Thallic phialides - David Moore's World of Fungi: where ...€¦ · Radical views on development in some conidial fungi were expressed by Minter, Kirk & Sutton (1982).Thepresentpapercontinuesthe

Trans. Br. mycol. Soc. 80 (1), 39-66 (1983)

[ 39 ]

Printed in Great Britain

THALLIC PHIALIDES

By D W. MINTER, P. M. KIRK AND B. C. SUTTONCommonwealth Mycological Institute, Ferry Lane, Kew, Surrey TW9 lAF

The reappraisal initiated recently in these Transactions of development in conidial fungi andthe language used to describe is continued. The terminology proposed in 'Holoblasticphialides' is applied to a wider range of fungi. Two further developmental stages (conidialmaturation and collarette production) and three types of cell-wall building (apical, diffuse andring) are identified, and the terms used to describe certain conidial chains are revised. Analysisof selected deuteromycetes shows that sympodial, false- or no-chain phialidic, retrogressive,true-chain phialidic and certain types of thallic development are all parts of a continuum. Itis concluded that the terms phialidic, blastic and thallic are at present ill defined and cannotbe applied to fundamental divisions of conidial fungi, and that if a new system of classificationis to be devised for these fungi, it should not rest on such terms.

Radical views on development in some conidialfungi were expressed by Minter, Kirk & Sutton(1982).The presentpaper continues the reappraisal,using as a foundation the new ideas set outpreviously. Definitions of the five stages ofdevelopment which can occur when conidiogenouscells produce conidia, namely conidial ontogeny,conidial delimitation, conidial secession, proliferat-ion and regeneration, are given in Minter et al.(1982). Information derived from the transmissionelectron microscope is distinguished from thatderived from the light microscope by the samedrawing convention as used by Minter et al. (1982).

AIM AND ARRANGEMENT OF THE PRESENT

WORK

In deuteromycetes described as having phialidestwo developmental patterns can be recognized.Conidiogenous cells with one pattern produceconidia in gummy masses or in chains which are notheld together by wall material, but result simplyfrom the chance accumulation of one conidium ontop of another. Such conidiogenous cells aredescribed as false- or no-chain phialides. Conidio-genous cells with the other pattern produce conidiain chains which are held together by wall material.These are described as true-chain phialides (Minteret al., 1982).

The development of false- or no-chain phialidescan be explained in terms of the same five stages asare required to describe development in sympodialand annellidic fungi (i.e. conidial ontogeny,conidial delimitation, conidial secession, proliferat-ion and regeneration). The reason for this is that acontinuumexists with knownintergradingexamplesfrom false- or no-chain phialides, through annel-

Iides, to sympodial fungi. These fungi thereforeform a group related developmentally, and theunifying factor is that conidiogenous cells of allmembers, if they are to produce more than oneconidium, must proliferate between producingeach conidium (Minter et al., 1982). The develop-ment of true-chain phialides cannot be explainedin such terms alone, however, because they canproduce a large number of conidia in successionwithout intervening proliferation. Minter et al.(1982) suggested that additional stages to these fiveneed to be defined if development in true-chainphialides is to be described adequately. They alsoobserved that no intergradation has hitherto beendemonstrated between false- or no-chain phialidesand true-chain phialides.

This paper attempts to identify these extra stagesby assessing additional species . In many cases thesehave been chosen because they are the subject ofpreviously published research. They are describedand discussed in turn, and each new stage ofdevelopment identified is discussed immediatelyafter the example in which it first became apparent.It will become clear that intergradation exists notonly between false- or no-chain phialides and true-chain phialides, but also between these and somefungi in which development has generally beendescribed as thallic.

Examples will be discussed in four parts. Tomaintain continuity, the first part begins withCladobotrym varium Nees which Minter et al.(1982) showed to be located at one extreme of thesympodial, annellidic and false- or no-chainphialidic continuum. Intergradation will be shownbetween C. varium and certain conidiogenous cellswhich are not phialides but which produce conidiain true chains. In the second part intergradation

40 Thallic phialides

will be shown between these conidiogenous cellsand fungi such as Oidiodendron truncatum Barronwhich have generally been described as thallic. Inthe third part intergradation will be shown betweenconidiogenous cells producing conidia in truechains and fungi such as Aspergillus clavatus Desm.,which was cited by Minter et al. (1982) as havingtypical true-chain . phialides. Intergradationbetween thallic fungi and those with true-chainphialides will be demonstrated in the fourthpart; this integradation provided the title of thispaper. At the end of the four parts a generaldiscussion follows in which the implications ofthis work, particularly the significance of trueand false chains, .will be evaluated.

Fig. 1. Cladobotryum varium .

PART 1

Cladobotryum varium

Cladobotryum varium (Fig. 1) produces its firstconidium holoblastically (conidial ontogeny) anddelimits it (conidial delimitation) (F ig. 2). Thisconidium quickly becomes detached (conidialsecession) and the conidiogenous cell proliferatesenteroblastically (proliferation) to produce anotherconidium (conidalontogeny). The second conidiumis delimited by a septum (conid ial delimitation)which occurs lower down the axis of the conidio-genous cell than the base of the new inner walllayer produced by the conidiogenous cell during thefirst proliferation (Fig. 2, arrow). When the secondconidium becomes detached (conidial secession) ittakes with it all wall layers above the delimitingseptum and the conidiogenous cell becomesphysically shorter. These stages may then berepeated (succession of holoblastic conidialontogeny, retrogressive conidial delimitation,conidial secession and enteroblastic proliferation).

Wall building

Early growth of the conidiogenous cell and the firstconidium in this and many other deuteromycetesoccurs because wall material is produced (Fig. 3) in

Fig. 2. Interpretation of development inCladobotryum varium.

D. W . M inter, P. M. Kirk and B. C. Sutton

•

Fig . 3. Production of wall material at the apex of ahypha . Arrows locate secretory organelles. Blackened wallindicates reference point.

a small region of the hyphal cell apex which isundergoing modification to form the fertile element(Cole & Samson, 1979). Transmission electronmicroscopy shows that certain organelles areconcentrated in the cytoplasm at the cell apex,adjacent to the walls being laid down. It has beensurmised that they have a secretory function inassociation with wall production (Cole & Samson,1979). Although these organelles cannot individua-lly be observed with the light microscope, they cansometimes be seen collectively as a refractive bodycalled the "spitzenkorper ' (Cole & Samson, 1979),and the presence of the small apical region wherewall is being produced can be inferred by observinggrowth or because cell walls are th inner and, indematiaceous hyphomycetes, less pigmented in thisarea .

This region has been called a rneristematic zone(e.g. Cole & Samson, 1979; Kendrick, 1971) byanalogy with meristematic zones of higher plants.This is misleading because the meristematic zoneof higher plants, as the derivation of the wordsuggests (Greek: pep{~et", to divide into parts), isa multi-cellular region where growth occurs by celldivision. In fungi the term has been adoptedappropriately to describe growth by cell division inrhizomorphs of some basidiomycetes, and con-fusion could result from its use in deuteromycetesto describe a region within a single cell wheregrowth occurs by the laying down of wall material.It therefore seems sensible to avoid the termmeristematic altogether when describing cell wallproduction in individual cells of deuteromycetes,and the words ' wall building' are preferred .

Conidiogenous cells of C. varium grow to full sizeand produce a first conidium by apical wallbuilding (use of the word apex and its derivatives

Fig . 4. The secretory organelles of the first wall-buildingapex are separated from the conidiogenous cell when thefirst conidium is delimited. A replacement wall-buildingapex then occurs at the top of the conidiogenous cell.

in this context implies that the wall building isstrongly localized in that region). When the firstconidium is delimited this wall-building apex is lostby the conidiogenous cell (i.e. the apex does notremain in the conidiogenous cell: it mayor may notremain in the conidium), and is replaced by a newapex at the top of the conidiogenous cell, whichthen produces a second conidium when the firstsecedes (Fig. 4). The replacement wall-buildingapex in turn is lost when the second conidium isdelimited, and a third wall-building apex arises ina sequence which can be repeated. Since C. variumis a member of the continuum of fungi withsyrnpodial, annellidic and false- or no-chainphialidic proliferation (M inter et al., 1982), it seemslikely that development in those fungi generallyfollows this pattern of replacement of successivewall-building apices.

Trichothecium roseum

In Trichothecium roseum (Pers.) Link (Fig. 5) thefirst conidium is produced holoblastically by thefirst wall-building apex (conidial ontogeny) anddelimited (conidial delimitation) (Fig. 6). Thisconidium remains attached, and the conidiogenouscell proliferates enteroblastically (proliferation) toone side below the first conidium by the activity ofa new wall-building apex, which takes the line ofleast resistance. A second conidium is thusproduced (conidial ontogeny) and delimited by aseptum (conidial delimitation), which howeveroccurs lower down the axis of the conidiogenouscell than the base of the new inner wall layerproduced by the conidiogenous cell during the first


Fig. 5. Trichothecium roseum.

proliferation (Fig. 6, arrow) . These stages may thenbe repeated (succession of holoblastic conidialontogeny, retrogressive conidial delimitation andenteroblastic sympodial proliferation) and conidiamay become detached (conidial secession) at anytime . When conidia become detached, each (like itscounterpart in C. varium) takes with it all wall layersabove the septum delimiting it after it wasproduced. The conidiogenous cells of T. roseumthus, like those of C. varium, become physicallyshorter as a succession of conidia is produced (Cole& Samson, 1979).

Trichothecium roseum shares many features withC. oarium. It differs only because conidial secessionis delayed and proliferation as a result is entero-blastic-syrnpodial, the replacement wall-buildingapex taking the line of least resistance, whereas inC. varium conidial secession is earlier in thesequence and proliferation as a result is entero-blastic-percurrent. This difference is of minordevelopmental significance (Minter et al., 1982),and is probably not of major taxonomic importance

Fig. 6. Interpretation of development in Trichotheciumroseum (modified from Cole & Samson, 1979).

either : Cladobotryum Nees and Trichothecium Linkboth contain anamorphs of Hypomyces Tul.However, it has a profound effect on the appearanceof the fungus. Conidia of T . roseumadhere by innerand outer walls in a true chain, while conidia ofC. varium accumulate in gummy masses. Tricho-thecium roseum can accord ingly be regarded eitheras the sympodial counterpart of C. oarium (in thissense it is more extreme than C. varium in itsposition in the continuum of sympodial, annellidicand false- or no-chain phialidic fungi ) or, becauseof its unusual combination and sequence ofdevelopmental stages, as the first example ofanother continuum, comprising some of the fungiwith conidia in true chains .

Basipetospora rubra

In Basipetospora rubra G. Cole & Kendrick (Fig. 7)the first conidium is produced holoblastically by thefirst wall-building apex (conidial ontogeny), delim-ited (conidial delimitation) and remains attached(Fig . 8). Just as in T. roseum, the first wall-buildingapex is lost by the conidiogenous cell when the firstconidium is delimited. Unlike in T. roseum,

D. W. Minter, P. M. Kirk and B. C. Sutton

10 /LIn

Fig. 7. Basipetospora rubra.

43

however, this lost apex is not replaced, and in theproduction of the second and all subsequentconidia wall-building apices cannot be observed.Instead, the upper portion of the conidiogenous cellswells on all sides (proliferation) below the septumdelimiting the first conidium, and a secondconidium is formed which, like the first, remainsattached. These stages may then be repeated

Fig . 8. Interpretation of development in Basipeto sporarubra (modified from Cole & Samson, 1979).

(alternation of proliferation by swelling of theupper part of the conidiogenous cell to produce aconidium, and retrogressive conidial delimitation)and conidia may become detached at any time(conidial secession). As a result ofthis combinationof developmental stages, the conidiogenous cells ofB. rubra, like those of C. varium and T. roseum,become shorter with each conidium produced (Cole& Samson, 1979).

Until after the first conidium is delimited, thedevelopment of B. rubra is indistinguishable fromthat of T. roseum, C. uarium and any of the fungiin the sympodial, annellidic and false- or no-chainphialidic continuum. Even after the first conidiumisdelimited, thereare obvious and strong similaritiesbetween B. rubra and all of these fungi (butparticularly C. varium and T. roseum). There is alsohowever an important difference: B. rubra lacks areplacement wall-building apex. This is evidenteven under the light microscope, as there is nosympodial growth, no annellidic line across theconidiogenous cell and no periclinal thickening.The proliferation is therefore of a type notinvolving a wall-building apex, and as such it isdifferent from that in T. roseum, C. varium and allthe fungi cited by Minter et at. (1982). It isimportant to establish the natureofthis proliferationand whether or not it enables a significantdistinction to be made between B. rubra and theseother fungi .

Fig. 9. Tritirachium oryzae.

Thallic phialides

occur so that spheres and intermediates betweenspheres and cylinders can be produced. Becauseboth types of wall building have been observed ina wide variety of deuteromycetes, and both havebeen confirmed at an ultrastructural level in manyspecies, it seems reasonable to generalize that thediverse shapes in many conidia can be explained bypostulating different relative levels ofactivity of thetwo wall-building types. Cole & Samson (1979)have gathered an impressive array of evidencesupporting this generalization.

Fungi may now be considered in which all wallbuilding to produce a given conidium occurs at onlyone time in the developmental process. Amongthese fungi are Tritirachium oryzae (Vincens) deH-oog (Fig. 9), Belemnospora epiphylla P. M. Kirk(Fig. 10) and Cladobotrym varium (Fig. 1). If thegeneralization is true, apical and diffuse wallbuilding occur concurrently, but in differentrelative proportions in each species because theirconidia differ in shape. In T. oryzae diffuse wallbuilding plays an important part in determiningconidial shape because conidia are almost globose,and thus much wider across their middle than thepointat which they are attachedto the conidiogenouscell. In C. varium diffuse wall building plays a lessimportant part in determining conidial shapebecause the conidia are elongated ellipsoidal,although they are still about twice as wide acrosstheir middle as the point at which they are attachedto the conidiogenous cell. In B. epiphylla diffusewall building plays little or no part in determiningconidial shape, and the conidia are cylindrical androughly the same width across their middle as the

44

Diffuse wall building

In B. rubra proliferation occurs not by thereplacement of a wall-building apex, but byswelling on all sides of already existing walls.Research has shown that this form of wall buildingis not associated with high concentrations ofcytoplasmic organelles adjacent to the cell apex.Instead they are randomly distributed at a lowconcentration throughout the cytoplasm (Cole &Samson.iyrc). This is to be expected given thenon-polarized nature of this form of wall building,and is compatible with the surmise that they havea secretory function in association with wallproduction. It is therefore possible to recognize inthe production ofsecond and subsequent conidia ofB. rubra another type of wall-building activity,different from apical wall building, and describedhere as 'diffuse'.

Diffuse wall building differs from apical wallbuilding in that there is no region of high-activitywall production localized in the apex of the cell.Instead there is general, lower-activity wall pro-duction over a wide area of the cell, giving rise togrowth by conversion ofexisting wall, which is evenin all directions. This difference is correlated withand is probably the result of the variation indistribution within the cytoplasm of the relevantorganelles.

In B. rubra and many other deuteromycetes bothtypes of wall building occur at different timesduring conidial production. In all of these fungi, fora given length of cell wall, apical wall buildingalways precedes diffuse wall building (thus inB. rubra the wall of the conidiogenous cell, origin-ally produced by the first wall-building apex, isconverted into the second and subsequent conidiaby diffuse wall building). By examining fungi suchas B. rubra in which the two wall-building typesoccur at different times, it is possible to deducewhat contribution each type ·makes to the finalshape of the conidium.

In B. rubra the conidiogenous cell is produced bythe activity in isolation of the first wall-buildingapex and has a cylindrical shape on account of thepolarized growth. In the first conidium apical anddiffuse wall-building activity are concurrent andtheir effects cannot separately be evaluated. Thisconidium is globose. The second and subsequentconidia result from the activity of diffuse wallbuilding upon wall already produced by the wall-building apex. The diffuse wall building causesspecific areas of the cylindrical conidiogenous cellto swell, thus forming globose conidia. From thisit seems reasonable to infer that apical wall buildingtends to produce cells cylindrical in shape, whilediffuse wall building enables lateral swelling to

D. W. Minter, P. M. Kirk and B. C. Sutton 45

10/lln

Fig. 10. Belemnospora epiphylla.

pointat which they are attachedto the conidiogenouscell.

It is thus evident that in many deuteromycetesboth apical and diffuse wall building are involvedin forming conidia, even if these two stages occurconcurrently and cannot visually be distinguishedexcept by deduction from the conidial shape (thetwo types are concurrent in all examples discussedby Minter et al. (1982) and in this paper, except forB. rubra, which is why until now there has been noneed to distinguish the two wall-building types).Apical wall building can precede diffuse wallbuilding (e.g. Basipetospora rubra), the two canoccur concurrently (e.g. Tritirachium oryzae), andapical wall building can produce conidia with littleor no contribution from diffuse wall building (e.g.Belemnosporaepiphylla). There appears however tobe no case known where, for a given length of cellwall, diffuse wall building precedes the apical type.This is not surprising, since the conversion ofalready existing walls appears to be a fundamentalfeature of diffuse wall building.

Conidial maturation

From the preceding discussion it is evident that thepresence of diffuse wall building in B. rubra is nothighly significant in delimiting it from fungi of thesympodial, annellidic and false- or no-chain

phialidic continuum. What is important however isthat in B. rubra and similar fungi the apical wallbuilding occurs at an earlier and separate stage ofdevelopment from diffuse wall building. Minter etal. (1982) have catalogued some of the confusionswhich can arise when a stage occurring at one timeduring the development of a conidial fungus isdescribed in terms applicable only to a differentstage occurring at another time. These are exactlythe sorts of confusions which could arise if, as hasoften been done in the past, the diffuse wall-buildingstage in B. rubra is described as conidial ontogeny.Such a description implies that the diffusewall-building stage is directly comparable withconidial ontogeny in, for example, Belemnosporaepiphylla, where the term refers to the productionof conidia by apical wall building alone.

It is therefore confusing to use the term conidialontogeny for both apical and diffuse wall-buildingstages, particularly when the two stages occur atdifferent times. A decision must be made as towhich stage the term concerns, and to judge fromits general use in the past, maximum clarity will bemaintained if conidial ontogeny is used in future torefer only to wall-building stages in which walls areproduced where none existed before. Conidialontogeny can therefore be brought about by apicalwall building (although this is not necessarily theonly type of wall building which can result inconidial ontogeny), but it cannot be brought aboutby diffuse wall building. A new term must be foundto describe the stage characterized by diffuse wallbuilding, and it is suggested here that a suitableterm would be 'conidial maturation'. In this paperconidial maturation is used to refer to theproduction or modification of conidia by diffusewall building. Using the light microscope thetiming of a conidial maturation stage in a givenfungus can be deduced from the shape of theconidium or conidial initial.

Conidial maturation, when it occurs, can beconcurrent with or can follow conidial ontogeny. Itdoes not necessarily follow immediately, however,and between conidial ontogeny and conidialmaturation intervening stages may occur. InBasipetospora rubra the conidial ontogeny of allconidia is incorporated in the initial growth of theconidiogenous cell (this is possible because ofretrogressive delimitation), and conidial ontogenyand conidial maturation occur concurrently only inthe case of the first conidium. In the second andsubsequent conidia the duration of time betweenconidial ontogeny and conidial maturation isprogressively greater, and more stages intervene(Fig. 11).

Two further features of conidial maturationworthy of note can be observed in B. rubra. First,

Thallic phialides

KJHGFEDcBAn

Fig. 11. Development in Basipetospora rubra showing how, with each conidium produced, the stages ofconidial ontogeny and conidial maturation become increasingly separated in time. A. Conidiogenous cell. B.Conidial ontogeny of the fourth conidium. C. Conidial ontogeny of the third conidium. D. Conidial ontogenyof the second conidium. E. Simultaneous conidial ontogeny and maturation of the first conidium. F.Delimitation of the first conidium. G. Maturation of the second conidium. H. Delimitation of the secondconidium. I. Maturation ofthe third conidium. J, Delimitation of the third conidium. K. Maturation ofthefourth conidium.

proliferation and conidial maturation occur con-currently, after conidial ontogeny, on the same areaof cell wall, so that it is impossible to distinguishthe two stages in terms of the time or place in whichthey happen. This compares strikingly withdevelopment in fungi, where conidial ontogeny andconidial maturation occur concurrently immedi-ately after proliferation (e.g. in false- or no-chainphialides). In these fungi the temporal and spatialdistinctions between proliferation and conidialontogeny, though small, are of crucial importancein understanding the true affinities of the develop-ment (Minter et al., 1982) (Fig. 12). Secondly, andlinked to the first observation, with diffuse wallbuilding the distinction between holoblastic andenteroblastic proliferation becomes meaningless, asthere is no strongly polarized growth to make anidentifiable breach in a pre-existing wall barrier.

To summarize, therefore, in fungi where prolif-eration precedes concurrent conidial ontogeny andconidial maturation (e.g. in false- or no-chainphialides), there is no great value in recognizing asseparate conidial ontogeny and conidial maturation,but the distinction between these two and prolifer-ation is essential. By comparison, in fungi whereconidial ontogeny precedes concurrent prolifer-ation and conidial maturation (e.g. Basipetosporarubra), there is no great value in recognizing

proliferation and conidial maturation as separate,but the distinction between these two (particularlyconidial maturation, because it has not beenrecognized before) and conidial ontogeny is essen-tial. There is only one advantage to be gained inrecognizing and listing as separate each stageregardless of whether or not it is significant in theparticular fungus under observation. Such carefuland systematic listing enables the true affinities tobe seen between fungi such as Tritirachium oryzaeand Basipetospora rubra. Omission of a stage fromthe list, simply because it is not obvious, can totallyobscure these affinities.

Divergence

Cladobotryum varium, Trichothecium roseum andBasipetospora rubra show intergradation indevelopmental characteristics between fungi ofthe sympodial, annellidic and false- or no-chainphialidic continuum and certain fungi producingconidia in true chains. Both can now be related toa different type of wall-building activity. AlthoughB. rubra and similar fungi produce conidia in truechains, they cannot be said to have true-chainphialides for three reasons. The first is that theenteroblastic mode of development is not involvedin production of conidia in these fungi. The second


PART 2

Basipetospora variabilis

No time-lapse or other developmental studies havebeen carried out on this species, and this accountis based on analysis of the original detaileddescription and illustration by Matsushima (1975)·In Basipetospora variabilis (F ig. 13) the conidialontogeny stages of the second and subsequentconidia are incorporated in the initial activity of thefirst wall-building apex. The first conidium isproduced holoblastically by the concurrent activityof the first wall-building apex, and diffuse wallbuilding(concurrent conidial ontogeny and conidialmaturation), is delimited (conidial delimitation)

Ontogeny!maturation

Proliferation

Proliferation!maturation

..

is that they proliferate between producing eachconidium (even though the proliferation is difficultto distinguish from conidial maturation). The thirdis that the number of conidia which a givenconidiogenous cell can produce in these fungi islimited ultimately by the original length of thatconidiogenous cell. In many fungi with develop-mental patterns closely comparable with those ofB. rubra mechanisms have evolved which circum-vent this limitation. They may be divided into twobroad categories, perhaps evidence of an evolu-tionary divergence. The first leads from Basipeto-spora variabilis Matsushima to fungi such asOidiodendron truncatum which in the past havebeen described as thallic. The second leads fromBasipetospora chlamydosporis Matsushima eventu-ally to fungi with true-chain phialides like those ofAspergillus clavatus.

Fig. 12 . In false- orno-chain phial ides (top ) the distinctionbetween proliferation and ontogeny/maturation is im-portant. In fungi like Basipetospora rubra (bottom) pro-liferat ion and maturation occur simultaneously long afterontogeny has ended.

Ontogeny of next conidiumalready in existence----

Fig. 13. Basipetospora variabilis.

Thallic phialides

Fig. 15. Hypothetical later development ofBasipetospora variabilis.

Fig. 14. Hypothetical early development ofBasipetospora variabilis.

and remains attached (Fig. 14). The first wall-building apex is thus lost into the conidium by theconidiogenous cell, and is not replaced. Theconidiogenous cell then proliferates by diffuse wallbuilding below the septum delimiting the firstconidium, to produce a second conidium (concur-rent proliferation and conidial maturation) which isdelimited (conidial delimitation) and remainsattached. This sequence of stages can be repeated(alternation of concurrent proliferation/conidialmaturation and retrogressive conidial delimitation),and conidia may become detached (conidialsecession) at any time.

Thus far development in B. variabilis isindistinguishable from that in B. rubra. Thedifference between them becomes apparent onlyafter several conidia have been produced. In B.variabilis, unlike in B. rubra, the succession ofstages to produce new conidia does not stop whenthe conidiogenous cell is totally converted intoconidia: instead it continues by moving retrogres-sively to the cell belowthe conidiogenous cell, whichbecomes a new conidiogenous cell and is in turngradually converted into conidia. When this cell istotally converted, the succession moves to the cellbelow in a sequence which can be repeated. In thisway, if sufficient conidia are formed, branchedchains of conidia will be seen as a result ofretrogression to a point beyond the hyphalbranching below the original conidiogenous cell(Fig. 15).

In B. variabilis therefore the number of conidiaproduced is not limited by the original length of theconidiogenous cell. The continued retrogressionwhich enables more conidia to be produced caneasily be detected with the light microscope simplyby comparing the length of a mature chain ofconidia with that ofa young original conidiogenouscell which has produced only one conidium (themature chain will be much longer than the youngconidiogenous cell), and by looking for branchedchains of conidia. As was observed in thecomparison of Cladobotrym varium and Tricho-thecium roseum, a small difference in developmentcan often have a profound effect on the appearanceof a fungus. In the cases ofB. variabilis and B. rubrathe small difference is that retrogression continuesbeyond the original conidiogenous cell in the one,and does not in the other. The profound effect onappearance is that one has branched chains ofconidia and the other does not. This however, likethe difference in appearance between C. varium andT. roseum, is not necessarily of great taxonomicsignificance.

When B. variabilis is compared directly with anymember of the sympodial, annellidic and false- orno-chain phialidic continuum (e.g. Belemnosporaepiphylla (Fig. to) or Tritirachium oryzae (Fig. 9)),there are such marked and apparently fundamentaldifferences in development that it is difficult tobelieve the fungi could be in any way developmen-tally related. Belemnospora epiphylla, for example,produces conidia singly by apical wall building,with percurrent proliferation after each secedes andthus with a growing conidiogenous cell, where-as Basipetospora variabilis produces conidia in

D. W. Minter, P. M . Kirk and B. C. Sutton 49

~:~~

Fig . 16. Oidiodendron truncatum.

branched chains by diffuse wall building, withretrogression and hence conidiogenous cells whichbecome smaller.When they are compared indirectly,however, through a series of intergrading examples,as in this paper and Minter et al. (1982), it becomesapparent that all are developmentally related. Anytwo adjacent members of this series have incommon almost all developmental features, anddiffer only slightly in one or two features which, onexamination, are clearly not significant in distingu-ishing these adjacent species developmentally. Theapparently fundamental differences observed whenBelemnospora epiphylla or Tritirachium oryzae and

Basipetospora variabilis are compared are thusclearly only a gradual accumulation of minordifferences, and it is evident that continuousintergradation exists between all of these fungi.

Although Belemnosporaepiphylla or Tritirach iumoryzae and Basipetospora variabilis are verydifferent in appearance, belonging as they do togreatly differing parts of a spectrum, it is agreedthat the genera to which they belong share at leastone important feature in common, namely all areblastic. Blastic is one of two modes of developmentwhich , with few dissenting voices (Ingold, 1981b),are generally believed to be of fundamentalsignificance in deuteromycete classification (Cole &Samson, 1979; Ellis, 1971, 1976; Kendrick, 1971 ;Sutton, 1980). The other is thallic. An example ofthallic development will now be described andcompared with blastic development of B . variabilis.This example, Oidiodendron truncatum (Fig. 16),has been the subject of a developmental study fromwhich time-lapse pictures have been published(Cole & Samson, 1979; Kendrick, 1971). It isgenerally agreed to be thallic (Cole & Samson,1979; Ellis, 1976; Kendrick, 1971) and to becorrectly placed in Oidiodendron, a genus which alltreatments in recent years have regarded as havingthallic development (Cole & Samson, 1979; Ellis,1971, 1976; Kendrick, 1971; Sigler & Carmichael,1976).

Oidiodendron truncatum

In O. truncatum the conidial ontogeny stages of thesecond and subsequent conidia are incorporated inthe init ial activity of the first wall-building apex(Fig. 17). The first conidium is produced holo-blastically by the activity of the first wall-building

Fig. 17. Development of Oidiodendron truncatum. Large arrows point to conidial ontogeny stages. Whitearrows point to delimitation stages. Small arrows point to maturation stages.

Fig. 19. Marked thickening of delimiting septain Oidiodendron truncatum.

Thallic phialides

to blastic development in B. variabilis. Indeed inmany respects development in the two species is thesame. Both produce conidia in branched chains asa result of diffuse wall building acting upon wallsalready built by a wall-building apex. In both theyoungest conidium is at the base, and is producedas a result of a sequence involving retrogressivedelimitation and conidial maturation, with noreplacement wall-building apices. In both, after theoriginal conidiogenous cell is totally converted intoconidia, the cell below and then the cell below thatin tum become conidiogenous cells and areconverted into conidia in a sequence which can berepeated. Therefore thereare manymore similaritiesin development between O. truncatum and B.variabilis than there are between B. variabilis and,for example, Belemnospora epiphylla.

It is however also important to observe whatdifferences exist between the developments ofO. truncatum and B. variabilis. First, conidialdelimitation does not occur in a perfectly regularretrogressive sequence in O. truncatum. Secondly,in O. truncatum conidial delimitation occurs earlier,so that conidia are usually delimited before theonset ofdiffuse wall building and always well beforediffuse wall building has finished (examination ofthe time-lapse photomicrographs published byKendrick (1971) shows some enlargement beforedelimitation of the pro-penultimate conidium ofthe right-hand branch). Thirdly and related to the

5°

Fig. 18. Development of Oidiodendron truncatum. Arrowspoint to delimitation stages not occurring in exactretrogressive sequence.

apex (conidial ontogeny), is delimited (conidialdelimitation) and remains attached. The first wall-building apex is thus lost into the conidium by theconidiogenous cell, and is not replaced. The secondconidium is then delimited (conidial delimitation)retrogressively and remains attached, while the firstconidium begins to mature slowly (conidial matur-ation). This sequence of stages can be repeated(retrogressive conidial delimitation followed byretrogressive gradual conidial maturation), andconidia may become detached (conidial secession)at any time following completion of maturation.The succession of stages to produce new conidiadoes not stop when the conidiogenous cell is totallyconverted into conidia; instead it continues bymoving retrogressively to the cell below theconidiogenous cell, which becomes a new conidio-genous cell and is in tum gradually converted intoconidia. When this cell is totally converted, thesuccession moves to the cell below in a sequencewhich can be repeated. In this way, if sufficientconidia are formed, branched chains of conidia willbe seen as a result of retrogression to a point beyondthe hyphal branching below the original conidio-genous cell. Development in O. truncatum some-times deviates from the sequence of stages outlinedabove in that production of septa which delimitconidia does not always occur in a strict retro-gressive sequence (Fig. 18).

In the foregoing account, no new term has beenintroduced and much of the wording is very similarto that used to describe the conjectured developmentin B. variabilis and even the well-studied develop-ment of B. rubra. Thus thallic development in O.truncatum can be described in a way broadly similar

D. W. Minter, P. M. Kirk and B. C. Sutton 51second difference, because delimitation is earlier,second and subsequent conidia are usually pro-duced without the conidiogenous cell prolifer-ating . Fourthly, and not mentioned in theforegoing description, the delimiting septa ofO. truncatum become markedly thickened as theconidium matures (Fig. 19).

Of these differences, the third need not receivemuch discussion as it is clearly dependent upon thesecond, and it has already been suggested in thispaper that in some deuteromycetes with retrogres-sive development the distinction between prolifer-ation and conidial maturation is not greatlysignificant. It is difficult to comment on thesignificance of the first difference, because this sortof information is available for only a small numberof deuteromycetes, and little is known about suchvariation. There appear to have been no serioussuggestions that the order in which conidia aredelimited should be the basis for the fundamentaldivision of deuteromycetes into blastic and thallic.The same can be said about the fourth difference:secondary thickening of septal walls during conidialmaturation is well known in many strong chainphialides (e.g. in Aspergillus Mich. ex Fr.) and thesehave always in the past been regarded as blastic.The only difference remaining to be discussed isthe second.

In essence, the second difference is that thesequence of developmental stages is altered, so thatconidial delimitation occurs earlier when comparedto conidial maturation. In the series of intergradingexamples considered in this paper and Minter et al.(1982), alterations in the sequence of developmentalstages have been observed between several adjacentexamples. Conidial secession in Acrogenosporasphaerocephala (Berk. & Br.) M. B. Ellis is delayedin comparison with Stigmina angusiana M. B. Ellis(Minter et al., 1982), and conidial maturationoccurs later in Basipetospora rubra than in Triche-thecium roseum. When these examples wereexamined, it was concluded that the alterationsin sequence were of minor significance: it wascertainly evident that no fundamental distinctionin developmental classification could be based onthem . Although the following statement questionsthe basis for recognizing thallic development asdistinct from blastic, there is no intention ofentering into a detailed discussion of this sub ject inthe present paper. There appears to be no cogentreason why the alteration in sequence of conidialdelimitation between B. variabilis and O. truncatumshould be considered of more significance than anyof the other alterations in sequence alreadyencountered, especially in view of the fact that thetwo species share a great majority of other develop-mental features. It is therefore concluded that a

continuous intergradation exists between thallicfungi such as O. truncatum and members ofthe sympodial, annellidic and false- or no-chainphialidic continuum.

PART 3

Basipetospora chlamydosporis

As with B. variabilis, no time-lapse or otherdevelopmental study of B. chlamydosporis (Fig . 20)

is available, and the following account is based onanalysis of the original detailed description andillustration by Matsushima (1975). The deductionsfrom this analysis rely greatly on Matsushima's useof the word 'meristem', and may appear at firstsight to have little justification. These deductionsare, however, backed by further evidence which

Fig. %0. Basipetospora chlamy dosporis.


Fig. 21. Hypothetical early development of Basipetospora chlamydosporis.

will become apparent when other fungi areanalysed later in the text.

In B. chlamydosporis the conidial ontogeny stagesof the second and some subsequent conidia areincorporated in the initial activity of the firstwall-building apex. The first conidium is producedholoblastically by the concurrent activity of the firstwall-building apex and diffuse wall building(concurrent conidial ontogeny and conidial matur-ation), is delimited (conidial delimitation) andremains attached (Fig. 21). The first wall-buildingapex is thus lost into the conidium by theconidiogenous cell, and is not replaced. Theconidiogenous cell proliferates by diffuse wallbuilding below the septum delimiting the firstconidium, to produce a second conidium (concur-rent proliferation and conidial maturation) whichis delimited (conidial delimitation) and remainsattached. This sequence of stages can be repeated(alternation of concurrent proliferation/conidialmaturation and retrogressive conidial delimitation),and conidia may become detached (conidialsecession) at any time. The succession of stages toproduce new conidia does not stop when theconidiogenous cell is totally converted into conidia:instead it continues by moving retrogressively tothe cell below the conidiogenous cell, whichbecomes a new conidiogenous cell and is in turngradually converted into conidia. When this cell istotally converted, the succession moves to the cellbelow in a sequence which can be repeated. In thisway, if sufficient conidia are formed, branchedchains of conidia will be seen as a result of retro-gression to a point beyond the hyphal branchingbelow the original conidiogenous cell.

Thus far development is indistinguishable fromthat of B. variabilis, and the two species are

therefore very similar in appearance, but at thispoint an important difference can be observed:whereas in B. variabilis retrogression continuesindefinitely, resulting in more and more of theoriginal hyphae being converted into conidia, in B.chlamydosporis after a while the original hyphaecease to be converted into conidia, and retrogressionappears to halt. At this point a new wall-buildingzone comes into being beneath the last conidium tobe produced and delimited by retrogression. Thiswall-building zone produces new wall material

Fig. 22. Hypothetical later development of Basipetosporachlamydosporis. Black arrows indicate movement ofreference conidium. White arrows indicate position ofnew wall-building zone.

D. W. Minter, P. M. Kirk and B . C. Sutton 53continuously at its upper end, and it is th is wallmaterial rather than the walls of already existinghyphae which forms the walls of all subsequentconidia. A steady state therefore comes about inwhich the continuous upward conidial ontogenybalances the continuous downward maturation anddelimitation, with the result that the conidiogenouscell in which the new wall building zone has arisenceases to become shorter, but the chain of conidiaincreases in length (Fig. 22 ).

Ring wall building

The continuous upward wall production in the finalconidiogenous cell of B. chlamydosporis is the resultof wall-building activity. This activity cannot bedescribed as apical, because there is no cell apex inthe region in which it occurs, yet it has obvioussimilarities with apical wall building in all theexamples previously examined: i.e. it produces anew cylindrical wall from a localized region.Because of these similarities, the type of wallbuilding found in B. chlamydosporis has rarely been

Fig. 23. Distribution of secretory organelles in Chalarasp., and semi-diagrammatic illustration of the hypo-thetical shape of the wall-building ring in Chalara sp.

clearly separated in the past from apical wallbuilding. Their differences are however highlysignificant: the wall building apex produces wallsdownwards (so that it always remains above thewalls it has produced) whereas this wall buildingzone produces walls upwards (so that it alwaysremains below the walls it has produced). The wallbuilding found in B. chlamy dosporis is also differentfrom diffuse wall building, because diffuse wallbuilding produces swelling over a wide region byconversion of already existing walls (in any case inB . chlamydosporis diffuse wall building can beobserved separately causing conidial maturation).The activity in the upper part of the finalconidiogenous cell of B. chlamydosporis is thus theresult of a new and different type of wall building.

The ultrastructural basis for this wall buildingcannot be discussed for B. chlamydosporis as notransmission electron microscope study is available.It was, however, originally discovered in anotherhyphomycete, a species of Chalara (Cda) Rabenh.,and electron microscopy of this species (Hawes &Beckett, 1977b) has made it clear that this wallbuilding, like apical and diffuse wall building, canbe correlated with variation in concentration ofsupposed secretory organelles in the cytoplasm.Whereas in diffuse wall building theyare distributedrandomly and at a low concentration throughoutthe cytoplasm, and in apical w111 building they areconcentrated at the cell apex, in the new type of wallbuilding they are concentrated along the sides of thecell (Fig. 23 ) (Hawes & Beckett, 1977b). Onaccount of the shape of this distribution, the newtype is described here as 'ring wall building ' (theword ring being used here either as an adjective ora noun, and intended to imply that the wall buildingis highly localized). The ultrastructural explanationof the three different types of wall building makesit very likely that they all intergrade. It shouldtherefore be remembered that each is no more thana convenient highlight in a spectrum.

In ring wall building, as in apical and diffuse wallbuilding, the distribution of the organelles cannotbe detected using light microscopy. It is howeverpossible to deduce whether or not ring wallbuilding is in operation from the appearance of afungus under the light microscope. Just as apicalwall building is characterized by sympodial growth,annellidic lines across the conidiogenous cell orpericlinal thickening, and diffuse wall building ischaracterized by deviation in shape from thecylindrical, so the presence ofring wall building canbe deduced by observation of true chains whichgrow in length, with the youngest conidium at thebase, from a conidiogenous cell which remains thesame size. In the following descriptions anddiscussions, when ring wall building is mentioned,


its presence has been deduced using this criterion,unless specifically stated otherwise.

In apical wall building, the wall-building apexalways remains above the walls it has produced. Ittherefore moves up into any conidium producedand is lost by the conidiogenous cell. In fungi withconidiogenous cells in which apical wall building isthe only means of producing new walls, theproduction of more than one conidium from aconidiogenous cell is dependent either on continu-ous retrogression (as in B. variabilis) or on areplacement wall-building apex coming into exist-ence after each conidium is produced. Replacementwall-building apices cause not only conidialontogeny but also proliferation, and this explainswhy proliferation is unavoidable in the sequence ofdevelopmental stages in all fungi of the sympodial,annellidic and false- or no-chain phialidic continu-um. By comparison, in ring wall building thewall-building ring always remains below the wallsit has produced. The ring does not move up intothe conidium and is thus not lost by theconidiogenous cell when a conidium is delimited.Once a conidiogenous cell has been modified toproduce a wall-building ring, it can produce asuccession of conidia without further proliferation.Recognition of the wall-building ring is thusconsidered here to be the step of crucial importancetowards understanding how true-chain phial idescan produce many conidia without interveningproliferations.

Sagenomella striatisporaThis account of development of striated conidia inS. striatispora (Onions & Barron) W. Gams (Fig.24) interprets the detailed developmental studies bylight microscopy of Subramanian & Pushkaran(1975). In S. striatispora the first conidium isproduced holoblastically by the concurrent activityof the first wall-building apex, and diffuse wallbuilding (concurrent conidial ontogeny and con-idial maturation), is delimited (conidial delimita-tion) and remains attached (Fig. 25). The firstwall-building apex is thus lost into the conidium bythe conidiogenous cell, and is not replaced . Theconidiogenous cell immediately produces a wall-building ring below the septum delimiting the firstconidium, and the activity of this ring withconcurrent (or more likely slightly delayed) diffusewall building produces the second conidium(conidial ontogeny with slightly delayed conidialmaturation) which is delimited (conidial delimita-tion) and remains attached. This sequence can thenbe repeated (alternation of conidial ontogeny (byring wall buildingj/slightly delayed conidial matu-ration and retrogressive delimitation) and conidia

Fig. 24. Sagenomella striatispora.

may become detached (conidial secession) at anytime.

The development of S. striatispora is obviouslyvery similar to that of B. chlamydosporis. Bothproduce conidia in true chains with the youngest atthe bottom. In both the chains begin by the usualholoblastic production of the first conidium, and inboth chain production eventually occurs as a resultof ring-wall building. The main difference betweenthem is that in S. striatispora the wall-building ringcomes into operation immediately after the firstconidium has been produced, whereas in B.chlamydosporis, the appearance of the wall-buildingring is delayed until after the original and a variablenumber of substitute conidiogenous cells have beenconverted into conidia. Thus, whereas in B .chlamydosporis the original conidiogenous cell (i.e,the one from which the first conidium wasproduced) and the final conidiogenous cell (i.e. theone which has the wall-building ring) are different,in S. striatispora the two are the same. Thedifference in developmental terms between thesetwo species is thus in essence one of the timing ofappearance of the wall-building ring .

Several similar differences in timing have alreadybeen observed in this paper and the work of Minter

D. W . Minter, P. M . Kirk and B. C. Sutton 55

Fig. 25. Development in Sagenomella striatispora.Arrow locates wall-building ring.

et at. (1982). None has been considered to be ofgreat taxonomic significance. There appears noreason why the present case should be unlike theothers . Indeed, like the others, the presentdifference in timing, although minor, has aprofound effect on the appearance of S. striatisporawhen compared with B. chlamydosporis. In B.chlamydosporis retrogression has often travelledbeyond the hyphal branches below the originalconidiogenous cell before the wall-building ringcomes into operation, so that branched chains ofconidia are frequently observed. In S. striatisporaby comparison the wall-building ring comes intooperation at a very early stage, and once it issupplying upwards continuous new wall material,branched chains cannot develop.

In the second part of this paper the observationwas made that deuteromycetes from widely diff-erent parts of the spectrum of developmentalvariation can appear to differ fundamentally whencompared directly with each other, and that the trueaffinities between them can only correctly beassessed when the intergrading examples betweenthem are also taken into consideration. This sameobservation can be made about S. striatispora.When compared with, for example, Belemnosporaepiphylla (Fig. 10), the differences between thetwo appear fundamental ; but after intergradingexamples have been examined, it is evident thatthese apparently major differences are only agradual accumulation of a host of minor differences .Thus a continuous intergradation exists betweenS. striatispora and fungi of the sympodial,annellidic and false- or no-chain phialidiccontinuum.

It is interesting to recall the three ways in whichBasipetospora rubra (the last fungus to be considered

in the first part ofthis paper) was said to differ fromtrue-chain phialidic species. When S. striatispora isexamined with these three ways in mind, it will beseen to differ from true-chain phialidic species inonly one, being the same in respect of the other two:that is, S. striatispora has conidiogenous cells whichcan produce conidia in numbers not limited by theoriginal size of the conidiogenous cells, and it can

.produce them without intervening proliferation,but the enteroblastic mode of development isnowhere present. It is clear that development in S.striatispora is more similar to the true-chainphialide than is development in B. rubra. Develop-ment of the true-chain phialide will now bedescribed and compared with development in S.striatispora, using as an example Aspergillusclauatus.

Aspergillus clavatus

This account of development in A. clavatus (Fig.26) is based on, and interprets, the detailedtransmission electron microscope study by Hanlin(1976). In A. clavatus the first conidium isproduced holoblastically by the activity of the firstwall-building apex (conidial ontogeny) which laysdown in this conidium and at the top of theconidiogenous cell an extra inner wall (Fig . 27,arrow : this wall is shown as double-layered inaccordance with the drawing convention in use here(Minter et al., 1982), but in the transmissionelectron micrographs on which this drawing isbased only one layer can be detected). This newinner wall is not present in the lower part of theconidiogenous cell. The first conidium is thendelimited (conidial delimitation). The first wall-building apex ceases to operate, and is not replaced;instead, a wall-building ring occurs below theseptum delimiting the first conidium. This wall-building ring lays down wall material on the innerwall arrowed in Fig. 27, causing it to grow upwardsand forcing the outer wall between the firstconidium and the conidiogenous cell to break, thusproducing a collarette (Fig. 27, arrowhead) (part ofthe stage of secession of the first conidium). Thefirst conidium still remains attached however by theinner wall. Continued wall production by thewall-building ring then produces the walls of thesecond conidium (conidial ontogeny), which isdelimited (conidial delimitation) and remainsattached. This sequence of stages can be repeated(alternation of conidial ontogeny by ring-wallbuilding and retrogressive conidial delimitation),conidia gradually mature (retrogressive conidialmaturation with associated marked thickening ofdelimiting septa) and may become detached(conidial secession) at any time after maturation.

Development in A . clavatus, as interpreted here,

Thallic phialides

Fig. 26. Aspergillus clavatus.

is very similar to development in S. striatispora.Both produce their first conidium holoblasticallyby the activity of the first wall-building apex, andthen both switch to produce all subsequent conidiaby ring-wall building. Conidia are in true chainswith the youngest at the base, produced fromconidiogenous cells which remain the same length,and in both the original conidiogenous cell (i.e. theone from which the first conidium was produced)is the same as the final conidiogenous cell (i.e. theone with the wall-building ring). Sagenomellastriatispora is thus obviously developmentally moresimilar to A. clavatus than to Belemnospora epiphyllaor any other member of the sympodial, annellidicand false- or no-chain phialidic continuum.

Aspergillus clavatus and S. striatispora do,however, differ in some respects. Firstly, conidial

Fig. 27. Development in Aspergillus clavatus. Arrowlocates new (double-layered) inner wall. Arrowheadlocates collarette.

maturation is slightly later in A. clavatus, and isassociated with some thickening of the delimitingseptal walls. This difference has already beenshown as not significant in other examples and isnot considered significant here either. Secondly, inA. clavatus a new inner wall is produced inside thefirst conidium and at the top of the conidiogenouscell, whereas in S. striatispora no such wall has beenobserved. Thirdly, in A. clavatus the wall-buildingring produces growth of this inner wall only,necessitating partial secession of the first conidium.Furthermore, in A. clavatus walls of second andsubsequent conidia are produced by the wall-building ring operating on this inner wall only, withthe result that there is a discontinuity between thetop of the conidiogenous cell wall and the wall ofthe conidium currently being produced (Fig. 28).In S. striatispora however no such discontinuity canbe observed with the light microscope (Fig. 28)(Subramanian & Pushkaran, 1975). This providesstrong evidence that in S. striatispora the secondand subseqent conidia are produced by action ofthewall-building ring on the original conidiogenous


Fig. 28. Comparison of Aspergillus clauatus (arrowsindicate discontinuity of walls at apices of conidiogenouscells) and Sagenomella striatispora (no discontinuityobservable).

cell wall, that no new inner wall is produced, andthat there is no partial secession of the firstconidium before production of the second.

The difference in development between A.clavatus and S. striatispora thus basically concernsthe presence or absence of a new inner wall. Whenthis inner wall is absent there is no partial secessionof the first conidium, no collarette and nodiscontinuity between the conidiogenous cell walland the wall of the conidial chain: i.e. theenteroblastic mode has no part in the development.When the inner wall is present however there ispartial secession of the first conidium, a collaretteand discontinuity between the conidiogenous cellwall and the wall of the conidial chain: these are allfeatures indicating the presence of the enteroblasticmode in the development. It follows therefore thatthe presence or absence of the new inner walldetermines whether the enteroblastic mode is or isnot present, and hence whether or not the fungushas true-chain phialides.

The presence of a new inner wall in true-chainphialides is, like ring-wall building, a feature whichhas only recently been discovered. It has beenreported with evidence from transmission electronmicrographs only in A. clavatus (Hanlin, 1976), theChalara anamorph of Ceratocystis adiposa (Butl.) C.Moreau (Hawes &Beckett, 1977b)and Thielaviopsisbasicola (Berk. & Br.) Ferraris (Hawes & Beckett,1977c). There are also, however, similar accountsbased on light microscopy of development ofChalara and Thielaviopsis basicola which provide

additional evidence strongly suggestive that thistype of development is more widespread in specieswith true-chain phialides (Hawes & Beckett,1977a, c; Ingold, 1981a); the work of Hawes &Beckett using fluorescent staining techniques alsoprovides further evidence for the existence ofring-wall building. It is very difficult to assesswhether the production of this new inner wallis significant either developmentally ortaxonomically.

What is certain, however, is that the differenceis not necessarily great between development intrue-chain phialides and in fungi with conidia intrue chains produced holoblastically with theyoungest at the top. This has been very clearlydemonstrated in Aspergillus aureolatus Muntafiola-Cvetkovic & Bata and its Cladosarum Yuill & Yuillmutant (Madelin, 1979; Vujicic & Muntafiola-Cvetkovic, 1973) and a wide variety of othertrue-chain-phialidic species and their mutants inthese and similar genera (Clutterbuck, 1969; Raper& Fennell, 1953; Yuill & Yuill, 1938; Zachariah &Metitiri, 1970, 1971). The case of A. aureolatus andits Cladosarum mutant will be discussed here as anexample.

Transmission electron micrographs publishedby Vujicic & Muntafiola-Cvetkovic (1973) showthat in normal growth A. aureolatus producestrue-chain phialides (Fig. 29). The mutant,however, when grown at 26°C, produces conidia intrue chains with the youngest conidium at the top(Fig. 29). These conidia remain thin-walled unlessthe temperature is lowered to 18°, when theterminal conidium in each chain becomes thickwalled and indistinguishable from conidia of thewild type. Madelin (1979) discussed variouspossible explanations for these observations, andconcluded that the unusual development of themutant is a result of its failure to produce beneathconidia delimiting septa characteristic of the wildtype. While this explanation is accepted as ofsignificance, it is also suggested that the differencein appearance between mutant and wild type can beexplained much more satisfactorily in terms of wallbuilding. In the wild type, ring-wall buildingcomes into operation immediately after the firstconidium is produced, and diffuse wall buildingcauses conidial maturation. In the mutant, the firstwall-building apex remains active and produces anapically elongating conidial chain in which much ofthe diffuse wall building is suppressed.

If this interpretation is correct, at least somefungi with true-chain phialides can be made to keeptheir apical wall building operational instead ofswitching to ring-wall building by mutation of asmall number of genes (Madelin, 1979) or evensimply by manipulation of temperature. Since the

Fig. 29. Aspergillus aureolatus. A. Normaldevelopment oftrue-chain phialides. B. Mutant development at 26°Cproducing true chains of thin-walled conidia with theyoungest spore at the apex. C. Mutant developmentinitiallyat 26°C, laterat 18 °C: theapicalsporeofthe truechain becomes thick walled, and indistinguishable fromconidiaof the wild type.

PART 4

Geotrichum candidum

This account of development in G. candidum Link(Fig. 30) is based on the time-lapse study by Cole& Kendrick (1969). In G. candidum the conidiogen-ous cell is produced by the activity of the originalwall-building apex (Fig. 31), incorporating theconidial ontogeny stages of all conidia. Thiswall-building apex then ceases to function and isnot replaced, so that at this point the structure hasthe appearance of a hypha which has ceased togrow. Conidial delimitation then takes place by theformation of septa in an apparently random order.This delimitation is not restricted to the apicalconidiogenous cell, but can be observed in all partsof the hypha so that, as in B. oariabilis, branchedchains of conidia occur. There is no conidial

Thallic phialides

C. Moreau (Nag Raj & Kendrick, 1975), this beliefbecomes considerably substantiated. It is thereforeconcluded that continuous intergradation existsbetween development in, for example, Bas ipeto-spora rubra and species with true-chain phialides.It follows from this and from the conclusion of thesecond part of this paper that continuous inter-gradation also exists between fungi with true-chainphialides and species with development which isgenerally described as thallic. In the fourth part ofthis paper some other relevant examples will beexamined to show how this intergradation is notmerely based on the few fungi examined so far .

58

presence of ring-wall building in true-chainphialides has been correlated in this paper with thepresence of a new inner wall, it seems highly likelythat, if a switch from producing to not producinga wall-building ring can easily be made, a similarswitch to not producing a new inner wall could bemade with the same facility. Madelin (1979) madethe following observation about such switches: 'iffungi in nature occasionally make at least some ofthe changes between these ... modes, the taxonomicsignificance to be attached to their possessiondiminishes'. This is strongly supported, and it isbelieved that the presence of a new inner wall in thetrue chain phialide is not necessarily any greater adifference developmentally than any other differ-ence already observed between adjacent membersof a continuum. When it is remembered thatSagenomel/a striatispora can produce its conidia intrue chains by holoblastic ring-wall building (asdescribed above) or by enteroblastic ring-wallbuilding (i.e, as a true-chain phialide) (Subraman-ian & Pushkaran, 1975), and so can the Thielaviopsisanamorph of Ceratocystis paradoxa (Dade) Fig. 30. Geotrichum candidum.


Fig. 32. Chalara hughesii.

genous cell, until often more than half of theoriginal conidiogenous cell wall is lined by this newinner wall. This new inner wall, although stilltotally enclosed by the original outer wall of theconidiogenous cell, is destined to be the cell wallsof the first conidia (conidial ontogeny) ; accordingly,as it is laid down, retrogressive conidial delimitationfollows, and conidia can be discerned formingwithin the unbroken original outer wall of theconidiogenous cell. After a while the retrogressivelaying down of the new inner wall comes to a halt,and at its lowest part a wall-building ring is formed,which then adds continuously to it in an upwarddirection (conidial ontogeny). At this point theoriginal outer wall of the conidiogenous cellruptures almost at the apex, leaving on theuppermost conidium a small cap of wall materialsometimes visible with the light microscope. Theactivity of the wall-building ring causes the chainof conidia (marked off by continuous retrogressive

~o%oon

Fig . 31. Development in Geotrichum candidum.

Chalara

Developmental studies of a variety of species ofChalara have been made (Hawes & Beckett,1977a, b, c; Ingold, 1981a; Nag Raj & Kendrick,1975), and these collectively form the basis of thefollowing account . In species of Chalara (Fig . 32)the conidiogenous cell is produced by the activity ofthe original wall-building apex (Fig. 33). When thiswall-building apex ceases to function, a new innerwall (double-layered in accordance with thedrawing convention in use (Minter et al., 1982)) isimmediately laid down, as in Aspergillus claoatus,at the top of the conidiogenous cell. Unlike in A.claoatus, however, this inner wall is extensive,being produced not just at the apex, but alsoretrogressively well down the sides of the conidio-

maturation, and conidial secession can happen atany time after conidial delimitation.

Production of conidia by G. candidum is thussimilar to that by Oidiodendron truncatum (thethallic species examined in the second part of thispaper), except for the absence of conidial matura-tion and randomness of conidial delimitation. Sinceall major treatments in recent years have agreed thatboth of these species have thallic development(Cole & Samson, 1979; Kendrick, 1971; Sigler &Carmichael, 1976), it seems reasonable to concludethat neither of these exceptions is of significance indetermining whether or not a fungus has thallicdevelopment. As has already been stated, it is notintended in this paper to undertake a detaileddiscussion of what constitutes the fundamentaldistinction between blastic and thallic modes ofdevelopment. It is nevertheless interesting tocompare this supposedly thallic development of G.candidum with the development of species ofChalara (Cda) Rabenh. which recent major treat-ments have described as phialidic (Cole & Samson,1979; Nag Raj & Kendrick, 1975).


I I

..

Fig . 33. Development in Chalara sp . Straight black arrow locates wall-building ring. White arrow indicatesconidial delimitation. Curved black arrow locates cap.

delimitation) to be pushed out of the cylindricalcollarette formed by the original outer wall. Afterleaving the collarette, a small degree of conidialmaturation may be observed in some species, butis normally absent, and conidial secession can occurat any time, sometimes even before emergence ofthe conidium from the collarette.

From the foregoing account development ofChalara species is fundamentally similar to that ofA. clauatus . The main distinctions are that inspecies of Chalara conidial maturation is usuallyabsent, and more inner wall is laid down before theproduction of a wall-building ring, so thatcollarettes are much larger : both of these are merelydifferences of degree. What is interesting, however,is the striking similarity in appearance betweenconidia of Chalara and G. candidum and also, apartfrom the branching in one, between conidial chainsof these two taxa (compare fig. 4.25 with fig. 7.8 inCole & Samson (1979» . The similarity occursbecause conidia are formed by the same cocktail ofdevelopmental stages; in both a replacementwall-building apex is lacking, conidial maturationis absent, and conidia can only be recognized afterthe activity of conidial delimitation on a cylindricalregion of fertile hypha. This similarity in appear-ance, of course, no more necessarily indicates thatthe two taxa are closely related phylogenetically

than did the dissimilarity between Cladobotryumvarium and Trichothecium roseum indicate theopposite. What this similarity does indicate, how-ever, is that something is wrong with the presentrigid practice of categorizing fungi as either thallicor blastic : because this practice, while purportingto be based on developmental principles, fails torecognize the considerable number of way in whichthese two taxa are developmentally analogous.Geotrichum candidum does, certainly, differ fromspecies of Chalara in a number ofhighly significantdevelopmental features. Examination of the nexttwo relevant examples shows, however, howunsatisfactory it would be to try to use suchdifferences to distinguish the thallic and blasticmodes of development.

Wallemia sebi and Vouauxiella lichenicola

This account of development in W . sebi (Fr.) Arx(Fig. 34) is based on the interpretation by Cole &Samson (1979) ; the detailed electron microscopestudies by Madelin &Dorabjee (1974) are also takeninto account. In W . sebi the conidiogenous cell isproduced by the activity of the first wall-buildingapex (Fig. 35). The first wall-building apex thenceases to function, and a new inner wall isimmediately laid down at the top of the conidio-

61D. W . M inter, P. M. Kirk and B. C. Sutton

oo

Fig. 34. Wallemia sebi.Fig. 35. Development in Wallemia sebi. White arrowlocates collarette. Black arrow locates wall-building ring .

genous cell by a new wall-building apex. This newinner wall then grows up through the original outerwall, which ruptures, resulting in a collarette. Thenew inner wall is deduced from the presence of thecollarette (the validity of thi s deduction will beassessed in the discussion) and from the abrupttransition at this point between rough and smoothornamentation. A small amount of original outerwall material may be carried away as a cap by theupward-growing inner wall. After a short time thenew wall-building apex ceases to function and isreplaced by a wall-building ring around the insideof this new inner wall a little above the collarette.The activity of this wall-building ring nowprovides the stage of conidial ontogeny. After newwall material sufficient for several conidia has beenproduced by the wall-building ring, conidialdelimitation begins in a regular but not strictlyretrogressive order. During and following conidialdelimitation, a small amount of diffuse wallbuilding (conidial maturation) occurs and, afterthat, conidial secession may occur at any time .When several conidia have been produced in thisfashion, the wall-building ring may cease tofunction, and may be replaced by a new wall-building apex (sympodial or percurrent prolifer-ation). After some growth, the resultant elongatedconidiogenous cell usually produces a new innerwall again, and the sequence of developmentalstages is repeated, producing more conidia. Thedevelopment of V. lichenicola (Linds.) Petrak &Sydow, to judge from the appearance of its conidio-genous cells and conidia, is an example in the

coelomycetes of a similar type of development tothat of W. sebi (Morgan-Jones, 1971; Sutton,1980).

Wallemia sebiand V.lichenicola have both in thepast been classified as thallic (Ellis, 1971; Sutton,1980), but it has subsequently become obvious toa number of researchers that this is far fromsatisfactory. At least in the case of W. sebireclassification close to phialidic species such asSporoschisma Berk . & Br. (a genus similar indevelopment to Chalara) has been proposed(Ma delin & Dorabjee, 1974). The uncertaintysurrounding the taxonomic positions of bothspecies clearly exists because each has certaindevelopmental features which resemble those offungi conventionally regarded as thallic, and otherswhich resemble those of fungi conventionallyregarded as phialidic. Thus, like the thallic speciesOidiodendron truncatum, conidial maturation ismostly delayed until after conidial delimitation,which itself follows conidial ontogeny irregularlyand only after an interval; similarly, like thetrue-chain-phialidic species Aspergillus claoatus,they produce a wall-building ring and have at somepoint in their development the enteroblastic mode.In other words, both W . sebi and V . lichenicola havea combination of developmental features whichmakes them intermediate between thallic andtrue-chain-phialidic species. They are thus fungiwith , in the words of the title of this paper, thallicphialides.


Fig. 37. False chain of conidia with no connective.

Fig. 36. True chain of conidia with connective joiningupper conidia.

Fig. 38. False chain of conidia with connective joiningupper conidia.

3837

oo

36

The actual words 'connected' and 'disconnected'are misleading, however: some true chains, e.g.those of Chalara species (Fig. 39), have adjacentconidia with firm wall connexions but withoutconnectives, and in these cases the highly specificuse of words like 'connected' would be confusingbecause such words are also used in a very generalsense in the English language. It is proposed thatthese words are replaced with the more descriptiveand precise terms 'with connectives' and 'withoutconnectives' respectively.

It is important to note that these terms describemorphological features only. It is therefore impos-sible to deduce the way in which a chain hasdeveloped solely from the combination of termsused to describe the appearance of the finishedchain. Thus false chains without connectives (Fig.37) could be produced by a false- or no-chainphialide (e.g. Mariannaea elegans(Cda) Samson) orby an annellidic conidiogenous cell (e.g. inScopulariopsis brevicaulis (Sacc.) Bain.), Similarlytrue chains without connectives could be producedby ring-wall building (e.g. Chalara species), byreplacement wall-building apices (e.g. Triche-thecium roseum) or by simple retrogressivedelimitation (e.g. Basipetospora rubra). Many moreexamples could be cited.

DISCUSSION

Conidial chains

The terms 'true' and 'false', used here and byMinter et al. (1982) to describe chains of conidia,were originally proposed by Subramanian (1972),who defined them on the basis that in true chains, the wall around the successive conidia in the chainis a continuum', whereas in false chains it is not.Gams (1978) rejected Subramanian's terminology,but made a similar attempt to distinguish between'connected' and 'disconnected' chains which hedefined by the presence or absence of connectives(i.e. secondary thickening of septal walls duringconidial maturation). These two sets of terms,'true' and 'false', 'connected' and 'disconnected',are now evaluated and compared.

Aspergillus clauatus has chains of conidia whichadhere by connectives, with a continuous outer wallbetween adjacent conidia (Fig. 36). Such chains areobviously' true' chains in the sense of Subramanian(1972) and, although Gams (1978) did notrecognize chain development of this type, it seemsreasonable to suppose on the basis of the exampleshe used that such chains would fall into his'connected' category. Similarly the fungi examinedby Minter et al. (1982) which have conidia notadhering by any wall connexions (Fig. 37) clearlyfall into Subramanian's 'false' and Gams' 'discon-nected' categories. There is therefore a generalcorrespondence between the different ways ofclassifying chains. It is not however a completecorrespondence, as the following example shows.

Gams (1978) illustrated a third type of conidialchain produced by a phialide, shown here using thetransmission electron microscope drawing conven-tion in Fig. 38. There is no continuity ofthe outerwall between adjacent conidia, but there iscontinuity between the inner wall of the older andthe outer wall of the younger of the adjacentconidia. Such development, in this example, alsoforms a chain in which there are connectives, butthey are in the form of an incomplete wallcontinuity between adjacent conidia. Such a chainis 'connected' in Gams' sense, because his chaincategories are defined by the presence or absenceof connectives. It is, however, 'false' inSubramanian's sense, because the outer wall is notcontinuous between two adjacent conidia.

There are thus at least three categories of chainsto describe (Figs 36, 37, 38), and since there areonly two sets of terms, each containing two words,it is obvious that one set alone will not be enoughto describe the three categories, especially sinceeach set covers ground which although similar isnot identical. It is believed in this paper that thereis good reason to retain both sets at least in concept.

Fig. 39. Collarette production(after Cole & Samson,1979)·

Cellarettes

The manner in which collarettes are produced hasbeen explained in Minter et al. (1982) and thispaper by two different theories. The first explana-tion , attributed by Minter et al. (1982) to Cole &Samson (1979) is illustrated in Fig. 39. Accordingto this theory, the cellarette in anatomical terms ispart of the outer wall of the first conidium, and thisouter wall is left behind, still attached to theconidiogenous cell, when that conidium secedes. Ifthis is correct, then the first conidium, over muchof its surface, has as an integument only its originalinner wall. The second explanation, used here, isshown in Fig. 40. According to this explanation thecollarette is composed of the same inner and outerwalls as the conidiogenous cell, and the firstconidium has inner and outer walls both of whichare totally new and not the same as any wall of thecellarette.

The differences between these two explanationsmay appear at first sight minute and trivial, butcareful reflection shows that they lead to profoundlydifferent conclusions. If the first explanation iscorrect, then the collarette is merely a by-productof conidial ontogeny, and the first conidium to beproduced is holoblastic. In developmental terms,therefore, as Minter et al. (1982) pointed out, thisconidium would not differ from any subsequentlyproduced conidium. If the second explanation iscorrect, however, the collarette is not merely aby-product of conidial ontogeny. Instead it is theresult of wall-building activity actually precedingconidial ontogeny of the first conidium. The firstconidium is accordingly produced enterobiasticallyand thus differs in developmental terms fromsucceeding conidia which have holoblastic conidialontogeny.

It is, ofcourse, possible that among the vast arrayof deuteromycetes examples will be found which

Periclinai thickening

Phialides producing false chains with connectivesdiffer from those producing true chains withconnectives because they lack a wall-building ringand there is no outer-wall continuity betweenadjacent conidia. It is often, however, not possiblereliably to tell with the light microscope whetheror not outer-wall continuity is present, and so incertain critical cases the problem of determiningwhether a phialidic conidial chain with connectivesis true or false depends on telling whether or nota wall-building ring is present. In theory this canbe done by looking for periclinal thickening(Sutton, 1980): its presence indicating a successionof replacement wall-building apices and its absenceindicating a wall-building ring. Unfortunately, inpractice a certain degree of periclinal thickening canoften be observed in fungi such as A. clauatus,which are firmly established as having true chainswith connectives. It is not known why, but it ispossible to speculate. For example, it may perhapsbe that conidial production in such fungi has amarked 24 h periodicity, conidia being producedduring the day and not during the night. Ifthis werethe case, it might well be necessary for the fungusto produce a replacementapex and then immediatelya new wall-building ring each day as conidialproduction recommences. Periclinal thickeningwould then build up over several days. This sort ofphenomenon is unlikely to have been observed,however, by researchers carrying out transmissionelectron microscope studies, simply because thefungal cultures used in such studies tend (asMadelin (1979) pointed out) to be grown in highly 'artificial conditions which may make impossiblethose pauses in development which might haveoccurred under natural conditions. Until furthercriteria become available the distinction betweentrue and false chains with connectives (whileadmittedly of great value) should be applied withcaution.

D. W. Minter, P. M. Kirk and B. C. SuttonBy the same token, it is also impossible to deduce

the final appearance of a chain solely from thecombination of terms used to describe its mode ofdevelopment. Thus, for example, although acombination of retrogressive delimitation withreplacement of wall-building apices can produce atrue chain without connectives, e.g. Trichotheciumroseum, it can equally well produce no chain at all,e.g. Cladobotrym uarium. A full description of anydeuteromycetc should therefore contain not onlydetails of its mode of development but also,distinguished from these, an accurate account of itsactual physical appearance.

Thallic phialides

Fig. 40. An alternative explanation ofcollarette production.

are appropriate for each explanation. Unfortuna-tely, however, both of these explanations purportto be based on the same evidence (contained intransmission electron microscope studies of, forexample, Trichoderma saturnisporum Hammill(Hammill, 1974) and A. clauatus (Hanlin, 1976));and where both purport to explain development ofthe same species, based on the same evidence, at leastone of them must be wrong. Which one becomesclear when the original publications on which thetheories are based are re-examined. Hanlin (1976)stated that, 'as the phialide neared maturity theapex became narrower .. .the phialide wall wasuniformly thin, but became much thicker at theapex due to the formation of a secondary wall layerinside the outer primary wall '. Similarly, Hammill(1974) observed, 'when a phialide produced its firstconidium ... the first conidial initial was positionedinside of the phialide wall, while the phialide walloutside of it was disintegrating and presumablybeing sloughed off. The minute collarette of aphialide was that portion of the original phialidewall which remained after completion of thedisintegration and sloughing-off processes'. Whenthese descriptions are compared with the twoexplanations, there can be no doubt that the secondtheory (Fig. 40) fits the facts more closely than doesthe first (Fig. 39). It is therefore concluded thatcollarette production is not merely part of theontogeny of the first conidium, but must beregarded as a totally independent stage.

Although this conclusion does not affect the truthof the observation that false- or no- chain phial idesintergrade with annellides (Mi nter ec al., 1982), itis interesting to note that collarettes tend to bepresent in the former and absent in the latter. Thissuggests that there is some evolutionary advantagein having collarettes for those fungi which producea large number of conidia from the same or almost

the same locus. The fact that collarettes are alsogenerally present in true-chain phialides supportsthis observation, although what this evolutionaryadvantage is can, at present, only be a subject ofspeculation. Certainly the apparatus of a wall-building ring or a succession of replacementwall-building apices in a phialide must be ratherdelicate, and perhaps the collarette is a means ofproviding such a fragile but important mechanismwith some degree of protection.

If collarettes have a protective role and areproduced as a 'one-off' item before phialidic sporeproduction commences, it is interesting to speculateon how they came about in the first place . Theirstructure, and particularly the fact that the firstconidium is produced by totally new inner walls, isstrongly reminiscent of the 'cups' produced byaborted conidia, and the subsequent regenerationobserved in Endophragmiella Sutton by Hughes(1979). If this similarity is significant, a theorycould be advanced that collarettes are the same as'cups', i.e, in evolutionary terms, collarettes are nomore than aborted conidia, and that naturalselection in phialidic fungi has generally favouredthe abortion of a first conidium and consequentialregeneration, because this is an easy way to protectthese delicate phialidic mechanisms.

This sort offeature could well have evolved morethan once, and so there seems to be no need toregard the presence of a collarette as evidence thatphialides are monophyletic. In fact, although it hasbeen suggested on a number of occasions (e.g.Madelin, 1979) that the phialide represents aunique line ofevolution, the evidence considered inthis paper indicates that this is hardly likely :phialides resulting from a succession of replacementwall-building apices have surely come down adifferent evolutionary line from those with wall-building rings . Indeed if these two (or more) linesare compared, it is even possible to suggest that oneis more recent than the other, since to replace awall-building apex which has been lost is primitivein comparison to the sophistication of producingthe wall-building ring, a mechanism which neverneeds replacing.

CONCLUSIONS

The arguments presented in the four parts of thispaper have made it abundantly clear that inter-gradation exists developmentally between thallicand blastic fungi in general and, in particular, be-tween true- and false- or no-chain phialidic fungiand between thallic and true-chain-phialidic fungi.Representatives of these groups are therefore indevelopmental terms merely members of one largecontinuum. The existence of such a continuum has

D. W. Minter, P. M. Kirk and B. C. Suttonlong been known (e.g, Kendrick, 1971), butthe lackof a suitable vocabulary to describe it has made itdifficult for taxonomists to act upon this knowledge.The result has been that deuteromycetes havecontinued to be classified according to ill-definedgroups long after the limitations of these groupshave been acknowledged, simply because no viablealternative has been offered.

This paper and its predecessor (Minter et al.,1982) have attempted to generate some of thevocabulary needed if such a viable alternative is tocome about. Existing terminology has beenrationalized and, where necessary, supplementedwith new but easily understood terms. Thus it isnow possible to recognize seven developmentalstages (conidial ontogeny, conidial maturation,conidial delimitation, conidial secession, prolifera-tion, regeneration and collarette production) whereformerly only one, conidial ontogeny, was incommon use. The terms 'holoblastic' and 'enter-oblastic' have been redefined, and the terminologyfor describing conidial chains has been put on arational foundation. Similarly the confusing term'meristem' has been discarded and in its place threetypes of wall building have been recognized.

Not all difficulties have been solved, however. Ifanything, these papers have only revealed morelinguistic problems which are holding backdeuteromycete systematics. It is now obvious, forexample, that the terms 'phialidic' and 'thallic'under their present definitions contain inherentconfusions and contradictions, and if they are tocontinue in use in any meaningful sense they needto be subjected to thorough scrutiny. Similarly, theproposed reforms have not yet been applied to allknown types ofdevelopment in the deuteromycetes,but this must be done to see if there too they reallywork. In particular, conidia produced in acropetalchains and conidia produced by so-called' basauxic'conidiophores need to be re-examined carefully.

It is hoped that the ideas put forward here, forall their acknowledged limitations, will stimulatemycologists to realize that current terminology canbe improved, and that if this is done, perhaps someviable alternative method of classifying these fungiwill be discovered. One day terms like 'phialidic'and 'thallic' may even be viewed in the same wayas 'phlegmatic' and' mercurial', which the chemistnow regards as belonging in the out-of-date worldof alchemy.

Colleagues at C.M.I. and elsewhere are thankedfor their comments, opinions and lively discussions,without which this paper could not have beenwritten. Professor C. V. Subramanian is thankedfor hospitality at the Centre for Advanced Studiesin Botany, University of Madras, where on a recent

3

visit the senior author prepared some of thismanuscript. Most of all, the senior author thankshis wife Helen for her continued patient support,encouragement and perceptive observations.

REFERENCES

CLUTTERBUCK, A. J. (1969). A mutational analysis ofconidial development in Aspergillus nidulans. Genetics63,3 17-327.

COLE, G. T. & KENDRIcK, W. B. (1969). Conidiumontogeny in hyphomycetes. The arthrospores of Oidia-dendron and Geotrichum and the endoarthrospores ofSporendonema. Canadian Journal of Botany 47,1773-1780.

COLE, G. T. & SAMSON, R. A. (1979). Patterns ofDevelopment in Conidial Fungi. London, San Francisco,Melbourne: Pitman.

ELLIS, M ..B. (1971). Dematiaceous Hyphomycetes. Kew:Commonwealth Mycological Institute.

ELLIS, M. B. (1976). More Dematiaceous Hyphomycetes.Kew: Commonwealth Mycological Institute.

GAMS, W. (1978). Connected and disconnected chains ofphialoconidia and Sagenomella gen. nov. segregatedfrom Acremonium. Persoonia 10, 97-112.

HAMMILL, T. M. (1974). Electron microscopy of phialidesand conidiogenesis in Trichoderma saturnisporum.American Journal of Botany 61, 15-24.

HANLIN, R. T. (1976). Phialide and conidium develop-ment in Aspergillus claoatus. American Journal ofBotany 63, 144-155.

HAWES, C. R. & BECKETT, A. (1977a). Conidiumontogeny in the Chalara state of Ceratocystis adiposa.1. Light microscopy. Transactions of the British Myco-logical Society 68, 259-265.

HAWES, C. R. & BECKETT, A. (1977b). Conidiumontogeny in the Chalara state ofCeratocystis adiposa. II.Electron microscopy. Transactions of the BritishMycological Society 68, 267-276.

HAWES, C. R. & BECKETT, A. (1977c). Conidiumontogeny in Thielaviopsis basicola. Transactions of theBritish Mycological Society 68, 304-307.

HUGHES, S. J. (1979). Relocation of species of Endo-phragmia auct. with notes on relevant generic names.New Zealand Journal of Botany 17, 139-188.

INGOLD, C. T. (1981 a). The first-formed phialoconidiumof Thielaoiopsis basicola. Transactions of the BritishMycological Society 76, 517-519.

INGOLD, C. T. (1981b). The validity of the concept ofconidia as either blastic or thallic. Transactions of theBritish Mycological Society 77, 194-196.

KENDRICK, W. B. (1971). Taxonomy of Fungi lmperfecti.Toronto and Buffalo: University of Toronto Press.

MAnELIN, M. F. (1979). An appraisal of the taxonomicsignificance of some different modes of producingblastic conidia. In The Whole Fungus, vol. 1 (ed.W. B. Kendrick), pp. 63-80. Ottawa: NationalMuseum of Natural Sciences, National Museums ofCanada and the Kananaskis Foundation.

MAnELIN, M. F. & DORABJEE, S. (1974). Conidiumontogeny in Wallemia sebi. Transactions of the BritishMycological Society 63, 121-130.

MYC 80


MATSUSHIMA, T. (1975). leones Microfungorum a Matsu-shima lectorum, Published by the author, Kobe, Japan.

MINTER, D. W., KIRK, P. M. & SUTTON, B. C. (1982).Holoblastic phialides. Transactions of the BritishMycological Society 79, 75-93.

MORGAN-JONES, G. (1971). Conidium ontogeny inCoelomycetes. III. Meristem thalloconidia. CanadianJournal of Botany 49, 1939-1940.

NAG RAJ, T. R. & KENDRICK, W. B. (1975). A Monographof Chalara and Allied Genera. Waterloo, Ontario:Wilfred Laurier University Press.

RAPER, K. B. & FENNELL, D. I. (1953). Heterokaryosis inAspergillus. Journal of the Elisha Mitchell ScientificSociety 69, 1-29·

SIGLER, L. & CARMICHAEL, J. W. (1976). Taxonomy ofMalbranchea and some other hyphomycetes witharthroconidia. Mycotaxon 4, 349-488.

SUBRAMANIAN, C. V. (1972). Conidial chains, their natureand significance in the taxonomy of Hyphomycetes.Current Science 41, 43-49.

SUBRAMANIAN, C. V. & PuSHKARAN, M. (1975). Conidium

ontogeny in some monophialidic Hyphomycetes.Kavaka 3, 77-99·

SUTTON, B. C. (1980). The Coelomycetes. Kew, Com-monwealth Mycological Institute.

VUJICIC, R. & MUNTA:llrOLA-CVETKOVIC, M. (1973). Acomparative ultra-structural study on conidium dif-ferentiation in the Cladosarum-like mutant 22B ofAspergillus aureolatus. Journal of General Microbiology79,45-51.

YUILL, E. & YUILL, J. L. (1938). Cladosarum olivacearum,a new Hyphomycete. Transactions of the BritishMycological Society 22, 194-200.

ZACHARIAH, K. & METlTlRI, P. O. (1970). The effect ofmutation on cell proliferation and nuclear behaviour inPenicillium claoiforme Bainier. Protoplasma 69, 331-339·

ZACHARIAH, K. & METlTlRI, P. O. (1971). The organiza-tion of the penicillus of Penicillium claviforme Bainier.In Taxonomy of Fungi lmperfecti (ed. W. B. Kendrick),pp. 120-129. Toronto and Buffalo: University ofToronto Press.

(Received/or publication 28 January 1982)

Documents

Thallic phialides - David Moore's World of Fungi: where ...€¦ · Radical views on development in some conidial fungi were expressed by Minter, Kirk & Sutton (1982).Thepresentpapercontinuesthe