STAMEN DEVELOPMENT
Lfo^ n Promotoren: dr.M.T.M .Willems e hoogleraari nd eplantkund e
dr.J.L .va nWen t hoogleraari nd eplantkund e tJfiJO??o\l (0%
C.J.Keijzer
STAMEN DEVELOPMENT
Proefschrift ter verkrijgingva n de graad van doctor in deLandbouwwetenschappen , op gezag van de rector magnificus, dr. C.C.Oosterlee , inhe t openbaar te verdedigen opvrijda g 17oktobe r 1986 des namiddags tevie r uur in de aula van deLandbouwuniversitei tt e Wageningen
sW l^lb^l tAiliiisOi: .VKOGESCHOOL WAGENINGIM <• /^N0P2Q\, 10^^
Stellingen
1. Voor het anthere-openingsproces in diverse plantesoorten dienen de 0- vormig verdikte endotheciumcelwanden de loculus niet alleen open, doch in een vroegere fase ook dicht te buigen. Dit proefschrift.
2. Het is onjuist te veronderstellen dat recente wetenschappelijke literatuur de kennis van alle voorafgaande omvat. Bij herhaald refereren blijkt informatie verloren te gaan. Dit proefschrift.
3. De vraag vanuit de veredelingspraktijk naar mannelijke steriliteit in diverse gewassen rechtvaardigt een intensivering van het onderzoek naar de belnvloedbaarheid van het anthere-openingsproces. Dit proefschrift.
4. Ten onrechte zijn de bij micro- en macrogametogenese aan elkaar grenzende genetisch identieke haploide cellen tot voor kort niet met een snelle productie van inteeltlijnen in verband gebracht. C.J. Keijzer (1984). Landbouwkundig Tijdschrift 7/8: 21-26.
5. Mede door het gebruik de ogen te sluiten wanneer men de neus in een bos bloemen steekt, zijn diverse eenvoudig waarneembare fasen van het plantaardige voortplantingsproces slechts bij een relatief klein publiek bekend.
6. De mate waarin de Wageningse promovendus erin slaagt in het kwartier voorafgaand aan de openbare verdediging van haar of zijn proefschrift de inhoud ervan aan een ondeskundig publiek duidelijk te maken, dient mee te wegen in de beoordeling.
7. Wie een Van Leeuwenhoek-microscoop namaakt volgens de reconstructie van Walter en Via, kan onvoldoende nagaan waarom onze Delftenaar een betere kijk op de zaken had dan zijn tijdgenoten. W.G. Walter and H. Via (1968). The American Biology Teacher 8: 537-539. 8.Doo rbi jd etenorbanjo ,naas td ereed sbekend eaanpassingen ,d esnaar -e n velspanning tebenutte no m deklankbode m tegen dekete lt e spannen, ontstaat een instrument dat zelfs de koperblazers van beginnende Dixieland-orkestjeska noverstemmen .
9.Voo rhe toptimalisere nva nd esnelwandeltechnie kka nee nanalys eva n zoweld ebiped eal squadruped evoortbewegin gva nsommig eamfibiee ne n reptielenva nbelan gzijn . R.McNeil lAlexande r (1975).Th eChordates .Cambridg eUniversit yPress .
10.D einvoerin gva nelectronisch ehulpmiddele nvoo rscheidsrechterlijk e beslissingenkom td etennisspor tal sspektake lnie tte ngoede . F.L.M.Vossenaa re nC.J .Keijze r (1983)in :D eideee nligge no pstraat . J.W.Thompson ,Amsterdam ,p.37 .
11.D e gewoonte vanhe tAN P om eenkruispun t tegenwoordigal sknooppun tt e betitelen/ maakthe tverde roverbodi gme et edele nda td ezaa kdaa rwee r eensvas tzit .
12.He ti see nbeleef dgebaa rva nhe tIO Co mbelangrijk ebeslissinge npa st e nemennada tee ndaari ngeinteresseer dpromovendu sd everdedigin gva nzij n proefschrifti nall erus theef tkunne nvoorbereiden .
Wageningen, 17oktobe r198 6 C.J.Keijze r In herinnering aan mijn vader Dankwoord
Gaarne maak ik van de gelegenheid gebruik on eenieder te bedanken die bijgedragen heeft aan het tot stand komen van dit proefschrift. Vrijwel alle medewerkers van de vakgroep Plantencytologie en -morfologie werden meerdere malen geraadpleegd om mij met hun uiteenlopende kundigheden advies of daadwerkeli jke hulp te geven. Ik wil hen alien vanaf deze bladzijde bedanken, slechts een momentopname uit een voortdurende samenwerking die ook in de toekomst de voornaamste pijler van een enthousiaste en productieve vakgroep dient te zijn. Slechts enkelen treft het lot hier bijzonderheden over hun relatie met ondergetekende gepubliceerd te zien. Hooggeleerde Willemse, mijn promotor, ik dank U allereerst voor het mij aanbieden van de betrekking waardoor dit proefschrift mogelijk werd. De verplichte contacten tussen een bioloog en een plantenveredelaar werden, na een gewenningsproces, een uiterst leerzame periode. Dankzij dit fundamentele proefschrift is een meer toegepast vervolg nog slechts een kwestie van fantasie gebruiken en invullen. Hooggeleerde Van Went, mijn co-promotor, ik dank U voor de mogelijkheid die U mij bood om enige tijd in Italie te kunnen werken. Ook waren de vele uren op Uw kamer tijdens de afifonding van dit proefschrift onmisbaar. Carmen Reinders, jouw - nog relatief korte - medewerking betrof de moeilijkste aspecten van ons onderzoek. Dankzij je enthousiasme en doorzettingsvermogen sloeg je je door vele negatieve, bruine resultaten van in-vitro experimenten heen. Het is geen toeval dat onze samenwerking pas in het laatste hoofdstuk naar voren komt. Het geeft daarmee het accent voor de toekomst aan. Professor Cresti, il loro modo di lavorare e organizzare e stato piu istruttivo per me di quello que pensavo possibile in tre mesi. Anche il loro corso "Palio per Olandesi e Americani" era benissimo. Dr. ir. H.J. Wilms, His, onze toegenomen samenwerking in zowel onderzoek als onderwijs heeft ons vakgebied voor de directe oplossing van landbouwkundige problemen toegankelijker gemaakt. Ir. A.H. Bouma ben ik erkentelijk omdat hij zonder aarzeling beaamde dat de aanwezigheid van een deskundig oud-01 tijdens de openbare verdediging van een proefschrift onontbeerlijk is. Contents
Introduction 1
1.Hydratio n changes during anther development 3 C.J. Keijzer (1983). In: Pollen, biology and implications for plant breeding,edite d by D.L.Mulcah y and E.Ottaviano ,pp .197 - 201. Elsevier Biomedical,Ne w York, (changed).
2. Theprocesse s of anther dehiscence andpolle n dispersal. 1. The opening mechanism of longitudinally dehiscing anthers 9 C.J. Keijzer (1986). TheNe w Phytologist (inpress) .
3.Th eprocesse s of anther dehiscence andpolle n dispersal. 2. The formation and the transfer mechanism ofpollenkitt , cell wall development inth e locule tissues and a function of the orbicules inpolle n dispersal 31 C.J. Keijzer (1986).Th eNe w Phytologist (in press).
4. A comparison of anther development inmal e sterileAlo e vera and male fertileAlo e ciliaris 49 /, C.J. Keijzer andM . Cresti (1986). Annals of Botany (in press).
5.Functiona l morphological relationships inth e developing locule of Gasteria verrucosa 63 C.J. Keijzer and M.T.M. Willemse (1987). Submitted for publication.
6. The filament development of Gasteria verrucosa 89 C.J. Keijzer, I.H.S.Hoe k andM.T.M . Willemse (1987). Submitted forpublication .
7. Theprocesse s of anther dehiscence andpolle n dispersal. 3. The dehydration of the filament tip and the anther in some monocotylous species 107 C.J. Keijzer, I.H.S. Hoek andM.T.M .Willems e (1987). Submitted forpublication . 8.I n vitro culturean dintraspecifi ctransplantatio n of tetrads intoolde ranther s 123 M.C.Reinder san dC.J .Keijzer .
Generalconclusion san ddiscussio n 135
Summary 145
Samenvatting 149
References 153
Curriculumvita e 165
Ifa chapte r ofthi sthesi s isuse d asa referenc e inth e text,th e abbrevation "ch."i s oftenused . Introduction
A detailed knowledge of the development and dispersal of angiosperm pollen is important for both biologists and plant breeders. The first reports on this topic appeared as early as in the 17th century in the letters of Van Leeuwenhoek and the work of Grew. During the second half of the last century, the different stages of pollen development were discovered. Besides, the importance for plant breeders became evident after the (re-)discovery of the Mendel laws. Around the turn of the century French and German biologists published many works on the different steps of the anther dehiscence process, with a strong teleological approach. These works concerned light microscopical observations of living tissues together with simple, but ingenious micromanipulations. After 1930 the attention on the dehiscence process faded, whereas the light microscopical investigations on the generative reproduction processes continued. From the early 1960's the electron microscope and modern physiological techniques revealed many of the processes of microsporo- and microgametogenesis in great detail. These works mainly focussed on the pollen and tapetum development and were summarized in some excellent reviews. The mechanical processes in the stamen still received less attention in these years and most of the older works seemed to be forgotten, given the flaws and the call for research on this topic in some botanical standard works. Accordingly, detailed work on the entire stamen did not exist. In this thesis the classical and modern techniques were combined to investigate the stamen as a functional unit for pollen production and dispersal respectively. Chapter 1
HYDRATION CHANGES DURING ANTHER DEVELOPMENT
C.J. Keijzer
Summary
The hydration changes of the anther tissues of Gasteria verrucosa are followed from meiosis to dehiscence. The water percentage of the locule wall is about constant during this development, in spite of remarkable histological changes. The water percentage of the locule contents is about constant during the microspore stages. Shortly before maturation the pollen grains take up water from the surrounding tissues. The water percentage of the locule wall decreases strongly at anthesis, that of the locule contents at dehiscence. Ventilation inside the flower bud leads to desiccation of the anthers by evaporation. Shortly before anthesis the locule wall can evaporate much faster than in early developmental stages, whereas it protects the pollen grains against dehydration as long as the anther is closed.
Introduction
With regard to anther development little attention has been paid to water displacements. Humidity plays an important role since for example after meiosis the locular fluid is the connection between the microspores and the surrounding tapetum. Changes in water content were often mentioned in the explanation of the opening of the anther. Strassburger (1902) and Schneider (1911) related the swelling pollen mass to anther opening. Furthermore the outward bending of the locule wall, caused by the dehydrating endothecium cells, is a well known process. Schmid (1976) reports the disruption of the xylem in stretching filaments of different species at anthesis. This would handicap the further supply of water to the anthers. Erickson (1948) determined the fresh and dry weights of Lilium anthers during development and founda shar pdecreas eo fth efres hweigh ta tdehiscence . Inthi schapte rth ewate rcontent so fdifferen ttissue sdurin ganthe r development are determined in order to clarify some aspects of tissue interactionsan danthe rdehiscence .
Materialsan dmethod s
Developing anthers ofGasteri averrucos a (Mill.)H.Duva l were decapitated andsom eo fth etissue swer eseparate da sfollows .Th elocul econtents ,bein g pollenan dlocula rfluid ,wer ecollected .Nex tth econnectiv etissu ewa s removedwit ha razo rblade ,s oleavin gth elocul ewall .Thi s"wall "consist s ofa nepidermis ,a nendothecium ,on emiddl e layeran da tapetum .A tth e connectivesid eth eepidermi si sabsent .Ligh tmicroscopica l controlensure d thecorrec tseparatio no fth etissues .Fres hweight so fth ewall so ftw o adjacent loculesa swel la sth elocula rcontent s werequickl y measured.Dr y weightswer emeasure dafte rheatin gth etissue sfo r1 6hour sa t7 0degree sC . The freshweight s were measured 45,6 0 and 75 seconds after the anther openingan dthes edat awer eextrapolate dt o0 seconds ,i.e .th emomen to fth e antheropening .Thi swa sdon et oovercom eth eeffec to fevaporatio ndurin g thepreparatio n work. From these data the waterpercentag e was calculated. Thefres hweigh twa stake na s100% . Inorde rt otes tth eanthe revaporatio nabilit yi ndifferen tdevelopmenta l stages,th einsid eo fth eclose dflowe rbu dwa sventilated .Fo rthi spupos ea micropipettewa spushe dthroug hth ebas eo fth eperiant hint oth eflowe rbud . Nexta nai routle twa smad eb ymean so fa needl ehol ei nth eto po fth ebud . Thepipett ewa sconnecte dwit ha nairpum pa ta pressur eo f0.0 5kg/squar ecm . Thistreatmen t ledt oanthe r dehydration and lethality inal lteste dstage s within4-1 0 days.
Resultsan ddiscussio n
1.Wate rpercentage s inth edevelopin ganther .
In fig.1 the water percentages of the tissues are presented in the differentstage so fanthe rdevelopment .Ou rvalue sagre ewit hth efinding so f Linskens (1967),wh ofoun dvalue saroun d82 %fo rth eentir eanthe rdurin gth e mainpar to fth edevelopmen ti nPetuni ahybrida .I nth eyounges tstages ,i.e . frommeiosi st ojus tafte rmicrospor e mitosis (untillin e 1), thewate r percentage ofth e loculewal li sa littl ehighe rtha ntha to fth elocul e contents.Th efres han ddr y weights ofth eentir eanthe rar e increasing duringthes estages .Th elocul ewal lundergoe sn oremarkabl ehistologica l changes.O nth e contrary,th e microspores vacuolate andunderg o theirfirs t mitosis.Besides ,th epolle nwall sar eforme dan dreserv esubstance sar e stored.I nspit eo fthes eprocesse sth ewate rpercentage so fbot hth elocul e wallan dth elocul econtent sremai nabou tconstant . Betweenth eline s1 an d2 ,i.e .th eperio do fpolle nswelling ,th ewate r percentageo fth elocul econtent sincreases ,wherea stha to fth elocul ewal l remains aboutconstant .Thi s disagrees with the findingso fLinsken s (1956) whofoun da decreas eo fa fe wpercen tfo rth eentir eanthe ri nthi speriod . Inthes estage ssom e remarkablehistologica l changes occur. Within a short period of time epidermis and endothecium cells swell by vacuolation. Moreover, theendotheciu m cells stretch in radial direction, enlarge their vacuoles and deposit their characteristic U-shaped wall thickenings. Thevolum eo fth etapetu m cellsdecreases ,a proces swhic hstarte dsomewha t earlier.Thes echange sappea rt ob edu et oa nequa lincreas eo fbot hwate r anddr ymatter .Besides ,th ewall sbetwee ntw oadjacen tlocule sa swel la s thestomiu mar eopene ddurin gthi speriod .Thes eopenin gprocesse soccu r insideth eclose dflowe rbud . Inth elocul e thecapacit yo fth epolle ngrain sincreases ,whil eth e surroundinglocula rflui ddisappears ,a proces s which iscomplete d atabou t line2 .Thi sflui di spresumabl ytake nu pb yth epolle ngrains ,tha tfinall y areth eonl ycontent so fth elocule ,togethe r withsom e tapetum-derived pollenkitt. The increase of the water percentage of the entire locule contentsindicate sa transpor to fliqui dfro mo rthroug hth elocul ewal lint o the locule.Thi s canb eexplaine d inthre edifferen t wayswhic h areno t mutuallyexclusive .Firstly ,wate r may come from thedisappearin g tapetum cells.Secondly ,parenchym a cells arerelease d duringth eenzymati c opening of the tissue between two adjacent locules, as a preparation on the dehiscence.Thes ecell sar erupture dbetwee nth epolle ngrain san dlos ethei r water.Thirdly ,transpor to fwate rfro mth efilamen to rth econnectiv etissu e toth elocul ei sprobabl ea sth efres hweigh to fth eanthe ri sstil lslightl y increasingdurin gthi speriod . Shortly beforeanthesis ,betwee nth eline s2 an d3 ,th ewate rpercentag e ofth elocul econtent sdecrease sver yslightly ,whic hma yb edu et osom e 100 GASTERIA VERRUCOSA
i % 3i H 0 2 ""1- 1 loculuswall i ^*''"' 1 loculuscontents —-'''" * \ i \ 1 1 \ 1 1 \ 1 I M 50 I 2 \\ \\ \\ \\ \\ 1 \\
'l l' 1 anthesisi 0 . [dehisc. ^Pp^ ;jptgiaf ipasias TOTOTOTC ^orgtoic ioVoTc JQlaoIc i°@34 laasL ~§£~ #6 ^1 ^@ €t en free gen.cel l
GASTERIA VERRUCOSA
loculuscontents
Sea' Tjffist isissrs 3paoS TSTQIOIC i^OSss VooTP p.'ooTc jSnjfc^c ^o! lossthroug hth eopene dstomiu mint oth eclose dflowe rbud . Atanthesi s(lin e3 )th erelativ ehumidit yaroun dth eanther sdecrease s sharplyan dth elocul ewal llose sit swate rrapidl yb yevaporation .Thi s agrees with thefinding s ofErickso n (1948).Th eresul ti sth eoutwar d bendingo fth elocul ewall ,i.e .dehiscenc e (line 4), byth ewal lthickenin g mechanism ofth eendotheciu m cells.Th epolle n grains startt oshrin k once theyar ei ncontac twit hth eope nair .Fro mtha tmomen tth ej*ate rpercentag e ofth elocul econtent s decreases sharply.Th epolle n grainsiremai nstuc kt o each other andt o the outward bending locule wall. The final water percentageso fth etissue sdepen do nth erelativ ehumidit yo fth eair .
2.Effect so fth eventilatio no fth eflowe rbuds .
Fig.2show sth eeffect so fthre e ventilation treatments (startinga t points1 ,2 an d3 )o nth ewate rpercentages .Th eanthe rtissu
THEPROCESSE SO FANTHE RDEHISCENC EAN DPOLLE NDISPERSAL . 1.THEOPENIN GMECHANIS M OFLONGITUDINALL Y DEHISCINGANTHERS .
C.J.Keijze r
Summary
Theproces so fanthe ropenin gi sfollowe di ndifferen tspecie sb ymean so f lightan delectro nmicroscopica ltechniques ,stere omicroscopica lobservatio n inside livingtissue san dmicromanipulation .Fro mthes edat aan da revie wo f themainl yver yol dliteratur eo nthi ssubject ,th eopenin gproces sha sbee n outlined. Itconsist s of thefollowin g steps: 1.Expansio no fth eepidermi s andendotheciu mcell san ddepositio no fU-shape d wallthickening sinsid eth e latter.2 .Enzymati copenin go fth eseptu mbetwee ntw olocules .3 .Mechanica l rupture of the tapetum. 4. Mechanical opening of the stomium by the centripetalforc eo fth ehighl yturgescen tepidermi san dendothecium .5 . Outward bending of the locule walls by the centrifugal force of the dehydrating epidermis and endothecium. Asth e steps4 an d5 clearl yreflec t themechanica lfunction so fth etw olatte rtissues ,th erol eo fth eothe r anthertissue san dth efilamen tar ediscussed .
Introduction
Sinceth ediscover yan dnamin go fth eendotheciu mb yPurkinj e (1830),man y botanistshav einvestigate dth eprocesse so fanthe r dehiscence.Mos to fthi s workwa scarrie dou taroun dth etur no fth ecentur y withlivin gtissues , using light microscopy and micromanipulation. French authors (Chatin,1870a, b,c;Lecler cd uSablon,1885 )wer eth efirs tt odistinguis htw o different stepsi nth eprocess :firstl yth etissu ebetwee ntw oadjacen tlocules ,calle d theseptu m (Venkatesh,1957),wa sdisrupted .Next ,th ethu screate dcavit ywa s openedb ymean so fon eslit ,calle dth estomiu m (Coulteran dChamberlain , 1903).Th eearlies tGerma ninvestigator sclaime dth eforc eo fth edesiccatin g endothecium to cause the opening of both tissues (Schrodt,1885,1901; Steinbrinck,1895,1909) .O nth econtrary ,Chati n(1870b )ha dfoun dtha tth e anther was alsoopene d in caseth eendotheciu m lackedan ywal lthickenings , asi nLycopersico nsp .Strasburge r (1902)reporte dth egrowin gpollenmas st o disrupt the septum and lytic enzymes to open the stomium. Schneider (1908,1909,1911)relate d theopenin g ofbot h tissues to thepressur e ofth e growingpollen .However ,Steinbrinc k (1909)foun da norma lopenin gproces si n antherstha twer eno tcompletel yfille dwit hpollen .Namikaw a (1919)an d Becquerel(1932 )suppose dth eactivit yo flyti cenzyme si nbot hth eseptu m and the stomium. Woycicki (1924a) prevented pollen pressure with a longitudinalcuttin gi nth elocul ewall ,whic hwa sfollowe db yth eopenin go f the septum. He reported pectinase activity for the septum opening. Furthermoreh ereporte dtha tth estomiu m wasopene dmechanicall yb yth e suddenswellin go fth eborderin gepidermi scell s (Woycicki,1924b).Fift y yearsearlie rChati n(1870a )ha dalread ysuppose dtha tth eswellin gepidermi s cellsi nsom especie smigh thav ea functio ni nth eopenin gprocess . Most of theauthor s related therol eo fth edesiccatin gendotheciu monl y toth efina loutwar dbendin go fth etw odisconnecte dlocul ewalls .Chati n (1870b)coul dpreven t thisbendin gb yraisin gth erelativ ehumidit y and concludedtha tevaporatio n causedth edehydratio no fth eendothecium .Hanni g (1910)demonstrate d that daylight couldrais eth etemperatur ei npigmente d anthers, which increased the evaporation,eve n inhumi d circumstances. Contrary toevaporation ,Burc k (1906)reporte d retraction of water fromth e anthersthroug hth efilamen tt oth enectaries ,bu tthi sobservatio ndisagree d with laterwork so fHanni g (1910), Wolff (1924)an dSchmi dan dAlper t(1977) . Anotherquestio nwa swhethe rth eendotheciu mwa sth eonl ytissu eresponsibl e forth ebendin go fth e locule wall.Bonnie ran dLecler cd uSablo n(1903 ) testedthi sb ymechanicall y removingth eepidermis ,whic hdi dno tinfluenc e thebehaviou ro fth elocul ewalls .Schneide r (1908)obtaine dth esam eresul t with isolated endothecium wall thickenings. Moreover he found the isolated epidermis event oben d centripetally inbot h wet anddr yconditions ,s o contraryt oth efina lforc eo fth eendothecium .Thes efinding sagree dwit h thebehaviou r ofanther s with anaturall yreduce depidermi s(Chatin ,1870a ; Staedler,1923). On thequestio n whether thedesiccatio nproces si nth eendotheciu m cells hasa hygroscopi co ra cohesivebackground , many reportsar eavailable .Th e formermechanism ,locate di nth ecel lwall ,wa saccepte db yLecler cd uSablo n (1885),Brodtman n(1898) ,Steinbrinc k (1909)an dSchip s (1913).Schwendene r
10 (1899)an dCollin g(1905 )foun ddehiscenc et ooccu rafte ral lth ewate rha d disappearedfro mth ecell san dconclude da hygroscopi cmechanism .However ,i n lateryear scohesiv eforces ,locate di nth ecel llumen ,wer emor ean dmor e accepted(Schneider ,1908 ;Hannig,1910 ;Steinbrinck ,19 15 ;Haberland ,1924 ;Vo n Guttenberg,1926; Straka,1962).Brodtman n (1898)an d Schneider (1908)isolate d theU-shape d wall thickenings from theendotheciu m cells. The latter found them toben d inward (closingth eU-configuration )durin g desiccation,whic h infac tmean sa hygroscopi ceffect .Canrigh t(1952 )believe dbot htype so f mechanismst opla ya role ,a conclusio ntha ti sals odraw ni nth ewor ko f Schmid(1976) . Inrecen treport sth eprecedin gstep so fth eopenin gproces swer eonl y incidentally reported (Horneran d Wagner,1980; Cheng, Chengan dHuang,1981) . Theonl y authors who recently focussed onthi stopic ,investigate dth erol e ofth efilamen t(Schmid,1976 ;Schmi dan dAlpert,1977) .Unti lnow ,n odetaile d descriptiono fth eproces so fanthe rdehiscenc eca nb efound .Moreover ,mos t ofth eol dreport ssee mt ob eforgotte no runread .A precis eknowledg eo f cytologicalan dphysiologica lprocesse so fth edehiscenc emechanis m canope n newway sfo rmanipulations .A simpl ewa yt opreven tdehiscenc ei swelcom ea s itwil l alsopreven tpolle nexposure ,whic hha sth esam eeffec ta smal e sterility.Thi si sinterestin gfo rplan tbreeders ,a sWyat t(1984 )showe d recently with aselecte d nondehiscen tPhaseolu s vulgarisL .line .Non - dehiscent anthers are frequently observed after chemical induction of male sterility (Laible,1974). Inthi schapte rmos texperiment sfro mth eol dreport swer echecke dan dne w experimentswer eadded .Th emai nstep so fth eopenin gproces stak eplac e betweentw oadjacen tlocule san dar eno tvisibl efro mth eoutsid eo fth e anther.T o observe this site,th e anther had tob e opened, which might disturbo rpreven tphysiologica lprocesses ,includin gmechanica lforce s caused by tensions of the tissues.Moreover , cutting the living tissue appearedt obrea kth ewea kconnection si nthi sare aeasily .T oovercom ethes e problems special techniques were used to approach the opening site. Experimental and micromanipulation work was combined with current microscopical techniques. On this topic three lecture abstracts were published before(Keijzer,1983b,1984b,1985a) .
11 Materialsan dmethod s
Inthi sstud yo nlongitudinall ydehiscin ganther smos to fth eobservation s werecarrie dou to nGasteri averrucos a (Mill.)H.Duva lan dLiliu mhybri d cv. Enchantment,bot hLiliaceae .T oclarif ycertai nspecia lstep so fth eprocess , some specially chosen species were used, like Fuchsia coccinea Ait. (Onagraceae), Ipomoeapurpure aRot h (Convolvulaceae),Hippeastru mhybrid acv . Apple Blossom (Amaryllidaceae),Euchari s grandiflora Planch, et Lind. (Amaryllidaceae),Nicotian atabacu mL .an dLycopersico nesculentu mMill , (bothSolanaceae) . Forligh tmicroscopica lobservations ,carefull y madethic kfres hcutting s (0,5mm )wer eimmerse d inth eclearin gsolutio n ofHer r (1971), afterwhic h the desired depth in the tissue could be chosen with aNiko n Optiphot microscopewit hinterferenc econtras tequipmen t(Nomarski) . Forultrastructura l observations intactanther swer efixe di n5 %gluta r aldehyde during4 hour s and 1%osmiu m tetroxide during 1hour ,bot h in cacodylatebuffer ,p H7. 2a troo mtemperature .Th especimen swer edehydrate d ina grade dethano lseries ,embedde di nth elo wviscosit yresi no fSpur r (1969)an dobserve d ina Philip sE M30 1a t6 0kV . Toobserv eprocesse s inside livinganthers , flower buds were decapitated witha razo rblade ,covere dwit ha heav ycove rsli pan dobserve dwit ha hig h magnificationstere omicroscop e (upt o 160x)unde r glassfiberdirected ,col d light (150W )(figs.1, 2an d 3).Unde rthes econdition santhe r development continued in anorma l way. For manipulatory experiments a Leitz micro manipulatorwit hultrathi nglas spipette swa sused . Toestimat e theosmoti cpressur e inside living cells the visual melting pointmetho do fBearc ean dKoh l(1970 )wa sused .
Results
1.Cytologica l changesconcernin ganthe rdehiscence .
Theanthe ro fGasteri averrucos aconsist so ftw othecae ,boun dtogethe rb y a connective tissue with a centralvascula rbundle .Eac hthec acontain stw o locules whichbot h dehisceb y meanso fon elongitudina l slit.A nanthe r morphology like this is representative for most of the theAngiosper m species.Th e"wall "o feac hlocul econsist so ffou rdifferen t singlelayere d
12 tissues: anepidermis ,a nendothecium ,a middl e layeran da tapetu m (fig.4). Facingth econnectiv etissu eth eepidermi si sabsen tan dher eparenchym a bordersth eendothecium . Thefirs tobservabl ecytologica lchange sconcernin gth eanthe ropenin g startwhe nth egenerativ ecel lo fth epolle ngrai ni sbein gtie dof ffro mth e intine. In this stage the cells of the epidermis and the endothecium graduallylos epar to fthei rstarc han dstar tt oexpan di nbot htangentia l andradia ldirectio nb ymean so fth eenlargemen to fthei rcentra lvacuole .A t thesam emomen tth echaracteristi cU-shape dwal lthickening sar edeposite di n theendotheciu m cells (fig.5).Bot hthes eprocesse soccu rrathe rquickly . However,i na smal lzon ebetwee ntw oborderin glocules ,a fe wendotheciu m cellsremai nsmal lan ddon' tdevelo pU-shape dwal lthickening sbu tposse s thickened corners (fig.5).Th ecell si nthi szon e lack any starch (fig.6)an d aremuc hlonge ri nlongitudina ldirectio ntha nth eborderin gcell so fth e endothecia,th emiddl e layersan d theconnectiv e tissue (fig.7). Giventhes e differencesan dth edeviatin gbehaviou r lateron ,on eca ndefin ethes ecell s nott obelon gt oth elatte rthre etissue san dw ecal lthi szon eth eseptum , accordingt oVenkates h (1957).Accordingly ,th eonl ytissu etha tdevelop s uniformly aroundth ewhol e loculei sth etapetum .A fe w epidermis cells facingth eseptu mdon' texpan dlik eth eres to fth eepidermis .Als othes e cellsar e much longer in longitudinal direction, asit e calledth estomiu m (Coulteran dChamberlain,1903) . Thesecon d important change concerninganthe rdehiscenc eoccur sshortl y afterthes echange si nth eepidermi san dendothecium .Al lth ecell so fth e septum dissociate from each other and from theborderin g tissues.Thi s processstart swit hth eformatio no fgap si nth emiddl elamella ,whic hexten d allaroun dth ecell s (fig.8).Th etapetu mcells ,however ,remai nconnecte dt o eachother .The ykee pth elocule sclose dan dhol dth edissociate d septum cellsi nthei rorigina lpositio n (fig.9).I nthi sstag eth etapetu mcell sar e only surrounded by sporopollenin containing membranes,th e innertangentia l one bearing the orbicules (fig.8). Afterthi sstag eth etapetu m cellsborderin gth eseptu mdisrup tafte rth e expansion of the individualpolle n grains.Thi s continuous expansion starts afterth ereleas eo fth emicrospore sfro mth emeioti ccallos ewal l(Keijzer , 1983b).Fro mth estag ei nwhic hth egenerativ ecel li scompletel ytie dof f fromth eintine ,th evolum eo fth epolle nmas sexceed sth ecapacit yo fth e loculesan dth esli ti nth eseptu mwiden sb ywhic hth etapetu m isbroken . Thismomen tca nb esee nclearl yi nth eanther so fFuchsi acoccinea ,whic h
13 have a very wide septum (fig.10). From this moment the pollen populations of each two bordering locules make contact with each other (fig.11). The separated cells of the septum may have been dissolved completely by this time; if not so, they can be observed between the pollen grains (fig.12). Anyway they can be dislocated in an earlier stage after artificially breaking the tapetum by micromanipulation (fig-13). From this moment the anther is closed only by the small epidermis cells of the stomium (fig.14). The tangential swelling of the epidermis and the endothecium cells enlarges the circumference of the locule wall. However, because the inner tangential walls of the endothecium cells are rigid due to the interconnected U-shaped wall thickenings, the circumference of the inside of the locule wall
1. A method to observe processes inside living anthers with a stereo microscope (SM). A decapitated flowerbud of Lilium hybrid cv. Enchantment is held in its position by a clamp (C), closed with a coverslip in plastic rings (CS) and illuminated with glassfiber directed cold light (L). 2. View through the coverslip into the flower bud of fig.1. The base of the flower is surrounded by wet paper (P), which is held in its position between two lower situated plastic rings. (4x) 3. A still completely closed anther in the flower of fig.2. Note the large epidermis cells (LE) near the stomium. (25x) 4. The opening area of the Gj_verrucos a anther. Ep=epidermis, En= endothecium, ML=middle layer, T=tapetum. Thick fresh cutting cleared with Herr's solution and observed with an interference contrast microscope. (625x) 5. The epidermis and endothecium cells of G^ verrucosa swell and U-shaped wall thickenings are deposited in the endothecium cells, the cells of the septum (S) possess thickened corners. Same technique as fig.4. (560x) 6. In the septum (S) of Lilium hybrida cv. Enchantment the absence of starch is clearly visible, contrary to the bordering tissues. Ep=epidermis, En=endothecium, C=connective tissue. Stained with IKI. (800x) 7. View from inside the locule towards the septum (S) of G^ verrucosa. Note the longitudinal length of the cells as compared with those of the endothecium (En). Same technique as fig.4. (1400x) 9. The cells of the septum (S) of G. verrucosa continue their dissociation while the tapeta (T) keep the locules closed. (560x)
14 15 isfixed .Thi s combination of anenlargin g outsidean da fixe dinsid ebend s the locule wall inward. Next the small epidermis cells of the stomium dissociate and the two locule walls bend further inwards (fig.15). The epidermiscell sborderin gth e smallone s work asa turnin gpoint , whichi s also clearly visible by the complementary molds of their cuticles (fig.16). Insom especie s (e.g.Liliu mspp. )thes estomiu mborderin gcell sexpan dt oa largerexten ttha nth eothe repidermi s cells (fig.3).Th edissociatio no fth e smallepidermi scell si sno tprecede db yvisibl egap si nth emiddl elamell a likebetwee nth ecell so fth eseptu m (fig.16).The yremai nunchange dunti l thesudde nopenin go fth e stomium, after which they allremai npresen tan d connectedt oth etw olocul ewalls .Fro mthi smomen tth elocule sar eopen , although their walls remain inth e inward bent position (fig.15). This configurationo flocul ewall spresse dtightl yagains tth epolle nremain s untilanthesis .Al lth estep smentione dabov eoccu rinsid eth eclose dflowe r bud. Atanthesi sth ecell so fth eepidermi san dendotheciu mlos emos to fthei r wateran dshrink ,th ethickene dwall so fth eendotheciu m toa lesse rextent . Contraryt oth einwar dbendin gmechanis m ofth eforme r step,thi sbend sth e loculewall soutward .A sevaporatio nappear st ob eth emai nreaso n forthi s rapidlos so fwater ,thi sproces sca nb eshow ni nth ehig hvacuu mo fth eSE M
8.Gap sappea ri nth emiddl elamella eo fth ecell sbetwee nth eseptu m(S ) and the epidermis (Ep) of G^_verrucosa . Deeper in the septum the intercellular spaces have enlarged and are filled with pollenkitt(Pk) . The tapetum cells containonl ypollenkit t and tryphinean dar ebordere d by sporopollenin containing membranes,partiall y coveredwit h orbicules (0). (1925x) 10.Th e cells of the septum of Fuchsia coccinea have disappeared. The expandingpolle nhav ebroke nth etapetu m ofth e leftlocul e (L),th e othertapetu m (R)i sstil lclosed ,a si sth estomiu m (St).Fres hcuttin g inwater .(440x ) 11.Th etw opolle npopulation so fG .verrucos ahav ebroke nth etapet aan d makecontac twit heac hother .Sam etechniqu ea sfig.4 .(500x ) 12.Detai lo ffig .11 .Th ecell so fth eseptu m (S)continu ethei rdissociatio n betweenth epolle ngrain s (Pg).(1980x ) 13.A dissociated septum cell (S) of G^ verrucosa is set free after artificially breakingth eseptum .Sam etechniqu ea sfig.4 .(780x )
16 17 (fig.17). As the length of the bar-supported endothecium walls is not absolutelyfixed ,th elocul ewall sshrin ki nbot htangentia lan dlongitudina l direction (figs.17an d 18).Besid eth eepidermi san dth eendothecium ,th e otheranthe rtissue sa swel la sa smal lapica lpar to fth efilamen tals o dehydrate. InG ^verrucosa , as inmos to fth eanima lpollinate d species,pollenkit t andtryphin estic kth epolle no nth eanther ,equall ydistribute dove rth e entire, shrunken locule walls. This is the final configuration of the dehisced anther (figs.17an d18) .
2.Additiona lobservation s inothe rspecies .
Apartfro mGasteri averrucos aan dLiliu mhybrida , weobserve d someothe r species, because of their declining behaviour during different stepso fth e openingprocess . InIpomoe apurpure aRot hth evolum eo fth epolle ndoe sno texcee dth e
14.Th esam econfiguratio na sfig.11 .afte rremova lo fth epollen .Th eanthe r isclose d only by the smallepidermi s cells (Ep).Sam e techniquea s fig.4.(780x ) 15.Th estomiu mo fG .verrucos ai sopene ddu et oth einwar dbendin glocul e walls.Facin gthi sstomium ,som eseptu mcell s(S )ar estil lconnecte dt o the connective tissue (C).Sam etechniqu ea s fig.4.(780x ) 16.Th eoute rtangentia lwal lo fth estomiu m bordering epidermis cells(Ep ) of G. verrucosa work as a turning point and form complementary templates.Th eradia lwall s(W )betwee nth esmal lepidermi scell sar e stillintact . (1100x) 17.A fresh anther ofG ^verrucos a bends open duet oevaporatio n inth e vacuum of the scanning microscope. This is soon followed by the desiccationo fth epollen ,resultin gi nchargin gi nth efina lstages . (24x) 18.Th e transversely cutdesiccate d anther ofLiliu m hybrid cv.Enchantmen t shrinksstrongly .I ti scovere dwit hpollen , thanks toth e sticking pollencoatings.Techniqu e of fig.1.(45x ) 19.Th efina lconfiguratio ninsid eth eclose dflowe rbu do fLiliu mhybri dcv . Enchantment.Th einwar dben tlocul ewall shav ebroke nth etapet aan dkee p thepolle ninsid eth eopene danther .Techniqu eo ffig.1 . (42x)
18 19 capacity of the locule and accordingly cannot break the tapetum. The latter is broken contemporary with the opening of the stomium. Premature opening of the stomium before the opening of the septum, that we could induce by manipulation, may occur spontaneously in Ipomoea purpurea Roth, sometimes resulting in asynchronous opening of the two bordering locules. In Nicotiana tabacum L. we found the opening of the stomium to take place after anthesis. In Ipomoea purpurea Roth and Eucharis grandiflora Planch, et Lind. the outward bending of the locule walls occurs before anthesis in the closed flower bud. On the contrary, in Hippeastrum hybrida cv. Apple Blossom and Allium spp. this process may take place a few days after anthesis.
3. Experimental approach of the processes.
To reveal the direct causes of the processes that we observed, some manipulations were carried out, using stereo, light and electron microscopy. a. The opening of the septum.
To test whether the outward bending endothecium can open the septum, the perianth was removed from G. verrucosa and L. hybrida flowers in situ just before the stage in which we observed the first gaps in the middle lamellae of the septum cells. Desiccation by open air, dry air (r.h. 35%), or even proceeded heating did not open the septum. To control whether the force of the endothecium was indeed directed outward in these cases, the septum was opened by microsectioning. Given this situation, the created slit in the septum widened immediately, indicating an outward directed force. On the contrary, when dehydration of the endothecium was prevented by filling the flower bud with water via a micropipette, the septum opened in a normal way after one or two days. To test the theory of the pressing pollen (Strasburger, 1902; Schneider, 1908,1909,1911) we removed the pollen from the anthers of G. verrucosa in situ by micromanipulation. After two days the septum was found open, while a control showed that immediately after the manipulation it was still closed. In case the pollen plays a non-mechanical role in the opening process, the experiment was repeated with just partial removal of pollen, leaving most of it in the locules, leading to the same result.
20 b. The disruption of the tapetum.
The opening of the tapetum as a separate process has not been investigated before. At the stage of its rupture this cell layer only consists of the tapetal membranes and a small amount of pollenkitt and tryphine. To observe the opening process in living anthers we used the stereo microscopical technique (figs.1,2 and 3). In G^_ verrucosa and Fuchsia coccinea the swelling pollen mass was seen to enlarge the opening of the septum, stretching the elastic tapetal membranes to a considerable extent until they broke, in general not synchronously in two adjacent locules. In L^_ hybrida the pollen does not exceed the capacity of the locule and here the tapetum was ruptured together with the stomium by the inward bending of the locule walls (fig. 19). As this is a sudden movement, it occurs synchronously in two adjacent locules. In general it is easy to demonstrate that the force needed to break the tapetal membranes is very low.
c. The opening of the stomium.
Also for the opening of the stomium we tested the proposed causes from the literature, being the outward bending endothecium (Schrodt, 1885,1901; Steinbrinck, 1895, 1909), the pollen pressure (Schneider,1908,1909,1911) and the lysis of the cell walls (Strasburger,1902; Becquerel,1932). For the first two we used the same methods as for the septum opening, leading to comparable results. Besides, we controlled the suggestion of Woycicki (1924b) that the stomium is broken mechanically by the sudden expansion of the bordering epidermis cells. As has been mentioned before, we observed not only these, but all the epidermis cells to expand. In our stereo microscopical observations this was followed by the inward bending of the locule walls and the mechanical disruption of the stomium (fig.19). We could artificially effect these changes by filling a decapitated flower bud with water, after it had suffered some chilling during 24 hours. Immediately after the water supply, the epidermis cells expanded and the locule walls bent inward, suddenly breaking the stomium. In this way premature opening of the stomium could be induced, even before the opening of the septum, a situation which is shown in f ig.20. The extent of the movement could be shown by removing the pollen from the anthers. Now the locule walls bent further inward, filling the empty locules (fig.21). We also tested the statement of Woycicki (1924b) that an increasing
21 osmotic pressure in the epidermis cells precedes the water uptake. Indeed the visual melting point method of Bearce and Kohl (1970) showed an increase of the osmotic pressure from 16.8 to 28.1 bar over a period of 72 hours.
d. The outward bending of the locule walls.
Concerning the outward bending of the locule walls we carried out some experiments to show the way of water loss from the anthers. We removed the tepals from an L^_hybrid a flower bud shortly before anthesis and removed one stamen from the receptacle. Within half an hour all the anthers started to bend open, the anther of the dissected stamen just 3-5 minutes earlier. Larger differences were found when not a whole stamen, but only an anther was removed. The latter started to bend open 10-12 minutes before the anthers in situ. All these experiments were carried out in a relative humidity of about 60%.
20. Premature opening of the stomium (St) of Lilium hybrid cv. Enchantment after the flower bud had been filled with water. The septum (S) is still closed. Fresh cutting in water. (48x) 21. The same configuration as in fig. 19., after removal of the pollen. The locule walls of Lilium hybrid cv. Enchantment bend into the empty locules (L). Fresh cutting in water. (27x) 22. The different steps of the dehiscence process. A. The deposition of U-shaped wall thickenings in the endothecium reveals the existence of a special tissue between two bordering locules: the septum. B. The cells of the septum dissociate by enzymatic lysis of their middle lamellae, the bordering tapeta remain intact. C. The two tapeta are broken mechanically, in this case by the expansion of the pollen mass. D. The small epidermis cells facing the septum, called the stomium, are broken mechanically by the inward bending of the two adjacent locule walls, which is due to water uptake by the epidermis and the endothecium cells. E. The locule walls bend outward due to evaporation from the epidermis and the endothecium. The pollen is equally distributed over the desiccated anther.
22 20 WM
Anatomical Forces caused changes byth etissue s
23 Toexclud ea possibl eeffec to fth eartificia lopenin go fth eperiant hi n thisprocess ,tw omicropipette swer epushe dthroug hth eperiant hint oth eti p and thebas e of theflowe rbu drespectively .Fro m oneo fthes epipette sdr y air(r.h.35% )wa spumpe dthroug hth ebud ,b ywhic hth eothe rpipett eserve d asa nexit .Whe nsuc hflower swer eopene dafte r3 0minutes ,th eanther sha d bentope ncompletely .Heate dai raccelerate dth eprocess .O nth econtrary , whennewl y openedflowe rbud so rdissecte danther swer eplace di na hig h relativehumidit y (98%)o reve nunde rwater ,th eproces swa sretarde do r suppressed.
Discussion
Ourmicroscopica lobservation ssho w ananthe ropenin gproces stha tca nb e dividedint ofiv estep sa spresente di nfig.22 .Thes estep swil lb ediscusse d in chronological order, except the changes in the epidermis and the endotheciumcells ,whic hwil lb ediscusse dtogethe rwit hth estomiu mopenin g and loculewal l bending, asthe yappea rt ob efunctionall y relatedt othes e twoprocesses .
1.Th eopenin go fth eseptum .
Ourobservation ssugges ta nenzymatica lseparatio no fth ecell si nth e septumrathe rclearly .Th edevelopmen to fhole snea rth emiddl e lamella,no t preferentially startingnea rth ecorner so fth ecells ,indicat esuc ha cause . Moreoverth epresenc eo fcompletel yseparate dintac tcell smake sa mechanica l cause unlikely, as does theobservatio ntha tth eborderin gtapetu m remains intact. This conclusion agrees with observations of Woycickl (1924a), Becquerel (1932),Namikaw a(1919 )an dHorne ran dWagne r(1980) .Th elatte r tworeport seve ndescrib eth etota ldisappearanc eo fth ecell si nth eseptu m ofSolanaceae .Woycick i (1924a)showe d thatth e collenchyma-like cellwall s mainlyconsist so fpectin ,indicatin gth enecessit yo fpectinas et oope nth e septum.Ou rexperiment ssho wtha tth eforc eo fth edesiccate dendotheciu mca n notope nth eseptum .O nth econtrary ,th eseptu mi sbein gopene di ncas ethi s force is artificially prevented or even when wall thickenings in the endothecium are absent as in Lycopersicum esculentum. Both Schrodt (1885,1901)an d Steinbrick (1895,1909)wh o hypothesize this desiccating endotheciumt obrea kth eseptum ,forge tan ylytica lpossibilit y inthei r
24 effortst oden yth etheor yo fth eswellin gpollenmass .I nou rexperiment sth e septumopene di na norma lwa yafte rw eremove d(par tof )th epolle nfro mth e locules. The opened septum that was found in male sterile anthers by Steinbrinck (1909)wa sals ofoun di nth emal esteril eAlo ever a(L. )Burm . fil. (ch.4)wit h degenerating, notexpandin gpollen .Fro mthes efinding san d theexperiment so fWoycick i(1924a )th econclusio nca nb edraw ntha tals oth e pollen grainsar eno tth ecaus eo fth edissociatio no fth eseptu m cells,i n spiteo fth ereport so fStrasburge r (1902)an dSchneide r(1908,1909,1911) . Enzymaticallysi so fth emiddl elamell a(an dlate ro nth ecel lwalls )remain s themos tprobabl ecause .
2.Th eruptur eo fth etapetu mnea rth eseptum .
Inth e literature theruptur eo fth etapetu mha sneve r been reported beforea sa separateprocess .I n generalthi stissu ewa sconsidere dt ohav e disappeared by the time of septum opening.Indeed ,th e only observable resistingstructure si nthi sstag ear eth esporopolleni ncontainin gmembrane s (ch.3)a sth ecel lwall shav edisappeare ddurin gmeiosis .Ou rdat asho wbot h theinwar d bending loculewall s (inL .hybrida )an dth eoutwar dpressin g pollen (inG^ _verrucosa )t obrea kthes emembranes .I nou ropinio nn ospecia l mechanisms are developed tobrea kthem , butruptur eoccur s mechanically by anysligh tforc etha tthe yreceiv eafte rth eseptu mha sbee nopened .
3.Th eopenin go fth estomium .
Asmentione dbefore ,som eauthor shypothesize dth esam ecause sfo rth e openingo fbot hth eseptu man dth estomium , i.e.th eoutwar d directedforc e of theendotheciu m (Schrodt,1885,1901 ; Steinbrinck,1895,1909 )o rth epressur e of thepollenmas s (Schneider,1908,1909,1911) . Our observations and experiments show thatth eforc eo f theoutwar d bending locule walls isno t strongenoug ht oope nth estomium .Furthermore ,i fSteinbrinck' stheor yi s right,i ti sstrang etha tth elocul ewall sremai nben tinwards ,bein gth e finalconfiguratio ni nth eclose dflowe rbu do fmos t species.Moreover ,th e openingo fth estomiu m ofLycopersicu mesculentum , lackingendothecia lwal l thickenings (Chatin,1870b), makes such a cause unlikely. We have some indicationstha ta nexceptio no nthi spoin tma yb efoun di nth ePoaceae ,bu t thisi sstil lunde rinvestigation . Schneider (1908,1909,1911)reporte dtha tth eexpandin gpolle npopulatio n
25 openedth estomium .W ecoul drejec tthi stheor ywit hth esam eexperiment sa s weuse dfo rth eseptu mopening ,i.e .polle nremova lan dth ecase so fmal e sterility. Enzymatical cellwal llysi si nth estomiu m epidermiswa sreporte db y Strasburger (1902).Ou rfinding scanno tsho wan ywal llysi si nthi slayer .W e never found intermediate stages, comparable with the gaps and loosely arranged walls in the dissociating septum. A second observation also disagrees withenzymatica l lysis.Th eopene dstomiu m always containsal lo f itssmal lcells ,connecte dt oeithe ron eo rbot ho fth elocul ewalls .Thi s meanstha talway sjus ton eradia lepidermi scel lwal li sopene dan dtha tthi s walli srandoml ychosen . Our stereo microscopical observations and artificial, water-induced opening of the stomium, even before the septum was opened, suggest a mechanicalruptur eb yth einwar dbendin glocul ewalls .Th etightl yclose d loculesi nth eflowe rbu dafte rth estomiu mi sopened ,th emutua lcuticl e templates,a swel la sth einwar dbendin glocul ewall safte rpolle nremoval , supportthi stheory .Woycick i (1924b)teste dthi sb ymicromanipulatio nafte r earlier indications by Chatin (1870a). However, both researchers only observedth elarg eepidermi scells ,tha tar eborderin gth esmal lones ,t o expand.I nG« _verrucosa ,a si nothe rspecies ,w efoun dth ewhol eepidermi sa s well as the endothecium cells to expand, the latter mainly in radial direction.Ou rexperiment swit hth evisua lmeltin gpoin tmetho drevea la n increaseo fth eosmoti cpressur ei nbot htissues .Woycick i(1924b )als ofoun d this increase,usin ga KN03-plasmolys emethod , andassociate dthi swit hth e decrease of starch and increase of sugars. Also our findings show a dissolutiono fstarc hi nbot hth eepidermi san dth eendothecium .Althoug h thisma ycaus eth eincreas e ofth eosmoti cpressure ,par to fth estarc h derivativesma yals ob euse dfo rth ewal lthickening si nth eendothecium . Thisconclusio ni ssupporte db yth eabsenc eo fstarc hi nth eseptu mcells , whichd ono tdevelo pwal lthickenings . Itremain sdifficul tt oprov ewhic hpar to fth eepidermi san dendotheciu m is (mainly)responsibl e for the stomium rupture. Presumably there are specialisations,dependin go nth especies ,fo rexampl eth elarg ecell so f Liliumspecies .Staedtle r(1923 )describe da rang eo fplan tspecie swit ha gradualreductio no rdisappearanc eo fth eanthe repidermis .Remarkabl ei shi s findingtha tth eepidermi snea rth estomiu mpersist si nmos tcases . Theradia l stretching which canb eobserve d inth eendotheciu m cells duringth edepositio no fth ewal lthickening sma yb eusefu lt oenlarg eth e
26 distancebetwee nth eswellin gepidermi scell san dth efixe dinne rtangentia l endothecium wall. In this way the bending effect due to the epidermal expansion isenlarged . Besidesth eruptur eo fth estomium ,th einwar dbendin go fth elocul ewall s mayhav ea secon dfunction .I tals oprevent sth eprematur eexposur eo fpolle n insideth eclose dflowerbud ,whic hmigh tlea dt osel fpollination .
4.Th eoutwar dbendin glocul ewalls .
Thisi sth emos teasil yobservabl ean dmos tdescribe dste po fth eprocess . Theoutwar ddirecte dbendin go fth edesiccatin g locule walls works contrary toth einwar ddirecte dmovemen to fth epreviou sstep .Th eman yreport so n this process, which are often results of ingenious micromanipulatory experiments,ar emainl yfocusse do nthre etopics .Firstl yth eanatomical - physiological reason ofth e water loss; secondly thephysica l processes in thedehydratin g cellsan dthirdl y thebendin gpattern s ofth e different speciesi nrelatio nt othei rmod eo fpollination . Ourdryin gexperiment ssho wtha tventilation ,periant hremova lan dheatin g canaccelerat eanthe ropening ,eve nbefor eanthesis .O nth econtrary ,a hig h relativehumidit yo rwate raroun dth eanther sca npreven topening .Fro mthes e resultson e canconclud etha t evaporation from theanthe r surface isa n importantfacto rfo rth elos so f water(Chatin ,1870b ;Hannig,1910) .However , iti sremarkabl etha tanther sdesiccat ewhil eth eothe rpart so fth eflower , like the rest of the plant, remain fully turgescent. This means that differenceso rbarrier s betweenth eanthe ran dth efilamen t mustexist . Althougha stud yo nsuc hfactor si spresente di nchapte r7 ,som epreliminar y resultsar ebriefl ydiscusse dhere ,du et othei rimportanc efo rth eprocess . Manyspecie sposses sstomat ao nth eanthers ,whic hw eneve rfoun dt ob e closed, independently from the developmental stage.The y will be important ways for water exchange. As not all species possess stomata, more possibilitieswil lexist .Probabl yth eundulatin gcuticl ea tmaturity , which hasbee n observed indifferen tplan t species (Chenge tal.,1981 ; ch.4an d5 ) isimportant .I tma yaccelerat eevaporation ,whic hwoul dagre ewit hou r findingstha tanther sca nlos ewate rmuc hfaste rtoward s maturity thani na young stage (ch.1). Onth e contrary, the cuticle thickens towards maturity, whichmake ssuc ha nexplanatio nuncertain .A si nmos tspecie sth eanther sar e bentope njus tafte ranthesis ,th erelativ ehumidit y ofth e environment controlsth espee do fthi sproces sa swel la sth efina lwate rconten to fth e
27 anther (ch.1). In some species (e.g.Ipomoe a purpurea and Eucharis grandiflora)th eanther sar ealread yopene dinsid eth eclose dflowe rbud . This may bedu et oeithe r alo w relativehumidit y insidethi sbu do ra n activeretractio no fliqui dfro mth eanthe rint oo rthroug hth efilament . However, apossibl e water transport from the anthers to the nectaries (Burck,1906)wa s rejected by Hannig (1910), Wolff (1924)an d Schmidan d Alpert (1977). InPoacea ether ear eindication stha tth esuddenl y stretching filamentretract swate rfro mth eanther .Moreover , this stretching disrupts thevascula rtissue ,whic hpresumabl yprevent sfurthe rwate rtranspor tt oth e anther(Schmid,1976) . Inth edehiscin ganthe rth eeffec to fth eevaporatio ni sdiminishe db yth e watercapacit yo fth eanthe ritself .Besid eth econnectiv etissue ,th emiddl e layer(s)ma yhav ea functio ni nthi sprocess .I fso ,thei rspecie sdependen t moment of degeneration is important (ch.5). Itwil l be clear that many factorsca ninfluenc eth elos so fwate ran dit sspeed . Inspecie swit hpersisten tmiddl e layers,lik eL^ _hybrid a (Reznickovaan d Willemse,1980),cel lwall sar eshrinkin go nbot hside so fth efixe dinne r tangentialwal lo fth eendotheciu mcell safte ranthesis .O nth eoutsid ethes e areth ewall so fth eepidermi san dth eoute rtangentia lendotheciu m wall,o n the inside the walls of the middle layer(s).However , the force at the outsideappear s tob estronger ,presumabl y fortw oreasons .A tfirst ,a s mentioned before,th edistanc ebetwee nth e shrinking cell walls atth e outsidean dth einne rtangentia lendotheciu mwal li sfixed .Moreove rthi s distancei srelativel y largei fcompare dwit hth edistanc ea tth einside . Secondly, the epidermal and endothecial cell walls arei ngenera lthicker , whichi simportan tfo rhygroscopi cmechanisms .However ,jus tth eendotheciu m itself may be responsible forth emovement . This statement issupporte db y thenatura lepiderma lreductio ni nsom especie s (Chatin,1870a;Staedler , 1923)o rmechanica lremova lo fthi stissu e(Bonnie ran dLecler cd uSablon , 1903;Schneider ,1908) .Th ecentripetall y bending epidermis reported byth e latterautho reve nworke dcontrar yt oth efina lforc eo fth eendothecium . Fromthes efinding sth econclusio nca nb edraw ntha tth ecel llayer sothe r thanth eendotheciu mar eno tnecessary ,o reve nwor kcontrar yt oth efina l bendingo fth elocul ewall . Inou ropinio nth epressur eo fth epolle nagains tth epollenkit tsupplyin g loculewall sbefor eanthesi s equally distributesth epolle nove rth eentir e locule wall, which causesa noptima lexposur eafte rdehiscence .Thi sca nb e seeni nfig.1 8i nwhic hth edehisce danthe rform sa massive bar, completely
28 surroundedb ypollen .However ,th efina lshap eo fth eanthe ran dth esit eo f pollen exposurevarie s greatly depending onth e speciesan d itspollinator s (Coulteran dChamberlain ,1903 ; Richter,1929;Vo nGuttenberg ,1926) .I tfirstl y dependso nth einitia lshap eo fth eanthe ran dsecondl yo nth esite ,patter n and shape of the endothecial wallthickenings .Stanle y andKirb y (1973) report that the evolutionary change from entomophylic to (secundary) anemophylic pollination is accompanied bya shif tfro m longitudinal to poricidaldehiscence .Thi sturn sth eanthe rint oa polle nsac krathe rtha n the sticky bar of fig.18. Apart from bending the walls, we found the desiccation also to decrease the length of the locule walls in both longitudinalan dtangentia ldirection .I nth esecon dpar to fthi swor kth e importanceo fthi sshrinkag ewil lb erelate dt opolle ndispersa l(ch.3) .
Conclusions
This chapterha s shown thatth eproces s ofanthe rdehiscenc econsist so f fourmai nstep so f differentphysiologica l origin.Mos t ofth eobservation s and experiments were carried out withG .verrucos a andL^ _hybrida .However , theprocesse s insom eothe rteste dspecie sappeare dt odiffe r only onmino r pointso rtimin gfro mthes etw ospecies ,s otha ta genera lsequenc eca nb e given(fig.22) : 1.Th eepidermi san dendotheciu mcell sstar tt oexpan dan dU-shape dwal l thickeningsar edeposite di nth elatter ,leavin ga parenchymatica lzon e betweentw oadjacen tlocules ,calle dth eseptum . 2.Th ecell so fth eseptu mdissociat eb yenzymati clysi so fthei rmiddl e lamellae,th etw otapet aremai nintact . 3.Th etw o tapeta disrupt followingth efirs toccurrin gsligh tmechanica l forcei nth erelevan tdirection . 4.Th esmal lepidermi scell sfacin gth eseptum ,calle dth estomium ,ar e brokenmechanicall yb yth einwar dbendin go f thetw o adjacent loculewalls . This movement is effected by the tangentially expanding epidermis (and endothecium?)versu sth efixe dthickene dinne rtangentia lendotheciu mwall . 5.Th elocul ewall sben doutwar db yth ecohesiv ean dhygroscopi cshrinkag e of thetangentia loute r wall of theendotheciu m versus its fixed thickened tangentialinne rwall .Th elos so fwate ri sdu et oevaporation ,ofte nbecaus e ofanthesis ,an dprobabl ysupporte db ycertai nbarrier si nth efilament . Itha sbecom e cleartha tth eepidermi san dth eendotheciu m aredirectly .
29 mechanically involved inth eprocesses .Th e endotheciumcause stw ocontrar y directedforces :centripetall yt oope nth estomiu man dt okee pth epolle ni n theanther ,centrifugall yt oope nth eanthe ran dt oexpos eth epollen .Th e latterforc ei swel lknown ,th eforme rwa sunrecognize dunti lnow .A possibl e roleo fth emiddl elayer sconcernin gth espee do fevaporatio n isbriefl y discussed.Th etapetu mappear srathe rpassive ,bu tma ypla ya rol ei nth e inductiono fth eprocesses ,a sd oth eothe ranthe rtissues .I ngenera lth e problemo finductio nan ddirectio no fsignal si soutsid eth escop eo fthi s study. In the literature onthi s subject, many experiments with livingtissue s arereported .Especiall y Woycicki (1924a,b)presente d the most complete description of the dehiscence processes until now. Itshoul db e emphasized thatfo rprocesse slik ethese ,micromanipulator ywor kwa san dwil lb eo f greatvalue . Giventh eproces sdescribe dhere ,manipulation s canb eoutline dfo rplan t breedingpurposes .Eas ymethod st odistur bth eproces sma yb ehelpfu li n cropstha tar edifficul tt oemasculate .Fo rexample ,prematur eopenin ga youngflowe rbu do fLiliu msp. ,simpl yb ycutting ,le dt oa decreas eo fth e relativehumidit yan dkep tth estomiu mclose di nsom ecases .I nthi sligh t theunderlayin gstud yma yb eo fpractica limportance .
30 Chapter 3
THE PROCESSES OF ANTHER DEHISCENCE AND POLLEN DISPERSAL. 2.THE FORMATION AND THE TRANSFER MECHANISM OF POLLENKITT, CELL WALL DEVELOPMENT IN THE LOCULE TISSUES AND A FUNCTION OF THE ORBICULES IN POLLEN DISPERSAL.
C.J. Keijzer
Summaxy
The development of the locule of Gasteria verrucosa (Mill.) H.Duval and Lilium hybrid cv. Enchantment, especially the border between the sporophyte and the gametophyte, is investigated by means of light and electron microscopy, histochemistry and micromanipulation and related to pollen dispersal. The pecto-cellulosic cell walls of both the middle layer and the tapetum and a part of the endothecium disappear and the tapetum cells are covered with sporopollenin containing tapetal membranes and orbicules. The cell content of the tapetum turns into the hydrophobic pollenkitt. In the locule similar changes are observed as the pecto-cellulosic and callose walls of the meiocytes disappear and the future pollen grains are covered with the sporopollenin containing exine. These hydrophobic substances on both sides of the sporophyte-gametophyte surface are related to pollen dispersal, due to their non-wettability. Secondly the transfer of pollenkitt from the tapetum to the locule is investigated. It appears to be due to capillary forces in the locule after the continuous expansion of the pollen, not to an active flow or a direct attraction by the pollengrains.
Introduction
In many plant species the tangential inside of secretory tapeta is covered with orbicules (Bhandari,1984). These acetolysis resistant, sporopollenin containing structures are often connected to an acetolysis resistant tapetal
31 membrane,whic hwa sfirs treporte db yBanerje e(1967 )i ngrasses .Th erol eo f thesetapeta lcovering sappeare ddifficul tt oexplain .Orbicule swer esee na s relicts (Echlin,1968)o ra nai di nexin eformatio n (Maheshwari,1950 )an d tapetal membranes asa culture sac (Heslop-Harrison,1969).However , Heslop- Harrison (1968)reporte d apossibl e role fororbicule s inpolle ndispersal , duet othei rnon-wettability .I nthi schapte rw e checkth elatte rhypothesi s with lily anthers.Furthermore , the changes inth ewal l composition ofth e loculetissue so fGasteri a verrucosa (Mill.)H.Duva lar einvestigate d in relationt osuc ha rol ean dpossibl eothe rfunctions . Alsoth etransfe r mechanism ofth eals ohydrophobi cpollenkit tt oth e pollengrains ,whic hi simportan tfo rentomophyli cpolle ndispersal ,wa sno t clear.I nspecie swit hperiplasmodia ltapet ath ecell sinvad eth elocul ean d pollen sticking substances remain on the pollen grains after the desintegration of these tapetum cells (Mepham and Lane,1968,1969a,b). However,th ecell so fsecretor ytapet aremai no nth elocul ewal lan dthei r pollenkitt istransporte d toth e locule shortly before dehiscence. Sincen o locularflui di spresen taroun dth epolle ngrain sa ttha tmomen t(ch.1 )a liquidtranspor tmediu mi sabsen tan dmerel ypollenkit ti stransferred .Thi s transfermechanis mi sinvestigate di nG .verrucosa .
Materialsan dmethod s
For electron microscopical observations intact anthers of Gasteria verrucosa (Mill.)H.Duva lwer efixe di n3 %gluta r aldehydefo r4 hour san d1 % osmium tetroxidefo r 1hour ,bot hi n0. 7M cacodylat ebuffer ,p H7. 2a troo m temperature.Afte rdehydratio ni na grade dethano lseries ,th especimen swer e embeddedi nth elo wviscosit y resino fSpur r (1969).Ultrathi nsection swer e poststainedwit hlea dcitrat ean durany lacetat e (Reynolds,1963)an dobserve d witha Philip sE M 301a t6 0kV . Forobservatio no fth einterna lstructur ewit hth eSEM ,th efixe dan d dehydratedanther swer e criticalpoin tdrie d and mounted on stubs.Nex tth e specimenswer ebroke nwit h forcepsan dcoate dwit hgold .Th ebroke nsurface s wereobserve dwit ha Jeo lSE M 35Ca t1 5kV . For stereomicroscopica l observation ofprocesse s in livinganthers , flowerbuds of Gj_verrucos a andLiliu m hybrid cv. Enchantment were treated and observed asreporte d before (ch.2). Forhistochemistr y antherswer efixe di nFA A(5:7:90 )fo r1 8hour sa troo m
32 temperature and after dehydration in a graded ethanol-TBA-water series embedded in paraplast. 8um Sections were washed with xylol to remove the paraplast and next stained with zinc-chloro-iodine for cellulose, ruthenium red for pectins, aniline blue for callose (all according to Jensen,1962), calcofluor white M2R for both cellulose and callose (Hughes and MacCully, 1975) and phloroglucinol for lignins (Jensen, 1962). Sporopollenin structures were revealed with the acetolysis method of Erdtman (1960). The developmental stages of the locule were related to the nuclei after staining with acetocarmine (Jensen, 1962).
Observations
1.Th echange si nth ecel lwalls .
In fig. 1a the locule development of G. verrucosa is divided into equal intervals. The locule wall consists of an epidermis, an endothecium, one middle layer and a tapetum. The latter surrounds the developing pollen grains. During this development the cell walls of the tissues undergo remarkable changes (fig.1b-f). Before meiosis all the walls contain cellulose and pectins, the latter mainly situated near the middle lamella. During the meiotic prophase the sporogenous cells develop a callose wall between their pectin-cellulose walls and the plasma membrane, which gives the cell a spherical shape (fig.ld). During the second meiotic division pro-orbicules are deposited in the tapetum between the pecto-cellulosic walls and the plasma membrane on the inner tangential and bordering radial and transverse sides of the cells (fig.2). At the same time the pecto-cellulosic walls around both the tapetum and the meiocytes are lysed rather quickly (fig.1b,c) and the pro-orbicules are covered with drops of sporopollenin by means of membrane-like lamellae (fig.1f,3). Moreover, black lines appear outside the plasma membrane of the tapetum cells, which turn into a continuous granular sporopollenin containing sheet, surrounding the whole cell and interconnecting the orbicules (fig.1f,3). Although the outer part of this sheet has been referred to as "peritapetal or extratapetal membrane or wall" and the inner part as "orbicular wall" (Bhandari,1984), we just use the term "tapetal membrane", as in G. verrucosa it concerns a very thin structure which is morphologically similar on all sides of the cell. Just before this sporopollenin appearance around the tapetum cell, each of
33 the four microspores arising from one sporogenous cell is surrounded by an extra, thin callose wall. Next the cellulosic primexine is laid down, interrupted by the baculae of the sporopollenin containing exine (fig.1c,f,4), according to the process described for this species by Willemse (1972). After this callose has been quickly lysed (fig.ld), the tapetum cells and the microspores are merely covered by sporopollenin containing structures. In the exine cavities some cellulosic primexine remnants remain present (fig.1c,5). In stage 11 of fig.1 the pecto-celluloslc intine is deposited in the pollen grains. After the first pollen mitosis the generative cell is separated from the vegetative cell by a transitory callose wall (fig.ld). During the flattening of the tapetum, callose appears in the outer part of the radial and transverse walls bordering the middle layer (fig.ld), soon followed by the total lysis of the pecto-cellulosic walls of the latter. This callose disappears by the time of dehiscence (fig.ld). In stage 13 the cellulosic wall thickenings are deposited in the endothecium cells, which may become lignified (fig.1c,e). In the walls of the endothecium and the middle layer remarkable ultrastructural changes can be observed during this development. In the meiotic stage, gaps appear in the middle lamella between these two layers
1. Changes in the constitution of the cell walls in the locules of Gasteria verrucosa. The development is divided into intervals of approximately 30 hours. Continuous lines indicate the presence of the substance, dotted lines represent the circumference of the cells. A. The determination of the stages. 1=pollen mother cells. 2=leptotene- zygotene. 3=pachytene-diplotene. 4=metaphase-1. 5=metaphase-2. 6= late tetrad. 7=free microspores. 8-10=pollen vacuolation. 11= first pollen mitosis. 12=young gametophyte. 13-14=the generative cell dissociates from the wall. 15=mature pollen. 16=after dehiscence. Ep=epidermis, en=endothecium, ml=middle layer, tp=tapetum, sp= sporogenous tissue B.Th e localizationo fpectins . C.Th elocalizatio no fcellulose . D.Th elocalizatio no fcallose . E.Th elocalizatio no flignins . F.Th elocalizatio no fsporopollenin .
34 !€
I '""•-->:
>-f-i<-' rcttO -. '
H^
...;iri-... j
1HR •^ >
-c j—--?" J H_J>
$£-!£•$ UJ T- < DO
35 (fig.6a,b).Durin g thefurthe r developmentthes egap senlarg ei ntangentia l direction and during the deposition of the wall thickenings in the
2.Th etapetu man dth emeiocyte sdurin gth elat etetra dstage .Pro-orbicule s (PO)hav e appeared outsideth eplasm a membrane of the tapetum cell. Inside the callose wall ofth etetra dth eearl yexin e (Ex)i sdeposited . (2750x) 3.Th e orbicules are covered with sporopollenin by means of membrane-like lamellae.Th eplasm a membrane isreplace d by a sporopollenin-containing membrane (tm),whic hwil lb econnecte dt oth eorbicule si na late rstage . (60000x) 4.Insid e the callose wall the tetrads are covered with a cellulosic primexine (Pe).Th elatte ri sinterrupte db yth ebacula eo fth eexine , whichar einterconnecte db yth etectum .(104000x ) 5.Afte rth elysi so fth ecallose ,th etapetu man dth emicrospore shav ebee n coveredwit hsporopolleni n structures.Th etapetu mha sbee ncovere dwit h orbiculesan da tapeta lmembran e (TM).O nth emicrospore sth eexin e(Ex ) hasbee ndeposited ,stil lbearin gth e(hardl yvisible )primexin eremnant s in itscavities .(6000x ) 6.Th echange si nth eradia lwal lbetwee nth eendotheciu man dth emiddl e layer, (allfigure s22000x ) a.Th ewal ldurin gth epolle nmothe rcel lstage . b.Durin gmeiosi sgap sappea rnea rth emiddl elamell a(arrows) . c.Thes e gaps enlarge until the early stage ofwal lthickenin g (WT) depositioni nth eendotheciu mcells . d.Durin g dehiscence most of the wall has disappeared, the wall thickeningsremain .Th etapeta lmembrane s havebee npresse d against theendothecium .Betwee nthes etw ostructure scytoplasmi cremnant so f bothth etapetu man dth emiddl elaye r(ML )ar evisible . 7.Th e same stage as fig.6d. The U-shaped wall thickenings of the endothecium are interconnected by tangentially directed bars (TB).Not e the orbicules (0).(SEM )(11000x ) 8.Th etapeta lplastid s inth eyoun gmicrospor estage .Th estarc hgraduall y disappears,whil eelectro ndens emateria l(ED )wit hglobule sappears . (48000x) 9.I nth esam estag ea sfig. 8strand so felectro ndens emateria l (arrow) appeari nth etapetum ,bordere db yswolle nER-cisterns .(78000x )
36 •2* #"£•'??'
37 endothecium,th eentir ecel lwal laroun dth emiddl elaye ra swel la sth e innertangentia lendotheciu mwal lstar tt odesintegrat e (fig.6c).Finall yth e latter consists mainly of theU-shape d wall thickenings (fig.6d), whichar e connected by tangentially directed bars (fig.1e,7).
2.Th eformatio nan dtransfe ro fth epollenkitt .
Shortly after meiosis the largestarc h grains inth eplastid s of the tapetum gradually disappear,whil e moreelectro ndens emateria lgraduall y appears (fig.8). At the same time strandso f electron dense materialappea r inth ecytoplasm ,bordere db ystrongl yswolle nRE Rcistern s (fig.9).Nex t the electron dense material inth eplastid s turns into globules.Thes e globulesstar tt ofuse ,formin glarge ,les selectro ndens edroplets ,b ywhic h theplastid s strongly expand (fig.10) . Next the plastid membranes disappear andth edroplet sfus ewit hth estrand si nth edegeneratin gcytoplas mt ofor m hugeglobule s (fig.11).Finall ythes eglobule san dsom e cytoplasmic remnants (tryphine)ar eth eonl yconten to fth esack sforme db yth etapeta lmembrane s
10.Th eelectro ndens emateria li nth etapeta lplastid sturn sint olarg e droplets,th estarc hha sdisappeared .(12600x ) 11.Th etapeta lplasti dmembrane shav edisappeare dan dth edroplet so ffig .1 0 fusewit hth estrand s offig. 9t ofor m large globuleso fpollenkitt . (30000x) 12.Th etapetu m"cells "consis tmerel yo fth epollenkit tglobule s(Pk )an d somecytoplasmi cremnant s(tryphine) ,surrounde db yth etapeta lmembrane s withorbicule s (TM).Apar tfro mth eremnant so fth enucleu s (N),plastid s ina nearlie rstag eo fdegeneratio n(P )ar evisible .Th ecytoplas m ofth e middle layer (ML)ha sdegenerated .(2880x ) 13.Pollenkit tan dtryphin ehav e beenpresse d through thetapeta l membranes intoth elocula rcavity .Th epolle ngrain sar estuc kt oeac hothe ran dt o the orbicules. (6400x) 14.Drop s ofpollenkit t intrude theexin ecavitie s ofa naborte dpolle n grain.Th eprimexin e (P)an dth eweakl ydevelope dintin e (I)ar eals o visible.(28800x ) 15.Th eintercellula r spacebetwee ntw oendotheciu m cellsi sfille dwit h pollenkitt.Th etapeta l membranes andth eorbicule s arepresse dagains t theendothecium . (3000x)
38 39 (fig.12). During the last stages of tapetal degeneration the cytoplasm of the middle layer cells desintegrates in a relatively short time, quicker than the tapetum (fig.12). A few of these cytoplasmic remnants remain between the endothecium and the bordering tapetal membrane (fig.6d). After the lysis of the callose walls the developing pollen grains gradually expand up to anthesis (fig.1), pushing the almost degenerated tapetum and middle layer towards the rigid endothecium wall. This presses the pollenkitt and the tapetum and middle layer derived tryphine through the tapetal membranes into the spaces between the pollen grains and the exine cavities (fig.13), even of aborted pollen grains (fig.14). Furthermore, it can be found in the intercellular spaces between the endothecium and epidermis cells (fig.15). It sticks the pollen grains against each other and against the orbicule covered part of the tapetal membrane, which now borders the outer tangential tapetal membrane and the reduced endothecium wall (fig.6d,7). After the opening of the septum and the stomium the pressure between the pollen grains and the endothecium remains, due to the centripetal force of the locule walls (ch.2). Fig.16 shows the final situation in the mature anther, comparable with
16. The final composition of the mature locule. The expanding pollen have pressed the tapetal membranes with the orbicules against the endothecium and are covered with pollenkitt and tryphine. These pollen coatings can also be found between the swollen epidermis and endothecium cells (arrowheads). (1650x) 17. During dehiscence the orbicules approach each other following the shrinkage of the locule wall. (45000x) 18. The darkly stained pollenkitt of Lilium hybrid cv. Enchantment flows from the tapetum (T) into the locules, according to the arrows. Darker coloured, pollenkitt-covered pollen grains (d) can be distinguished from still uncovered, light coloured ones (1). In the left locule the flow of pollenkitt is retarded if compared with the right one. Stereo microscopical observation. (60x) 19. Pollen grains of Lilium hybrid cv. Enchantment have been reattached to a locule wall from which the orbicules were removed. Wet circumstances can shed the pollen. (3375x) 20. The intact situation on the dehisced anther. The pollen grains are stuck to the orbicules (0) and are not influenced by wet circumstances. (3375x)
40 £ " -(*** _.,ir>- }' • •
41 stage1 6o ffig.1 .Th eepidermi san dendotheciu mcell sar eswollen ,thei r wallsar epecto-cellulosic ,thei rintercellula rspace sma ycontai npollenkit t andtryphine .Th ewal lthickening so fth eendotheciu m sometimes contain lignin.Th emiddl elaye rha sdisappeare dan dth esporopolleni ncontainin g tapetalmembrane san dorbicule sborde rth elocule .Th estrongl yexpande d pollen grains are covered with the sporopollenin containing exine, bearing remnants ofth e cellulosicprimexin einsid e itscavities .Thi sexin ei s partially coveredwit hpollenkit tan dtryphine .A tit sinsid eth epecto - cellulosic intinei sfound . Afterdehiscenc eth eshrinkag eo fth elocul ewal lmake sth eorbicule s approacheac hother ,creatin ga mor econtinuou s layer (fig.17).
3.Experimenta lapproac ho fth eprocesses .
Themicroscopica lobservation si nfixe dmaterial swer e supportedb y some experimentswit hlivin gtissues . Therelationshi pbetwee nth epressur eo fth epolle nmas san dth etransfe r ofpollenkit t was checked withth e stereo microscopical observation of decapitated Lilium hybrida flowers (ch.2).I fth eflowe r isdecapitate d beforethi stransfe rha sstarted ,th ewhol eproces s canb eobserved .A sth e decapitationcause ssom e losso fpollen ,th eglas scovere danther snee dsom e lagtim et orecove rth eincreasin gpressure .Fro mth emomen to ftapetu mcel l repression, the dark red pollenkitt can be seen invading the locules. Moreover,thi smomen tappear st ob easynchronou sfo rtw oborderin glocules , althoughthi sma yb edu et oth emanipulatio n (fig.18).Artificia lpressur eo n theoutsid eo fth eanther s canaccelerat e thisprocess .Diminishin gth e pressureb yremovin gpar to fth epolle nca nretar do rpreven tit . Todetermin ewhethe ran yunderpressur emigh tsuc kth epollenkit tint oth e cavity,th elocul ewa sconnecte dt oth eatmospher eb ya thi nglas scapillary . Inspit eo fthi streatment ,th epollenkit ti stransporte di na norma lway .A n important observation is the absence ofpollenkit t inth eextrem e apexan d baseo fth elocul e after dehiscence. Inthes e sitesthi s substance isstil l foundinsid eth etapetu mcells . We alsoteste dth eide ao fHeslop-Harriso n (1968)an dHeslop-Harriso nan d Dickinson (1969)tha ta functio no fth eorbicule s mightb eth ecreatio no fa non wettable surface to enable an easy lifting of the pollen by their vectors.Fo rthis ,w eremove dth emixtur eo fpolle nan dpollenkit tfro ma dehiscedanther .Nex tw eremove dmos to fth eorbicules ,togethe rwit hth e
42 tapetalmembrane sfro mth einne rtangentia lendotheciu mwall .No wth emixtur e ofpolle nan dkit twa sreattache dt othi sreduce dlocul ewal l (fig.19).Give n thissituation ,mos to fth epolle nfel lof fth eanthe runde rextremel ywe t circumstances (rel.hum .98%) ,whil eintac tanther s(fig.20 )di dno tlos e pollen.Finall y bothtype s of dehiscedanther s reclosedunde r these wet conditions.
Discussion
1.Th echange si nth ecel lwalls .
Thechange si nth ecel lwall so fth elocul eca nb erelate dt oth eproces s ofpolle ndispersal . Themai nchang ei nth ewall so fth eepidermi si sth edevelopmen to fwal l ridges on the outside of the cells (ch.2)an d in the endothecium the depositiono fth eU-shape dwal lthickenings .Th erol eo fth ecel lwall so f thesetw olayer sha sthoroughl ybee ndiscusse di nth efirs tpar to fthi swor k (ch.2). Thechange si nth etapetu man dth edevelopin gpolle ngiv eris et obot h finalan dtransitor ystructures .O nth eborde rbetwee nth etapetu man dth e pollen grains the pectins and cellulose are gradually replaced by sporopollenin.Thi sals omean sa shif tfro mhydrophyli c intohydrophobi c properties (Heslop-Harrison,1968).Whil e the role ofth eexin eha sbee n studiedfrequentl y (seebelow) ,th efunctio no fth eorbicules ,appeare dmor e difficult to explain. Maheshwari (1950)suggeste d a role in the exine synthesis,probabl y asa sporopollenin stock.Echli n (1968)an dChristensen , Horner and Lersten (1972) regarded them as a relict, referring to a phylogeneticrelationshi pbetwee nth etapetu man dth esporogenou stissue . Dickinson andBel l (1972)regarde d them justa sa by-produc to fth etapetu m cells.A rea lautonomou sfunctio nwa sgive nb yHeslop-Harriso n (1968)an d Heslop-Harrisonan dDickinso n (1969),wh oindicate dtha tth eorbicule sforme d ano nwettabl esurfac eo nth elocula rinside ,effectin ga neas ydispersa lo f pollen. Our experiments inL ^hybrida , which has a secretory tapetum, strongly supportthi side aa sth eremova lo f orbicules changesth estickin g propertieso fth epollen(kitt )t oth elocul ewall .Heslop-Harriso n (1969) founda tapeta l membrane outside theperiplasmodia l tapetum ofCompositae . This was acetolysis resistant as are the exine and the orbicules. As
43 periplasmodialtapeta lprotoplast s invade the locule,thi stapeta lmembran e finally forms theborde r between the locule wall and thepolle n grainsan d may have a similar function as the orbicules in the secretory tapeta. However, thepolle n sticking substance in this species is mainly tryphine, i.e.hydrophyli c remnantso fth etapeta lcytoplasm ,whic h makesth estickin g complex different fromL .hybrida .Pacin i andJunipe r (1983)repor t the deposition ofPAS-positiv epyramid-shape d structures onbot h theexine san d thetangentia loutsid eo fth eperiplasmodia ltapetu mo fAru mitalicu mMiller . Although this material isacetolysi s susceptible, itspresenc e supportsou r hypothesistha ta tdehiscenc eth einsid eo fth elocul eha sa simila rphysica l constitutiona sth eoutsid eo fth epolle ngrains .Anothe ragreemen twit hthi s hypothesis isgive nb yCarnie l (1966)wh oreporte d orbicules facing the locular cavity onth etangentia l outsideo fth esecretor y tapetum ofGnetu m gnemonL . Moreover,thi ssit ei s not situated on thepathwa y from the tapetum toth epolle ngrains ,whic hmake sMaheshwari' side ao fa rol ei nth e exinesynthesi s lesslikely . Given ourhypothesi s of the similarconstitutio no nbot hth elocul ewal l andth epolle ngrains ,th equestio narise sa st owh yth emorpholog y ofthes e structures iss odifferent .Thi swil lb edu et oothe rfunctions .Apar tfro m the similarity mentioned above,th e exine is a continuous, protective structure, which iselasti can denable sth eharmomegath yo fth ecel ldurin g hydration changes. Besides, it has sites to transport proteins for recognitionreaction s(Knox ,1984) ,whil eit ssculpturin gma yb eimportan tfo r vectors andstigm a contact (Woittiezan dWillemse,1979 ; Knox,1984).Th e tapetalmembrane swit hth eorbicule s dono thav ethes efunctions .I fou r hypothesis is right, their main role is the formation of a dispersive surface.Durin gth eshrinkag eo fth elocul ewal la tdehiscence ,th eorbicule s canapproac heac hother ,coverin gth ewettabl e cellwall s ofth e locule almostcompletely . Contraryt oth esporopollenin ,al lth ecallos edeposit sar etransitory . Callosei softe nfoun da sa nisolatio n between cellstha tstar ta divergen t development or in sites wherea transport mustb eblocked . Heslop-Harrison andMacKenzi e (1967)an dSouthwort h (1971)demonstrate dtha tth ecallos e walls during meiosis form a selective barrier for certain molecules, presumably relatedt oth eindependen tdevelopmen to f thepolle n grainsfro m that moment. Owens andWestmucket t (1983)studie dth e callose wallbetwee n thevegetativ ean dth egenerativ ecel li ndifferen tspecie san drelate di tt o cytoplasmic rathertha nnuclea rdifferentiation .Th ecallos eoutsid eth e
44 tapetum justbefor edehiscenc ema ypreven ttranspor to fliqui dt oth elocule , asth espac ebetwee nth epolle ngrain si s"dry "a ttha tmomen t (ch.1),whil e theepidermi s andendotheciu m arefull yturgescent .However ,thi sneed s furtherinvestigatio n inrelatio n topossibl e barrier capacities ofth e tapetal membranes, according to Heslop-Harrison's (1969)ide ao f a"cultur e sac".
2.Pollenkit tan dtryphine .
Aplastidia l origin of part ofth epollenkit t andth e fusiono fthi s materialwit hcytoplasmi c lipidswa sals ofoun d inL .hybrid ab yReznickov a andWillems e (1980)an dReznickov aan dDickinso n (1982). Theobservatio n thatth epollenkit t is mixed with a small amount of cytoplasmicremnant s(tryphine )durin git stransfe rt oth elocul eagree swit h thefinding si nchapte r4 ,wher ew erelate dth epollenkitt/tryphin e ratiot o thephylogeneti c moment of transfer toth e locule. Wepresum e thetryphin e breakdown (ortur nint opollenkitt? )t oceas eafte rit sreleas efro m the tapetumcell .Hess e(1979a,b ;1980 )relate dth eshif tfro mentomophyli ct o secundaryanemophyli cpolle ntranspor tfo rsom especie st oa decreas eo fth e amounto fpollenkitt ,whic hi nou ropinio nmigh tdecreas eth epollenkitt - tryphineratio .Consequently ,thi sshif tma yb erelate dt oa phylogeneti c accelerationo fth epollenkitt-tryphin e transfert oth elocule . Although many researchers reportth edepositio n ofpollenkit t onth e pollen grains (see the review of Bhandari,1984), little data about the transportmechanis m from thesecretor y tapetum toth e loculear eavailable . Reznickova andDickinso n (1982)presum e arelationshi p with the increaseo f thesurfac eare ao fth etapeta lplasm amembrane ,bu ti nG^ _verrucos a sucha n extension isno tfound .Ou r experiments suggesta relationshi pwit hth e pressureo fth epolle npopulatio ninsid eth elimite dcapacit yo fth elocule . Duet oa nincreas eo fth ecytoplas m (Willemse,1972)an dth edisappearanc eo f thelocula rflui d (ch.1)th epolle nmas sexpand st ofil lal lo fth espac e insideth einne rtangentia lendotheciu mwalls .Th eonl ypossibilitie sfo rth e pollenkitt-tryphinemixtur ear ewhethe rt ofloa tbetwee nth epolle ngrain so r to leaveth e locule,bot h of whichar eobserved .Th epresenc eo fpollenkit t inth e exine cavities of aborted pollen grains as well as between the epidermisan dth eendotheciu mcell smake sa nattractiv erol eo flivin gpolle n grainsfo rth etransfe runlikely .It sabsenc ei nth eape xan dth ebas eo f someL .hybrid aanther san dth eeffec to fartificia lpolle nremova lagre e
45 with thisstatement .I nou ropinion ,du et oth elac ko fspace ,capillar y forcesi nth ecrowde d loculepermi tth epollenkit tt oexten dove rth ewhol e pollenpopulation ,instea do fonl yth epolle ngrain stha tar ei ncontac twit h thetapetum .Als oit spresenc ebetwee nth eepidermi san dth eendotheciu m cellsmake sa capillar yforc elikely .Th epossibilit yo fa vacuu ma sa resul t ofth euptak eo fth elocula rflui db yth epolle n grains (ch.1)i sexclude db y our experiments with artificially opened locules.Moreover , if vacuum existed,th epolle ngrain swoul dlos ethei rwate r immediately (ch.2). Hesse (1978,1979b,1980) found a decrease of the sticking effect of pollenkittwhe nshiftin g fromentomophyli ct osecundar yanemophyli cspecies . Dependingo nth especie sh efoun deithe rmos to fth epollenkit tt oremai ni n thetapetu mcells ,pollenkit tdee pinsid eth eexin ecavitie so rth epresenc e oflittl eo rn opollenkit ta tall .I twoul db einterestin gt orelat ethes e phenomena toth etranspor t mechanism hypothesized here,fo r example to convertentomophyli cint oanemophyli cspecie sb ymanipulation . Besidesa rol ei nsticking ,whic hi srathe rclear ,especiall yafte rth e comparative observations ofHesse ,an dth eabsenc eo fpollenkit ti nth e (primary anemophylous)Gymnosperm s (Hesse, 1984), pollenkitt mayhav eothe r functions.I tma y takepar ti nth eexin eformatio n (Willemsean dReznickova , 1980)an dma yattrac tan dfee dpollinator s (Willemse,1985).Preventio no f bothdehydratio nan dU Vdamag e(Willemse,1985 )i sunlikel yi nG .verrucos aa s weobserve dth ecoatin gno tt ocove rth eentir epolle ngrai ni nthi sspecies . Giventh eresult so fthi spaper ,th ecentripeta lforc eo fth eendotheciu m beforedehiscenc eca nb e related to more functions thanw e reportedbefore . Inth efirs tpar to fthi swor k (ch.2)w efoun dthi sforc et oope nth estomiu m andt okee pth elocule sclosed ,preventin gprematur e losso fpollen .A thir d function may bet ocontinu e theeffec to fth epolle npressur e afterth e stomium isopened .Moreover ,thi spressur elead st oth eequa ldistributio no f pollen on the locule walls at dehiscence. The rigid inner tangential endotheciumwal lenlarge sth eeffec to fthi spressure ,althoug hthi swil lb e lessi nspecie swit hpersistin g middle layers,lik eL .hybrid a(Willems ean d Reznickova,1980).
Conclusions
Changesi nth ecel lwall so fth elocul etissue so fG .verrucos alea dt o thepresenc e ofhydrophobi c substances on both sides of the border between
46 the sporophyte andth e gametophytea tdehiscence .Thes e areth etapeta l membranesan dth eorbicule so nth etapeta lsid ean dth eexine so nth eother . Alsodat afro mth eliteratur eabou tothe rspecie ssho wa tendenc yt ofor m similarsubstance so nbot hside so fthi sborder . InG^ _verrucos ath ehydrophobi cpollenkit tstick sth epolle ngrain st o eachothe ran dt oth eorbicule san dth etapeta lmembranes .I nL^ _hybrid athi s complexappear st oresis twe tcircumstances ,b ywhic hi tprevent spolle n dispersalb ywater , i.e.a nundesire dvector . Thepollenkit ti spresumabl y transferredb ycapillar y forcesdu et oa limitedspac ei nth elocule ,create db yth eexpansio no fth epolle ngrains .
47 Chapter 4
ACOMPARISO N OFANTHE RTISSU EDEVELOPMEN TI NMAL ESTERIL EALO EVER AAN DMAL E FERTILEALO ECILIARI S
C.J.Keijze ran dM .Cresti(* )
* Departmento fEnvironmenta lBiology ,Botan ySectio n Universityo fSien a- Vi aP.A.Mattiol i4 - 5310 0Sien a(Italy )
Researchsuppor tb yCNR-IPR Aprogram ,su bprojec tno.1 ,pape rno .770 .
Summary
Cytological differences between theanthe r developmento fa mal esteril e anda mal efertil eAlo especie sar euse dt oexplai ninteraction s between anther tissues. Deviations in the layers of the locule wall and the microsporeso fth emal esteril eanthe rar erelate dt oeac hothe ran dthei r biological functionsar ediscussed . The cytological development ofth emal esterility , which canb eobserve d shortlyafte r meiosis,seem st ob erestricte dt oth elocula rcavity .Th e tapetaldevelopmen tan dbreakdow nar enormal ,apar tfro mth esiz eo fsom e orbicules. However, thepollenkit t isno ttransferre d toth epolle ngrains , which strongly supportsou rtheor y thatthi sproces si smechanicall ypollen - controlled.Th edevelopmen to fth eepidermi san dendotheciu mcell si snormal , except in parto fth eanther swher ethes ecell sd ono texpand , afterwhic h dehiscence isincomplete .Th elatte rproces si sdiscusse d inrelatio nt oth e deviations insideth elocula rcavity .
Introduction
Inth elarg eamoun to fliteratur eo nbot hgeneti can dcytoplasmi cmal e sterility (see the reviews by Edwardson,1970 ; Laser and Lersten,1972; Gottschalk andKaul,1974 ;Bhandari,1984 )man yauthor srepor tirregularitie s
49 inth etapetu m andsometime sth eothe rlocul ewal ltissues .Difference s betweenth elocul edevelopmen to ffertil ean dsteril eanther so ftw oAlo e species offer apossibilit y to study functions of and interactions between thesetissues .Suc ha stud ymigh tb eusefu li npolle nresearch ,a sdevelopin g pollengrain sar eprobabl y dependentupo nmor eanthe rwal ltissue stha nonl y thetapetu m (Reznickovaan dWillemse,1980) .Moreover ,i tmigh telucidat eth e function of these tissues in relation to anther dehiscence and pollen dispersal.Th elatte rprocesse shav ebee ndescribe di nchapte r2 an d3 .A s malfunctioningo fth eprocesse sleadin gt oanthe rdehiscenc eca npreven t pollenexposure ,whic hha sth esam eeffec ta smal esterility ,thi sstud yma y beo f greatinteres tfo rplan tbreeding . Thischapte rdeal swit hsom efunction so fth elocul ewal ltissues ,enable d by differences between fertile and sterile anthers and developmental gradients insideth elatter .
Materialsan dmethod s
Anthers ofth e malesteril eAlo ever a (L.)Burm .fil .(Liliaceae )wer e followeddurin gthei rdevelopmen tusin gligh tan delectro nmicroscopica l techniques.A sn ofertil eplant so fthi sspecie swer eavailable ,comparison s weremad ewit hAj _ciliari sHaw. ,th eanthe rdevelopmen to fwhic happear s ultrastructurally indistinguishable fromth esteril especies .Bot hspecie s grewi nth eBotanica lGarde no fth eUniversit y ofSien a(Italy) . Fordar kfiel dmicroscop yintac tanther swer ecleare dfo r5-1 0minute s with the solution described by Herr (1971). In this way crystals, sporopolleninan dlignifie dstructure sbecam evisible . Forelectro nan dinterferenc econtras tmicroscop yanther swer efixe di n3 % glutaraldehyd ei n0.07 Mcacodylat ebuffe ra tp H7, 4fo rtw ohours .Afte r rinsingi nth ebuffer ,th emateria lwa spostfixe di n 1%osmiu m tetroxidei n buffer for4 5minutes .Bot h fixations werecarrie dou ta troo mtemperature . After dehydration ina graded ethanol series, the material was embeddedi n thelo wviscosit yresi no fSpur r(1969) . For transmission electron microscopy ultrathin sections werepoststaine d withurany lacetat ean dlea dcitrat e (Reynolds,1963) .Polysaccharide s were tracedb yTCH-Ag-proteinate ,accordin gt oThier y (1967).Observation swer e madewit ha Jeo lJE M 100Ban da Zeis sE M9 ,bot ha t6 0kV . Forinterferenc econtras tmicroscop y2u mcutting so fth esam eblock swer e
50 enclosedi na dro po fSpurr' sresi no na microscop e slidean dcovere dwit ha coverslide . Sporopollenin structures were isolated by enclosing such 2um cuttingsi n concentrated sulphuric acid.Thi s isa varian to fth eacetolysi s method describedb yErdtman n (1960).Th eligh tmicroscopica lobservation swer emad e witha Niko nOptipho twit hinterferenc econtras tequipment . Forscannin gelectro nmicroscop ynaturall ydehisce dan ddrie danther swer e sputteredwit hgol dan dobserve di na Philip sSE Ma t7, 2kV .
Results
TheAlo e anthers consist of four locules,boun d togetherb ya connectiv e tissuewit ha centra lvascula rbundle .Tw oborderin glocule sdehisc eb ymean s ofon elongitudina lslit .Th elocul ewal lconsist so ffou rcel llayers :a n epidermis,a nendothecium ,on emiddl elaye ran da tapetum . InA; _ver ath eabortio nproces swa sfollowe di nintac tanther safte ra treatment with clearing solution (Herr,1971). Beforeth e sporopollenin depositionstarts ,onl yth exyle man dsom ecrystal sar evisibl e(fig.la) . Onceth e exineformatio nha sstarted ,sporopolleni nbecome svisibl ean d afterdissolutio no fth ecallos ewalls ,th emicrospore stightl yfil lth e locules(fig .1b) .Afte rthi sstag ethe ystar tt odegenerat ean dgraduall y shrink,leavin ggap si nth epopulatio nb yth etim eth ewal lthickening si n theendotheciu mcell sappea r (fig.lc).U pt odehiscenc emos to fth ecollapse d microspores retracttogethe r with thelocula rflui dint oth etermina lpart s ofeac hlocul e (fig.ld).Th ecentra lpar tbecome sfille dwit h gas (air?), which is not visible with this technique. Just before dehiscence the microsporesar eonl ypresen ti nth eextrem eto pan dbase ,wher ethe ystil l can be foundafte r dehiscence (fig.le). Inth elat etetra dstag eth emicrospore sinsid eth ecallos ewall sstil l developi na simila rwa yi nbot hspecie s (fig.2).I nth ecytoplasm-ric hcell s smallvacuoles ,spherica l mitochondria andplastid s arepresent .Th efutur e colpusi sline db yon elon gshee to fER ,i nth eres to fth ecytoplas mE Ri s scarce. Inthi sstag eth etapetu mo fbot hspecie sconsist so fbinucleat ecell s with much dispersedE R andsmal lvacuole s throughout the cytoplasm (fig.3). Plastidscontainin gstarc ha swel la smitochondri aar epresent .Th ecel lwal l isstil lintact ,toward sth elocul epro-orbicule sla youtsid eth eplasm a
51 membrane (fig.2). Afterth ereleas efro m thecallos ewall ,th efirs tsign so fsterilit yca n beobserve di nth emicrospore so fA .vera .Th epolarit y inthes emicrospore s differsfro mth efertil esituation .Th enucleu smigrate sfro mth ecolpu s side,wher ei ti ssituate di nth eyoun gtetra dstage ,t oa centra lpositio n inth e cell (fig.4). Inth e fertile speciesthi smigratio n iscontinue d towardsth eopposit ewal l(fig.5) .Stack so fER ,surroundin gth enucleu si n A.ciliari s (fig.5),ar eabsen ti nA .ver a(fig.4) .Besid ethes edifference s someprocesse s inth e sterile development still takeplac e like inth e fertilesituation .Th epremitoti cvacuole sappea ran dth eplastid sstor eman y small starch grains (figs.4an d 5).Th eexin ethicken s andth e intinei s deposited. Inthi sstag eth etapetu mo fth esteril especie sshow sonl yon evisibl e deviation from thefertil e situation (fig.6), as some ofth eorbicul ecore s swell enormously andar eirregularl y coveredwit hsporopolleni n (fig.7). Apartfro m this, nodifference sbetwee nth esteril ean dth e fertiletapeta l development canb eobserved .
1.Th edevelopin gsteril eanther .Figs.a-d :dar kfiel dilluminatio nafte r clearingwit hHerr' ssolution ,fig.e :SEM . (17x) a.Durin gmeiosi sonl yvascula rtissu ean dcrystal sar evisible . b.A tth eyoun g microspore stageth eforme d exineshav ebecom evisible . Theysho wtha tth emicrospore stightl yfil lth elocules . c.Th ewal lthickening shav e appeared inth eendotheciu m (arrow), empty spacesappea ra sblac kdot sbetwee nth edegeneratin gmicrospores . d.Th ematur eanthe rwit hcollapse dmicrospores ,mainl ysituate dnea rth e apexan dth ebas eo fth elocules . e.Afte rdehiscenc eth emicrospore sar estuc konl yt oth eextrem eape x andbas eo fth eanther . 2.Th e late tetrad stage of the sterile species. P=plastid, M= mitochondrion,V=vacuole ,C=callose ,E=exine ,Cp=colpus .(4080x ) 3.Th e tapetum of the sterile species duringth e latetetra d stage.P = plastids,M=mitochondria ,V=vacuoles ,PO=pro-orbicules .(1375x ) 4.Th e degenerating micropore of thesteril e species.Th enucleu sha s migratedt oth ecentr eo fth ecell .P=amyloplasts ,V=vacuole .(3500x ) 5.Th e microspore ofth e fertile species.Th enucleu sborder s thewal l oppositeth ecolpu s (Cp),surrounde db ystack so fER .(3500x )
52 53 After this young microspore stage the microspores continue their degeneration, which is visible in all the cytoplasmic structures. Contrary to the expanding cells of the fertile species (figs.8 and 10), the sterile cells lose their turgescence (fig.11). Their nuclei shrink and the plastid membranes disrupt by which the starch grains are set free (fig.9). Finally the remaining cell contents are lysed, the released starch grains being the last recognizable structures. The tapeta in the sterile and the fertile anthers follow a similar degeneration process, in which no ultrastructural differences between the species can be observed. The cells finally contain mainly pollenkitt and some cytoplasmic remnants (tryphine). However, in the sterile species these substances are not transferred to the locule and remain inside the acetolysis resistant tapetal membranes (figs.12, 13 and 14), after which the tryphine disappears. On the contrary, in the fertile species the tapetum cells are flattened and the pollenkitt and tryphine are deposited on the pollen grains (fig.15). In the sterile species most of the pollenkitt is still inside the tapetal membranes after dehiscence (fig.16). In both species some small drops
6. The orbicules of the fertile species. (21600x) 7. A deviating huge orbicule (left). The core is much larger than in the normal orbicule (right), the sporopollenin layer is discontinuous. (21600x) 8. The microspores of the fertile species divide into a vegetative (V) and a generative cell (G). Only the former contains plastids (P). (4000x) 9. Degenerated microspore of the sterile species, stained with TCH-Ag- proteinate. Most of the cytoplasm has disappeared, the amount of vacuoles has increased. The plasma membranes disrupt and the starch grains are set free. (5200x) 10. The fertile anther just before the first pollen mitosis. The enlarging microspores tightly fill the locule. (900x) 11. The sterile anther from the same flower length as in fig. 10. The microspores degenerate and do not expand like in the fertile species. (900x) 12. The locule wall of the sterile species shortly before dehiscence. The epidermis and bar-supported endothecium cells have expanded, the middle layer has disappeared. The pollenkitt is surrounded by the tapetal membranes, the degenerated microspores have collapsed. (900x)
54 iis»-;fc- J© - >*• *»&4 9r>,
¥m*#•o 12r-r^
55 ofpollenkit tar efoun di nth eintercellula rspace sbetwee nth eepidermi san d theendotheciu m (fig.17). Inth esteril e species the locular fluid with TCH-Ag-proteinatepositiv e material,whic hi sno tfoun dbetwee nth etapeta lmembrane s (fig.13),persist s until dehiscence, gradually retracting together with the degenerating microspores intoth etermina lpart so feac hlocul e (fig.ld).Whe nth ega si n thecentra lpar to fthes elocule sappears ,th estomiu mi sstil lclosed . Inth esteril eanther sth eepidermi san d endothecium develop likei nth e fertilespecies .Tw oday sbefor edehiscenc ethei r cellsswel lb yvacuolatio n (figs.12an d 15).Wal lridge sdevelo pi nth eoute rtangentia lwal lo fth e epidermis cells, enlarging the cuticular surface (fig.18). In the endotheciumpar to fth estarc hdisappear s fromth eplastid san dth eU-shape d wallthickening s aredeposite d (figs.12an d15) . Inbot hth e fertilean dth e sterile speciesth e middle layer cellsdon' t
13.I n the sterile species a large drop ofpollenkit t is surrounded by tapetal membranes.Th eres t of the cytoplasm has almost completely disappeared. Remnants of the disappeared middle layer are visible (arrow).Insid eth elocul e(L )flui dwit hgranula rmateria li svisible . (5200x) 14.A transvers epar to fth etapeta lmembrane safte racetolysi so fa 2p m crosssection .O nth elocula rsid eorbicule sar evisible ,attache dt oth e (unfocussed)inne rtangentia lpar to fth etapeta lmembrane .(3400x ) 15.Th ematur eanthe ro fth efertil especie sjus tbefor e dehiscence.Th e pollenkitt (arrow)ha sbee ntransferre dt oth epolle ngrains ,th etapeta l membraneswit hth eorbicule sar epresse dagains tth eendothecium . Inth e pollen grainsth evegetativ e nucleus andth eelongate dgenerativ ecel l are visible. (900x) 16.Th e anther of the sterile species after dehiscence. Most of the pollenkitti sstil li nth esam esit ea sbefor edehiscence ,i.e .fig.12 . (900x) 17.Afte r dehiscence small drops ofpollenkit t are found intercellular betweenth eepidermi san dth eendothecium .(3600x ) 18.Th ecuticl ei slifte db yth eformatio no fwal lridges ,a phenomeno nwhic h isals ovisibl ei nfigs.1 2an d 15.(35000x ) 19.A partiall ydehiscin ganthe ro fth esteril especies .Th estomiu mi sstil l closed,a si sth e(invisible )septu ma littl edeepe ri nth eanther .(60x )
56 v.' ^rs
57 divide anticlinallyan dar eflattene ddurin gth econtinuou senlargemen to f the locularcircumferenc e (figs.10,11 ,1 2an d 15).Shortl ybefor e dehiscence the cell disappears in both species,a fe w remnants ofth e wall remain visible(fig.13) . About60 %o fth esteril eanther sdehisc ei na norma lway ,accordin gt oth e process described inchapte r 2an d 3 forGasteri averrucos a (Mill.)H.Duval . Thecell si nth etissu e betweentw oadjacen tlocules ,calle dth eseptu m (Venkatesh,1957),dissociate ,afte rwhic hth etapeta l membranes arebroke n andth estomiu mi sopened .Afte ranthesi sth elocul ewall sben doutwar db y whichth eanthe rdehisces ,bearin gth edegenerate dmicrospore snea rit sape x andbas e (fig.le),accordin gt othei rfina lpositio ninsid eth ematur eclose d anther. Theremainin g40 %o fth esteril eanther sonl ydehisc ei nth eregio nwher e themicrospore sar efound .I nthei rcentra lpar tth eseptu man dth e stomium remain closed(flg.19) .
Discussion
Althoughthi schapte rdeal swit hmal esterility ,i twil lb efocusse do n thefunction so fth edifferen t loculetissues .Th eproces so fmicrospor e abortionitsel fwil lonl yb ediscusse dwhe nnecessar yfo rthi stopic .
1.Th etapetu man dth elocula rcavity .
Inmal esteril especie spolle nabortio ni softe nprecede db ydifferen t kinds of irregularities in the tapetal development, like organelle deformations,increase d vacuolation andpersistenc eo renlargemen to fth e cells (see the review of Bhandari,1984) . In A;_ vera all the visible deviations,i.e .th elarg eorbicules ,th epersistenc eo fth elocula rflui d and the microspore degeneration are restricted to the locular cavity. Althoughth eorbicul eformatio ni si nth emai ncontrolle db yth etapeta l cells(Reznickov aan dWillemse,1980) ,th epresenc eo fsom ehug eorbicule sma y be associated with the deviations inth e locule.A s the orbicule cores containpolysaccharide s (Reznickova and Willemse,1980), the accumulation of TCH-Ag-proteinatepositiv e substances inth e locule may lead toa continue d supply ofcompound s for the cores,causin g their enormous growth. The appearanceo fthi sdeviatio ni njus ta numbe ro fth eorbicule sma yb edu et o
58 theirsequentia lformatio nb yth etapetum .Th efac ttha tthi sdeviatio ntake s place inside the locule means thatth etapeta l development is normalunti l thestag ei nwhic hth epollenkit tha st ob etransferred . Thefindin gtha t thepollenkit t isno ttransferre d toth epolle n grains agrees with manipulationexperiment s inlivin ganther so fth emal e fertile Gasteriaverrucos a (Mill.)H.Duva l andLiliu mhybri d cv. Enchantment (ch.3). These species show that thepollenkit t is transferred by capillary action, duet oth eexpansio no f the developingpolle n grains insideth e limited locular space, by which the tapetum is repressed. Accordingly, this repressioncoul db eprevente db yprematur eremova lo fpar to fth epolle nfro m the locule, after which the pollenkitt remained between the tapetal membranes.Als oi nth emal esteril eA .ver ath eabortin gmicrospore sdon' t expandt ofil lth eentir elocul ean dcaus en opressure .Th eonl ycapillar y forcetha tremain sappear sa tdehiscenc ei nth eintercellula r spacesbetwee n thedehydratin gendotheciu man depidermi scells .Indeed ,i nman yo fthes e spacespollenkit tca nb efound ,lik ei nth emal efertil eA^ _ciliari san d earlierobservation s inG .verrucos a (ch.3).I nA .ciliari sth erepressio no f thetapetu mb yth eexpandin gpolle nmas soccur swhe nth elysi so fcytoplasmi c remnants (tryphine)i sstil li ncourse .Her ea mixtur eo fpollenkit tan d tryphinei ssucke dint oth elocule ,afte rwhic hth elysi so ftryphin estops , likei nG ^verrucos a (ch.3).I nth esteril especie sth etryphin ebreakdow ni s continuedbetwee ntapeta lmembranes ,resultin g inmerel ypollenkitt .Thi s meanstha tth emomen to fth etapeta lmembran eruptur edetermine swhethe rth e finaltapeta lproduc t isa mixtur eo fpollenkit tan dtryphin eo rpollenkit t only. Thedevelopmen ti nth esteril especie sshow stha tinteraction sbetwee nth e tapetuman dth emicrospore so rth elocula rflui dapparentl y occuri non e direction, asth e development ofth eforme ri sno t influencedb ydeviation s inth elatter ,wit hth eexceptio no fth emechanica lproces so fpollenkit t transfer. Theexistenc e of acetolysisresistan ttapeta lmembrane sa tal lside so f thetapeta lcells ,foun dhere ,wa sals oreporte di nG .verrucos a (ch.3an d 5).Apar tfro mthes etw ospecie si twas ,a sfa ra sw eknow ,no treporte d before. Banerjee (1967)wa s the first to show an acetolysis resistant membrane,interconnectin gth eorbicules ,alon gth esecretor ytapetu mo fsom e grasses.Heslop-Harriso n(1969 )foun da tapeta l membrane onth e tangential outsideo fth eperiplasmodia ltapetu mo fCompositae .Willems ean dReznickov a (1980)foun dbot hmembrane si nLiliu mhybri d cv.Enchantment ,whic hha sa
59 secretorytapetum .I nou rrepor tth einne ran doute rtangentia lmembrane s appear tob e interconnected by radialan d transverse membranes, the latter clearlyvisibl ei nfig.14 ,b ywhic hth esphere so fpollenkit tar ekep ti n theirorigina lposition .I ngeneral , aclea robservatio n of these membranes is difficult because of the crowded structure ofth e locule,mentione d before.However ,th eacetolysi s method clearly revealsthes e structures,a s this treatment destroys the whole anther, except these membranes, the orbicules and the exines.I n our opinion,a n important function of the tapetalmembrane s ist okee pth eorbicule sconnecte dt oth elocul ewall , wherethe ypla ya rol ei npolle ndispersa lafte rdehiscence ,a sindicate db y Heslop-Harrisonan dDickinson , (1969)an ddemonstrate di nchapte r3 .
2.Th eoute rlayer s
After Gupta and Nanda (1973)reporte d the independent, asynchronous depositiono fendothecia lwal lthickening s inindividua l loculeso fon e anther,i tca nb eassume dtha tthi sproces si sinitiate d insideeac hlocule . Moreover,thi sasynchronit y means thatth e signal for depositionoriginate s frominsid eth eanthe ran dno tfro mothe rflora lparts .I nth esteril eA . veraw e founda norma ldevelopmen t ofendothecia lwal lthickening sove rth e wholelengt ho fth elocules ,whic hmean stha tthi sdepositio ni sneithe r dependentupo npolle nmaturation .Thi s conclusion agrees with results ofD e Fossard (1969),Gupt aan dNand a (1973),Chauna n(1979 )an dChaunan ,Dingr a andKinoshit a(1982 )i nbot hnatura lan dchemicall y inducedmal esteril e species.Thes eresult srestric tth einitiatio no fthi sproces st oth elocul e walltissues .D eFossar d(1969 )presume sth eflo wo fan yinhibito ro fwal l thickeningdevelopmen tfro mth etapetu mt oth eendothecium ,whic hwoul dceas e from the start of the tapetal degeneration, i.e. the moment that the sporopollenin deposition has finished. However, Gupta andNand a (1973) demonstratedtha tth eappearanc eo fth ewal lthickening sha sn ochronologica l relationshipwit hth een do fth esporopolleni ndeposition .Besides ,afte rth e depositiono fth ewal lthickenings ,w e findi nA^ _ver aobviousl y normalRE R andpolysome si nth etapeta lcytoplasm .Thi smean stha ti fth einitiatio n originatesfro m thetapetum , thesynthesi so fan ystimulu si sals opossibl e insteado fth edisappearanc eo fa ninhibitor . Aninterestin gphenomeno n isth eretractio no fth elocula rflui dtogethe r with themicrospore s toth e distalan dproxima l ends ofeac hlocule ,whil e thecentra lpar tbecome sfille dwit hgas .Sinc ethi sretractio nconcern stw o
60 oppositeends ,an yeffec to fth egravit yi sexcluded .I tseem st ob ecause d bya kin do fcapillar ysuctio no fth einitiall ysmal lamoun to flocula rflui d towardsth enarrowes tpart so fth econtinuousl yenlargin glocule ,i.e .th e apexan dth ebase .I nth efertil especie sth elocula rflui ddisappear sbefor e dehiscence,presumabl y intoth eexpandin gpolle ngrain s (ch.1an d 5). Iti s replacedb ygas ,indicatin gn ofurthe rsuppl y of liquid from or throughth e tapetumint oth elocula rcavity .Th eappearanc eo fga si nth esteril especie s fromth emomen ttha tth eamoun to flocula rflui dbecome sinsufficien tt ofil l theenlargin g locule,agree swit hthi sendin gsuppl y inth efertil especies . Thequestio nremain sa st oho wga sca nente rth egrowin glocule ,whil eth e stomiumi sstil lclosed .Th estomat aan dintercellula rspace sma ypla ya rol e inthi sprocess . Fromth edevelopmenta lgradien ttha twa sfoun di nth elocul ewal lo fabou t 40% ofth esteril eanthers ,som eimportan tconclusion sca nb edrawn .Als oi n theseanther sth elocula rflui dca nb efoun di nth etermina lpart so fth e locules.I nth ecentra lpar tth eepidermi san dendotheciu m dono texpan dan d bothth eseptu man dth estomiu mremai nclose da tdehiscence .O nth econtrary , inth etermina lpart so fthes e anthersth eepidermi s and endothecium cells doexpand ,th eseptu mcell sar edissociate dan dth estomiu m isopened .S o epidermisan dendotheciu mexpansio nan danthe rdehiscenc eonl yoccu ri nsite s where locular fluid is present. This suggests that the epidermis and endotheciumcell sus e( apar to f? )th elocula rflui dfo rthei rexpansion . Thishypothesi si ssupporte db yth eobservatio ntha ti nthi sstag eth emiddl e layerha slos tmos to fit scontent san dth etapetu mmainl yconsist so fth e hydrophobicpollenkitt .Moreover ,i tmean stha tth epolle ngrain sar eno tth e onlycell stha ttak eu plocula rflui dan dtha tals oepidermi san dendotheciu m influenceth e dehydration ofth e locular cavity.However ,thes ehypothese s requirelabellin gexperiment st ob eproved . Asbot h the swelling of the epidermis and endothecium cells andth e rigidityo fth einne rtangentia lendotheciu mwal lcaus eth einwar dbendin go f the locule walls,b y which the stomium isopene d (Woycicki,1924b ; ch.2),i t isclea rtha tth e lessexpandin g epidermis andendotheciu m cells inth e centralpar t of thedeviatin g antherskee pth estomiu m closed.Th equestio n remainsa st owh yals oth eseptu mcell sd ono tdissociate .Contrar yt oth e stomium, theopenin go fthi stissu ei sa lytica lproces swithou tmechanica l forces (Woycicki,1924a; Becquerel,1932; ch.2). In generalth efunctio no f an openingseptu m mightb eth ecreatio no fmor espac efro mth emomen ttha tth e volume ofth eexpandin g pollen exceeds the capacity of the two locules
61 (ch.2).Therefore ,a nopenin gsigna lmigh toriginat efro mth epollen ,th e locularflui do rth etapetum .A sth eforme rtw oar eabsen ti nth enon-openin g zoneo fth edeviatin ganthers ,thei rinfluenc eseem st ob erathe rclear .
62 Chapter 5
FUNCTIONAL MORPHOLOGICAL RELATIONSHIPS INTH EDEVELOPIN G LOCULEO FGASTERI A VERRUCOSA
C.J.Keijze ran dM.T.M .Willems e
Summary
The development of the locule tissues of Gasteria verrucosa (Mill.) H.Duvali sinvestigate dusin gelectro nmicroscop y andvisualize d indetaile d pen-drawings. Mosto f theanthe r openingprocesse stak eplac ei nth eepidermi san dth e endothecium.Th estarc hi nthes etw olayer si suse dfo rdifferen tstep so f the dehiscence process.Th e (single)middl e layer is isolated from the endothecium and finally degenerates, which may be related to tapetal functions. Thespherica lmeioti ccallos ewall swit hthei rcytomicti cchannel senabl e thepolle n mother cellst odivid e intofou r equally shapedmicrospores . Sporopollenin containing structures arise on both the tapetum and the microspores, forpolle n dispersal andprotectio n respectively. At the same time thetapetu m regulatesth edisappearanc eo fth eorigina lcel lwall san d thecallos ewall sfro mth emicrospore san ditself .Thes e wall materials partlytur nint ostarc hi nbot htissues .I nth eforme ri ti sstore di nth e vegetative cell,i nth elatte ri tturn sint opar to fth epollenkitt .Th e polarity in the microspores isrelate dt oth e formation and developmento f thegenerativ ecell .Ultrastructura lchange si nth edifferen tcel llayer sar e discussed inrelatio nt othei rdevelopmen tan dinteraction .
Introduction
During the last decennia many reports appeared on tapetum and pollen development and their relationship (see the reviews of Linskens, 1964,1967; Puri,1972; Vasil,1973; Mascarenhas,1975; Shivanna et al.,1979; Bhandari,1984;
63 Knox,1984; Shivanna and Johri.,1985 and the many works of Heslop-Harrison). Few investigations were carried out on the development of the other locule tissues and these were mainly focussed on anther dehiscence (many old works and ch.2) or deal with the systematics of anther tissues or endothecium patterns (Richter,1929; Eames,1961; Davis,1966; Stanley and Kirby,1973). Although the different locule tissues are sometimes described in reports on male sterility, where they often show deviations (see the reviews of Edwardson,1970; Laser and Lersten,1972; Gottschalk and Kaul,1974), investigations on the fertile developing locule as a functional unit are scarce (Reznickova and Willemse, 1980). In this chapter the ultrastructural development of the locule of Gasteria verrucosa (Mill.) H.Duval is investigated. This work was preceded by some studies on different aspects of the anther development in the same species (Willemse,1972; Van Lammeren et al.,1985; Schroder, 1985; ch.1, 2 and 3).
Materials and methods
For electron microscopy intact anthers of Gasteria verrucosa (Mill.) H.Duval in different stages of development were fixed in 3% glutar aldehyde for 2 hours and 1% osmium tetroxide for 45 minutes, both in 0.07M cacodylate buffer (pH 7.2) at room temperature. After dehydration in ethanol they were embedded in the low viscosity resin of Spurr (1969). Ultrathin sections were stained with lead citrate and uranyl acetate (Reynolds,1963 ) and observed in a Philips EM 301 at 60 kV. From each locule tissue twenty cells were investigated and the amount, size, shape and position of nuclei, vacuoles, vesicles, plastids, mitochondria, dictyosomes, ER, ribosomes, membranes, cell walls and in a single clear case microtubules were determined in the transverse plane of the nucleus. Locule segments were reconstructed by means of pen-drawings on the scale 2200:1, reflecting the mean values of these observations.
Results
1.Th epolle nmothe rcel lstage .
Fig.1show sth epolle nmothe rcel lstage ,i.e .afte rth emitoti cdivision s
64 inth e sporogenous tissuehav e finished. Inthi sstag eth elocul eo fG . verrucosaconsist so f5 tissu e layers:eac hth e epidermis,endothecium , middle layeran dtapetu m consisto fon elayer ,th epolle nmothe r cellsfil l thecentr eo fth eanther .Th eepidermis ,endothecium ,an dmiddl elaye rar e ultrastructurally similar inman y aspects.Thei rcell sar evacuolate dwit ha centralnucleus .Th eplastid scontai n thylakoids, staining theanthe rgreen . Mosto fthe mbea rstarch ,excep ti nth emiddl e layerwher eth eplastid sar e smaller. The number of mitochondria is almost equal to the number of plastids. Bothth eamount so fRE Ran dribosome sar erathe r low,dictyosome s and lipiddroplet sar escarce . Alsoth etapetu man dth epolle nmothe rcell sstrongl yresembl ean dar e only distinguishable by cytoplasmic details. Both cell types are isodiametric, lackingan ypolarit y andhav e aprominen t centralnucleus .I n thepolle nmothe rcell sth enucleole sar emuc h larger.Th emitochondri ahav e thesam esiz ea sthos eo fth eoute rlayers ,th eplastid sar emuc hsmalle ran d containsingle-membran einclusion san dribosomes ,lackin gan ystarch .Th e ribosome density is larger than in the outer layers.Th e tapetum cells containmor eRE Ran ddictyosome s thanth epolle nmothe rcells ,bu tlac kan y lipiddroplets .Th eamoun to flipi ddroplet si nth epolle nmothe rcell si s comparablewit hth eoute rlayers .Bot hth enucleole san dth evacuole so fth e pollenmothe rcell sar elarge rtha nthos eo fth etapetu mcells . Plasmodesmataar epresen ti nal lth ecel lwall si nth elocul eexcep ti n those between the bordering pollen mother cells.Th epart so fth ewal l borderingth eplasmodesmat aar ethicke rtha nth eres to fth ewall ,excep t betweenth eepidermi san dth eendothecium .
2.Th ezygoten estage .
Inth ezygoten estag e (fig.2)th eepidermi san dendotheciu m cellshav e growni nbot htangentia lan dradia ldirection ,a shav ethei rvacuoles .Th e amountan dcompositio no fthei rcytoplas mi sunchanged ,onl yth esiz eo fth e starchgrain sha sincreased .Als oth emiddl e layercell shav ebee nstretche d intangentia l direction, but these cells have been flattened in radial directionb ywhic hthei rcapacit yha sno t increased.Th etapetu m cellshav e strongly grown and became binucleate after endomitosis. The amount ofRER , dictyosomes, lipiddroplets ,mitochondri aan dplastid s has increased asha s theshap ean dth eamoun to fmembranou s inclusions insideth elatter .Th e plasma membrane undulates slightly. Bothth epolle n mother cells andthei r
65 nuclei have grown strongly, inside the latter synaptonemal complexes are visible. Their original walls have become more electron dense and the site of the middle lamella has widened, leaving intercellular spaces at the corners. Between the plasma membrane and the original wall callose deposition has started, giving the cell a spherical shape. The plasmodesmata towards the tapetum have disappeared, the meiocytes are mutually interconnected by the much wider, newly formed cytomictic channels in which organelles may be present. The amounts of cytoplasm and organelles have increased. A few multi membrane-bound inclusions containing ribosomes are visible in the cytoplasm.
3. The early tetrad.
During meiosis no remarkable changes occur in the epidermis, the endothecium and the middle layer cells. The circumference of the locule increases slightly by incidental anticlinal divisions in the epidermis and the endothecium, by which the middle layer cells continue their tangential stretching together with a radial flattening. The tapetum cells grow slightly in radial direction. Their amounts of dictyosomes, RER and ribosomes increase sharply, the plastids are filled with starch. The vacuolation increases. Electron dense droplets appear between the cell wall and the stronger undulating plasma membrane. The meiocytes have become spherical by the deposition of callose, which has also closed the cytomictic channels. They are divided into four equally sized and shaped microspores by the formation of additional thin callose layers, which are bordered by dictyosomes. The haploid nuclei are situated near the periphery of the original diploid cell. The plastids contain a few single-membrane structures and small globules. In the cytoplasm a few nucleoids have appeared and the multimembrane-bound inclusions are still present.
4. The late tetrad.
The ultrastructure of the epidermis, the endothecium and the middle layer remained unchanged, the circumference of the locule has increased slightly and the middle layer continued its flattening. The tapetum cells have stretched in radial direction accompanied by an increase of both cytoplasm and vacuoles. The amounts of RER and ribosomes have increased. Along the centripetal part of the cells pro-orbicules appear between the wall and the
66 plasma membrane. The middle lamella around both the tapetum and the meiocytes widens and dissolves, between both tissues electron dense material appears. The nuclei of the microspores have moved to the centre of the cell complex. The callose wall formation around the cells has finished and a primexine is deposited. The latter is interrupted by the baculae of the exine, which is accompanied by the fusion of dictyosome vesicles with the plasma membrane. Near the future colpi the exine is very thin and here RER lines the plasma membrane. The multimembrane-bound inclusions have disappeared, the cytoplasmic nucleoids are still present. The plastids contain more membranes than in the previous stage.
5. The young free microspore.
The ultrastructure of the epidermis, the endothecium and the middle layer remained unchanged, the circumference of the locule increased slightly and the middle layer continued its radial flattening. The tapetum cells have reached their maximal size, accompanied by a continued vacuolation. In their plastids grey material has appeared and the starch grains have enlarged. Dark droplets are visible between the mitochondrial membranes. The amount of lipid droplets has increased strongly. The pro-orbicules are covered with fusing drops of sporopollenin and are interconnected by an electron dense, granular tapetal membrane outside the plasma membrane. This surrounds the whole cell and is associated with the extracellular dark droplets which appeared in stage 3. The original cell wall has disappeared. The microspores have lost both their original (pollen mother cell) walls and the callose walls, while the exine thickens. They are floating in locular fluid. Their plastids grow strongly and store many small starch grains. Stacks of RER lay against the nucleus. The cytoplasmic nucleoids have disappeared, near the colpus a few large vesicles and the plasma membrane are confluent.
6. The vacuolated microspore stage.
Both the slight circumferencial growth and the radial flattening of the middle layer continue. The amount of starch in the plastids of the epidermis and the endothecium has decreased, while the number of plastoglobules has
67 increased,stainin gth eanthe ryellowish-green .Hole sappea ri nth ewal l betweenth eendotheciu man dth emiddl elayer ,wherea sth eplasmodesmat ahav e disappeared here. The vacuolation of the tapetum cells has strongly decreased.Th estarc h disappeared from mosto fthei rplastids ,th egre y material gradually turns into droplets which fuse to larger units.Th e plastids swell strongly, eventually followed by the disappearance ofthei r membranes.Thes echange soccu rasynchronousl y inth edifferen tplastid san da developmentalrang eca nb eobserve dwithi neac hsingl ecell .I nth ecytoplas m long strands of electron dense materialhav e appeared, bordered by swollen RERcisterns .Th eamoun to felectro ndens edroplet sbetwee nth emitochondria l membranes has increased, the dictyosomeshav e disappeared. The nucleihav e becomesmaller ,th eheterochromati nmor eelectro ndense ,contrar yt oth e euchromatin.Th eplasm a membrane has disappeared by which the cytoplasm is bordered by the tapetal membranes. Further sporopollenin deposition onth e orbicules cannotb eobserved . Themicrospore shav eenlarged ,mainl yb ya growt ho fthei rvacuoles ,whic h fusewith ,an dsometime scontain ,dictyosom evesicles .Th eexin eha sbee n thickened and the intine is deposited in association with dictyosome vesicles,whic h fusewit h theplasm a membrane.Th ethickene d intine ofth e colpusi sinterrupte db yradiall ydirecte dplasm amembran eextensions .Durin g theirformatio nthes ear ebordere d byRE R cisterns,finall y they become isolated. Both the amounts of RER stacks around thenucleu s and starch containingplastid shav e increased. Most ofth e cytoplasm is found between the colpus and thenucleus ,whic h isstil lborderin gth eopposit ewall .Th e vacuolesfil lth eothe rzone so fth ecell .
7.Th egenerativ ecel lagains tth eintine .
Thegrowt ho fth elocula r circumferencean dth eflattenin g ofth emiddl e layer have continued as did the decrease of starch and the increase of plastoglobules in the epidermal and endothecial plastids. In the epidermis cellsth enumbe r of dictyosomesha sincreased .I nth eendotheciu m cellsth e nucleusan dmos to fth eplastid smove dtoward sth einne rtangentia lwall .I n the middle layerth enuclea rchromati nbecome smor eelectro ndense ,th e plasmamembran eundulate so nal lside so fth ecell . The tapetum cells schrank by which the concentration of ribosomes increased. Againth enucle ihav ebecom esmaller .Th eelectro ndens estrand s fusewit hth egre ydroplet st olarg edrop so fpollenkitt .Th eamoun to fbot h
68 mitochondria and RER decreased. A mitotic division and the deposition of a separation wall have divided the pollen grain into a vegetative and a much smaller generative cell. The latter is mainly filled with two terminal vacuoles and its nucleus with electron dense chromatin. The small amount of cytoplasm contains mitochondria, dictyosomes, single stranded RER, lipid droplets and ribosomes. The content of the vegetative cell is comparable with the previous stage, with the exception of the organelles that were deposited in the generative cell. The amount of stacked RER has diminished. It has partially been transported towards the connection between the separation wall and the intine.
8. The generative cell is set free from the intine.
The epidermis and endothecium cells have expanded, mainly in radial direction, by means of enlarging vacuoles. In the plastids of both layers the amounts of thylakoids and globules have increased. The cuticular surface is enlarged due to the formation of electron opaque wall ridges. In the endothecium cells U-shaped wall thickenings are deposited, covered with RER. The holes in the wall between the endothecium and the middle layer have enlarged. The plasma membrane undulates along both this wall and the wall thickenings, bordered by dictyosome vesicles. In the middle layer the plastids, mitochondria and vacuoles have disappeared, the nucleus has degenerated and the amounts of RER and lipid droplets have increased. In the tapetum cells the turnover from plastids and electron dense strands into pollenkitt has continued. The mitochondria have disappeared and the nucleus has degenerated. In the remaining cytoplasm the concentrations of RER and ribosomes are rather high, their total amount decreased. The pollen grains swell strongly, the amount of cytoplasm increased while the vacuoles disappeared. The colpus stretched by which some of the plasma membrane vesicles have been opened towards the locular cavity. The generative cell is tied off from the intine as the contact sites of the separation wall and the intine approached each other. The separation wall itself has disappeared and the two cells are only separated by their plasma membranes, without any pores. The cytoplasmic content of the generative cell is unchanged, the cell is surrounded by a large amount of lipid droplets in the vegetative cell. In the latter the concentration and amount of all types of organelles has strongly increased. However, the plastids are smaller than in
69 thepreviou s stage and containa fe w smallstarc hgrain san delectro ndens e material.Al lth eRE Ri ssingl estrande dwit hswolle nends .Th evegetativ e nucleus ishighl y euchromatic.Th e colpusi sstretche d andha sa nirregula r surface.
9.Th ematur epollengrain .
Theepidermi san dendotheciu mcell shav eexpanded .I nth eplastid so fbot h layers the amount of starch and thylakoids decreased, contrary to the plastoglobules. The cuticular ridges have become higher. Between the epidermis andth eendotheciu m theplasmodesmat ahav e disappeared. The wall thickenings inth eendotheciu m cellshav ebee nthickened ,th eamoun to f bordering RER has decreased. The inner tangential wall has largely disappeared,wit hth eexceptio no fth ewal lthickenings .Bot hth emiddl e layerwall san dcel lconten ta swel la sth etapeta lconten thav edisappeared , the tapetal membranes borderth eendotheciu m cells.Th epollenkit tan dsom e cytoplasmicremnant so fbot hth etapetu man dth emiddl e layer (tryphine)ca n befoun dbetwee nth epolle ngrain san dinsid eth eexin ean dcolpa lcavities . Moreover iti spresen t inman y of the intercellular spaces between the epidermisan dth eendotheciu mcells . Thepolle n grainshav e reachedthei rlarges tsize .Th espindl eshape d generativecel lfloat sfreel yi nth evegetativ ecell ,surrounde db ystack so f RERan dth elobe dvegetativ enucleus .Th egenerativ enucleu si sspindle - shaped,mos to fth ecytoplas mi sfoun di nth etermina lpart so fth ecell .I n the vegetative cellth eamount so f the different organelleshav e increased. Themitochondri aar efille dwit hswolle ncristae .Th esmal lplastid scontai n electron dense materialan d are lined with semi-rough ER,th e smoothside s facingth eplastids .
Discussion
1.Th edevelopmen to fth eepidermi san dth eendothecium .
Someo fth eprocesse sdurin gth eepidermi san dendotheciu mdevelopmen tca n berelate dt oeac hother . Theincreas eo fth econsiderabl eamoun to fstarch ,alread ypresen tbefor e thepolle nmothe rcel lstage ,ma yb edu et ophotosynthesi si nth ethyllakoid -
70 rich plastids themselves. The gradual disappearance of this starch from stage 6 precedes structural changes in both layers. The increase of the osmotic value, which causes the vacuolation of the cells from stage 8 (Woycicki,1924; ch.2) is presumably due to the breakdown of starch into smaller molecules (Woycicki,1924). Secondly, breakdown products of the starch may be used for the deposition of the wall thickenings in the endothecium (Panchanksharappa and Syamasendar,1974), which is also indicated by the migration of most of the plastids to the inner tangential wall, where most of the thickenings can be found. Thirdly, components of the starch may be used for the formation of the carotenoid containing plastoglobules, which stain the anther yellow. Hannig (1910) demonstrated that even weak light could raise the temperature in pigmented anthers, increasing the evaporation, necessary for dehiscence (ch.1 and 2). Besides, Reznickova (1983) shows a carbohydrate pathway from the epidermis and endothecium to the tapetum and the pollen grains. The role of the suddenly appearing epidermal wall ridges, reported before by Cheng et al. (1981) in Oryza sativa, in chapter 2 in Gasteria verrucosa and in chapter 4 in Aloe spp., is unclear. The accompanying enlargement of the cuticular surface can explain the improving ability to evaporate, demonstrated in chapter 1, necessary for dehiscence (ch.2), but the continuously slight thickening of the cuticle contradicts with this theory. Moreover, similar structures can be found in the epidermis of non-desiccating floral parts, like filaments and pistils (Willemse and Franssen- Verheijen, 1986) which may indicate a function in the rigidity of these organs.
2. The development of the middle layer.
The cytoplasm of the middle layer hardly changes during most of the development, shortly before its degeneration the amounts of RER, ribosomes and lipids increase. Apart from the continuous tangential stretching of the cells, due to both the enlarging circumference of the locule and the continued absence of anticlinal divisions (Davis,1966), the final isolation from the endothecium is remarkable. The disappearing plasmodesmata between these two tissues in stage 5, followed by the dissociation of the middle layer from the endothecium, indicate a deviating development, possibly related to the tapetal development from this stage. Although the middle layer degenerates from a later stage and quicker than the tapetum, the preceding increase and relatively long persistence of RER, ribosomes and lipids are
71 comparable with the tapetal degeneration. During the transfer of the pollenkittt oth elocul eth ecytoplasmi cremnant so fth emiddl elaye rar e mixedwit hthos eo fth etapetu man ds obecom epar to fth etryphin e(ch.3) . Boththes edat asugges ttha tth emiddl e layercontribute st oth etapeta l production.Th e latterhypothesi s issupporte d byth eundulatin g plasma membranei nstag e7 ,indicatin gexcretio nactivity ,mayb ean ysporophyti c substancet ob emixe dwit hth edevelopin gpollenkitt .I tma yals oinclud e lyticenzyme sfo rth edegradatio no fth ecel lwalls .Thi sdigestio nma yals o be caused and continued by the endothecium, which isvisualise d by its undulatingplasm amembrane . A contribution of themiddl e layer inan ytapeta lproductio nagree swit h thefinding so fReznickov aan dWillems e (1980)fo rth einne rmiddl e layero f Liliumhybrida .However ,contrar yt othei rfindings ,similaritie sbetwee nth e plastidial contentso fth emiddl e layeran dth e tapetum areno t found inG . verrucosa.
3.Th erole so fth etapetum .
Thetapeta lfunctions ,recentl yreviewe db yPacin ie tal .(1985) ,ca nb e dividedint othre emai ngroups ,dependin g onthei r final goal. Firstly, the regulationo fcel lwal lchange san dcoating s(th esuppl yo flyti cenzymes , sporopollenin precursors,pollenkit t and tryphine). Secondly, the supplyo f energy rich substances to the developing pollen grains (PAS-positive materials).Thirdly ,th epreparatio no fth epolle ngrain so nrecognitio nan d germination (thetransfe ro f sporophyticpolle nwal lenzymes) .Som eo fthes e aspectsca nb erelate dt oou robservation san dwil lb ediscusse dhere .
a.Th eorganizatio no fth ecel lwal lchanges .
Our previous histochemical data (ch.3) and present ultrastructural observationsindicat ea norganizin grol eo fth etapetu mi nth echange so fth e cellwall s ofbot h thetapetu m itself andth e developingpolle ngrains . Shortly afterth e callose depositionha sstarted ,th emiddl e lamellaaroun d bothth etapetu man dth epolle nmothe rcell sdesintegrates .Th emor eelectro n denselayer so fth ecel lwall si nbot htissue sremai nan ddisappea rlate ron , synchronously with the callose of the meiocytes.Thes e ultrastructural observationsagre ewit hth ehistochemicall y testeddisappearanc eo fpectins , cellulosean dcallos e respectively (ch.3).Th esynchronitie s inbot hth e
72 tapetumcell san dth emeiocyte sconcernin gth epecti nan dnex tth ecellulos e breakdown,indicat ea simila rorigi nfo rth enecessar y enzymes.Th epresenc e ofth ecallos ewalls ,whic h mayb eimpermeabl e forlarg emolecule s(Heslop - Harrisonan dMacKenzie,1967) ,durin gth epecti nan dcellulos edigestio nmake s anenzym esuppl yfro mth emeiocyte sunlikely .Therefore ,th etapetu mwil lb e thesourc efo rth epectinase san dcellulase sfo rbot htissues .Mepha man d Lane (1969b)demonstrate dtha tals oth ecallase ,necessar yt odiges tth e callosewalls ,originate sfro mth etapetum .Thi smean stha tth etapetu m directsth edifferen t stepst oreleas eth emicrospores .Thi sma yb erelate d toth epresenc eo fdictyosom evesicle soutsid eit sundulatin gplasm amembran e inth emeioti c stages (Echlin,1971). Beside thesewal l digestions thetapetu m directsa par to fth esynthese s of new walls.Th e sporopollenin containing tapetal membranes andorbicule s andpar to fth esporopolleni nprecursor sfo rth eexine sar etapeta lexcretes , whereasth eearl yexin ean dintin eoriginat e from themicrospor e itself (Heslop-Harrison andDickinson ,1969 ;Heslop-Harrison,1971) .Thi smean stha t the tapetum regulates the shift from hydrophylic (pectins, cellulose, callose)t ohydrophobi c (sporopollenin)wal lsubstance so nth eborde rbetwee n thesporophyt ean dth egametophyte ,whic h isimportan t forpolle n dispersal (ch.3). b.Th etapeta ldegeneration .
During the young microspore stage the tapetum cells reach their largest size.Th e enlargingvacuoles ,presumabl y filled withPAS-positiv e materials (Pacinian dFranchi,1983) ,th eincreas eo fcytoplasmi clipid san dER ,a swel l as the plastidial increase of starch and grey materials, presumably carotenoids (Reznickova andWillemse,1980) ,reflec tth eaccumulatio no f differentsubstances .Afte rthi sstag eth ecell sstar tt odegenerate .Th e starchan dlipi dgraduall ydisappea ran dar epartl yinvolve di nth esynthesi s ofth e largeamoun t ofpollenkit t (ch.3). Mitochondria and dictyosomeshav e disappearedafte rth epolle nmitosis ,th enucleu sdegenerate sslower ,wherea s mosto fth eRE Ran dribosome sremai nunti lth elas tdegeneratio n stagesan d arepartl y depositedo nth epolle ngrain sa stryphine .I nman yspecie sther e isa clos e correlation between this degeneration and the appearance of reserve substances inth e developingpolle ngrain s (Echlin,1972;Christense n andHorner ,1974) .I nG^ _verrucos ath emicrospore shav estore da larg eamoun t ofstarc hbefor eth etapeta ldegeneratio nbecome svisible ,rathe rdu et oth e
73 callose breakdown than totapeta l supply (seebelow) .Thi samoun t increases slightly up toth e mitotic division.However , afterthi s division, halfway thetapeta ldegeneration ,th eamoun to fcytoplas mi nth evegetativ ecel l increases sharply,whil e the locular fluid gradually disappears (ch.1).Thi s would mean that tapetalbreakdow nproduct s arestore d temporary inth e locular fluid, according to Pacini and Franchi (1983), and are only transferred toth epolle n grainsafte r the cytoplasm ofth egenerativ ecel l hasbee nisolate dfro mth evegetativ ecell .Th epersistin gtapeta lRE Ran d ribosomes indicatea continuin gsynthesi so fproteins ,presumabl y lytic enzymesfo rth ecel litsel fan dsporophyti cpolle nwal lprotein s(Heslop - Harrisone t al.,1973).However , theincorporatio no fenzyme sint oth epolle n coatingmus tno tb eexcluded ,a sRE Rofte nborder sth edevelopin gglobule so f pollenkitt (ch.3)an dribosome s canfinall yb etrace di nth etryphin eo nth e pollen grains.Th e transfer of the final tapetal content, i.e. mainly pollenkitt and a small amount of tryphine, to both the locule and the intercellularspace sbetwee nth eepidermi san dth eendotheciu mi sa resul to f capillary forces inth ecrowde d locule (ch.3an d4) . Besidespossibl ecorrelation sbetwee ndisappearin gtapeta lsubstance san d appearingpolle nsubstances ,th eapparentl ynorma lcontinuatio no fth epolle n developmento nrathe rsimpl ecultur emedi ai nvitr o(Tanak aan dIto,1980 , 1981;Tanak ae tal.,1980 )indicate sa considerabl eautonom yfo rman yo fthes e processes.
4.Th edevelopin gpollen . a.Th ecytoplas mdurin gmeiosis .
Duringmeiosi sdedifferentiatio no fribosomes ,plastid san dmitochondria , asreporte db yDickinso nan dHeslop-Harriso n (1977)i nth erelate dLiliu m hybrida,i sno tobserved .Membrane-boun dcytoplasmi cinclusion sdevelo pt oa smallextent .Thi smean stha ti nG .verrucos acytoplasmi cchange sfro mth e diplophaset oth ehaplophas ear eobservable ,bu tno ts oclea ra si nLiliu m (Dickinsonan dAndrews,1977) .O nth econtrary ,th epresenc eo fnucleoids , relatedt ocytoplasmi c redifferentiaticn (Dickinsonan dHeslop-Harrison,1970 ) isevident .
74 b.Possibl efunction so fth emeioti ccallos ewall .
Themeioti c callose wall, alsohistochemicall y demonstrated in this species (ch.3), presumably originates from materials in the dictyosome vesiclesi nth eproximit yo fth eplasm amembran e(Echli nan dGodwin,1968 ; Willemse,1971a,b). About the role of these walls and their cytomictic channels different reportshav e appeared. Waterkeynan dBienfai t (1970) demonstrated anegativ ereplic ao fth eexin epatter no nth einsid eo fth e callosewal lo fIpomoe apurpure a(L )Roth ,an dsuppose da possibl erol ei n thedeterminatio no fthi spattern .I nG .verrucos asuc ha patter nwa sals o found(Willemse,persona lcommunication) .Heslop-Harriso nan dMacKenzi e(1967 ) andSouthwort h (1971)demonstrate d thatcertai nmolecule scanno tpenetrat e thecallos ewhil eother scan ,whic h suggeststha tth ewal l may acta sa selective barrier during meiosis. Any selective role during development is likely after the observation thatprematur e callose digestion is often accompanied bymal esterilit y (Izharan dFrankel,1971) . Weassum ea thir dfunctio nfo rthi swall .I nG ^verrucosa ,a si nman y species, mosto fth ecallos ei sdeposite d inth ecorner so fth eorigina l cellwall san dturn sth edifferentl yshape dmeiocyte sint ospherical ,equall y shapedcell s (VanLammere ne tal.,1985) .Thi senable sthe mt odivid eint o fourequall yshape dan dsize dmicrospores ,whic h excludesan y selective influenceo fsiz ean dshap edurin gpolle ndispersal .Thi sequa lshapin gi s alsoeffecte db yth esymmetri c arrangement ofth ecolp ii nth e callosic sphere.Furthermore ,th eamoun to fcallos enecessar yt oreac ha spherica l insidedepend so nth esiz eo feac hindividua lpolle nmothe rcel lan ddiffer s betweenth ecells .Thi scause sdifferen tvolumina lchange so fth ecell s duringth ecallos e deposition.Fo rthis ,th ecytomicti c channels mayb e necessaryt oenabl ea cytoplasmi credistributio nbetwee nth ecells ,b ywhic h alsoorganelle sar etransported . Heslop-Harrison (1968)suggeste da differen t function forthes e channels, beingth e creationo f a syncytium which might synchronizeth emeioti cdivisions .Thi sma yb etrue ,althoug hsynchronou s processes alsooccu r inspecie swithou tcytomicti c channels,lik ePinu s sylvestrisL .(Willemse,1971a,b) . Therear eindication s forth erecyclin go fth econsiderabl e amounto f callosebreakdow nproduct sdurin gth efurthe r development.Larso nan dLewi s (1962)reporte dth esuppl yo fcompound sfro mth ecallos ewal li nfavou ro f thecellulosi cprimexine .Besides ,w efin di nG^ _verrucos aa sudde nincreas e ofth eamoun to fstarc hi nbot hth etapetu man dth emicrospore sshortl yafte r
75 thecallos edigestion .W epresum etha tthes etw otissue stak eu pcallos e derivedsugar sfro m thelocula rflui dfo rthei r starchproduction , by which the callose canb eregarde d asa nearl y reserve substanceo fth e future pollengrains .Als oth epecto-cellulosi cwall so fbot hth etapetu man dth e pollenmothe rcell sma ypla y sucha rol e(Pacin ian dFranchi,1983) .
c.Th emeanin go fpolarity .
Polarity appears immediately upon the formation of the microspores. The first signs are the centrifugal position of both the nucleus and ER along the plasma membrane of the young tetrad, the ER presumably preventing the deposition of a fully developed exine over the future colpus (Willemse,1972). During the late tetrad stage the nucleus migrates to the wall opposite the colpus, due to the activity of microtubules (Van Lammeren et al.,1985). After the callose digestion most of the cytoplasm has moved to the area between the nucleus and the colpus, surrounded by the rapidly enlarging vacuoles. In this way the absence of plastids in the generative cell, reported before in this species by Schroder (1985), is an important result of the polarity. Furthermore, the polarity may play a role in the deviating development of the generative cell. The presence of callose in the wall between the two cells (ch.3) is a first indication of this deviation (Owens and Westmuckett, 1983). Moreover, the confluence of vesicles (secundary lysosomes?) with the plasma membrane near the colpus in the young microspore indicates that the flow of locular fluid to the pollen grains occurs mainly through the colpus, suggested before by Rowley and Flynn (1971) and Christensen and Horner (1974). This may be facilitated by the radially directed plasma membrane extensions. Consequently, the generative cell is isolated from the locular fluid, which may effect its deviating development and explains why the callose is restricted to the division wall.
d.Th epolle nwalls .
Theexin edevelopmen t ofG .verrucos ai sthoroughl y describedb yWillems e (1972). Besides aprotectiv e function with storage sites for proteins (Knox,1984), its hydrophobic natureplay s a role inpolle n dispersal (ch.3). Theelasticit y ofth eexin ei smainl yrestricte dt oit sreduce dpar to nth e colpus,whic hallow sa considerabl eharmomegathy .I nth efina ldevelopmenta l stages (8an d 9)thi s thin sporopollenin layer is disrupted due to the
76 enormousstretchin go fth ecolpus ,b ywhic hth etie dof fplasm amembran e undulationso fth eintin ear eopene dtoward sth elocule .A sth elatte ri s rather dry justbefor edehiscenc e (ch.1),thi swil lpresumabl y notlea dt oa losso fth ehydrolyti cenzyme san drecognitio nsubstance swhic har estore di n thesesite s (Knox,1984) .Par to fth elatte rsubstance sma yoriginat efro mth e ER which is bordering these plasma membrane undulations during their formationi nth evacuolate dmicrospore . e.Th edevelopmen to fth egenerativ ecell .
Soon after its deposition, the generative cell istie d off from the intine,b ywhic hit scallos ewal l disappears (ch.3).Associatio n ofthi s process with microtubule activity (Dickinson,1975)wa s not found inG . verrucosa until now (Van Lammeren et al.,1985). On the contrary, the preferentialpresenc e of RERstack s nearth e edge ofth egenerativ e celli n stage7 indicate san yrol eo fthes estructure si nth e disjunction.Thei r preceding position around the nucleus has presumably to dowit h their formation.I nth ematur epolle ngrai ncomparabl estack sappea ralon gth e entiresurfac eo fth egenerativ e cell,whic hmake sthei rrol ei nth estorag e of substances for germination (Jensene tal.,1974 ; Cresti et al.,1985)i nG . verrucosa lesslikely . Thefina lspindl eshap eo fth efre egenerativ ecel li smaintaine db yman y microtubulesparalle lt oit slon gaxi s(Va nLammere ne tal.,1985) ,whic hi sa general feature (Sanger and Jackson,1971). The close association ofth e generativecel lwit hth evegetativ enucleus ,als oa genera lfeatur e(Wilm se t al.,1986),remain safte rpolle ndehydratio n(Keijze re tal.,1986 ;Wilm se t al.,1986)an dma y servea s atranspor tconfiguration . f.Th estructure si nth ematur epolle ngrain .
Althoughstarc hi sstil lpresent ,th emai nreserv esubstanc ei nth ematur e pollen grain ofG; _verrucos a islipid , which agreeswit hth eclassificatio n ofBake r andBake r (1979)fo rLiliaceae .Afte r the mitotic division the plastidsgraduall y losemos to fthei rstarc han dtur nint oa nincreasin g number of elaioplasts, most of them containing a small starch grain. Quantitativedat ao fth echangin gplasti dpopulatio no fth edevelopin gpolle n graini nthi sspecie sar ereporte db yWillems e (1972).A t the same timeth e amounto fcytoplasmi clipi ddroplet sincreases ,initiall yonl yaroun dth e
77 dissociating generative cell,bu t latero ndisperse dthroughou tth eentir e vegetativecell .Thi smigh tindicat etha tth elipid shav ethre epossibl e sources:firstly ,th ecallos efro mth egenerativ ecel lwall ,secondly ,th e starch and thirdly, soluble precursors from the degenerating tapetum, transportedvi ath elocula rfluid .Apar tfro mthi slipi dreserves ,th eE R whichi slinin gth eplastid sa tmaturit y indicatesth efutur ereleas eo f plastidialreserve s(partl ystarch )durin gpolle ntub egrowt h(Crest ian d Keijzer,1985).
Conclusions
Theultrastructura l changes inth eepidermi s andth eendotheciu m are stronglycomparabl ean dca nb eassociate dwit hth edehiscenc eprocess .Role s ofthes etissue si nth epolle ndevelopmen tca nprobabl ybette rb etrace dwit h labellingexperiment s(Reznickova,1983) . Although the development ofth emiddl e layerindicate ssom e similarities toth etapetum ,it scytologica lchange sar efew ,whil e its degeneration startsto olat et ogiv ea nimportan tsuppl yo fbreakdow nproduct st oth e developingpollen .A nexceptio nma yb eth epresenc e of cytoplasmic remnants inth etryphine . Partly accordingt oth emai ntapeta lfunction s reviewed byPacin ie tal . (1985), the cell wall changes in both the tapetum and the developing pollengrain smak eth etapetu mth emai norganize ro fbot hth eseparatio nan d dispersalo fth emal egametophyt efro mth esporophyte .Th e sequenceo fcel l wall digestions,th etapeta l degeneration,th elocula rflui ddisappearanc e andth epolle ndevelopmen tsugges ta shif to fcarbohydrate sfro mth eorigina l meiocytewall st oth etapetum ,followe db ya nopposit etransfe rfro mth e tapetumt oth edevelopin gpolle ngrains ,b ywhic hth elocula rflui dprobabl y servesa sa temporar ystorag esite .A fina lfunctio no fth etapetu mi sth e synthesiso fpollenkitt ,apar tfro m (sporophytic)protein sa swa sreviewe db y Knox(1984) . Animportan trol eo fth emeioti ccallos ewal li sth eformatio no fequall y sizedan dshape dpollengrains ,beside sa possibl erol ea sa selectiv ebarrie r ora nexin emold ,know nbefore .Th epolarit yi nth eyoun gmicrospore sresult s in the exclusion of plastids from the generative cell as well as the temporaryisolatio no fth elatte rfro mth elocule . From thisan dearlie r studies inthi s species different interactions
78 between the investigated anther tissues canb efound .Th e collaboration betweenth eepidermi san dth eendotheciu m seems tob e rather clear, mainly supportingth edehiscenc eproces s (ch.2).Interaction sbetwee nth emiddl e layeran dth eborderin gtissue sar edifficul tt odemonstrate ,contrar yt oth e results ofReznickov aan dWillems e (1980)i nLiliu m hybrida.I nou ropinio n interactionsbetwee nth etapetu man dth edevelopin gpolle n occur intw o directions. Apart from the often described transfer of substances fromth e tapetumt oth epolle ngrains ,th emeiocyte sinfluenc eth etapeta ldevelopmen t withth eoffe ro fderivative sfro mthei rorigina lwalls .I nlate rstage sa directingrol eo fth epolle ngrain si nth etapeta ldevelopmen tca nno tb e demonstrated and isprobabl y absent (ch.4).However , forth etransfe ro fth e pollenkitt towards the pollen grains the presence of the latter is indispensable (ch.3an d4) .
79 80 81 1
82 83 84 85 86 87 88 Chapter 6
THE FILAMENT DEVELOPMENT OF GASTERIA VERRUCOSA
C.J. Keijzer, I.H.S. Hoek and M.T.M. Willemse
Summary
The developing filament of Gasteria verrucosa is investigated using interferencecontrast ,transmissio nan dscannin g electron microscopy.Thi s developmenti srelate dt oth eanthe rdevelopmen tan dth eobservation sar e discussedi nrelatio nt opossibl efunctions . Thefilamen textend sdurin git sentir edevelopment ,i nth eyounge rstage s mainlyi nth ebasa lregion ,i nth eolde rstage si nth etip .U pt omaturit y thesolidit yi sgraduall yimprove db yth edepositio no fwal lmaterial si nth e intercellular spacesi nbot hth eepidermi s andth eparenchyma . Besides,th e outer tangential epidermal wall andth e cuticle areprogressivel y thickened fromth eti pt oth ebas eo fth efilament ,togethe rwit ha nextensio no fth e surfaceo fth elatte rb ymean so fth eformatio no fridge so nth ewall . Theamoun to ftrachear yelement sincrease su pt omaturit y asdoe sth e starch contento fth eepidermi san dth eparenchyma .Fro mth epolle nmitosi s stagea progressiv ecytoplasmi cdegeneratio nca nb eobserve dfro mth eti pt o thebas ean dfro m thecentra lparenchym a toth eepidermis .Afte ranthe r dehiscenceth efilamen tti pha sshrivele dan dth estarc hha sdisappeare dfro m theentir e filament.Bot hth elatte rphenomeno nan dth edegeneratio nma yb e duet oredistributio no fsubstance st oothe rflora lparts .Th epresenc eo f prominentintercellula r spacesma yb eimportan tfo rth esuppl yo fga st oth e maturinglocule .
Introduction
Contraryt oth ehug eamoun to fwor ko nth edevelopmen to fanthe rtissues , investigations concerningth edevelopmen to fth eborderin g filamentar e scarce.Schmi d (1976)wrot ea prominen twor ko nth esystematic so ffilamen t
89 anatomy, focussingo nth edifferen ttissue san dfoun dsom etypica lfilamen t characteristics.Besides ,h ediscusse dth epossibl erole so fth efilamen ti n antherdehiscence .Burc k(1906 )reporte dth eretractio no fwate rfro mth e anthersthroug hth efilamen ti nfavou ro fanthe r dehiscence,bu texperiment s of Hannig (1910)an d Schmidan dAlper t (1977)contradictin g withthes e results. A few works deal with the vascular tissue of filaments (Ivancich,1906;Leinfellner ,1956 )o rth eanatomica lbackgroun do ffilamen t movements (Pfeffer,1904;Knoll,1914) .I nsom ecereal sinterrelationship swer e foundbetwee n apoorl y developed vasculartissu ean dth eoccurranc eo fmal e sterility (Shivanna andJohri,1985) . However, these authors also reportth e normaldevelopmen to fthi stissu ei nmal esteril eplants .Anderso n(1980 ) reportsth egrowt ho fpolle ntube sfro mth enon-dehiscen tanther sthroug hth e filament,th e receptacle and thecarpel st oth eovule si nth ecleistogamou s flowerso fsom eMalpighiacea espp . Inthi s chapter we investigated thefilamen tdevelopmen t ofGasteri a verrucosa (Mill.)H.Duva l from thepremeioti cstag eu p toanthe rdehiscence . Differentaspect so fanthe rdevelopmen ti nthi sspecies ,whic hwer ereporte d inchapte r1 ,2 ,3 an d5 ,ar erelate dt oth efilamen tdevelopment .
Materialsan dmethod s
Stamens of Gasteria verrucosa (Mill.)H.Duva l in different stageso f developmentwer efixe di n3 %gluta raldehyd efo r3- 4hour san d 1%osmiu m tetroxide for2- 3 hours,bot h in0.1 M cacodylate buffer (pH7.2 )a troo m temperature.Afte rdehydratio n inethano lthe ywer e embedded inth elo w viscosityresi no fSpur r(1969) .Ultrathi nsection swer estaine dwit hlea d citratean durany lacetat ean dobserve di na Philip sE M30 1a t6 0kV .Fo rth e detectiono fcallos eth egrid swit hultrathi nsection swer eenclose di na mixtureo fanilin eblu e(Jensen,1962 )an da highl y dilutedaquou ssolutio no f calcofluor whiteM2 R (Hughesan dMacCully,1975 )unde ra cove rglass .Th e former dyestain sth ecallose ,th elatte rstain sbot hcallos ean dcellulose , butt oa les sexten ttha nth eformer ,thank st oth ever ylo wconcentration . Forinterferenc econtras tmicroscop y2 u msection swer ecu tfro mth esam e blocksan dobserve di nwate runde ra coverglass .Fo rth esam emicroscop e intact filamentswer e immersed inth e clearing solution of Herr (1971)an d directly observedunde ra coverglass . Toobserv eth einterna lstructur eusin gscannin gelectro nmicroscopy ,
90 filamentswer etreate daccordin gt oth eprocedur eo fCrest ie tal .(1986) , andobserve di na Jeo lSE M35 ca t1 5kv .Als ofres hfilament swer eobserve d inth eSE M forth ecuticula rstructure -
Observations
1.Genera ldevelopment .
Theflowe ro fG .verrucos aha stw owhorl so fthre estamens .Th estamen so f theoute rwhor lar eshorte rtha nthos eo fth einne rone ,i nwhic hth eape xo f theoute ranther sreache sth ebas eo fth einne ranther sdurin gth eentir e development. Thefilamen to f Gj_verrucos aconsist so fa nepidermis ,surroundin ga mai n tissueo fparenchym awit ha centra lvascula rbundle .Durin gth egrowt ho fth e flowerbu dth efilamen textend scontinuousl y inlongitudina ldirection , keeping the anther in the top of the bud. In the younger stages this extensionoccur smainl ynea rth ebas eo fth eanther ,late ro ni ti sgraduall y replacedt oth etip .I nth eolde rstage sth eapica lregio no fth efilamen t extends muchfaste rpe runi to ftim etha nth ebas ei nth eyounge rstages , whichi sals osee ni nth etepal san dth epistil .Th eanthe r itselveonl y stretches slightly.Th eti po fth efilamen ti sconnecte dt oth eanthe ra t abouta thir d lengthfro m theanthe r base.It svascula r bundle iscontinue d inapica ldirectio ni nth econnectiv etissu eo fth eanther .I tconsist so f onestrai no fspirall ythickene dtrachear y elementsan dtw ostrain so fsiev e elements (fig.1).Th ebundl e terminates atabou ta thir d lengthfro mth e antherapex .Jus tunde rthi sendin gth eamoun to ftrachear yelement si ncros s sectionincrease su pt otwic eth emea nnumbe ri nth eres to fth efilament , thes ocalle dtrachei dmaximu m (fig.2).
2.Th epolle nmothe rcel lstage .
Inth epolle nmothe rcel lstag ebot hth eepidermi san dth eparenchym aar e characterizedb yvacuolize dthin-walle dcell swit hprominen t intercellular spaces (fig.3).Nea rth eti po fth efilamen tsom eo fthes espace sar eabsen t underth eepidermi san dthickene dcel lwal lcorner si nthei r sites giveth e tissue a collenchyma-likeappearanc e (fig.4).Alon gth eentir elengt ho fth e filament the cuticle isthi n (fig.5). The epidermis and the peripheral
91 parenchyma layerscontai nsom estarch ,th eamoun tincreasin gtoward sth eto p region.Lipi di sscarc ei nal lth etissues ,excep ti nth eaccompanyin gcell s ofth esiev eelements .E Ri sscarce ,dictyosom evesicle sar epresen tnea rth e plasmamembran ei nbot hth eepidermi san dth eparenchyma ,especiall ynea rth e baseo fth efilamen t (fig.5).Th ecell sborderin gth evascula rtissu ear e richi nmitochondri aa scompare dwit hth eoute rlayers .I nthi sstag eth e xylem containsfou rt osi xtrachear yelement si ncros ssection .Eac ho fth e twophloe mstrain sha stw ot othre esiev eelements .
3.Th eyoun gmicrospor estage .
Inth eepidermi sth eamoun to fstarc hha sdecreased .O nth econtrary , starch has appeared in the central parenchyma layers (fig.6), still increasingprogressivel y towards thetip .Lipi d dropletsar eno w onlyfoun d inth eepidermi san dth ecentra lparenchyma .Unde rth eepidermi s thenumbe r of intercellular spaces thathav e beenfille d with wall materials,ha s increased.Thi sphenomeno ni sstil lmainl yfoun dnea rth etip .Th ecuticl e hasth esam estructur ea si nth epreviou sstage . Thewall so fth esiev eelement shav ebee nthickene db ywhic hthe ybecom e visibleafte rclearin gth eintac tfilamen twit hHer rsolutio n(fig.7) .Th e amounto ftrachear yelement sha sslightl yincreased .
1.Genera lcompositio no fth evascula rbundl eo fG .verrucosa .Th ehooke d helicalthickene dtrachear y elementsar esituate dopposit etw oseparate d strainso fsiev eelements ,whic har ehooke da swell .(12 1Ox ) 2.Afte rclearin gwit hHerr' ssolution ,th eendin go fth evascula rbundl ei s visiblei nth econnectiv etissu eo fth eanther .I nbasipeta ldirectio n parto fth etrachei dmaximu mi svisible .(6 1Ox ) 3.Th ebas eo fth efilamen ti nth epolle nmothe rcel lstage .Th ethi nwalle d epidermis andparenchym a cells surround prominent intercellularspaces . Inthi sexampl eth exyle mcontain sfou rtrachear yelements .(200x ) 4.I nth epolle nmothe rcel lstag eth ecel lwall so fth eepidermi so fth e baseo fth efilamen thav ethickene dcorners .(1825x) . 5.Detai l of fig.4. Dictyosome vesicles can be seen near the plasma membrane.(19300x ) 6.I nth eyoun gmicrospor estag estarc hdecrease di nth eepidermis ,wherea s itappeare di nth ecentra lparenchym a (arrows).(625x )
92 93 4.Th eyoun gpolle ngrai nafte rmitosis .
Inth eepidermi so fth eti po fth efilamen tth eamoun to fglobule si nth e amyloplastsha sincreased .Shortl yunde rth eto pth eepidermi scell shav e strongly expanded inal l directions by the growtho fth evacuole .Starc hi s stillpresen ti nth esam etissue sa si nth epreviou sstage .I nth elowe rhal f of the filament lipid has slightly increased inth ecentra lparenchym a and appeared in the peripheral parenchyma. Along the entire length of the filament signs of degeneration canb e observed insom e individual cellso f thecentra lparenchym a (fig.8),increasin gprogressivel yfro mth e baseu pt o the tip. In the central parenchyma of the filament tip most of the intercellular spaceshav e been replaced by cell wall materials, comparable withth ecorner so fth eepidermi scel lwall si nth eyounges tstages .Als o towardsth eti po fth e filament theoute rtangentia l wall ofth e epidermis hasbee nthickene dan dridge shav ebee nforme do nit soutside .Thi senlarge d thesurfac eo fth ecuticle ,whic h itselfha sbee nthickene da swel l(fxg.9) .
7.I nth eyoun gmicrospor estag eth esiev eelement shav ebecom evisibl e afterclearin gintac tfres hfilament swit hHerr' ssolution .I nth ecentr e tracheary elements (T)ca nb e seen,althoug h they are out offocus . (1325x) 8.Afte rth emicrospor emitosi sdegeneratin g cells (right)ca nb efoun d borderingobviousl yvita lones . Inth e former clusters of ribosomesan d plasmaar evisible ,bot hth eplasm amembran ean dth etonoplas tar estil l present.(18500x ) 9.Afte r the microspore mitosis both the cuticle andth eoute rtangentia l epidermalwal lhav ethickened .Th ecuticl ei slifte db yundulation so n the wall. (21600x) 10.Afte rth emicrospor emitosi snewl yformed'trachear yelement sca nb efoun d inth eti po fth efilament .I nthes eelement sth edictyosom eactivit yi s high.(6050x ) 11.I nth e matureanthe r stageth eamoun t and sizeo fth estarc h grains increasedi nth efilamen ttip .Thi sphotograp hwa smad e20 0u m underth e antherattachmen tsite .(970x ) 12.I nth ematur eanthe rstag elarg eamount so fSE Rappeare di nth eepidermi s of the upper 350 um of the filament. The undulation process of the cuticleha scontinue da scompare dwit h fig.9.(18500x )
94 95 In the tip of the filament tracheary elements are still being formed (fig.10), raising their total number up to 10 in cross section, thus extending the tracheid maximum downward from the anther into the filament. In the rest of the filament the amount of tracheids remained constant as did the number of sieve elements. In the accompanying cells a large amount of lipid droplets has appeared.
5. The mature anther stage.
In this stage the development of the main part of the filament differs from the upper 50 um. In the region from 50 to 500 um under the tip starch reappeared in the epidermis and increased strongly in the tissues where it was already present (fig.11), contrary to a decrease in the upper 50 um. From 500 um to the base of the filament the amount of starch remained about constant. In the upper 350 um the amount of RER has decreased, while large amounts of tubular SER have appeared in both the epidermis and the parenchyma (fig.12). In this stage the replacement of intercellular spaces by cell wall materials under the epidermis and in the central parenchyma has extended to about 350 um from the anther attachment site. In the lower parts the intercellular spaces persist in all the tissues. Most of the lipid, that appeared in the previous stage in the parenchyma of the lower half (=750 urn) of the filament, has disappeared. Signs of degeneration, in the previous stage only in some cells of the central parenchyma, can now also be found in the peripheral parenchyma of the entire filament, although still more frequently near the tip. In both the epidermis and the parenchyma the amount of dictyosomes has increased over the entire length of the filament, their vesicles are often close to the plasma membrane. The SER is still present in both tissues. The changes of the outer tangential epidermal wall and the cuticle, which started in the tip in the previous stage, have continued (fig.12), showing gradual differences from the tip to the base (fig.13). In the former zone the
13. The undulating cuticle with underlaying wall ridges appeared progressively from the base to the tip of the filament. Fig.13a shows the tip with a part of the anther, fig.13e shows the base. The f igs.13b-13d represent the intermediate stages. Besides, the longitudinal cell size increases from the tip to the base. (1425x)
96 97 cuticle sometimes dissociates from the wall (fig.14). Also the cuticle, which is present in some of the intercellular spaces, shows such changes (fig. 15). The amount of tracheary elements has still slightly increased, by which young tracheids can be seen between older stretched ones in this mature stage (fig. 16). The pores between bordering sieve elements and between sieve elements and accompanying cells have been filled with callose, which could be histochemically detected with aniline blue and calcofluor white staining on the same grid as was used for the TEM-photograph (figs.17 and 18).
6. Twelve hours after anther dehiscence.
In this stage the upper 50 um of the filament has dehydrated and shriveled as did the entire dehisced anther (fig.19). In the following 300 um some of the protoplasts detached from the walls. The cytoplasm appears disorganized in all the tissues (fig.20). The lower parts of the filament are turgescent, the cell shapes are comparable with the previous stage, although the cytoplasm of part of them shows signs of degeneration (fig.21). In the epidermis and parenchyma cells of the lower half (=750 um) of the filament dictyosome vesicles can be seen near both the plasma membrane and the tonoplast. Over the entire length of the filament the starch and most of the lipids have disappeared from these tissues. In the lower half plastoglobules appeared in the epidermis cells.
14. In the mature anther stage the undulating cuticle near the tip sometimes dissociates from the epidermal wall. (13500x) 15. In the filament tip the intercellular cuticles change like the epidermal cuticle, they form undulations and sometimes dissociate from the wall. (9200x) 16. In the mature anther stage both young and old stretched tracheary elements can be found in the vascular bundle of the filament. (1000x) 17. In the mature anther stage callose deposits are visible in the pores between two bordering sieve elements. (27000x) 18. The same grid as in fig.17 after staining with aniline blue and calcofluor white. The callose site (arrow), which is yellow due to aniline blue, can be distinguished as a white spot on this picture. The fainty (blue) calcofluor white staining of the rest of the cell walls serves as an orientation of the tissue in this ultrathin section. (3600x)
98 99 The cuticle has the same structure as in the previous stage, apart from extreme undulations on the epidermis of the dehydrated zone (fig.22). In the dehydrated tip the tracheary elements remained open (fig.23).
Discussion
1. The vascular tissue.
A vascular bundle with separated strains of sieve elements and relatively much phloem is rather common in filaments (Schmid,1976) . The presence of two bundles of sieve elements in both the filament and the connective tissue may have to do with the two-lobed structure of the anther. This is likely, since Ivancich (1906) reported the tangential splitting of both phloem and xylem in the filaments of some Amentaceae. Schmid (1976) related the emphasis on phloem to the considerable nutritional needs of the developing pollen. Indeed in G. verrucosa, as in most species, the pollen grains are gradually loaded with reserve substances. Also the development of the tapetal pollenkitt and the endothecial wall thickenings in this species (ch.5) justify such a hypothesis in favour of a considerable phloem development. Besides, the stamens develop inside the closed flower bud in which the relative humidity is very high and the evaporation low (ch.2). Magendans (1983) demonstrated that the amount of tracheary elements in vein endings of Hedera canariensis Willd. leaves, grown
19. Longitudinal cut through the critical point dried filament tip after anther dehiscence. A sharp border can be seen between the zones of collapsed and turgescent cells. (600x) 20. After anther dehiscence the cytoplasm in the tissues of the filament tip has degenerated. (305Ox) 21. The degenerating cytoplasm in part of the parenchyma cells halfway the filament after anther dehiscence. The plasma membrane lays against the cell wall. (18500x) 22. In the dehydrated apical zone the cuticular undulations approached each other by the shrinkage of the epidermal wall. (1650x) 23. In spite of the shrinkage of the dehydrated filament tip, the tracheary elements remain open. (1500x)
100 101 in a relative humidity of 97%, is half as much as in leaves grown in a relative humidity of 70%, thus decreasing the xylem to phloem ratio in humid conditions. This might be another reason for the emphasis on phloem in the stamen. Although experimental evidence for this hypothesis should be very interesting for anther development studies, such evidence is difficult to obtain as decreasing the relative humidity inside the flower bud appeared to be lethal for the stamen (ch.1). The presence of exclusively helical wall thickenings in the tracheary elements is often found in filaments (Schmid,1976) . These structures, which are characteristic for extending plant parts, remain open during the continuous extension (Grootaarts,1978) of the G. verrucosa filament. This means that the transporting capacity increases continuously. Important water sinks in the anther are the swelling epidermis and endothecium cells near maturity, which are resposible for the anther opening (ch.2), and the pollen grains themselves (Willemse,1972; ch.5). All these cells may receive water from the filament, although they also retract water from the locular cavity (ch.1 and 4), which must have been transported to the anther in an earlier stage. Besides, a continuous supply of water is necessary for the enlargement of the stamen, but this occurs rather slowly, given the total time of stamen development in Gj_ verrucosa (ch.3 and 5). The question remains as to why new xylem elements are deposited up to maturity. Maybe the new elements serve as a replacement for the older extending ones, which indeed remain open, but decrease their capacity as is visible in cross sections. The increase of tracheary elements near the ending of the vascular bundle (tracheid maximum) resembles the storage tracheids ("Speichertracheiden") near the tip of the vascular bundle of other plant organs (Heinricher, 1885; Haberland, 1924). Schmid's (1976) suggestion that these tracheids might store water to prevent premature anther dehiscence seems unlikely, given the large amount of water in the epidermis and endothecium cells (ch.2) in comparison with the low capacity of this tracheid complex. Prevention of premature dehiscence, if necessary, will presumably need a considerable supply of water from or through the filament, in spite of the extension of the tracheid maximum to the filament tip. Apart from the vascular tissue, a second transport system might exist. The persisting intercellular spaces in the peripheral parenchyma and the connective tissue, which are often found, at least in filaments (Straw,1956; Schmid, 1976), may play such a role. We related the appearance of gas in the locular cavity, before the stomium is opened, to the transport through
102 stomata and intercellular spaces of the anther (ch.4). The presence of these spaces in the filament may also have to do with this phenomenon, which means that the gas might originate from lower plant parts. The cuticle that we found in some of the intercellular spaces, makes the transport of gas rather plausible. On the other hand, Pfeffer (1904) and Knoll (1914) report the transfer of water from the cells to the intercellular spaces, causing voluminar changes on behalf of filamental movements.
2. The extension and rigidity of the filament.
In the extending parts of the filament (i.e. the base in the young stages and later on the tip) dictyosome vesicles are present near the plasma membrane. They probably have a function in the deposition of wall materials (Chrispeels,1980). Besides, the presence of tubular SER in the filament base near maturity is remarkable. Cresti and Keijzer (1985) related such SER to the transport of either phospholipidic components in favour of membrane formation or growth regulating substances in the rapidly growing pollen tubes of Nicotiana alata Link and Otto. Although the rate of growth in the filament tip is far less than in pollen tubes, it is relatively quick if compared with the filament base in earlier stages in which we found no SER. From this point of view a comparable role might be possible. From the pollen mother cell stage and especially towards maturity the intercellular spaces under the epidermis and in the central parenchyma are gradually replaced by wall materials. A further wall thickening to form real collenchyma does not take place. Schmid (1976) even reports the general absence of collenchyma in filaments. However, these depositions improve the rigidity of the filament, which may be necessary from the moment that the anthers are no more tightly surrounded by the tepals, i.e. from the microspore mitosis stage. A second process which may be related to rigidity is the considerable thickening of the outer tangential epidermis wall and the thickening and enlargement of the cuticle. In the filament these changes occur gradually from the tip to the base. In the anther the epidermis and cuticle change similar to the tip of the filament, without gradual differences between its tip and base (ch.5). Apart from a role in rigidity, we carefully related these structural changes, and especially the enlarging cuticular surface, to the improved ability to evaporate in favour of dehiscence (ch.1 and 5), although the thickening of the cuticle contradicted with such an explanation
103 (ch.5). However/ since the cuticular undulations on the filament can also be found on the region that does not dehydrate and that remains more or less turgescent after dehiscence, a direct relationship with evaporation has become less likely. Moreover, the gradual change from the thin smooth cuticle on the base to the thick undulated configuration on the tip (comparable with the cuticular types 6 and 2 respectively from the classification of Holloway, 1982) contradicts with the sharp border between the shriveled and the swollen part of the filament. On the contrary, this gradual change is accompanied by the proceeding cytoplasmic degeneration from the tip to the base. This might suggest an interrelationship between these processes, possibly a rigidity function for the outer tangential epidermal wall and its cuticle during a decreasing turgidity. Although these ideas remain speculative, the cuticular changes will certainly have a function, as we observed them currently in many randomly chosen species and on other reproductive organs like pistils (Willemse and Franssen-Verheyen,1986). Also the finding that the cuticle in intercellular spaces develops in a similar way, is remarkable. A third rigid structure will be the swollen epidermis cells. As their increasing vacuolation is preceded by starch breakdown, osmotical water uptake is likely. Schmid (1976) explained the often observed prominent intercellular spaces, xylem lacunae as well as a thin cuticle, as evaporation promoting factors of the filament. He wondered whether the turgidity in the filament after anther dehiscence is real turgidity, thus emphasizing the importance of other rigidity factors like cell walls. In G^ verrucosa we observed both signs of degeneration and plasma membrane retraction in the last developmental stages. These phenomena support Schmid's idea. However, the total absence of rigidity a few days after anthesis in our species suggests the contrary. From a functional point of view, one could say that the filament of G. verrucosa does not need any rigidity after anthesis, since the flower is hanging in this stage (Grootaarts,1978). In spite of the development of different mechanisms to keep the anther rigid after the degeneration of the cells, the question remains as to why this degeneration occurs relatively long before anthesis. In the anther the epidermis, endothecium and the parenchyma of the connective tissue remain turgescent and ultrastructurally intact until dehiscence. An explanation might be that breakdown products of the degenerating cells, which are for the first time observed shortly after microspore mitosis, are redistributed to
104 other floral parts, for example to the pollen grains that are strongly growing from that moment. Linskens (1974) found a redistribution of amino acids and proteins from the stamen to the style after fertilization in Petunia hybrida. As this shift occurred rather quickly, transport through the sieve elements is likely. In this light the prominent callose deposition in the pores between the sieve elements shortly before maturity, is remarkable. However, since we could not find such depositions between all the sieve elements, this transport system may remain (partly?) intact. Apart from breakdown products of degenerating cells, also the sudden disappearance of the starch during anther dehiscence may be a comparable mobilization of reserve substances for redistribution. Besides, the sudden breakdown of starch to sugars may cause an osmotical retraction of water from bordering tissues. In this way dehydration of the dehiscing anther and the filament tip may be promoted, apart from the role of evaporation in this process (ch.1 and 2). Burck (1906) proposed such a water retraction to the nectaries, but his experiments contradict with later findings of Hannig (1910), Wolff (1924) and Schmid and Alpert (1977). Associations between changes of reserve substances (starch and lipid) and processes in the anther are difficult to find. In spite of the considerable metabolic activities during anther development (ch.2,3 and 5), the amount of starch in the filament increases gradually up to maturity. On the contrary, the connective tissue lacks starch during the entire development. This difference in starch content between the filament and the connective tissue (i.e. the continuation of the filament in the anther) is quite remarkable. In our opinion it makes a role for the starch in the filament as a storage product for anther development less likely. The decrease of starch in the filament epidermis in the younger stages might have to do with both the thickening and the enlarging of the cuticle from that moment, since similar changes in the anther are accompanied by starch dissolution as well (ch.5). Besides, the starch dissolution may cause the vacuolation of the cells, as was mentioned before.
105 Chapter 7
THEPROCESSE SO FANTHE RDEHISCENC EAN DPOLLE NDISPERSA L 3.THEDEHYDRATIO NO FTH EFILAMEN TTI PAN DTH EANTHE RI NSOM EMONOCOTYLOU S SPECIES
C.J.Keijzer ,I.H.S .Hoe kan dM.T.M .Willems e
Summaxy
The development of the filament tip and the connective tissue during anther dehiscence were followed in Gasteria verrucosa. Allium cepa and Lilium hybrida using different microscopical techniques. Experiments with living tissues in the SEM and different kinds of micromanipulation were carried out to observe the behaviour of these dehydrating tissues. In Gj_ verrucosa anthesis is immediately followed by anther dehiscence, in which the filament tip dehydrates and shrivels as well. In L. hybrida the filament tip dehydrates and shrivels before anthesis. After anthesis the anthers start to dehisce immediately. In A^ cepa the filament tip starts to dehydrate and shrivel slowly immediately after anthesis. Anther dehydration and dehiscence occur up to days after anthesis and appear to depend upon the rate of extension of the filament. In all three species the epidermis cells of the stomiums have dissociated at anthesis. The delayed dehiscence of A. cepa is associated with the absence of stomata on the anther, in comparison with their presence on the anthers of the other two species. The continuously open stomata mainly determine the moment of dehiscence after anthesis, in spite of the presence of an open stomium. The shrinking filament tip functions as a hinge and enables the anther to dangle on the filament after dehiscence. This is supported by the presence of all-sided thickened cells in the anther epidermis opposite the filament, which prevent the inside-out bending locule walls to embrace the filament. This improved mobility of the anther may have to do with an optimal collaboration with the pollinators.
107 Introduction
Observations on the behaviour of the filament and theconnectiv etissu e duringanthe rdehiscenc ear esometime scurrentl ydescribe di nwork stha tdea l withanthe rdevelopmen t(Mepha man dLane,1969b ;Pacini,1986) .Specia lstudie s onthi stopi car escarc e(Schmid,1976 ; ch.6).Schmi d (1976)wrot ea prominen t work on filament development. He related different typical filament characteristicst oanthe rdehiscence ,althoug hh edi dno tespeciall yfocu so n the filament tip.Kend a (1950,1952)observe d stomata onth estamen so fman y speciesan dreporte dthei rpreferentia lpresenc e onth econnectiv etissue , contrary tothei rgenera labsenc eo nfilaments . Recently we investigated the filament development of Gasteria verrucosa andfoun dtha tth etrachear yelement si nth ecollapsin gfilamen tti premaine d opendurin gdehiscenc e (ch.6).Besides ,w erelate dth esudde ndecreas eo f starcha tdehiscenc et oa possibl eosmotica lwate rretractio nfro mth eanthe r toth efilamen ti nfavou ro fanthe rdehydratio nan ddehiscence .Apar tfro m itsstron g dehydration and shrinkage atanthe r dehiscence,th efilamen tti p formsa relativel ywel ldevelope d solidity system bythickenin g different siteso fth eepidermi scel lwall sa swel la sth ecuticle . Inthi schapte rth ebehaviou ro fbot hth efilamen tti pan dth econnectiv e tissue were investigated in threemonocotylou sspecie s which differo n essentialpoint sconcernin gth edevelopmen to fthes etissues .
Materialsan dmethod s
In this investigation stamens ofGasteri a verrucosa (Mill.)H.Duval , Liliumhybri d cv.Enchantmen tan dAlliu mcep aL .wer eused . To observe dehydration processes, fresh stamens were observed ina Jeo l SEM35 Ca t1 5k vfo ron ehour . Theconductiv eabilit y ofth exyle mwa sdetermine db yplacin gth ebas eo f dissectedstamen si na naquou ssolutio no f 1%aci dfuchsin .Thi ssolutio ni s sucked by the xylem and directly observable, thanks to the colourless epidermisan dparenchym ao fth efilaments .Th einfluenc eo fdifferen torgan s on anther dehiscence was determined byplacin gthes eisolate dorgan si nth e same relativehumidity , afterwhic h anther dehiscence was followed with a stereo microscope.I na simila rwa y alsoth einfluenc eo fstomat ao nanthe r dehiscencewa sdetermine db ysealin gth estomat awit hacid-fre evaselin .
108 Forelectro n microscopy pieceso fstamen s indifferen t developmental stageswer efixe di n3 %gluta raldehyd efo rtw ohour san d1 %osmiu mtetroxid e for4 5minutes ,bot hi n0.07 Mcacodylat ebuffe r(p H7.2 )a troo mtemperature . Afterdehydratio ni nethano lthe ywer eembedde di nth elo wviscosit yresi no f Spurr(1969) .Ultrathi nsection swer estaine dwit hlea dcitrat ean durany l acetatean dobserve di na Philip sE M 301a t6 0kV . Forinterferenc econtras tmicroscop y 2u m sectionswer ecu tfro mth esam e blocksan dobserve di nwate runde ra coverglass .
Observations
1.Gasteri averrucosa .
Toobserv eth edehydratio npattern ,fres horgan swer eobserve d inth e vacuum ofth eSEM , combiningth eadvantage so fa detaile dobservatio nan da promotedevaporation .Fo rthis ,th etepal swer eremove dfro misolate dflowe r buds justbefor eanthesis , after which thereceptacl e with the stamensan d thepisti lwa splace di nth eSEM .Fig. 1show sth edorsa lsid eo fa stame na fewminute safte renterin gth emicroscope .Fig.2 ai sa detai lo ffig. 1an d wasphotographe dabou ttw ominute safterwards .I nth eti po fth efilamen t someo fth eepiderma lcell sslightl ycollapsed .Thi sconcern smainl ycell so f theextremel y dorsalsid eo fth efilament .Fig.2 bwa smad ete nminute slater . The dehydration process continued. Some of the cells collapsed sostrongl y that the radial walls protrude.Apar t from the tangential direction,a longitudinalshrinkag eca nals ob eobserved .Afte r2 2minute s(fig.2c )th e dehydration process continued. In some cells bordering the initial dehydrating cells,sign s of shrinkage canb eobserve d aswell .Fig.2 dshow s the situation 36 minutes after the beginning. The shrinkage of the dehydratingepidermi scel lproceeded .I nlate rstage sth eoutwar dbendin g "walls"o f the loculespreven t thefurthe r observation of the filamenttip . Duringth ewhol eproces sth ecuticle-covere d wall ridges undulate much strongeri nth eregio nwher eth ecell sborde rthei rtangentia lneighbour s (fig.3). The number of ridgesremain s constant during dehydration. Inspit e of the strong dehydration, the stomata, which are only found on the connectivetissu eo fth eanthe r inthi sspecies ,remai nope n(fig.4) . Thesituatio ni nfixe dmateria lconfirm sth edorsiventra lpolarizatio no f thedehydratio nproces so fth efilamen t (fig.5).Th eepiderma l cellsopposit e
109 the tepals collapsed whereas opposite the anther they still have a considerablelumen . Inth elowe rpar to fth eanthe rth ewall so fth eepiderma lcell sar eall - sidedlythickene dlik esklerenchyma .Afte rdehiscenc eth eshape so fthes e cellsremai nunchange d (fig.6).Accordingly ,thi ssit edoe sno tshrin ko r bendinside-ou ta sdoe sth eres to fth eanthe ran d especially thelocul e walls. Todetermin eth e conductive ability ofth exyle m andth eevaporatio n ability of the stamen, isolated flowers or flower buds of different developmentalstage swer eplace d inaci dfuchsin .Thre edifferen tsituation s weretested :intac tflowe rbuds ,decapitate dbud san dbud sfille dwit hwater . The resultsfro m the former twotreatment sca nhardl yb edistinguished . However,th eflowe rbud sfille dwit hwate rsho wa slowe rrat eo ffuchsi n transporttha ni nth eothe rtw otreatments . Alsoth einfluenc eo fth estomat ao ndehiscenc ewa stested .Fo rthis ,th e tepals were removed from mature closed flower budsi nsit uan dth esit eo f thestomat a (fig.4)o fsom eo fth eanther swa s sealedwit hacid-fre evaselin . Within onehou r theuntreate danther sstarte dt odehisce ,whil eth etreate d onesremaine dclose dfo rman yhour san dfinall ydehisce dver yslowly .
1.Th e hydrated fresh stamen ofG .verrucos a from aflowe r budshortl y beforeanthesis ,a fe wminute safte rit sentranc ei nth eSEM .(60x ) 3.Detai lo fth efilamen tepidermi so fG .verrucos afro mth esam estag ea s fig.1b .Nea rth eradia lwall sth ecuticl eundulate sstronge rtha no nth e resto fth ecell .O nth econtrary ,nea rth etransvers ewall si tundulate s toa lesse rextent .(1100x ) 4.Th e (fresh)anthe r ofG^ _verrucos a after 20 minutes in the SEM.Th e stomatao nth econnectiv etissu eremai nope ndurin gth edehydration . (85x) 5.Th e filament tip of G^_verrucos a shortly after anthesis.Not e the dorsiventral polarization. On the dorsal side the epidermal cells collapsed (arrow) while opposite the anther they still have a considerablecel llumen .(200x ) 6.Th eall-sidedl ythickene d epidermis cellso nth e lowerpart so fth e anthero fG .verrucosa ,opposit eth efilament .(425x )
110 ^ ,.-\K \ ' . -.•:.^y
111 2.Alliu mcepa .
Contrary toth epreviou s speciesth eanther so fA^ _cep a lackan ystomata . Alsoth ebehaviou ro fth efilamen tti pdiffer sfro mG^ _verrucosa , whichca n beobserve di nth eSEM .Onc eanthesi sha sstarte d (fig.7),th efilament so f theinne rwhor lo fstamen sstar tt oextend ,liftin gth eanther sou to fth e openingbud .I nthi sstag eth estomiu mo feac hthec ai sopen .On eda ylate r anthesis has proceeded (fig.8)an d the mentioned anthers dehisced simultaniously. Theanther s of theoute r whorl stillma y remain closedfo r manyhour su pt oon eo rtw odays ,bu tfinall ythe yals ostretc han ddehisc e (fig.9).Althoug hthi si sth emos tcommo nsequenc eo fevents ,a considerabl e numbero fflower ssho wdifference sbetwee nth estamen so fon ewhor l(fig.10) . Inthes eflower sth eanther sdehisc ewit hinterval so fon et oman yhours , with theremarkabl e characteristic thatth e filaments of the dehiscing anthers stretch toa larger extent than those of the closed ones. When finallyth elatte ranther s dehisce,thei r filamentshav estretche d aswell , leadingt oth e situationi nfig.8 . Duringthi sproces s theti po fth efilamen tstart st oshrin k immediately uponth estar to fanthesis .Fig.1 1 showsa par to fth estame ni nthi sstage , comparablewit h fig.7.Th eshrinkag eo fth efilamen tti pca nb esee nbetwee n thetw otheca eo fth eanther .Figs.12a- dsho wcros ssection so fthi sstage . In the connective tissue both the vascular bundle and the bordering parenchymaar eturgescen t (fig.12a).I na lowe rsite ,wher eth e filament stillcontact sth eturgescen tconnectiv etissue ,th eepidermi scell san dsom e
2.Th efresh ,dehydratin g filament tip ofG^ _verrucos a inside theSEM . (600x) a.Detai lo ffig .1 ,photographe dtw ominute slater .Th eslightl yshrunke n epidermalcell s(S )o nth eextrem edorsa lsid eca nb edistinguishe d fromthei rmor eturgescen tneighbour s(T) . b.Te nminute slate rth edehydratio nha sproceede di ntha tth eradia l walls(arrows )extrud ei ncompariso nwit hth ecollapse dtangentia l wall. c.Afte r 22 minutes the epidermal cells borderingth eextremel y dorsal onesals ostar tt ocollaps e (arrow). d.Afte r3 6minute sth edehydratio nha scontinued .Th eoutwar dbendin g locule"walls "star tt omas kth efilamen ttip .
112 113 ofth eparenchym acell so nth edorsa lsid eo fth efilamen thav ecollapse d (fig.12b).Thi sselectiv eproces si ssharpl y restricted toindividua l cells (fig.13).I nth esit ewher eth efilamen tjus tseparate dfro mth eanther ,mos t ofth eepidermi scell so fth efilamen tcollapse d(figs.12 can d 14),excep t thoseopposit eth econnectiv etissue .Jus tunde rth eanthe rbas eth efilamen t looksturgescen t (fig.12d),a sdoe sth eres to fth efilamen t downt oit s base. Todetermin eth econductiv e ability ofth exyle m andth eevaporatio n ability ofth eStamen , isolated stamens with dehisced and non-dehisced anthersfro mth esam ewhor lo fstamen s (asi nfig.10 )wer eplace di naci d fuchsin.I nth estamen s with dehisced anthers the fuchsin transportoccure d twicea sfas ta si nth enon-dehisce dones . Tocompar eth etim eneede dt odehisce ,bot hanther swit han dwithou tthei r filamentwer ecu tof ffro mflower sjus tafte ranthesi s (asi nfig.7 )an dwer e laiddow nbesid eth edono rplant ,i.e .i nth esam erelativ ehumidity .Th e isolatedorgan sdehisce dslowe rtha nthos eo nth eplant .Th epossessio no fa filamentdi dno tinfluenc eth edehiscenc eproces so fth eisolate danthers .
3.Liliu mhybrida .
Theanthe ro fL ^hybrid aha sstomat ao nbot hth econnectiv etissu ean dth e loculewalls .Thes ewer ealway sfoun dope ndurin gth eentir edevelopmen t untilafte rdehiscence . Contraryt oth eothe rspecie sth efilamen ti sattache dt oth eventra lsid e
7.Th eflowe ro f Aj_cep a shortly after anthesis.Th e anthers are still turgescentan dremai nclose dfo ra relativel ylon gtime .(18x ) 8.Th eflowe ro fA ^cep aon eda yafte rth estag eo fth epreviou sfigure . Parto fth epetal s and sepals havebee nremoved .Th einne rwhor lo f anthersha sdehisced .(18x ) 9.Th e flower ofA^ _cep a one day after thepreviou s stage.Par to fth e petalsan dsepal shav ebee nremoved .Bot hwhorl so fanther sdehisced . (18x) 10.A flowe ro fA .cep awit hasynchronousl y dehiscinganther swithi non e whorl.Th eanthe ri nth e fronti sstil lclosed , contrary tohi s two dehiscedcompanion si nth ewhorl .On eo fth eanther so fth eothe r(non - dehisced)whor li sno tvisible .(18x )
114 115 ofth eanther .O nthi ssid eth esli tbetwee nth etw otheca ei smuc hdeepe r thano nth edorsa lside .Befor eanthesi sth efilamen tti pi sembrace db yth e two thecae of the mature, closed anther (fig.15),almos tinvisibl efro mth e outside. Theepidermi scell so fth econnectiv etissu eopposit eth efilamen t are small and their walls are all-sided thickened, compared with the epidermis cells of the two bordering locules (fig.15). Two days before anthesis the filamentti pshrivels ,whil eth eanthe rremain s turgescent (fig.15).A ttha tmomen tth estomiu mcell so fth eanthe rhav edissociated . Immediately afteranthesi sth eanther sstar tt odehisce ,whic h iscomplete d within a few hours. Thanks to the strong tangential shrinkage of the dehydrating loculewall san dth epresenc eo fth enon-shrinkin gthickene d epidermiscell so fth econnectiv etissue ,th einitiall yembrace dfilamen tti p isse t free (fig.16).I nthi swa y theanther sdangl efreel yo nth efilament s onceth edehiscenc eha scompleted . Thepatter ni nwhic hth efilamen tti pcollapse sagree swit hth eproces si n theothe rtw ospecies .Th eepidermi san dth eoute rparenchym a cellsar eth e firstt ocollaps e (fig.17),a proces stha tstart so nth eoutside-directe d side(i.e .i nth ecas eo fL .hybrid aventral )o fth efilament .
Discussion
Themorpholog yo fth eanther si nth edifferen tstage so fdehydratio ni n theSE M issimila rt othos ei neac ho fthes estage sfro mnaturall y dehiscing anthersi nsit u(Keijzer ,unpublishe d data).Consequently ,th eforce dSE M
11.Th eti po fth estame no fA .cep ashortl yafte rth estag eo ffig.6 .Th e filamentti pslightl ystarte dt oshrin ka sdi dth eti pan dbas eo fth e anther.(80x ) 12.Cros ssection so fth econnectiv etissu ean dth efilamen to fA.cep ai n thestag eo ffig.11 .(400x ) a.I nth econnectiv etissu eal lth ecell sar eturgescent . b.I nth esit ewher eth efilamen tjus tcontact sth eanther ,th eepidermi s and someparenchym a cellso fth edorsa lsid ehav ecollapse d (arrows). Theconnectiv etissu ei sturgescent . c.A littlebi tlowe rth eproces so ffig.12 bha sincreased . d.Underneat hth eanthe rbas eth efilamen ti sstil lturgescent .
116 f* " ,.- I. '. "•
.-^
117 dehydration represents the naturalprocess , with the exceptiono f themuc h higher rate of evaporation. This means that a comparison between thetim e andth erat eo fshrivelin gdoe sno treflec tth enatura lsituation . Untilanthesi sth esi xanther so fG .verrucos aar ecompletel y hydrated (ch.1,2 an d 5).Th esligh tshrinkag ei nfig.2 ai sprobabl ydu et oth efre e evaporationbetwee nth emomen to fflowe rbu dopenin gan dth emomen ti nwhic h thepictur e couldb emade .I nthi sspecie santhesi si simmediatel y followed bydehiscenc e (ch.1an d 2).Th eanthe rdehydrate s together with thefilamen t tip (ch.6), inwhic hth erat eo fevaporatio ndiffer sbetwee nborderin gcells . Theshar pborde rbetwee nth ehydrate dan dth edehydrate dzone sindicate stha t the dehydration isrestricte d tocertai n cells.A sth ecuticula r ridges approacheac hothe ran dthei rnumbe rremain sunchanged , iti sclea rtha tth e outer tangential wall shrinks during thisprocess .Thi s turns thecuticle - wall-complexint oa ver ycompac tstructure ,whic hprobabl yha sa functio ni n rigidity (ch.2an d 6). No ultrastructural differences could be observed in either the cytoplasm, thewall s orth e cuticlebetwee n individualcell so f thedehydratin gzon eo rbetwee nth etw ozone s (ch.6).Thi smean stha tth e backgroundo fthi sdifferentiatio ncoul dno tb edetecte dunti lnow . Fromth efuchsi nexperimen tsom econclusion sconcernin gevaporatio nca nb e
13.A detai lo fth eepidermi san dth eperipher eparenchym ao fth efilamen to f A. cepa inth ezon eo f fig.12b .Th edehydratio nproces si srestricte dt o certain cells.(1650x ) 14.A detail of the epidermis and periphere parenchyma of the A^_cep a filamenti nth estag eo ffig .12c .Th edehydratio ni srestricte dt oth e outer layers.(3200x ) 15.I n the closed flower bud of L.hybrid a the filament tip shriveled strongly and isembrace d byth econnectiv e tissuean d the bordering loculewalls .Th eepidermi scell so fth econnectiv etissu ear eall-side d thickened.(24Ox ) 16.Afte rdehiscenc eth efilamen tti po fL .hybrid ai sse tfre ethank st oth e shrinking locule walls andth e rigid, thickened epidermis cells of the connective tissue.Th euppe rpar to fth efilamen tha sshriveled ,th e lowerpart sar eturgescent .(36x ) 17.Th e outside-directed sideo ffilamen tti po fL .hybrid adehydrate s ina n earlierstag etha nth esid eopposit eth econnectiv etissue .Th ecollaps e ofcell sproceed sfro mth eepidermi stoward sth ecentre .(240x )
118 17
• T-YV *•• ^—r.> Aht lam m
V,
119 drawn.Althoug hth erelativ ehumidit yi nth eclose dflowe rbu di shig henoug h topreven tanthe rdehydratio nan ddehiscenc e(ch.2) ,i ti slowe rtha n100% , given the delayed fuchsin uptake after filling the bud with water.Th e transporto f fuchsin toth eanthe r agrees withth eobservatio n thatth e trachearyelement si nth eshrivelin gzon eremai nope n(ch.6) . Contrary to the situation inG^ _verrucos a (ch.2)an d L^_hybrida , the anther ofAj _cep adoe sno tdehisc eimmediatel y afteranthesis .Give nth e resultso fth eevaporatio n experiments, thisretardin g ismainl y controlled by the anther itself, as dissectedanther skee pthei rwate rfo rman yhours , contrary to the immediately dehiscing G^verrucos a anther (ch.2). The possessiono fa filamen tdoe sno tinfluenc eth espee do fdehiscenc eo fth e dissectedA .cep aanthers .Th esligh tdela yo fdehiscenc eo fth edissecte d organsa scompare dwit hth eintac tflowe ro nth eplant ,indicate san ywate r retractionfro mth eanthe rthroug hth efilamen tt oth elowe rplan tparts . Anotherpossibilit yi sa hormona lsigna lfro mthes epart st oth eanthe rt o promoteevaporatio nan ddehiscence . Thedifferenc ebetwee nth emomen to fanthe rdehiscenc eo fextende dan dno n extendedstamen swithi non eflowe ro fA^ _cep ai sconsiderable .U pt ono wou r current microscopical observations could not yet reveal any decreasing structural water barrier inth e filaments ofth etw o extension stages.Th e duringth eextensio nnarrowin gtrachear yelement sprobabl yd ono tpla ya rol e inthi sdifference ,give nth esligh tdecreas eo fthei r capacity.Preliminar y resultsfro mexperiment swit hTriticu maestivu mL .stamens ,carrie dou t recently,reveale da specia llaye ri nth efilamen ttip .Togethe rwit hth e xylem gapsi nth eextendin gPoacea e filaments (ch.2),thi s would meana remarkablespecializatio n inth egrasses ,whic h mighthel pu st oanswe rth e questions inth emonocotylou sspecie si nth efuture . InA^ _cep aanthe rdehiscenc e isprecede d by degeneration anda sligh t shrinkageo fth efilamen ttip .I nthi srespec tA .cep aca nb econsidere da s an intermediate between Gj_verrucosa , which doesno tsho w thisphenomenon , and L.hybrida ,whic hha s adehydrate d filament tipbefor eanthesi san d dehiscence.I nA^ _cep aan dL^ _hybrid a this dehydrationoccur s relatively slow,s otha tth edifferen tstage scoul db eobserved .Th eepidermi scell sar e thefirs tt ocollaps ,followe d byth eperiphera lparenchym a cells.Thi s suggeststha tevaporatio nplay sa rol ei nthi sprocess .Durin gdehiscenc e bothth efilamen tti pan dth eanthe rshrive lstrongly . Theall-side d thickened cells ina par to fth eanthe repidermi so fG . verrucosa andL .hybrid a oppositeth e filamentpreven t theshrinkag e ofth e
120 anther afterdehiscenc e inthes e sites.I nthi swa yth eshrivelin gfilamen t tipi sno tembrace db yth einside-ou tbendin glocul ewalls ,b ywhic hi tca n worka sa hinge .Thi senable sth eanthe rt odangl emor eo rles sfreely ,whic h determinesit swa yo fexpositio ntoward spolle nvectors .Th epresenc eo fsuc h cellsappear st odepen dupo nth especie san dma yfunctio ni nfavou ro fa n optimalcollaboratio nwit hth epollinators .A compariso nwit hthi saspec to f theflora lbiolog ywil lb einteresting . Inou ropinio nth enee do fa hing ei sa mai nfunctio no fth edessiccatio n andshrivelin g ofth efilamen t tip.A rol e incuttin gof f further water supply toth eanthe r isunlikely , giventh eope ntrachear yelement s inthi s zone. Thequestio nremain sa st owh yth efilamen tti po fL .hybrid ashrivel s relatively long before anthesis. This apparently does not accelerate dehiscence as L^_hybrid a dehisces much longer after anthesis than G. verrucosa,whic hha sn oprematur eshrivelin gfilamen ttip .Thi ssupport sou r theory concerningth efunctio no fth eshriveling ,mentione dabove .Moreover , inbot h A. cepa and L.hybrid a the effect (a dangling anther)i s only present(an dneeded )afte ranthe rdehiscence ,sinc eth eturgescen tlocul e wallsbefor edehiscenc etightl yembrac eno tonl yth edehydrate dfilamen ttip , butals oa par to fit sturgescen tzone .I nthi swa yth estame nbecome sa rigid,turgescen tro dove rit sentir elength . Thegenerall ydelaye dcollaps eo fth eanther-directe dsid eo fth efilamen t tip, ascompare dwit hth eothe rside ,ma yb edu et oth e micro-environmental humidity aroundth efilament ,a sultrastructura ldifference sbetwee nth etw o sidescoul dno tb efound .I nou ropinio nthi spolarit y isa phenomeno no f minorimportance . It must be emphasized thatth easpect smentione d above only concernth e dehydrationan dshrivelin go fth etissu ean dno tth ecytoplasmi cdegeneratio n or the decreasingturgescence , which we observed before in largerzone so f thefilamen t (ch.6).Th elatte rphenomen a mayhav et od owit hredistributio n ofsubstance st oothe rplan tpart s (Linskens,1974;ch.6) ,apar tfro mpossibl e functionsi nfavou ro fanthe rdehiscence . The observation that the stomata ofG^ _verrucos a andL; _hybrid a are continuouslyope ndurin gth edehydratio nprocess ,agree swit hearlie rreport s onothe r species (Kenda,1952;Schmid,1976) .I nman y speciesthe yremai nope n duringth eentir estame ndevelopment ,whic hi sreflecte di nthei runchangin g plastidial content (Kenda,1952) .The yprobabl yonl ypromot eevaporatio ni n favour of anther dehiscence (Schmid,1976).Thei r absence on theA .cep a
121 anther coincides with the delayed dehiscence after anthesis, as compared with the other investigated species, which do have stomata on the anther. Also the fuchsin experiment in A^_cep a supports this observation, as the dehisced anther evaporates faster than the closed one. Especially the clear results from the vaselin experiment with G^ verrucosa demonstrates their strong influence on the speed of dehiscence. In spite of the opened stomium, the tightly inward bent locule walls (ch.2) prevent a rapid dehydration of the anther via the tapetal side of the locule walls, in both this vaselin experiment and in A. cepa. In our opinion the presence and amount of stomata mainly determine the moment of pollen exposition to a large extent. This conclusion agrees with our idea (ch.6) that the enlargement of the cuticular surface by the wall ridges hardly promotes evaporation. This supports our hypothesis that solidity is the main goal of either the cuticular extension, the formation of wall ridges as well as the thickening of the epidermal wall (ch.5 and 6). A comparable vaselin experiment to determine the stomata activity could not be carried out with L^ hybrida anthers. Given the distribution of stomata over the entire surface of both the connective tissue and the thecae, sealing the stomata would also mean sealing the complete anther in this species. According to Kenda (1952) who investigated over 350 species, stomata are mostly found on the connective tissue. In some species they are present on the thecae, while they are seldom found on filaments. The number of stomata on anthers ranges widely, from one to many. Kenda reported Lilium henryi L. to have the highest number of stomata (224 per square mm). Their influence on dehiscence will certainly be decreased by the capacity of the anther, which is for example relatively high in case of the large Lilium anthers. Given this interrelationship between stomata, water capacity and a probable water supply through or retraction into the filament (ch.6), stomata on anthers determine the moment of pollen exposition and accordingly may influence the occurrence of phenomena like protandry or protogyny. In this light they may be important for the occurrence of self- or cross pollination. From that point of view a floral biological approach of the thesis of Kenda (1950) may be interesting.
122 Chapter 8
IN VITRO CULTURE AND INTRASPECIFIC TRANSPLANTATION OF TETRADS INTO OLDER ANTHERS
M.C. Reinders and C.J. Keijzer
Summary
Torevea lth einfluenc eo fth ecallos ewal ldurin gmicrosporogenesis / tetradso fGasteri averrucos awer etransplante d toolde ranther si nth efre e microsporestage .Besides ,tetrad swer eculture di nvitro . In the transplantation experiments the callose walls of the tetrads sometimes dissolved, inothe r cases theypersisted . Inth elatte rcase sth e tetrads could be recognized 48 hours after the transplantation in the acceptorpopulation . They hadforme d starch and vacuoles,accordin gt oth e behaviour ofth eaccepto rpopulation ,bu tt oa lesse rexten ttha n inth e lattercells . Tetradsi nvitr oi na Nagat aan dTakeb emediu mwit h10 %sucros eforme d largeamount so fstarc han dvacuoles .Afte r 12day sthe y expanded andbrok e outo fth ecallos ewall ,afte rwhic hthe yappeare dt ohav eforme da cellulos e wall.Th elatte rphenomeno nals ooccurre d whenartificiall yenzymaticall y releasedprotoplast sfro mtetrad swer eculture do nthi smedium . Tetradsculture di na naquou ssolutio no f 1%glucos e didno t form starch andvacuoles .I nthi smediu m thecallos ewall so fth eolde rtetrad s(wit ha n exine)wer edissolve dan dth ecell sreleased . The relationships between the callose wall andth e different changesi n theexplante dcell sar ediscusse dan dcompare dwit hth enorma ldevelopment .
Introduction
As a part of our studies on tissue interactions in stamens, an in-vitro and semi-vitro program was started to clarify aspects of the pollen tapetum interaction and the role of the meiotic callose wall.
123 Recently,Pacin ie tal .(1985 )reviewe dth eimportan trol eo fth etapetu m inpolle ndevelopment .O nth econtrary ,Tanak aan dIt o(1980,1981 )an dTanak a etal .(1980 )reporte d theapparentl y normal development of microsporesan d polleno fLiliu mlongifloru mThunb .o nrathe rsimpl ecultur emedi ai nvitro . This indicates that the supply of very specialized substances from the tapetumi sno tabsolutel ynecessar yan dtha tth edevelopin gpolle ngrain sar e rather autonomous.Althoug h these contradicting resultsma yno tb emutuall y exclusive,man yaspect so fth erol eo fth etapetu m andth elocula rflui d remained stillunclear , inspit e of our investigations onthi stopi c(ch.5) . Toanswe rsuc hquestions ,w echange dth esynchronit ybetwee nth emicrospore s andth etapetum .Fo rthis ,w etransplante dtetrad so fGasteri averrucos a (Mill.)H.Duva lt ofurthe rdevelope danther so fth esam especie san dfollowe d their development during a few days.Furthermore , we cultured isolated tetradso ndifferen tmedi ai nvitro . Giventh edevelopmen to fth etapetu mi nth emal esteril eAlo evera ,whic h appeared tob e independentupo nth epresenc e of vital microspores (ch.4), a normal and continuing tapetal development may be expected during a confrontationwit hyounge rmicrospore si nth etransplantatio n experiments. Alsoth erathe rasynchronou sdevelopmen to fth epolle ngrain si nanther s duringth ewinte rperio dmake ssuc ha nexpectatio nplausible .Willems e(1981 ) transferred G^_verrucos a microspores to Lilium hybrida anthers and demonstratedchemica lsimilaritie si nth ecompositio no fth epolle nwall so f bothspecies .Thi sindicate stha tth e(tapetal )depositio no fexin ematerial s continues,als oo nnon-specifi cpolle ngrains . Asa nai dt ocompar eth emicrospore sfro mthes eexperiment swit hthos ei n the normal situation, the ultrastructural description of the pollen development inthi sspecie swa suse d(ch.5) . Inthi s chapter the first results from in-vitro and transplantation experiments arepresented . The final goalo fthes e experiments willb eth e complete controlo fpolle n development, inorde r to influence thisproces s forbreedin gpurposes .Thi si sa rathe runcommo napproach ,a spolle nculture s in vitro are mostly used for androgenic purposes (see the review of Maheshwarie tal.,1982) .
Materialsan dmethod s
Forin-vitr o cultureexperiments ,anther s ofGasteri averrucos a(Mill. )
124 H.Duval inth etetra dstag ewer esurface-sterilize d in70 %ethano lfo r10-3 0 secondsan di n2 %potassiu mhypochlorid efo r2- 5minutes .Afte rrinsin gtwic e inaqu abides tfo r1 0minutes ,th eanther swer edecapitate dwit ha razo r bladean dth etetrad swer ecarefull yremoved .Nex tthe ywer eculture d intw o differentmedia : 1.A droplet of the liquid medium ofNagat a andTakeb e (1971)wit h10 % sucrose (0.3M)bu twithou tmannitol .Thi sdrople twa ssituate dbetwee na microscopeslid ean da cove rslide ,i nwhic hth elatte rwa splace do ntw o othercove rslides ,thu sformin ga bridge .Thi sconstructio nwa sseale dwit h sterilized acid-freevaselin . 2.A drople to fa naquou ssolutio no f1 %glucos e(0.0 6M) .Thes eculture s werecarrie dou ti ncultur eslide sno .66516 1o fGreine ran dSons . Youngmicrospor eprotoplast swer eartificiall yrelease dfro mthei rcallos e wallsb yincubatin gtetrad sfo r4 5minute si nth ementione dmediu mo fNagat a andTakeb ewit h10,00 0units/m lSnai lDigestiv eJuic e(Koch-Light) ,whic h containscallase . Theprotoplast san ddebri swer erinse dtwic ei nth emediu m without this enzymemixture .Apar tfro m anutritiona l function,th esucros e acteda sa nosmoliticu mthu spreventin gth eprotoplast sfro mbursting . Allth eculture swer ecarrie dou ta t2 7degree sC i na 16/ 8hour sda yt o nightratio .Bot htype so fcultur e slides allow a directobservatio n ofth e cellsunde r the lighto rinterference-contras tmicroscope .Th evitalit yo f culturedcell swa sdetecte dwit hfluorescein-diacetat e(FDA )accordin gt o Heslop-Harrisonan dHeslop-Harriso n(1970) .Cellulos ewa strace dwit hchloro - zinc-iodine,starc hwit h iodinei npotassium-iodin e (Jensen,1962) . Forth etransplantatio nexperiment sw eremove don etepa lfro ma flowe rbu d andmad ea smal lhol ei nth eaccepto ranther ,usin ga glas scapillar .Nex t tetradswer esucke dfro ma dono ranthe ro fa younge rflowe rbud ,als ousin ga glasscapillar ,an dinjecte d intoth eaccepto ranther ,jus tbesid eth efirs t holewhic hserve da sa noutle ti ncas eo fa to ohig hpressure .Finall yth e flowerwa sreclose dwit hth edissecte dtepa lan dseale dwit hMowilit hDM2K L (Hoechst).Thi si sa modificatio no fth etechniqu eo fWillems e(1982) .
Observations
1.Th e tetradtransplantations .
Asmal lamoun to fyoun gtetrad s (fig.1 )wa stransplante dt olocule so f
125 anthers filled with young free microspores (fig.2). After 48 hours the mixtureo fdono ran daccepto rcell swa sremove dfro mth elocul ean dobserved . The acceptor microsporesha d turned into bicellatepolle ngrain s (fig.3).I n 15experiment sth edono rtetrad sha dlos tthei rcallos ewall san daccordingl y had turned into free microspores.I n another 10experiment s they still possessedthei rcallos ewall san daccordingl ywer eclearl ydetectable .I nth e latterexperiment stw oremarkabl estructura lchange sha doccurred .Firstly ,a smallamoun to fstarc hha dappeare di nth ecell s (fig.4). Secondly, vacuoles werefound ,randoml y dispersed throughout thecel l (fig.5).Th eFD Ates t showeda vitalit yo f80 %fo rthes etetrad safte rthei r4 8hour ssta yi nth e acceptoranther .
2.Th etetrad si nvitr oo nth eNagat aan dTakeb emedium .
Tetradsculture do nth eNagat aan dTakeb emediu m immediately storeda largeamoun t of starch,disperse d over justa fe w hugeplastid s (fig.6). Besides, randomly dispersedvacuol eformatio nwa sobserve d (fig.7).Afte r 12 days incultur e someo f thecell sha dexpande dt obrea kthroug hth ecallos e wall and freed themselves (fig.8). These released cells had been surrounded bya rathe rthic kwal lwhic h appearedt ob ecellulosi caccordin gt oth eChl - Zn-Itest .Afte rthes e1 2day s30 %o fth ecell swer estil lalive ,a sreveale d withth eFD Atest .Thes elivin gcell swer eonl yfoun dnea rth eedg eo fth e drop of medium. After three weeks in culture all the cells had died. Artificially enzymatically releasedtetra dprotoplast sdevelope da cellulos e wallafte rtw o daysi ncultur e (figs.9,1 0an d 11).
1.Tetrad so fth edono rpopulatio nbefor etransplantation .(95Ox ) 2.Microspore so fth eaccepto rpopulatio n beforeth etransplante d tetrads wereinjected .(1020x ) 3.Tw o daysafte rth etransplantatio n theaccepto r microsporeshav eturne d intobicellat epolle ngrains .(680x ) 4.Tw o daysafte r thetransplantatio n starchgrain sca nb etrace dwit hth e IKItes ti nth etransplante dtetrads .Als osmal lvacuole sar epresen t (arrows).(1425x ) 5.Th e same population as in fig.5. In some cells larger vacuoles are present(arrow) .(1425x )
126 127 3.Th etetrad si nvitr oo nth eglucos esolution .
Youngexine-les stetrad swit ha nintac tcallos ewal lwer e culturedi nth e glucose solutions (fig.12).Afte rtw o daysi ncultur estil l90 %o fth ecell s appearedt ob ealive ,accordin gt oa contro l with theFD A test.Besides , little holes had appeared in the callose walls of many of the tetrads (fig.13).Afte rsi xday si ncultur ebot hth enumbe ran dsiz eo fthes ehole s had increased, whereas the vitality was still 80%(fig.14) .Lik e inth e previousexperiment ,th ecell snea rth eedg eo fth edro po fmediu m appeared tob eth elonges tlivin gones .Th ecallos edigestio nha dcontinued ,releasin g some of the microspores (fig.15). However, all of the free microspores possessed anexine , contrary toth e stillencapsulate d microspores (fig.15). Contraryt oth epreviou sin-vitr oexperiment ,formatio no fneithe rstarc hno r vacuolescoul db eobserved .Afte r1 5day smos to fth ecell swer edead .
Discussionan dconclusion s
Thetetra dtransplantatio nexperimen twa scarrie dou tt oavoi dth enatura l dissolutiono fcallos eafte rmeiosis .Besides ,th ecallos ewal lacte da sa marker to recognize the donor cells.However , the wall did not show a
6.Tetrad safte rtw oday si ncultur ei nth eNagat aan dTakeb e medium.Hug e starchgrain sar edeposited .(1425x ) 7.Th esam epopulatio na si nfig.6 .A numbe ro fcell sfor mlarg evacuoles . (1425x). 8.Tetrad s after 12 days in culture in the Nagata and Takebe medium. Expandingmicrospore sbrea kthroug hth ecallos ewall .The yhav eforme da cellulosewall .(950x ) 9.Tetra d protoplasts plasmolyse inside their callose wall due toth e sucroseconcentratio n (10%), two minutes after incubation inth eNagat a andTakeb emediu mwit hth eenzym emixture .(950x ) 10.Afte r4 5minute si nth eenzym emixture ,th emai npar to fth ecallos eha s beendigeste dan dth eprotoplast sar ese tfree .Vita lone s(arrows )ca n bedistinguishe dfro mletha lshrivele dones .(9 5Ox ) 11.Afte rtw oday si ncultur ei nth eNagat aan dTakeb emediu mwithou tth e enzymemixture ,th eprotoplast shav eforme da cellulos ewall .(800x )
128 129 consequentreaction .Th edissolutio no fcallos ei n60 %o fth eexperiment s means that the callase remains active for some time after the callose digestionstag ei nth eaccepto rplant .Th epersistin gcallos ewall si nth e other 40%ma y be an effect of the technique itself.Th e more cells are transplanted incompariso nwit hth ecapacit yo fth eacceptor-locule ,th e strongerwil lb ethei rinfluence .A si nth etetra dstag eth ep Hi nth elocul e ishighe rtha ni nth efre emicrospor estag e (Keijzer,unpublishe d data)/ the pHo fth e locular fluid may havechanged .Thi s candecreas e the callase activity,whic hi soptima la tlo wp H (Izharan dFrankel,1971) . Moreover/an y wounding effect due to a too large transplantation hole may have a deregulatingeffect . The amount of developed vacuoles is comparable with those inyoun g microsporestha tdevelo pi n4 8hour sfro mth einitia lstag eo fth edono r tetrads (ch.5).O nth econtrary ,th eamoun to fstarc hi sfa rles stha ni nth e naturalcase .A sth eaccepto r microspores storemuc hstarc hi nthes estages , the substrate, which originatespresumabl y from the digested callosewall s (ch.5),wil lb eabundantl ypresent .Th ever ysmal lamoun to fstarc hi nth e transplanted tetrads willno tb edu et oendogenou s factors of thecell s themselves,give nth estarc h synthesis duringcultur e inth eNagat aan d Takebemedium .Accordingly ,exogenou sfactor sma ypla ya role ,fo rexampl ei n case the locular fluid contains no sucrose inthi s stage (which is the substrate inth e in-vitro experiment), butothe r oligosaccharides whichca n notpas sth ecallos e wall.Thi swoul d agree withth einabilit y of large moleculest opas sth ecallos ewal l(Heslop-Harriso nan dMacKenzie,1967) . Theabundan tstarc hproductio no fth e microspores inth e medium with sucrosemean stha tthi ssuga rca npas sth ecallos ewall .Th eamoun to fstarc h storingplastid sagree swit hth enatura lsituatio n inth emicrospore s after
12.Youn g tetrads immediately after thestar t ofth e culture in 1%glucose . (800x). 13.Afte rtw oday si ncultur ei nth eglucos esolutio nlittl ehole sappea ri n thecallos ewalls .(400x ) 14.Afte r six days in culture in the glucose solution the holes have enlarged.(800x ) 15.Afte r 15 days in culture in the glucose solution most of the exine bearingmicrospore sar ese tfree .Th ecallos ewall so fth eexine-les s microsporesar estil lintact .(400x )
130 131 callose digestion (ch.5).Althoug h thevacuolatio n in the cultured cells is alsoa norma lphenomeno nafte rth ecallos edigestion ,th erandoml ychose n positiono fthes evacuole s doesno tcoincid ewit hth epolarize dsituatio ni n thenatura lproces s (ch.5).Thi s apolarization was also observed inth e transplanted tetrads.Thi smean stha tth epolarit yo fth ecell si sdisturbed . Thismigh tb edu et oa reactio no nit senvironmen t (i.e.th elocula rflui do r thecultur emedium) , whichpresumabl y contactsth ecel lonl yvi ath ecolpus , whichha sa polarize dpositio ni nth ecel l (ch.5).Anyway , thevacuolatio ni s anactiv eprocess ,whic hoccur si nspit eo fth ehig hosmolarit yo fa mediu m with10 %sucrose .Als oth esoli dcellulos ewall so fth eescapin gcell sar e not comparable with the natural situation in which the sporopollenin containing exine isdeposite d (ch.3an d5) . Inconclusio non ecoul dsa ytha tth ereactio no fth ecell si nth eNagat a andTakeb emediu m isa mixtur e ofth enorma l development and declining processes. Especially the deposition of acellulos ewal li scomparabl ewit h thevegetativ edevelopmen to fandrogeni cpolle n(a si sth eintin edevelopmen t inth enorma lsituation?) .Thi sma yb eimportan tfo randrogeni cpurposes .I n generalth e tetrad stage is seldom used as asourc e forhaploi d plants (Maheshwari etal.,1982) , inspit eo fpossibl eadvantage slik eth eabsenc eo f theexin e( apossibl egrowt hlimitation )o rth erelativel y lessproceede d generativedevelopment .Wit hrespec tt oth elatte rargument ,a possibilit yt o detectpredestinate dvegetativel y developingpolle n grains (Sangwanan d Sangwan-Norreel,1985)a searl y asi nth etetra dstag ei so fgrea timportance . Themos tremarkabl edifference sbetwee nth eculture so nth eNagat aan d Takebe medium on onehan d and the sugar solutions on theothe rar eth e selectivecallos edigestio nan dth eabsenc eo fstarc han dvacuole si nth e lattermethod .Th erelatio nbetwee ncallos edigestio nan dth epresenc eo f exinesindicate sa developmental-stage-dependen t endogenous factor forth e digestion.Althoug hth emomen to fth efirs tobservabl ecallos edigestion , aftertw odays ,make sa nexogenou s (tapetal?)origi no fth eenzym erathe r unlikely,an yeffec to fremaine d callasefro mth edono rlocul ecanno tb e excluded. Callaseproductio n by the tetrads themselves would disagree with the assumption that the tapetum does supply the callase (Mepham and Lane,1969b; ch.5).Th equestio nremain sa st owh y thisphenomeno n isno t observed inth eNagat aan dTakeb emedium .Furthe rexperiment swil lsho w whetherthes edifference sar eeffect so feithe rth edifferen tcarbo nsources , theinorgani cpart so fth emedi ao rothe rfactor slik eth epH .Unfortunately , iti sno tye tclea rwhethe rth eexine sha dbee ndeposite ddurin gth ecultur e
132 orwer einitiall ypresent . Theabsenc eo f starch duringth eentir ecultur eperio di nth eglucos e solution indicates the inability of the cell totur n glucose into this sucrosepolymere .Th elinke dabsenc eo fvacuoles ,a scompare dwit hthei r presencei nth estarc hstorin gcell si nth eNagat aan dTakeb emedium , mayb e duet oa nosmoti ceffect .I ti splausibl ethat ,apar tfro mstarch ,th ecell s in the latter medium store sucrose as well, which stimulates their vacuolation. Inal lth e culture experiments astron grelationshi pexiste dbetwee nth e lethalityan dth epositio no fth ecell si nth edro po fmedium .I ti srathe r clear that insufficient gas exchange is the reason for the premature lethality andaccordingl y the culture method has tob e improved inth e future.
133 General conclusions and discussion
In this thesis the stamen development is mainly discussed from our findingsi nGasteri averrucos aan dth etw oAlo especie so fchapte r4 .Bot h latterspecie sappeare dt ob ealmos tindistinguishabl efro mG .verrucosa , concerning the ultrastructural aspects of anther development. Further investigated species were only used to a lesser extent, mainly for comparisonswit hou rresult sfro mth ementione dthre eones .Fo rthis ,th e nameso fth eplan tspecie swil lgenerall yno tb euse di nthi spar to fth e work.I fnecessary ,the yca nb efoun di nth efrequentl yreferre dchapters . Sinceou robservation sar erathe rthoroughl y discussedi nth eindividua l chapters,thi spar twil lb efocusse do nth emai nlines .
Thestame ndevelopmen t
Thestame ndevelop so nbehal fo ftw omai npurposes :th eformatio nan d exposureo fpolle nrespectively .Th efirs tpurpos ei smainl yth etas ko fth e tapetuman dth esporogenou stissu e(ch.1,4, 5an d8) .Th esecon dimplie sth e differentstep so fanthe rdehiscence ,mainl y located inth eepidermi s (with or without stomata) and the endothecium (ch.1,2,4 and 7),an d pollen dispersalwhic hi senable db yspecia lsubstance s onth esurfac ebetwee nth e (sporophytic)tapetu m andth e (gametophytic)polle n grains (ch.3an d 5). The filamentdetermine sth epositio no fth ematur epolle ni nth eflower ,th e developmento fit sti pimprove sth emobilit yo fth edehisce danthe r(ch. 6an d 7). Duringmeiosi sth enucle io fth emeiocyte schang efro mdiploi dt ohaploid , which isaccompanie d bycytoplasmi c changes,lik eth epresenc e of some multimembrane-boundvesicle s andnucleoid s (ch.5).Thes ecytoplasmi cchange s occur to a lesser extent than inLiliu m hybrida (Dickinson and Heslop- Harrison,1977).Immediatel yafte rth esecon dmeioti cdivisio nth ehaploi d nucleistricktl yborde rth eperipher yo fth eorigina lpolle nmothe rcel l (ch.4an d 5). Here ashee t of RER is laid down alongth ewall ,probabl y decreasingth eexin edepositio nt ofor mth ecolpu s(Willemse,1972 ; ch.5). Duringth elat etetra dstag eth enucle imigrat et oth eopposit ewal l(ch. 4 and 5), probably duet omicrotubul eactivit y (VanLammere ne t al.,1985).Th e
135 meiotic callose walls turn the originally irregular cell shape intoa sphericalone ,b ywhic h thecel lca ndivid eint ofou requall y sizedan d shapedmicrospores ,thu sexcludin gan yselectiv einfluenc eo fsiz ean d shape duringpollinatio n (ch.5).Besides ,Waterkey nan dBienfai t (1970)suppose d that this callose wall might serve as amol d for the exinepatterning . Furthermore, itwork s asa filteragains t large molecules,wherea s smaller molecules can pass the wall (Heslop-Harrison and MacKenzie,1967 ; Southworth,1971) . The latter topic was investigated using in-vitro and transplantationexperiment san dsom epreliminar yresult swer epresente d (ch.8),indee dindicatin gan yselectiv einfluenc eo fth ewall .Th ecytomicti c channels inth e callosewall sma ypla ya rol ei nth esynchronizatio no fth e meiotic divisions (Heslop-Harrison,1968)a s well as inth e cytoplasmic reallocationi norde rt omak eth emicrospore s equally shaped(ch.5) . Duringmeiosi sth e (pro-)orbicules are deposited on the inner tangential wallo fth etapetu mcells .The yar econnecte dwit ha tapeta lmembrane ,whic h surrounds the entire tapetal cell. Both structures will contain the acetolysis-resistant sporopollenin ina laterstag e (ch.3,4an d 5).Nex tth e meiotic callosewall s are digested, together with thepecti n and cellulose wallso fbot hth epolle nmothe rcell san dth etapetum .Afte rthes edigestion s the tapetum cells are only covered with the orbicule-bearing tapetal membranes andth emicrospore swit hth eexines .Th eenzyme sfo rthes echange s originatefro mth etapetu m(ch. 3an d5) .However ,som ecallas esuppl yfro m the microspores themselves may notb eexclude d (ch.8). The callaseactivit y may be continued for arelativel y longperio d after the release ofth e microspores (ch.8). Thebreakdow nproduct so fth ecel lwall sar eprobabl y usedfo rth estarc hstorag ei nbot hth etapetu mcell san dth emicrospores ,b y which the callose canb e seena sa nearl yreserv esubstanc efo rth efurthe r development of the pollen grains (ch.5). At least after this starch depositionth e tapetum candevelo p independently uponth epresenc eo fvita l microspores (ch.4).Apar t from thestarc hproduction ,th etapetu m grows strongly inthi sstag eb ybot hth eincreas eo forganelle san dvacuolatio n (ch.1an d5 ) Themicrospore sdivid eint oa vegetativ ean da generativ ecell ,whic har e initiallyseparate db ya callos ewal lwithou tplasmodesmat a(ch. 3an d5) . Thanks to itspolarity , which isdu e tobot h microtubule activity (Van Lammeren etal.,1985 )an dvacuolation , the generative cellreceive s asmal l parto fth ecytoplasm ,withou tan yplastid s (Schroder,1985 ; ch.5).Apar tfro m the absence ofplastid s from the sperm cell after fertilization, this
136 phenomenon may be important for manipulations with isolated sperm cells for breeding purposes (Wilms et al.,1985; Keijzer et al,1986). As the uptake of locular substances presumably occurs through the colpus (Rowley and Flynn,1971; Christensen and Horner,1974; Pacini,1986), the presence of a callose wall temporary cuts the generative cell off, not only from the vegetative cell, but also from locular supply, which may be related to its declining development. This is a direct result of the polarity in the pollen grain (ch.5). After the dissolution of the callose wall the generative cell is tied off from the intine, which occurs in close association with stacks of RER (ch.5). We could not find microtubule activity in this process until now (Van Lammeren et al.,1985), contrary to other species (Dickinson,1975) . Once this process has finished, the generative cell turns into a spindle shape, which is maintained by microtubules (Van Lammeren et al.,1985). Finally it is surrounded by the vegetative nucleus (ch.5), in which tubulin may be involved (Van Lammeren et al.,1985). This association is maintained in the dehydrated pollen grain after dehiscence (Keijzer et al.,1986) and may be a preparation for the transport of the nuclei through the pollen tube. From the young microspore stage the starch in the tapetal plastids gradually turns into a lipid-like substance, after which the plastids fuse with each other and with the content of the RER to form a large amount of pollenkitt (ch.3). The other cell contents degenerate (ch.3,4 and 5). The amount of cytoplasm in the pollen grains increases sharply. The starch in the vegetative cell is dispersed over many smaller elaioplasts. Furthermore, a large amount of lipid droplets appear and form the main part of the reserve substance in the mature pollen grain (ch.5), according to the classification of Baker and Baker (1979). During this process the locular fluid disappears (ch.1). It may be taken up by the pollen grains (ch.1) or by the epidermis and endothecium cells (ch.4). The pollen grains expand due to their cytoplasmic increase, in spite of the disappearance of the vacuoles (ch.5). Finally they exceed the capacity of the locules and press themselves together with the orbicule-covered tapetal membrane into the tapetum cells. In this way the pollenkitt is transferred by capillary action into the locule, between the pollen grains (ch.3). The moment and duration of transfer during the tapetal development may determine whether the pollen grains are covered with merely pollenkitt or with a mixture of pollenkitt and tryphine (ch.4). The pressure of the pollen grains mentioned above, sticks them to the orbicule-covered locule wall. Since either the pollenkitt, the tapetal membranes with the orbicules and the exines are hydrophobic, a hydrophobic
137 sticking complex betweenpolle nan dlocul ewal li screate d (ch.3).Afte r antherdehiscenc eth estickin gpropertie so fthi ssyste mappea rt oresis twe t circumstances,thu spreventin ga nundesire d losso fpolle n (ch.3). Asmentione dbefore ,th eprocesse so fanthe rdehiscenc ear emainl y duet o theactivit yo fth eepidermi san dth eendothecium ,apar tfro ma speciall y developingtissu ebetwee neac htw oborderin glocules :th eseptum .Mos to f theseprocesses ,whic hwil lb edescribe di nth enex tparagraphs ,hav ebee n reportedi nchapte r2 . The firstsign s of the dehiscenceproces s canb eobserve d inth eseptum . Beforeth edepositio no fwal lthickening si nth eendothecium , starch canb e observed inth eendotheciu m (ch.5),excep ti nth esit ewher etw olocule s bordereac hother ,whic hi sth eearlies tpossibilit y torecogniz eth eseptu m as such.I nth estag ei nwhic h thegenerativ e celli stie dof f from the intine,th ecell so fth eseptu m dissociate byth e lysiso fthei r middle lamellae.Fro mtha tmomen tth elocule sar ekep tclos eonl yb yth etapeta l membranes,whic har erupture di na late rstag eb yth emechanica lforc eo f either the expandingpolle n grains (ch.1,3an d 5)o r the inward turning loculewalls .Afte rthi s rupture the stomium isonl y kept close byth e epidermis.No w the twopolle npopulation s canus eth espac eo f thetw o locules,whic hthe yca nenlarg ei ncas ethe yexcee dth ecapacit yo fth etw o loculesdu et othei rexpansion . The openingo f the stomium isprecede db ystructura lchange si nbot hth e epidermis and the endothecium. In these layers a part of the starch disappearsi nth elate rstage so fanthe rdevelopmen t (ch.5)an dth ecell s swell byvacuolatio n (ch.1an d 5).Breakdow n products of the starch may increaseth eosmoti cpressure ,thu sincreasin gth evacuolation .Thi swa s demonstrated by Woycicki (1924b)i nth ehug e epidermis cells which are borderingth e stomium of lily anthers.Th eneede dwate r may originate from the locular cavity (ch.4).Besides ,U-shape d wall thickenings are deposited in the endothecium cells, causing a rigid inner tangential wall but preservingth eelasticit y ofth eoute ron e (ch.3,4an d 5).Th ethickening so n the radial and transverse walls maintain afunctiona l distancebetwee nth e twodifferentl yreactin gtangentia l walls.Th e signal forth e depositiono f these thickenings originates from insideeac h individual loculewal l(ch.4) , presumably from the tapetum (De Fossard,1969). The collaboration of a swellingepidermi san dendotheciu mo non ehan dan dth erigi dinne rtangentia l endothecium wall on the other bends the locule walls in centripetal direction.Th eforce so ftw oborderin glocul ewall scaus eth emechanica l
138 rupture of one of the radial walls between the small epidermis cells of the stomium, which is opened in this way. Next these inward bending locule walls keep the anther closed, in spite of the opened stomium. This prevents the premature loss of pollen into the cavity of the closed flower bud as well as premature dehydration and dehiscence of the anther (ch.7). This configuration is the final situation in the closed flower bud in many species. The next step is the dehiscence of the anther, due to the outward bending locule walls after the dehydration of the epidermis and the endothecium. This mechanism works contrary to the inward bending movement of the previous step. In most of the investigated species this dehydration is a direct result of anthesis, as the relative humidity of the air around the anthers decreases at that moment, thus increasing the evaporation. The presence of pigments in the epidermis may increase the temperature in these cells after light absorption and improve the evaporation (Hannig, 1910). The evaporation is strongly accelerated by continuously open stomata on the anther, whose presence depends upon the species (ch.7). In this way they influence the moment of pollen exposition and may determine phenomena like protandry or protogyny (ch.7). If present, stomata are mostly restricted to the connective tissue (Kenda,1952). Apart from evaporation, water may be retracted from the anther to the filament, in which a sudden breakdown of starch is observed shortly before anther dehiscence (ch.6). In this way an osmotical uptake of water may occur, comparable with the process in the anther epidermis before the stomium is opened. The filaments of the grasses presumably posses special mechanisms to prevent further water supply to the anther at dehiscence (ch.2 and 7). In the epidermis cells another remarkable structural change can be observed, as the outer tangential wall progressively forms ridges from the microspore mitosis stage, by which the cuticular surface is enlarged (ch.2,4 and 5). Synchronously the cuticle is thickened (ch.2,4 and 5). We related these changes to both rigidity and an evaporation promoting function (ch.5). The latter function became less likely after we found these phenomena also to occur progressively from the tip to the base of the filament (ch.6) and on other floral parts (Willemse and Franssen-Verheijen,1986) which do not dehydrate after anthesis (ch.6). However, the question remains as to why the anther has a better ability to evaporate in older than in younger stages (ch.1). After dehiscence the pollen of entomophylous species is sticked by the pollenkitt to the orbicules on the inside-out bent locule walls, as described before. The position of the so formed pollen source may be optimalized in
139 relation to the pollen vectors by the strong shrinkage of the also dehydratingfilamen ttip ,whic henable sth edehisce danthe rt odangl eo nth e filament (ch.7). Apartfro mth erelationshi pbetwee ndehiscenc ean dth ewal lthickening si n theendothecium , mosto f the dehiscence-related mechanisms were unknowno r forgottenunti lnow .Woycick i(1924 aan db )approache dth edehiscenc eproces s inth ebes twa y withhi s description ofth e septum and stomium opening mechanisms inth elil yanther .I ngeneral ,mos to fth eliteratur eo nanthe r dehiscence isver y old and deals with light microscopy and often very ingeniousmanipulatio nexperiments .
Therol eo fth edifferen ttissue so fth estame n
Asa resul to fth efunctional-morphologica lstud yw eca npresen tth emai n functionso fth edifferen ttissues . The anther epidermis regulates the opening of the stomium and the movements ofth e loculewall s by itschangin g water volume, in aclos e collaboration with the endothecium. This water may originate from the dehydratinglocula rcavit ytoward smaturity .Th ethickenin goute rtangentia l epidermis wall and thickening and extending cuticle probably act asa solidifyingstructure .It scontinuousl yope nstomat aprobabl yenabl eth e exchangeo f gas.Par to fthi s gasma yb euse dt ofil lth elocula rcavit y afterth edisappearanc eo fth elocula rfluid ,bu tma yals ob etransporte d throughth eprominen tintercellula rspace si nth efilamen tan dth econnectiv e tissue.Besides ,th estomat aar eth emai npathwa yfo revaporatio nan dthu s canregulat eth espee do fdehiscenc ei nthi sway . Theendotheciu mchange sit samoun to fwate rsynchronousl yan dwit hth e sameeffec ta sth eepidermis ,thu ssupportin gth efunctio no fth elatter .Th e depositiono fth eU-shape dwal lthickening so nth einne rtangentia lwall so f thesecell senabl eth echangin gvolume so fth eepidermi san dendotheciu m cellst ocaus ea bendin gmovement .Th e continuation of thesethickening so n the transverse andradia l walls of the cells maintain the distancebetwee n themovin gan drigi dstructure san daccordingl ystrengthe nth ementione d effecto fth evolumina lchanges . The middle layer cells may have different functions.The y mightsuppor t tapetalfunction swhic hca nb eobserve d byth emor e or less synchronous degeneration ofthes etw o layers.Besides ,remnant so fth emiddl elaye r
140 cytoplasm arepar to fth etryphine .I nspit eo fthei r continuous radial flattening,th ecapacit yo fth emiddl elaye rcell spartl y determinesth e capacity of the locular cavity. This means that the moment of their degenerationma yinfluenc eth eamoun to fpollenkitt ,whic hi stransferre dt o the loculeb yth epressur eo fth eexpandin gpolle ngrains .I fth emiddl e layercell spersis tunti lmaturity , theiramoun to fwate r may influenceth e speedo fdehydratio nan dthu sth espee do fdehiscence . Thetapeta lfunction shav erecentl ybee nreviewe db yPacin ie tal . (1985). Beside these functions we presume a main role of this tissue in the regulationo fth ecel lwal lchange so fbot hit sow n cellsan dth edevelopin g pollengrains .Thi sals oinclude sth echang efro mhydrophyli c tohydrophobi c wallmaterial so nth eborde rbetwee nth esporophyt ean dth e gametophyte.I n ouropinio nth emai nfunctio no fth etapeta lmembrane swit hth eorbicule si s thecreatio no fsuc ha surface ,whic h behavesphysicall y similar toth e exines.Togethe rwit hth epollenkit ti tform sa non-wettabl estickin gcomple x infavou ro fa noptima ldispersa lo fpollen . The function of the sporogenous tissue isth e development of mature pollen.Thi sproces sha softe nbee nreviewe d (seech.5) .Apar tfro mthes e general lines, we propose two additional mechanisms during pollen development.Firstly ,bot hth eshap eo fth emeioti ccallos ewall san dth e presenceo fcytomicti cchannel senabl eth eformatio no fequall ysize dan d shaped microspores from each pollen mother cell, thus preventing the influenceo fthes efactor sdurin gpollination .Secondly ,th epolarit y inth e developingpolle ngrain safte rmeiosi sma ytemporar y cutof f thegenerativ e cellfro mbot hth evegetativ ecel lan dth elocula rflui dan dtherefor ema y supportth estar to fit sdeclinin gdevelopment . Therol eo fth econnectiv etissu eremaine ddifficul tt oexplain ,apar t fromit sepidermi swit ho rwithou tstomata ,th erol eo fwhic hwa sdescribe d before.Substance s from thevascula r bundle towardsth elocule smus tb e transportedthroug hthi stissue ,bu tth estorag eo freserv esubstance shardl y takesplace .It scapacit y willinfluenc eth espee do fevaporatio nan danthe r dehiscence,comparabl ewit ha possibl erol eo fth emiddl elaye r (andth e otherwate rcontainin glocul etissues )i nthi sprocess . Theepidermi so fth efilamen tshow sa developmenta lgradien tfro mth eti p toth ebase .Nea rth eti pit soute rtangentia lwal lan dcuticl echang ei nth e samewa y asth eanthe repidermis ,bu tthes e changes disappearprogressivel y down to the base.A s premature degeneration also proceeds in both the epidermisan dth eparenchym a fromth eti pt oth ebase ,thes eepiderma lwal l
141 structuresma yhav ea rigidit y function,lik ew epropose dfo rth eanthe r epidermis.Moreover ,th ethickene dcorner so fth einne rtangentia lepidermi s wallsma ypla ya simila rrole . Theparenchym a ofth efilamen t stores starchan dlipids , which suddenly disappear shortly before dehiscence.Thi s may effect an osmotic water retraction from theanthe ri nfavou ro fdehydratio nan ddehiscence .Th e prominentintercellula rspace sma yb ecovere dwit ha cuticle ,whic hsuggest s thetranspor to fgas .Thi sma yb euse dt ofil lth eincreasin gcapacit yo fth e growingflowe rbu dan dth edehydratin glocula rcavit ybefor edehiscence . Inth eti po fth efilamen tal lth etissue sdehydrat ean dshrive lbefor eo r during dehiscence, presumably topromot eth emobilit y ofth eanthe ro nth e filament infavou ro fa noptima l interrelationship withth epolle nvectors . Duringthi sproces sth etrachear yelement si nthi sregio nremai nope nan dn o waterbarrie rbetwee nth edehydratin ganthe ran dth eturgescen tfilamen tca n befound . Therol eo fth evascula rbundl eha sno tthoroughl ybee n investigatedi n thisthesis ,no rha sbee nth ecompositio no fit scontents .On eremarkabl e processi sth eappearanc e ofcallos ebetwee nsom eo fth esiev eelement s shortly beforedehiscence .
Possibleapplication sfo rplan tbreedin gpurpose s
Apartfro mthei rscientifi cvalue ,som eo fou rfinding sca nb euse dt o develop applicabletechniques . A decreaseo fth erelativ ehumidit y insideth eflowe rappeare dt ob e lethalfo rth eanthers .Moreover ,th eopenin g mechanism ofa tleas tth e stomium appearedt ob eblocke d(ch. 1an d2) . Whenw eartificiall ydecapitate d a flowerbu do fth elil ybefor eth eopenin go fth estomium , theseprocesse s alsooccurred , whereasth episti lremaine dvital .Suc ha ninductio no fmal e sterilitymigh tb ea neffectiv ewa yt opreven tself-pollinatio ni ncrop stha t aredifficul tt oemasculate . Selectionagains tth epossessio no fstomat ao nth eanther sma ylea dt o protogynlcplant s("temporar ymal esterile" )whic hca nfacilitat ehybri dsee d production.Th esam eeffec tma yb eobtaine dafte rincreasin gth erelativ e humidityi na greenhouse ,whic hkeep sth eanther sclose dfo ra longe rtime . The use of tetrads for androgenic purposes may be advantageous in comparisonwit hth egenerall yuse dfre emicrospores .Th elo wresponse s from
142 the latter cells are presumably due to their irreversible gametogenic differentiation.Youn gtetrad sma yno thav ereache dthi s irreversiblestage . Theknowledg eo fth enorma lproces so fmicrosporo -an dmicrogametogenesi si s indispensablefo ra nearl yselectio no fsuc hcell si nfavou ro fandrogeni c characteristics. This knowledge isals o necessary tous e the different applicationpossibilitie s of generative and sperm cells, like the raiseo f inbredline sfro m microsporesdurin gth efirs tmitosi s(Keijzer,1984a ;Wilm s andKeijzer,1985 ; Keijzer etal.,1986 )
Anot eo nth etechnique s
During these investigations some techniques have been developed or existingone swer eimproved . The microscopical observation of developing livingtissue sha s been improved in two different ways. Firstly, the stereo microscopical observations inside livinganther s (ch.2)offer spossibilitie sfo rresearc h ona histologica llevel .Secondly ,th edehydratio nprocesse si nth evacuu mo f theSE Mpresente dclearl yobservabl esimulation so fth enatura lprocesse s (ch.2an d7) . The clearing solution of Herr (1971)wa s used succesfully to reveal sporopollenin structures in anthers (ch.4). Thefluorescen tdye sCalcofluo rwhit eM2 Ran danilin eblu eappeare dt ob e applicableo nultrathi n(TEM- )section s(ch.6) .Th epart so fthes esection s thatar esituate do nth ebar so fa TE M gridbecam evisibl eb ythi stechnique , whenobserve dunde ra microscop ewit hinciden tU Vlight . Thesectionin go f criticalpoin t dried tissuesenable du st ous eth eSE M forth eobservatio no fth einterna lmorpholog yo ftissue s(ch. 3an d 6).I n thenea rfutur ethi stechniqu ewil lb eextende dt oth eleve lo fcell san d organelles,havin gsom eadvantage si ncompariso n with thecryo-techniqu eo f Tanakaan dMitsushim a (1984). Theintraspecifi ctransfe ro fdevelopin gpolle nt oolde rlocules ,i norde r to asynchronize the development of bordering tissues,hav e given some preliminaryresults .Als othi stechniqu ewil lb eimprove di nth efuture .
143 Summary
Inthi s thesis the development of the stamen is investigated, using structuralan dhistochemica lobservations ,micromanipulatio n and in-vitro culture. Formationan dexposur eo fpolle nar eth etw omai ngoal so fth edevelopin g stamen.Th emai nprocesse so fth epolle nformatio nwer ealread yknown .I n this thesis they are critically evaluated and some additional functionso f theobserve d structuresan dprocesse sar eproposed .Moreover ,th epolle n formationproces sappear st ohav edirec tconnection swit hth eprocesse s leadingt opolle nexposure ,includin ganthe rdehiscence .Althoug hth e latter processwa sinvestigate db ysom eresearcher saroun dth etur no fou rcentury , itwa s inth emai nunknow nunti lnow .I nthi sthesi sth eol dwork s were reviewed andcomplete d with additionalresearc ht oa descriptio no fth e entire anther dehiscenceprocess .Als o the development of the filamentha d beenhardl ydescribe dunti lno wan dappear st ohav especialize drelationship s withpolle nexposure . Inchapte r 1th echange si nth ewate rconten to fth edevelopin ganthe r tissuesar edescribed .Th emos tstrikin g changesoccu r atanthesi s whenth e dehydrationo fth elocul ewal lcause santhe rdehiscence .Thi si simmediatel y followed by the dehydration of the pollen, indicating that both these dehydrationsar edu et oevaporation .I ncas eth erelativ ehumidit yi nth e flowerbu di sartificiall y decreased indifferen tstage so fit sdevelopment , theevaporatio nabilit yo fth eanther sappear st ob ebette rshortl ybefor e anthesis than in younger stages. This indicates any preparation on dehydration inth eoldes tstage si nfavou ro fdehiscence . Inchapte r2 th eproces so fanthe rdehiscenc ei sanalysed .I tappear st o consisto f4 majo rsteps . 1.Th eenzymatica lopenin go fth etissu e (septum)betwee neac htw oadjacen t locules. 2.Th emechanica lruptur eo fth etapeta lmembrane stha tar eborderin gthes e sites. 3.Th e inward bendingo fth e loculewalls ,du et oth eexpansio no fth e epidermisan dendotheciu mcell san dth erigidit yo fth einne rtangentia l endothecium wall. This movement dissociates theepidermi s cells ofth e stomiummechanicall yan dkeep sth ethu sopene danthe rclosed ,preventin g
145 prematurelos so fpollen . 4.Th eoutwar dbendin go fth elocul ewalls ,du et oth edehydratio no fth e epidermisan dendotheciu mcell san dth erigidit yo fth einne rtangentia l endothecium wall.Thi s mechanism works contrary toth eforme rmovemen t andi scause db yevaporation ,i nmos tspecie s duet oth eentranc eo f relatively dryai ri nth eflowe rbu dafte ranthesis . Afterdehiscenc eth epolle ngrain so fmos t animalpollinate d speciesremai n stucko nth einside-ou tben tlocul ewal lb ymean so fth etapetum-derive d pollenkitt. Chapter3 describe sth esynthesi so fthi spollenkit ti nth etapetu mcells . Nextthi ssubstanc e istransferre db y capillary actiont oth e locule,afte r theexpandin gpolle n grainshav epresse d themselves into thetapetu mcells . Thisstick sth epolle ngrain st oth elocul ewall ,fro mwher ethe yca nb e pickedu p after dehiscence by apollinator .A safte rmeiosi sth ecel lwall s onth eborde rbetwee nth e loculewal lan dth epolle n grains change from hydrophylic tohydrophobi c andth epollenkit t isals ohydrophobic ,thi s sticking complex is resistant against moisture. This prevents anundesire d losso fpolle nafte rdehiscence .I fth eexpansio no fth epolle ngrain si s prevented artificially, thepollenkit t stays inside the tapetum cells, turningth eplan tint oa win dpollinator . Inchapte r4 th edegeneratio no fth emicrospore si na mal esteril especie s isrelate dt oaccompanyin gdeviation sa swel la snorma lprocesse si nth e loculewall .Th eprematur e degenerationprevent sth eexpansio no fth epolle n grainsi nthi sspecie san dals oi nthi scas eth epollenkit tremain sinsid e thetapetu mcells ,whic hagree swit hth eproce sdescribe di nth epreviou s chapter.Th etapetu m developsi na norma lwa yan dappear st ob eindependen t uponth epresenc eo fdevelopin gpolle ngrains ,a si smos to fth edevelopmen t ofbot hth eepidermi san dth eendothecium .However ,i nth eregion so fth e antherwher eth elocula rflui di ssucke dawa yi na nearl ystage ,du et oth e male sterility, dehiscence does nottak eplace , indicatingtha tth eneede d expansiono fth eepidermi san dth eendotheciu mfo rthi sproces s(ch.2 )ma yb e duet owate rretractio nfro mth elocula rcavity . Inchapte r5 a nultrastructura lanalysi so fth edevelopin glocul etissue s ispresented . Ultrastructural changes in the epidermis and the endothecium canb erelate dt oth edehiscenc eprocesse so fchapte r2 .Th echange so nth e borderbetwee nth elocul ewal lan dth epolle ngrain sfro mhydrophyli ct o hydrophobicproperties ,outline di nchapte r3 ,appea rt ob edu et otapeta l activity.
146 Themeioti ccallos ewall sequaliz eth esiz ean dshap eo fth efutur epolle n grains,excludin gan yinfluenc eo fthes efactor sdurin gpollination . Thedevelopin gpolle ngrain sar epolarize dcell sfro mthei rformatio n upon meiosis. This may result in the deviating differentiation of the generative cell, firstly by the exclusion ofplastids ,secondl y by the isolationo fthi scel lfro mbot hth evegetativ ecel lan dth elocule .Thi s isolationi sa resul to fth epresenc eo fa callos ewal lbetwee nbot hcell s and thepositio n of the generative cell inrelatio n toth e site of the colpus. Inchapte r6 a nimpressio ni sgive no fth efilamen tdevelopment .Thi s organshow sa developmenta lgradien tfro mit sti pt oit sbase .Prematur e degenerationproceed sfro mth eti pt oth ebas ean di saccompanie db yth e thickening ofbot hth eoute r tangentialepiderma l wall and the cuticle and theenlargemen to fth esurfac eo fth elatter .Thi ssupport sou ride atha t theseepiderma lchange simprov eth erigidity .Th epresenc eo fa cuticl ei n someo fth eintercellula rspace so fth efilamen tindicate sga stransport , possiblytoward sth edryin glocul etha twa sdescribe di nchapte r1 . Inchapte r7 structura ldifference si nth efilamen tti pan dth econnectiv e tissueo fthre edifferen t speciesar erelate d tothei r speedo f dehydration andanthe rdehiscence .Closin gth estomat ao fa nanthe rappear st oretar d dehiscence,thu s indicatingthei r importantrol e inthi sprocess .Thi side a issupporte db yth eslo wdehiscenc eo fa stomata-les santher .Apar tfro mthi s accelerationo fdehiscence ,th estomat ama ypla ya rol ei nth edryin go fth e locule (ch.1),whic hwa s alsorelate dt oth eintercellula r spaceso fth e filament (ch.6).I nsom especie sth efilamen tti pdehydrate san dshrivel s togetherwit hth edehydratio no fth edehiscin ganther ,whic henable sth e latter to dangle on the filament, probably in favour of an optimal collaborationwit hth epollinators .An yrol eo fthi sshrivelin gfilamen tti p inth epreventio no fwate rsuppl yt oth eanthe ri sunlikely ,a sth etrachear y elementsremai nope ni nthi szone ,whic hi sdemonstrate di nchapte r6 . Inchapte r8 thereaction so fexplante dtetrad si nolde ranther so ri n vitro indicate a selective influence of the callose onth e transfer of substancest oth emicrospores .Thi ssupport sth ealread yexistin gtheor y about the role of this wall, apart from its shaping function that we describedi nchapte r5 .Th epolarit yo fth emicrospores ,a sdescribe di n chapter 5,disappear safte rth eexplantations ,indicatin ga grea t influence ofth eenvironmen to nthi sphenomenon .
147 Samenvatting
In dit proefschrift wordt de ontwikkeling van de meeldraad onderzocht door middel van structurele en histochemische waarnemingen, micromanipulatie en in vitro cultuur. De twee hoofddoelen van de zich ontwikkelende meeldraad zijn de vorming en afgifte van stuifmeel. Van de stuifmeelontwikkeling waren de hoofdlijnen reeds bekend. In dit proef schrift worden deze kritisch bezien en wordt een aantal nieuwe functies aan de waargenomen structuren en processen toegekend. Voorts blijkt het proces van stuifmeelvorming directe verbanden te bezitten met de processen die leiden tot afgifte van het stuifmeel, inclusief het open gaan van de helmknop. Hoewel dit laatste proces reeds door enkele onderzoekers rond de laatste eeuwwisseling was onderzocht, was het toch voor het grootste deel tot op heden nog onbekend gebleven. In dit proefschrift werden deze oude publicaties samengevat en aangevuld met nieuwe gegevens, waardoor het gehele proces van het opengaan van de helmknop beschreven kon worden. Ook de ontwikkeling van de helmdraad was tot op heden nauwelijks beschreven en blijkt gespecialiseerde relaties te hebben met de afgifte van stuifmeel. In hoofdstuk 1 worden de veranderingen in het watergehalte van verschillende helmknopweefsels beschreven. De belangrijkste veranderingen vinden plaats tijdens het openen van de bloem, wanneer de uitdrogende wanden van de helmhokjes de helmknop openen. Dit wordt onmiddellijk gevolgd door het uitdrogen van de stuifmeelkorrels, hetgeen aangeeft dat beide uitdrogingsprocessen het gevolg zijn van verdamping. Wanneer de relatieve luchtvochtigheid in de bloemknop kunstmatig wordt verlaagd in verschillende ontwikkelingsstadia, blijkt de mogelijkheid tot verdamping vanuit de helmknoppen vlak voor het opengaan van de bloem beter te zijn dan in de jongere stadia. Dit geeft een of andere voorbereiding op vochtverlies in de oudere stadia aan, ten gunste van het opengaan van de helmknoppen. In hoofdstuk 2 wordt het openingsprocess van de helmknop geanalyseerd. Het blijkt uit 4 hoofdonderdelen te bestaan. 1. Het enzymatisch openen van het weefsel (septum) tussen iedere twee aangrenzende helmhokjes. 2. Het mechanisch verscheuren van de tapetummembranen die aan deze plaats grenzen.
149 3.He tnaa rbinne nbuige nva nd ewande nva nd ehelmhokjes ,te ngevolg eva n hetuitzette nva n deepidermis -e nendotheciumcelle ne n destughei dva n debinnenst e tangentiale endotheciumwand. Deze beweging scheurt de epidermiscellenva nhe tstomiu mmechanisc hva nelkaa rlo se nhoud tteven s deo pdez ewijz egeopend ehelmkno pgesloten ,waardoo rvroegtijdi gverlie s vanstuifmee lvoorkome nwordt . 4.He tnaa rbuite nbuige nva n detwe ehelmhokwanden ,tengevolg eva nd e uitdrogingva nd eepidermis -e nendotheciumcelle n en destughei d vand e binnenstetangential eendotheciumwand .Di tmechanism ewerk tprecie s tegengesteldaa nd ebewegin gui td evorig esta pe nword tveroorzaak tdoo r verdamping, ind emeest egevalle nte ngevolg eva nhe tbinnenstrome nva n relatiefdrog eluch ti nd ezic hopenend ebloemknop . Nahe topene nva nd ehelmkno p blijvend estuifmeelkorrel sva nd emeest e dierbestuivende soorten op de binnenstebuiten gekeerde helmknopwand vastgeplakt zitten door middel van de, uit het tapetum afkomstige, pollenplakstof. Hoofdstuk 3beschrijf td esynthes eva ndez eplaksto f ind etapetumcellen . Vervolgensword tdez esto fdoo rcapillair ezuigin gd ehelmhokje si ngezogen , nadatd ezwellend estuifmeelkorrel szichzel f ind etapetumcelle nhebbe n gedrukt.Di tplak the naa nd ewan dva nhe thelmhokj evast ,va nwaa rz en ahe t openenva nd ehelmkno pdoo ree nbestuive rkunne nworde nmeegenomen .Aangezie n na de reductiedeling de celwanden op de grens tussen helmhokwand en stuifmeelkorrel s vanwateropnemen d inwatera fstoten dverandere ne noo kd e pollenplakstof waterafstotendis ,i sdi tplaksystee mvochtbestendig .Di t voorkomtongewens tstuifmeelverlie snada td ehelmkno pi sopengegaan .Wannee r het zwellen van het stuifmeel kunstmatig voorkomen wordt, blijft de pollenplakstofi nd etapetumcellen ,waarbi jd eplan ti nee nwindbestuive r verandert. Inhoofdstu k4 wordtd e degeneratieva nd ejong estuifmeelkorrel si nee n mannelijksteriel esoor ti nverban dgebrach tme tzowe lafwijkend eal snormal e processeni nd ewan dva nhe thelmhokje .D evroegtijdig edegenerati evoorkom t hetzwelle nva nd estuifmeelkorrel si ndez esoor te noo ki ndi tgeva lblijf t depollenplaksto fi nd etapetumcelle nachter ,hetgee novereenkom tme the t procesui the tvorig ehoofdstuk .He ttapetu montwikkel tzic ho pnormal ewijz e enblijk tdu sonafhankelij kt ezij nva nd eaanwezighei dva nzic hontwikkelen d stuifmeel,hetgee noo kgeld tvoo rhe tgrootst edee lva nd eontwikkelin gva n deepidermi se nhe tendothecium .Allee ni nd ezone sva nd ehelmkno pwaa rd e helmknopvloeistofvroegtijdi gword tweggezogen ,te ngevolg eva nd emannelijk e
150 steriliteit, springt de helmknop niet open, hetgeen aangeeft dat de benodigde zwelling van de epidermis en het endothecium een gevolg kan zijn van waterontrekking aan de helmhokholte. In hoofdstuk 5 wordt een ultrastructurele analyse van de zich ontwikkelende helmhokweefsels gegeven. Ultrastructurele veranderingen in de epidermis en het endothecium kunnen in verband worden gebracht met het openspringen van de helmknop, zoals dat in hoofdstuk 2 werd beschreven. De veranderingen op de grens tussen de helmknopwand en de stuifmeelkorrels van wateropnemend in waterafstotend, die in hoofdstuk 3 werden behandeld, blijken het gevolg te zijn van tapetumactiviteit. Tijdens en na de reductiedeling maken de callosewanden de toekomstige stuifmeelkorrels gelijk van grootte en vorm, daarmee elke invloed van deze factoren tijdens de bestuiving uitsluitend. De zich ontwikkelende stuifmeelkorrels zijn gepolarizeerde cellen, direct vanaf hun vorming na de reductiedeling. Dit kan een oorzaak zijn van de afwijkende ontwikkeling van de generatieve eel, ten eerste door de uitsluiting van plastiden, ten tweede door de isolatie van deze eel van zowel de vegetatieve eel als de helmhokholte- Deze isolatie is een gevolg van de aanwezigheld van een callosewand tussen beide cellen en de positie van de generatieve eel ten opzichte van de plaats van de kiempore. In hoofdstuk 6 wordt een indruk gegeven van de ontwikkeling van de helmdraad. Dit orgaan vertoont een ontwikkelingsgradient van de top naar de basis. Vroegtijdige degeneratie treedt op van de top naar de basis, vergezeld door de verdikking van zowel de buitenste tangentiale epidermiswand als de cuticula en de oppervlaktevergroting van de cuticula. Dit versterkt ons idee dat deze veranderingen dienen ter vergroting van de stevigheid. De aanwezigheid van een cuticula in sommige van de intercellulaire holten van de helmdraad geeft het transport van gas aan, mogelijk naar de uitdrogende helmhokholte die in hoofdstuk 1 werd beschreven. In hoofdstuk 7 worden structurele veranderingen in de top van de helmdraad van drie verschillende plantesoorten in verband gebracht met de snelheid van waterverlies en dus het openen van de helmknop. Het sluiten van de huidmondjes van een helmknop blijkt het opengaan van deze helmknop te vertragen, waarmee hun belangrijke rol in dit process wordt aangegeven. Dit idee wordt nog versterkt door het langzame opengaan van een helmknop zonder huidmondjes. Behalve bij het versnellen van het openspringen, kunnen de huidmondjes ook een rol spelen bij het uitdrogen van de helmhokholte (hoofdstuk 1), hetgeen ook in verband werd gebracht met de intercellulaire
151 holten van de helmdraad in hoofdstuk 6. In sommige plantesoorten droogt en krimpt de top van de helmdraad samen met de openspringende helmknop, waardoor deze kan bungelen aan de helmdraad, mogelijk ten behoeve van een optimale samenwerking met de bestuivers. Een rol voor deze verschrompelende helmdraadtop in het voorkomen van watertoevoer naar de helmknop is onwaarschijnlijk, aangezien de tracheale elementen in deze zone open blijven, hetgeen in hoofdstuk 6 wordt aangetoond. In hoofdstuk 8 geven de reacties van geexplanteerde tetraden in vitro of in oudere helmknoppen en in vitro aan dat de callosewand een selectieve invloed uitoefent op het doorgeven van stoffen aan de zich ontwikkelende stuifmeelkorrels. Dit ondersteunt de reeds bestaande theorie over de rol van deze wand, afgezien van zijn vormgevende functie die we in hoofdstuk 5 beschreven. De polariteit van de cellen (hoofdstuk 5) verdwijnt na de explantaties, hetgeen een duidelijke invloed van de omgeving op dit verschijnsel aangeeft.
152 References
Anderson, W.R. (1980). Cryptic self fertilization in the Malpighiaceae. Science 207: 892-893. Baker, H.G. and I. Baker (1979). Starch in angiosperm pollen and its evolutionary significance. American Journal of Botany 66 (5): 591-600. Banerjee, U.C. (1967). Ultrastructure of tapetal membrane in grasses. Grana Palynologica 7: 365-377. Bearce, B.C. and H.C. Kohl Jr. (1970). Measuring osmotic pressure of sap within living cells by means of a visual melting point apparatus. Plant Physiology 46: 515-519. Becquerel, P. (1932). La dehiscence de l'anthere du Lis blanc. Comptes Rendus 195 (2): 165-167. Bhandari, N.N. (1984). The microsporangium. In: Embryology of Angiosperms (Ed. B.M. Johri). Springer Verlag, Berlin, New York. pp.53-121. Bonnier, G. and Leclerc du Sablon (1903). Cours de Botanique, Librairie Generale de l'Enseignement, Paris. Brodtmann, F. (1898). Uber die Funktion der mechanischen Elemente beim Farnsporangium und bei der Anthere. Dissertation, Erlangen. Burck, W. (1906). On the influence of the nectaries and other sugar containing tissues in the flower on the opening of the anthers. Recueil des Traveaux Botaniques Neerlandais 3: 163-172. Canright, J.E. (1952). The comparative morphology and relationships of the Magnoliaceae. 1. Trends of specialization in the stamens. American Journal of Botany 39: 484-497. Carniel, K. (1966). Uber die KSrnchen-Schicht in den Pollensacken von Gnetum gnemon. Osterreichisches Botanisches Zeitschrift 113: 368-374. Chatin, A. (1870a). Causes de la dehiscence des antheres. Comptes Rendus 70: 201-203. Chatin, A. (1870b). Causes de la dehiscence des antheres. Role de la deuxieme membrane. Comptes Rendus 70: 410-413. Chatin, A. (1870c). Causes de la dehiscence des antheres. Role de la troisieme membrane, de la cloison et du connectif. Experiences sur l'action de l'eau. Comptes Rendus 70: 644-648. Chaunan, S.V.S. (1979). Development of endothecium in relation to tapetal behaviour in some male sterile plants. Phytomorphology 29: 245-251.
153 Chaunan, S.V.S., R.K. Dingra and T. Kinoshita (1982). Effect of phytogametocidal compounds on the development of endothecium in some plants. Japanese Journal of Breeding 32 (2): 139-145. Cheng, P.C., Y.K. Cheng and C.S. Huang (1981). The structure of anther cuticle in cultivated rice. National Science Council Monthly 9 (11): 983- 994. Chrispeels, M.J. (1980). The endoplasmatic reticulum. In: Biochemistry of plants: a comprehenseve treatise (Eds. P.K. Stumpf and E.E. Conn), Vol1. The Plant Cell (Ed. N.E. Tolbert). Academic Press, New York, pp.390-412. Christensen, M.C. and H.T. Horner Jr. (1974). Pollen development and its special orientation during microsporogenesis in the grass Sorghum bicolor. American Journal of Botany 61: 604-623. Christensen, M.C, H.T. Horner Jr. and N.L. Lersten (1972). Pollen wall and tapetal orbicular wall development in Sorghum bicolor (Gramineae). American Journal of Botany 59 (1): 43-58. Colling, J.F. (1905). Das Bewegungsgewebe der Angiospermen Staubbeutel. Dissertation, Berlin. Coulter, J.M. and C.J. Chamberlain (1903). Morphology of Angiosperms, part 2. Appleton, New York. p.41. Cresti, M., F. Champolini, D.L. Mulcahy and G. Mulcahy (1985). Ultrastructure of Nicotiana alata pollen, its germination and early tube formation. American Journal of Botany 72: 719-727. Cresti, M. and C.J. Keijzer (1985). The structure of ER in pollen grains and pollen tubes after osmium tetroxide potassium ferricyanide staining. Journal of Submicroscopical Cytology 17 (4): 615-620. Cresti, M., C.J. Keijzer, A. Tiezzi, F. Champolini and S. Focardi (1986). The stigma of Nicotiana: ultrastructural and biochemical studies. American Journal of Botany, in press. Davis, G.L. (1966). Systematic Embryology of the Angiosperms. John Wiley, New York. p.12. Dickinson, H.G. (1975). Post meiotic stages in pollen development. Proceedings of the 12th International Botanical Congress, Leningrad. Dickinson, H.G. and L. Andrews (1977). The role of the membrane-bound cytoplasmic inclusions during gametogenesis in Lilium longiflorum Thunb. Planta 134: 229-240. Dickinson, H.G. and P. Bell (1972). The role of the tapetum in the formation of sporopollenin containing structures during microsporogenesis in Pinus banksia. Planta 107: 205-215.
154 Dickinson, H.G. and J. Heslop-Harrison (1970). The ribosome cycle, nucleoli and cytoplasmic nucleoids in the meiocytes of Lilium. Protoplasma 69: 187- 200. Dickinson, H.G. and J. Heslop-Harrison (1977). Ribosomes, membranes and organelles during meiosis in angiosperms. Phylosophical Transactions of the Royal Society of London B 277: 327-342. Eames, A.J. (1961). Morphology of the Angiosperms. MacGraw-Hill Book Company, New York, Toronto, London. Echlin, P. (1968). Development of the pollen grain wall. Berichte der Deutschen Botanischen Gesellschaft 81: 461-470. Echlin, P. (1971). The role of the tapetum during microsporogenesis of angiosperms. In: Pollen: development and physiology (Ed. J. Heslop- Harrison). Butterworths, London, pp. 41-61. Echlin, P. (1972). The ultrastructure and ontogeny of pollen in Helleborus foetidus L. 4. Pollen grain maturation. Journal of Cell Science 11: 111- 129. Echlin, P. and H. Godwin (1968). The ultrastructure and ontogeny of pollen in Helleborus foetidus. 2. Pollen grain development through the callose special wall stage. Journal of Cell Science 3: 175-186. Edwardson, J.R. (1970). Cytoplasmic male sterility. Botanical Review 36: 341- 420. Erdtman, G. (1960). The acetolysis method. A revised description. Svensk Botanisk Tidsskrift 54 (4): 561-564. Erickson, R.O. (1948). Cytological and growth correlations in the flower bud and anther of Lilium longiflorum. American Journal of Botany 35: 729-739. Fossard, R.A. de (1969). Development and histochemistry of the endothecium in the anthers of in vitro grown Chenopodium rubrum. Botanical Gazette 134 (2): 125-129. Gottschalk, W. and M.L.H. Kaul (1974). The genetic control of microsporogenesis in higher plants. The Nucleus 17:133-166. Grootaarts, H. (1978). Generatieve en vegetatieve ontwikkeling van de bloem van Gasteria verrucosa. Internal report Dept. of Plant Cytology and Morphology, Agricultural University, Wageningen, The Netherlands. Gupta, S.C. and K. Nanda (1973). Fibrous endothecium, tapetum and pollen development in Belamcanda chinensis Dc. Botanical Gazette 134 (2): 125- 129.
155 Guttenberg, H. von (1926). Bewegungsgewebe und Perzeptionsorgane. In: Handbuch der Pflanzenanatomie (Ed.K.Linsbauer) . Borntraeger, Berlin,pp . 93-100. Haberland,G . (1924).Physiologisch e Pflanzenanatomie ,6t hEdn . Engelmann, Leipzig. Hannig, E. (1910). Uberde nOffnungsmechanismu sde rAntheren .Jahrbiiche rfu r WissenschaftlichenBotani k47 :186-218 . Heinricher, E. (1885).Ube r einige imLau b dikotylerPflanze n trockenen Standortesauftretend eEinrichtungen ,welch emuthmassi gein eausreichend e Wasserversorgung desBlattmesophyll sbezwecken .Botanisch eZentralblat t 23:25-33 ,56-61 . Herr,J.M.Jr .(1971) .A ne wclearin gsquas htechniqu efo rth estud yo fovul e developmenti nangiosperms .America nJourna lo fBotan y58 :785-790 . Heslop-Harrison,J . (1968).Polle nwal ldevelopment .Scienc e 161:230-237 . Heslop-Harrison,J . (1969).A n acetolysis resistant membrane investing tapetum and sporogenous tissue inth eanther s ofcertai nCompositae . CanadianJourna lo fBotan y47 :541-542 . Heslop-Harrison,J .(1971) .Th epolle nwall :structur ean ddevelopment .In : Pollen: developmentan dphysiolog y (Ed.J . Heslop-Harrison). Butterworths, London,pp .7 5-98 . Heslop-Harrison, J. and H.G. Dickinson (1969). Time relationships of sporopolleninsynthesi sassociate dwit htapetu man dmicrospore si nLilium . Planta84 :199-214 . Heslop-Harrison, J.an dY .Heslop-Harriso n (1970).Evaluatio n of pollen viabilityb yenzymaticall y inducedfluorescence ; intracellular hydrolysis offluorescei ndiacetate .Stai nTechnolog y45 :115-120 . Heslop-Harrison,J. ,Y .Heslop-Harrison ,R.B .Kno xan dB .Howlet t (1973). Pollenwal lproteins :gametophyti can dsporophyti cfraction si nth epolle n wallso fth eMalvaceae .Annal so fBotan y37 :403-412 . Heslop-Harrison,J .an dA .MacKenzi e(1967) .Autoradiograph yo fsolubl e(2 - 14C)thymidine derivatives duringmeiosi san d microsporogenesis inLiliu m anthers.Journa lo fCel lScienc e2 :368-400 . Hesse,M .(1978) .Entwicklungsgeschicht e undUltrastruktu r desPollenkitt s beiTili a(Tiliaceae) .Plan tSystematic san dEvolutio n129 :13-30 . Hesse,M .(1979a) .Entwicklungsgeschicht e undUltrastruktu r vonPollenkit t undExin ebe inah everwandte nentomo -un danemophile nAngiospermen .Flor a 168:558-577 .
156 Hesse, M. (1979b). Entwicklungsgeschichte und Ultrastruktur von Pollenkitt und Exine bei nahe verwandten entomo- und anemophilen Angiospermen. Flora 168: 540-557. Hesse, M. (1980). Entwicklungsgeschichte un Ultrastruktur von Pollenkitt und Exine bei nahe verwandten entomo- und anemophilen Angiospermensippen der Alismataceae, Liliaceae, Juncaceae, Cyperaceae, Poaceae und Araceae. Plant Systematics and Evolution 134: 229-267. Hesse, M. (1984). Pollenkitt is lacking in Gnetaceae: Ephedra and Welwitschia; further proof for its restriction to the angiosperms. Plant Systematics and Evolution 144: 9-16. Holloway, P.J. (1982). Structure and histochemistry of plant cuticular membranes: an overview. In: The Plant Cuticle (Ed. D.F. Cutler, K.L. Alvin and C.E. Price). Academic Press, London, New York, pp.1-32. Horner, H.T. and B.L. Wagner (1980). The association of druse crystals with the developing stomium of Capsicum annuum (Solanaceae) anthers. American Journal of Botany 67 (9): 1347-1360. Hughes, J. and M.E. MacCully (1975). The use of an optical brightener in the study of plant structure. Stain Technology 50: 319-329. Ivancich, A. (1906). Der Bau der Filamente der Amentaceen. Osterreichisches Botanisches Zeitschrift 8: 385-394. Izhar, S. and R. Frankel (1971). Mechanisms of male sterility in Petunia: the relationship between pH, callase activity in anthers and the breakdown of the microsporogenesis. Theoretical and Applied Genetics 41: 104-108. Jensen, W.A. (1962). Botanical Histochemistry. Freeman & Company, San Francisco and London. Jensen, W.A., M. Ashton and L.R. Heckard (1974). Ultrastructural studies of the pollen of the subtribe Castelleinae, family Scrophulariaceae. Botanical Gazette 135: 210-218. Keijzer, C.J. (1983a). Hydration changes during anther development. In: Pollen, biology and implications for plant breeding (Eds. D.L. Mulcahy and E. Ottaviano). Elsevier Biomedical, New York. pp. 197-201. (=ch.1) Keijzer, C.J. (1983b). The mechanisms of anther dehiscence. Acta Botanica Neerlandica 32 (4): 348. Keijzer, C.J. (1984a). Voortplantingsonderzoek, pijler voor planten- veredeling. Landbouwkundig Tijdschrift 7/8: 21-26. Keijzer, C.J. (1984b). Pollen dispersal and the function of the orbicules. Acta Botanica Neerlandica 33 (2): 244.
157 Keijzer, C.J. (1985a). The functions of the endothecium. Acta Botanica Neerlandica 34 (1):130-131 . Keijzer, C.J. (1985b). Anther development of Gasteria. In: Sexual Reproductioni nSee dplants .Fern san dMosse s (Eds.M.T.M .Willems ean d J.L.va nWent) , Pudoc,Wageningen ,Th eNetherlands ,p.20 . Keijzer, C.J. (1986a). The processes of anther dehiscence and pollen dispersal. 1.Th e opening mechanism of longitudinally dehiscinganthers . TheNe w Phytologist, inpress .(=ch.2 ) Keijzer, C.J. (1986b). The processes of anther dehiscence and pollen dispersal.2 .Th emechanis m ofpollenkit t transfer and afunctio no fth e orbicules inpolle n dispersal. TheNe w Phytologist,i npress .(=ch.3 ) Keijzer, C.J. (1986c). Functional morphological relationships in the developing locule of Gasteria verrucosa. Submitted for publication. (=ch.5) Keijzer,C.J .an dM .Crest i (1986).A compariso no fanthe rtissu e development inmal e sterile Aloever aan dmal e fertileAlo e ciliaris.Annal s of Botany,i npress .(=ch.4 ) Keijzer, C.J., I.H.S. Hoek and M.T.M. Willemse (1987a). The filament developmento fGasteri averrucosa .Submitte dfo rpublication .(=ch.6 ) Keijzer,C.J. , I.H.S.Hoe k and M.T.M.Willems e (1987b).Th e processes of antherdehiscenc ean dpolle ndispersal .3 .Th edehydratio no fth efilamen t tip and the anther in some monocotylous species. Submitted for publication.(=ch.7 ) Keijzer, C.J., H.J. Wilms,A.C .va nAels t and H.B.Leferink-te n Klooster (1986).Toward sth eisolatio no fsper m cells forandrogeni cpurposes .In : Genetic Manipulation inPlan tBreeding . (Ed.W .Horn ,C.J .Jensen ,W . Odenbachan dO .Schieder) .Walte rd eGruyter ,Berlin ,Ne wYork ,i npress . Kenda,G . (1950).Ube rStomat aa nAntheren .Dissertation ,Graz . Kenda,G . (1952).Stomat aa nAntheren .Phyto n4 :83-96 . Knoll, F. (1914). ZurOkologi eun dReizphysiologi e desAndroceum svo nCistu s salvifoliusL .Jahrbiiche rfu rWissenschaftliche nBotani k54 :498-527 . Knox,R.B .(1984) .Th epolle ngrain .In :Embryolog yo fAngiosperms .(Ed .B.M . Johri)Springe rVerlag , Berlin,Ne w York.pp .197-271 . Knox, R.B.an d J. Heslop-Harrison (1970).Polle nwal lproteins : localization andenzymi cactivity .Journa lo fCel lScienc e6 :1-27 . Lammeren, A.A.M.van ,C.J .Keijzer , M.T.M.Willems e and H.Kief t (1985). Structurean dfunctio no fth emicrotubula rcytoskeleto ndurin gpolle n developmenti nGasteri averrucos a (Mill.)H .Duval .Plant a165 :1-11 .
158 Larson, D.A. and C.W. Lewis (1962). Pollen wall development in Parksonia aculeata. Grana Palynologica 3: 21-27. Laser, K.D. and N.R. Lersten (1972). Anatomy and cytology of microsporogenesis in cytoplasmic male sterile angiosperms. Botanical Review 38 (3): 245-454. Laible, C.A. (1974). Chemical methods of pollen control. Proceedings 29th Annual Corn and Sorghum Research Conference 29: 174-184. Leclerc du Sablon (1885). Recherches sur la structure et la dehiscence des antheres. Annales des Sciences Naturelles, Botanique 7 (1): 97-134. Leinfellner, W. (1956). Die Gef assbiindelversorgung des Lilium Staubblattes. Osterreichisches Botanisches Zeitschrift 103: 346-352. Linskens, H.F. (1964). Pollen Physiology. Annual Review of Plant Physiology 15: 255-270. Linskens, H.F. (1967). Pollen. Handbuch der Pflanzen Physiologie 18: 368-406. Linskens, H.F. (1974). Translocation phenomena in the Petunia flower after cross- and self-pollination. In: Fertilization in higher plants (Ed. H.F. Linskens). North-Holland Publishing Company, Amsterdam, Oxford, pp. 285- 292. Magendans, J.F.C. (1983). Anatomy of vein endings in Hedera leaves; influence of wet and dry conditions. Mededelingen Landbouwhogeschool Wageningen 83- 6. Maheshwari, P. (1950). An introduction to the Embryology of Angiosperms. MacGraw-Hill, New York, Toronto, London. Maheshwari, P., A. Rashid and A.K. Tyagi (1982). Haploids from pollen grains; retrospect and prospect. American Journal of Botany 69 (5): 865-879. Mascarenhas, J.P. (1975). The biochemistry of angiosperm pollen. Botanical Review 41 (3): 259-314. Mepham, R.H. and G.R. Lane (1968). Exine and the role of the tapetum in pollen development. Nature 219: 961-962. Mepham, R.H. and G.R. Lane (1969a). Role of the tapetum in the development of Tradescantia pollen. Nature 221: 282-284. Mepham, R.H. and G.R. Lane (1969b). Formation and development of the tapetal periplasmodium in Tradescantia bracteata. Protoplasma 68: 175-192. Nagata, T. and I. Takebe (1971). Plating of isolated tobacco mesophyl protoplasts on agar medium. Planta 99: 12-20. Namikawa, I. (1919). fiber das Offnen der Antheren bei einige Solanaceen. Botanical Magazine (Tokyo) 33: 62-69.
159 Owens,S.J .an dA.D .Westmucket t (1983).Th estructur ean ddevelopmen to fth e generative cell wall in Gibasis karwinskayana, G. venustula and Tradescantia blossfeldiana (Commelinaceae).In : Pollen, biology and implications forplan t breeding (Eds.D.L . Mulcahy andE . Ottaviano) ElsevierBiomedical ,Ne w York.pp .149-157 . Pacini, E. (1987).A n approach to harmomegathy. In: Proceedings onth e Congresshel do nth ehonou ro fB.A .Afzeliu san dH.F .Linsken s inSien a (Italy),i npress . Pacini, E.an d G.G. Franchi (1983). Pollen grain development in Smilax asperaan dpossibl efunction so fth eloculus .In :Pollen :biolog yan d implications for plant breeding (Ed.D.L .Mulcah y and E.Ottaviano) . ElsevierBiomedical ,Ne w York.pp .183-190 . Pacini, E.an d G.G. Franchi (1984). Harmomegathy, unproblem a aperto e misconosciuto.Giornal ed iBotanic aItalic a118 :271-282 . Pacini,E. ,G.G .Franch ian dM .Hess e (1985).Th etapetum : itsform ,functio n andpossibl ephylogen yi nEmbryophyta .Plan tSystematic san dEvolutio n 149: 155-185. Pacini,E .an dB.E .Junipe r (1983).Th eultrastructur eo fth eformatio nan d developmento fth eamoeboi dtapetu m inAru m italicum Miller.Protoplasm a 117: 116-129. Panchanksharappa,M.G .an dJ .Syamasenda r (1974).A cytochemica lstud yo f antherdevelopmen ti nIphigeni apallid aBak .Cytologi a39 :133-138 . Pfeffer,W . (1904).Pflanzenphysiologie ,2n ded .Wilhel mEngelmann ,Leipzig . Puri, V. (1972). Anther sacs and pollen grains: some aspects of their structure andfunction .Journa lo fPalynolog y6 :1-17 . Purkinje,J.E .(1830) .D ecelluli santheraru mfibrosi sne eno nd egranoru m pollinarumformis .Bratislava . Reynolds, E.S. (1963).Th e useo f leadcitrat ea thig hp H asa nelectro n opaquestai ni nelectro nmicroscopy .Journa lo fCel lBiolog y 17:208-212 . Reznickova, S.A. (1983). Metabolism of reserve substances inth e developing anther.In :Fertilizatio nan dembryogenesi s inovulate dplant s (Ed.O . Erdelska). SlovakAcadem y ofSciences , Bratislava,pp .57-62 . Reznickova, S.A. and H.G. Dickinson (1982). Ultrastructural aspects of storage lipid mobilization in the tapetum of Lilium hybrida var. Enchantment.Plant a 155:400-408 .
160 Reznickova, S.A. and M.T.M. Willemse (1980). Formation of pollen in the anther of Lilium. 2. The function of the surrounding tissues in the formation of pollen and pollenwall. Acta Botanica Neerlandica 29 (2/3): 141-156. Richter, S. (1929). Uber den Offnungsmechanismus der Antheren bei einigen Vertretern der Angiospermen. Planta 8: 154-184. Rowley, J.R. and J.J. Flynn (1971). Migration of lanthanium through the pollen wall. Cytobiology 3: 1-12. Sanger, J.M. and W.T. Jackson (1971). Fine structure study of pollen development in Haemantus katherinae Baker. 2. Microtubules and elongation of the generative cells. Journal of Cell Science 8: 303-315. Sangwan, R.S. and B.S. Sangwan-Norreel (1985). Early indicators of in vitro transformation from gametophyte to sporophyte in Datura. In: Sexual Reproduction in Seed plants, Ferns and Mosses (Ed. M.T.M. Willemse and J.L. van Went). Pudoc, Wageningen, The Netherlands, pp. 76-78. Schips, M. (1913). Zur Offnungsmechanik der Antheren. Beihefte zum Botanischen Zentralblatt, Abteilung 1, 31: 119-208. Schmid, R. (1976). Filament histology and anther dehiscence. Botanical Journal of the Linnean Society 73: 303-315. Schmid, R. and P.H. Alpert (1977). A test of Burck's hypothesis relating anther dehiscence to nectar secretion. The New Phytologist 78: 487-498. Schneider, J.M. (1908). Der Offnungsmechanismu s der Tulipa Anthere. Berichte der Deutschen Botanischen Gesellschaft 26: 394-398. Schneider, J.M. (1909). Zur ersten und zweiten Hauptfrage der Antherenmechanik. Berichte der Deutschen Botanischen Gesellschaft 27: 196- 201. Schneider, J.M. (1911). Uber das Offnen des Nahtgewebes der Antheren. Berichte der Deutschen Botanischen Gesellschaft 29: 406-416. Schroder, M.B. (1985). Ultrastructural studies on plastids of generative and vegetative cells in Liliaceae. 3. Plastid distribution during pollen development in Gasteria verrucosa (Mill.) H. Duval. Protoplasma 124: 123- 129. Schrodt, J. (1885). Das Farnsporangium und die Anthere. Regensburg. Schrodt, J. (1901). Zur Offnungsmechanik der Staubbeutel. Berichte der Deutschen Botanischen Gesellschaft 19: 483-488. Schwendener, S. (1899). Uber das Of fnungsmechanismus der Antheren. Sitzungberichte der Koniglichen Preussischen Akademie der Wissenschaften. Berlin.
161 Shivanna, K.R., B.M. Johri and D.C. Sastri (1979). Development and physiology of angiosperm pollen. Today and Tomorrow's Printers and Publishers, New Dehli. Shivanna, K.R. and B.M. Johri (1985). The angiosperm pollen, structure and function. Wiley Eastern Ltd., New Dehli. Southworth, D. (1971). Incorporation of radioactive precursors into developing pollen walls. In: Pollen, development and physiology (Ed. J. Heslop-Harrison). Butterworths, London, pp.115-120. Spurr, A.R. (1969). A low viscosity epoxy resin embedding medium for electron microscopy. Journal of Ultrastructure Research 71: 173-184. Staedtler, G. (1923). tiber Reduktionserscheinungen im Bau der Antherenwand von Angiospermen-Bluten. Flora 116: 85-108. Stanley, R.G. and E.G. Kirby (1973). Shedding of pollen and seeds. In: Shedding of Plant Parts (Ed. by T.T. Kozlowski). Academic Press, New York and London, pp.295-340. Steinbrinck, C. (1895). Grundziige der Offnungsmechanik von Bliitenstaub- und einigen Sporenbehaltern. Botanisch Jaarboek der Dodonaea (Gent) 7: 26-27. Steinbrinck, C. (1909). fiber den ersten Offnungsvorgang der Antheren. Berichte der Deutschen Botanischen Gesellschaft 27: 300-312. Steinbrinck, C. (1915). tiber den Nachweis von Kohaesionsfalten in geschrumpfelten Antherengeweben. Berichte der Deutschen Botanischen Gesellschaft 33: 66-72. Straka, H. (1962). Nicht durch Reize ausgeloste Bewegungen. In: Handbuch der Pflanzenphysiologie (Ed. W.Ruhland). Springer Verlag, Berlin, pp. 716-835. Straw, R.M. (1956). Adaptive morphology of the Penstemon flower. Phytomorphology 6: 112-119. Strasburger, E. (1902). Ein Beitrag zur Kenntniss von Ceratophyllum submersum. Jahrbiicher fur wissenschaftlichen Botanik 37: 477-524. Tanaka, I. and M. Ito (1980). Induction of typical cell division in isolated microspores of Lilium longiflorum and Tulipa gesneria. Plant Science Letters 17: 279-285. Tanaka, I. and M. Ito (1981). Studies on microspore development in Liliaceous plants. 3. Pollen tube development in lily pollens cultured from the uninucleate microspore stage. Plant and Cell Physiology 22 (1): 149-153. Tanaka, I, T. Taguchi and M. Ito (1980). Studies on microspore development in Liliaceous plants. 2. The behaviour of explanted microspores of the lily, Lilium longiflorum. Plant and Cell Physiology 21 (4): 667-676.
162 Tanaka,K .an dA .Mitsushim a (1984).A preparatio n method forobservin g intracellular structures by scanningelectro n microscopy.Journa lo f Microscopy 133:213-222 . Thiery,J.P . (1967).Mis ee nevidenc ede spolysaccharide s surcoupe sfine se n microscopieelectronique .Journa ld eMicroscopi e6 :987-1018 . Vasil,I.K .(1973) .Th ene wbiolog yo fpollen .Naturwissenschafte n60 :247 - 253. Venkatesh,C.S .(1957) .Th eform ,structur ean dspecia lway so fdehiscenc eo f antherso fCassia .3 .Subgenu sSenna .Phytomorpholog y 7:253-273 . Waterkeyn, L.an dA .Bienfai t (1970).O na possibl efunctio no fth ecallosi c specialwal li nIpomoe apurpure a (L)Roth .Gran a10 :13-20 . Willemse, M.T.M. (1971a). Morphological andquantitativ e changes inth e populationo fcel lorganelle sdurin gmicrosporogenesi so fPinu ssylvestri s L. 1.Morphologica l changes from zygoteneunti lprometaphase-1 .Act a BotanicaNeerlandic a 20 (3):261-274 . Willemse, M.T.M. (1971b). Morphological andquantitativ e changes inth e populationo fcel lorganelle sdurin gmicrosporogenesi so fPinu ssylvestri s L.2 .Morphologica lchange sfro mprometaphase- 1unti lth etetra dstage . ActaBotanic aNeerlandic a2 0(4) : 411-427. Willemse, M.T.M. (1972). Morphological and quantitative changes in the population of cell organelles during microsporogenesis of Gasteria verrucosa.Act aBotanic aNeerlandic a2 1(1) :17-31 . Willemse, M.T.M. (1981). Microsporogenesis in vivo and in vitro. Autofluorescenc e of pollenwal lo fLiliu man d changesi npolle nwal lo f Gasteriai nLiliu manther .Act aSocietati sBotanicoru mPolonia e 50 (1/2): 103-110. Willemse, M.T.M. (1982).Transfe r of developingbea npolle n in lilyanthers . ActaBotanic aNeerlandic a3 1 (3): 221-225. Willemse,M.T.M .(1985) .Pollenkitt ,a glue ?Act aBotanic aNeerlandic a 34 (1):132 . Willemse, M.T.M.an d M.A.W.Franssen-Verheije n (1986).Structura l changes in the style during pollination of Gasteria verrucosa. Acta Botanica Neerlandica,i npress . Willemse, M.T.M. and S.A. Reznickova (1980).Formatio n of pollen inth e anther of Lilium. 1.Developmen t of the pollen wall. Acta Botanica Neerlandica2 9 (2/3):127-140 .
163 Wilms, H.J. and C.J. Keijzer (1985). Cytology of pollen tube and embryo sac development as possible tools for in vitro plant (re)production. In: Experimental manipulation of ovule tissues (Ed. G.P. Chapman, S.H. Mantell and R.W. Daniels). Longman, New York, London, chapter 3. Woittiez, R.B. and M.T.M. Willemse (1979). Sticking of pollen on stigmas: the factors and a model. Phytomorphology 29 (1): 57-63. Wolff, G.P. (1924). Zur vergleichende Entwicklungsgeschichte und Biologischer Bedeutung der Blutennektarien. Botanische Archive 8: 305-311. Woycicki, Z. (1924a). Recherches sur la dehiscence des antheres et le role du stomium. Revue Generale de Botanie 36: 196-212. Woycicki, Z. (1924b). Recherches sur la dehiscence des antheres et le role du stomium. 2. La formation de la fente dans les antheres et les phenomenes qui l'accompagnent. Revue Generale de Botanie 36: 253-268. Wyatt, J.E. (1984). Indehiscent anther in the common bean: description, inheritance and uses. Hortscience 19: 544-545.
164 Curriculum vitae
Christiaan Jacobus Keijzer, geboren 24 maart 1950 te Voorburg, doorliep de HBS-B aan het Huygens Lyceum in zijn geboorteplaats en begon in 1970 zijn studie Plantenveredeling aan de Landbouwhogeschool te Wageningen. In 1975 behaalde hij het candidaats- en in 1979 het doctoraalexamen met als hoofdvakken: plantenveredeling, plantencytologie en -morfologie en tuinbouwplantenteelt. Na zijn afstuderen was hij enige maanden werkzaam als wetenschappelijk medewerker bij de vakgroep Tuinbouwplantenteelt. Vanaf September 1979 is hij wetenschappelijk medewerker aan de vakgroep Plantencytologie en -morfologie. In 1984 werkte hij drie maanden bij de vakgroep Botanie van de Universiteit van Siena in ItaliS o.l.v. professor M. Cresti. Naast dit proefschrift is hij auteur en co-auteur van diverse andere publicaties binnen zijn vakgebied.
165