316 Plant Disease / Vol. 81 No. 4

John C. Sutton, De-Wei Li, Gang Peng, Hai Yu, and Pinggao ZhangUniversity of Guelph, Guelph, Canada

R. M. Valdebenito-SanhuezaEMBRAPA/CNPUV, Bento Gonçalves, Brazil

Antagonistic microorganisms are an im-portant focus of current attempts to im-prove control of gray mold and other dis-eases caused by the pernicious pathogenBotrytis cinerea Pers.:Fr. in fruits, vegeta-bles, field crops, tree seedlings, ornamentalplants, and other crops (56). Microbialantagonists of B. cinerea have potential tohelp counter inadequacies, periodic fail-ures, and concerns associated with presentcontrol practices, which generally involvecultural and sanitation measures, regulationof the microclimate, and heavy dependencyon fungicides. Frequent resistance of B.cinerea to fungicides and inadequate hostresistance (14) contribute to precariousvulnerability of many crops to the patho-gen. Biological control by microorganismsalso is seen as a means to avoid any risksassociated with occupational exposure ofworkers to fungicides and with fungicideresidues in harvested crops and in the envi-ronment. The filamentous fungus Gliocla-dium roseum Bainier has emerged as aneffective and versatile antagonist of B.cinerea that, with other outstanding agents,promises to improve disease managementin crops. This article attempts to provide abroad overview of G. roseum and insightsinto the roles of this remarkable fungus innatural systems, in crops, and as a biologi-cal control agent.

Morphology and TaxonomyG. roseum is an unusual hyphomycete

that produces one-celled conidia on twodistinct types of conidiophores, one peni-cillately branched and the other verticil-

lately branched (Fig. 1A) (13,44). Coloniesof the fungus on agar media are generallywhitish, orange, or salmon (Fig. 1B). G.roseum is the anamorph of Nectria ochro-leuca (Schwein.) Berk, an ascomycete ofthe Hypocreales (45). The fungus is con-sidered generically distinct from the bio-logical control agent Gliocladium virensJ.H. Miller, J.E. Giddens, & A.A. Foster,now referred to as Trichoderma virens, andfrom the type species Gliocladium penicil-lioides. Domsch et al. (13) proposed thatG. roseum belongs in the genusClonostachys and suggested that it shouldbe regarded as a collective species.

EcologyG. roseum is common in an extraordi-

nary range of habitats in tropical, temper-ate, subarctic, and desert regions of theworld. It was reported in cultivated, grass-land, woodland, forest, heathland, fresh-water, and coastal soils, particularly thoseof neutral to alkaline pH (4,13,15,43). G.roseum is frequently associated with cystsof Heterodera spp., Globodera spp., andother nematodes in soil (6,18,24,26,36) andwith sclerotia of Sclerotinia sclerotiorum,Phymatotrichum omnivorum, Rhizoctoniasolani, Botrytis spp., Verticillium spp., andother fungi in soil and plant materials(25,51,61,62). Over the past 35 years, G.roseum has gained a distinctive reputationas a mycoparasite of a broad spectrum offungi (3,30,35,37,59).

Equally remarkable are diverse associa-tions of G. roseum with roots, stems,leaves, fruits, and seeds of plants. Thefungus occurs on plant surfaces, such asthe phylloplane of strawberry (31) and themycorrhizoplane of European silver fir(Abies alba) (33). Reports point to anabundance of G. roseum within senescentand dead roots and foliage of a wide vari-ety of plants and in plants weakened by

stress factors such as herbicides and dis-ease (5,9,11,12,13,16). Intriguingly, G.roseum is known to colonize, withoutsymptom production, apparently healthyroots, stems, pods, and seeds of soybean(34,46), roots of red clover (50), and leavesof strawberry (55) and raspberry (63). Inthese instances, the fungus apparentlycolonized the host as a nonpathogenicparasite, and in soybean at least, the asso-ciation is systemic. Reported evidence thatG. roseum is pathogenic to apple fruits,potato tubers, conifer seedlings, Exacumaffine, and faba beans is considered incon-clusive by the present authors (1,2,17,20,21,58). The teleomorph of G. roseum,N. ochroleuca, is found most often onbranches of recently dead trees, but also onherbaceous and fleshy tissues of plants andfungi (19,44).

Collectively, the aforementioned reportsunderscored the cosmopolitan nature andextraordinary ecological versatility of G.roseum. Given this perspective, it is notsurprising that the fungus has appearedrepeatedly in assemblages of microbesrecovered from living and dead plant tis-sues for evaluation in biocontrol testsagainst B. cinerea and other pathogens(28,29,40,63).

Socioeconomic Expectationsof Biological Control

Development of antagonistic organismsas biological control agents is a demandingtask given socioeconomic expectations thatbiocontrol should be efficient, dependable,cost-effective, and safe for humans, thecrop, and the environment (53,54). Likeother methods of disease management,biological control ideally aims to suppressdisease enough that yield losses are mini-mized and crop quality is maintained at anacceptable level. To ascertain whetherbiological control agents meet the expecta-

Dr. Sutton’s address is: Department of Environ-mental Biology, University of Guelph, Guelph,Ontario, N1G 2W1, Canada;E-mail: jsutton@evbhort.uoguelph.ca

Publication no. D-1996-1206-05F© 1997 The American Phytopathological Society

C@F:C8;@LD�IFJ<LD

A Versatile Adversary of�FKIPK@J�:@E<I<8 in Crops

Plant Disease / April 1997 317

tions, we generally must evaluate themunder epidemiological conditions similarto those of well-managed crops. Antago-nists introduced to suppress B. cinereamight have to interact appropriately withthe crop plants, pathogen, and other micro-organisms under the prevailing microcli-matic conditions, and in face of fertilizerand pesticide applications, tillage opera-tions, irrigation, pruning, and other“human interferences” (27,56). Biocontrolsystems of B. cinerea are highly dynamic,embracing growth and development of thehost, infection cycles and serial dispersalsof the pathogen, quantitative shifts in thebiocontrol agent and indigenous organisms,and microclimatic fluctuations over timeand in space.

Selection of G. roseumThe task of selecting an organism as a

biological control agent is formidablegiven the staggering numbers of microbialisolates that can justify evaluation, theneed to optimize effectiveness of biologi-

Fig. 1. Gliocladium roseum. (A) Verticil-late and penicillate conidiophores andconidia. (B) Colony on potato-dextroseagar medium.

Fig. 2. Signs and symptoms of gray mold caused by Botrytis cinerea. (A) Lesionsbearing conidiophores and conidia of B. cinerea on stems and needles of blackspruce seedlings. (B) Lesions on stem and leaves of Exacum affine bearing conidio-phores and conidia of B. cinerea. (C) Lesion arising from infection of a deleafingwound in a tomato stem. (D) Collapse of tomato plants affected by stem lesions. (E)Petal flecking in cyclamen. (F) Lesion produced in a cyclamen leaf invaded by B. cine-rea from an adhering petal. (G) Gray mold of raspberry fruits.

318 Plant Disease / Vol. 81 No. 4

cal control, and investments normally re-quired to develop a microbe for commer-cial use. Whether a truly outstandingorganism has been found and whether asuperior microbe will turn up are lingeringconcerns of biocontrol investigators.

In selecting organisms against B. cine-rea, we subscribed to the rationale thatecological adaptation of the microbe to thehost plant is advantageous for effective andsustained biological control in the croppingsystem, and accordingly used host plants aschief sources of test organisms. Mycelialfungi, yeasts, and bacteria were isolatedfrom living and dead leaves, stems, flow-ers, and fruits of pesticide-free host plants.The organisms included saprotrophs, weakpathogens, and nonpathogenic epiphytesand endophytes. Greenhouse plants, whichlacked desirable diversity of microflora,were sprayed with aqueous suspensions ofsoils and incubated in high humidity priorto recovery of test isolates. Host plants, ineffect, functioned as selective substratesand substrata of indigenous organisms,thus narrowing the field of organisms forevaluation in screening tests. For somehosts, this field was narrowed further byapplying conidia of B. cinerea to plantsand later recovering organisms only fromportions of tissue on which the pathogenfailed to sporulate given favorable condi-tions. G. roseum was isolated repeatedlyfrom living and dead tissues of almost allhosts investigated.

Isolates of G. roseum were among hun-dreds of organisms tested against B. cine-rea in each host. Sequences of tests de-signed to increasingly mimic the epidem-iological conditions of commercial cropswere conducted using detached plant partsin the laboratory, pot-grown plants in thegrowth room and greenhouse, and whenappropriate, plants in field plots. Standardfungicide treatments served as yardsticks

of biocontrol effectiveness. As it turnedout, isolates of G. roseum consistentlyranked among organisms of outstandingeffectiveness against B. cinerea in straw-berry, raspberry, black spruce, and a rangeof flowers and vegetables produced ingreenhouses, including begonia, geranium,cyclamen, poinsettia, Exacum affine, to-mato, cucumber, and pepper. In almost allinstances, G. roseum was as or moreeffective than fungicide treatments inleaves, bracts, stems, flowers, and fruits. Aperspective of biocontrol testing against B.cinerea is now given as background tofurther consideration of biocontrol per-formance of G. roseum.

Customizing Biocontrol TestsBecause relationships of B. cinerea and

various hosts differ widely in respect topathways and conditions of infection, kindsof tissues affected, host age, inoculumconcentration, and other factors, testingmethods normally are best customized foreach host. Leaves and stems require pro-tection in some hosts to suppress graymold on the foliage, as in conifer seedlings(Fig. 2A) and the flowering pot plant Ex-acum affine (Fig. 2B). In hosts such asstrawberry, leaves are protected to controlsymptomless infection and spore produc-tion of B. cinerea when the tissues die(55,57,66). In greenhouse-grown tomato, itis important to protect wounds made instems when lower leaves are removed tofacilitate ventilation and to remove inocu-lum sources of B. cinerea (23). Thedeleafing wounds are prime sites of entryof the pathogen, from which lesions candevelop and lead to collapse of the plants(Fig. 2C and D). Similarly, wounds madein mother and daughter plants of geraniumwhen cuttings are prepared require protec-tion against gray mold. Flowers requireprotection for a variety of reasons. Protec-tion against petal flecking is of prime im-portance in flowering ornamentals likecyclamen (Fig. 2E), gerbera, and rose.Flowers of tomato, cyclamen, and gera-nium should be protected in part becausecolonized petals, when fallen, are utilizedas a food base by the pathogen to aggres-sively attack leaves (Fig. 2F), stems, caly-ces, and fruits (23). Protection of straw-berry and raspberry flowers blocksinvasion of fruits (Fig. 2G) by mycelium ofB. cinerea in flower tissues (52,56). Prin-cipal effects of temperature, humid peri-ods, inoculum concentration of the patho-gen, and other key variables on the host–pathogen interaction frequently can beaccounted for in the design of biocontroltests.

In a majority of tests developed in ourlaboratory, the pathogen and test organismswere each applied once to host tissues,usually at different times, and control wasestimated in terms of signs or symptoms ofdisease. Conidiophore production of B.

cinerea provided an indirect but convenientmeans for quantifying infection and colo-nization of host tissues by the fungus (40).Tissues of leaves, petals, and stem seg-ments used as test materials or removedfrom treated plants usually were incubatedin high humidity or on an agar mediumcontaining paraquat and chloramphenicol(PCA; Fig. 3A and B). By killing the tis-sues, this procedure circumvented prob-lems associated with long and variableperiods of quiescence or latency of B. cine-rea and enabled the pathogen to sporulaterapidly so that tests were completed expe-ditiously and with improved reproducibil-ity. Monocyclic tests were employed whenthe objective was to control portions ofepidemics involving a single infectioncycle of B. cinerea, but polycyclic testsinvolving serial infection cycles of thepathogen often were needed for evalua-tions under conditions similar to those ofcommercial crops. Because of complexity,logistics, and space requirements, polycy-clic tests usually were reserved for leadingbiocontrol candidates like G. roseum.

Biological Control of B. cinereaStrawberry. G. roseum first emerged as

a powerful antagonist of B. cinerea instudies in strawberry in the late 1980s(39,40). Isolates of the antagonist consis-tently suppressed B. cinerea by more than98% in assays on attached and detachedleaves, petals, and stamens, and invariablyperformed as well as or better than didother leading organisms, such as Tricho-derma viride, Alternaria alternata, My-rothecium verrucaria, and Penicilliumspp., and a standard fungicide (captan).Tissues used in assays usually were inocu-lated with test organisms (107 conidia perml of water plus surfactant for fungi),challenged after 24 h with B. cinerea (106

conidia per ml), and later incubated onPCA and assessed for sporulation of thepathogen.

The most realistic evaluations, however,were conducted in field plots established inmatted-row strawberry plantings (40,53,54).Test fungi (106 conidia per ml) or captanwas applied weekly to protect the flowers,the chief pathway to fruit invasion by B.cinerea in strawberry (52). Treatmentswere applied shortly before nightfall anddew onset, based on conjecture that dewand darkness could facilitate survival andactivity of the biocontrol organisms. Infour tests in cv. Redcoat, G. roseum sup-pressed incidence of B. cinerea in stamensand fruits by 79 to 93% and 48 to 76%,respectively, a performance equal to orbetter than that of other leading antagonistsand the fungicide. G. roseum also was aseffective as captan against B. cinerea infruits of eight strawberry cultivars that didnot differentially affect biocontrol by theantagonist. As in commercial plantings, B.cinerea generally was recalcitrant and

Fig. 3. Sporulation of Gliocladiumroseum on tissues incubated on an agarmedium containing paraquat. (A) Leafdisk. (B) Stamens.

Plant Disease / April 1997 319

sporulated on 15 to 30% or more of fruitswhether the plants were treated with fungi-cide or G. roseum. A plausible explanationwas that flowers of the tested cultivarsremained open for only a few days, so thatmany would have escaped weekly treat-ments or were treated too late for effectivecontrol. In addition, the tests were strin-gent, in that plots were artificially inocu-lated with B. cinerea, and sporulation inci-dence of the pathogen in harvested fruitswas estimated only after several days inroom temperature and high humidity. Im-proved methods for timing and applyingtreatments may well improve biocontrol inthe field.

Weekly applications of G. roseum (106

conidia per ml) gave good control of B.cinerea in Chandler strawberries grown ina plastic-covered production greenhouse inthe extremely humid climate of the SerraGaucha mountains of southern Brazil (60).Fruit losses at harvest and 4 to 5 days afterharvest, averaged over 8 weeks, were re-duced by 73 and 48%, respectively, instrawberries treated with G. roseum com-pared with 64 and 36% by a weekly fungi-cide program.

Besides flowers and fruits , G . roseumproved remarkably effective against B.cinerea in strawberry foliage, the chiefinoculum source of the pathogen in straw-berry plantings (52,55,57). In six field testsin Ontario, variously conducted in spring,late summer, and early fall, foliage wasinoculated with B. cinerea (105 to 106

conidia per ml) and 2 to 5 weeks later wassprayed with inoculum of G. roseum orother biocontrol fungi (107 conidia per ml)or with chlorothalonil, a fungicide of ex-ceptional effectiveness against B. cinereain strawberry leaves (52). In most in-stances, G. roseum suppressed spore pro-duction of B. cinerea on the leaves by 90 to100% and was always as effective aschlorothalonil, whereas Penicillium sp. andT. viride matched the performance ofchlorothalonil in only three of the six tests.Both G. roseum and the fungicide sup-pressed the pathogen by only about 60% insemisenescent overwintered leaves.

Raspberry. Isolates of G. roseum gen-erally equaled or outperformed other mi-crobes and captan when tested against B.cinerea in raspberry field plots in Ontario.Strong suppression normally requiredtreatment of flowers and of fruits, pre-sumably because B. cinerea is able to in-fect drupelets directly as well as by inva-sion from flower tissues (8). In tests ofleading biocontrol candidates, flowerclusters of seven summer-bearing rasp-berry cultivars were treated with B. cinerea(5 × 103 conidia per ml) only, or with thepathogen in combination with G. roseum,T. viride, or Trichothecium roseum (all at106 conidia per ml). Disease severity wasestimated after each harvested flower orfruit was incubated in high humidity at 20to 22°C for 10 days. Observations for the

cultivars, combined in the absence of inter-active effects, indicated that the antago-nists and captan suppressed B. cinerea inflowers and ripe fruits by 57 to 73% and 37to 61%, respectively. In subsequent tests,conducted in both summer- and fall-bear-ing cultivars, G. roseum suppressed fruit

rot by 27 to 54% when applied once toflower clusters, and by 48 to 61% whenclusters were treated at flowering andagain when fruits were developing. Re-spective values for captan-treated raspber-ries were 18 to 25% and 23 to 40%. Underconditions of protracted flowering and

Fig. 4. Biological control of Botrytis cinerea in various hosts by Gliocladium roseumand Trichoderma koningii. (A) Profuse sporulation of B. cinerea in raspberry shootsinoculated with the pathogen only (left) compared with sporulation chiefly of G.roseum in shoots inoculated with the pathogen and the antagonist. (B) Suppressionof gray mold symptoms in black spruce seedlings by a fungicide (chlorothalonil) andvarious inoculum concentrations of G. roseum: (left to right) seedlings treated withfungicide, 106 conidia per ml, 108 conidia per ml, 104 conidia per ml, and water. (C) (leftto right) Begonia plants inoculated with B. cinerea (106 conidia per ml) and treatedwith water, T. koningii, G. roseum, and chlorothalonil. (D) (left to right) Cyclamenplants inoculated with B. cinerea (106 conidia per ml) and treated with water, G.roseum, T. koningii, and chlorothalonil.

320 Plant Disease / Vol. 81 No. 4

fruiting, more than two treatments may beneeded to optimize fruit rot control inraspberry crops. G. roseum was found alsoto be a powerful antagonist of B. cinerea inleaves and canes of raspberry (Fig. 4A;63).

Conifer seedlings. G. roseum, althoughnot among several hundred microbes re-covered from conifer trees in northernOntario, was evaluated against B. cinereain black spruce seedlings because of previ-ous strong performance in strawberry (65).In growth room tests, isolates of G. roseumsuppressed B. cinerea almost completelyand outperformed 132 microbial isolatesfrom conifers; only an isolate of T. virideand one of Trichothecium sp. matched theeffectiveness of G. roseum (65). In green-house studies, treatment programs wereinitiated to coincide with closure of seed-ling canopies, when gray mold epidemicsusually begin. Four applications of G.roseum (106 or 108 conidia per ml) at 2- to4-week intervals suppressed spore produc-tion of B. cinerea and gray mold symptomson seedlings as or more effectively thandid chlorothalonil treatments applied at thesame times (Fig. 4B). In subsequent green-house tests, programs of G. roseum re-duced incidence of killed shoots by 63 to81% compared with reductions of 44 to55% for chlorothalonil programs. A singleapplication of G. roseum (106 conidia perml) at canopy closure was found to sup-press the pathogen as effectively as didprograms of two to six applications of the

antagonist, or six of chlorothalonil, all at 1-to 2-week intervals starting when thecanopies closed (67). Biological controlactivity of G. roseum was extraordinarilypersistent, lasting 8 to 12 weeks or more.

G. roseum also gave promising results inred pine seedlings (66). Inoculum applica-tions of the antagonist (108 conidia per ml)suppressed B. cinerea by 88% in first-yearseedlings inoculated with the pathogen(106 conidia per ml). The antagonist alsowas strongly therapeutic against gray moldin second- and third-year seedlings grownin the presence of natural inoculum of B.cinerea in greenhouses and outdoor nurs-ery compounds.

Greenhouse flowers and vegetables.The strong performance of G. roseumagainst B. cinerea in hosts as diverse asstrawberry and conifers fueled speculationthat the antagonist also might be effectiveagainst the pathogen in the multitude ofsuscepts grown in greenhouses. For com-parative studies, microbial isolates associ-ated with greenhouse crops were obtainedfrom foliage and flowers of plants withmicroflora enriched by previous treatmentwith suspensions of natural soils. Isolatesof G. roseum were obtained frequentlyfrom each of 11 kinds of greenhouse plantsemployed in biocontrol investigations andevaluated on the same hosts from whichthey were recovered.

G. roseum suppressed B. cinerea by 94to 100% in leaf disk assays of begonia,cyclamen, and geranium, and in stem-piece

assays of cucumber, pepper, and tomato.No tested microbe was superior to G.roseum, and few microbial isolatesmatched the performance of G. roseumisolates, except in cyclamen (Table 1).Isolates of Trichoderma spp. were amongthose as effective as G. roseum in somehosts, including one identified as T. kon-ingii in begonia and another in cyclamen,and three of T. harzianum in geranium.Among isolates of Stilbella, a genus rarelyif ever mentioned in biocontrol literature,one of S. aciculosa performed well in cy-clamen, cucumber, and pepper. A diversityof microbes was highly effective in cycla-men, an unusual circumstance in our expe-rience. Although isolates of G. roseuminvariably gave good protection of leavesand stems against conidia of B. cinerea,they were less effective against petal in-oculum. Petals colonized by B. cinerea andadhering to plant foliage are an importantand formidable form of inoculum of thepathogen in cyclamen, geranium, petunia,tomato, and other hosts (23,49). In ourexperience, biological control is achievedmore easily by treating flowers againstconidia of B. cinerea than by treating foli-age against mycelium of the pathogen indiseased petals. In cyclamen, for example,treatment of flowers with G. roseum priorto inoculation with B. cinerea and subse-quent use of the petals as inoculum onfoliage suppressed the pathogen in theleaves by more than 90%; in contrast, only40 to 50% control was attained in leaves

Table 1. Proportion of tested isolates of various microbes that suppressed Botrytis cinerea as effectively as did isolates of Gliocladium roseum inleaf disk assays of begonia, cyclamen, and geranium, and in stem-piece assays in cucumber, pepper, and tomatoa

Proportion of effective isolates

Microbes Begonia Cyclamen Geranium Cucumber Pepper Tomato

Acremonium sp. 0/3 … 0/1 0/2 0/2 0/2Alternaria sp. 1/10 3/4 0/5 0/3 0/2 0/2Aspergillus sp. 0/1 … 0/3 0/2 0/2 0/2Bacteria (unidentified) 0/10 0/2 0/2 0/3 0/4 0/3Botryosporium longibractiatum … … 0/1 … … …Cladosporium sp. 0/6 … 0/4 0/1 0/1 …Colletotrichum gloeosporioides 1/1 0/1 … … … …Epicoccum purpurascens 1/4 2/3 0/3 0/2 0/2 0/2Fusarium sp. 0/4 1/3 0/1 0/1 0/1 0/2Gliocephalis sp. 0/1 … … … … …Myrothecium sp. 0/3 … … … … …Paecilomyces sp. … 1/1 0/1 0/2 0/2 0/2Penicillium sp. 0/15 1/1 0/5 0/4 0/6 0/5Phialophora sp. … 0/1 … … … …Phoma sp. 0/1 3/6 … 1/2 0/2 0/1Physalospora sp. 0/1 1/1 0/1 0/1 0/1 0/2Pithomyces sp. 0/2 … … 0/1 0/1 …Stilbella sp. … 1.1 … 1/1 1/1 0/1Trichoderma sp. 6/15 1/3 3/8 0/5 0/4 0/5Verticillium sp. 0/1 … … … … …Yeast (unidentified) 1/6 0/2 0/4 0/3 0/2 0/3Mycelial fungi (unidentified) 0/2 … 0/2 0/3 0/2 0/2Totals 10/86 14/29 3/41 2/36 0/35 0/33

a Inoculum concentrations were: 106 conidia per ml for B. cinerea, Alternaria spp., and Epicoccum purpurascens; 107 cells per ml for yeasts; and108 CFU/ml for bacteria. Number of G. roseum isolates tested in begonia, cyclamen, geranium, cucumber, pepper, and tomato were four, four,four, three, two, and three, respectively. These isolates suppressed B. cinerea in the respective hosts by 100, 98, 100, 100, 100, and 100%. Noneof the other microbial isolates was more effective than G. roseum.

Plant Disease / April 1997 321

treated with G. roseum and inoculated withcolonized petals.

Performance of G. roseum against B.cinerea varied considerably in flowers ofornamental plants in the greenhouse, and inseveral instances was inferior to that of thebest isolates of Trichoderma spp. and tostandard fungicides (Fig. 4C and D). Forexample, G. roseum suppressed B. cinereaby 49 and 68%, respectively, in fullyopened and senescent flowers of begonia(Begonia × hiemalis) compared with 95and 93%, respectively, for an isolate of T.koningii, and 85 and 95% for chlorothalo-nil. Substantial differences were found inbiocontrol effectiveness among isolates ofG. roseum from begonia and seven otherhosts when tested against B. cinerea inpetals but not in leaves of begonia. In cy-clamen petals, G. roseum and an isolate ofT. koningii suppressed the pathogen by 75and 90%, respectively. In geranium petals,G. roseum and T. harzianum were 32 to44% and 44 to 86% effective, respectively.Petal flecking caused by B. cinerea in ger-bera was reduced by >90% by G. roseumand by T. harzianum, but neither antagonisteffectively controlled petal blight. Undersome conditions, however, G. roseum sup-pressed the pathogen in petals of varioushosts by as much as 85 to 95%. Flowerage, pollen, temperature, humid periods,inoculum concentration, and method ofinoculum application are among the manyvariables that substantially influence bio-control by G. roseum and other biocontrolagents in flowers and should be consideredin a biocontrol program.

B. cinerea often utilizes fallen tomatopetals as food bases from which it invades

green or ripening tomato fruits, either di-rectly or by first attacking calyces attachedto the fruit (23). When petals were inocu-lated with the pathogen (5 × 105 conidiaper ml) and positioned near calyces onattached green fruits, almost all calycesbecame blighted and 95% of the fruitsdeveloped rot (Fig. 5A and B). Calyxblight and fruit rot were controlled almostcompletely, however, when the petals weretreated with G. roseum a few hours prior toinoculation with the pathogen (Fig. 5C andD).

Plant wounds are another theater of ag-gressive biological control by G. roseum.The antagonist strongly suppressed B. cin-erea in cuttings and mother plants of gera-nium, in wounds on tomato stems whenlower leaves are removed, and in sunscorch wounds on leaves of Exacum affine.In geranium, dipping fresh cuttings bearingsnapping and deleafing wounds into in-oculum of G. roseum (107 conidia per ml)controlled gray mold almost completelyand as or more effectively than did iprodi-one (Fig. 6A and B). Cuttings treated withG. roseum developed a profusion ofhealthy roots and were more vigorous thanwere the fungicide-treated cuttings. Incontrast, cuttings inoculated with B. cine-rea developed large lesions, and about one-quarter of them quickly died. In tomato,application of G. roseum (107 conidia perml) to fresh deleafing wounds protectedthe stems for several months after applica-tion. B. cinerea was recovered from 70%of wound sites 3 months after deleafingwhen fresh wounds were inoculated withthe pathogen (5 × 105 conidia per ml) only,but from less than 2% of wounds treated

with G. roseum 24 h before inoculationwith the pathogen. G. roseum was extraor-dinarily persistent in the living stems andwas recovered from more than 98% ofwound sites 3 months after inoculation.Sun scorch wounds in Exacum leaves aremajor sites of invasion by B. cinerea (Fig.7A). Treatment of artificial scorch woundsmade by momentary use of a flame and ofnatural sun scorch wounds using G. roseum(5 × 107 conidia per ml) reduced gray moldseverity induced by B. cinerea (105 conidiaper ml) by 97 to 100% (Fig. 7B).

Variables InfluencingEffectiveness of G. roseum

An understanding of interactions amongG. roseum, B. cinerea, host plants, micro-climatic factors, and other components ofthe biological control systems is funda-mental for optimizing suppression of thepathogen in crops. Highlights of what isknown of these interactions are now con-sidered.

Development stage of host organs.Development of leaves from the newlyexpanded stage to early senescence gener-ally had little effect on performance of G.roseum against B. cinerea in foliage ofstrawberry, begonia, and cyclamen. Ex-panding leaves of begonia and cyclamenwere of such low receptivity to B. cinereathat biocontrol was difficult to measure. Inall three hosts, biocontrol effectivenesswas 30 to 70% lower in senescent leavescompared with presenescent leaves, andalmost zero in dead leaves.

Biological control of B. cinerea in flow-ers presents special challenges because ofthe complexity of flower growth, devel-opment, and structure. In general, variousorgans of flowers develop and senesce overdifferent times, are differentially suscepti-

Fig. 5. Protection of calyces and fruits of tomato against Botrytis cinerea in colonizedtomato petals by treatment of petals with Gliocladium roseum. (A) Calyx blightinitiated from diseased petals. (B) Fruit rot developing from diseased calyces. (C)Calyces protected by G. roseum applied to petals. (D) Fruits protected by G. roseumin adhering petals.

Fig. 6. Biological control of gray mold ingeranium cuttings by Gliocladiumroseum. (A) Cuttings inoculated withBotrytis cinerea only. (B) Cuttingsinoculated with B. cinerea and with G.roseum.

322 Plant Disease / Vol. 81 No. 4

ble to B. cinerea, and interact differentlywith microbial antagonists (49,63). In-sights into effects of flower age on bio-control of B. cinerea were provided instudies of partially opened, fully opened,and senescent flowers of begonia (Begonia× hiemalis) that were inoculated with thepathogen (106 conidia per ml) and chal-lenge-inoculated after 24 h with G. roseumor T. koningii (each at 107 conidia per ml)(Fig. 8). Susceptibility to B. cinerea in-creased in the successive stages of flowerdevelopment. G. roseum increasingly sup-pressed the pathogen by 0, 49, and 68%;whereas T. koningii was consistently sup-pressive (93 to 95%) at the successivestages.

Inoculum concentration. An under-standing of inoculum concentration of G.roseum in relation to control of B. cinereais fundamental to establishing appropriatebiocontrol recommendations in crops. Asmight be anticipated, the relationships varywith crop, type and age of plant organ,pathogen concentration, microclimatic con-ditions, and other factors. By inference,optimal concentration of G. roseum fluctu-ates in the crop over time and spatially onfoliage, flowers, and fruits. As is common-place in disease management, settling onan inoculum concentration of G. roseumthat provides satisfactory control of B.cinerea in a crop will inevitably demandcompromise and approximation.

Studies under controlled conditions pro-vided perspectives of inoculum require-ments of G. roseum for controlling B. cine-rea in several crops. When foliage andflowers of begonia and cyclamen wereinoculated with various concentrations ofB. cinerea (0 to 106 conidia per ml) and ofG. roseum (0 to 108 conidia per ml) in allcombinations, good biocontrol wasachieved mainly when concentration of theantagonist was the same as or greater thanthat of the pathogen. In similar studies inraspberry, high levels of control (90 to100%) were obtained in leaves for all

combinations of 103 to 106 conidia of B.cinerea per ml and 104 to 108 conidia of G.roseum per ml, but in stems, stamens, andstigmas, only when concentration of theantagonist was 10 or 100 times greater thanthat of the pathogen (64). Unexpectedly, G.roseum at 108 conidia often was less effec-tive than at 107 or 106 conidia in stamensand stigmas, a circumstance also observedin foliage of black spruce in the growthroom. Apparently, it can be important notto exceed optimal concentrations of G.roseum for biocontrol in crops, especiallywhen conditions favor survival of the an-tagonist. In the greenhouse and growthroom, application of 107 conidia of G.roseum per ml strongly suppressed B. cine-rea applied at 106 conidia per ml in variousorgans of begonia, cyclamen, Exacum,geranium, gerbera, cucumber, pepper,raspberry, strawberry, tomato, and severalother hosts. In each instance, concentrationof the pathogen was sufficient to produceconsistently severe disease in the absenceof the antagonist.

Inoculum concentration of 106 to 108

conidia of G. roseum per ml normally pro-vided good control of B. cinerea in crops inthe field and in greenhouses. Inoculumcontaining 106 conidia per ml usually wassufficient in strawberry, begonia, and cy-clamen; 107 conidia per ml gave best re-sults in raspberry; and 107 to 108 was nearoptimal in black spruce (28,29,40,55,67).Applications of G. roseum at 5 × 106 con-idia per ml, timed using a disease fore-casting system, suppressed leaf blightcaused by B. squamosa in field plots ofcooking onions by 50 to 58% (22).

Temperature. Observations of tem-perature in relation to biocontrol activity ofG. roseum against B. cinerea were com-piled from separate studies in five hosts tofacilitate comparison of data (Fig. 9). Ingeneral, the level of biocontrol was high at20 and 25°C, and also at 30°C in hosts inwhich measurable infection was obtained,but was progressively less at 15 and 10°C.

Cool temperatures reduced biocontrol onlymarginally in leaves of strawberry andbegonia and in raspberry stamens, moder-ately in leaves of raspberry, cyclamen, andgeranium and in begonia petals, but mark-edly in petals of cyclamen and geranium(55,63; unpublished). Differential effectsof temperature on biocontrol in differenthost organs were evident in each of fourhosts in which two organs were investi-gated. From additional tests of several G.roseum isolates on flowers of cyclamenand geranium, different patterns of reducedbiocontrol at cool temperatures were at-tributable chiefly to the host but not toisolates of the antagonist.

Environmental water. Relationships ofatmospheric humidity, dew, rain, irrigation,or other forms of environmental water tobiological control of B. cinerea in crops byG. roseum are poorly understood. Avail-able evidence indicates that environmentalwater is of paramount importance in thesurvival, germination, and growth of G.roseum on plant surfaces and in penetrationof plants by the fungus. Conidia of G.roseum failed to germinate on dry foliageof black spruce seedlings in growth cham-bers at 12, 20, and 28°C, and their abilityto germinate, estimated by recovery onpotato-dextrose agar medium, decreasedsharply with time after application to seed-lings (Fig. 10). In black spruce and proba-bly other hosts, a humid period is neededwithin a few hours of application of G.roseum in order to optimize growth of theantagonist and biological control. Severalhours of surface wetness suffices in leavesand flowers of hosts that we have investi-gated, but relationships of form (droplets,films) and duration of wetness to biocon-trol have yet to be explored. Similarly, it isnot known whether humidity near satura-tion in the absence of surface moisture isconducive to biological control. In fieldstudies, natural and simulated rain shortlyafter application of G. roseum to straw-

Fig. 7. Exacum affine. (A) Sun scorch wounds on leaves. (B) Biological control of graymold in plants with scorch wounds. At 3 weeks after treatment, disease was severe inwater checks (left column), moderately severe in plants treated with iprodione (Rovral)or Trichoderma harzianum (second and third columns, respectively), and controlledalmost completely by G. roseum (right column).

Fig. 8. Conidiophore production of Bo-trytis cinerea in petals of begonia flow-ers of different growth stages that wereinoculated with the pathogen andtreated with water plus surfactant, Glio-cladium roseum, Trichoderma koningii,or chlorothalonil.

Plant Disease / April 1997 323

berry plants depleted inoculum density ofthe antagonist on the leaves and reducedthe effectiveness of biological control (7).Besides effects on G. roseum, effects ofenvironmental water on B. cinerea and onhost plants also could influence biologicalcontrol (48).

Relationships of G. roseumwith Host Tissues

Conidia of G. roseum are able to germi-nate in moisture on the host surface, pro-

duce simple germ tubes and superficialhyphae, and penetrate into leaves, stems,petals, and stamens of various kinds ofplants. Superficial development of theantagonist is best known from studies con-ducted under continuous high humidity at21 to 23°C on leaves, stems, and stamensof raspberry (63). A large proportion (70 to90%) of conidia germinated at 4 to 12 hafter inoculation and produced narrow (1 to1.5 µm diameter) germ tubes that elongatedrapidly after 12 and 16 h on stamens andstems, respectively, and slowly at 12 to 24h on leaves. Short branches (1 to 5 µmlong) developed on germ tubes and hyphaeafter 16 h and penetrated host tissues di-rectly. Remarkably, verticillate and peni-cillate conidiophores of G. roseum devel-oped from thickened superficial hyphae onstems, stamens, and leaves at 32 to 72 h(Fig. 11A) and produced abundant conidiaat 40 to 72 h. The antagonist thus has po-tential to reproduce in the epiphytic phase,a situation that, to our knowledge, is novelfor biocontrol agents applied to foliage andflowers and that could contribute to sus-tained control of B. cinerea in the crop.Evidence of penetration by G. roseum inseveral hosts is presumptive and based onrecovery of the antagonist from tissues thatwere inoculated and later surface-steril-ized. Such evidence was obtained in blackspruce, red pine, Exacum affine, geranium,gerbera, petunia, cucumber, and tomato, aswell as in raspberry.

Postpenetration development of G.roseum in plant foliage and flowers gener-ally is unclear. Whether the antagonistremains strictly localized or colonizes tis-sues more extensively may be a function ofthe type, age, and physiological status ofthe host tissues. Circumstantial evidence

based on recovery of G. roseum from tis-sues at various distances from inoculationsites indicated that the antagonist remainedlocalized for at least 3 months in stems ofvigorously growing tomato plants (G.Peng, unpublished). The antagonist pro-gressively colonized senescent leaves andflowers of strawberry, begonia, cyclamen,and geranium, and frequently sporulatedwithin a couple days after the tissues died(38,56; D-W Li and G. Peng, unpublished).While senescence and stress apparentlyfavor progressive colonization of host tis-sues by G. roseum, the antagonist also isknown to colonize vascular elements ofvigorous soybean and red clover plants(34,46,50). In our experience and in almostall other reports, host tissues infected by G.roseum remained symptomless. In the fewexamples of lesion production by G.roseum, such as in bean hypocotyls (20),host plants were under stress-inducingconditions and were treated with heavydoses of inoculum, often as mycelium.Long-term persistence in symptomlesstissues is a key feature of the parasiticactivity of G. roseum in stems, leaves,petals, and some other tissues of numeroushost plants.

Modes of Actionof Biological Control

Competition for nutrients or substrateand mycoparasitism are presumed modesof antagonism of G. roseum toward B.cinerea in host plants. Nutrient competitionwas thought to be important on the phyl-loplane of strawberry, raspberry, begonia,and cyclamen (55,57,63; D.-W. Li, unpub-lished). Substrate competition was consid-ered a key mode of biocontrol of B. cinerea

Fig. 10. Germination of Gliocladiumroseum conidia after periods of incuba-tion at various temperatures on dry foli-age of black spruce seedlings. Percentgermination was estimated after conidiawere recovered from the foliage andplaced on potato-dextrose agar medium.

Fig. 9. Observations from separatestudies in five hosts for effects of tem-perature on conidiophore production ofBotrytis cinerea in leaves and flowerparts that were inoculated with thepathogen and treated 24 h later withGliocladium roseum. Isolates of thepathogen and antagonist used in eachstudy were from the respective hosts.

Fig. 11. Development and interactions of Gliocladium roseum and Botrytis cinerea inraspberry. (A) Verticillate and penicillate conidiophores of G. roseum on living leaftissue. (B) Conidial germination and appressorium formation of B. cinerea on a leaf inthe absence of G. roseum. (C) Lack of conidial germination of B. cinerea on a leaf inpresence of G. roseum hyphae. (D) Narrow hyphae of G. roseum entwining and in-vading larger hyphae of B. cinerea.

324 Plant Disease / Vol. 81 No. 4

by G. roseum in senescing leaves of straw-berry and black spruce (55,66). Evidentlythe antagonist colonized and exploited thesenescent tissues much more rapidly thandid B. cinerea and largely precluded colo-nization by the pathogen. This mode ofaction is important chiefly because it con-tributes to inoculum suppression of B.cinerea. G. roseum is a well-known myco-parasite of hyphae, spores, sclerotia, andfruiting bodies of numerous fungi, includ-ing B. cinerea (3,10,61). While a majorityof studies were conducted on agar media,recent observations on raspberry indicatedthat mycoparasitism can be an importantmode of biocontrol of B. cinerea on hostplants (63).

Notably different interactions of G.roseum and B. cinerea were found onleaves, stems, and stamens of raspberrymaintained under continuous high humid-ity (63). On leaves, G. roseum stronglysuppressed germination and germ tubegrowth of B. cinerea but rarely parasitizedthe pathogen (Fig. 11B and C). On stems,however, conidia of B. cinerea frequentlygerminated and produced well-developedgerm tubes and superficial hyphae regard-less of nearby conidia, germ tubes, or hy-phae of G. roseum. Antagonism on stemswas manifested as modest suppression ofgerm tube growth and intense parasitism ofB. cinerea by G. roseum. During persistenthigh humidity (32 to 72 h), short hyphalbranches of G. roseum contacted, grew on,or coiled around, and in many instances,invaded conidia and hyphae of B. cinerea(Fig. 11D). Affected cells generally lackedcontents, and some hyphae collapsed.Whether or not cells of B. cinerea werealive when invaded was not determined.Continuity of superficial and intracellularhyphae indicated that many hyphae withinthose of B. cinerea were those of G.roseum, as opposed to fine regrowth hy-phae of the pathogen developing withindamaged cells from septal pores of adja-cent undamaged cells (10). On surfaces ofstamens, G. roseum generally did not sup-press germination, growth, or productionof appressoria and infection cushions of B.cinerea, and was not observed to parasitizethe pathogen. However, the antagoniststrongly reduced colonization incidence ofstamens by the pathogen. Differentialavailability and composition of nutrientsand of antifungal substances in exudates onthe host surface were advanced as hypothe-ses to explain the extraordinary variation ininteractions of G. roseum and B. cinerea onthe different raspberry tissues.

Besides competition and mycoparasit-ism, G. roseum possibly antagonizes B.cinerea through antibiosis and inducedresistance. The antagonist produces anarray of fungal inhibitors and wall-de-grading enzymes (10,37,38), some ofwhich could function in overlapping sce-narios of antibiosis and mycoparasitism.Loss of turgor and lysis of B. cinerea hy-

phae, each commonly induced by G.roseum, can be considered in the context ofboth of these modes of antagonism. Ingeneral, antibiosis appears to function onlyover short distances, implying that theantagonist must have access to nutrientsclose to the pathogen and that nutrientcompetition could underpin antibiosis (10).While G. roseum is able to produce at leastone antifungal metabolite, no evidence wasfound to indicate a role of the metabolite inbiological control. On the contrary, mu-tants of G. roseum that produced high orintermediate levels of the metabolite, ornone at all, did not differ in biocontroleffectiveness against B. cinerea in straw-berry leaves (38). Induced resistance inhost plants by G. roseum is an intriguingpossibility, especially in view of prolongedassociations of the antagonist with livingtissues of many hosts. Collectively, avail-able information underscores a high flexi-bility of G. roseum in modes of biocontrol.

Application Methodsand Strategies

A critical challenge in practical biologi-cal control is development of reliablemethods and strategies for applying thebiological control agent in the croppingsystem. Extraordinary attention to detail isneeded to optimize biological control un-der the full range of conditions in the cropand to minimize risks that errors will bemade when the system is in the hands ofgrowers. Methods and strategies generallyhave to be worked out on a crop-by-cropbasis to account for crop-specific condi-tions such as production methods and epi-demiologic factors. Integration of biologi-cal control with other crop protectionpractices is often called for and may differfrom farm to farm or region to region. Foroptimizing effectiveness of G. roseumagainst B. cinerea, considerable progresshas been made in terms of formulation ofinoculum and methods to time, target, anddeliver the inoculum in several crops.There should be no illusions, however,regarding the enormity of tasks that lieahead if we are to take full advantage ofthe antagonist in disease management.

Inoculum of G. roseum can be producedeasily and cheaply on sterilized grains ofwheat and other cereals. In our laboratory,high concentrations of the antagonist (1 to5 × 109 conidia per g) are routinely pro-duced within 35 days on wheat grainsmaintained at 20 to 23°C. After an initialperiod of mycelial growth (14 to 18 days),cultures are allowed to slowly dry, a proc-ess that promotes production of penicil-lately branched conidiophores, which yieldmore conidia than do verticillate conidio-phores. The verticillate form usually pre-dominates under persistent high humidity.Besides conidia, mycelial fragments of G.roseum recovered from wheat grains after15 days were found to be effective against

B. cinerea in strawberry. Conidia fromwheat grains have been stored for morethan a year at 3 to 5°C and for 3 months at20 to 23°C without significant loss in vi-ability or biocontrol effectiveness, findingsof considerable importance for commer-cialization of the antagonist. Prior to appli-cation to crops, conidia of G. roseum canbe suspended in water plus a surfactantsuch as Triton X-100 or formulated as apowder with talc or another carrier (41,54).

The best method of application of G.roseum ordinarily depends on the kinds ofhost tissues that require protection. Sprayapplication to deposit fine droplets of in-oculum uniformly on target surfaces isappropriate for general treatment of foliageand flowers in many crops, but more spe-cialized techniques often are needed fortargeting specific tissues. Wounds on gera-nium cuttings are easily treated by mo-mentary immersion of cuttings in inocu-lum, while deleafing wounds of green-house tomato and cucumber can be treatedwith a hand sprayer, brush, or cloth.Devices that simultaneously deleaf andspray inoculum onto the deleafing woundcan contribute to efficiency of these opera-tions. Flowers of strawberry and raspberryhave been successfully treated with G.roseum by means of bees employed asvectors of inoculum (41,63).

Few concepts in plant disease manage-ment can be considered more “environ-mentally friendly” than the use of bees todeliver a biological control agent to flow-ers for controlling a flower-infectingpathogen. In recent studies (41,63), honey-bees (Apis mellifera L.) effectively vec-tored inoculum of G. roseum to flowers ofstrawberry and raspberry, and bumblebees(Bombus impatiens Cresson) efficientlyvectored the antagonist to raspberry flow-ers (Fig. 12A and B). For vectoring, apowder formulation of G. roseum wasplaced in an inoculum dispenser mountedon the hive. Dispensers for use with hon-eybees and bumblebees were of differentdesign to accommodate structural differ-ences in hives (Fig. 12C to E); in bothinstances, however, bees were contami-nated with inoculum only when leaving thehive. In field plots, honeybees and bum-blebees trapped when emerging from dis-pensers containing inoculum with 5 × 108

to 1 × 109 CFU of G. roseum per g eachcarried several hundred thousand colony-forming units of the antagonist. Honeybeesdelivered 300 to 27,000 CFU per flower instrawberry and 600 to 2,100 CFU perflower in raspberry; whereas bumblebeesvectored 450 to 2,400 CFU per flower inraspberry. Bee-vectored inoculum effec-tively controlled B. cinerea in petals, sta-mens, and flowers of strawberry and inflowers of raspberry. However, raspberryfruits were not adequately protected, pre-sumably because B. cinerea is able to in-fect drupelets directly as well as by inva-sion from flowers.

Plant Disease / April 1997 325

Efficiency of bee vectoring can be af-fected by competing nectar sources outsidethe crop, periods when individual flowersare open, climatic factors, spatial relation-ships of hives and the crop, and other vari-ables (53,54). Honeybees are easily at-tracted away from strawberries, butraspberries are a favored nectar source.Bumblebees readily patronize both cropsand, unlike honeybees, forage in coolweather (e.g., 6 to 16°C). When conditionsare favorable, bee vectoring can have thepotential advantage of daily delivery ofinoculum to freshly opened flowers. Com-bination of inoculum delivery and hostpollination by bees is feasible in manycrops.

Optimal timing of G. roseum applica-tions against B. cinerea remains to beworked out for almost all crops. Goodtiming of treatments sometimes is easy toestablish, as in protection of deleafing andother artificial wounds, but is more diffi-cult for fast-growing foliage and flowers.Residual biocontrol activity of a G. roseumtreatment may depend heavily on epiphyticand endophytic relationships of the an-tagonist with specific host tissues and theability of the antagonist to remain activeand spread in host populations. Endophyticestablishment can obviate the need forfurther application of G. roseum to treatedtissues (55,66), but does not necessarily

protect new growth. In many, perhaps amajority, of instances, empirical studiesunder conditions representative of com-mercial crops are suitable for developingtreatment programs. In some crops, it maybe feasible to adapt disease predictionsystems to facilitate timing of G. roseumapplications, as was done in onion (22).

Outlook and Future ResearchA strong case can be made for continued

exploration and development of G. roseumas a biological control agent. The antago-nist has a remarkable record of satisfyingsocioeconomic expectations of efficiency,dependability, cost-effectiveness, andsafety in biological control. Strong per-formance of G. roseum against B. cinereain hosts as taxonomically diverse as straw-berry, geranium, tomato, and black sprucesuggests that the antagonist may effec-tively suppress the pathogen also in manycrops in which biocontrol has yet to beinvestigated. G. roseum also justifies fur-ther evaluation against other pathogens(32,42,47), particularly in view of its broadecological adaptation in plants and itswide-ranging mycoparasitic competence.Indeed, knowledge of the activity spectrumof the antagonist, at present fragmentary, iscentral to integration of the antagonist intodisease management programs. G. roseumhas advantages of abundant production and

long-term viability of inoculum, attributesof key importance for commercial use.Moreover, stickiness of the spores facili-tates handling and formulation of inoculumwhile minimizing any risks such as aller-gies associated with occupational exposureof workers. Fortunately, G. roseum doesnot grow at human body temperature, andstandard ocular, dermal, and feeding testsof G. roseum inoculum failed to producetoxicological effects in rabbits and rats. Invectoring studies, no evidence was foundthat G. roseum was harmful to bees.

Introduction of G. roseum into cropsystems should not present ecologicalrisks, especially in view of wide distribu-tion of the fungus in plants and soils inmany regions of the world and abundantevidence that G. roseum is not normallypathogenic to plants. In many instances,treatment of crops would simply augmentnatural populations of the antagonist. Ef-fectiveness of G. roseum against B. cinereain a broad range of crops can be expectedto increase market potential of the antago-nist, counter difficulties of commercialinterests with minor or specialty cropscommonly attacked by B. cinerea, andgenerally improve prospects for successfulimplementation of biological control.

G. roseum beckons numerous lines ofintriguing and useful research, both fun-damental and applied. A much clearer per-

Fig. 12. Bee vectoring of Gliocladium roseum inoculum from inoculum dispensers to flowers. (A) Honeybee on raspberry flower. (B)Bumblebee on raspberry flower. (C) Inoculum dispenser (painted black) mounted on a honeybee hive. (D) Inoculum dispensermounted on the front of a bumblebee hive. (E) Inoculum dispenser for bumblebees with top removed and inoculum-filled tray partlyopened. When leaving the hive, bees enter the dispenser through the lower of two rear holes, crawl through the inoculum and intothe upper chamber, and emerge from the front hole.

326 Plant Disease / Vol. 81 No. 4

John C. Sutton

Dr. Sutton is a professor in the Depart-ment of Environmental Biology at theUniversity of Guelph, Ontario, wherehe teaches plant pathology andconducts research. He received his B.Sc.from the University of Nottingham in 1965and his Ph.D. in plant pathology from theUniversity of Wisconsin in 1969. Hiscurrent research interests include epide-miology and management of diseases ingreenhouse ornamentals and vegetables,conifer seedlings, strawberries, and rasp-berries, with emphasis on biological andcultural control of Botrytis cinerea.

De-Wei Li

Dr. Li received his B.Sc. in forestprotection from Nanjing ForestryUniversity in 1982, his M.Sc. in forestpathology from Lakehead University in1990, and his Ph.D. in mycology fromthe University of Waterloo in 1994. Heworked as a postdoctoral researchassociate in the Department ofEnvironmental Biology at the Universityof Guelph from 1994 until 1996, whenhe joined the Agricultural College inOlds, Alberta. His research interestsinclude biological control of diseases ofvegetables and ornamentals in green-houses and aeromycology.

Gang Peng

Dr. Peng is a research anddevelopment specialist with AgriumInc. in Saskatoon. He received hisB.Sc. from Anhui College ofAgriculture and his Ph.D. in plantpathology from the University ofGuelph in 1991. He worked as a pestmanagement specialist with theOntario Ministry of Agriculture, Food,and Rural Affairs until 1994 and wason study leave at the University ofGuelph in 1995. Since joining AgriumInc. in 1996, he has worked oncommercial development of biologicalcontrol agents in cropping systems.

Hai Yu

Dr. Yu is a research associate in theDepartment of Environmental Biologyat the University of Guelph. Hereceived his B.Sc. in agronomy atQinghai University in 1982 and workedin crop physiology and breeding at theNorthwest Plateau Institute of Biology,Qinghai, until 1990, when he came tothe University of Guelph as a visitingscientist. He received his Ph.D. in plantpathology from the University ofGuelph in 1996. His research interestsinclude biological and cultural controlof diseases in small fruit andgreenhouse crops.

Pinggao Zhang

Dr. Zhang received his B.Sc. from theCentral China Agricultural University in1983, his M.Sc. from the ChineseAcademy of Agricultural Sciences in1986, and his Ph.D. in plant pathologyfrom the University of Guelph in 1992.Since 1992, he was a researchassociate at the Great Lakes ForestryCentre, Sault Ste. Marie, Ontario, andat the University of Guelph, where hereceived his M.Sc. in statistics in 1995.His research interests include epide-miology, modeling, and biologicalcontrol of diseases in conifer seedlingsand other crops. He is employed withConnaught Laboratories Inc. inToronto.

Rosa M. Valdebenito-Sanhueza

Dr. Valdebenito-Sanhueza receivedher B.Sc. from the University of Chilein Santiago in 1975 and her M.Sc.and Ph.D. degrees in plant pathologyfrom the University of São Paulo,Brazil, in 1979 and 1983, respectively.She worked in plant pathology in theChilean Institute for AgriculturalResearch (INIA) from 1970 to 1976,and since 1985 she has served as aplant pathologist at the BrazilianInstitute for Agricultural Research(EMBRAPA) in Bento Gonçalves andVacaria, Rio Grande do Sul, Brazil.Her research interests include bio-logical control and integrated manage-ment of diseases in temperate fruitcrops.

Plant Disease / April 1997 327

spective is needed of microclimatic vari-ables, host factors, fungicides, insecticides,and other components of cropping systemsin relation to population density and activ-ity of G. roseum in crops, especially fromthe perspective of integrating biologicaldisease control with other production andprotection practices. Genetic variability,stability, and relationships of the pool offungi referred to as G. roseum are largelyunexplored but could harbor solutions toproblems such as reduced effectiveness ofbiocontrol of B. cinerea at low tempera-ture. Highly tantalizing is the endophyticphase of G. roseum that is known to existin some plants (46,50,55,63). How wide-spread is this phenomenon among plantspecies? What relationships exist betweenthe endophyte and the host tissues? Can theendophyte induce resistance in the host topathogenic organisms such as B. cinerea?Almost nothing is known of G. roseum inplants except that it did not alter electrolyteleakage, chlorophyll level, and photosyn-thetic rate in black spruce (68). From theviewpoint of human needs, priority shouldbe given to exploiting G. roseum as a bio-logically and ecologically flexible antago-nist for suppression of B. cinerea and otherdestructive pathogens in cropping systems.

AcknowledgmentsPublication of this article and considerable re-

search reported therein were supported by theNatural Sciences and Engineering Research Coun-cil of Canada (grant OGP0006119 to the seniorauthor). Other portions of the research were sup-ported by the Ontario Ministry of Agriculture,Food, and Rural Affairs. We acknowledge scien-tific contributions of Guoxing Xue and thank D. J.Hamilton for photographic assistance and K.Dipersio and L. Bird for typing the manuscript.

Literature Cited 1. Al-Hamdany, M. A., and Salih, M. M. 1986.

Wilt causing fungi on broad bean. IndianPhytopathol. 39:620-622.

2. Arie, T., Namba, S., Yamashita, S., Doi, Y.,and Kijima, T. 1987. Stem blight, a new dis-ease of Exacum caused by Nectria glio-cladioides Smalley et Hansen. Ann. Phyto-pathol. Soc. Jpn. 53:570-575.

3. Barnett, H. L., and Lilly, V. G. 1962. A de-structive mycoparasite, Gliocladium roseum.Mycologia 54:72-79.

4. Bhatt, G. C. 1970. The soil microfungi ofwhite cedar forests in Ontario. Can. J. Bot.48:333-339.

5. Brandsberg, J. W. 1969. Fungi isolated fromdecomposing conifer litter. Mycologia61:373-381.

6. Carris, L. M., Glawe, D. A., Smyth, C. A.,and Edwards, D. I. 1989. Fungi associatedwith populations of Heterodera glycines intwo Illinois soybean fields. Mycologia 81:66-75.

7. Carvalho, R. A. 1994. Effect of microclimateon biological control of grey mould (Botrytiscinerea) by Gliocladium roseum in strawber-ries. M.Sc. thesis. University of Guelph,Guelph, Canada.

8. Dashwood, E. P., and Fox, R. A. 1988. Infec-tions of flowers and fruits of red raspberry byBotrytis cinerea. Plant Pathol. 37:423-430.

9. Dashwood, E. P., Fox, R. A., and Duncan, J.M. 1993. Effect of substrate and plant matur-

ity on the incidence of infection of potatoroots by pathogenic and nonpathogenic fungi.Mycol. Res. 97:733-745.

10. Deacon, J. W., and Berry, L. A. 1992. Modesof action of mycoparasites in relation to bio-control of soilborne plant pathogens. Pages157-167 in: Biological Control of Plant Dis-eases; Progress and Challenges for the Future.E. C. Tjamos, G. C. Papavizas, and R. J.Cook, eds. Plenum Press, London.

11. Dix, N. J. 1964. Colonization and decay ofbean roots. Trans. Br. Mycol. Soc. 47:285-292.

12. Dix, M. J. 1967. Mycostasis and root exuda-tion: Factors influencing colonization of beanroots by fungi. Trans. Br. Mycol. Soc. 50:23-31.

13. Domsch, K. H., Gams, W., and Anderson, T.-H. 1980. Compendium of Soil Fungi. Aca-demic Press, London.

14. Elad, Y., and Evensen, K. 1995. Physiologicalaspects of resistance to Botrytis cinerea.Phytopathology 85:637-643.

15. Foley, M. F., and Deacon, J. W. 1985. Isola-tion of Pythium oligandrum and other ne-crotrophic mycoparasites from soil. Trans. Br.Mycol. Soc. 85:631-639.

16. Gourley, C. O., and MacNab, A. A. 1964.Verticillium dahliae and Gliocladium roseumisolation from strawberries in Nova Scotia.Can. J. Plant Sci. 44:544-549.

17. Hanso, M., and Karoles, K. 1980. Someagents of damping-off in conifer seedlings inEstonian forest nurseries and associated mi-crofungi. Metsanduslikud Uurim. 16:120-134.

18. Hay, F. S., and Skipp, R. A. 1993. Fungi andactinomycetes associated with cysts of Het-erodera trifolii Goffart (Nematoda: Ty-lenchida) in pasture soils in New Zealand.Nematologica 39:376-384.

19. Houston, D. R., Mahoney, E. M., andMcGauley, B. H. 1987. Beech bark disease:Association of Nectria ochroleuca in W. VA.,PA., and Ontario. (Abstr.) Phytopathology77:1615.

20. Huber, D. M., and Finley, A. M. 1959. Glio-cladium roseum, a causal agent in the beanroot rot complex in Idaho. Plant Dis. Rep.43:626-628.

21. Ibnu, Z., and Supriana, N. 1987. Inventory ofdiseases at three Pinus merkusii nurseries inNorth Sumatra. Bul. Penelitian Kehutanan3:37-44.

22. James, T. D. W., and Sutton, J. C. 1996. Bio-logical control of Botrytis leaf blight of onionby Gliocladium roseum applied as sprays andby fabric applicators. Eur. J. Plant Pathol.102:265-275.

23. Jarvis, W. R. 1992. Managing Diseases inGreenhouse Crops. American Phytopatho-logical Society, St. Paul, MN.

24. Jones, G. M., Rodriguez-Kabana, R., andJatala, P. 1986. Fungi associated with cysts ofpotato cyst nematodes in Peru. Nematropica16:21-31.

25. Keinath, A. P., Fravel, D. R., and Papavizas,G. C. 1991. Potential of Gliocladium roseumfor biocontrol of Verticillium dahliae. Phyto-pathology 81:644-648.

26. Kennerley, C. M., Jeger, M. J., Zuberer, D. A.,and Jones, R. W. 1987. Populations of fungiassociated with sclerotia of Phymatotrichumomnivorum in Houston black clay. Trans. Br.Mycol. Soc. 89:437-445.

27. Kranz, J. 1974. Comparison of epidemics.Annu. Rev. Phytopathol. 12:355-374.

28. Li, D.-W., Sutton, J. C., and Peng, G. 1995.Gliocladium roseum effectively suppressesBotrytis cinerea in greenhouse-grown bego-nia. (Abstr.) Phytopathology 85:1191.

29. Li, D.-W., Sutton, J. C., and Peng, G. 1996.Effective suppression of Botrytis cinerea ingreenhouse-grown cyclamen by Gliocladium

roseum. (Abstr.) Can. J. Plant Pathol. 18:93-94.

30. Makkonen, R., and Pohjakallio, O. 1960. Onthe parasites attacking the sclerotia of somefungi pathogenic to higher plants and on theresistance of these sclerotia to their parasites.Acta. Agric. Scand. 10:105-126.

31. McLean, M. A., and Sutton, J. C. 1992. My-coflora of strawberry in Ontario. Can. J. Bot.70:846-852.

32. Moody, A. R., and Gindrat, D. 1977. Biologi-cal control of cucumber black root rot byGliocladium roseum. Phytopathology67:1159-1162.

33. Mosca, A. M. L., and Marchisio, V. F. 1985.Mycorrhizoplane fungal populations of Abiesalba seedlings. Allionia (Turin) 27:30-39.

34. Mueller, J. D., and Sinclair, J. B. 1986. Occur-rence and role of Gliocladium roseum infield-grown soybeans in Illinois. Trans. Br.Mycol. Soc. 86:677-680.

35. Mulligan, D. F. C., and Deacon, J. W. 1992.Detection of presumptive mycoparasites insoil placed on host-colonized agar plates.Mycol. Res. 96:605-608.

36. Novotna, J. 1989. Microscopic fungi in cystsof the potato nematode Globodera rosto-chiensis Wollenw. Ceska Mykol. 43:96-107.

37. Pachenari, A., and Dix, N. J. 1980.Production of toxins and wall degradingenzymes by Gliocladium roseum. Trans. Br.Mycol. Soc. 74:561-566.

38. Peng, G. 1991. Biological control of greymold (Botrytis cinerea) on strawberries. Ph.D.thesis. University of Guelph, Guelph, Canada.

39. Peng, G., and Sutton, J. C. 1990. Biologicalmethods to control grey mould of strawberry.Pages 233-240 in: Proc. Brighton Crop Prot.Conf. Pests Dis. Vol. 3. Br. Crop Prot. Counc.,Farnham, England.

40. Peng, G., and Sutton, J. C. 1991. Evaluationof microorganisms for biocontrol of Botrytiscinerea in strawberry. Can. J. Plant Pathol.131:247-257.

41. Peng, G., Sutton, J. C., and Kevan, P. G.1992. Effectiveness of honey bees for apply-ing the biocontrol agent Gliocladium roseumto strawberry flowers to suppress Botrytiscinerea. Can. J. Plant Pathol. 14:117-129.

42. Pratella, G. C., and Mari, M. 1993. Effective-ness of Trichoderma, Gliocladium, and Pae-cilomyces in postharvest fruit protection.Postharvest Biol. Technol. 3:49-56.

43. Pugh, G. J. F., and Dickinson, C. H. 1965.Studies on fungi in coastal soils. VI Gliocla-dium roseum Bainier. Trans. Br. Mycol. Soc.48:279-285.

44. Rehner, S. A., and Samuels, G. J. 1994. Tax-onomy and phylogeny of Gliocladium ana-lysed from nuclear large subunit ribosomalDNA sequences. Mycol. Res. 98:625-634.

45. Samuels, G. J. 1976. A revision of the fungiformerly classified as Nectria, subgenus Hy-phonectria. Mem. N.Y. Bot. Gard. 26:1-126.

46. Schiller, C. T., and Sinclair, J. B. 1984. Mi-croorganisms associated with soybean vascu-lar exudate and plant parts. Int. J. Trop. PlantDis. 2:1-4.

47. Seifert, K. A., Breuil, C., Rossignol, L., Best,M., and Saddler, J. N. 1988. Screening of mi-croorganisms with the potential for biologicalcontrol of sapstain in unseasoned lumber.Mater. Org. (Berl.) 23:81-95.

48. Sirjusingh, C., and Sutton, J. C. 1996. Effectsof wetness duration and temperature on in-fection of geranium by Botrytis cinerea. PlantDis. 80:160-165.

49. Sirjusingh, C., Sutton, J. C., and Tsujita, M. J.1996. Effects of inoculum concentration andhost age on infection of geranium by Botrytiscinerea. Plant Dis. 80:154-159.

50. Skipp, R. A., and Christensen, M. J. 1990.Selection for persistence in red clover: Influ-

328 Plant Disease / Vol. 81 No. 4

ence of root disease and stem nematode. N.Z.J. Agric. Res. 33:319-333.

51. Stewart, A., and Harrison, Y. A. 1989. Myco-parasitism of sclerotia of Sclerotium ce-pivorum. Australas. Plant Pathol. 18:10-14.

52. Sutton, J. C. 1990. Epidemiology and man-agement of Botrytis leaf blight of onion andgray mold of strawberry: A comparativeanalysis. Can. J. Plant Pathol. 12:100-110.

53. Sutton, J. C. 1994. Biological control ofstrawberry diseases. Adv. Strawberry Res.13:1-12.

54. Sutton, J. C. 1995. Evaluation of microor-ganisms for biocontrol: Botrytis cinerea andstrawberry, a case study. Pages 173-190 in:Advances in Plant Pathology. Vol. II, Sustain-ability and Plant Pathology. J. H. Andrewsand I. C. Tommerup, eds. Academic Press,London.

55. Sutton, J. C., and Peng, G. 1993. Biocontrolof Botrytis cinerea in strawberry leaves.Phytopathology 83:615-621.

56. Sutton, J. C., and Peng, G. 1993. Manipula-tion and vectoring of biocontrol organisms tomanage foliage and fruit diseases in croppingsystems. Annu. Rev. Phytopathol. 31:473-493.

57. Sutton, J. C., and Peng, G. 1993. Biosuppres-

sion of inoculum production by Botrytis cine-rea in strawberry leaves. Pages 47-52 in:Biological Control of Foliar and PostharvestDiseases. Proc. Workshop, Wageningen, TheNetherlands, 1992. N. J. Fokkema, J. Köhl,and Y. Elad, eds. IOBC/SPRD Bull. Vol. 16.Montfavet, France.

58. Theron, D. J., and Holz, G. 1991. Dry rot ofpotatoes caused by Gliocladium roseum. PlantPathol. 40:302-305.

59. Vakili, N. G. 1985. Mycoparasitic fungi asso-ciated with potential stalk rot pathogens ofcorn. Phytopathology 75:1201-1207.

60. Valdebenito-Sanhueza, R. M., Sutton, J. C.,Perazzolo, I., and Czermainski, A. B. C. 1996.Controle biológico de Botrytis cinerea, omofo cinzento dos morangueiros cultivadosem estufa. Fitopatol. Bras. (In press.)

61. Walker, J. A., and Maude, R. B. 1975. Naturaloccurrence and growth of Gliocladiumroseum on the mycelium and sclerotia of Bo-trytis allii. Trans. Br. Mycol. Soc. 65:335-338.

62. Whipps, J. M. 1987. Behaviour of fungiantagonistic to Sclerotinia sclerotiorum onplant tissue segments. J. Gen. Microbiol.133:1495-1501.

63. Yu, H. 1996. Relationships of epidemiologic

factors, Gliocladium roseum, and bee vectorsto gray mold of raspberry caused by Botrytiscinerea. Ph.D. thesis. University of Guelph,Guelph, Canada.

64. Yu, H., and Sutton, J. C. 1994. Inoculumconcentration of Botrytis cinerea and of thebiocontrol agent Gliocladium roseum in rela-tion to suppression of the pathogen in rasp-berry. (Abstr.) Phytopathology 84:1377.

65. Zhang, P. G., Sutton, J. C., and Hopkin, A. A.1994. Evaluation of microorganisms for bio-control of Botrytis cinerea in container-grownblack spruce seedlings. Can. J. For. Res.24:1312-1316.

66. Zhang, P. G., Sutton, J. C., and Hopkin, A. A.1995. Gliocladium roseum suppresses Bo-trytis cinerea in red pine seedlings. (Abstr.)Can. J. Plant Pathol. 17:364.

67. Zhang, P. G., Sutton, J. C., and Hopkin, A. A.1996. Inoculum concentration and time of ap-plication of Gliocladium roseum in relation tobiocontrol of Botrytis cinerea in black spruceseedlings. Can. J. For. Res. 26:360-367.

68. Zhang, P. G., Sutton, J. C., Tan, W., and Hop-kin, A. A. 1996. Gliocladium roseum reducesphysiological changes associated with infec-tion of black spruce seedlings by Botrytis cin-erea. Can. J. Plant Pathol. 18:7-13.