evidence forconversion of stearic acid to palmitic and then tomyristic. In )3-oxidation two carbons are removedfrom the carboxyl end of the Cn fatty acid and theproducts are an acid of CR _ 2 chain length, aceticacid (as acetyl CoA), and 4H. The 4H are normallyoxidized with those from the TCA cycle. Althoughthis mechanism would account for the acidicbreakdown products of lipids in digesting sludge,the bacteria responsible and the actual mechanismof the reaction remain unknown. So far as theauthors are aware there is no known anaerobicbacterium which will cause this reaction. It seems

    140 CRC Critical Reviews in Environmental Control

    Dow

    nloa

    ded

    by [

    Cor

    nell

    Uni

    vers

    ity L

    ibra

    ry]

    at 2

    2:20

    13

    Nov

    embe

    r 20

    14

  • most likely that this is a symbiotic reaction,"pulled" by the transfer of hydrogen in a waywhich will be discussed later. The hydrogen-accepting bacterium would then be a methano-genic one reducing carbon dioxide. The hydrogen-generating bacterium could be an unknown onewhich degraded long chain acids completely toacetic. This would not seem to be overall anenergy-generating reaction so that the bacteriumwould presumably rely on fermentation of someother substrate to provide growth energy. On theother hand it may not be necessary to postulate anunknown bacterium. If some degradation ofabsorbed long chain acids were normal in pro-duction of acids needed for cell synthesis, andhydrogen were a normal fermentation product,then under conditions where the hydrogen wasbeing utilized by a methanogenic bacterium thismight deplete the bacterium's level of reducednucleotides needed for synthesis of cell. com-ponents and so the normal one or two stepdegradation of fatty acids might continue com-pletely to acetate to make good this deficiency ofelectrons. Since Chynoweth and Mah7 foundpalmitate degradation to continue in the presenceof chloroform as inhibitor of methanogenesis theysuggested that electron-accepting systems must beable to function in the absence of carbon dioxidereduction, and there was some evidence thatbutyrate formation was increased (see below).Further experimental evidence from digesters isobviously needed. The reason why such reactionsdo not occur in the rumen where the reduction ofcarbon dioxide is the only pathway of methaneproduction is not clear. A possible explanationmight lie in feedback inhibitions of acetate pro-duction from degradation of long chain acids bythe concentration of acetate normally present inthe rumen fluid. In the rumen acetate is notconverted to methane.

    Minor Bacterial Metabolites in DigestersThe breakdown of the carbohydrate, protein,

    and lip'id components of sewage waste providesenergy and materials for cell synthesis for thebacteria, but the bacteria also need salts andvarious elements and vitamins. In sewage waste thetrace metals are unlikely to be a limiting factor formicrobial growth as they will be provided not onlyby autolysis or hydrolysis of dead intestinalbacteria, but by food residues, intestinal secre-tions, and urine contained in the sewage. Suffi-cient anions, such as chloride and phosphate,

    should also be present from the same sources.Industrial effluents and materials dissolved fromsewer pipes or leached from the ground by rainwater will also contribute. Sulphur, necessary formicrobial amino acids will be present either as thesulphur amino acids in the waste proteins, or assulphate in salts. Some of the digester bacteriamay incorporate the sulphur amino acids as such;others, utilizing ammonia as source of nitrogenwill, as in the rumen, utilize sulphate or, morelikely, sulphide. Sulphide is present in digestercontents. This can come from degradation ofsulphur amino acids, hydrogen sulphide produc-tion from these sources being a common reactionof anaerobic bacteria such as are found indigesters. Sulphate will also be reduced to sul-phide, either by desulphovibrio-type of bacteria,or, as is found in the rumen30 '31 by carbohy-drate-fermenting bacteria.

    Vitamins may be present in sewage as a residuefrom production of vitamins by intestinal bacteria.But, again by analogy with the rumen, sufficientsynthesis of the B vitamins and vitamin K (seeReference 32) for the needs of the microbial popu-lation might be expected from the activities ofcertain species in the bacterial flora. The produc-tion of vitamin B1 2 by digesting sewage has evenbeen set on a commercial footing.33

    Carbohydrate Fermentation and Some FactorsAffecting Bacterial Metabolism

    Although all the materials for growth of amixed bacterial population are either present indomestic sewage or produced by some of thebacteria for use by others, they may not be inoptimum proportion. It seems likely that somespecialized types of industrial waste might be sounbalanced in one or other of the main com-ponents, carbon and energy, or nitrogen, thatcomplete digestion would be impossible. They areless likely to be limiting in trace elements or salts,but for adequate digestion of these wastes it mightbe necessary to mix them with others, or add, forinstance, a cheap nitrogen source.

    The first stage of digestion, then, results in thehydrolysis of polymeric substrates into simplercompounds which can be used as sources of energyor cell components by the bacteria. But this firststage also results in the conversion of suspendedsolids of the waste into a soluble form and so itcontributes to the overall reduction and stabiliza-tion of the waste.

    Deamination of amino acids to volatile fatty

    June 1974 141

    Dow

    nloa

    ded

    by [

    Cor

    nell

    Uni

    vers

    ity L

    ibra

    ry]

    at 2

    2:20

    13

    Nov

    embe

    r 20

    14

  • acids, carbon dioxide, and ammonia appears to bein the rumen mainly a side reaction of carbohy-drate-fermenting bacteria and not a source ofenergy, and such rumen-type bacteria have beenisolated from digesters (Section IV.). However,clostridia appear to form a larger proportion of thebacterial population in digesters than they do inthe rumen (Section IV.), and it is possible thatsome of these, and perhaps other bacteria, canobtain energy by dismutation of amino acids, forinstance in Stickland reactions. But all presentevidence points to the carbohydrate as being theprincipal source of energy for digester bacteria.The metabolism of carbohydrates will be con-sidered in detail. As mentioned before, althoughbreakdown of long chain fatty acids apparentlyoccurs in digesters, the mechanism of this reactionand its significance as an energy source remainunexplained.

    The fermentation of carbohydrates results inthe formation of one or more of the followingproducts: hydrogen, carbon dioxide, ethanol;formic, acetic, propionic, butyric, valeric, caproic,succinic, and lactic acids. The higher volatile fattyacids valeric and caproic are seldom found, onlyone out of the many rumen bacteria isolated beingshown to form these acids. The actual fermenta-tion products vary with the species, or strain, ofbacteria, but they also vary with the conditions ofgrowth of the bacterium and it seems unlikely thatany bacterium produces only one or a particularcombination of fermentation products at all times.Recent work has also shown some mechanisms ofinteraction of mixed cultures which can affect thefermentation products of constituents of themixture. These effects are all relevant to a con-sideration of the biochemical pathways ofanaerobic digestion and will be discussed here.

    Effects of Growth Rate, Substrate Concentrationand pH

    The results quoted in Section IV. show that theindividual bacteria isolated from digesters whengrown in laboratory culture generally producemore than one fermentation product and col-lectively one might assume that the whole digesterpopulation would produce all the fermentationproducts mentioned in the previous paragraph.However, analysis shows that in a functioningdigester the volatile fatty acid concentration is lowand this residual acid is mainly acetic. The authorsanalyses of some stable, piggery-waste digesters

    showed volatile acid concentrations around 300ppm as acetic, and about 79% of this was actuallyacetic acid with about 15% propionic and nobutyric acid, but about 6% isovaleric, probablyfrom amino acid deamination. Other workers haveshown a similar preponderance of acetic acid indomestic sewage digesters, although there is, as theauthors have also found, some variation in thefigures for propionic and butyric acids. Formicacid is usually absent or in negligible amount indigesters. For instance Pohland and Bloodgood34

    quoted 27 ppm formic out of a total acidconcentration of 780 ppm. Lactic and succinicacids were not detected in piggery-wastedigesters,35 and have not been reported to bepresent in other digesters. Ethanol is not generallylooked for, but is not known to be present indigester contents. Negligible amounts of hydrogenare present in digester gases, although carbondioxide is always present. It would seem, then,that either the bacteria ferment waste constituentsto the mixture of products found in pure culturestudies but that all these products, with theexception of a small amount of acetic acid, areutilized, presumably by the methanogenicbacteria; or that the fermentation pathways are somodified in the mixed digester culture that little,if any, of most of the pure culture products isformed. Consideration of available evidence sug-gests that the latter explanation is probably thecorrect one.

    Fermentation products of bacteria in pureculture are generally determined from growth in amedium of high sugar concentration and the pH ofthe culture may fall considerably during thefermentation. In these cases the initial sugarconcentration is above that required for maximumgrowth rate of the bacteria, and in many cases,since growth is stopped by the lowering of theculture pH rather than by exhaustion of nutrients,the sugar concentration may never fall to a growthrate-limiting value.

    Studies of carbohydrate-limiting continuouscultures have shown that the growth rate of abacterium may affect the fermentation products.At low growth rates the rumen bacterium Sele-nomonas ruminantium produces almost entirelyacetic and propionic acids, whereas at high growthrates the fermentation products of glucose areabout 50% lactic acid and 50% acetic plus pro-pionic acid.36 Lactobacillus casei in continuousculture produced a high proportion of lactic acid

    142 CRC Critical Reviews in Environmental Control

    Dow

    nloa

    ded

    by [

    Cor

    nell

    Uni

    vers

    ity L

    ibra

    ry]

    at 2

    2:20

    13

    Nov

    embe

    r 20

    14

  • only at high growth rates.3 7 In batch cultures thegrowth rate of bacteria may be controlled by theconcentration of limiting nutrient added initiallyto the culture. Hishinuma, Kanegasaki, andTakahashi38 showed that batch culture of S.ruminantium with high glucose concentrations(high growth rate) gave a high proportion oflactate and low glucose gave low lactate with highacetate plus propionate. It is difficult to equatethe mixed substrate and mixed culture conditionsof an anaerobic digester, where bacteria may begrowing on different substrates at different timesand where some bacteria may be dependent onothers for substrates or growth factors, with pureor mixed cultures under defined laboratory condi-tions. But it would not seem likely under mostdigester conditions that nitrogen is limiting growthof the bacteria. Certainly in piggery-waste digestersammonia, which many of the bacteria are able touse as nitrogen source,3 5 '3 9 is always somehundreds of times higher than the concentrationsfound in ammonia-limited continuous cultures ofthe rumen anaerobe Bacteroides amylophilus.40

    Thus energy supply would seem to be the limitingfactor. This may be controlled by the rate ofaddition of the energy source to the digester (i.e.,the dilution rate, or turnover time of the digester)if the source is a mono- or disaccharide. But, aswas previously suggested, such readily availableenergy sources are likely to be low in most digesterfeedstocks and the rate of supply of the energysource will be governed by the rate of hydrolysisof insoluble substrates such as cellulose, and notprimarily by the rate of supply of the substrate.That this can be so for cellulose is shown by thefollowing results. Van Gylswyk and Roch4 ' foundthat the products of fermentation of cellobiose bya strain of the rumen bacterium Butyrivibriofibrisolvens contained three times the proportionof lactic acid as those of the fermentation ofcellulose. In other experiments10 B. fibrisolvenswas found to have a lower maximum growth rateon cellulose than on cellobiose in batch cultures,although both substrates were in the same excessconcentration. Thus the rate of hydrolysis of thecellulose was controlling the rate at which theactual fermentable substrate, cellobiose, was beingsupplied to the bacterium and so the growth rate.Cultures on the two substrates then exhibited thesame effects of growth rate on fermentationproducts as those mentioned previously.

    The growth rate of bacteria can thus affect

    fermentation products and the foregoing observa-tions suggest that digester conditions of lowsubstrate concentrations and low growth ratewould not be conducive to the formation of lacticacid, a fermentation product of many bacteria inlaboratory culture. The other nonvolatile acidmentioned in the list of possible fermentationproducts was succinic. With the rumen bacteriumAnaerovibrio lipolytica growing in continuousculture on glycerol, the fermentation productswere almost entirely acetic and propionic acids atlow growth rates, whereas at high growth ratessuccinic acid, with some lactic, formed 40% ofthe fermentation products. This would accordwith little, or no, succinic acid being found indigester contents. On the other hand Bacteroidesamylophilus tended to produce a higher propor-tion of succinic acid at low growth rates.42 Theevidence would thus appear conflicting, but whileB. amylophilus which as a bacterium which canferment only starch or maltose might not beexpected in a digester treating sewage wastes, A.lipolytica is one of the few known anaerobicbacteria which can hydrolyze glycerides of longchain fatty acids and so might be a constituent ofdigester populations.

    The previous experiments showed changes inproportions of lactic or succinic acids and volatilefatty acids in response to substrate concentrationsin pure cultures, which made it seem unlikely thatthe former acids were formed in large amounts indigesters. The fact that the proportions of theindividual volatile fatty acids can vary with growthrate (substrate concentration) has also beenshown.

    The experiments of Walker and Monk43

    showed changes in proportion of volatile fattyacids in the mixed culture of rumen contents inresponse to substrate concentration. In rumencontents incubated in vitro with glucose at aconcentration of 20 mg/15 g rumen contents,acetic, propionic, and butyric acids were formed inthe proportions of 60, 25, and 15%, respectively.But when glucose was only 0.6 mg/15 g rumencontents, acetic acid was 95% of the total, whilepropionic and butyric acids were only 4 and 1.5%.These results again show how postulated digesterconditions would produce the fermentation pro-ducts found but in the mixed culture the explana-tion may be complex. It could be partly an effecton lactic and succinic acid production due tosubstrate concentrations, as described above, as

    June 1974 143

    Dow

    nloa

    ded

    by [

    Cor

    nell

    Uni

    vers

    ity L

    ibra

    ry]

    at 2

    2:20

    13

    Nov

    embe

    r 20

    14

  • these acids are fermented in rumen contents topropionic and acetic acids, but hydrogen utiliza-tion could have an effect, as described later.

    The pH of growth of bacteria can also affectfermentation products. With lactic acid-producingbacteria decreasing culture pH produces an in-creasing proportion of lactic acid in the fermenta-tion products. Some experiments42 withAnaerovibrio lipofytica suggested that decreasingpH increased succinate production, although withBacteroides amylophilus succinate production wasmaximum at pH 6.8 (the maximum pH forgrowth) but this, it has been suggested, is not alikely digester bacterium. The general conditionsof digester operation, i.e., low substrate and a pHabout 7 might then seem conducive to low lactateand succinate production. Any of these acidsformed would probably be fermented under nor-mal conditions, as in the rumen, to acetic andpropionic acids, since lactate-fermenting bacteria,at least, have been found in digesters (Section IV.).However, the sudden addition of a waste of highfermentable carbohydrate concentration could, byincreasing the growth rate of the bacteria increaselactic acid production above that which could bedealt with by the lactate-fermenting bacteria, andas lactic is a strong acid, so contribute to a fall inpH. The fall in pH would then, in turn, furtherincrease lactate production and digester "souring."

    It was mentioned earlier that formic acid is notfound in large amounts in digester contents.Although formic acid is a substrate for methano-genic bacteria its formation may also be minimizedat digester pH. Micrococcus lactilyticus (Veillo-nella gazogenes) will ferment pyruvate to aceticacid, carbon dioxide, and hydrogen at pH 6.5;only at pH 8.5 are acetic and formic acids formedin equimolar proportions.44

    Growth rate and culture pH affect not onlyfermentation products but also extracellular en-zyme concentrations and the composition ofbacteria40 (see reviews45'50). Since sewage di-gesters appear to run naturally at a pH of about7.2, one must presume that this pH produces theoptimum balance between the growth rates of thevarious bacteria and extracellular enzymes such ascellulase. But with some trade wastes composedprimarily of one substrate the optimum pH ofdigestion could be different. For instance, theproduction of amylase, an enzyme which would beimportant in digestion of starchy wastes by Bac-teroides amylophilus had a rather sharp maximum

    at pH 6.1.42 An effect of growth rate orpHonthe composition of bacteria could reflect in thedigested sludge. Digester conditions which favoredformation of intracellular storage polysaccharidescould have a sludge of higher than normal COD,while the formation of extracellular polysac-charide could affect not only COD but the settlingproperties of the sludge.

    The growth-limiting substrate can also affectbacteria. Although storage polysaccharide forma-tion is usually increased at low growth rates, theusual digester conditions, which we have suggestedare carbon-limiting, would not be conducive tointra- or extracellular polysaccharide formation,but they could be conducive to maximum forma-tion of the hydrolytic extracellular enzymes. Forinstance, the specific amylase and protease activi-ties at a particular growth rate were much higherin carbon-limited than in nitrogen-limited con-tinuous culture of a rumen bacterium. Apparently,when nitrogen is low the bacterium divertsnitrogen from extracellular enzyme formation tomore essential activities.40 Since in a mixedculture like a digester, extracellular enzymes suchas cellulases are always subject to attack byextracellular proteases from other bacteria, a highproduction of cellulase is essential. The carbon-nitrogen balance in digestion of trade wastes ofhigh polysaccharide content must then be care-fully considered, and some mixing of such wasteswith high-nitrogen wastes might be desirable forthis as well as other reasons.

    Observations such as these are, of course,largely speculative with regard to digester opera-tion, as the presence or absence of the particularbacteria have not been demonstrated. But they areput forward to show the kind of reactions whichmight be influencing bacterial action in digestersand the kind of information which should besought if a proper understanding and control ofdigestion is to be obtained.

    Maintenance and Uncoupled FermentationIn the foregoing paragraphs some ways in which

    conditions of growth may influence microbialmetabolism have been shown, but there is afurther consideration which might influencedigester reactions. This is maintenance and un-coupled fermentations. It has been shown with theanaerobic rumen bacteria as well as with otherbacteria growing under aerobic or anaerobic con-ditions that a proportion of the energy generated

    144 CRC Critical Reviews in Environmental Control

    Dow

    nloa

    ded

    by [

    Cor

    nell

    Uni

    vers

    ity L

    ibra

    ry]

    at 2

    2:20

    13

    Nov

    embe

    r 20

    14

  • by oxidation or fermentation of substrate sugarsmay be diverted from production of new cellmaterial to maintenance of the integrity of thebacterial cell (e.g., see References 36, 51, 52). Di-version of energy for maintenance thus results inless cell mass being formed for metabolism of agiven amount of energy source. This is most appar-ent at low growth rates of the bacteria. In experi-ments with rumen bacteria in carbohydrate-limitedculture the bacterial yield from fermentation of agiven weight of carbohydrate could be abouthalved by reducing the growth rate from nearmaximum to half maximum.36'42 Production ofbacterial mass is undesirable in anaerobic digestionas it leads to bigger sludge volumes. So forminimum production of bacteria it would seembest to keep the bacterial growth rate as low aspossible. The turnover time or dilution rate of thesystem controls the growth rate of the bacteriaunder true continuous culture systems with onesoluble growth-limiting substrate. A digester withthe usual heterogeneous substrates will be morecomplex than this simple system, but the turnovertime must have some influence on the overallgrowth rate of the bacteria if the digester iscontinuously loaded and consideration of themaintenance phenomenon suggests that this couldbe one factor in decreasing treatment efficiency ifthe turnover time is decreased too much. Digestersystems treating more easily degradable or moredissolved waste than normal sewage will approachmore the ideal continuous culture system and herethe effect of turnover time on maintenance energyrequirements and production of bacterial massmight have a more obvious effect. However,infrequent feeding of a digester with large volumesof fermentable waste will, as has been mentionedbefore, set up conditions for rapid batch growth ofthe bacteria even if the overall turnover time ofthe digester is long. The high growth rates thenobtained will lead towards minimum maintenanceenergy requirements and maximum production ofcell mass.

    A further factor helping to control residualsludge volume and production of gaseous met-abolic products could be endogenous metabolismof the bacteria. Endogenous metabolism can takeplace when bacteria are in a resting phase afterutilization of available exogenous substrate; fer-mentation of cell substance then takes place. Thiscell material can be a storage substance which isusually a glucose polymer in the case of anaerobic

    bacteria. The storage material is built up duringgrowth under conditions of carbohydrate excess,and its metabolism may provide energy only tomaintain the integrity of the cell during thesubsequent period without substrate or it may belinked to formation of new cell material. Walkerand Nader53 suggested that energy from un-coupled fermentation of excess food carbohydratein the first three hours after feeding a sheep couldbe channeled into forming reserve polysaccharidein the bacteria which could then be fermented toprovide energy for protein synthesis later in thefeeding cycle when the energy supply becamelimiting.

    However, endogenous metabolism may alsoresult in degradation of the cell structural materialin the absence of storage material and, if exogen-ous substrate is not supplied, final death of the cellwill result. Such endogenous metabolism may be apart of the cell lysis which appears to be very rapidin the anaerobic rumen bacteria and which prob-ably occurs in similar digester bacteria. Maximumgrowth of many rumen bacteria in pure culture isfollowed by a rapid decline in optical density ofthe suspension and in viable cells. Death ofbacteria in continuous cultures of very low dilu-tion rate has been demonstrated,54 and death andlysis of bacteria is very likely during digestion ofwastes. If this endogenous metabolism and celllysis takes place in the digester then the bacterialmass will be decreased and the lytic products willbe available for fermentation by other bacteria andnitrogenous components could be incorporatedinto these other bacterial cells. It may be assumedthat digester bacteria are capable of utilizing cellcomponents of other bacteria, as a large part ofthe feed to sewage-waste digesters consists ofbacterial cells. The overall effect of this processwill be a gain in fermentation of the waste and anoverall loss of microbial cells, and so a reduction insludge volume. On the other hand there could bean increase in soluble nitrogenous material in thedigester fluid, as the resynthesis of cells may notutilize all the products of lysis of the initialbacteria.

    Lysis of bacteria will take place in secondarysludge-conditioning tanks, but because of the lowtemperature little fermentation or cell synthesismight be expected. Here decrease in sludge volumedue to death of the bacteria might be expected toincrease the dissolved organic material in thesupernatant liquid.

    June 1974 145

    Dow

    nloa

    ded

    by [

    Cor

    nell

    Uni

    vers

    ity L

    ibra

    ry]

    at 2

    2:20

    13

    Nov

    embe

    r 20

    14

  • Uncoupled fermentation, that is fermentationnot linked to growth, is unlikely to take place insewage digesters, but it could be important indigestion of high carbohydrate wastes, as theprerequisite for this is an excess of fermentablesubstrate and growth limited by some othernutrient, usually nitrogen. The necessity of supply-ing nitrogen to high carbohydrate trade wastes toachieve digestion has already been mentioned, butthe concentration of nitrogen might not be thatcalculated to be necessary to achieve full bacterialgrowth from fermentation of the carbohydrate.Gasser and Jeris19 found in their studies ofnitrogen sources for digestion of a high carbohy-drate, corn starch syrup waste that while digestersreceiving a high nitrogen concentration used about21 mg nitrogen for each gram of COD removed (afigure in agreement with other workers) digestersreceiving a low nitrogen input (ammonium chlor-ide) used only about 11 mg nitrogen for the sameCOD removed. As the COD was almost all carbo-hydrate uncoupled fermentation could haveplayed a part here. These digesters had a turnovertime of 12 days. For the full utilization ofuncoupled fermentation low digester turnovertimes may be best. Henderson, Hobson, andSummers4 showed in experiments with growth ofa rumen bacterium in nitrogen-limited culture thatwhereas only about half the available carbohydratewas fermented at growth rates near the maximumfor these bacteria, at about 14% of maximum rate100% of the carbohydrate was fermented, thebacterial concentration being the same in eachcase. This result could have been a consequence ofa combination of a true uncoupled fermentationand a fermentation linked to maintenance require-ments at the low growth rate. But the result indigester terms would have been increased produc-tion of methanogenic substrates and increaseddigestion of the carbohydrate waste with noincrease in residual sludge mass.

    Growth on Mixed SubstratesIn the previous sections some effects of growth

    conditions on the metabolism of one substrate bya bacterium have been considered. But wasteadded to digesters will usually contain a number ofpossible substrates. Diauxic growth of a bacteriumin a medium containing two substrates, that ispreferential utilization of one substrate before theother, is a well-known phenomenon. It might besupposed to take place in digester contents, and

    Bhatla and Gaudy55 have discussed it in relationto activated sludge systems. But with a mixture ofsoluble carbohydrates it would not seem as ifdigestion would be affected by diauxic growth.For instance, Mtales, Chian, and Silver56 showedthat preferential utilization of one substrate oc-curred only at high growth rates of a number ofbacteria. In these cases glucose and fructose wereused together in continuous cultures at low growthrates, while glucose was used preferentially only athigh growth rates. Low growth rates will generallybe the rule in normal sewage digestion. A similareffect of substrate concentrations was found inexperiments with a rumen bacterium. Solublestarch and lactate were utilized together by Se-lenomonas ruminantium in batch culture, whereaswith glucose and lactate, glucose was preferentiallyutilized and lactate fermentation occurred onlyafter a lag period when the glucose had been used.However, when glucose was in very low concentra-tion it was rapidly used and lactate fermentationfollowed with no appreciable lag period.38 '57

    With starch and lactate the amount of fermentablesugar would be low, as its concentration would begoverned by the rate of hydrolysis of the starch,and so conditions would be similar to those whereglucose was present in low concentration. If thiseffect were general then under the normal lowsugar concentrations of sewage digestion the sugarsand one of their possible fermentation products,lactate, would both be fermented. But high solublesugar concentrations in a waste might not onlylead to high production of lactic acid, as previous-ly discussed, but might suppress its further me-tabolism, increasing the possibilities of digesteracidity.

    However, the presence of soluble carbohydratecould have an effect on cellulose degradation. Thepresence of easily fermented carbohydrate in thefeed (and starch can come into this class) can leadto suppression of cellulolysis in the rumen.S8

    Some experiments with the rumen cellulolyticbacterium Butyrivibrio fibrisolvens showed that itcould grow much faster on cellobiose than cellu-lose.1 Thus, with a mixed substrate growth willtake place on soluble sugar rather than cellulose,and so cellulolysis will be suppressed if bacterialike the Butyrivibrio are the cellulolytic bacteria.The suppression of cellulolysis by the ruminococciin the presence of cellobiose appears to be adifferent phenomenon, as the strains of Rumin-ococcus tested have the same growth rate on

    146 CRC Critical Reviews in Environmental Control

    Dow

    nloa

    ded

    by [

    Cor

    nell

    Uni

    vers

    ity L

    ibra

    ry]

    at 2

    2:20

    13

    Nov

    embe

    r 20

    14

  • cellulose as cellobiose.10 The mechanism seems tobe an inhibition of the cellulase formation not itsactivity.59

    Oscillations in Continuous CulturesDifferent growth conditions can thus affect the

    metabolism of bacteria. Some substrates or sub-strate concentrations, or turnover rates, may thenset up conditions conducive to digester failure. Butin any continuous culture system a change in someparameter will lead to a change in the cultureconditions. And although the new conditions maybe those of a new, stable steady-state, the bacteriamay take some time to adapt and there may beintermediate, transient conditions in the culture.More is known now about the transient conditionsin continuous cultures (e.g., see References 60 and61) but what is important from the point of viewof anaerobic digestion is that an abrupt change in,say, pH or dilution rate of a pure continuousculture can lead to oscillations in bacterial activitywhich can last for a number of turnover times. Theoscillations in a mixed continuous culture such asa digester can be even more complicated, asoscillations in metabolism of one bacterium mayaffect growth of another, and could lead to poorperformance, or even failure, of the digestion. Therumen may again be cited as a parallel. It is wellknown that abrupt changes in diet can lead todigestive disturbances, for instance bloat andacidosis in ruminants and so feed changes arealways better made gradually over a period of aweek or two (the turnover time of a rumen isabout a day). And with a change of diet from hayto certain starch concentrates the rumen flora hadnot adapted to the new diet even after threeweeks.62 Similarly, a change in balance of bacteriawas shown in domestic digester contents adaptingto digestion of piggery waste over a period of someweeks.35 '63 A rapid change in digester tempera-ture from 35 to 30 led to an immediate decreasein gas production and although this was notpermanent it took about eight days (at a ten-dayturnover time) for the digester to recover from theabrupt change in temperature (authors, unpub-lished). If the temperature were changing continu-ously with a comparatively short periodicity thenone might never expect digestion to be optimumand conditions for digester failure might build up.

    All these results lend support to the hypothesisthat digester conditions must be kept as constantas possible, and any changes made gradually.

    Growth of Mixed CulturesAll the effects of growth conditions on the

    metabolism of pure cultures of bacteria may affectthe growth of particular members of the digesterflora, but in addition these bacteria are growing ina mixed culture and this introduces more variables.Some aspects of the interactions of mixed cultureshave been reviewed by one of the authors.64

    There is no doubt that in any mixed culture,such as that in a digester, many bacteria aredependent on others for their growth. We have sofar cited the provision of fermentable substratesand nitrogen sources for one bacterium by an-other. But other growth factors may be involved.Digester fluid has been found to be a usefulconstituent of media for viable counts of thecomplete digester flora, just as rumen fluid has forrumen bacterial counts. The volatile fatty acidscontained in this fluid have been shown to be themain growth-promoting factors for rumen bacteriaand Maki6 showed that digester fluid containedgrowth-promoting factors for cellulolytic digesterbacteria. Acetate has been shown to be utilized incell synthesis by a number of rumen bacteria andthe branched-chain fatty acids isobutyric andisovaleric and 2 methyl butyric acid have beenfound to be needed as carbon skeletons for aminoacid synthesis by the ammonia-utilizing, cellulo-lytic Ruminococcus and other rumen bacteria.Acetate, propionate, butyrate, and valerate havealso been shown to be incorporated into theprotein of mixed rumen bacterial cultures. Thebranched-chain fatty acids are also incorporatedinto the long chain, branched acids and aldehydesof rumen bacterial lipids. The branched-chainvolatile fatty acids are not products of carbohy-drate fermentation in the rumen, but are productsof deamination of amino acids. The bacteriarequiring these growth factors have either no, or alimited capacity to incorporate amino acids. Car-bon dioxide is required by many rumen bacteriafor synthesis of the carboxyl groups of cellularamino acids, for incorporation into other cellconstituents, or for production of the fermenta-tion product succinic acid (see References 65 and66 for a general review of these topics andreferences to original papers). Succinate has alsobeen shown to be the precursor of glutamic acid,proline, and arginine in Bactetoides ruminicola andpossibly Selenomonas ruminantium.61

    The interactions of a mixed culture may alsoaffect the fermentation products formed. Fermen-

    June 1974 147

    Dow

    nloa

    ded

    by [

    Cor

    nell

    Uni

    vers

    ity L

    ibra

    ry]

    at 2

    2:20

    13

    Nov

    embe

    r 20

    14

  • tation product repression of its own formation hasbeen shown. For instance, propionate added to acontinuous culture of Propionibacterium sherm-anii decreased propionate production and thegrowth rate of the bacteria, while acetate produc-tion was not greatly affected.68 Acetate had to beadded in much greater concentration than thepropionate to affect acetate production. So in adigester some fermentation products may controltheir own formation if they tend to build up, andthus digester conditions may be helped towardstability. In particular these experiments suggest amechanism for minimizing propionate productionand maximizing acetate production in digesters.Another factor having the same effect is discussedbelow.

    Recent work on the metabolism of hydrogen inmixed cultures is very relevant to theories ofdigester function. In anaerobic fermentation ener-gy for the bacteria is ultimately derived fromoxidation reactions, as it is in aerobic bacteria, butas no oxygen is available the hydrogen ions mustusually be added to some other electron acceptorto allow the oxidation reactions to proceed. For asmall number of bacteria the electron acceptor canbe an inorganic ion such as nitrate or sulphate, andsulphate-reducing bacteria have been demonstratedin digesters (Section IV.). But the majority offermentative bacteria cannot transfer electrons inthis way, so the electrons are transferred to one ofthe products of the carbohydrate degradation (forinstance pyruvate to make the more reducedlactate) or are released as hydrogen gas by ahydrogenase system. The accumulation of hydro-gen in solution and in the gas phase could inhibitgrowth of hydrogen-producing bacteria by a feed-back mechanism, just as propionate or acetateinhibits its own production (see above). Theutilization of metabolic hydrogen might thenenhance the rate of bacterial growth and theprimary fermentation in the digester. But recentexperiments with rumen bacteria have shown thatutilization of metabolic hydrogen can alter fer-mentation patterns. Iannotti, Kafkewitz, Wolin,and Bryant69 found that a glucose-limited con-tinuous culture of a strain of the cellulolyticRuminococcus albus formed as fermentation pro-ducts 69 mol of ethanol, 74 of acetate, 237 ofhydrogen and an unmeasured amount of carbondioxide/100 mol glucose fermented. Vibrio suc-cinogenes grows by reduction of fumarate withhydrogen to form succinate. In a continuous

    culture of the two bacteria on glucose plusfumarate the only products formed were 147 molacetate and 384 mol succinate. Each mole ofsuccinate was equivalent to the production of onemole of hydrogen by R. albus. The utilization ofhydrogen formed by R. albus increased the pro-duction of hydrogen at the expense of ethanolformation thus leading to the formation of moreacetate.

    Chung70 found that accumulation of hydrogeninhibited the growth of the hydrogen-producingClostridium cellobioparus, although he consideredthat the inhibition could not be explained purelyby a mass action effect, as suggested above.Growth of the Clostridium with the hydrogen-utilizing Methanobacterium ruminantium (undercarbon dioxide) increased the growth of theClostridium and led to production of more aceticacid and less lactic acid, butyric acid, and ethanolthan in pure culture. Again utilization of hydrogenled to diversion of hydrogen from the productionof reduced end products.

    The bacteria used in these experimentsproduced hydrogen as part of their normal fermen-tation products. With other bacteria no gaseoushydrogen can be detected and all "hydrogen" (orreduced nucleotides) can be accounted for asreduced fermentation products such as butyric orpropionic acids. Whether these bacteria do have ahydrogenase system which in the presence ofhydrogen-utilizing bacteria can produce gaseoushydrogen and so less reduced end products, orwhether bacteria such as the methanogenic onescan utilize hydrogen on some form of electron-carrier which can be transferred between cells isunknown. But what has been shown in theauthors' laboratories (Henderson, unpublished) isthat the reverse reaction can take place withbacteria which normally do not produce hydrogen.Selenomonas ruminantium and Anaerovibriolipolytica growing on, respectively, glucose orfructose under carbon dioxide produced acetic andpropionic acids in a 1:1 ratio. Under a hydrogen-carbon dioxide atmosphere ratios of acetate topropionate of 1:2 and 1:4 were obtained, indi-cating an uptake of hydrogen into fermentationproducts. And with Bacteroides ruminicolagrowing under hydrogen-carbon dioxide anincrease succinate production also indicatedhydrogen uptake.

    The implications of these experiments indigester running are then twofold. Although the

    148 CRC Critical Reviews in Environmental Control

    Dow

    nloa

    ded

    by [

    Cor

    nell

    Uni

    vers

    ity L

    ibra

    ry]

    at 2

    2:20

    13

    Nov

    embe

    r 20

    14

  • digester bacteria in pure cultures produce mixturesof fermentation products, in the complete digestersystem hydrogen utilization by the methanogenicbacteria will pull the carbohydrate fermentationtowards maximum acetic acid production andmaximum hydrogen production. Hydrogen utiliza-tion may also enhance the breakdown of longchain fatty acids and alter purine metabolism.Thus the preponderance of acetate as residual acidin digesters is explained, as are also the changes inacid found on starting up a digestion. For instance,in piggery waste digesters started from water,acetic, propionic, and butyric acid concentrationsbuilt-up, the proportions of acids being about10:3:2. With the onset of methanogenesis acidconcentrations fell and the residual acid graduallychanged to acetic with no detectable propionicacid or butyric acid.3S>63 Propionic acid, how-ever, did remain in the same concentration forsome three or four weeks after gas productionstarted and acetic acid concentrations fell. At thistime, although it was in small amount it wasactually the predominant acid.

    The formation of acetate from pyruvate can beaccompanied by formation of formic acid insteadof hydrogen plus carbon dioxide. As formic acid isalso a substrate for methanogenic bacteria itsutilization could have the same consequences asutilization of hydrogen. However, as previouslysuggested, the digester pH might tend to favorhydrogen plus carbon dioxide formation and notformic acid. Also, Iannotti et al.69 found thatformate was not formed in preference to hydrogenby the Ruminococcus in the continuous cultureexperiments mentioned above. Formate wasformed only in batch cultures with high glucoseconcentrations; conditions unlike those indigesters.

    A further route by which hydrogen may bemetabolized, although this is probably minorcompared with methane production, is by conver-sion of hydrogen plus carbon dioxide into aceticacid. Chynoweth and Mah7 found labeled acetateto be produced from radioactive carbon dioxide inthe previously mentioned experiments on fattyacid degradation. This reaction has been shownwith some clostridia and Hobson and Summers(unpublished) have found evidence from fermenta-tion balances that it occurs under certain condi-tions in mixed continuous cultures of two(nonclostridium) rumen bacteria of the type whichmight occur in digesters.

    On the other hand if conversion of hydrogen tomethane fails due to inhibition of the methanebacteria, or there is over production of hydrogen,then this will push the primary fermentation in thedirection of increased production of the reducedacids. Utilization of "hydrogen" to form reducedacids means that hydrogen need not necessarilyappear in the gas phase. In the authors' experi-ments on piggery waste digestion difficulty wasexperienced on changing the digester (100 1capacity) input from one type of piggery waste toanother. During the period after the change gasproduction fell off roughly exponentially, sug-gesting death and a washout of the active bacteriaby the new waste, possibly because of its lowersolids content. On continued loading gas produc-tion fluctuated between zero and a few liters aday, falling off with the build-up in acid concen-tration. The pH varied between 6.6 and 7.1 andvolatile acid concentrations were as high as about7,000 ppm when the pH was lowest. At one pointwhere gas production was not measurable the gasin the digester was still 72% methane, 28% carbondioxide with negligible hydrogen, but 94.3% of thevolatile acid concentration of 2,100 ppm waspropionic acid, and only 4.3% was acetic (butyricacid was the remaining 1.4%). The pH was 6.9 to7.0. Digestion was later recovered and the residualacid returned to its normal concentration of about300 ppm by changing the waste to one of highersolids content.

    The Methanogenic Stage of Digestion andInhibition of DigestionMethanogenesis

    In this stage the originally polluting materialsof the waste are finally converted to nonpollutinggases. This is done at the expense of convertingsome of the soluble digester products into micro-bial cells, but as methanogenesis is inefficient as asource of energy for cell synthesis the weight ofbacteria added to the sludge is comparatively smalland some of the ammonia of the digester liquid isremoved in forming the bacteria.

    In the primary stage of digestion there is abreakdown of the incoming waste material by amixed population of bacteria. This mixed popula-tion of bacteria is partly fortuitous as the feed tothe digester not only contains a very large popula-tion of bacteria from its original source if it besewage, but the methods of running digesters ofnecessity allow for exposure of the feed material

    June 1974 149

    Dow

    nloa

    ded

    by [

    Cor

    nell

    Uni

    vers

    ity L

    ibra

    ry]

    at 2

    2:20

    13

    Nov

    embe

    r 20

    14

  • to contamination from many sources. However,although a number of the types of bacteria foundin digesters may be of no apparent use in theprocess, the type of heterogeneous materialnormally used as substrate in digestion demands amixed bacterial flora, with a wide spectrum ofenzymic activities for its complete breakdown.Although each of the bacteria when alone mayproduce a range of fermentation products, theprevious sections have shown that the conditionsunder which the bacteria are growing in thedigester may induce the bacteria of the mixedpopulation to produce a limited number of endproducts, and these form the substrates for thesecond stage of the process, methanogenesis. Thedetailed biochemistry of the formation of methaneby the methanogenic bacteria has, at least in thecase of the substrates hydrogen plus carbondioxide or formate, been elucidated to a greatextent over the past few years. However, this isnot really relevant to the biochemistry of anaer-obic digestion. What will be considered here arethe overall reactions which are likely to be ofimportance in digester function. For the detailedbiochemistry the reader is referred to the originalpapers and the detailed review of Wolfe71 sum-marizing the position in 1970.

    In the primary stage, nitrogen metabolismproduced by breakdown of the various nitro-genous constituents of the feed wastes, aminoacids, and ammonia, which acted as nitrogensources for growth of the bacteria of this stage.And these same bacteria provide the nitrogensource for growth of the methanogenic bacteria, asammonia seems to be the main, or only, nitrogencompound which can be used by the knownmethanogenic bacteria.72 Since the methanogenicbacteria can only grow under conditions of verylow Eh and have no ability to produce this low Ehthemselves, they are also dependent on somecomponents of the primary flora to producesuitably reduced conditions for growth.

    Thus the methanogenic bacteria are completelydependent on the primary stage bacteria forgrowth in a digester. On the other hand, though,the activities of the methanogenic bacteria cancontrol the primary fermentation. Although therewere early reports that methane could be pro-duced from a large number of substrates, thesereports were later shown to be the result of workwith mixed cultures. Further work narrowed downthe range of substrates for methanogenic bacteria

    to a number of the lower volatile fatty acids, thetwo lower primary alcohols and hydrogen pluscarbon dioxide. Even more recent work73 hasshown that the apparent utilization of ethanol bya methanogenic bacterium (Methanobacillusomelianskii) was the result of the long accepted"pure culture" of the bacterium being a symbioticassociation of two bacteria. One bacterium oxi-dizes ethanol to acetate and hydrogen and theother, the actual methane bacterium, reduces thecarbon dioxide atmosphere of the culture mediumwith the hydrogen to form methane. Utilization ofthe hydrogen is essential for continued growth ofthe first bacterium. This result and the fact that anumber of reported isolates of methanogenicbacteria, even if they had once been in pureculture, then no longer existed led Wolfe in areview71 of the methanogenic bacteria to con-clude that only eight methanogenic bacteria wereknown and were in pure culture at that time(1970) and the only substrate utilized was hydro-gen plus carbon dioxide or formate, except for onebacterium Methanosarcina barkerii, which usedhydrogen plus carbon dioxide, methanol, oracetate.

    Methane with carbon dioxide are the principalcomponents of the rumen gas. The proportionsvary, but are usually about 30% methane, 70%carbon dioxide; only very small traces of hydrogenare usually found. The only methanogenic bacteriaisolated by dilution techniques in pure culturefrom the rumen are Methanobacterium ruminan-tium and Methanobacterium mobilis."1*'15 Thesetwo bacteria occur in high numbers and theyutilize only formate or hydrogen plus carbondioxide. Little formate is probably formed inrumen contents and the rumen methane can beaccounted for by production from hydrogen pluscarbon dioxide, and the numbers of the twobacteria mentioned are sufficient to account forthis production. Although acetic, propionic, andbutyric acids are formed in the rumen there isevidence for only little, if any, formation ofmethane from these substrates. By injection of14C-labeled acetate into the rumen, Oppermann,Nelson, and Brown76 found that, under theconditions of feeding of their experimental cattle,about 2 to 5.5% of the rumen methane might havebeen derived from the methyl or carboxyl groupsof the variously labeled acetates used. By enrich-ment culture techniques carried out over severalweeks of addition of acetate to rumen contents in

    150 CRC Critical Reviews in Environmental Control

    Dow

    nloa

    ded

    by [

    Cor

    nell

    Uni

    vers

    ity L

    ibra

    ry]

    at 2

    2:20

    13

    Nov

    embe

    r 20

    14

  • vitro Oppermann, Nelson, and Brown77 obtainedcultures which produced methane from formate oracetate. By transfer in agar media and then intoliquid culture two bacteria which fermented for-mate or acetate were obtained and these weretentatively identified as Methanobacterium for-micium and a possible variety of Methanobacter-ium shngenii (isolated by Barker78). It is notclear whether any cultures were kept when theformation of methane had been demonstratedafter four or five subcultures, but the properties ofthe bacteria were not examined in detail. Theseauthors also observed that in the mixed culturesthe presence of formate inhibited acetate utiliza-tion and with the mixed substrates acetate was notutilized until formate had been dissimilated.Carbon dioxide also reduced acetate utilization.These same workers also isolated cultures ofbacteria from rumen contents, by enrichmentmethods, which produced methane from butyricand valeric acids,79 but again the cultures werenot examined in detail. However, all attempts toobtain cultures utilizing propionate failed.

    In Section IV the reported isolation fromanaerobic digesters of bacteria producing methanefrom all the lower fatty acids is described, as wellas some utilizing hydrogen plus carbon dioxide.The position some 18 years ago regarding thespecies of methane bacteria and their reportedsubstrates is well summarized by H. A. Barker, oneof the pioneer workers in the field.80 This may becontrasted with the list of methane bacteria in thepreviously mentioned review of Wolfe.71 Thediscrepancies and uncertainties in the field of themethane bacteria are due to the difficulty ofgrowth and culture of the bacteria and especiallythe difficulties of keeping many of them in culturefor long periods so that definitive tests may bemade. What is certain is that the hydrogen- orformate-utilizing bacteria are the easiest to isolateand probably have the fastest growth rate of themethane bacteria. These bacteria undoubtedlyexist in pure culture and a number have beenextensively investigated. The position of bacteriautilizing the other substrates is indefinite, butthere seems to be overwhelming evidence for theexistence of some bacteria or symbiotic associa-tions of bacteria which will produce methane fromC2 to C5 volatile fatty acids. Even if the bacteriano longer exist in culture the formation ofmethane from these substrates has been demon-

    strated in the content of the rumen or digesters,and in mixed or apparently pure cultures.

    The fermentation of 14C-labeled benzoic acidto methane and carbon dioxide by an enrichmentculture of methane bacteria from digesting sewagesludge was demonstrated by Nottingham andHungate,81 thus confirming the earlier observa-tions of Fina and Fiskin82 that anaerobic splittingof the benzene ring is possible. In this case thereaction agreed approximately with the equation 4C6H6CO2 + 18H2O -> 15CH4 + 13CO2 and thedistribution of the radioactivity suggested that thebenzene ring was converted to methane andcarbon dioxide directly without the intermediateformation of other substrates or participation ofcarbon dioxide from the culture atmosphere. Thissuggested that only one bacterium might beinvolved, but of course, it was impossible toconfirm this with the mixed cultures used. Theimportance of this reaction in digestion is difficultto assess because it is not known if aromatic ringslinked in more complex molecules or with otherthan the carboxyl side chain can be degraded. Itdoes suggest that phenolic trade wastes in sewagemight be digested and there is also the possibilityof some digestion of lignin in vegetable or farmanimal wastes. A slight decrease in "lignin" wassuggested in the authors' experiments with piggerywaste,63 but the methods of determining "lignin"are so imprecise that the results cannot be taken asproof of the degradation of the aromatic ring.

    Since the volatile acid fermentation productsdo not accumulate in functioning digesters, but dobuild up if methanogenesis ceases, it has seemedobvious that these fermentation products must bethe substrates for methane formation. Acetate hasbeen generally regarded as the major substrate formethane production, the work of Smith andMah83 for instance showing that 73% of themethane formed in their experiments (with labeledacetate) came from acetate in a domestic sewagedigestion. The observations discussed in the pre-vious section suggested that the digester fermenta-tions are biased towards acetate production.Hydrogen plus carbon dioxide are also substratesfor methane formation and these two gases areformed in equimolar amount with acetate in thedissimilation of pyruvate during carbohydrate fer-mentation. However, the fermentation of a carbo-hydrate to acetate, hydrogen and carbon dioxidecan be represented in outline as follows:

    June 1974 151

    Dow

    nloa

    ded

    by [

    Cor

    nell

    Uni

    vers

    ity L

    ibra

    ry]

    at 2

    2:20

    13

    Nov

    embe

    r 20

    14

  • C6 H, 2 0 -> 2CH3 COCOOH + 2H22CH3 COCOOH + 2H, O - 2CH3 COOH + 2H2 + 2CO24H2 + CO 2 -*CH 4 + 2H2O

    2CH3COOH->2CH4 + 2CO2C

  • in digester contents. The hydrogen-utilizingbacteria show comparatively rapid growth in liquidmedia, a few days being sufficient for mostcultures. In the authors' experience the acetate-utilizing bacteria are also more susceptible tounexplained death than are the hydrogen-utilizers.

    Whether the apparent rate of growth of theacetate-utilizing bacteria is an artifact of inade-quate media or whether they are generally slowgrowing even in the mixed culture of the digesteris important in considering the rate-limiting re-action in digester operation. It was argued pre-viously that cellulolysis could be the rate-limitingreaction in many digestions. This was done on thebasis of the relatively rapid growth of the hydro-gen-utilizing bacteria compared to cellulolyticbacteria in cultures. The apparent rate of growthof the acetate-utilizing bacteria is, however, slowerthan that of the cellulolytic bacteria growing onprepared cellulose in cultures. Growth of theacetate-utilizing bacteria in culture suggests amaximum growth rate measured in days ratherthan hours as in the case with hydrogen-utilizers.This conclusion is supported by the work ofMcCarty88 who determined the rates of break-down of various substrates in laboratory digestersseeded with domestic digester sludge, but adaptedto a synthetic feed containing the substrate. Hefound that the efficiency of conversion of acetateor propionate to gas fell off rapidly as the turnovertime was reduced below 8 days and approached90% only at about 16 days. With formate, whichcan be equated with hydrogen plus carbondioxide, efficiency was still about 90% at atwo-day turnover time (the minimum tested).These results also explain why hydrogen pluscarbon dioxide or formate are the only, or by farthe major, substrates for methanogenesis in therumen, which has a turnover time of about a day.McCarty also showed that long chain fatty acidsfrom hydrolysis of glycerides added to laboratorydigesters were almost totally degraded only atturnover times of ten days or more, but thepreviously-mentioned work of Chynoweth andMah7 suggested a more rapid breakdown. Heconcluded that methanogenesis was the rate-limiting step in anaerobic digestion, as have others.On the other hand Hammer and Borchardt con-cluded that hydrolysis and fermentation of inputwas the rate-limiting step (see below). However,the whole question of rates of growth of bacteriain their natural habitats, and especially in situa-

    tions such as a digester, is difficult to determine asit is affected so much by the association of otherbacteria and by the state of the substrates. Forinstance the growth of cellulolytic bacteria onprepared cellulose in cultures was just mentioned.This will be faster than growth on natural cellulose(see previous discussion). In a digestion growth ofcellulolytic bacteria on the treated cellulose oftoilet and other papers in domestic sewage will bemore rapid than on the natural cellulose ofundigested plant fibers in farm animal sewage. It isalso difficult to equate completely the results ofdifferent experiments. Not only may substratesvary, some being pure compounds, some true wastematerials, but the bacteria may vary, as even inexperiments with mixed cultures in digesters, insome experiments the bacterial population willhave been modified by continued feeding of aparticular substrate. Some digester inputs havebeen waste plus the compound under test, somesynthetic mixtures made up of protein hydroly-sates and so on. The only conclusion that might bedrawn is that turnover times will vary withdifferent wastes, but that about ten days isprobably the minimum time for a digester treatingthe usual municipal or farm wastes.

    Failure of conversion of acetate to methane dueeither to death or washout of the acetate-utilizingbacteria or an increase in production of acetate sosudden that the slow growing acetate-utilizerscould not adjust their numbers to cope with it(e.g., a sudden addition of fermentable carbohy-drate) could, however, set in train digester failure.Results quoted previously showed that accumula-tion of acetate could inhibit acetate production bybacteria. This would then diminish the productionof the hydrogen that accompanies acetate forma-tions. Alternately, the growth of the bacteriacould be maintained by moving the fermentationpattern towards the production of more pro-pionate, succinate, or lactate which are propionateprecursors in mixed cultures (see previous discus-sion). This would not only diminish the hydrogenavailable to the hydrogen-utilizing bacteria butalso propionate has been shown to be an inhibitorof the hydrogen-utilizing methanogenic bacteriaand so methane production by these bacteriawould diminish. Accumulation of propionatecould probably be helped by the fact that theremay be little formation of methane from pro-pionate, although dilution counts of propionate-utilizing bacteria from municipal digesters have

    June 1974 153

    Dow

    nloa

    ded

    by [

    Cor

    nell

    Uni

    vers

    ity L

    ibra

    ry]

    at 2

    2:20

    13

    Nov

    embe

    r 20

    14

  • been reported, (e.g., see Reference 85). It has al-ready been mentioned that no evidence for forma-tion of methane from propionate by the rumenbacteria was found, although butyrate-utilizing bac-teria were obtained. In the authors' laboratoriesShaw35 '39 demonstrated methane productionfrom butyrate in dilution cultures from piggery-waste digesters, but propionate was not utilized. Inthe enrichment cultures from piggery-waste di-gesters (Section IV.) utilization of propionateseemed to be a property of only a minority of the

    cultures. And, it has already been mentioned thatpropionate was the last of the three volatile fattyacids to diminish in concentration during build upof digestion. Nelson et al.79 in their experimentson rumen methanogenesis found that whilebutyrate was evidently fermented via acetate, asduring the fermentation acetate appeared and thendisappeared, valerate fermentation did not go tocompletion as the reaction appeared to proceed viaacetate and propionate, and the latter accumu-lated. The reactions can be represented as follows:

    2CH3 CH2 CH2 COOH + 2H2 O + CO2 -* 4CH3 COOH + CH4 -> 4CH4 + 4CO + CH42CH3(CH2)3COOH + 2H2O + CO2 - 2CH3CH2COOH + 2CH3COOH + 2CH4

    2CH3 COOH - 2CH4 + 2CO2

    Accumulation of acids would then tend to inhibitthe fermentative and the methanogenic bacteriaand inhibition of all bacterial action would beincreased by any fall in pH which accompaniedaccumulation of acid and the digestion wouldgradually "sour."

    This argument suggests a way in which failureof one set of reactions could cause digester failure.The initial circumstance which caused the re-actions was either increase in turnover rate of thedigester or sudden increased loading of the fer-mentable substrate. These would both lead to thesame gross symptoms of decreased gas productionand increased acid concentration. From analysis ofdigesters it is generally impossible to tell which iscause and effect, increased acid concentration ordecreased gas production. The previous discussionshows that reactions could be complex. However,there is experimental evidence for inhibitions ofsome specific reactions in digester failure, andevidence that some particular materials do inhibitdigestion, if not by a specific action, by action onthe digester bacteria in general.

    The easiest to measure parameter of digesterfunction is total gas production. In the authors'experience a change in composition of the gas, i.e.,decrease in methane and increase in carbondioxide, giving the usual total gas volume, seldom",if ever, takes place in a functioning digester. Adecrease in total gas volume always precedesdigester failure (except as mentioned below).Griffiths89 suggested that increase in carbondioxide indicated incipient digester failure. On theother hand, the sudden addition of a large amountof fermentable material could lead to a very rapidincrease in gas production and a greater than usualproportion of carbon dioxide. The exception to

    the rule mentioned in the previous sentence hasbeen found by the authors only in the case of apilot plant where the digester was heated bycirculating the contents through an external heatexchanger. An initially undetected minute hole inthe pipe work was sufficient to draw appreciableair into the system because of the rapid circulationof digester contents. No apparent decrease in gasproduction was noted at first, but the gas becamenonflammable. Analysis a little later showed thatthe gas in the head-space of the digester, instead ofits previous composition of 78.2% methane, 21.8%carbon dioxide, with negligible hydrogen andnitrogen and no oxygen, was then 5.2% methane,23.1% carbon dioxide and 71.7% nitrogen, with nooxygen. The oxygen in the air leaking into thesystem had been metabolized by the bacteria.However, at this stage the oxygen-metabolizingcapacity of the bacteria was evidently at its limitbecause some five hours later the gas in thedigester had almost the same oxygen to nitrogenratio as air, and methane and carbon dioxide werevery low. The leak in the system was thendiscovered and stopped. But over the next fourdays gas production was virtually nil. Loading was,however, continued and gas production graduallyincreased. About eight days of increasing gasproduction was needed to flush all the residualnitrogen out of the system (the oxygen notedabove had been metabolized by a few hours afterthe leak had stopped). However, the build-up ofmethanogenesis could be shown by the change inproportions of methane and carbon dioxide in thegas. Taking these nonnitrogen components of thegas as 100%, the methane and carbon dioxideconcentrations were, a day or two after small gasproduction was apparent 35% and 65%, respec-

    154 CRC Critical Reviews in Environmental Control

    Dow

    nloa

    ded

    by [

    Cor

    nell

    Uni

    vers

    ity L

    ibra

    ry]

    at 2

    2:20

    13

    Nov

    embe

    r 20

    14

  • tively; next day 40% and 60%, two days later 54%and 46%, three days after this 68% and 32%. Theresidual nitrogen was then very small and waswashed out by this normal digester gas in thecourse of the next few days.

    These results demonstrate the previously men-tioned capacity of the mixed digester flora to takeup oxygen, but they also show that there is a limitto this process and that too much oxygen will raisethe Eh of the digester contents to such an extentthat microbial metabolism will practically cease.However, some metabolism must still carry on toagain reduce the digester contents sufficiently foranaerobic metabolism to start, and after a fewdays methanogenesis starts. In this case loadingwas continued as the digester contents werenormal and it was felt that substrate was needed torestore microbial activity, so this was an exceptionto the general "rule" that allowing a digester to"rest" without loading for a few days is the firstand probably the best way to try to restoredigestion. That the digester bacteria will remainviable for days, or even weeks has been shown bythe authors' piggery-waste digesters, which eitherbecause of experimental changes in design or oncebecause of an outbreak of swine vesicular diseasehave been stopped with or without heating, forperiods of up to six weeks. On reloading at a lowrate, digestion and gas production started againalmost immediately and normal loading was re-sumed in a week or two.

    It seems unlikely that any mixed continuousculture system will attain the steady state of apure continuous culture. With a continuous cul-ture of two rumen bacteria growing at constantdilution rate for many weeks in a medium withurea as sole nitrogen source, and in which only onebacterium was ureolytic, although both could useammonia, oscillations in proportions of the twobacteria occurred (Hobson and Summers, un-published). This dependence of one bacterium onanother for nutrients is a situation found indigesters. In the similar rumen habitat both short-term and long-term fluctuations in relative num-bers of different bacterial species have been found(see Reference 65). The short-term fluctuationsoccur over a few hours after addition of a feed asbacteria dependent on either the primary feed orits breakdown products grow. Fluctuations alsotake place over some days, and long-term fluctua-tions occur over weeks or months. These lattermay be due to the bacteria in question not being

    completely adapted to life in the rumen; thepopulation is only being kept up by continuedinoculation with the feed.

    Thus oscillations in a digester population wouldbe expected and so would oscillations in theparameters used to measure the performance ofthe digesters. Such oscillations are found even inexperimental digesters, like those run by theauthors, where input is kept as constant as possibleover periods of months. Since there are theseoscillations an increase in volatile fatty acidconcentration, or a decrease in gas productionoccurring over a day or two is not necessarily asign of imminent digester failure. Unless thesechanges increase over some days then no actionneed be taken. In the authors' digesters volatileacids sometimes built up to as much as 2,800 ppmfor a day or two, and sometimes as much as 1,000ppm acids were present for some days, althoughthe normal fluctuations were between 100 and500 ppm. The former values were sometimesconnected with exceptionally high concentrationsof volatile acids in the input waste. The feedsludge to a piggery-waste digester can contain upto 8,000 ppm volatile acids, but such sludge isadequately digested. However, the principal acid inthe feed sludge is acetic, a typical analysis being63.6% acetic, 18.0% propionic, 9.8% butyric, 4.2%isobutyric, 1.4% valeric, 2.8% isovaleric, and thehigh acid concentrations in the digesters men-tioned above were principally acetic. It wasmentioned previously that in another case, wheredigestion failed and high acid concentrations werepresent, the principal acid was propionic. Pohlandand Bloodgood34 also found that during retardeddigestion propionic acid concentrations increasedrelative to acetic and butyric acids, and theyconcluded that the "mechanism of degradation ofpropionic acid to either gas or lower acids wasaffected to the greatest degree."

    The difficulty in assessing the reports of acidinhibition of digestion is then composed of thefact that the accumulated acid may be only asymptom of some initial inhibition of methano-genesis and, although acid concentrations of about2,000 ppm have been suggested as inhibitory todigestion (e.g., see Reference 90), the compositionof the acid mixture may be more important thanthe total concentration of acid. McCarty andMcKinney9 ! observed that high concentrations ofacetic acid were not inhibitory to digestion, inaccordance with the observations above. But the

    June 1974 155

    Dow

    nloa

    ded

    by [

    Cor

    nell

    Uni

    vers

    ity L

    ibra

    ry]

    at 2

    2:20

    13

    Nov

    embe

    r 20

    14

  • work of Andrews92 suggested that propionate wasinhibitory to methanogenesis in laboratorydigesters. Shaw35 showed that the growth ofMethanobacterium formicicum was not inhibitedby up to 10,000 ppm (as acetic acid) of sodiumacetate or butyrate, but that sodium propionate atall concentrations from the lowest studied (1,000ppm as acetic acid) was inhibitory, in that al-though growth of the bacteria eventually com-menced, the lag phase was extended as thepropionate concentration was increased. Thesecultures were all at pH 7.1. As previously men-tioned, M. formicicum appears to be one of themost common hydrogen-utilizing bacteria indigesters. The utilization of volatile fatty acids asgrowth factors by the nonmethanogenic bacteriahas already been mentioned and recent work hasshown that acetate is, contrary to being inhibitory,a growth factor for some, at least, of the hydro-gen-utilizing methanogenic bacteria. Bryant etal.72 found that a rumen strain of Methanobac-terium ruminantium formed about 60% of its cellcarbon from acetate and needed between about900 and 1,200 ppm acetate in batch media foroptimal growth. A digester strain of M ruminan-tium and the hydrogen-utilizing bacterium fromthe symbiotic association Methanobacillusomelianskii also required acetate.

    The question of inhibition of methanogenesisfrom substrates other than hydrogen, or the acidsas growth factors for the bacteria concerned, isobviously important, but such work awaits theisolation and growth in stable cultures of theappropriate bacteria. It was previously mentionedthat formate inhibited production of methanefrom acetate.

    Apart from the effects of acid concentrationthere is also controversy about the effects of pHeither per se or as a factor in acid inhibition ofdigestion. The fall in pH of a digester due to acidaccumulation will be related to the bufferingcapacity of the digester contents and not directlyto the acid concentration. However, a compara-tively small fall in pH from the usual value ofabout 7 will inhibit bacterial action. Smith andHungate 7 4 showed that Methanobacteriumruminantium, a hydrogen-utilizing bacteriumfound in digesters as well as the rumen, that therange of growth of the bacterium was from aboutpH 6.5 to 7.7. The growth of a number of thebacteria most concerned in the digestion of thenormal roughage feed of ruminants also diminishes

    rapidly when the pH falls below about 6.5, and theminimum pH for growth is about 6.0.57 Thesebacteria are similar to those found in digesters. ApH as low as 6.5 usually indicates a faileddigestion.

    The results above were obtained in cultureswhere the pH was altered by addition of mineralacids, alkalis, or buffer salts. Changes in pH due tothe volatile fatty acids are also accompanied bychanges in the concentration of ionized andunionized forms of the acids and these forms mayhave different inhibitory effects. The experimentsof Andrews92 for instance, indicated that union-ized propionic acid was more inhibitory thanionized propionic acid, as inhibition was greater atpH 6.5 than at pH 7.2. The greater inhibitoryeffect of volatile acids as the pH is lowered hasbeen noted with nonmethanogenic bacteria (e.g.,see Reference 64 for a review of some papers), andthere is also evidence that propionate is more in-hibitory than acetate or butyrate (e.g., see Refer-ence 93).

    The question of inhibition of digester functionby the lower volatile fatty acids is, therefore,complex and whether any general rules about thenormal concentration of acid in a digester can begiven appears doubtful. Also called into question isthe correct procedure for recovery of a "sour"digester. Increasing the pH to more normal levelsof just over 7 could help, as this will decreaseadverse effects of the acid conditions and willprobably decrease the inhibitory effects of theacids themselves. This raising of pH could helpnatural recovery. Natural recovery often occurswhen loading is stopped, but a continued loading,possibly at a lower than usual rate, could helprecovery in that it would dilute the inhibitoryconcentrations of acids in the digester andprobably help to stabilize the pH. However, if thelow pH were not caused by a unique event, cyclingof digestion could then occur. Hobson65 showedthat cycles of growth occurred in a pure con-tinuous culture where the pH was controlled onlyby the buffering action of the medium. A low pHand decreased bacterial activity was followed by aperiod of increasing bacterial growth as the pH wasrestored by input of fresh medium, but maximumbacterial activity again lowered the pH and so thecycle started again. A situation where continuedloading appeared the best treatment has beenmentioned, but here there was no question of acidinhibition. Adjustment of pH by addition of limehas been the subject of debate. A recent review of

    156 CRC Critical Reviews in Environmental Control

    Dow

    nloa

    ded

    by [

    Cor

    nell

    Uni

    vers

    ity L

    ibra

    ry]

    at 2

    2:20

    13

    Nov

    embe

    r 20

    14

  • the subject was given by Kirsch and Sykes1 andthe general opinion seemed to be that addition oflime was not the best, although perhaps thecheapest, method of restoring the pH of a digestor,because of the formation of insoluble calciumcarbonate. Addition of bicarbonate appeared to bea better method of dealing with digester acidity.These views are also supported in a recent paperby Mosey et al.94 Ammonium salts are contrain-dicated because of the dangers of ammoniatoxicity.

    As discussed earlier, ammonia in digesterscomes either directly from the input or from thebreakdown of proteins and other nitrogenouscompounds in the input. The ammonia is used as anitrogen source by the fermentative bacteria of theprimary stage of breakdown and by the methano-genic bacteria, or at least by the hydrogen-utilizingmethanogenic bacteria which have been studied indetail.72 As in the case of the volatile acids, anincreased ammonia concentration in a digestermay be only a secondary symptom of failure ofgrowth of the bacteria and not a primary cause offailure, but there is evidence that high ammonialevels can cause inhibition of the methanogenicbacteria. Again as in the case of the volatile acids,digester pH can affect the concentrations of freeammonia and ammonium ion in the digestercontents and there is some evidence that thesehave different effects on bacteria. But since totalammonia nitrogen is usually measured, toxic totalammonia concentrations are probably of mostinterest for digester control. McCarty95 statedthat if the concentration of ammonia nitrogen isbetween 1,500 ppm and 3,000 ppm then at pHvalues of 7.4 and above ammonia itself is inhibi-tory, but if the concentration is above 3,000 ppmthen the ammonium ion is toxic regardless of pH.Since the input to the authors' piggery-wastedigesters contains rather high concentrations ofammonia nitrogen (about 1,200 to 1,800 ppm onaverage and sometimes as high as 2,500 ppm), andthis ammonia is not all utilized in the digestion(the digester concentrations vary from about1,100 to 1,600 ppm), Shaw35 investigated theeffect of ammonia on growth of Methanobac-terium formicicum. At a culture pH of 7.1 little orno inhibition of growth and methanogenesis wasobserved at concentrations of 1,000 and 2,000ppm ammonia, but at 3,000 ppm ammonia (2,466ppm ammonia nitrogen) about 40% inhibition ofgas production was observed and at 4,000 ppm

    ammonia complete inhibition of growth and gasproduction took place. These results are in accordwith those of McCarty, and also show that theammonia concentrations in the piggery-waste di-gesters are below toxic levels.

    Other constituents of waste or products ofdigestion of the waste could be inhibitory todigestion. For instance, long chain fatty acids havebeen shown to have inhibitory effects on bacterialgrowth, and long chain acids have been shown toinhibit methane production and affect othermetabolic pathways when added to the rumen (seeReference 96 for reference to original papers).Henderson,96 in the authors' laboratories, ex-amined the effects of long chain acids on somepure cultures of rumen bacteria, using capric,lauric, myristic, palmitic, stearic, and oleic acids.None of the acids inhibited a number of theGram-negative rumen bacteria producing propio-nate or succinate, and this was in accord with theobservation that long chain acids added to rumencontents increased the proportion of propionate inthe rumen acids. The Gram-positive to Gram-variable cellulolytic Ruminococcus (which pro-duce acetic acid but not propionic) was inhibitedby all acids, most strongly by lauric. A strain ofButyrivibrio (a Gram-negative bacterium producingacetic and butyric acids) was inhibited by stearic,palmitic, and myristic acids, but oleic, lauric, andcapric acids stimulated growth at concentrationsof less than 100 mg/1 and inhibited at higherconcentrations. The Gram-positive Methano-bacterium ruminantium was inhibited by all theacids, especially oleic, as in the case of theRuminococcus and Butyrivibrio. The inhibitoryconcentrations of acids used in this work werefrom 5 mg/1 upwards, marked inhibition wasfound at 50 mg/1. Prins et al.97 also showedinhibition of M. ruminantium by long chain acids,but not by the corresponding glycerides. AlthoughM. ruminantium is found in digesters, the ap-parently more common M. formicicum is Gram-negative and may not be affected to the sameextent as M. ruminantium by fatty acids. However,these rumen experiments do show the possibilitiesof inhibition of digestion by long chain fatty acids,as hydrolysis of glycerides and the presence of freefatty acids has been shown in digesters (see above).Heavy metal ions have an inhibitory effect onmost microbial processes and the possibility oftrade wastes getting into domestic digesters and offeed additives such as the copper given to pigs

    June 1974 157

    Dow

    nloa

    ded

    by [

    Cor

    nell

    Uni

    vers

    ity L

    ibra

    ry]

    at 2

    2:20

    13

    Nov

    embe

    r 20

    14

  • getting into farm-waste digesters make the possibil-ities of heavy metal inhibition of digestionpossible. Fortunately these metals, exceptchromium, form extremely insoluble sulphides andwith the digester conditions of hydrogen sulphideformation the metals are likely to be removedfrom solution and so not to interfere with diges-tion. Shaw3 5 found in studies with M. formicicumthat total copper concentrations in excess of 200ppm (of which about 128 ppm would have beenprecipitated as copper sulphide) were needed toshow appreciable inhibition of methanogenesis,and it was concluded that piggery waste of veryhigh total solids would have to be fed to digestersbefore inhibition of digestion due to residualcopper from the pig feed became apparent. Re-ported copper concentrations needed to inhibitanaerobic digestion have varied quite widely (e.g.see References 98, 99). Mosey et al.94 haverecently discussed the detection of heavy metals insewage and the effects of digester sulphide, car-bonate, pH, etc. on their toxicity. They also suggestmethods of treating a digester failing because ofmetal toxicity, of which precipitation of the metalsby addition of sodium sulphide was preferred, al-though soluble sulphide itself in excess of 100mg/1 was toxic to digestion. But they also dis-cussed the effects of operating conditions on metaltoxicity.

    With the possibilities of treating factory wastesby anaerobic digestion other inhibitory factorsmay become apparent. For instance, McNary etal.100 found that D-limonene in citrus peel in-hibited digestion of citrus wastes, and laterBuswell and Mueller101 found that the compoundalso inhibited domestic sewage digestion.

    The questions of inhibition of anaerobic diges-tion are complicated by the mixed flora and themixed constituents of the wastes and theirmetabolic products in the digester. While furtherexperiments on pure cultures of bacteria maythrow light on the mechanisms of inhibition, itwould seem that neither these nor experiments oncomplete digester contents will give more thanvery general rules to be applied to the running ofdigesters. Each digestion will probably show in-dividual characteristics. Similarly, it seems thatfew rules for salvage of an inhibited digestion aregenerally applicable. Experience in runningdigesters with different waste inputs appears to bethe only true guide.

    III. THEORETICAL DIGESTERSYSTEMS AND MODELING OF

    DIGESTER SYSTEMS

    Anaerobic digesters are essentially continuousculture systems, although they may depart inmany ways from the ideal. Since about 1956 whenHerbert, Elsworth and Telling produced a simplemathematical model based on Monod kinetics,which broadly explained the working of a singlestage, continuous fermenter in which one bacterialspecies was growing with one soluble limitingnutrient, the number of models of even this simplesystem has increased yearly. That so many modi-fications to the original model are thought neces-sary, and can be suggested, shows that the be-havior of even one bacterium on a single substrateis far from simple. Although the usual high-ratedigester is a single stage continuous culture itcontains a vast number of species of bacteria. Thegrowth of these bacteria is controlled by threemain reactions: hydrolysis of macromolecular, andoften insoluble substrates, fermentation, or incor-poration into cellular constituents of the productsof this hydrolysis, and attack on the products ofthis fermentation. In addition many of the speciesdepend on others for the provision of growthfactors other than major nutrients or fermentablesubstrates. A further complicating factor is thatnot all the bacteria are in suspension in the"culture medium." Many must be attached toparticulate material and their immediate surround-ings and kinetics of growth will be different fromthose bacteria free in the liquid phase. This allpresents a very complex situation and most of thefactors have been ignored in modeling digestersystems. A model system given one set of data toproduce constants in equations should be able topredict the results of changing some parameter inthe data. The ability to predict will depend on thedata used in forming the model; if this is onlygeneral then the model will not be able to predictthe result of a small change in the input, or achange not covered by the original data. Forinstance a model system could be programmed toforecast the effect of a change in loading repre-sented by a term in COD units, but COD could bemade up in practice of dissolved material, insolu-ble material, or material of different biodegrad-abilities. The model could not predict the effect ofchanging the character of the COD input. A

    158 CRC Critical Reviews in Environmental Control

    Dow

    nloa

    ded

    by [

    Cor

    nell

    Uni

    vers

    ity L

    ibra

    ry]

    at 2

    2:20

    13

    Nov

    embe

    r 20

    14

  • number of models have been tested with labora-tory systems using synthetic, soluble substrates,again the application of such models to systemsusing heterogeneous, partly solid wastes is diffi-cult.

    Although models may have some use in specificcircumstances, it is doubtful whether at thepresent moment any models have done more thanagree to a limited extent with some of the dataalready furnished by experimental work. Theproduction of a model which will produce detailedpredictions of use in the design and running ofdigesters depends on the obtaining of much moredata on the microorganisms and substrates involv-ed in digestion than we have at the moment.

    All anaerobic digesters are varieties of continu-ous cultures of microorganisms, and a very largenumber of papers on the theory and practice ofcontinuous cultures have been published. Al-though these deal with a variety of designs ofapparatus, they are mostly concerned with purecultures growing on dissolved substrates such asare found in industrial applications, but which arenot usually found in waste treatment. It is notpossible to give lists of these papers here, butmany aspects are covered in the proceedings of theinternational symposia on continuous culturewhich have been held in various centers, and thesecan be consulted, e.g., see Reference 102 and pre-vious volumes. Detailed lists of papers covering allaspects of the theory and practice of continuousculture, including industrial applications and wastetreatment, are issued yearly by members of staffof the Czechoslovak Institute of Microbiology,under the title "Continuous Culture of Micro-organisms A review," and are published in FoliaMicrobiologica. The lists include a brief summaryof each paper and the papers are classified undervarious headings.

    Kirsch and Sykes in their review of anaerobicdigestion1 included a detailed discussion onmathematical models and referred to a number ofpapers both on the theory of continuous culturesand its application to modelling of digester sys-tems, and there have been other reviews anddiscussions on the subjects. There seems to be nopoint in repeating these here, so the main pointsdealt with by Kirsch and Sykes will be used as abasis for discussion and amplification in thissection, where some possible anaerobic digestionsystems, and some applications of theory to thesesystems will be described. The practical applica-

    tions of the systems will be dealt with to a greaterextent in the last section.

    The High-rate DigesterThe usual high-rate digester is in theory, if not

    always in practice, a single stage chemostat cul-ture. In the true chemostat a population ofbacteria is maintained in a steady state at aparticular growth rate, which is equated with thedilution rate (or turnover rate) of the medium ofthe culture, by the concentration of one growth-rate-limiting nutrient. It has been shown thatwithin limits the concentration of substrate enter-ing the culture vessel governs the bacterial densityobtained, while the residual concentration ofsubstrate present in the culture is adjusted by thebacterial growth to that which will maintain thegrowth rate of the bacteria at the imposed dilutionrate. Since the growth rate of a culture has amaximum value for any rate-limiting substrate,then a continuous culture will have a maximumdilution rate at which a steady-state can bemaintained. At dilution rates above this thepopulation, if already developed by growth at alower dilution rate, will wash out of the culturevessel, or an inoculum added to the culturemedium will not grow.

    A simple mathematical theory of growth underthese conditions, based on Monod's expressionsfor bacterial growth, was developed by Herbert,Elsworth, and Telling103 and this type of expres-sion is worked out by Kirsch and Sykes. Theseauthors also emphasize the point that this theoryis applicable only to completely mixed cultures.The mixing of incoming nutrients must be almostinstantaneous and the bacteria must be uniformlydispersed in the medium and all have an equalchance of flowing out of the vessel. Also theaddition of the medium must be in as nearcontinuous a stream as possible. The fact thatmost digesters are operated with additions ofsubstrate at discreet intervals, which may varyfrom a few minutes (for example in the authors'experimental units) to a few days in some muni-cipal digesters, has already been mentioned, as hasthe effect of large feed additions in promotion of"batch" growth. But recently Brooks andMeers104 showed experimental proof of the ef-fects, on one bacterium at least, of the discontinu-ous addition of substrate to a continuous culture.In a methanol-limited culture of a Pseudomonasgrowing at a dilution rate of 0.23 hr"1, when the

    June 1974 159

    Dow

    nloa

    ded

    by [

    Cor

    nell

    Uni

    vers

    ity L

    ibra

    ry]

    at 2

    2:20

    13

    Nov

    embe

    r 20

    14

  • interval between additions of methanol was in-creased beyond 20 sec, the yield of bacteria pergram of methanol fell significantly and cycling in anumber of parameters measured in the cultureoccurred. This change in yield was explained bythe fact that the bacterium gave higher yieldsunder carbon-limitation than carbon excess andthe method of addition of the medium (i.e., abigger drop of medium at longer intervals) gaveconditions of growth under alternate methanolexcess and methanol limitation. This phenomenonsuggests that uncoupled growth (see above) wastaking place when methanol was in excess becausethe bacterium could not adjust to the growth ratedefined by the substrate available for the shorttime interval. While this particular aspect ofintermittent addition of substrate is perhaps to bewelcomed in digester operation, in that a low yieldof cells decreases sludge formation, the generalprinciple that intermittent addition of substrateproduces oscillations in a culture which couldimpair efficiency of operation or lead to completebreakdown is important. Also important is thatsuch phenomena produce deviations from continu-ous culture theory and make the theory lessapplicable in practice. While continuous loading ofa digester may be desirable, especially in the caseof small digesters, such as the authors are using forfarm wastes, limitations on pump size imposed bythe physical nature of the waste and the compara-tively long turnover time of anaerobic digestersmake intermittent addition of substrate the onlypractical way of running.

    As said above, continuous culture theory de-mands complete mixing, and this is seldom if everattained in digesters, except in small experimentalplants. Mixing of large-scale digesters is usuallyintermittent and cannot be guaranteed to keep allthe contents completely uniformly stirred. Thuswhile the high-rate digester approaches the idealmixed culture far more than the unstirred digesterwhere stratification definitely occurs, the systemmust produce deviations from the theoreticalbehavior of the mathematical model. Imperfectmixing will lead to pockets where the bacteria maybe in conditions of pH or nutrient concentrationaway from the overall values, and the chances ofwashout of the bacteria will vary, so producingnonideal residence times. In addition to grossimperfect mixing there is the question of solids.These can have two effects. Particles can produceionic effects at their surfaces105 which may alter

    the environment round the particle. Bacteriagrowing attached to a solid will have morecomplex kinetics than that considered by thechemostat theory, and there will be two possiblesystems. One type of bacterium, for instance thecellulolytic ones, will be utilizing the material towhich they are attached as an energy source, but ifthey use ammonia as nitrogen source and volatilefatty acids as growth factors they will be depen-dent for supply on the rate of diffusion of these ifmixing is not perfect and the solid material is in astagnant pocket, or by movement of liquid overthe particle if mixing is better. In either case it ispossible that while growth of the digester popula-tion as a whole is carbohydrate-limited, growth ofsome cellulolytic bacteria could be nitrogen-limited. If the bacteria are not attacking theparticulate material, but are merely entrapped inor on it, then their growth will be similar to thatof, for instance, Sphaerotilus attached to the wallsof an activated sludge system and will be governedby the movement of nutrients to the solid particle.In addition these attached bacteria will tend togrow more as a colony than as the dispersedgrowth required by continuous culture theory.Pirt106 has considered colonial growth of bacteria.

    The various factors considered here could affectthe yield of bacteria per unit of substrate utilized,and as Kirsch and Sykes point out, the originaltheory of continuous culture demands a constantyield factor. These authors noted the fact thatmost continuous cultures deviate from the plot ofsteady-state cell concentrations against dilutionrate predicted by the simple chemostat theory. Atlow growth rates cell yields are smaller thanmaximum, and at high growth rates instead of asudden change from full yield to no yield as themaximum growth rate is passed there appears tobe a number of steady states with decreasing yieldas the dilution rate is increased. This latterphenomenon could only be of importance inconsidering the theory of digesters run at shortturnover times near the maximum growth rate ofsome of the bacteria, and this is unlikely to occurwith the normal high-rate sewage digester. Someworkers have ascribed the phenomenon to artifactsproduced by imperfect mixing or wall growth ofthe bacteria which in effect continuously inocu-lates the medium. However, one of the authors isof the opinion that this is not the case and that itis a true departure from theory.36

    The fall in growth yield at low dilution rates is

    160 CRC Critical Reviews in Environmental Control

    Dow

    nloa

    ded

    by [

    Cor

    nell

    Uni

    vers

    ity L

    ibra

    ry]

    at 2

    2:20

    13

    Nov

    embe

    r 20

    14

  • discussed by Kirsch and Sykes. This occurs mark-edly with some anaerobic rumen bacteria and wasascribed to maintenance requirements, and notendogenous metabolism as previously sug-gested,107 by one of the authors in an earlierpaper on the rumen bacteria.36 Maintenance andendogenous metabolism have been consideredfrom the practical point of view earlier in thisreview. The various theoretical treatments ofendogenous and maintenance metabolism and theapplication of these equations to digester modelsare discussed by Kirsch and Sykes. There have alsobeen other reviews and extensions to the theoriesdiscussed by these authors. For instance, van Udenreviewed equations for nutrient-limited growth ofbacteria involving maintenance requirements, com-petitive inhibition, and transfer of nutrients.52

    Whether maintenance or endogenous metabolismare considered the mathematics are similar, and ofcourse the resultant phenomenon of low yield atlow growth rates is the same. A newer approach tomaintenance and uncoupled growth and its affecton cell yields is discussed by Stouthamer andBettenhaussen108 but this does not alter the needto account for low growth yields in continuousculture theory. A further theoretical considerationof maintenance and uncoupled metabolism isrelated to the growth of Azotobacter in continu-ous culture by Nagai and Aiba.109 They weredetermining the maintenance coefficient and truegrowth yield (see also Pirt110), and they found ahigh maintenance coefficient and an apparentnegative value for the true growth yield in glu-cose-limited cultures. This could be accounted forby uncoupled metabolism.

    The theory of continuous culture also assumesno loss of medium other than through the over-flow. King et al.1 ' * gave a theoretical treatmentof the effects of loss of liquid by evaporation ofcontinuous cultures and showed that at lowdilution rates errors in calculation of parameterssuch as cell yields and maintenance coefficientscould be induced. Evaporation is, of course, mostserious in aerobic cultures sparged by large vol-umes of air, and it will not be so important inslowly mechanically-stirred anaerobic digesters,although some evaporation must take place withthe effluent gas. It might be more important ingas-stirred digesters, but it does illustrate anotherpossible deviation from theory in the running ofdigesters.

    A further consideraton discussed by Kirsch and

    Sykes is that of death of bacteria growing at lowgrowth rates and the possibility of a minimumgrowth rate for a bacterial culture. This firstphenomenon has previously been mentioned inconnection with digester function because it oc-curs in laboratory pure cultures at dilution ratessimilar to those at which digesters operate. Totalcounts of bacteria in a heterogeneous jnixture likedigester contents are difficult or impossible to do,but because of the overall low growth rate set bythe digester dilution rate and the possibilities ofintermittent, or diauxic, growth and lysis assubstrates become available, it is probable that theviable count of bacteria in digesters will be a verysmall proportion of the total count, as it usually isin the rumen. These phenomena again complicatetheoretical treatment of digester function.

    A further complication discussed by Kirsch andSykes with respect to the possibilities of somebacteria in a population being nonviable, whichinfers that to maintain a steady population densitysome cells must be growing faster than the overallgrowth rate, is that simple theories of continuousculture assume that all bacteria are growing at thesame overall rate. However, Powell112 suggestedthat low viability could be represented nearlyenough by the same equations as for endogenousmetabolism.

    But a further consideration in a mixed culturesuch as a digester is that the components of thebacterial population may at any one time begrowing at different rates because they are growingon different substrates and are limited by differentfactors. This latter was suggested as a reason forthe ability of a mixed population such as therumen to exist,64 because the simple theory ofgrowth of organisms competing for the samelimiting substrate shows that at only one dilutionrate may the bacteria grow as a stable mixture.113

    The few experiments which have been done onbacteria competing for one substrate do notalways bear out this theory, as mixed populationsof the same bacteria seem able to grow at differentdilution rates (e.g., see Reference 114; Hobson andSummers, unpublished). The situation is againcomplicated. However, there is a further compli-cation in mixed culture growth. Meers andTempest1 i 5 found that with two bacteria inocula-ted into a magnesium-limited, aerobic, continuousculture at a certain dilution rate, although thetheory mentioned above predicted that one par-ticular species should become dominant, the actual

    June 1974 161

    Dow

    nloa

    ded

    by [

    Cor

    nell

    Uni

    vers

    ity L

    ibra

    ry]

    at 2

    2:20

    13

    Nov

    embe

    r 20

    14

  • dominant species was the one in highest propor-tion in the mixed inoculum; the other washed out.Similar results have been obtained by Hobson andSummers (unpublished) with two rumen bacteriain ammonia-limited anaerobic culture. A steady-state was obtained with one bacterium at aparticular dilution rate and the culture inoculatedwith the other bacterium. But, contrary to theory,the dominant bacterium was always the onepresent in the original culture, the other bacter-ium, although it remained in the culture, neverdeveloped beyond about one thousandth of theoriginal population. The results were explained byMeers and Tempest, and later expanded into amathematical theory of "hypertrophie" growth byPowell,1 ' 6 on the basis that the organisms secreteinto the medium substances which enhance theirown growth. This phenomenon could be of im-portance in digesters because if a particular com-ponent of the digester population is inhibited orkilled by some toxic substance, it may be verydifficult for this component to regenerate, unless avery large inoculum of the bacteria is supplied.However, if the whole population of a digester isaffected to much the same extent by adverseconditions then presumably all the componentswill have an equal chance of regenerating.

    A further consideration brought into somemathematical expressions of digester function isthat of substrate inhibition. Kirsch and Sykesdiscuss the theories, and the experimental evidencefor substrate inhibition, with respect to the meth-ane bacteria and decide that the question ofsubstrate inhibition is still open. The discussion ina previous section would suggest that substrateinhibition as such does not occur, as propionateseems to be the main inhibitory substance and themain methane production is from acetate orhydrogen plus carbon dioxide. The production ofpropionate was suggested to be the results of achange in fermentation pathways, and the inhibi-tion of methanogenesis a cumulative effect re-sulting from this. Product inhibition has not beenconsidered in models of digester systems althoughsome models have been suggested for continuouscultures in general, e.g., see Reference 117.

    The simple models used for digester simulationassume that no viable bacteria enter with thefeedstock. This is not true and it was suggestedpreviously that continued massive inoculation mayserve to keep up numbers of some of the digesterpopulation and thus their growth kinetics will be

    different from those of the stable population.Even if these bacteria play no key role indegradation of the waste they will contribute tothe overall fermentation and some could be acontrolling factor in growth of the strictanaerobes.

    Kirsch and Sykes discuss the theory of thechemostat as applied to experimental digesters andconclude that if methane production from thevolatile fatty acids is the rate-limiting step, thenmodel systems based on the comparatively simpletheory of continuous cultures can predict certainexperimental results such as the minimum turn-over time. They consider, rightly, that the kineticsof acid production from sewage solids are un-known, but that this does not matter if methano-genesis is the rate-limiting step. But they also pointout some drawbacks to the experimental work onwhich many of these conclusions are based, suchas use of soluble substrate, use of enrichmentcultures, and lack of suitable length of time for thedigester to attain a steady state. This latter point isa criticism of a number of papers on continuousculture and oscillatory phenomena on changingconditions in a culture or on build up of acontinuous culture have been referred topreviously. It seems to the authors that thesecriticisms apply to other papers not mentioned inKirsch and Sykes's review and that most of thepapers are merely extensions or reconsiderationsof previous ones, and the working out of themodels is done from much the same experimentaldata. For instance, Lawrence118 in a considera-tion of the application of kinetic data to digesterdesign and function found different values forcoefficients for the breakdown of long chain fattyacids in an enrichment culture, where the acidsonly had been fed to a digestion, and thebreakdown of lipid material in a digester fed onsewage sludge. (The data are taken from otherpublished work and the authors have not been ableto obtain the thesis in which some of the datawere given.119) He, rightly, takes the latter asbeing probably more representative of digestion ingeneral. But then he takes results from enrichmentculture experiments for calculations on the forma-tion of methane from the lower volatile fattyacids. Andrews in a "Dynamic model of theanaerobic digestion process"92 has introduced anequation for growth rate based on a function givenby Haldane for the inhibition of enzymes by highsubstrate concentrations, which at low substrate

    162 CRC Critical R eviews in Environmental Control

    Dow

    nloa

    ded

    by [

    Cor

    nell

    Uni

    vers

    ity L

    ibra

    ry]

    at 2

    2:20

    13

    Nov

    embe

    r 20

    14

  • concentrations approximates to the Monod equa-tion. He then applies this model to the conversionof volatile acids to methane, assuming this to bethe rate-limiting step. But laboratory studies inconnection with the model were made on anenrichment culture fed on a synthetic solublemedium containing propionic acid. The model didthen simulate some results "commonly observed inthe field:" 1. Increasing the quantity of seedsludge or increasing pH will decrease the timerequired for start-up of a continuous flow digester;2. digester failure may occur if insufficient seedsludge is present or the pH is too low when acontinuous flow digester is started; 3. digesterfailure may be caused by sudden changes indigester loading; 4. digester failure during start-upmay be avoided by slowly bringing the digesterloading to its full value; 5. pH control or decreaseof process loading are effective techniques forcuring a failing digester.

    But 1. could be predicted from known micro-biolgy; a bigger inoculum stands a better chance ofgrowing and will give a certain cell density fasterthan a small one and the pH of the medium shouldsuit the bacteria growing in it. Point 2. followsautomatically from Point 1. Point 3. was discussedon the basis of known continuous culture resultsin a previous section of this review. Point 4. is, asthe author implies, what has been found byexperiment and can be predicted from othersituations and consideration of the points notedunder Point 3. Point 5., bringing back the pH tothe correct value for growth is of course the basisfor buffering or controlling the pH of all cultures,continuous or batch - the problem with digestersis how to do it. Decreasing loading is an obviousway of trying to metabolize some intermediateproduct built up in a culture.

    However, the author does say that his model isonly a beginning and that it may be a basis forfurther work, and indeed, he points out somefactors mentioned previously which should beincorporated in future models. In a further paperLawrence and McCarty120 try to provide a "uni-fied basis for biological treatment design andoperation. In this they consider the completelymixed chemostat system, a similar system withsolids recycle, and a plug-flow system with solidsrecycle. These systems approximate to most anaer-obic or aerobic treatment plants, but again themathematical models are based on the simpletreatment of Monod, and Herbert et al.103 previ-

    ously mentioned, with all the assumptions that arenot actually attained in practice. And in theintroduction to the paper where some basicformulae are discussed they make the pertinentpoint that "while the merits of the two modelshave been debated, it should be emphasized thatboth are empirical and a choice between the twoshould properly be based more upon convenienceand ability to furnish a satisfactory solution thanupon any fundamental consideration." This seemsto the present authors to be a legitimate criticismof the present state of the art of mathematicalmodeling of digester systems: it is comparativelyeasy to obtain mathematical expressions that willproduce an approximation to some experimentalresult, without the theory being based on anyproved biological reactions. Again, in using theirmodel to predict minimum retention (or turnover)times for substrate utilization for various pure andmixed synthetic substrates or natural wastes, theynote a discrepancy between two sets of valueswhich they say may be due to unknown toxic orother rate-limiting factors in some substrate mix-tures, or to experimental conditions, or "thedifficulties in evaluating kinetic coefficients," andthis brings the argument back to the point madepreviously, that much data used in evaluatingmathematical models is based upon experimentswhich do not truly simulate the conditions ofactual digesters. The authors conclude that, "Inessence, the models presented herein are onlymathematical formulation of what has been obser-ved to be the important parameters by designers,operators and investigators in the past. Suchformalization, hopefully, will furnish relationshipswith predictive value to serve not only in thedesign and control of existing treatment processes,but also will aid in the development of biologicalprocesses for other purposes." This is very true,but for true prediction much more must be knownabout the microbiology and the biochemical pro-cesses of anaerobic digesters. For instance if, aspreviously suggested, cellulolysis and not methano-genesis is the rate-limiting factor, at least in thedigestion of some wastes of high natural cellulosecontent, then new kinetic constants will have to beobtained for any model system. The modelsthemselves may have to be amplified. Baldwin andhis co-workers have attempted computer modelingof the rumen system in which the very largeamount of data on rumen microorganisms, sub-strates, fermentation products, and fermentation

    June 1974 163

    Dow

    nloa

    ded

    by [

    Cor

    nell

    Uni

    vers

    ity L

    ibra

    ry]

    at 2

    2:20

    13

    Nov

    embe

    r 20

    14

  • pathways, etc., can be incorporated. This modelused the computer simulation technique developedby Garfinkel121 and is described by Baldwin etal.122 The model cannot be described in full here,but some indication can be given of its scope. Theinput included equations depicting association ofcellulolytic organisms with fibers, colonization offibers by bacteria, passage of microbes out of therumen, etc. The bacteria were considered invarious groups, such as the cellulolytic group, withproperties based on those experimentally deter-mined for various rumen bacteria. Known fermen-tation balances and growth yields and chemicalequations for biosynthetic reactions in cells, and soon, were also incorporated. Using experimentallydetermined data to provide values for the variousequations the model was used to predict volatilefatty acid production from various feeds, diurnalpatterns of rumen volatile acids, ammonia, andmicrobial concentrations under different feedingregimes, microbial interactions, and so on. Themodel made very good predictions consideringthat, as the authors say, "a number of theassumptions made are tentative and subject tochange with the accumulation of additional experi-mental data and further theoretical development."

    This type of modeling could not have beendone when some of the original models of con-tinuous cultures were formulated, as the opera-tions can only be carried out with the aid of acomputer, and in fact it is probably not necessaryfor the ideal systems of a single stage, pure culturechemostat. But the computer model just describeddoes show the way to obtain really meaningfulresults from a complex microbial system such as arumen or anaerobic digester, although its full usein either, and especially the digester, system muststill await the provision of more data by themicrobiologist and biochemist.

    The Two-stage DigesterAn extension of the chemostat which was

    suggested for, and is being used in, some industrialprocesses, is the multiple stage fermenter. Here theoverflow from the first fermenter passes to asecond fermenter and possibly to a third or fourth.By suitable adjustment of the volumes of thevessels the dilution rates can be made to differ inthe vessels, even though the flow rate is the same.The system can, for instance, be used in opti-mizing the production of a substance formedduring the resting stage of a culture, in that growth

    would be limited in the first stage by exhaustionof a nutrient, and in the second stage the cellswould continue to metabolize in the depletedmedium. Growth could take place quickly in thefirst stage, say under nitrogen limitation, and inthe second stage a slower, resting metabolism ofexcess carbohydrate in the medium could takeplace, or some metabolite not in the originalmedium could be fed into this stage.

    In the usual layout the standard high ratedigester has a type of second stage in the settlingtanks which usually take the outflow. Here,besides settling and consolidation of the sludgesome microbial activity takes place, but as thetanks are at ambient temperature this microbialactivity is slow, because the bacteria in theoverflow from the digester will have an optimumgrowth round about the digester temperature. Formaximum activity of the population already devel-oped in the digester then the second stage temper-ature should be the same as the first stage. This, ofcourse, involves construction of a second stage ofequal mechanical complexity to the first andunless a second stage can definitely be shown to beneeded for optimizing some particular digestionprocess its construction would be barred oneconomic grounds. Although digestion is a two orthree stage process biochemically, the processesare interlinked and spacial separation would bedifficult and probably not more productive. Forexperimental purposes Hammer and Borchardtproduced a two stage laboratory domestic sludgedigester in which the hydrolysis and fermentationstage was separated from the methanogenic stageby a dialysis system.123 This type of system is toocostly and complex for large scale general use, butit did provide useful experimental data. Enrich-ment cultures were developed in the two stages byfeeding raw sludge to the first stage and acidsalone to an inoculum of digested sludge in thesecond stage. The first stage was then fed rawsludge and the acids formed transferred to thesecond stage through the dialysis unit. The experi-ments showed, among other things, that thehydrolytic and acid-forming bacteria were notoperating at full efficiency under the conditions ofEh and pH needed for the methanogenic bacteria,and that, as previously mentioned, the initialfermentation was the rate-limiting factor and notmethanogenesis.

    The separation of the two stages in a large-scaleplant could only be done on the basis of altering

    164 CRC Critical Reviews in Environmental Control

    Dow

    nloa

    ded

    by [

    Cor

    nell

    Uni

    vers

    ity L

    ibra

    ry]

    at 2

    2:20

    13

    Nov

    embe

    r 20

    14

  • retention times in the two fermenters. If one weredealing with a very easily fermented waste then ashort retention time in a primary fermenterfollowed by a longer retention time in a secondaryfermenter to allow the methanogenic bacteria togrow on the effluent from the primary fermentermight work, as the methanogenic bacteria wouldwash out of the first fermenter. However, dealingwith the usual wastes where cellulolysis or otherprimary reactions may be rate-limiting then thedetention time in the first stage would be suchthat a methanogenic population would certainlydevelop. It would seem quite impracticable forwaste disposal work to inoculate each tank withthe appropriate bacteria and keep the bacterialpopulations stable. In any case increase in efficien-cy over a properly designed, single stage systemrun at the optimum turnover time for the particu-lar waste being treated, shown by a two stagesystem would have to balanced against increasedcapital cost and running costs in the form ofincreased heating and so on.

    High Rate Digester with FeedbackThe single stage system with feedback of

    bacteria from the effluent to the inflow is a systemwhich was suggested early as a means of increasingthe efficiency of the chemostat for some reactions.It should allow increased bacterial concentrationin the fermenter and a dilution rate greater thanthat represented by the maximum growth rate ofthe bacteria. The activated sludge system is anexample of cell recycle.

    Mathematical models for a chemostat systemwith cell recycle have been made by Herbert124

    and Fencl,12 5 and this type of mathematicaltreatment is discussed by Kirsch and Sykes. Morerecently Pirt and Kurowski126 have extended thetheory and subjected the theoretical approach toexperimental investigation in a laboratory fer-menter. Pirt discusses four practical systems. Inthe simplest system the outflow from the fer-menter passes through a separator tank, from thebottom of which medium containing the sedi-mented cell concentrate is returned to the fer-menter, while the rest overflows from the top. Inthe second system a part of the concentrated cellsuspension is fed back while the rest passes on, andthe dilute cell suspension is taken from the top ofthe separator. This is a means of controlling therate of feedback of cells. The third system was theone used to obtain concentrated and dilute efflu-

    ent streams in a laboratory fermenter. In this, partof the medium is pumped out of the fermenterthrough a filter which obstructs the passage ofmost of the cells in the fermenter, thus the diluteeffluent stream can be, and was in the experi-mental apparatus, almost a cell-free medium. Theconcentrated effluent stream flows from the fer-menter over a normal weir system. The fourthsystem was a two-compartment fermenter with abaffle separating a bottom stirred active zone, anda top unstirred zone in which the cells couldsediment and fall back into the bottom zone.Dilute effluent was taken from the top of thevessel and concentrated effluent from the bottom.A theoretical treatment of the second and thirdsystem is given; the fourth is theoretically identicalto the third. Aerobic growth of a yeast culture wastested in a laboratory version of the third systemand agreement with theory was obtained. Anumber of steady-states at different dilution ratesand feedback rates were examined. The fraction ofthe total effluent stream containing concentratedcells flowing over the weir was designated C andthe fraction of dilute effluent passing out throughthe filter was (1-C). When the culture wasoperated under feedback conditions when g = 4.0then the yeast concentration in the fermenter wasabout 3.3 g/1 compared with 0.8 g/1 in the culturerun without feedback, and the washout dilutionrate was about 1.6 hr"1 compared with about 0.4hr"1. The residual glucose concentration was keptat the same low level up to a dilution rate of about1.1 hr"1 in the culture, with feedback, comparedwith about 0.3 hr"1, before it increased in theusual manner as the maximum dilution rate wasapproached.

    The experimental system used here sufferedfrom the expected disadvantage of clogging of thefilter. Frequent back-flushing was necessary. Thesystem would be almost impossible to use on alarge scale. However, a plant operating on theprinciples of the first system described above hasbeen used for meat wastes, following pilot plantstudies (see below). The process is generallyreferred to as the "anaerobic contact process."These experiments suggested that detention timesof about one half to four days in the digester unitscould be used for different wastes. In discussingthese results, Kirsch and Sykes pointed out thatsteady-state conditions were probably not ob-tained. Difficulties were found on a large scale insettling the digester outflow for feedback because

    June 1974 165

    Dow

    nloa

    ded

    by [

    Cor

    nell

    Uni

    vers

    ity L

    ibra

    ry]

    at 2

    2:20

    13

    Nov

    embe

    r 20

    14

  • of disturbance of the particles by gassing, and adegasifier had to be incorporated in the meat-waste plant. Thus there are practical difficulties inthe process, although feedback has been used inother fermentation systems including activatedsludge systems (see References 125 and 127 forexamples), and a form of feedback is practiced insome sewage works where incoming sludge ismixed with some of the settled, digested sludge.However, the theory of feedback assumes that thefeedback is almost instantaneous, i.e., there is nochange in the metabolism of the bacteria betweenthe outflow and return to the fermenter. If thetime between outflow and return is prolonged, orthe bacteria are cooled, then even if no death ofthe bacteria occurs there must still be a period ofadjustment to the digester conditions when thebacteria are returned and the process will deviatefrom theory. The theory also assumes a solublesubstrate with bacteria freely suspended in themedium. If the biodegradable material is a recalci-trant solid (e.g., the cellulose previously men-tioned) then the active bacteria will be attached tothe solid. The mechanical processes of recycling ofthe mass of partially degraded solids may detachbacteria and it might be that the time taken toregain active colonization of the solids on returnto the digester would detract from the efficiencyof the process. Also in this case there is deviationfrom the theory, as this assumes that there is norecycle of substrate with the biomass.

    A true recycle system may, then, not be themost suitable plant for digestion of wastes of high,but not readily biodegradable, solids content andthe costs of the plant could be high. But on theother hand, the system is theoretically and practi-cally suited to treating dilute wastes at high flowrates as it is possible to have a long bacterialresidence time with high cell density and a shortliquid turnover time within limits.

    The unstirred digester in which sludge accumu-lates is to some extent an application of Pirt'sfourth theoretical scheme. But here process effi-ciency is low as the activity of the sludge layermay be lowered because lack of stirring may leadto build-up of pockets of conditions unfavorableto the bacteria because of accumulation of pro-ducts or lack of diffusible nutrients. But, anotherprocess similar to Pirt's fourth scheme is the towerfermenter. This has been investigated on a small

    scale and is now in use on a large scale for beermaking.128 The process here is different in that itis partly aerobic and partly anaerobic, but thereseems no reason why it should not be applied toanaerobic digestion, and it is to some extentsimilar to the anaerobic, upflow filter system.

    The Anaerobic Filter SystemThe usual anaerobic digester feedstock contains

    a large amount of solid waste, and indeed anadvantage of the process is that it will treat thesesolids. But the solids also seem to be necessary forthe running of the process: the authors, forinstance, have found washout of a digester fedwith piggery waste of only 1 to 2% solids content.The solids probably act as a matrix for growth ofthe bacteria in microcolonies and may aid concen-tration of nutrients around the bacteria. There isalso some suggestion that growth of the methano-genic bacteria is aided by dispersed solids. Thusthe process is not usually suitable for dissolvedwastes.

    In the old established vinegar process, or thetrickling filter for aerobic treatment of dilutewastes the microorganisms grow attached to solidsin a matrix of twigs, stones, or coke, or nowadaysplastic shapes, and the substrate runs over them.McCarty and co-workers have developed to a largelaboratory scale an anaerobic upflow filter basedon the same principle of medium flowing overbacteria attached to, or lodged in the crevices of acolumn of small stones. A detailed discussion ofresults and some theoretical aspects, which aresimilar to those of the trickling filter, and anotherpaper giving the experimental layout and someresults are those by Young and McCarty.129 '130

    The process is described later, but the main pointsto be considered here are that the process is similarto the recycle process in that a large mass of activebacteria is retained in the filter and so the liquidretention time in the filter is only a matter ofhours. Suspended solids in the effluent were low,and the experimental filters also operated effi-ciently at lower temperatures (25C) than that ofthe normal high-rate digester (so on a large scaleheating requirements would be low) and the filtersoperated for long periods without clogging. Thefilter process appears to be theoretically desirableand, at least on a small scale, to have few practicaldifficulties.

    166 CRC Critical Reviews in Environmental Control

    Dow

    nloa

    ded

    by [

    Cor

    nell

    Uni

    vers

    ity L

    ibra

    ry]

    at 2

    2:20

    13

    Nov

    embe

    r 20

    14

  • IV. THE BACTERIAL POPULATIONSOF ANAEROBIC DIGESTERS

    Definitive study of the bacteriology of anaero-bic digesters has still to be carried out and muchless is known about the bacteria of digesters thanis known of, for instance, the very similar rumen.Although some workers have attempted to relatetheir cultural observations to the main activitiesknown, or presumed, to be occurring in digesters,there have not been sufficient observations to beable to determine normal variations in numbers ofthe bacteria during steady state digestion, changeswhich may occur in the bacteria before or duringdigester failure, changes during or after shockloading, washout of particular bacteria as relatedto turnover times, and so on. In addition some ofthe recorded cultural observations have been madeon laboratory digesters with synthetic or partlysynthetic substrates and the relation of these toobservations made on laboratory digesters ofnatural substrates and large-scale digesters is diffi-cult to decide. Furthermore, most isolations havebeen made from mesophilic, high-rate type diges-ters: systematic experiments are also needed onthermophilic digesters and the various other diges-ter systems. For instance does feedback in adigester lead to higher concentrations of all bacte-ria or are specific types favored? More knowledgeof the bacteria in digesters with different sub-strates and the properties of the bacteria mighthelp in suggesting ways of improving digestion byinoculation of selected bacteria, but it would helpin the understanding of digester function, theprediction of possibilities of digestion of differentsubstrates, and the modeling of digester systems ina more meaningful way. Pure and mixed continu-ous cultures of important digester bacteria wouldalso help in the understanding of digester function,but such experiments cannot meaningfully becarried out until the most important species ofdigester bacteria have been identified by culturalwork on digesters themselves.

    Although all wastes fed to digesters will containthe major groups of substrates for growth of themicrobial population concerned in the first hydro-lytic and fermentative stage of digestion, theamounts and types of these substrates will vary, asdescribed in a previous section. In the same waythe feed of ruminants varies: grass, hay, mixedconcentrates, etc. Although all these feeds containcarbohydrate, protein, and lipid, the varying forms

    in which these occur lead to well-documentedchanges in the relative numbers of the bacteria.However, although there are changes in the relativenumbers, most of the types of bacteria occur to agreater or lesser extent in all rumens. One mightsimilarly suppose, then, that all, or most meso-philic digesters of sewage-type wastes will containsome of the same types of bacteria, but that theirproportions will differ. Only in the digestion offactory wastes, which may be extremely differentin composition, say a meat-processing waste and apotato-processing waste, might one expect a ratherspecialized flora. There is, however, an importantdifference between the rumen and a sewagedigester. While the rumen is subject to inoculationby the bacteria present on the food and in the air,the greater proportion of these will be aerobicbacteria unable to live in the rumen, and so thesewill add little to the population. Although thereare some cases where the presence of a bacteriumin the rumen in numbers suggesting a possibleminor function has been attributed to continuedheavy inoculation from the food (see below), thereis no evidence of a major component of the florabeing present only because of a heavy inoculationfrom the food. On the other hand, the feed to asewage digester contains very large numbers ofbacteria which have been growing in the faecalmaterial either in the body or in collecting tanksfor the sewage, and should, therefore, be suited tolife in the digester. That such bacteria can becomenumerically a major part of the flora is shownlater, although functionally they may play a minorpart. Because of the nature of the feedstock, thenalmost any anaerobic or facultatively anaerobicbacterium might be expected to be found indigesters. The importance of a particular bacte-rium can only be assessed by considering itsnumbers in relation to the total population, itsproperties in relation to the known reactions ofdigestion, and its properties in relation to theneeds of the other bacteria. To take these in turn,it is difficult or impossible to assess total numbersof bacterial cells in digester contents, but this maynot matter, as many of the bacterial cells mightreasonably be presumed to be dead. Althoughsome nonviable bacteria might still be metaboli-cally active, total counts of viable bacteria must bean indication of the microbial activity in thehabitat. Total viable counts of bacteria from ahabitat such as a digester suffer in general from atleast three difficulties. One can only make assump-

    June 1974 167

    Dow

    nloa

    ded

    by [

    Cor

    nell

    Uni

    vers

    ity L

    ibra

    ry]

    at 2

    2:20

    13

    Nov

    embe

    r 20

    14

  • tions about the correct medium and presume thatthe medium which grows the largest number andgreatest variety of bacteria is giving something likethe total count. Since one must have visiblegrowth in a culture, media must of necessitycontain fairly large concentrations of nutrients andthe bacteria are then being transferred from thedigester where nutrient concentrations should, ifthe digester is behaving like a chemostat, be verylow to a medium of higher nutrient concentra-tions. This may have an adverse effect on bacterialviability. Third, there is the difficulty of separatingbacteria from particulate matter without damagingthe bacteria; complete separation is impossible. Sofor all these reasons the most that can be hopedfor is that samples of similar physical compositiontreated in the same way will give bacterial countsthat, if not total, will be at least relativelycomparable one with another.

    Since bacteriological investigations of digestershave been made not only on digesters withdifferent substrates, but by different methods andwith different media, it is difficult at present toassess the results, and the importance of thebacteria isolated. The later experiments of thoseworkers using anaerobic habitat-simulating mediawill tend to be most comparable and to give themost useful picture of the digester flora. When thebacteria from such experiments can be assignednumerical concentration and a function in digesterreactions then the results should fulfill the firstand second criteria mentioned above.

    The third criterion is more difficult to deter-mine. Since comparatively little is known of thedetailed nutrient and growth factor requirementsof the bacteria, and less of the production offactors such as vitamins, it is difficult to saydefinitely that a bacterium present in apparentlylow numbers is not producing some factor of vitalimportance to other bacteria.

    The problems of investigation of digester floraare, therefore, formidable and while some isolatesof bacteria may be assigned to a definite place inthe complex interactions that make up digestion,it is impossible to say that a bacterium definitelyhas no importance in the overall viability of theflora.

    Not only will the bacterial population vary withthe composition of the waste input to digesters,but the concentrations of bacteria would beexpected to vary with the concentration of wasteadded. Thus it is difficult, or impossible, to

    compare the concentrations of bacteria reportedby different workers. Some assessment of the roleand importance of the bacteria in particulardigestions may be possible, but only in the case ofthe methanogenic bacteria, where the substratesand the bacteria are probably the same in alldigestions, can the results be generalized. For themoment, then, it would seem best to recordresults, leaving a major assessment for later. Muchfurther work is needed to clarify the overallpicture.

    The bacteria which will be suited to thebreakdown of organic waste in a mesophilicdigester will be those adapted to an anaerobicenvironment, a pH of between 6.5 and 7.5, and atemperature of approximately 35C. These factorshave been taken into account by most investiga-tors, but with the habitat-simulating medium hascome the realization that the gaseous nature of theanaerobic environment may be important, thatgrowth factors may be present in the environmen-tal fluid, and that the nature of the substrates willplay a part in the attempted isolation of allbacteria from the particular system. Since carbondioxide is the metabolically important componentof the digester gas this is used as a cultureatmosphere, digester fluid adds growth factors,and mixed substrates, related to those presumablyutilized in the digester, attempt to provide energysources for all bacteria. One of the present authorshas discussed the basis of. the rumen bacteriologytechniques and their application to investigation ofdigester bacteria in another paper.13 '

    Anaerobic protozoa play a part in the fermenta-tion of food in the rumen, but there is no evidencethat protozoa are generally present, or play anybiochemical role in digesters. Young andMcCarty129 reported that free living ciliate proto-zoa and amoebae were abundant in their anaerobicfilters, but the reasons for their presence and theirrole in the process are not known. The authorshave never seen protozoa in their high-rate, pig-gery-waste digesters.

    Establishment of a Functioning DigesterPopulation

    Although the usual method of starting a diges-ter is to inoculate it with sludge from a function-ing digester, it is possible to build up a digestionby adding increasing amounts of waste to adigester originally filled with water. So, all thebacteria necessary for digestion must be present in

    168 CRC Critical Reviews in Environmental Control

    Dow

    nloa

    ded

    by [

    Cor

    nell

    Uni

    vers

    ity L

    ibra

    ry]

    at 2

    2:20

    13

    Nov

    embe

    r 20

    14

  • the incoming waste, although many seem to beonly in very low numbers.

    Work on the establishment of piggery-wastedigesters35'63 included an investigation of theflora during the development of digestion fromwater. Counts were done using the techniques ofrumen bacteriology and a habitat-simulating med-ium for the anaerobic cultures. The results showedthat in the piggery-waste of about 4% total solidsan average of 6.4 X 108 bacteria/ml (anaerobicand facultative) were able to grow anaerobically,while aerobic counts showed 2.4 X 108 bacte-ria/ml. Of the bacteria in the anaerobic counts,three main morphological groups were present:facultative Streptococcus sp. which comprised 43to 74% of the isolates from different batches ofwaste, Bacteroides which constituted 20 to 80% ofthe strict anaerobes, and clostridia (Cl. butyricurri)which made up the remaining anaerobes, exceptfor the occasional Gram-negative curved rod. Nocellulolytic or proteolytic bacteria were detectedin the predominant bacteria in the waste and nomethanogenic bacteria were detected. Amylolyticbacteria occurred in numbers greater than 4 X105/ml. However, it was from this material that afunctioning digester flora was built up over six toeight weeks. Changes in the predominant groups ofbacteria occurred and functional groups of bacte-ria undetectable in the piggery waste built up.Results quoted later suggest a similar buildup of acellulolytic flora in domestic digesters, althoughthere were no detectable celluloytic bacteria in theraw sludge. At first the digester flora correspondedto that of the piggery waste and only amylolyticbacteria were detected. After four to five weeks acellulolytic flora developed when counts of 104 to105/ml were obtained, and a proteolytic flora ofabout 4 X 104/ml also became apparent. The totalcount on an anaerobic, nonselective medium wasabout 6 to 8 X 106/ml, and the count on a similarmedium incubated aerobically was about 5 to 7 X106/ml. Methanogenic bacteria began to appearafter three weeks in numbers of 103/ml andreached 104/ml after seven weeks. By this timedigester gas composition was 60% methane, 40%carbon dioxide and a balanced digestion wastaking place.

    Nonmethanogenic BacteriaGeneral Observations

    In contrast to the highly specialized group ofbacteria which carry out the methanogenic stage

    of digestion, the fermentative stage is carried outby a widely diverse group of anaerobic andfacultative bacteria. The numbers, types, andspecies of this group will depend, as alreadysuggested, on the qualitative and quantitativecomposition of the waste fed to the digesters.

    Mah and Susman13 2 obtained counts of around1 X 108 anaerobic, acid-forming bacteria/ml froma domestic digester, and Kirsch133 reportedapproximately 7 X 108/ml. By comparisonanaerobic counts of between 6.4 X 10s and 8.4 X106/ml were obtained from piggery-waste digest-ers.35 The difference in bacterial numbers wasprobably due to the fact that, besides differentloading rates, domestic digesters in addition tofaecal material receive large amounts of organicmatter which has undergone little breakdown,whereas piggery-waste digesters receive only ma-terial which has undergone considerable microbialand enzymatic breakdown during its passagethrough the pig or in storage tanks under thepiggery. It would be reasonable to suppose thatpiggery waste may then be less readily degradableand support a lower bacterial population thandomestic waste during digestion.

    Early studies on the nonmethanogenic stage ofdigestion were mainly superficial, leading to thebelief that the bacteria consisted largely of faculta-tive anaerobes. Pohland,3 in reviewing the evi-dence of these early investigations, reported thatthe bacterial counts in digesting domestic sludgewere of the order of 1 to 2 X 106 /ml, the majorityof the bacteria being aerobes or facultativeanaerobes. Gaub,134 for example, isolated 16aerobic and 5 facultatively anaerobic bacteria,mostly of intestinal origin, and Buck et al.135 in1953 isolated Streptococcus diploidus, a faculta-tive anaerobe, from digesting sludge. This organismwas found by Keefer et al.136 to produce gasrapidly when inoculated into raw sewage sludge.Members of the Pseudomonadaceae, Enterobacteri-oceae, and Achromobacteriaceae were isolated byMcKinneyetal. I 3 7 inl958.

    It is only with the application of the strictlyanaerobic techniques of rumen bacteriology to thestudy of the nonmethanogenic stage of digestionthat a true picture is emerging. Using thesetechniques it has now been demonstrated that, atleast in the case of domestic digesters, obligateanaerobic bacteria are in fact present in far greaternumbers than are facultative anaerobes, and the

    June 1974 169

    Dow

    nloa

    ded

    by [

    Cor

    nell

    Uni

    vers

    ity L

    ibra

    ry]

    at 2

    2:20

    13

    Nov

    embe

    r 20

    14

  • major hydrolytic activities in digestion appear todepend on these obligate anaerobes.

    Cookson and Burbank138 used an inertatmospheric (nitrogen) chamber to isolate faculta-tive and obligate anaerobes from laboratory-scaledomestic-sludge digesters. Clostridium carno-foetidum and Escherichia coli were isolated, alongwith two bacteria for which the names Bacillusknefelkampi and Sarcina cooksonii were proposed.The direct count procedure showed the variationin the population of digesters, loaded once withmixtures of raw sludge and digesting sludge, asdigestion took place. The population of E. colidecreased substantially after eight days of diges-tion. Sarcina cooksonii was not present in detect-able numbers in the digesters, B. knefelkampi waspresent in high numbers throughout the digestionand the population of Cl. carnofoetidum becamesignificant after 11 days. The last 2 species werepredominant at efficient rates of digestion and inthe ratio 1:7, respectively. These authors suggestedthat B. knefelkampi was responsible for a majorportion of the volatile acid produced since the acidconcentration varied with the population of B.knefelkampi. In addition to the four species above,Burbank et al.139 isolated Pseudomonas deni-trificans, other pseudomonads, a Klebsiella sp., amember of the Neisseriaceae and Serrada indicans,by the same techniques. In these cultures the gasphase would be the nitrogen of the anaerobicchamber atmosphere.

    Torien and Siebert140 suggested that sincedigestion in anaerobic digesters and digestion inthe rumen follows a similar path, the anaerobictechniques used in rumen bacteriology could beapplied to enumerate the anaerobic bacteriapresent in digesting sludge. They developed suchmethods using practical techniques based on thoseused by Kistner14 ' for rumen bacteriology whichwere particular modifications or elaborations ofthe methods initiated by Hungate.142 '143 Theyintroduced 5% digester liquor into all their mediato stimulate growth and used a carbon dioxideatmosphere. Using this method, Torien et al.144

    enumerated the acid-forming, obligate anaerobesin the digester run on a synthetic substrate. Theyfound that although aerobes and facultativeanaerobes occurred at all stages of digestion,numerically they formed a minor part of theacid-producing population. Fastidious obligateanaerobes formed the major part of the popula-tion, usually in numbers one to two hundred times

    greater. These could only be cultured in prere-duced media in roll tubes with oxygen-free gasatmospheres. Torien et al.144 concluded that"The concept that the facultative anaerobicbacteria are the most important acid producers isincorrect. Obligate anaerobic bacteria seem to bethe major group in acid production." They sug-gested that the failure of previous workers such asCookson and Burbank138 and Burbank et al.139

    to report large numbers of obligate anaerobicacid-forming bacteria in digesters was due to thefact that petri dishes and anaerobic jars wereunsuitable for the culture of strict anaerobes andthat growth factors present in digester fluid wereessential for the growth of many of the strictanaerobes.

    Hattingh et al.14S studying the adaptation ofan anaerobic digester to a synthetic substrate, alsoconcluded that the aerobic and facultativeanaerobic bacteria did not form a large part of theacid-forming population. They found the role ofthe aerobic and facultative anaerobic bacteriadifficult to interpret, but felt that because thesebacteria do occur in anaerobic digesters despiteunfavorable conditions, they may play some minorpart in the digestion process. The possible role ofthese bacteria has been discussed earlier in thisreview.

    The findings of other workers confirmed thoseof Torien et al.,144 and Mah and Susman132 in1968 demonstrated ten to one hundred timesmore obligate anaerobic bacteria than facultativeand aerobic bacteria using the anaerobic roll-tubemethods of rumen bacteriology. They used ahighly reduced, habitat-simulating medium ofinorganic salts, digester sludge supernatant, andglucose. Kirsch133 used a complex, habitat-simulating medium similar to that developed byCaldwell and Bryant146 for the rumen bacteria,containing 16% digester fluid, and an anaerobicroll-tube method based on that of Hungate.143

    Counts two to five times higher than on the basalunsupplemented medium, or on typical generallaboratory media, were obtained. Examination ofthe bacterial types isolated on the supplementedmedium showed that the predominant flora con-sisted of a Gram-negative, nonsporing, obligateanaerobic rod.

    Unlike the results obtained for domestic sewagedigesters, Shaw35 '63 did not observe a predomin-ance of obligate anaerobes over facultativeanaerobes in piggery-waste digesters. Using a modi-

    170 CRC Critical Reviews in Environmental Control

    Dow

    nloa

    ded

    by [

    Cor

    nell

    Uni

    vers

    ity L

    ibra

    ry]

    at 2

    2:20

    13

    Nov

    embe

    r 20

    14

  • fication of the Hungate142 anaerobic techniquewhich has been outlined elsewhere13 ' and aerobicincubation, Shaw demonstrated that the bacterialcount of the input to the digesters was about onehundred times higher than that in the digestersthemselves. This would, therefore, influence thebacterial flora recovered from the digesters.Counts on anaerobic and aerobic media were verysimilar (see above). The influence of the incomingwaste was shown in the change in facultativeanaerobic bacteria found on starting a piggery-waste digester from a digester "seeded" withdomestic digester sludge. The anaerobic count ofthe seed sludge was 2.4 X 107/ml, and thefacultative bacteria in this were predominantlyEscherichia coli with some Staphylococcus sp.When the digestion had become adapted to piggerywaste the predominant facultative bacteriachanged to the Streptococcus sp. found in the rawwaste, and these were present in greater numbersin the cultures than were the obligate anaerobes.In the domestic sludge the main anaerobic bacteriaisolated on a nonselective medium were Bacte-roides and unidentified, nonsporing, Gram-positiverods. In the digestion adapted to piggery waste themain anaerobic bacteria were Clostridiumbutyricum and other unidentified clostridia, pre-sumptive lactobacilli and an unidentified Gram-positive small coccus occuring in chains.

    Since the bacterial inoculum given in the feedto the digesters influences the nature of the florain the digester, emphasis should be placed on theisolation of bacteria in relation to their propertiesand function in the digestion. Many of the bacteriapresent in large numbers in the digester input mayeither pass through the digester or may degradesimple substrates as they are unable to degrade themajor components of the waste.

    Sykes and Kirsch147 attempted to establishfermentation patterns of pure cultures growing onprimary sewage sludge. Isolations made byKirsch13 3 were grown on sterilized sewage sludgeand some of the fermentation products formedwere identified and quantified. Approximately aquarter of the isolates formed hydrogen andcarbon dioxide, with acetic, propionic, and butyricacids. The remaining cultures produced carbondioxide with acetic and propionic acids, but nothydrogen or butyric acid. Comparison of theactivity of single cultures on sterile sludge withthat of a mixed population led Sykes and Kirschto conclude that many naturally occurring

    fermentations are fundamentally ecological prob-lems and that to understand the behavior of anyorganism in its natural environment one must alsoconsider its interactions with other organisms inthe milieu. This aspect was considered previously,as was the question of analysis of pure cultures ofbacteria.

    Apart from these studies based on the isolationof bacteria on nonselective media, attempts havebeen made to isolate from anaerobic digestersbacteria which degrade the major components ofdigester input, namely cellulose, hemicellulose,starch, proteins and lipids, and these are con-sidered next. Experiments have shown that thebacteria with particular hydrolytic activities mayconstitute only a small part of the total flora in ahabitat like the rumen or anaerobic digester. Thus,even if these bacteria grow on a nonselectivemedium (as they should), unless a very largenumber of colonies are taken from such media it isunlikely that any representatives of specializedgroups will be found. This explains why thedescriptions of predominant bacteria found onnonselective media do not often agree withdescriptions of those found on selective media.Although a few hydrolytic bacteria can fermentonly a polysaccharide or its hydrolysis products,most ferment a variety of oligo- and monosac-charides and so can add to the general fermenta-tive ability of the flora.

    Cellulolytic BacteriaHungate,142 using an anaerobic roll-tube tech-

    nique, attempted to isolate cellulolytic bacteriafrom raw domestic sludge and digesting sludgefrom town digesters. Although none were isolatedfrom raw sludge, 0.8 to 2 X 103/ml were isolatedfrom the digesting sludge. The isolates consisted ofGram-negative rods and diphtheroids which variedconsiderably in size and were slightly curved.Large rods often exhibited internal granules at theends of the cells. The fermentation products oftwo isolates were determined after growth undernitrogen in a cellulose, yeast extract, phosphate-buffered medium. Both produced carbon dioxide,hydrogen, ethanol, and acetic acid, and one isolateproduced succinic acid. Other products might havebeen formed as the carbon recovery in the aboveproducts was only 50% of the degraded cellulose.These bacteria were similar to the rumenBacteroides succinogenes, except that they formedhydrogen and carbon dioxide.

    June 1974 171

    Dow

    nloa

    ded

    by [

    Cor

    nell

    Uni

    vers

    ity L

    ibra

    ry]

    at 2

    2:20

    13

    Nov

    embe

    r 20

    14

  • Maki,6 using similar anaerobic methods and amedium containing ground filter paper, salts, and30% digester liquor, obtained counts of 1.6 X 104

    to 9.7 X 10s cellulolytic bacteria/ml from domes-tic digesters. Like Hungate he isolated no cellulo-lytic bacteria from the raw sludge. The cellulolyticbacteria isolated by Maki resembled those isolatedby Hungate. Ten strains were isolated in pureculture and were separated into two groups bytheir colony characteristics. Bacteria of one grouphad high cellulolytic activity and producedprincipally hydrogen, carbon dioxide, ethanol, andacetic acid from cellulose. Mixed culture studiesusing a cellulolytic isolate and a noncellulolyticclostridium isolated from sludge showed that themixed culture fermented cellulose at more thantwice the rate of a pure culture, and indicated thatnoncellulolytic bacteria in sludge may increase therate of cellulose decomposition in digesters. It isinteresting to note that early experiments148 withthe rumen cellulolytic bacterium Ruminococcusflavifaciens showed a stimulatory effect of aclostridium on growth of the cellulolytic bac-terium, but clostridia are generally in very smallnumbers in the rumen.

    Harkness,149 using filter paper in culturesinoculated with digesting sewage, demonstratedthe presence of long chains of Gram-negativebacteria around the paper fibers, and Torien150

    found that although it was not possible to deter-mine, by microscopic observation, which morpho-logical types were cellulolytic, it was probable thatobligate anaerobic bacteria which produced apigmented layer on cellulose powder were themost important cellulolytic bacteria present.Facultative Bacillus spp., coliforms, pseudo-monads, and Gram-positive cocci possibly played asecondary role in the degradation.

    As previously mentioned cellulolytic bacteriacould not be cultured from piggery waste, butcould be isolated from established piggery-wastedigesters.35'36 Anaerobic cellulolytic bacteriawere found in numbers of 4 X 104 to 4 X 10s/ml.These compare with Hungate's and Maki's countsof 0.8 to 2 X 103 and 1.6 X 104 to 9.7 X 10s indomestic digesters, and counts of 4 X 103 indomestic digester sludge used to start a piggery-waste digester.3 s Since the number of cellulolyticbacteria increased as domestic digester sludge wasadapted to the high cellulose content pig waste, itseems likely that the variation in cellulolyticcounts of Hungate and Maki is a reflection of the

    paper content of the sewage. The cellulolyticbacteria from the piggery-waste digesters we--of11 different types, of which only 1 was Gram-positive. The most common fermentation productsof cellulose from these bacteria were acetic andpropionic acids; hydrogen was not determined andthe culture atmosphere was carbon dioxide.

    Although workers have endeavored to "shake"cellulolytic bacteria from fibers by differentmethods, it is likely that these counts will under-estimate the bacteria in digester contents. Also, aspreviously discussed, the use of filter paper sub-strates in media does not necessarily indicate thenumbers of bacteria able to digest untreated plantcellulose.

    Hemicellulolytic BacteriaShaw3 5 >6 3 found that considerable degradation

    of hemicellulose occurred during the digestion ofpiggery waste, but that no hemicellulose-degradingbacteria had apparently been isolated fromanaerobic digesters. Experiments showed thathemicellulolytic bacteria were present in thepiggery-waste digesters at similar levels to thecellulolytic bacteria (104 or more/ml). But unlikethe cellulolytic bacteria, the hemicellulolyticbacteria appeared to belong predominantly to onespecies identified with the rumen Bacteroidesruminicola. This bacterium could also have playeda major part in ammonia production throughdeamination of amino acids. A large number of thepiggery-waste digester bacteria were able to utilizeammonia as sole nitrogen source, and this includedthe hydrolytic and the generally fermentativebacteria.

    Amylolytic BacteriaTorien150 used enrichment cultures to deter-

    mine the aerobic and facultative anaerobic bacteriaparticipating in anaerobic digestion. He found thatstarch digestion was characterized by a dominantpopulation of rod-shaped organisms which some-times contained endospores. These were possiblyClostridium spp. Also isolated were Micrococcusspp., Bacillus spp., and Pseudomonas spp. whichmay possibly have contributed to the hydrolysis ofstarch.

    Amylolytic activity in piggery-waste digesterswas not confined to one group of bacteria;3s

    Bacteroides spp., Gram-negative coccobacilli, andClostridium butyricum were among the majorgroups. There were high counts of amylolytic

    172 CRC Critical Reviews in Environmental Control

    Dow

    nloa

    ded

    by [

    Cor

    nell

    Uni

    vers

    ity L

    ibra

    ry]

    at 2

    2:20

    13

    Nov

    embe

    r 20

    14

  • bacteria in the piggery waste fed to the digesters,probably a reflection of the fermentation of starchwhile the waste was standing under pig houses, andthe high count in the digesters was probably duemore to this than to the presence of large amountsof starch in the waste fermenting in the digester.

    Proteolytic BacteriaKirsch and Sykes1 quoted Agardy et a l . l s l as

    having measured the proteolytic enzyme activityin digesting liquor from a laboratory-scale digesterfed on a glucose-nutrient broth medium. Theenzymic activity was measured by "Thiogel" assayand was found to persist during normal digestion.When loading rate was increased to levels inducingdigester failure, onset of failure was accompaniedby an increase in proteolytic activity until theentire fermentation ceased and then the enzyme-activity fell to very low levels. They concludedthat changes in enzyme activity probably reflectedsuccessive alterations in bacterial types in thedigester ecosystem. It could, however, be a conse-quence of the effects of growth rate on theenzymic activity of the bacteria (see Section II.).

    Proteolytic Gram-positive bacteria were foundto be present in digesting sludge in numbersaround 7 X 104/ml by Harkness149 using serialdilutions into Robertson's meat broth. However,he concluded that these bacteria did not necessar-ily account for all the protein-decomposing bacte-ria present.

    Torien150 found that proteolytic activity wasassociated with the sporing rods which had alsobeen found to be the dominant proteolytic type indigesters fed with nutrient broth by McCarty etal.152 Torien1 s 0 also isolated Bacillus spp., pseu-domonads, and coliforms with some proteolyticactivity. These may have played a minor part inproteolysis, especially as Torien153 had earlier notobtained coliforms by direct-isolational proceduresfrom these anaerobic digesters. The role of Bacillusspp. is always difficult to determine as they areliable to be present in digester input from contam-ination of pipes, tanks, soil, etc., even if notpresent in the original faecal material. It is ofinterest to note that Appleby ls4 isolated proteo-lytic Bacillus sp. from the rumen, but concludedthat their numbers were mainly kept up from thepresence of about 106 spores and vegetative cells/gof the hay fed to the sheep, and later work1 s s

    showed that bacilli were not among the mainproteolytic flora of the rumen.

    Proteolytic enzyme level determinations and

    counts of proteolytic bacteria were carried out byKotze et al.156 They found that the enzyme levelsvaried with the nature of the waste being treated.Enzymic levels were low in a digester treating winewaste and one fed with industrial waste consistingprimarily of glucose and starch. A raw sewage-waste digester had a moderate proteinase level anda digester fed with a synthetic waste f dextrin,sucrose, nutrient broth, and casamino acids had ahigh enzyme level. These varied activities wouldappear to be more a reflection of substrate levelsavailable for microbial growth, especially availablenitrogen, than any response of the bacteria toincreased need for proteolytic activity, and theirrelevance to general digester analysis is doubtful.No bacterial details are available from this work.

    In 1969, Siebert and Torien157 enumeratedand studied the proteolytic flora of an anaerobicdigester using a high-protein medium containingskimmed milk, nutrient broth, digester super-natant liquor, vitamins, salts, and trace elements.Proteolytic bacteria producing a clear halo aroundthe colonies were present in numbers of 6.5 X107/ml. Forty-three isolates were randomly se-lected and classified. These consisted of 28 clostri-dia, 3 Gram-negative non-sporing bacilli, 8 Pepto-coccus anaerobus, 3 Bifldobacterium spp., and 1facultative anaerobic Staphylococcus sp.

    In his studies on piggery wastes Shaw35 '63

    found that analysis of input and output samplesfrom digesters showed little change in the totalprotein (N X 6.25) occurred during digestion.However, no distinction could be made betweenthe microbial and other proteins of the input oroutput, so it was not possible to say if anydegradation of nonmicrobial protein had takenplace. The presence of proteolytic bacteria sug-gested proteolytic degradation of the input, butthat this was balanced by utilization of thehydrolytic products in new bacterial synthesis.Counts of proteolytic bacteria on a mediumsimilar to that used for rumen proteolytic bacteria,but containing digester fluid for growth factors,gave results of the same order as those for otherhydrolytic bacteria (see above), and the principalproteolytic bacteria in the isolates examined wereclostridia.

    The results of examinations of digester contentsfor proteolytic bacteria in digesters shows apreponderance of Gram-positive bacteria, princi-pally clostridia. This is in contradiction to therumen where the principal proteolytic bacteria areall Gram-negative and of the same species as the

    June 1974 173

    Dow

    nloa

    ded

    by [

    Cor

    nell

    Uni

    vers

    ity L

    ibra

    ry]

    at 2

    2:20

    13

    Nov

    embe

    r 20

    14

  • other hydrolytic bacteria. On the other hand thecarbohydrate-hydrolyzing bacteria of digesters ap-pear to be similar to rumen types, and mainlyGram-negative. This seems unlikely to be anartifact as the isolation procedures for each habitathave often been similar, but why it should be so isnot immediately apparent.

    Lipolytic BacteriaDespite the fact that raw sewage sludges can

    contain large amounts of lipids (Table 1) there islittle information at present on the lipolyticbacteria present in digesters. In view of theprevious discussion on the breakdown of lipids(Section II.) it should be emphasized that lipolyticbacteria hydrolyze only the glycerides into glycer-ol and free fatty acids. There is no evidence forany lipolytic bacterium degrading a long chainfatty acid in the manner in which the acids areapparently degraded in digestion. It is also impor-tant to remember that true lipolytic activity is thehydrolysis of long chain fatty acid glycerides; thisis a much less widespread bacterial activity thanthat of hydrolysis of the glycerides of the shortchain, volatile fatty acids.

    Torien150 used sunflower oil enrichments toshow that vibrios, which were probably anaerobic,were the dominant lipolytic bacteria in anaerobicdigesters, but also suggested that Bacillus, Alcali-genes, and Pseudomonas might play a minor rolein lipolysis. These results may be compared withrumen studies where the only truly lipolyticbacteria so far characterized have been the Gram-negative, obligate anaerobic Anaerovibrio lipolyti-ca and some other vibrios differing slightly fromthe original isolates in fermentation reactions(Henderson, unpublished). But Bacillus sp. withesterase activity on short chain acid glyceridesformed a very minor part of the rumen flora.Anaerovibrio lipolytica also has esterase activity.

    Using the trilaurin medium of Henderson1 S 8

    which grows rumen bacteria of the A. lipolyticatype, the authors have found 104 or 10s lipolyticbacteria/ml in piggery-waste digesters, but have*not investigated the types of bacteria. A numberof bacteria from the piggery-waste digestions couldferment the glycerol which would be a primaryproduct of lipolysis.

    Other Specific BacteriaThe production of H2S or sulphides in media

    containing sulphur amino acids is a property of a

    number of anaerobic, carbohydrate-fermentingbacteria, but the sulphate-reducing Desulphovibriohave also been found in digesters. Torien et al.159

    using a modified Postgate medium found 3 to 5 X104 sulphate-reducing bacteria/ml. Desulphovibriodesulphuricans was an important part of the flora.When acid mine water containing high levels ofsulphate was added to the sewage digester sul-phate-reducing bacteria increased in number from6.6 X 103 to 9.5 X 107/ml.

    The possible formation of lactate as an inter-mediate in the fermentation of carbohydrates hasbeen mentioned above. The authors using ananaerobic lactate medium have found lactate-utilizing bacteria in numbers of about 3 X 107/mlin a piggery-waste digester. Ten representativecultures were tentatively identified as two Strepto-coccus sp., similar to S. lactis, five Bacteroidesspp., two clostridia, and one curved rod, possibly aDesulphovibrio. Their properties did not corre-spond exactly with known species. All producedmainly acetic and propionic acids (some withhigher acids) from lactate, but they could alsoferment other substrates.

    Torien1 s o reported the unexpected presence ofphotosynthetic bacteria in anaerobic digesters. Hesuggested that the development of numbers of theAthiorhodaceae would be favored by illuminationof the walls of transparent laboratory digesters andin view of their high efficiency of cell synthesisfrom fatty acids, hydroxy acids, keto acids, andalcohols, they might make a significant contribu-tion to the overall efficiency of the process.

    Methanogenic BacteriaThe possible routes of methane formation in

    digestion and the problems of isolation and charac-terization of methanogenic bacteria have alreadybeen discussed.

    What is given here is a resume of the counts andisolations of methanogenic bacteria as reported byvarious workers.

    Schnellen160 was the first to isolate methano-genic bacteria in pure culture from digestingsewage sludge. Two cultures were isolated andnamed. The first, Methanobacterium formicicumwas a Gram-negative rod variable in length andoften in chains. It produced methane from for-mate or hydrogen plus carbon dioxide. The sec-ond, Methanosarcina barkerii was a Gram-positivecoccus arranged in packets of eight or less, whichproduced methane from methanol, acetate, or

    174 CRC Critical Reviews in Environmental Control

    Dow

    nloa

    ded

    by [

    Cor

    nell

    Uni

    vers

    ity L

    ibra

    ry]

    at 2

    2:20

    13

    Nov

    embe

    r 20

    14

  • hydrogen plus carbon dioxide. Methanobacteriumformicicum was also isolated from digesters byMylroie and Hungate,161 and M ruminanthtm, aGram-positive short rod was isolated from diges-ters by Smith162 in 1961. This also producesmethane from formate or hydrogen plus carbondioxide.

    The first extensive isolations of methane bacte-ria were carried out by Buraczewski in 1964.163

    He used an enrichment technique in a liquidmedium followed by transfer to a solid medium.In addition to M. formicicum and Methanosarcinabarkerii he claimed to have isolated six moremethanogenic bacteria in pure culture. These wereMethanobacterium propionicum which producedmethane from propionate or hydrogen plus carbondioxide; M. suboxydans which utilized butyric,valeric, and caproic acids and also reduced carbondioxide; M. sohngenii which utilized acetaldehyde,acetic acid, butyric acid, and reduced carbondioxide; Methanococcus mazei which utilizedacetaldehyde, acetic acid, propyl alcohol, andbutyric acid; Methanococcus vanieli which pro-duced methane from formic acid, acetaldehyde,and hydrogen plus carbon dioxide. The sixth wasMethanobacillus omelianskii which has now beenshown to be an association of two bacteria, themethane producing one of which utilizes hydrogenplus carbon dioxide.73

    Smith85 also isolated pure cultures of methano-genic bacteria from digesting sewage sludge, in-cluding M. formicicum, M. ruminantium, andMethanosarcina barkerii. Three other hydrogen-utilizing strains were isolated, 2 Methanobacteriumspp. and 1 Methanococcus sp. Their utilization ofother substrates was not determined.

    Efforts have also been made to count thenumber of methanogenic bacteria in digestingsludge. Heukelekian and Heinemann86 developeda most probable number method. Serial dilutionsof a portion of digesting sludge were added to amineral salts, liquid medium supplemented with atrace of yeast extract and 1% of either acetate,butyrate, or ethanol. The medium was reducedwith sodium sulphide before inoculation. Invertedglass vials were placed in the tubes and the tubeswere sealed immediately after inoculation with"Vaspar." They found86 25 X 102 to 25 X 106

    acetic fermenters/ml, 6 X 102 to 25 X 106

    butyrate fermenters/ml, and 25 X 108 ethanol-utilizers/ml. Counts of 10s to 108 methanogenicbacteria/ml were obtained by Mylroie and

    Hungate16 ' using hydrogen plus carbon dioxide assubstrate. Using a method recording the highestdilution showing growth and methane productionSmith8 s found 107 hydrogen-utilizers/ml. He usedliquid and solid media with a single carbon sourceand single energy source. Liquid media gaveconsistently higher counts than solid media and heobtained counts of 105 to 106 acetate-utilizers/mland 107 ethanol-utilizers/ml.

    Siebert and Hattingh164 modified the mostprobable number technique of Heukelekian andHeinemann to estimate numbers of methane-producing bacteria. They suggested that formateplayed a central role in methane production andimplied that formate was an adequate substrate forenumerating methane production from acetic,butyric, and propionic acids, as well as formate.They used a mineral medium supplemented withvitamins and formate as sole source of carbon.Thiel et al.,16S however, thought this methodunsatisfactory and suggested the use of roll tubeswith a gas atmosphere of 20% carbon dioxide, 80%hydrogen, and a medium of mineral salts, vitamins,and trace amounts of fatty acids. Using this systemthey obtained counts of 1.5 X 108 to 3.03 X 108

    methanogenic bacteria/ml from a digester receivinga synthetic waste.

    Shaw35 '63 found 2 X 105 to 2 X 106

    methanogenic bacteria/ml in piggery-waste diges-ters using various media with hydrogen pluscarbon dioxide as substrate and counts of 2 X 104

    formate utilizers/ml and 2 X 104 butyrate utili-zers/ml using the appropriate substrates. No ace-tate or propionate utilization was found in thesedilution cultures which were incubated for amonth. In later work the authors used a techniqueto enrich acetate and butyrate-utilizing bacteria indigesting piggery waste. From the enrichmentcultures a number of cultures of bacteria wereobtained which produced methane from one ormore of the higher volatile fatty acids, acetic,propionic or butyric, as well as formate andhydrogen plus carbon dioxide, although only oneculture utilized propionate. The bacteria werepleomorphic but apparently consisted of two, orpossibly three, morphological types, a Gram-negative, variable length rod, sometimes in fila-ments (a short rod may have been the sameorganism), and a Gram-variable rod which occa-sionally showed spores. These bacteria were veryslow growing (8 to 12 weeks for cultures toproduce methane) and died off very easily. No

    June 1974 175

    Dow

    nloa

    ded

    by [

    Cor

    nell

    Uni

    vers

    ity L

    ibra

    ry]

    at 2

    2:20

    13

    Nov

    embe

    r 20

    14

  • factors added to the basic ammonia, mineral saltsmedium improved growth. Because of these diffi-culties the purity of the cultures has so far notbeen proved. The results do show that somebacteria, or association of bacteria, in the piggery-waste digesters will produce methane from acetic,propionic, or butyric acids.

    V. PRACTICAL ASPECTS OFANAEROBIC DIGESTION SYSTEMS

    IntroductionIn this section some practical aspects of digester

    design will be discussed, although the detailedaspects of construction and running of digestersare outside the scope of this review. Most of thefacts related to digester design have been takenfrom domestic sewage plants, as this is where thegreatest application of anaerobic digestion has sofar taken place. However, although domestic sew-age digesters have been in operation for manyyears the variety of design modifications andancillary equipment suggest that the perfect solu-tion to practical difficulties has not been obtained.In the authors' view it is practical difficultieswhich will hold up the application of anaerobicdigestion to other fields. While, for instance, thedigestion of farm waste has been shown to bepossible using laboratory-scale equipment, andalthough the authors have found that biochemi-cally, piggery-waste digestion has worked similarlyon each scale, they have had continued difficultieswith loading pumps and settlement of sludge infeed tanks when transferring their laboratorystudies to a continuous-loading 100-liter pilotplant scale.166 These same practical difficultiesand others have been found in the work of theauthors and their colleagues in building andrunning a small farmscale digester.167 The basicproblems are felt to be ones of scale. On mostfarms a digester system would, in terms ofdomestic digesters, be small - 100,000 gallonswould serve a large British farm. This means thatalthough the waste is of high solids content andcontains food residues, straw and other bedding,animal hairs or feathers, as well as general debrisfrom floors and tanks, and may be even moreintractable than domestic sludge, its rate of flowthrough pumps and pipes feeding digesters is lowcompared with the volumes pumped in the usualsewage system. And these problems, of smallpumps run intermittently, and settlement in pipes

    and tanks are some of the greatest to be overcomein the design of the smaller, easily run, but reliabletrouble-free digester system of defined perfor-mance. A further difficulty is one of cost ofequipment and servicing. While domestic systemsare financed from the taxes as part of the cost ofrunning a town, treatment of pollutional loadsequivalent to a small town must be financed fromthe costs of a medium size, intensive farm and runby the ordinary farm staff. The equipment thencannot be too costly or elaborate.

    Although, as previously discussed, a number oftheoretical considerations and results of laboratorywork can now be brought into the design andrunning of anaerobic digesters, many of these factsand theories (such as those of continuous culture)have only just recently been discovered or workedout, often on the basis of work unconnected withsewage digestion. So, many of the data for thedesign and running of present-day digesters havebeen obtained from experience in running full-scale plants. Some facts, mainly based on suchexperience, will be considered in this section.

    Conventional and High-rate DigestersIn this modern age anaerobic digestion is a far

    more sophisticated process than in the late 19thcentury when the cesspool, probably the firstanaerobic process applied to waste treatment, wasdeveloped in France by Louis Mouras.168 Whilethe "Mouras Tank" was designed for individualhouseholds, bigger tanks called "septic tanks"were designed and built and capable of dealingwith sewage from sizeable towns: e.g., the "AustinTank" in America, and the septic tank built byDonald Cameron for the town of Exeter inEngland from which he collected the gas for use instreet lighting. By the early 1900s the design ofseptic tanks had been improved and there were, forinstance, the "Travis Tank" in England and the"Imhoff Tank" in Germany.169 With increasingurban populations and stricter controls on pollu-tion, more houses with individual septic tanks andother sewage disposal systems became linked tomain sewers and centralized sewage works, and thesewage plant became more refined and morecomplicated. In 1911 the city of Birmingham inEngland installed a primary digestion plant whichwas capable of handling sewage from a very largepopulation and from which electricity could begenerated.170 This plant is one from which thepresent systems of anaerobic sewage digestiongrew.

    176 CRC Critical Reviews in Environmental Control

    Dow

    nloa

    ded

    by [

    Cor

    nell

    Uni

    vers

    ity L

    ibra

    ry]

    at 2

    2:20

    13

    Nov

    embe

    r 20

    14

  • The conventional anaerobic digester is basicallya holding tank into which wastes are pumpedeither continuously or periodically and retaineduntil a reasonably good reduction in pollutionalload has been achieved. The digestion tank may beof any shape, but circular ones, for obviousreasons, are by far the most common, and thesemay be constructed either of steel or of concrete.Concrete is generally used because of its lowercosts and resistance to corrosion. Conventionaldigesters are usually unheated and unstirred andthis leads to the formation of three zones ofactivity: a scum layer, an actively digesting layer,and a stabilized sludge and grit layer (Figure 2). Aslong as there was a 30- to 60-day retention time,coupled with a fairly low loading rate, the conven-tional digester performed, and will still perform,satisfactorily. However, as loading rates increasedand retention times were lowered, problems ofdegree of waste purification and control of scumlayers increased, and these led to other considera-tions being taken into account in the design anddevelopment of digesters,169 and so to the "high-rate" digester. Theoretical aspects of high-ratedigestion were discussed previously, but practicalconsiderations are the volume of sludge to bedigested per day, retention times, methods ofstirring, methods of heating, pumping and piping,gas collection, and insulation. Therefore, efficientanaerobic digestion requires a system of tanks,probably with floating covers, a means of stirring

    the digester contents, heat exchangers, a simpleyet effective pipe layout, circulation and transferpumps, and a method of collecting and using thedigester gas.

    For the digesters themselves the recent trendhas been towards deeper tanks (60 feet or morehigh) with smaller diameters.170 This design hasseveral advantages over the squatter tank, such as asmaller gas-holding volume, and easier sludgecirculation. Also operational requirements for thedigestion process seem to be vertical rather thanhorizontal, in that gas movement is upwards andsludge circulation is upward and downward toprevent stratification.

    The capacity of the tank should be sufficient tominimize the effect of variations in the characteri-stics of the raw sludge, and since, as discussedlater, the minimum detention time is about tendays, the capacity should be usually about ten totwelve times the maximum daily loading volumesexpected. The bottom of the tank should besloped at least 1 in 3 to allow for the removal ofsand and grit,171 and a means of stirring (seebelow) should be fitted. The stirring equipmentshould not involve machinery which cannot beremoved without seriously interrupting digesterfunctioning and should be of such a design that itinvolves very little maintenance (e.g., sealed bear-ings). Some form of weir overflow must be fittedsuch that the contents of the digester can bedischarged to balance the input without discharge

    CK*

    IXXXXXXXXXX SCUM REMOVAL

    SUPERNATANT

    INPUT

    SLUDGEREMOVAL

    FIGURE 2. Conventional sludge digestion unit.

    June 1974 177

    Dow

    nloa

    ded

    by [

    Cor

    nell

    Uni

    vers

    ity L

    ibra

    ry]

    at 2

    2:20

    13

    Nov

    embe

    r 20

    14

  • of gas, and usually at least two discharge pipes atdifferent levels are needed.171 The tanks can beabove, or partially or completely below ground.The latter helps insulation, especially in extremeclimates, and can be visually desirable, but it maylead to greater expense in making pipework etc.accessible. Although steel tanks may require insu-lation, thick concrete is generally a reasonableinsulator by itself.

    Whatever the material of construction of thetank, digester covers are usually made from about1/4 in. steel plates welded to form a gas-tight sheetand insulated. Although fixed volume digesters areused, many digesters have covers floating directlyon the digester contents, the liquid forming agas-tight seal with the skirt of the cover. A floatingcover can compensate for negative as well aspositive changes in volume of digester contentsand provides a reliable means of collecting the gasunder constant pressure. Gas take-off pipes areneeded and the cover should, for safety, have avacuum and pressure relief valve and gas-tight,glass inspection ports.

    Sludge input and output piping should be keptas simple as possible, with a minimum number ofvalves and bends. Pipe diameters should be select-ed on consideration of maintaining reasonablyhigh velocities to reduce sludge settlement, andlayout should provide easy access for internalclearing. Condensate pots and vents running toatmosphere should be provided for in gas pipelines.

    Sludge pumps should have adequate capacityfor the volumes to be pumped, but the size andtype is also dictated by the characteristics of thesludge being moved. Care should be taken toensure that pump glands are air-tight either mech-anically or with a form of water seal, as even aslight air leak can cause a large decrease in digesterefficiency and a gas consisting mainly of nitrogenand carbon dioxide (see Section II).

    A suitable gas holder is also needed, as, even ifthe gas is used only for heating the digester,fluctuations in gas supply must be evened out. Thegas is usually stored at a few inches water gaugepressure, which is that of the digester itself. Ameans of safely burning off excess gas is alsoneeded, if this cannot be put to use, and othersafety features are flame traps in gas lines con-nected to burners in heating systems. The rate ofgas production needs to be measured, as this is anindex of digester function. The rate can be

    measured by means of a steel or cast iron bodydiaphragm-type meter. The authors have foundcorrosion difficulties with wet-type gas meters ona pilot plant, and there have been reports ofdifficulties due to corrosion of bellows-typemeters. The experimental farm-scale piggery-wastedigester previously mentioned is fitted with rota-meter-type equipment with a millivolt output to arecorder. Corrosion problems with gas meters areconnected with traces of hydrogen sulfide, andpossibly ammonia, in the gas. To minimize cor-rosion of meters and other equipment (e.g., morethan 0.25 to \% hydrogen sulfide is said to causecorrosion of gas engine cylinders) the gas needs tobe water- and chemically-scrubbed. Ferric oxidesponge will remove hydrogen sulfide. Only in thelarger sewage works will the gas be regularlyanalyzed; changes in volume, or lack of flammabil-ity is generally a sufficient index of digesterfunction.

    Heating of the digester is done from thedigester gas in one way or another, a sludge heateror heat exchanger usually being used with externalcirculation of the digester contents. Modern heatexchangers can control digester temperature rea-sonably well as long as the circulation of thedigester contents is sufficient and thermostatprobes are correctly positioned in the tank. Thesludge can be heated by passing it in pipes througha water bath heated by direct gas flames, or thewater can be heated in an external boiler andpassed through other pipes in the exchanger.Internal heating by circulation of hot water from agas boiler might be more efficient, but problemscan arise in the event of corrosion or breakage ofthe pipes in the digester. In some large workswhere electricity is generated from the digester gasthe cooling water from the gas engines is used toheat the digester. Piping in the heating systemneeds to be well lagged as large heat losses canoccur in long pipe runs. Also the digester tempera-ture control needs to be as good as possible, as thisnot only keeps the microbial activity constant (seeSection II), but a very large amount of gas can bewasted in heating even a moderate size digestertwo or three degrees above its normal runningtemperature. The heat of fermentation is found inpractice to be not sufficient to maintain digestertemperature under normal temperate climatic con-ditions, but the amount of heat to be supplied tothe digester will obviously be governed by many

    178 CRC Critical R eviews in Environmental Control

    Dow

    nloa

    ded

    by [

    Cor

    nell

    Uni

    vers

    ity L

    ibra

    ry]

    at 2

    2:20

    13

    Nov

    embe

    r 20

    14

  • factors. Pohland172 has reported laboratory stu-dies on heat generation in anaerobic digestion.

    In general practice in large plants it is better tohave a number of smaller digesters rather than oneor two very large ones in case of failure of eitherthe digestion process or the plant.

    Many digesters are run between conventionaland high-rate, in that digester mixing is notcomplete. For instance, mixing by using tangen-tial-flow heat exchanger inlets and outlets may notbe sufficient to provide "high-rate" mixing. Thehigh-rate process is one that requires completemixing of the digester contents by either acontinuous or periodic system to maintain a homo-geneous "mixed liquor" stage in all parts of thetank. This approaches more the ideal, continuousculture system103 (see Section III). Some factorswhich are difficult to control in conventionaldigesters, but which are overcome in the high-rateplants, are thermal homogeneity, biological bal-ance, substrate-microorganisms contact, and sol-ids-water ratio.173 Therefore, high-rate digestionallows either higher solids loading per unit volumeor shorter retention times than the more conven-tional digestion.

    There are various methods of stirring used to

    maintain high rate digestion. Mixing by externalcirculation of digester contents was investigated byTorpey,174 but he concluded that internal mixingwas more efficient and practical.

    Gas recirculation can be used175 and with therecirculation of adequate volumes of gas, scumformation can be controlled and a uniform digest-ing mixture be maintained (Figure 3). Since thereis complete mixing and no scum formation, thereis no need for periodic cleaning. A 5 slope on thedigester floor is sufficient to allow any grit thatfalls out of suspension to be swept by thecirculating gas currents to the center for draw off.Grit removal should be a regular operating proce-dure.176 This system does involve extra handlingof gas, though, and would not be suitable for asmall plant.

    Mechanical mixing is more common and al-though early stirring devices did not give completemixing, later designs are more efficient. A simplepaddle system is, obviously, almost a mechanicalimpossibility in a large digester filled with amaterial of the consistency of sludge, so differentstirring methods have been developed, usually onthe principle of circulating the contents throughinternal pipes. Placing the draft-tube mixers tan-

    GAS

    SLUDGE

    GAS D1FFUSERSSLUDGE &

    GRIT REMOVAL

    FIGURE 3. High-rate digestion unit using gas recirculation. (Developed by theChicago Pump Company.)

    June 1974 179

    Dow

    nloa

    ded

    by [

    Cor

    nell

    Uni

    vers

    ity L

    ibra

    ry]

    at 2

    2:20

    13

    Nov

    embe

    r 20

    14

  • gentially across the bottom of the tank helps tokeep the entire contents mixed173 (Figure 4).Another form of mixing is by a large archimedeanscrew housed in a tube with a bell mouth, placedcentrally in the digester. This can be set to pull upor push down the digester liquor through the tubeand so set up a circulation.

    Since solution of oxygen is not the object ofstirring as it is in most industrial fermentationprocesses, the rate of stirring need not be fast andmost mechanical stirrers are run intermittently.

    With high-rate digestion satisfactory resultshave been obtained with domestic sewage withdetention times of between 7 and 12 days,177 '1 7 8

    and for piggery waste, 10 days, 1 6 6 ' 1 6 7 butattempts to shorten the retention times were metwith process breakdown.179 By loading the sludgeas densely as possible and operating on a ten-daydetention time the capacity of the digester tankcan be cut to about one fifth of that needed forconventional digestion. Therefore, it would seemthat by operating a high-rate digestion systemcapital costs would be cut by about 25% of theamount required for conventional treatments.Heat losses are reduced in almost direct relation tothe surface area of the tank,173 and the biologicalbenefits have already been discussed.

    Criteria of Good DigestionThe measurements used and general results ob-

    tained in good digestion are summarized below:

    1. Gas production in quantity and quality -Reports of gas production vary, and obviously thecomposition of the waste will be a factor, butsomewhere about 7 cu ft/lb volatile solids addedto the digester seems to be a reasonable figure fora number of domestic and farm wastes. A methanecontent of 55 to 72% in the digester gas indicatessatisfactory digestion, although higher figures havebeen reported for some specialized wastes (e.g., seeReferences 180 and 181).

    2. Volatile fatty acids (VFA) - Two to fourhundred mg/1 VFA as acetic would be normal in agood digestion. From the authors' experience awell-balanced farm waste digester can cope withregular VFA inputs of 4,000 to 6,000 mg/1.

    3. pH - A pH of 7.0 to 8.0 is in the range ofsatisfactory operating conditions. Below pH 6.5digestion falls off, and below pH 5.0 almostcompletely stops.

    4. Total and volatile solids - Inputs be-tween 3 and 6% total solids maintain satisfactorydigestion, but solids concentrations of up to about10% are used. Below about 2% the digester can

    DRIVE-UNIT

    *TP

    SLUDGEINPUT

    GAS COLLECTION

    cm LTD

    6/

    DEFLECTOR.VANES

    -CIRCULATINGMIXER

    - SLUDGEREMOVAL

    FIGURE 4. Draft-tube mixing.

    180 CRC Critical Reviews in Environmental Control

    Dow

    nloa

    ded

    by [

    Cor

    nell

    Uni

    vers

    ity L

    ibra

    ry]

    at 2

    2:20

    13

    Nov

    embe

    r 20

    14

  • begin to wash out. Total solids are usually reducedby 40 to 50% by digestion.

    5. Odor Well-digested sludge has a fainttarry smell and is blackish in color and solidsshould settle out reasonably well.

    6. Grease This should be almost totallyabsent from well-digested material.

    7. Temperature The optimum tempera-ture for high-rate digestion is about 95F (35C).Below 90 F and falling to 50 F digestion slowsdown and then almost ceases. No permanentdamage is done by raising the temperature to 100to 105F, but digestion, and consequently gasproduction, are slightly decreased.

    8. Biochemical and chemical oxygen de-mand - Reduction in these varies, but the valuesgive a measure of the extent of purificationattained. Digesters can cope with BOD's of over100,000 ppm.

    9. Alkalinity About 2,500 mg/1 is consid-ered normal, with that in the raw sludge about1,000 to 2,000 mg/1.

    Other Types of DigestersSince some of the theoretical aspects of two

    stage and recycling systems have already beendiscussed, a few practical aspects will be men-tioned here. Most high-rate digestions, as men-tioned before, run on a form of two stage principlein that the sludge settling tanks form a secondunheated stage. Recycling of digested sludge ispracticed to some extent in some high-rate sys-tems, but its real application has been in the

    "anaerobic contact system." A schematic diagramof the contact system is shown in Figure 5.

    The contact process is most efficient whentreating wastes of reasonable suspended solidscontent because the microbial growth becomesattached to the particles and so settles out easilyand is readily separated from the effluent stream.

    There have been various modifications in designof the contact process depending on the wastetreated. For example with meat-packing h'ousewas te s 1 8 0 ' 1 8 2 excellent results have beenachieved, but the problem of degassing the solidsin the settling tank had to be overcome. This wasdone by designing a vacuum degasifier which wassimply a steel tank situated on top of the digesterand into which the wastes were drawn by applyinga 20 in.-vacuum. Then the wastes were allowed toflow over a series of slats to permit the release ofgas, which was pumped off.180 The quality of thegas removed from the degasifier was very poor,being only about 36% methane and 63% carbondioxide (a reflection of the solubility of carbondioxide in the digester liquid), compared with thedigester gas coming off naturally, which had amethane content of 85% with 11% carbon dioxide.The high methane content of this gas comparedwith normal digester gas could be a reflection ofthe high rate of return of the degassed liquid andsolids to the digester. This could give a largevolume of liquid to be saturated with carbondioxide.

    Another modification of the anaerobic contactprocess has been used in the Bowery Bay Pollution

    HOLDING

    TANK

    DIGESTER DEGAS1 FIER SETTLING

    TANK

    EFFLUENTRAW

    SLUDGE

    RETURN SLUDGE WASTE SLUDGE

    FIGURE 5. Schematic flow diagram of the anerobic contact process as applied to packing house wastes.

    June 1974 181

    Dow

    nloa

    ded

    by [

    Cor

    nell

    Uni

    vers

    ity L

    ibra

    ry]

    at 2

    2:20

    13

    Nov

    embe

    r 20

    14

  • Control Plant.181 This is really an extension ofthe sludge return practiced to some degree in anumber of domestic plants. This system showedthat controlled thickening of the combined rawprimary sludge and digested sludge could giveadvantages in that the digester did not have to dealwith an excessive water load, that the digestionwas carried out in less than one third of the tankvolume formerly needed and that there were,therefore, substantial savings in capital costs fortank construction. The input of thickened sludgevaried from 8 to 10% total solids, while outputvaried from 4.6 to 6.1% total solids. Such ascheme works if the liquid from the input is nothigh in dissolved biodegradable materials, but inthe case of, for instance, the piggery wastepreviously mentioned, there is a high acid contentin the input liquid. This not only contributes togas production, but its reduction is one of theobjects of the anaerobic digestion.

    Another type of anaerobic contact process, inwhich wastes pass upwards through a mass ofconcentrated anaerobic biological solids, has beenused to treat fermentation process waste1 8 3 '1 8 4

    and domestic waste.185 This process could beused to treat a wide variety of wastes, especiallywhere they are concentrated, or naturally warm,but so far it has not been very satisfactory intreating wastes of less than 2,000 ppm BOD. Thesesystems are similar to the anaerobic filter processwith which Young and McCarty130 have carriedout detailed studies on a laboratory scale and thetower beer fermenten A diagrammatic anaerobicfilter is shown in Figure 6. For full-scale use theanaerobic filter has the advantages of beingsuitable for the treatment of low strength solublewastes,186 having minimal hydraulic head require-ments, and having very little sludge handling anddisposal. Construction, operation, and mainte-nance costs are low, and there is a by-product ofgas as in other digestion processes.

    The anaerobic lagoon has a number of disad-vantages, as was mentioned in the introduction,but it has been applied to the treatment ofpetrochemical wastes,187 where it has potential asa pretreatment step in overcoming some of thedifficulties encountered in treating this waste byconventional methods. These difficulties includehigh temperatures, surfactants causing foamingproblems, and high rates of oxygen demand. Theanaerobic lagoon is capable of absorbing majorchanges in the quality of the waste without

    OUTPUT

    INPUT

    FIGURE 6. Schematic diagram of the anerobic filterprocess.

    altering the overall system contents, the large areaallows cooling for later aerobic treatment, andsome problem materials, such as surfactants, canbe stabilized anaerobically so there can still be aplace for the anaerobic lagoon in the partialtreatment of specific organic wastes.

    VI. CONCLUSION

    In the previous pages we have attempted to givean overall picture of present knowledge of an-aerobic digestion and it is obvious that this isfragmentary. Work on anaerobic digestion can, likework on any other fermentation process, bedivided into "laboratory" and "practical" (pilot orfull-scale plant) aspects, but there is a fundamentaldifference between most industrial fermentationsand anaerobic digestion. In industrial fermenta-tions a particular product, be it cell mass ormicrobial metabolite, is desired. Some componentof the culture medium may be defined for reasonsof cost or availability, but the complete medium isformulated to give, along with the experimentally

    182 CRC Critical R eviews in Environmental Control

    Dow

    nloa

    ded

    by [

    Cor

    nell

    Uni

    vers

    ity L

    ibra

    ry]

    at 2

    2:20

    13

    Nov

    embe

    r 20

    14

  • optimized growth conditions, optimum formationof the product. This medium is then used inscale-up of the process from laboratory to pilotand large-scale plant. In most anaerobic digestionsalthough a product (stabilized degraded sludge andgas) is desired, the "medium" is defined andcannot be altered; the only things that can bechanged to optimize the process are the para-meters of operation of the digester and thesedepend on the kinetics of microbial attack on thatparticular medium. So although the medium for anindustrial-scale fermentation is defined by labora-tory experiments, the medium for an industrial-scale digester defines that for laboratoryexperiments. This means that laboratory experi-ments with artificial media can have only slightrelevance to the operation of full-scale plant.Artificial media here include actual wasteswhich have been ground, pulverized, sieved orotherwise physically treated to make them moreamenable to small-scale use, because apart fromobvious considerations there is much evidencefrom other experiments on, for instance, therumen, that physical change in a substrate canmarkedly alter its behavior under microbial attack.Unless some very specific point is being investi-gated, then laboratory experiments on anaerobicdigestion should use the substrate defined by thelarge-scale application if the results are to berelevant to the large-scale process. This applies notonly to the overall process of the waste degrada-tion but to the intermediate reactions. The intro-duction of large amounts of some postulatedintermediate into a digestion may not only alterthe kinetics of the reaction, as it may overloadsome step in the pathway, it may alter the rate ofsome usually minor reaction and be diverted intoother products, and it may alter the balance of theflora by acting as an enrichment culture. Further-more, the question of stability of experimentaldigesters must be taken into account. A contin-uous culture with a turnover time of 30 dayscannot be expected to give "steady-state" resultsafter two weeks of running; experiments onanaerobic digestion must necessarily be tediousand long-term.

    All these points may be said to be self-evidentbut they do seem to have been overlooked inmany experiments, and this, as has been pointedout, is one of the difficulties encountered incorrelating the results of different experimentsinto a coherent whole.

    Digestion of domestic sewage has probablybeen about optimized, or could be optimized, inthat the basic parameters for efficient digestion areknown. What set limits to the actual optimizationin practice are largely the physical aspects ofmoney and manpower available for building andrunning the plant and the nature of supply of thefeedstock. A fully efficient plant may have to beset against the costs and difficulties of stirring, andso on. Soieports (e.g., see Reference 188) on theoperation of full-scale domestic digesters are asimportant in obtaining the greatest efficiency inanaerobic digestion as are laboratory reports. Areport on anaerobic digestion of poultry-processing waste plus domestic sewage in a townplant189 said that "soon after initial operation itwas determined that the problems related tohandling and treating this waste were largelymechanical." The matter of practical difficultiesvs. costs and efficiency may be a large considera-tion in the application of digestion to otherwastes, such as farm wastes for example, wherealthough the pollutional load of an intensive farmcan be equivalent to that of a medium-sized town,the money available for waste treatment is, inpresent economic circumstances, much less thanthat available to the town. This must be kept inmind in research on anaerobic digestion; while itmay be relatively easy to demonstrate and opti-mize digestion in a laboratory apparatus, thedifficulties of scale-up to a relatively cheap,efficient, working size may be formidable.

    However, although the basic principles ofdomestic sewage digestion may be known, labora-tory and small pilot plant experiments can still beneeded, and be valuable, in investigation ofspecific matters brought up by modern conditions;one thinks immediately of the effects of tradewastes, including radioactive wastes,190 deter-gents, and so on. In the fields of extension ofdigestion to farm sewages, factory wastes, andmixtures of wastes (for example, domestic sewageand garbage) then there is obvious scope forsmall-scale experiments before larger schemes areembarked upon. The former will show if digestionis possible, optimum detention times, the correctmixtures of wastes, and other things, and maysuggest different designs of plant, but it must notbe overlooked that scale-up of the process maypose many more problems and take longer thanlaboratory experiments.

    Anaerobic digestion has been used, along with

    June 1974 183

    Dow

    nloa

    ded

    by [

    Cor

    nell

    Uni

    vers

    ity L

    ibra

    ry]

    at 2

    2:20

    13

    Nov

    embe

    r 20

    14

  • aerobic treatment, as part of systems for purifica-tion and stabilization of domestic sewage. It alsohas possibilities as a sole or primary treatment forfarm animal waste, or mixtures of animal andother farm wastes such as silage liquid (authors,unpublished), as these are of high pollutional andsolids content. A number of workers besides theauthors have shown that animal wastes will digest,at least on a laboratory scale (e.g., see References191 and 192). It has been, or could be, used forindustrial wastes in either the high-rate digester orthe modifications previously discussed. In all theseapplications the process is primarily one of wastepurification. The fact that combustible gas is alsoproduced has, of course, always been known, andthis gas has been a means of offsetting the costs ofwaste treatment by use in heating digesters,powering the plant, and sale to outside interests.But with the future prospect of scarcity and highprice of oil and related fuels, brought forward tothe present by the political events of the last fewmonths, attention has been increasingly focusedon the prospect of using gas from microbialdigestion of organic matter as a fuel. The prospecthas appeal and it has been the subject of many,often rather vague, talks and papers by peoplewith little expert knowledge of the field. Theauthors have been concerned in this problem fromthe point of view of their work on farm wastes;others have considered it, for example, from thepoint of view of mixing domestic garbage andsewage.1 9 3 All these processes are not novel; forinstance, digestion of domestic garbage and sewagewas reported on some years ago,194 but they needto be reexamined in the new light of present-dayoil and fuel costs and scarcity. While the possiblecontribution of microbial gas to the fuel require-ments of a nation may be only marginal, itscontribution to the requirements of individualfarms or factories could be useful. It must not beforgotten that these farms or factories have awaste-disposal problem and if they can alleviatethis and at the same time obtain at least somecontribution to fuel requirements, then the pro-cess is worthwhile. It is in this context of wastedisposal that the question of microbial gas pro-duction must, at present, be discussed. The sug-gestion has been made that crops could be grownespecially to provide substrates for production ofmicrobial gas. This might eventually become apossibility, but at the moment there are manyorganic wastes available in different kinds and

    quantities. Some of these wastes are produced allthe year round, in large quantities, at specificplaces. Others, while in large quantity in total, areavailable only in comparatively small amounts atany one place, or are available only seasonally. Thefirst could be digested on site, the second poseproblems of transport to central digester sites, orof digesters running on different substrates atdifferent seasons, and this would add to the costsof gas production. The use of pure substrates, suchas acetate, for methane production has beensuggested. This would theoretically require onlyone type of bacterium, but consideration of theproperties of the methanogenic bacteria showsthat the conditions of Eh needed for growthwould be almost impossible to obtain on a largescale (see Reference 195 for consideration of thelarge-scale growth of anaerobic rumen-type bac-teria). The production of the feedstock acetic acidwould also require energy. The optimum condi-tions for growth of the methanogenic bacteria aremost easily obtained in mixed culture. This mixedculture can then provide the substrates formethane production from a primary substrate oflow, or effectively no, cost.

    The questions of cost of microbial gas aredifficult to determine and in many cases furtherexperimental work is necessary before even abeginning of costing can be made, but if the costsof digestion as a necessary waste-treatment systemare comparable to other treatment systems (if suchare possible) then the cost of the gas can be takenas nil or negligible. However, with costs of fossilfuels rising almost daily, any comparisons aredifficult, but tend more and more to improve theeconomics of microbial gas. Scarcity of oil-basedfuels, however, to some extent outweighs thequestion of costs, and here we have the reason forsaying previously that some contribution to fac-tory or farm fuel is worthwhile, because if fuelallocations are cut by 10%, if even this 10% can bemade up by gas from the waste-disposal system,the primary process can be kept going at full rateand it must be remembered that the presentstandard of feeding of much of the world dependson intensive farming and food-processing factories,which require large power inputs.

    Generation of usable gas from small-scale farmsewage digesters is a feature of rural life in someeastern countries, but these simple plants (e.g., seeReference 196) give no guidelines for building oflarger scale plants and are generally inefficient. For

    184 CRC Critical Reviews in Environmental Control

    Dow

    nloa

    ded

    by [

    Cor

    nell

    Uni

    vers

    ity L

    ibra

    ry]

    at 2

    2:20

    13

    Nov

    embe

    r 20

    14

  • instance, figures of 0.5 to 1.5 cu. ft. gas/lb ofcattle waste were quoted for farm plants inIndia.19? Such plants also rely on the hot climatefor keeping the digestion at a reasonable tempera-ture, so no heating is required and all gas isavailable for use. There are, or have been, digestersof different kinds (batch and continuous) run inwestern countries but although these may providesome gas, those within the authors' knowledgeare not efficient from the point of view of gasproduction or reliability of working, and their useas sewage purification systems has, really, not beenconsidered. The design and operation of a farmplant of as small as possible capital cost, economicrunning (both in power inputs and manpower),safety, efficiency, and reproducibility in gas pro-duction and waste purification is still a matter ofinvestigations such as the authors and theircolleagues are carrying out.167 We have producedfigures, based on pilot plant experiments andsupposed heating requirements of input and di-gesters for the gas production from farms ofdifferent sizes, but while these give an indicationof power available as heat or mechanical orelectrical energy, they cannot be absolutely quan-tified until further experiments with small andlarge-scale plants have been carried out. Small-scaleexperiments will show if results from digestion ofone type of waste can be extrapolated to othersand if theoretical amounts of gas can be obtained.They will show detention times and the characterof the digester effluent, and so on. Only larger-scale working can show up problems of handlingthe waste, digester stirring, heating needs under allclimatic conditions, and materials and methods ofconstruction. Many of these things one wouldsuppose would be known from domestic plants,but the problem of relative costs has been men-tioned and there are also problems of scale. Onlythe very biggest farm plants would come withinthe usual range of sizes of domestic digesters andcommercial digesters and ancillary equipment,such as pumps, are related to this size. While farmwaste can be even more intractable in handlingthan domestic sludges, the amounts to be movedare well below the capacities of the usual sludgepumps and this causes problems. Such considera-tions of scale, handling of input, collection ofinput waste, and many others will have to be takeninto consideration if factory and other wastes are

    to be utilized. There is then the question ofdisposal of output and if this is a saleablecommodity then a contribution towards digesterrunning costs is obtained. Use as a fertilizer, soilconditioner, or land-fill is obvious, as the digestivewaste should be stable, compact, and of goodnitrogen content. The content of bacterial cellsalso suggests a use as microbial "protein" supple-ment for animal feeds. There could be difficultiesin possible pathogen contamination of "protein"from domestic or farm sewage digestion and thisneeds more investigation if such uses are envisaged.The presence of residual additives, such as thecopper in pig feeds, could also pose difficulties.However, the possibilities of using food-factorywastes as substrates for digestion make the pro-duction of definitely pathogen-free "proteins"easier. While the possibilities of selection of themost efficient bacterial population for breakdownof waste such as sewage seem remote because ofdifficulties of complexity of substrate and thelarge bacterial population of the waste, digestionof factory waste of defined composition should,theoretically, offer possibilities of buildup of adigester flora of known and optimum activity.This would lead to a microbial feed output ofdefined quality.

    Again, the shortage and cost of hydrocarbonsor derived substrates make microbial feed pro-duction from organic wastes a more viable pro-position. Anaerobic digestion does not give as highcell yields as aerobic growth of microorganisms onthe waste, but it is fundamentally easier andcheaper in having no necessity for vigorousaeration, and it can give a useful product in thegas, so it should be considered.

    Considering the nature of the problem,anaerobic digestion of domestic sewage has come along way towards an efficient "industrial" process.Present and future work has, or will, lead to evenmore stable systems. But the greatest areas forexperiment with anaerobic digestion are only justbeing opened up. The future possibilities ofanaerobic digestion have now taken hold of men'sthoughts. Whether these possibilities are attainablecan be proved only by experiments and suchexperiments must be done because it is only on abasis of facts and figures that the future ofanaerobic digestion can be determined.

    June 1974 185

    Dow

    nloa

    ded

    by [

    Cor

    nell

    Uni

    vers

    ity L

    ibra

    ry]

    at 2

    2:20

    13

    Nov

    embe

    r 20

    14

  • REFERENCES

    1. Kirsch, E. J. and Sykes, R. M., Anaerobic digestion in biological waste treatment, in Progress in IndustrialMicrobiology, 9, Hockenhull, D. J. D., Ed., J. and A. Churchill, London, 1971, 155.

    2. Pohland, F. G., Developments in anaerobic treatment processes, Am. Chem. Soc. Div. Microbial Chem. Technol.,160th A.C.S. Natl. Meeting, 1970.

    3. Pohland, F. G., General review of literature on anaerobic sludge digestion, Eng. Ext. Bull., Purdue Univ., 20, 583,1962.

    4. McCarty, P. L., Anaerobic waste treatment fundamentals, part 2 - environmental requirements and control, PublicWorks, 95, 124, 1964.

    5. Buswell, A. M. and Neave, S. L., Laboratory studies of sludge digestion, Ill. State Water Surv. Bull., 30, 1930.6. Maki, L. W., Experiments on the microbiology of cellulose decomposition in a municipal sewage plant, Antonie van

    Leeuwenhoek, 20, 185, 1954.7. Chynoweth, D. P. and Mah, R. A., Volatile acid formation in sludge digestion, in Adv. Chem. Ser., No. 105,

    American Chemical Society, 1971.8. Wood, T. M., Cellulose and cellulolysis, World Rev. Nutr. Diet., 12, 227, 1970.9. Halliwell, G. and Bryant, M. P., The cellulolytic activity of pure strains of bacteria from the rumen of cattle, J. Gen.

    Microbiol., 32, 441, 1963.10. Van Gylswyk, N. O. and Labuschagne, J. P. L., Relative efficiency of pure cultures of different species of

    cellulolytic rumen bacteria in solubilizing cellulose in vitro, J. Gen. Microbiol., 66, 109, 1971.11. Hungate, R. E., The Rumen and Its Microbes, Academic Press, New York and London, 1966.12. Kotz, J. P., Thiel, P. G., Torien, D. F., Hattingh, W. H. J., and Siebert, M. L., A biological and chemical study of

    several anaerobic digesters, Water Res., 2, 198, 1968.13. Cook, A. R., Ureolysis in the ovine rumen, Biochem. J., 127, 66P, 1972.14. Jurtshuck, P. J., Doetsch, R. N., and Shaw, J. C., Anaerobic purine dissimulation by washed suspensions of bovine

    rumen bacteria, J. Dairy Sci., 41, 190, 1958.15. Whiteley, H. R., The fermentation of purines by Micrococcus aerogenes, J. Bacteriol., 63, 163, 1952.16. Whiteley, H. R. and Douglas, H. C, The fermentation of purines by Micrococcus lactilyticus, J. Bacterial., 61, 605,

    1951.17. Barker, H. A. and Beck, J. V., Clostridium acidi-urici and Cl. cylindrosporum, organisms fermenting uric acid and

    some other purines, J. Bacteriol., 43, 291, 1942.18. Campbell, L. L., Jr., Reductive degradation of pyrimidines. I. The isolation and characterisation of a uracil

    fermenting bacterium Clostridium uracilicum, nov. spec, J. Bacteriol., 73, 220, 1957.19. Gasser, R. and Jeris, J. S., Comparison of various nitrogen sources in anaerobic treatment, J. Water Pollut. Control

    Fed., 41, R 91, 1969.20. Garton, G. A., Influence of the rumen on the digestion and metabolism of lipids, in Digestive Physiology and

    Nutrition of the Ruminant, Lewis, D., Ed., Butterworth and Co., London, 1961, 140.21. Garton, G. A., Lipids, World Congr. Anim. Feeding (Madrid), 1,255, 1972.22. Keeney, M., Fat metabolism in the rumen, in Physiology of Digestion and Metabolism in the Ruminant, Phillipson,

    A. T., Ed., Oriel Press, Newcastle upon Tyne, 1970, 489.23. Garton, G. A., Lough, A. K., and Vioque, E., Glyceride hydrolysis and glycerol fermentation by sheep rumen

    contents, J. Gen. Microbiol, 25, 215, 1961.24. Dawson, R. M. C. and Kemp, P., Biohydrogenation of dietary fats in ruminants, in Physiology of Digestion and

    Metabolism in the Ruminant, Phillipson, A. T., Ed., Oriel Press, Newcastle upon Tyne, 1970, 504.25. Wood, R. D., Bell, M. C., Grainger, R. B., and Teekell, R. A., Metabolism of labelled linoleic-1-C14 acid in the sheep

    rumen, J. Nutr., 79, 62, 1963.26. Henderson, C. and Demeyer, D. I., Incorporation of linoleic [1-14C] acid into the lipids of mixed rumen

    microorganisms, World Cong. Anim. Feeding (Madrid), 2,45, 1972.27. Allison, M. J., Bryant, M. P., Katz, I., and Keeney, M., Metabolic function of branched-chain volatile fatty acids,

    growth factors for ruminococci, J. Bacteriol, 83, 1084, 1962.28. McCarthy, R. D., Mammalian metabolism of straight-chain saturated hydrocarbons, Biochim. Biophys. Acta, 84, 74,

    1964.29. Heukelekian, H. and Mueller, P., Transformation of some lipids in anaerobic sludge digestion, Sewage Ind. Wastes,

    30, 1108, 1958.30. Emery, R. S., Smith, C. K., and Huffman, C. F., Utilisation of inorganic sulphate by rumen, microorganisms. 1.

    Incorporation of inorganic sulphate into amino acids, Appl. Microbiol, 5, 360, 1957.31. Henderickx, H., Incorporation of sulphate in rumenal proteins, Arch. Int. Physiol, 69, 449, 1961.32. Hobson, P. N., Rumen microorganisms, in Progress in Industrial Microbiology, 9, Hockenhull, D. J. D., Ed., J. and

    A. Churchill, London, 1971, 41.33. Szemler, L. L. and Szkely, A. D., Vitamin B12 from sewage sludge, Proc. Biochem., 4, 12, 1969.34. Pohland, F. G. and Bloodgood, D. E., Laboratory studies on mesophilic and thermophilic anaerobic sludge

    digestion, J. WaterPollut. Control Fed., 35, 11, 1963.

    186 CRC Critical Reviews in Environmental Control

    Dow

    nloa

    ded

    by [

    Cor

    nell

    Uni

    vers

    ity L

    ibra

    ry]

    at 2

    2:20

    13

    Nov

    embe

    r 20

    14

  • 35. Shaw, B. G., A practical and bacteriological study of the anaerobic digestion of waste from an intensive pig unit,Ph.D. thesis, University of Aberdeen, Aberdeen, Scotland, 1971.

    36. Hobson, P. N., Continuous culture of some anaerobic and facultatively anaerobic rumen bacteria, J. Gen. Microbiol.,38, 167, 1965.

    37. De Vries, W., Kepteijn, W. M. C., Van der Beek, E. G., and Stouthamer, A. H., Molar growth yields and fermentationbalances of Lactobacillus casei L 3 in batch cultures and in continuous cultures, J. Gen. Microbiol., 63, 333, 1970.

    38. Hishinuma, F., Kanegasaki, S., and Takahashi, H., Ruminai fermentation and sugar concentrations, a modelexperiment with Seienomonas ruminantium, Agric. Biol. Chem., 32, 1327, 1968.

    39. Hobson, P. N. and Shaw, B. G., The bacterial population of piggery-waste anaerobic digesters, Water Res., in press.40. Henderson, C, Hobson, P. N., and Summers, R., The production of amylase, protease and lipolytic enzymes by two

    species of anaerobic rumen bacteria, in Continuous Cultivation of Microorganisms, Proc. 4th Symposium, Academia,Prague, 1969, 189.

    41. Van Gylswyk, N. O. and Roch, C. E. G., Characteristics of Ruminococcus and cellulolytic Butyrivibrio speciesfrom the rumens of sheep fed differently supplemented teff (Eragrostis tef) hay diets, J. Gen. Microbiol., 64, 11,1970.

    42. Hobson, P. N. and Summers, R., The continuous culture of anaerobic bacteria, J. Gen. Microbiol., 47, 53, 1967.43. Walker, D. J. and Monk, P. R., Fate of carbon passing through the glucose pool of rumen digesta, Appl. Microbiol.,

    22, 741, 1971.44. Novelli, G. D., The exchange of H 14COOH with the carboxyl group of pyruvate by Clostridium butyricum and

    Micrococcuslactilyticus, Biochim. Biophys. Acta, 18, 594, 1955.45. Dean, A. C. R., Influence of environment on the control of enzyme synthesis, in Environmental Control of Cell

    Synthesis and Function, Dean, A. C. R., Pirt, S. J., and Tempest, D. W., Eds., Academic Press, London, 1972, 245.46. Bull, A. T., Environmental factors influencing the synthesis and excretion of exocellular macromolecules, in

    Environmental Control of Cell Synthesis and Function, Dean, A. C. R., Pirt, S. J., and Tempest, D. W., Eds.,Academic Press, London, 1972, 261.

    47. Demain, A. L., Cellular and environmental factors affecting the synthesis and excretion of metabolites, inEnvironmental Control of Cell Synthesis and Function, Dean, A. C. R., Pirt, S. J., and Tempest, D. W., Eds.,Academic Press, London, 1972, 345.

    48. Brown, C. M. and Stanley, S. O., Environment - mediated changes in the cellular content of the "pool"constituents and their associated changes in cell physiology, in Environmental Control of Cell Synthesis andFunction, Dean, A. C. R., Pirt, S. J., and Tempest, D. W., Eds., Academic Press, London, 1972, 363.

    49. Holme, T., Influence of environment on the content and composition of bacterial envelopes, in EnvironmentalControl of Cell Synthesis and Function, Dean, A. C. R., Pirt, S. J., and Tempest, D. W., Eds., Academic Press, 1972,391.

    50. Davies, H. C. and Rudd, J. H., Influence of environment on growth and cellular content of group A haemolyticstreptococci in continuous culture, in Environmental Control of Cell Synthesis and Function, Dean, A. C. R., Pirt,S. J., and Tempest, D. W., Eds., Academic Press, 1972, 401.

    51. Dawes, E. A. and Ribbons, D. W., Some aspects of the endogenous metabolism of bacteria, Bact. Rev., 28, 126,1964.

    52. van Uden, N., Kinetics of nutrient-limited growth, Annu. Rev. Microbiol., 23, 473, 1969.53. Walker, D. J. and Nader, C. J., Rumen microbial synthesis in relation to energy supply: diurnal variations with once

    daily feeding, Aust. J. Agric. Res., 21, 747, 1970.54. Tempest, D. W., Herbert, D., and Phipps, P. J., Studies on the growth of Aerobacter aerogenes at low dilution rates

    in a chemostat, in Microbial Physiology and Continuous Culture, Proceedings of the 3rd Int. Symp., Powell, E. O.,Evans, C. G. T., Strange, R. E., and Tempest, D. W., Eds., Her Majesty's Stationery Office, London, 1967, 240.

    55. Bhatla, M. N. and Gaudy, A. F., Jr., Sequential substrate removal in a dilute system by heterogeneous microbialpopulations, Appl. Microbiol., 13, 345, 1965.

    56. Mateles, R. I., Chian, S. K., and Silver, R., Continuous culture on mixed substrates, in Microbial Physiology andContinuous Culture, Proc. 3rd Int. Symp., Powell, E. O., Evans, C. G. T., Strange, R. E., and Tempest, D. W., Eds.,Her Majesty's Stationery Office, London, 1967, 233.

    57. Hobson, P. N., Physiological characteristics of rumen microbes and relation to diet and fermentation patterns, Proc.Nutr. Soc., 31, 135, 1972.

    58. Wood, T. M., Cellulose and cellulolysis, World Rev. Nutr. Diet., 12, 227, 1970.59. Fuisae, M. C. and Leatherwood, J. M., Regulation of cellulase from Ruminococcus, Can. J. Microbiol., 18, 347,

    1972.60. Dean, A. C. R., Oscillatory behaviour in the early stages of continuous culture, in Continuous Cultivation of

    Microorganisms, Mlek, I., Beran, K., Fencl, Z., Munk, V., iica, J., and Smrkov, H., Eds., Academic Press,London, 1969, 263.

    61. Gilley, J. W. and Bungay, H. R., Oscillatory growth rate response of Saccharomyces cerevisiae in continuousculture, Biotechnol. Bioeng., 9, 617, 1967.

    62. Sutton, J. D. and Johnson, V. W., Fermentation in the rumen of cows given rations containing hay and flaked maizeor rolled barley in widely different proportions, J. Agric. Sci., 73, 459, 1969.

    June 1974 187

    Dow

    nloa

    ded

    by [

    Cor

    nell

    Uni

    vers

    ity L

    ibra

    ry]

    at 2

    2:20

    13

    Nov

    embe

    r 20

    14

  • 63. Hobson, P. N. and Shaw, B. G., The anaerobic digestion of waste fiom an intensive pig unit, Water Res., 7, 437,1973.

    64. Hobson, P. N., Growth of mixed cultures and their biological control, in Symposia of the Society for GeneralMicrobiology, No. XIX, Meadow, P. M. and Pirt, S. J., Eds., Cambridge University Press, Cambridge, 1969, 43.

    65. Hobson, P. N., Rumen microorganisms, in Progress in Industrial Microbiology, 9, Hockenhull, D. J. D., Ed., J. andA. Churchill, London, 1971, 41.

    66. Allison, M. J., Biosynthesis of amino acid by ruminai microorganisms, J. Anim. Sci., 29, 797, 1969.67. Allison, M. J. and Robinson, I. M., Biosynthesis of -ketoglutarate by the reductive carboxylation of succinate in

    Bacteroides ruminicola, J. Bacteriol., 104, 50, 1970.68. Ierusalimsky, N. D., Bottle-necks in metabolism as growth-rate controlling factors, in Microbial Physiology and

    Continuous Culture, Proc. 3rd Int. Symp., Powell, E. O., Evans, C. G. T., Strange, R. E., and Tempest, D. W., Eds.,Her Majesty's Stationery Office, London, 1967, 23.

    69. lannotti, E. L., Kafkewitz, D., Wolin, M. J., and Bryant, M. P., Glucose fermentation products of Ruminococcusalbus grown in continuous culture with Vibrio succinogenes: changes caused by interspecies transfer of H2, J.Bacteriol., 114, 1231, 1973.

    70. Chung, K. T., An ecological significance of hydrogen utilisation in methanogenesis, Abstr. Annu. Gen. Meeting Am.Soc. Microbiol., P 64, 1972.

    71. Wolfe, R. S., Microbial formation of methane, Adv. Microb. Physiol., 6, 107, 1971.72. Bryant, M. P., Tzeng, S. F., Robinson, I. M., and Joyner, A. E., Jr., Nutrient requirements of methanogenic bacteria,

    in Adv. Chem. Ser., No. 105, American Chemical Society, 1971, 23.

    73. Bryant, M. P., Wolin, E. A., Wolin, M. J., and Wolfe, R. S., Methanobacillus omelianskii, a symbiotic association oftwo species of bacteria, Arch. Mikrobiol., 59, 20, 1967.

    74. Smith, P. H. and Hungate, R. E., Isolation and characterization of Methanobacterium ruminantium n. sp., J.Bacteriol, 75, 713, 1958.

    75. Paynter, M. J. B. and Hungate, R. E., Characterization of Methanobacterium mobilis sp. n. isolated from the bovinerumen, J. Bacteriol., 95, 1943, 1968.

    76. Oppermann, R. A., Nelson, W. O., and Brown, R. E., In vivo studies of methanogenesis in the bovine rumen:dissimilation of acetate, J. Gen. Microbiol., 25, 103, 1961.

    77. Oppeimann, R. A., Nelson, W. O., and Brown, R. E., In vitro studies on methanogenic rumen bacteria, J. Dairy Sci.,40, 779, 1957.

    78. Barker, H. A., Studies on the methane producing bacteria, Arch. Mikrobiol, 7, 420, 1936.79. Nelson, W. O., Oppermann, R. A., and Brown, R. E., In vitro studies on methanogenic rumen bacteria. II.

    Fermentation of butyric and valeric acids, J. Dairy Sci., 41, 545, 1958.80. Barker, H. A., Bacterial Fermentations, John Wiley and Sons, Inc., New York, 1956.81. Nottingham, P. M. and Hungate, R. E., Methanogenic fermentation of benzoate, J. Bacteriol., 98, 1170, 1969.82. Fina, L. R. and Fiskin, A. M., The anaerobic decomposition of benzoic acid during fermentation. II. Fate of carbons

    one and seven, Arch. Biochem. Biophys., 91, 163, 1960.83. Smith, P. H. and Mah, R. A., Kinetics of acetate metabolism during sludge digestion, Appl. Microbiol., 14, 368,

    1966.84. Thiel, P. G., The effect of methane analogues on methanogenesis in anaerobic digestion, Water Res., 3, 215, 1969.85. Smith, P. H., The microbial ecology of sludge methanogenesis, Dev. Ind. Microbiol, 7, 156, 1966.86. Heukelekian, H. and Heinemann, B., Studies on the methane-producing bacteria. I. Development of a method for

    enumeration, Sewage Works J., 11, 426, 1939.87. Smith, P. H., Pure culture studies of methanogenic bacteria, Proc. Ind. Waste Conf, 20, 583, 1965.88. McCarty, P. L., Kinetics of waste assimilation in anaerobic treatment, Developments in Industrial Biology, 7, 144,

    1966.89. Griffiths, I., The practice of sludge digestion, in Waste Treatment, Pergamon Press, Oxford, England, 1960.90. Buswell, A. M. and Hatfield, W. D., Anaerobic fermentations, Ill. State Waste Serv. Bull, 32, 59, 1936.91. McCarty, P. L. and McKinney, R. E., Volatile acid toxicity in anaerobic digestion, J. Water Pollut. Control Fed., 34,

    223, 1961.92. Andrews, J. F., Dynamic model of the anaerobic digestion process, J. Sanit. Eng. Div. Proc. Am. Soc. Civ. Eng., 95,

    95, 1969.93. Ichikawa, Y. and Kitamoto, Y., The inhibitory properties of propionate and related substances to the growth of E.

    coli, J. Agric. Chem. Soc. Jap., 41, 171, 1967.94. Mosey, F. E., Swanwick, J. D., and Hughes, D. A., Factors affecting the availability of heavy metals to inhibit

    anaerobic digestion, J. Inst. Water Pollut. Control, 6, 2, 1971.95. McCarty, P. L., Anaerobic waste treatment fundamentals, Part three, toxic materials and their control, Publ. Works,

    95, 91, 1964.96. Henderson, C, The effect of fatty acids on pure cultures of rumen bacteria, J. Agric. Sci., 81, 107, 1973.97. Prins, R. A., Van Nevel, C. J., and Demeyer, D. I., Pure culture studies of inhibitors for methanogenic bacteria,

    Antonie van Leeuwenhoek, 38, 281, 1972.98. Masselli, J. W., Maselli, N. W., and Burford, M. G., The occurrence of copper in water, sewage and sludge and its

    effect on sludge digestion, N. Engl. Interstate Water Pollut. Control Comm., 1961.

    188 CRC Critical Reviews in Environmental Control

    Dow

    nloa

    ded

    by [

    Cor

    nell

    Uni

    vers

    ity L

    ibra

    ry]

    at 2

    2:20

    13

    Nov

    embe

    r 20

    14

  • 99. Taiganides, E. P., Characteristics and treatment of wastes from a confinement hog production unit, Ph.D. thesis,Iowa State Univ. of Sei. and Technol., Ames, Iowa, 1963.

    100. McNary, R. R., Wolford, R. W., and Pulton, V. P., Experimental treatment of citrus waste water, Food Technol, 5,319, 1951.

    101. Buswell, A. M. and Mueller, H. F., Mechanism of methane fermentation, Ind. Eng. Chem., 44, 550, 1952.102. Environmental Control of Cell Synthesis and Function, Dean, A. C. R., Pirt, S. J., and Tempest, D. W., Eds.,

    Academic Press, London, 1972.103. Herbert, D., Elsworth, R., and Telling, R. C., The continuous culture of bacteria: a theoretical and experimental

    study, J. Gen. Microbiol., 14, 601, 1956.104. Brooks, J. D. and Meers, J. L., The effect of discontinuous methanol addition on the growth of a carbon-limited

    culture of Pseudomonas, J. Gen. Microbiol., 77, 513, 1973.105. McLaren, A. D. and Esterman, E. F., Influence of pH on the activity of chymotrypsin at a solid-liquid interface,

    Arch. Biochem. Biophys., 68, 157, 1957.106. Pirt, S. J., A kinetic study of the mode of growth of surface colonies of bacteria and fungi, J. Gen. Microbiol, 47,

    181, 1967.107. Herbert, D., in Recent Progress in Microbiology. VII. International Congr. for Microbiology, Stockholm.108. Stouthamer, A. H. and Bettenhaussen, C., Utilisation of energy for growth and maintenance in continuous and

    batch cultures of microorganisms, Biochem. Biophys. Acta, 301, 53, 1973.109. Nagai, S. and Aiba, S., Re-assessment of maintenance and energy uncoupling in the growth of Azotobacter

    vinelandii, J. Gen. Microbiol, 73, 531, 1972.110. Pirt, S. J., The maintenance energy of bacteria in growing cultures, Proc. R. Soc., Edinb. Sect. B. Biol., 163, 224,

    1965.111. King, W. R., Sinclair, C. G., and Topiwala, H. H., Effect of evaporation losses on experimental continuous culture

    results, J. Gen. Microbiol., 71, 87, 1972.112. Powell, E. O., The growth rate of microorganisms as a function of substrate concentration, in Microbial Physiology

    and Continuous Culture, Proc. 3rd Int. Symp., Powell, E. O., Evans, C. G. T., Strange, R. E., and Tempest, D. W.,Eds., Her Majesty's Stationery Office, London, 1967, 34.

    113. Powell, E. O., Criteria for the growth of contaminants and mutants in continuous culture, J. Gen. Microbiol., 18,259, 1958.

    114. Brunner, R., Oberzill, W., and Menzel, J., The continuous cultivation of Escherichia coli B and Serratia marcescensin common steady states, in Continuous Cultivation of Microorganisms, Mlek, I., Beran, K., Fencl, Z., Munk, V.,iia, J., and Smrkov, H., Eds., Academic Press, London, 1969, 323.

    115. Meers, J. L. and Tempest, D. W., The influence of extracellular products on the behaviour of mixed microbialpopulations in the magnesium-limited chemostat cultures, J. Gen. Microbiol., 52, 309, 1968.

    116. Powell, E. O., Hypertrophie growth, J. Appl. Chem. Biotechnol., 22, 71, 1972.117. Zines, D. O. and Rogers, P. L., The effect of ethanol on continuous culture stability, Biotechnol. Bioeng., 12, 561,

    1970.118. Lawrence, A. W., Application of process kinetics to design of anaerobic processes, in anaerobic biological treatment

    processes, Adv. Chem. Ser., 105, 23, 1971.119. O'Rourke, J. T., Kinetics of anaerobic treatment at reduced temperatures, Ph.D. thesis, Stanford University, Calif.,

    1968.120. Lawrence, O. W. and McCarty, P. L., Unified basis for biological treatment design and operation, J. Sanit. Eng. Div.,

    96, No. S.A.3, 1970.121. Garfinkel, D., A machine-independent language for the simulation of complex chemical and biochemical systems,

    Comput. Biomed. Res., 2, 31, 1968.122. Baldwin, R. L., Lucas, H. L., and Cabrera, R., Energetic relationships in the formation and utilisation of

    fermentation end products, in Physiology of Digestion and Metabolism in the Ruminant, Phillipson, A. T., Ed., OrielPress, Newcastle-upon-Tyne, 1969, 319.

    123. Hammer, M. J. and Borchardt, J. A., Dialysis separation of sewage sludge digestion, J. Sanit. Eng. Div., 95, 907,1969.

    124. Herbert, D., A theoretical analysis of continuous culture systems, in Continuous Culture, Monogr. No. 12, Soc.Chem. Ind., London, 1961, 21.

    125. Fencl, Z., A theoretical analysis of continuous culture systems, in Theoretical and Methodological Basis ofContinuous Culture of Microorganisms, Mdlek, I. and Fencl, Z., Eds., Academic Press, New York, 1966.

    126. Pirt, S. J. and Kurowski, W. N., An extension of the theory of the chemostat with feedback of organisms, itsexperimental realisation with a yeast culture, J. Gen. Microbiol, 63, 357, 1970.

    127. Hospodka, J., Industrial application of continuous fermentation, in Theoretical and Methodological Basis ofContinuous Culture of Microorganisms, Mlek, I. and Fencl, Z., Eds., Academic Press, New York, 1966.

    128. Royston, M. G., Tower fermentation of beer, Proc. Biochem., 1, 215, 1966.129. Young, J. C. and McCarty, P. L., The anaerobic filter for waste treatment, J. WaterPollut. Control Fed., 41, R160,

    1969.

    June 1974 189

    Dow

    nloa

    ded

    by [

    Cor

    nell

    Uni

    vers

    ity L

    ibra

    ry]

    at 2

    2:20

    13

    Nov

    embe

    r 20

    14

  • 130. Young, J. C. and McCarty, P. L., The anaerobic filter for waste treatment, Tech. Rep. No. 87, Dept. Civ. Eng.Stanford Univ., 1968.

    131. Hobson, P. N., The bacteriology of anaerobic sewage digestion, Proc. Biochem., 8, 19, 1973.132. Mah, R. A. and Susman, C., Microbiology of anaerobic sludge feimentation. I. Enumeration of the non-

    methanogenic anaerobic bacteria, Appl. Microbiol., 16, 358, 1968.133. Kirsch, E. J., Studies on the enumeration and isolation of obligately anaerobic bacteria from digesting sludge, Dev.

    Ind. Microbiol., 10, 170, 1969.134. Gaub, W. H., A bacteriological study of a sewage disposal plant, N.J. Agric. Expt. Sta. Bull, 394, 1, 1924.135. Buck, T. C, Keefer, C. E., and Hatch, H., Bacteriological studies of sludge digestion. I. A facultative aerobe isolated

    from digesting sludge, Sewage Ind. Wastes, 25, 993, 1953.136. Keefer, C. E., Buck, T. C., and Hatch, H., Bacteriological studies of sludge digestion. II. Optimum pH at which

    sludge digests when inoculated with Streptococcus diploidus, Sewage Ind. Wastes, 25, 1174, 1953.137. McKinney, R. E., Langley, H. E., and Tomlinson, H. D., Survival of Salmonella typhosa during anaerobic digestion.

    I. Experimental methods and high-rate digester studies, Sewage Ind. Wastes, 30, 1469,1958.138. Cookson, J. T. and Burbank, N. C., Isolation of anaerobic and facultative bacteria present in the digestion process, J.

    Water Pollut. Control Fed., 37, 822, 1965.139. Burbank, N. C., Jr., Cookson, J. T., Goeppner, J., and Brooman, D., Isolation and identification of anaerobic and

    facultative bacteria present in the (sewage sludge) digestion process, Air Water Pollut., 10, 327, 1966.140. Torien, D. F. and Siebert, M. L., A method for enumeration and cultivation of anaerobic acid forming bacteria

    present in digesting sludge, Water Res., 1, 397, 1967.141. Kistner, A., An improved method for viable counts of bacteria in the ovine rumen which ferment carbohydrates, J.

    Gen. Microbiol., 23, 565, 1960.142. Hungate, R. E., The anaerobic mesophilic cellulolytic bacteria, Bact. Rev., 14, 1, 1950.143. Hungate, R. E., Inter-relationships in the rumen microbiota, in Physiology of Digestion and Metabolism in the

    Ruminant, Phillipson, A. T., Ed., Oriel Press, Newcastle-upon-Tyne, 1969, 292.144. Torien, D. F., Siebert, M. L., and Hattingh, W. H. J., The bacterial nature of the acid forming phase of anaerobic

    digestion, Water Res., 1, 497, 1967.145. Hattingh, W. H. J., Kotz, J. P., Thiel, P. G., Torien, D. F., and Siebert, M. L., Biological changes during the

    adaptation of an anaerobic digester to a synthetic substrate, Water Res., 1, 255, 1967.146. Caldwell, D. R. and Bryant, M. P., Medium without rumen fluid for nonselective enumeration and isolation of

    rumen bacteria, Appl. MicrobioL, 14, 794, 1966.147. Sykes, R. M. and Kirsch, E. J., Fermentation patterns in the initial stages of sewage sludge digestion, Dev. Ind.

    MicrobioL, 11, 357, 1969.148. Sijpestein, A. K., On Ruminococcus flavifaciens a cellulose decomposing bacterium from rumen of sheep and cattle,

    J. Gen. Microbiol., 5, 869, 1951.149. Harkness, N., Bacteria in sewage treatment processes, J. Proc. Inst. Sewage Purif., Part 6, 542, 1966.150. Torien, D. F., Enrichment culture studies on aerobic and facultatively anaerobic bacteria found in anaerobic

    digesters, Water Res., 1, 147, 1967.151. Agardy, F. J., Cole, R. D., and Pearson, E. A., Enzyme activity as a parameter of digester performance, Proc. Ind.

    Waste Conf., Purdue Univ., Proc. 18th Ind. Waste Conf., Purdue Univ., p. 51, 1963.152. McCarty, P. L., Jeris, J. S., McKinney, R. E., Reed, K., and Vath, C. A., Microbiology of Anaerobic Digestion, Rept.

    No. R62-29, Sedgewick Lab. Sanit. Sci., Mass. Inst. Tech., Cambridge, Mass., 1962.153. Torien, D. F., Direct-isolational studies on the aerobic and facultative anaerobic bacterial flora of anaerobic digesters

    receiving raw sewage sludge, Water Res., 1, 55, 1966.154. Appleby, J. C., The isolation and classification of proteolytic bacteria from the rumen of the sheep, J. Gen.

    Microbiol., 12, 526, 1955.155. Blackburn, T. H. and Hobson, P. N., Further studies on the isolation of proteolytic bacteria from the sheep rumen,

    J. Gen. MicrobioL, 29, 69, 1962.156. Kotz, J. P., Thiel, P. G., Torien, D. F., Hattingh, W. H. J., and Siebert, M. L., A biological and chemical study of

    several anaerobic digesters, Water Res., 2, 198, 1968.157. Siebert, M. L. and Torien, D. F., The proteolytic bacteria present in the anaerobic digestion of raw sewage sludge,

    Water Res., 3, 241, 1969.158. Henderson, C., An improved method for enumerating and isolating lipolytic rumen bacteria, J. Appl. Bacterial., 36,

    187, 1973.159. Torien, D. F., Thiel, P. G., and Hattingh, W. H. J., Enumeration, isolation and identification of sulphate-reducing

    bacteria of anaerobic digestion, Water Res., 2, 505, 1968.160. Schnellen, C. G. T. P., Onderzoekingen over de methaangisting, Dissertation, Technische Hoogeschool, Delft,

    Holland, 1947.161. Mylroie, R. L. and Hungate, R. E., Experiments on the methane bacteria in sludge, Can. J. MicrobioL, 1, 55, 1954.162. Smith, P. H., Studies on the methanogenic bacteria of domestic sewage sludge, Bact. Proc., 1961, 60, 1961.163. Buraczewski, G., Methane fermentation of sewage sludge. I. The influence of physical and chemical factors on the

    development of methane bacteria and the course of fermentation, Acta Microbiol. Pol., 13, 321, 1964.

    190 CRC Critical Reviews in Environmental Control

    Dow

    nloa

    ded

    by [

    Cor

    nell

    Uni

    vers

    ity L

    ibra

    ry]

    at 2

    2:20

    13

    Nov

    embe

    r 20

    14

  • 164. Siebert, M. L. and Hattingh, W. H. J., Estimation of methane-producing bacterial numbers by the most probablenumber (MPN) technique, Water Res., 1,13, 1967.

    165. Thiel, P. G., Torien, D. F., Hattingh, W. H. J., Kotz, J. P., and Siebert, M. L., Inter-relations between biological andchemical characteristics in anaerobic digestion, Water Res., 2, 391, 1968.

    166. Bousfield, S., Hobson, P. N., and Summers, R., Laboratory and small pilot-plant scale anaerobic digestion of piggerywastes: some aspects of the microbiology of anaerobic piggery waste digestion, Proc. Farm Waste Disposal Conf.Glasgow, Agricultural Research Council, London, England, 1972.

    167. Bousfield, S., Hobson, P. N., and Summers, R., Where there's muck there's gas, Pig Fanning, 21, 79, 1973.168. Buswell, A. M., Biological Treatment of Sewage and Industrial Wastes, Vol. II, McCabe, J. and Eckenfelder, W. W.,

    Eds., Reinhold Publishing Company, New York, 1958.169. Babbit, H. E. and Baumann, E. R,, Sewerage and Sewage Treatment, J. Wiley and Sons Inc., New York, 1953.170. Townend, C. B., Proc. Inst. Civil Eng., 21,761,1962.171. Langford, L. L., Mesophilic anaerobic digestion, design and operating considerations, in Biological Treatment of

    Sewage and Industrial Wastes, Vol. II, McCabe, J. and Eckenfelder, W. W., Eds., Reinhold Publishing Company, NewYork, 1958.

    172. Pohland, F . G., Thermal energy interchange during anaerobic methane fermentation of waste organic substrates,Appl. Microhiol., 16, 1518, 1968.

    173. Sawyer, C. N., An evaluation of high-rate digestion, in Biological Treatment of Sewage and Industrial Wastes, Vol.II, McCabe, J. and Eckenfelder, W. W., Eds., Reinhold Publishing Company, New York, 1953.

    174. Torpey, W. N., Loading to failure of a pilot high-rate digester, Sewage Ind. Wastes, 27, 121, 1955.

    175. Storch, B., Proc. Inst. Civil Eng., Vol. 24, 135, 1963.176. Morgan, P. F. and Neuspiel, P. J., Environmental control of anaerobic digestion with gas diffusion, in Biological

    Treatment of Sewage and Industrial Wastes, Vol. II, McCabe, J. and Eckenfelder, W. W., Eds., Reinhold PublishingCompany, New York, 1953.

    177. Morgan, P. F., Studies of accelerated digestion of sewage sludge, Sewage Ind. Wastes, 26, 462, 1954.178. Torpey, W. N., High rate digestion of concentrated primary and activated sludge, Sewage Ind. Wastes, 26, 479,

    1954.179. Sawyer, C. N. and Roy, H. K., A laboratory evaluation of high-rate sludge digestion, Sewage Ind. Wastes, 27, 1356,

    1955.180. Steffen, A. J.( Treatment of packing house wastes by anaerobic digestion, in Biological Treatment of Sewage and

    Industrial Wastes, Vol. II, McCabe, J. and Eckenfelder, W. W., Eds., Reinhold Publishing Company, New York,1953.

    181. Torpey, W. N. and Melbinger, N. R., Reduction of digested sludge volume by controlling recirculation, J. WaterPollut. Control Fed., 39, 1464, 1967.

    182. Schroepfer, G. J., Fullen, W. J., Johnson, A. S., Ziemke, N. R., and Anderson, J. J., The anaerobic contact processas applied to packing house wastes, Sewage Ind. Wastes, 27, 4, 1955.

    183. Stander, G. J., Cillie, G. C , Ross, W. R., and Baillie, R. D., Treatment of wine distillery wastes by anaerobicdigestion, Proc. 22nd Ind. Waste Conf., Purdue Univ., 1967.

    184. Stander, G. J. and Snyder, R., Effluents from fermentation industries, part V., Proc. Inst. Sewage Purif., part 4, 447,458, 1950.

    185. Coulter, J. B. and Ettinget, M. B., A discussion of a paper by Fall, E. B. and Kraus, L. L., The anaerobic contactprocess in practice, J. Water Pollut. Control Fed., 33, 1038, 1961.

    186. Lovan, C. R. and Foree, E. G., The anaerobic filter for the treatment of brewery press liquor waste, Brew. Dig., 47,66, 1972.

    187. Hovious, J. C., Conway, R. A., and Ganze, C. W., Anaerobic lagoon for pretreatment of petro-chemical wastes, J.Water Pollut. Control Fed., 45, 71, 1973.

    188. Winsor, C. E., Operating experience at Minworth works, J. Water Pollut. Control Fed., 45, 560, 1973.189. Roberts, J. M., Combined treatment of poultry and domestic wastes, Sewage Ind. Wastes, 30, 1186, 1958.190. Grune, W. N., Bartholomew, D. D., and Hudson, C. I., Effects of radioactive materials on anaerobic digestion,

    Sewage Ind. Wastes, 30, 1123, 1958.191. Gramms, L. C., Polkowski, L. B., and Witzel, S. A., Trans. Am. Soc. Agric. Eng., 1971, 7.192. Sparling, A. B., Energy recovery from livestock wastes, in Proc. Intl. Biomass Energy Conf., Biomass Energy Inst.

    Inc., Winnipeg, Canada, 1973.193. Pfeffer, J., Processing organic solids by anaerobic fermentation, in Proc. Intl. Biomass Energy Conf, Biomass Energy

    Inst. Inc., Winnipeg, Canada, 1973.194. Ross, W. E., Dual disposal of garbage and sewage sludge at Richmond, Indiana, Sewage Ind. Wastes, 26, 140, 1965.195. Hobson, P. N., Rumen microorganisms, Proc. Biochem., 4, 53, 1969.196. Chung, P., Production of methane gas from manure, in Proc. Intl. Biomass Energy Conf, Biomass Energy Inst. Inc.,

    Winnipeg, Canada, 1973.197. Laura, R. D. and Idnani, M. A., Increased production of biogas from cow dung by adding other agricultural waste

    materials, J. Sci. Food Agric., 22, 164, 1971.

    June 1974 191

    Dow

    nloa

    ded

    by [

    Cor

    nell

    Uni

    vers

    ity L

    ibra

    ry]

    at 2

    2:20

    13

    Nov

    embe

    r 20

    14

Recommended

View more >