Floral Org a No Genesis in Five Genera of the Marantaceae and in Canna

8/2/2019 Floral Org a No Genesis in Five Genera of the Marantaceae and in Canna

1/17

Floral Organogenesis in Five Genera of the Marantaceae and in Canna

(Cannaceae)

Bruce K. Kirchoff

American Journal of Botany, Volume 70, Issue 4 (Apr., 1983), 508-523.

Kirchoff, B. K.1983. Floral organogenesis in five genera of the Marantaceae

and in Canna (Cannaceae). American Journal of Botany 70: 508-523.

Made available courtesy of the Botanical Society of America:

http://www.jstor.org/stable/i341826

Your use of the JSTOR archive indicates your acceptance of JSTOR' s Terms and Conditions of Use, available at

http://www.jstor.org/about/terms.html . JSTOR' s Terms and Conditions of Use provides, in part, that unless you

have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and

you may use content in the JSToR archive only for your personal, non-commercial use.

Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or

printed page of such transmission.

American Journal of Botany is published by Botanical Society of America. Please contact the publisher for further

permissions regarding the use of this work. Publisher contact information may be obtained at

http://www.jstor.org/journals/botsam.html.

American Journal of Botany1983 Botanical Society of America

JSTOR and the JSTOR logo are trademarks of JSTOR, and are Registered in the U.S. Patent and Trademark

Office. For more information on JSTOR [email protected].

2001 JSTOR

http://www.jstor.org/Tue Oct 9 12:27:39 2001
http://libres.uncg.edu/ir/uncg/clist.aspx?id=522http://www.jstor.org/stable/i341826http://www.jstor.org/stable/i341826http://www.jstor.org/about/terms.htmlhttp://www.jstor.org/about/terms.htmlhttp://www.jstor.org/journals/botsam.html.http://www.jstor.org/journals/botsam.html.mailto:[email protected]:[email protected]:[email protected]://www.jstor.org/http://www.jstor.org/http://www.jstor.org/mailto:[email protected]://www.jstor.org/journals/botsam.html.http://www.jstor.org/about/terms.htmlhttp://www.jstor.org/stable/i341826http://libres.uncg.edu/ir/uncg/clist.aspx?id=522


2/17

Amer. J. BOt. 70(4): 508-523. 1983.

FLORAL ORGANOGENESIS IN FIVE GENERA OF THE

MARANTACEAE AND IN CANNA (CANNACEAE) 1

BRUCE K. KIRCHOFF2

Department of Botany, Louisiana State University, Baton Rouge, Louisiana 70803

ABSTRACT

The paired flowers of all species of the Marantaceae studied, except Monotagma plurispicatum,are produced through the division of an apical meristem with a tunica-corpus structure. The

solitary flowers ofM. plurispicatum develop from a similar meristem which does not bifurcate.

The paired flowers of Canna indica are produced in the axil of a florescence bract through theformation of a bract and an axillary flower on the side of the primordium which gives rise to

the largest flower of the pair. The sequence of organ initiation for both families is: calyx, corolla

and inner androecial whorl, outer androecial whorl, gynoecium. The sequence of sepal formation

is opposite in the two families. In the Cannaceae it leads directly into the spiral ereated by theformation of the other organs, while in the Marantaceae the sequence of sepal formation follows

a spiral opposite to that of the other floral organs. The members of the corolla and inner

androecial whorl separate from common primordia. In general these common primordia sepa-rate into a petal and an inner androecial member through the initiation of two growth centers,

at the same level, in the dorsal and ventral flanks of the primordium. In Ischnosiphon elegans

and Pleiostachya pruinosa the stamen is initiated at a lower position than the petal in the ventralflank of the common primordium. A similar pattern of initiation is described for the callose

staminode in Marantochloa purpurea and Canna indica. This pattern is interpreted as a variationon the more generalized pattern of inner androecial formation found in the other genera.

FLoRAL oRGANoGENESIS and floral growth infive genera of the Marantaceae, and one speciesof Canna (Cannaceae) will be explored in a

series of two papers of which this is the first.The species under study include: Calathea leo-

pardinia, Calathea lancifolia, Calathea vinosa,Ischnosiphon elegans, Marantochloa purpurea,Monotagma plurispicatum, Pleiostachya prui-

nosa (all Marantaceae), and Canna indica(Cannaceae). This paper presents a descriptionand comparison of the patterns of organogen-

esis found in these species.The Marantaceae and Cannaceae are well

circumscribed pan-tropical groups of plantssome of whose members are of economic im-

portance (Maranta arundinacea

arrowroot' Received for publication 11 February 1982; revision

accepted 2 August 1982.This paper is based on a portion of a dissertation sub-

mitted in partial fulfillment of the requirements for the

degree of Doctor of Philosophy in the Department of Bot-

any in the Graduate School of Duke University. I amdeeply indebted to Dr. R. A. White for his advice and

guidance during the completion of my degree. I am also

grateful to Mr. Russell Goddard for his help in the prep-aration of this manuscript, Sally Baker for preparing Fig.

1, 3 and 4, and to Dr. Shirley Tucker, Dr. U. Posluszny,

and an unidentified reviewer for their suggestions to im-

prove the manuscript. The SEM work included in thispaper was supported by a Duke University Graduate

School Research Award.Present address: Dept. Botany, Hebrew University of

Jerusalem, 91904 Jerusalem, Israel.

flower; Maranta spp., Calathea spp. houseplants). Genera of these families were selected

for study for several reasons: l) floral devel-opment was not well known; 2) Andersson's(1977) revision of the genus Ischnosiphon in-cludes an assessment of the relationships of theneo-tropical genera which allowed the selec-tion of a natural subset of the family for study;and 3) the position of the Marantaceae in theorder Zingiberales allowed the identification ofthe Cannaceae as its sister group (that groupwhich shares a common ancestor with the Mar-antaceae and with no other third group).

The choice of genera within the Marantaceae

was based on Andersson's (1977) revision ofIschnosiphon. Selection of species within eachof the genera was based on availability of flow-ering material. Three species of Calathea werechosen to represent different sections of thislarge genus (Schumann, 1902). Collectors, col-lection numbers and location of voucher spec-imens are given in Table 1.

MATERIALS AND METHODSFlowerIng material was collected both in Costa Rica, duringthe summer of 1978, and from plants culti-

vated in the Duke University greenhouses. Dr.G. Prance was kind enough to supply materialofMonotagma plurispicatum from Brazil.

The floral apices for epi-illumination studyand paraffin sections were first fixed in for-malin-acetic acid-alcohol (FAA: 50 ml 95%

508


3/17

April, 1983] KIRCHOFFFLORAL ORGANOGENESIS 509

0

Fig. 1. Diagrams of an imaginary species of the Mar-antaceae showing the relationships of the floral parts. P,petals; FT, floral tube; C, sepals; 0, ovary; Sc, callosestaminode; A, anther; Pa, petaloid appendage to anther;

ST, style; STI, stigma; Sh, hooded staminode; Si, outerstaminode.

Calathea lancifoliaBoom

Calathea leopardinia(Bull) Reg.

Calathea vinosaKennedy

Ischnosiphon elegansStandl.

Marantochloa purpurea(Ridl.) M. -Redh.

Monotagma plurispicatum(Koernicke) Schum.

Pleiostachya pruinosa(Reg.) Schum.

Canna indica L.

KirchOff 281 DUKE

Kress 78-1001 DUKEKirchOff 381

Kress 77-879 DUKE

Kress 78-896 DUKE

Kress 78-894 DUKE

Prance 26320 NY

Kress 78-916 DUKEKress 76-541 DUKE

TABLE 1. Species examined

Collection VoucherSpecies no. locat ion

ethanol, 5 ml glacial acetic acid, 10 ml for-malin, 35 ml H 20) and dehydrated to 95%ethanol in an ethanol series. They were thenstained in 0.5% acid fuchsin (Sattler, 1968) for24 hr and destained in 98% ethanol for l-2wk. The apices were transferred to 100% eth-anol and photographed under epi-illuminationon a Leitz Orthomat microscope equipped

with an Ultropak Illuminator and dippingcones (Posluszny, Scott and Sattler, 1980).When all the buds from a species had beenphotographed they were transferred to tertiarybutyl alcohol and embedded in "Tissue-prep,"a paraffin embedding medium. Sections werecut at 6-9 kim on an American Optical 820Microtome. Buds were occasionally left in100% ethanol for as long as 3 months with noill effects in sectioning. The sections werestained in tannic acid-ferric chloride, safraninand fast green (Berlyn and Miksche, 1976) and

mounted in "Permount." Photographs weretaken using a Leitz Orthomat microscope.

Apices for scanning electron microscopy(SEM) were fixed in glutaraldehyde and de-hydrated to 100% freon in an ethyl alcohol-freon series. Critical point drying was carriedout in a Bomar SPC-900/EX Critical PointDryer. The apices were mounted on stubs,sputter coated with gold-palladium in a Film-Vac Inc. EMS-41 Mini-Coater, and observedand photographed at 15 kv on a Philips 501Scanning Electron Microscope.

RESULTSOrganographyFloral structurein both the Marantaceae and Cannaceae is sim-ilar. Both families possess asymmetrical, her-maphroditic flowers with a perianth of twotrimerous whorls. The main differences occur

in the form of the androecial members. Ac-cording to Pai (1965) the androecium of bothfamilies is composed of two trimerous whorls.The inner whorl contains the functional anther,which is reduced to two loculi with an asso-ciated petaloid appendage (Fig. l), and twopetaloid staminodes. In the Cannaceae these

petaloid members are the labellum and theinner staminode. In the Marantaceae they arethe hooded staminode, which encloses the styleand stigma before pollination, and a callosestaminode (Fig. l) on which the tripped stylerests after pollination (Fig. 2).

The outer androecial whorl is trimerous inconstruction but is seldom represented by threemature structures (for exceptions, see Costerus,1916). The members of this whorl are alwayspetaloid and are referred to as the outer stam-inodes. For most of the species included in this

study (Canna indica, Calathea vinosa, C. leo-ardinia, C lancifolia, Monotagma plurispi-catum, Pleiostachya pruinosa, Ischnosiphonelegans) only one member of this whorl is fullydeveloped (Fig. l). Marantochloa purpurea,however, has two.

The ovary in both families is inferior andtrilocular. In the Cannaceae each locule con-tains two series of anatropous ovules while theMarantaceae possess, at most, one anatropousto campylotropous ovule per locule. For manygenera of the Marantaceae (including Ischno-

siphon, Pleiostachya, and Monotagma) ovulesare not produced in two of the three loculi,leaving only one ovule per flower.

The corolla, androecium and style of bothfamilies are fused into a floral tube of variablelength. The sepals do not contribute to this


4/17

510 AMERICAN JOURNAL OF BOTANY [Vol. 70

P3 ad

STYLE

PI

CUCULLATESTAMINODE

SMB 7 11 81OFig. 2. Diagrams showing the relationships of the eal-

lose staminode, hooded (cucullate) staminode, stigma andstyle before (diagram 1, untripped) and after (diagram 2,

tripped) pollination. Stylar depression = stamp (Ander-

sson, I 9 8 1). Figure from Kennedy ( 1 9 77).tube, becoming free directly above the ovary

(Fig. l).The shape of the stigma and style is distinct

in the two families. In the Cannaceae it is a

petaloid structure with a terminal stigma. In

the Marantaceae the distal portion of the style

is sharply bent, displacing the stigma to a lat-

eral position, facing the callose staminode (Fig.2). The flattened, or depressed, terminal por-

tion of the style, the stamp (Andersson, 1981),

receives the pollen prior to anthesis (Kennedy,

1977; Andersson, 1981). Inflorescence structure in the Marantaceae

has been described by Eichler (1875) and by

Andersson (1976) and will not be reviewed

here except when pertinent to the develop-

mental descriptions which follow. Each bract

on the florescence (an indeterminate flowering

shoot bearing lateral flowers or groups of flow-ers and forming a repeated unit of an inflo-

rescence, Weberling, 1965) subtends a sym-

podia l sys tem of axes ( the f lo rescence

component, Andersson, 1976) each bearing a

Fig. 3, 4. 3. Diagram of the paired flowers of the Mar-

antaceae. F I , floreseence axis; ab, abaxial; ad, adaxial; C I ,abaxial sepal; C2, adaxial sepal, C3, lateral sepal; PI, lateral

petal; P2, abaxial petal; P3, adaxial petal; A, anther; Pa,

petaloid appendage of anther; Sh, hooded staminode; Sc,eallose staminode; S 1, abaxial outer staminode; S2, lateral

outer staminode; S3, adaxial outer staminode; G, gynoeeial

primordium. 4. Diagram of the paired flowers of the Can-naceae. F 1 , florescenee axis; ab, abaxial; ad, adaxial; C 1 ,

lateral sepal; C2, abaxial sepal; C3, adaxial sepal; P I , adax-

ial petal; P2, abaxial petal; P3, lateral petal; A, anther; Pa,petaloid appendage of anther; L, labellum; Si, inner stam-

inode; SI , lateral outer staminode; S2, abaxial outer stam-

inode; S3, adaxial outer staminode; G, gynoecial primor-

dium.

pair of flowers (the cymule). The flowers of a

pair are mirror images of each other and are

oriented so that the functional stamens are ad-acent to each other (Fig. 3). This orientation

results in one sepal lying distal to the main

florescence axis (the abaxial sepal) one proxi-

mal (the adaxial sepal) and one midway be-

tween these two in the lateral plane (the lateral

sepal). In a similar manner abaxial, adaxial and

lateral petals can be identified (Fig. 3). Thefunctional stamen is opposite the lateral petal,

the hooded staminode opposite the abaxial

petal, and the callose staminode opposite the

adaxial petal. When a single member of the


5/17

April, 1983] K I R C H O F F - F L O R A L O R G A N O G E N E S I S 511

outer androecial whorl is present it is locatedopposite the abaxial sepal. It is, thus, referredto as the abaxial outer staminode. When twomembers of the outer whorl are present thesecond is opposite the adaxial sepal and is des-

ignated the adaxial outer staminode. The thirdposition of the outer androecial whorl is vacantat maturity. However, the primordium whichoccupies this position during early develop-ment is designated the lateral outer staminodeas it is opposite the lateral sepal (Fig. 3).

Each florescence of the Cannaceae consistsof a main axis with spirally arranged bractseach of which subtends a single pair of flowers.In Canna indica one flower of this pair gen-erally does not complete development. In Can-na the flowers of a pair are not mirror images

of one another and the floral orientation withrespect to the florescence axis is also differentfrom the Marantaceae (Fig. 4). The stamen ofthe flower which most often completes devel-opment is adjacent to the florescence axis. Theother flower is oriented so that the solitaryouter staminode is adjacent to the floral axis.It is, thus, as if this second flower were rotatedapproximately 70 counterclockwise with re-spect to the first (Fig. 4).

As in the Marantaceae it is possible to iden-

tify abaxial, adaxial, and lateral perianth mem-

bers (Fig. 4). These terms, however, do notdesignate organs which are homologous to the

organs of the same name in the Marantaceae.

They are purely descriptive terms useful only

in locating a specific organ relative to the flo-

rescence axis. The homologies of the floral or-

gans will be taken up in the discussion. Three other terms require clarification. Ven-

tral and dorsal are used to indicate a side (flank)of a specific primordium within a flower. Theventral flank refers to the inner portion of a

primordium, adjacent to the central axis of theflower, while the dorsal flank refers to the outerportion of a primordium. The term medianplane will be used to designate the longitudinalplane which passes through the center of theflorescence axis and the subtending bract of thecymule. The paired flowers of the Maranta-ceae, thus, lie on either side of the medianplane.

OrganogenesisIschnosiphon elegans: Se-quence of initiation Apex becomes truncate;

abaxial sepal; adaxial sepal; lateral sepal; ring

primordium; stamen and petal; hooded stam-inode and petal; callose staminode and petal;

abaxial outer staminode; lateral outer stami-

node; adaxial outer staminode; gynoecium. Cymule The double flowered cymule orig-

inates from a single apical meristem (Fig. 5)

with a tunica-corpus structure (Fig. 6). Thisapex enlarges and divides medianly to producetwo asymmetric floral apices, each of whichproduces a mature flower. Lateral organ pro-duction begins with cell divisions in the lateral

flank of each apex away from the other flowerof the pair (Fig. 8). Although cell divisionspredominate in this flank they are taking placethroughout the whole of each apex. They giverise to a truncate floral primordium set at anangle to the vertical axis of the cymule (Fig. 7,8). Continued circumferential growth of thisapex produces a ring primordium which givesrise to the floral cup, corolla and androecium.

CalyxThe first lateral organs to be pro-duced are the sepals. The first and second sepalsarise, in rapid succession, on the abaxial and

adaxial flanks of the apex, respectively (Fig. 9).The sepal formed in the lateral position ariseslast (Fig. 9, 10). All of the members of thiswhorl are initiated through periclinal divisionsin the first and second corpus layers (Fig. 11).Growth of the ring primordium continues dur-ing the process of sepal formation to producean asymmetric ring of tissue. Even at this earlystage of growth the region which gives rise tothe functional stamen is slightly larger than therest of the primordium (Fig. 12).

Stamen and petalThe members of the co-

rolla and inner androecium do not arise di-rectly from the floral meristem, but from por-tions of the ring primordium. The functionalstamen with its attached petaloid appendagefirst becomes distinct from its adjacent petal(Fig. 10, 13) followed by the formation of thehooded staminode and petal, and the callosestaminode and petal (Fig. 14). The stamen isformed on the ventral flank of the commonstamen-petal primordium (part of the ring pri-mordium) through periclinal and anticlinal celldivisions in this region (Fig. 15). The stamen

first becomes apparent in the region of thepetaloid appendage and then in the region ofanther formation (Fig. 10). The terminal por-tion of the common primordium gives rise tothe petal. At a later stage of stamen growth itis evident that the lower (proximal) portion ofthe common primordium does not go into theformation of the anther or petal. Rather it formsthe primordial region (similar to the floral cupof Kaplan, 1967) which becomes part of thefloral tube and ovary (Fig. 16). Growth in thisregion is intercalary.

Hooded staminode and petalThe hoodedstaminode and petal are formed through theinitiation of two growth centers in the ventraland dorsal flanks of the common primordium.Cell divisions in these regions produce a trun-cate, slightly two-humped, primordium (Fig.


6/17

Fig. 5-13. organogenesis ofIschnosiphon elegans. 5. Apical meristem (Cy) which gives rise to a double floweredcymule. X294. 6. Longitudinal section of a young cymule primordium (Cy) showing tunica-corpus structure. X588. 7.

An abaxial view of a two-flowered cymule with floral primordia (F) set at an angle to the vertical. X181. 8. Section of

one of the floral primordia (F) shown in Fig. 7. The plane of section is parallel to the plane of the photograph in Fig.

7. The arrow indicates a region of cell division that causes the apex to flatten. X339. 9. Floral apex showing order ofsepal formation. Cl, abaxial (ab) sepal; C2, adaxial (ad) sepal; C3, lateral sepal. X233. 10. Floral apex showing theadaxial (C2) and lateral (C3) sepals. The abaxial sepal is below the plane of focus. The arrow indicates the region of

petaloid appendage formation. X233. 11. Longitudinal section showing the region of sepal formation (arrow). X727.12. Abaxial side of a two-flowered cymule. Arrows indicate the regions which produce the stamen. C1, abaxial sepal;

Cy, cymule primordium. X163. 13. Floral apex with distinct petal (P) and stamen (A) primordia. Arrow indicates the

region where hooded staminode formation begins. X219.

512


7/17

April, 1983] KIRCHOFF FLORAL ORGANOGENESIS 513

Fig. 14-22. Ischnosiphon elegans. 14. Sequence of formation (P1-3) of the inner androecial whorl. Arrow showsthe region of abaxial outer staminode formation. A, anther; Pa, petaloid appendage; PI, lateral petal; P2, abaxial petal;P3, adaxial petal; Sc, callose staminode; Sh, hooded staminode. X242. 15. Longitudinal section showing the region ofcell division (arrow) on the flank of the stamen-petal common primordium. X727. 16. Longitudinal section of petal(P), anther (A) and floral tube (FT). X640. 17. Longitudinal section of the truncate hooded staminode (Sh)petal (P)primordium. X623. 18. Longitudinal section showing a later stage of hooded staminode (Sh)petal (P) separation.X588.19. Longitudinal section of eallose staminode-petal eommon primordium. Arrows indicate regions of eell division.X678. 20. Longitudinal seetion of young abaxial outer staminode primordium (S I). CI, abaxial sepal. X657. 21. Crosssection at the level of the outer androeeial whorl showing vacuolation of the adaxial outer staminode (S3). S2, lateralouter staminode. X477. 22. Cross section of the abaxial outer staminode (S1) from the same bud as Fig. 21. X865.


8/17

),

Fig. 23-31. 23-27, 29. Ischnosiphon elegans. 28. Calathea vinosa. 30-31. Pleiostachya pruinosa. 23. Longitudinalseetion showing residual meristem (arrows) in the ventral margin of the floral cup. X498. 24. Longitudinal sectionshowing formation of a gynoecial primordium. The arrow indicates a periclinal cell division. X519. 25. Longitudinal

section of a later stage in the growth of a gynoeeial primordium (G). Cell divisions are visible in the subsurface layers.

X502. 26. Cross section showing three conduplicate gynoecial primordia (G) which produce the septa. X467. 27. Top viewof three gynoecial primordia (G, one conduplieate) whieh form the stigma and style. All other floral parts have been

removed. ab, abaxial; ad, adaxial. X224. 28. Lateral view of the fusion of the three gynoecial primordia (G) which form thestigma (STI) and style. X111. 29. Longitudinal section illustrating ovule initiation. The arrow shows a region of

anticlinal cell expansion. X779. 30. Longitudinal section of the anther-petal eommon primordium. Arrow indicates the

region of antielinal expansion. X554. 31. Longitudinal section of the anther (A) and petal (P) primordia in two adjaeentflowers. Arrow indicates a recent periclinal cell division. X329.

514


9/17

April, 1983] KIRCHOFF-FLORAL ORGANOGENESIS 515

17). Continued growth of the ventral and dorsalflanks gives rise to the staminode and petalrespectively (Fig. 18). In surface view it is clearthat cell division begins in the abaxial portionof the common primordium and extends adax-ially (Fig. 13).

Callose staminode and petalThe callosestaminode and its associated petal are initiatedthrough periclinal divisions in the ventral anddorsal flanks of the common callose stami-node-petal primordium (Fig. 14, 19). Contin-ued growth of these regions first causes theprimordium to flatten and then, gives rise tothe callose staminode ventrally and the petaldorsally.

Outer androecial whorl The outer (anti-sepalous) androecial whorl is formed in a se-quence that continues the spiral of the innerwhorl: abaxial outer staminode; lateral outerstaminode; adaxial outer staminode. Due tothe shape of the floral apex at this stage thesestaminodes are formed on the ring primordiumin the positions not occupied by the inner an-droecial members. The first signs that theyhave begun growth are slight protuberances atthese positions on the ring primordium (Fig.14). In longitudinal section they are visible assmall outgrowths opposite the sepals (Fig. 20).As development continues the lateral andadaxial outer staminodes lose their meriste-matic character and stop growing. Vacuolationis first apparent in the adaxial outer staminodeand then in the lateral staminode (Fig. 21).Throughout this process the abaxial staminoderetains its meristematic character (Fig. 22). Itscontinued development produces the solitaryouter staminode present in the mature flower.

Gynoecium The gynoecium is the last flo-ral organ to be initiated on the apex. It is firstvisible as a residual meristem which appearsin the inner margin of the floral cup (Fig. 23).

Soon after this meristem is visible, anticlinalcell expansion begins in the two outer cell lay-ers. This is followed by periclinal and obliquedivisions in the subsurface layers (Fig. 24, 25).As growth continues, cell divisions extendproximally to produce the septal primordia(Fig. 26). Distal expansion of the gynoecial pri-mordia produces the stigma and style (Fig. 27,28). Of the three primordia which contributeto the stigma and style the largest has a con-duplicate structure and faces adaxially (Fig.27). The remaining two primordia are simple

mounds which fuse to the adaxial margins ofthe first soon after they arise (Fig. 27, 28). Be-cause of unequal growth of these three pri-mordia the largest, conduplicate one contrib-utes approximately two-thirds of the stylartissue, while the smaller ones contribute the

remaining one-third. In a like manner, themore rapid growth of the conduplicate pri-mordium causes it to form the distal portionof the stigma while the proximal portion isformed by the remaining two primordia (Fig.28).

The single ovule is initiated basally, througha process very similar to that which gives riseto the septal-style primordia: anticlinal cell ex-pansion in the outermost layers of the corpus,followed by periclinal cell divisions in the samearea (Fig. 29). In genera, such as Ischnosiphon,which form only one ovule per flower this ovuleis found in the locule enclosed by the larger ofthe three septal-style primordia.

The additional species of Marantaceaewhich were investigated in this study show abasically similar pattern of organogenesis tothat found in Ischnosiphon elegans. For thisreason the organogenesis of these species willnot be discussed in detail. Only the organo-genesis of those species and floral parts whichcould not easily be summarized in tabular formare discussed below. Organogenesis of all ofthe other species of the Marantaceae is pre-sented in Table 2. In both cases only the dif-ferences with Ischnosiphon elegans are pre-sented. Unless otherwise stated all floral organsshow the same pattern of organogenesis asfound inI. elegans.

Marantochloa purpurea: Cymule The sep-aration of the paired flower primordia was notobserved. Ring primordia formation is as in

Ischnosiphon elegans.Outer androecial whorl The formation of

the outer androecial whorl is identical to thatofIschnosiphon elegans. However, soon afterinitiation the adaxial staminode begins to growrapidly and is soon larger than the earlier-formed lateral staminode (Fig. 33). This dif-ference in growth reflects the fact that the adax-

ial staminode develops into a mature floralorgan while the lateral staminode ceases de-velopment soon after initiation.

Calathea vinosa: Gynoecium The patternsof initiation of the gynoecial primordia andovules are the same as in Ischnosiphon elegans.One ovule is formed in each locule in all speciesof Calathea. The inner and outer integumentsare initiated, in that order, on the sides of theovule primordia. The inner integument is ini-tiated through periclinal divisions in the pro-toderm (Fig. 35). The outer integument is ini-

tiated soon after the inner, through anticlinalcell expansion and periclinal cell division inthe two outermost layers of the ovule primor-dium. Cell division begins in the outermostlayer before it begins in the inner (Fig. 36). Theinteguments are the only parts of the flower,


10/17

Fig. 32-40. 32-33. Marantochloa purpurea. 34. Calathea leopardinia. 35-36. Calathea vinosa. 37-38. Calathea

lancifolia. 39. Monotagma plurispicatum. 40. Canna indica. 32. Longitudinal section showing the initiation of theeallose staminode (Se) in the inner flank of the common primordium. P, petal. X709. 33. Floral primordium showing

the initiation and development of the adaxial outer staminode (S3). A, anther; Pa, petaloid appendage; S2, lateral outerstaminode. X191. 34. Longitudinal seetion showing separation of the hooded staminode (Sh) petal (P) primordium.

X541. 35. Longitudinal section showing initiation of the inner integument. The arrow indicates a periclinal divisionin the protoderm. X588. 36. Longitudinal section showing initiation of the outer integument (0). The arrow indicatesa region of cell division in the protoderm. I, inner integument. X571. 37. SEM showing displacement of the lateral

petal (P1) and relative sizes of the anther (A) and petaloid appendage (Pa). P2, abaxial petal. X164. 38. Longitudinal

seetion showing separation of the callose staminode (Sc)petal (P) primordium. X657. 39. Abaxial view of floreseeneeeomponent with two floral primordia (F) at different stages of growth. The abaxial sepal (C1) of the upper flower has

begun to reflex over the lower flower. X130. 40. Top view of florescence primordium showing initiation of flower pairs

(F1, F2) in the axils of the floreseenee bracts (FB). B, bract; FA, florescenee apex. X106. 516


11/17


TABLE 2. Summary of floral development in six species of Marantaceaea

Hooded staminode Callose staminodeSpecies Stamen and petal and petal and peta l

Pleiostachya pruinosa

Marantochloa purpurea

Calathea leopardinia

Calathea vinosa

Calathea lancifolia

Monotagma plurispicatum

As inI. elegans As inI. elegans

As inI. elegans Sc initiated in theventral flank Ofthe CP thrOughpericlinal divi-siOns in the out-er cOrpus (Fig.32).

As in I. elegans. Agreater amount

of growth in theventral flank ofthe CP producesa truncate pri-mordium whichis larger ventral-ly (Fig. 34).

As inI. elegans

As in I. elegans.only epi-illumi-nation observa-tions were possi-ble.

As in I. elegans.Growth in theventral flank ofthe CP exeeedsgrowth in thedorsal displaeingthe P to an al-most lateral po-sition on the Se(Fig. 38).

As inI. elegans. As inI. elegans.At an early stage only epi-illumi-of growth the Sh nation observa-appears larger tions were possi-than the P. ble.

As inI. e legans

As in I. elegans.Growth in theventral flank ofthe CP exeeedsgrowth in thedorsal whiehcauses the P toappear in a lat-eral position onthe Sh.

S initiated in ventral flank of

CP through antielinal cellexpansion (Fig. 30) and peri-

elinal cell division (Fig. 31)

in the outermost eorpus lay-er.

Two growth centers initiate in

the ventral (stamen) anddorsal (petal) flanks of CP.

The Pa arises before and

separately from the A toyield separate, subequal, pri-

mordia on the ventral flank

of the CP.As in Marantochloa purpurea.

only one stamen primor-

dium is produced and givesrise to both the A and Pa.

Widening of the CP begins

in the region of the A andextends abaxially to the re-

gion of the Pa. The largerpart of the CP goes to form

the S.As in Marantochloa purpurea.

It has proven impossible to

determine which of the two

stamen primordia is formedfirst.

Two growth centers are initiat-ed in the ventral (stamen)and dorsal (petal) flanks of

the CP. The A appears be-

fore and separately from thePa. Division of the CP is un-

equal, the major portion

going to form the S. Rapidgrowth of the S displaees the

P to a lateral position earlyin development (Fig. 37).

As in Marantochloa purpurea.only epi-illumination obser-vations were possible. It has

proven impossible to deter-mine which of the two sta-men primordia is formed

first although the A appears

larger than the Pa at an early

a When similarities exist between the development of an organ in two species the name of the similar species is givenfirst followed by a description of the differences found in the two species. CP, common primordium; P, petal; S, stamen;

A, anther; Pa, petaloid appendage; Sh, hooded staminode; Sc, callose staminode.

in the Marantaceae, whose initiation involvescell divisions in the outermost cell layef.

Monot agma pl urispicatum: Cymule

Acharacteristic feature of the genus Monotagmais the possession of one-flowered cymules. Thestages of cymule initiation reflect this aspectof mature structure in that the cymule pri-mordium does not divide, but gives rise to a

solitary floral apex (Fig. 39). The initial growthof this apex, which serves to flatten it and be-

gins the formation of the ring primordium,takes place in the abaxial margin, away fromthe florescence axis. In Ischnosiphon elegansgrowth of the floral apices begins in the lateralmargins, away from the other flower of the pair.

CalyxIn sequence and region of sepal ini-


12/17


Fig. 41-47. Canna indica. 41. Longitudinal section of young floral apex. Arrows indicate the regions of cell divisionin the medial flanks. X276. 42. SEM of young floral primordium showing initiation of the lateral (C1) and abaxial (C2)

sepals, and of the ring primordium (RP). The florescence axis is the upper left in this picture. X256. 43. Flower

primordium showing showing double stamen primordium and inner staminode (Si) formation. A, anther; L, labellum;Pa, petaloid appendage; P3, lateral petal. X200. 44. Formation of the outer androecial whorl on the ring primordium.

A, anther; L, labellum; S 1 , lateral outer staminode; S2, abaxial outer staminode; S3, adaxial outer staminode. X131.

45. View of the lateral side of the floral primordium after gynoecial initiation. Sepals and two petals have been removed.Note that the lateral petal is much smaller than the inner staminode. A, anther; P3, lateral petal; S2, abaxial outer

staminode; S3, adaxial outer staminode. X128. 46. Cross section showing the fusion of the conduplicate gynoecial

primordia (one labeled, G), proximal to the site of initiation, to produce the septa. X141. 47. Top view of threegynoecial primordia (G, one conduplicate) which fuse to form the stigma and style. Arrow indicates the region of fusion.

X108.

tiation Monotagma plurispicatum shows thesame pattern as Ischnosiphon elegans. How-ever, soon after the abaxial sepal is initiated itbegins to reflex so that it partially covers thenext flower of the sympodium (Fig. 39). Noother species studied shows this pattern ofabaxial sepal development.

Canna indica: Because of the difference inthe orientation of the flower with respect to

the florescence axis in the Cannaceae and Mar-antaceae the terms used below do not refer tothe same structures as do the correspondingterms in the Marantaceae. A comparison of theinitiation patterns in these two families is in

cluded in the Discussion section, which im-mediately follows this description.

Sequence of initiationApex becomes trun-cate; lateral sepal; abaxial sepal; adaxial sepal;ring primordium; stamen and petal; labellumand petal; inner staminode and petal; lateralouter staminode; abaxial outer staminode;adaxial outer staminode; gynoecium.

Origin of the flower pairEach bract on the

florescence axis produces, in its axil, a laterallyelongated primordium which gives rise to apair of flowers. The larger of these flowers isproduced directly from this primordium, whilethe smaller arises in the axil of a bract produced


13/17

April, 1983] KIR C HOF F - F LOR AL OR GANOGENES IS 519

on the side of the primordium (Fig. 40). Theformer gives rise to a mature flower while thelatter often ceases development while it is stillquite small. It is, however, occasionally pos-sible to find the two flowers of a pair fullydeveloped. Unlike the Marantaceae, the pairedflowers in the Cannaceae are not mirror imagesof one another. In the Marantaceae the stamenforms on the lateral portion of the floral apex,adjacent to the other flower of the pair. Thisresults in a medianly symmetrical pair of flow-ers. In Canna indica the stamen of the largerprimordium is produced perpendicular to themedial line of the flower, adjacent to the flo-rescence axis (Fig. 4), while the stamen of thesmaller flower arises at approximately a 20angle to the medial line (Fig. 4). Thus, if thelarger flower were rotated through approxi-mately 70 (counterclockwise) it would be inthe same relationship with respect to the flo-rescence axis as is the smaller flower. To avoidconfusion the terms used below all pertain tothe development of the larger flower primor-dium.

A truncate floral apex is produced throughcell divisions in the medial flanks of the apex(Fig. 41). Continued cell divisions around theperiphery of the apex produce the ring pri-mordium (Fig. 42). Even at this early stage ofgrowth the adaxial portion of this primordium,which gives rise to the stamen, is larger thanthe rest of the ring primordium.

CalyxSepal initiation begins while theabove processes are taking place. The sepalsare initiated by periclinal and anticlinal divi-sions in the corpus on the side of the floralapex, and slightly below the ring primordium.The lateral sepal appears first, followed by theabaxial and adaxial sepals in that order (Fig.42).

Stamen and petalThe common stamen-

petal primordium (part of the ring primor-dium) widens and separates into the stamenand petal due the formation of two growthcenters in its ventral and dorsal flanks. Thestamen primordium is first apparent in the re-gion which will produce the petaloid append-age and, then, as a separate primordium, inthe region of the anther. )In this way, a doublestamen primordium aries (Fig. 43). One pri-mordium produces the petaloid appendagewhile the other gives rise to the anther. Thedorsal portion of the common primordium

produces the petal (Fig. 43).Labellum and petal Cell divisions in theventral and dorsal flanks of the common pri-mordium give rise to two growth centers thatproduce the labellum ventrally and the petaldorsally. The labellum appears first on the in

ner flank of the common primordium in theregion adjacent to the first formed sepal andextends laterally to include the rest of the pri-mordium. There is some variation in the se-quence of inner androecial initiation describedto this point. An apex was found in whichseparation of the labellum-petal primordiumbegins before the anther primordium appears.However, even in this case stamen initiationbegins, with the formation of the petaloid ap-pendage, before the formation of the labellum.

Inner staminode and petalThe followingdescription is based only on material examinedby epi-illumination light microscopy: the in-ner staminode is produced on the ventral flankof the common primordium while the dorsalflank and part of the apex of the common pri-mordium produce the petal (Fig. 43). Soon af-ter the inner staminode is produced it appearslarger than the petal and remains largerthroughout organogenesis (Fig. 45).

Outer androecial whorlThe order of for-mation of the outer androecial whorl continuesthe spiral established by the other floral parts:lateral outer staminode; abaxial outer stami-node; adaxial outer staminode. Due to theshape of the floral apex at this time, the outerstaminodes arise on the ring primordium inthe positions not occupied by the inner an-droecial members. The first signs of these outerstaminodes are small protuberances on the ringprimordium (Fig. 44). In Canna indica onlythe lateral outer staminode normally developsinto a mature floral organ. However, severalflowers with an adaxial outer staminode havebeen observed.

Gynoecium The gynoecium is first visibleas a residual meristem which appears in theinner margin of the floral cup after all otherfloral organs have been initiated. The primor-dia which give rise to the septa and the stigma-

style appear on the distal margin of the floralcup, opposite the outer androecial members.They arise through a process of anticlinal cellexpansion in the tunica and first one or twocorpus layers, followed by cell divisions in thecorpus.

Proximal to the site of initiation the gy-noecial primordia become conduplicate andproduce the septa through marginal fusion ofadjacent primordia (Fig. 46). Distal to the ini-tiation site only one of the three primordiaremains conduplicate while the others appear

as undivided outgrowths. The two smaller pri-mordia fuse to the margins of the larger, con-duplicate, one (Fig. 47) to produce a single largeprimordium which goes on to form the stigmaand style.

The ovules are initiated axially in each locule


14/17


through anticlinal cell expansion in the tunica,

and periclinal divisions in the first layer of the

corpus.

DISCUSSIONHomologies of theflower in theMarantaceae and CannaceaeThe mainproblem in determining the homologies of thefloral parts is the interpretation of the androe-cium. The most widely accepted interpretationof flower structure in the Marantaceae, and theone supported by this study, is that the innerandroecial whorl consists of the fertile stamenwith its attached petaloid appendage, the hood-ed staminode, and the callose staminode. Theouter whorl is then composed of one or twopetaloid staminodes depending on the genus(Eichler, 1883; Holttum, 1951; Pai, 1965; Ti-lak and Pai, 1966, 1968, 1970). The existenceof these two whorls is clearly demonstrated inthe developmental sequences described here.If we ignore the perianth for the moment, thesequence of initiation is: fertile stamen; hoodedstaminode; callose staminode; abaxial outerstaminode; lateral outer staminode; adaxialouter staminode. Although the inner whorl ap-pears before the outer, this poses no problemto the above interpretation as the members ofthe inner whorl are, both in position and de-velopment, anti-petalous, as is common for theinner androecium.

The interpretation of the androecium of theCannaceae has sparked much more debatethan has that of the Marantaceae. Payer (1857),on the basis of development, interpreted theandroecium of Canna as consisting of a singleanti-petalous whorl. The stages of develop-ment described by Payer are similar to thosedescribed here, with two exceptions. First, hestates that the petals are initiated simulta-neously and are formed independently of theinner androecium. The developmental pat-

terns described here indicate that the petals areinitiated sequentially, not simultaneously,their initiation being tied to that of the innerandroecium. Second, Payer derives all thestaminodes in the mature flower from divi-sions of the primordia formed opposite thepetals. For instance, he derives the lateral outerstaminode (Fig. 44) fromp. division of the la-bellum primordium and the adaxial outerstaminode from a division of the inner stam-inode primordium. He does not mention a pri-mordium forming in the abaxial position of

the outer whorl between the labellum and innerstaminode (Fig. 44). From a consideration ofthe developmental patterns presented here itis clear that Payer has simply mistaken theinitiation of the outer androecial whorl for div-isions in the inner. In reality two whorls areformed.

Eichler (1873) postulates that the labellumand functional stamen, along with two or threestaminodes, united at the base with the stamen,represent the inner whorl. In this view the outerandroecium is also completely lacking. Thisinterpretation is derived from a study, and mis-interpretation, of floral development.

After the formation of the stamen and la-bellum Eichler describes the initiation of twoprimordia to the right and left of the stamenprimordium. Although he considers both to bemembers of the inner whorl only the oneformed on the left actually is. This is the pri-mordium of the inner staminode (Fig. 43). Theright primordium is a member of the outerwhorl: the lateral outer staminode (Fig. 44).Eichler, thus, misidentifies an outer androecial

member as an inner one. At this point in de-velopment two members of the outer androe-cial whorl remain uninitiated: the abaxial andadaxial outer staminodes, of which only theadaxial exists in the mature flower (both ofthese staminodes are lacking in Canna indica).The abaxial outer staminode is formed on thering primordium between the labellum and in-ner staminode, while the adaxial outer stam-inode is formed between the inner staminodeand the stamen (Fig. 44). Eichler makes twoerrors in observing the production of these or-

gans. First, at the stage after all the outer stam-inodes are formed he mistakes the adaxialstaminode primordium for the primordiumwhich he previously described as being initi-ated on the left of the stamen (Fig. 45). Underthis interpretation, the adaxial outer stami-node does not exist as a member of the outerwhorl. It is, rather, identified with a previouslyformed member of the inner whorl. Second,he identifies the primordia of the inner stam-inode and the outer abaxial staminode as twolobes of a single primordium that he says de-

velops from the primordium formed on theleft side of the stamen (Fig. 45). He claims thatthis double primordium gives rise to one ofthe petaloid floral parts which is, in the ter-minology used here, the inner staminode. Byidentifying two primordia as lobes of a singleprimordium he eliminates another member ofthe outer androecial whorl. The result of thesemisidentifications is that the outer androecialwhorl is perceived as missing, while all of thepetaloid floral members are identified as mem-bers of the inner whorl.

In his Bliitendiagramme, Eichler (1875) re-peats the above interpretation and puts for-ward a second hypothesis on the structure ofthe flower in the Cannaceae. Under the latterinterpretation, the stamen, labellum and oneof the staminodes (the inner staminode) rep-resent the inner whorl while the remaining one


15/17


or two staminodes represent the outer whorl.This hypothesis receives support from Rao andDonde (1955) and Pai (1963), and is supportedby the developmental work presented here.

Costerus (1916, 1917) also supports thisview but with some major modifications. Hesuggests that the anther and petaloid append-age are members of different whorls, and citesthe origin of the stamen from a double pri-mordium (Payer, 1857; Eichler, 1873) to sup-port this view. Under this interpretation theouter androecium consists of two petaloidstaminodes one of which (the adaxial outerstaminode) is produced from the same pri-mordium which gives rise to the anther. Theanther is thus seen as a member of the outerwhorl. The third member of this whorl is gen-erally missing as a mature structure, but is pres-ent in certain teratological forms. The innerwhorl consists of the petaloid appendage, la-bellum and one petaloid staminode (the innerstaminode). The connection of the anther withthe petaloid appendage is seen as a completelyarbitrary phenomenon which obscures its truerelationship as a member of the outer whorl.An analogous interpretation is put forward(Costerus, 1917) for the androecium of theMarantaceae.

The evidence used to support the above hy-

pothesis is drawn primarily from teratology.In the absence of the material examined byCosterus it is difficult to refute this theory.However, it receives no support from the de-velopmental point of view despite Costerus'claims to the contrary. The mere fact that thestamen arises as a double primordium doesnot prove that it is composed of two membersfrom different whorls. Its development is, rath-er, a reflection of the asymmetric develop-mental pattern found in the whole flower. Thisasymmetry of development is also evident in

the formation of the labellum which beginsseparation from its associated petal in the re-gion of the common primordium closest to thelateral sepal.

With the acceptance of the floral interpre-tations of the Marantaceae and Cannaceae dis-cussed above the task of establishing the ho-mologies of the floral parts pecomes an easyone. When two floral apices, one from eachfamily (Fig. 33, 43), are compared, it is clearthat there is an exact correspondence in num-ber and relative position of primordia (Table

3). The common construction plan is: calyx,corolla, outer androecium, inner androecium,gynoecium.

Comparison of organogenesisFloral devel-opment in the Marantaceae and Cannaceae is

not well known. Most previous studies of these

TABLE 3. Floral homologies in the Cannaceae and Mar-

antaceae

Floral member in Marantaceae Floral member in Cannaceae

Abaxial sepalb Lateral sepalLateral sepalb Abaxial sepa l

Adaxial sepalb Adaxial sepa lLateral petal Adaxial petalAdaxial petal Latera l petalAbaxial petal Abaxial petalAbaxial staminode Lateral staminodeLateral staminode Abaxial staminodeAdaxial staminode Adaxial staminodeStamen StamenHooded staminode LabellumCallose staminode Inner staminodeConduplicate primordium Conduplicate primordium

Remaining two Remaining twogynoecial primordiac gynoecial primordia

The difference in terminology used to identify ho-mologous organs is due to the difference in floral orien-

tation with respect to the florescence axis.b The homologies of the sepals are based on their po-

sition in the flower not the sequence of initiation which

is different for these two families.There are no convenient terms to distinguish these

primordia and no necessity to do so since they soon lose

their distinctness through fusion to the conduplicate pri-

mordium.

families have concentrated on mature floralmorphology (Eichler, 1875; Schumann, 1902;Kranzlin, 1912; Holttum, 1951; Pai, 1965),teratology (Costerus, 1916, 1917) and anatomy(Rao and Donde, 1955; Pai, 1963; Tilak andPai, 1966, 1968, 1970). Payer's (1857) mon-umental work on the organogenesis of flowerscontains a description of the development ofCanna indica as does Thompson's (1933)study of flower and inflorescence structure inthe Zingiberales. Eichler (1873) examined the

development of several species of Canna in-cluding Canna indica. Thompson (1933) andEichler (1883) are the only authors to discussfloral development in the Marantaceae andthey present strikingly different pictures of or-ganogenesis.

Eichler (1883) presents a description of de-velopment in Stromanthe sanguinea, Thaliadealbata and illustrates the development ofMaranta bicolor. His descriptions are similarto those presented here.

Thompson (1933) describes the develop-

mental patterns for many members of the Zin-giberales including Canna indica, Clinogynevirgata (Marantaceae), and Thalia geniculata(Marantaceae). However, since his descrip-tions are at variance with all previous and sub-sequent work and because he illustrates thesedescriptions only with diagrams, his work will


16/17

522 AM ER IC AN J OUR NAL OF B OTANY [Vo l . 70

not be discussed in detail. In essence, he de-scribes the initiation and fusion of 16 separateprimordia which give rise to the various floralorgans. For instance, the labellum of the Can-naceae is described as arising from a fusion of

primordia numbers 8, 13 and 16, all membersof different whorls. Nothing in the other lit-erature or in the present investigation suggeststhat this is so. His descriptions of gynoecialdevelopment, however, are in agreement withthose presented here.

All of the species examined in this study,except Canna indica show identical sequencesof formation of floral parts. Once the differencein orientation of the flowers in the Marantaceaeand Cannaceae is taken into account the onlydifference is in sepal formation. If we identify

the sequence of formation as Sepal l, Sepal 3,Sepal 2 in the Marantaceae the sequence in C.indica is Sepal l, Sepal 2, Sepal 3. This lattersequence leads directly into the spiral createdby the formation of the corolla and androeciumin Canna. Thus, Canna shows one continuousspiral of organ formation while in the Mar-antaceae the spiral of the corolla and androe-cium is opposite that of the calyx.

This discrepancy in sepal formation betweenthe Marantaceae and Cannaceae can be ex-plained by hypothesizing that the shape of the

floral apex during calyx formation (Fig. 9, 10)obscures the initial appearance of the lateralsepal in the Marantaceae. If this sepal wereinitiated before the adaxial sepal, then the ini-tiation sequence in the two families would beidentical. However, it has proven impossibleto confirm this hypothesis through the obser-vation of sectioned apices. Thus, it cannot beadopted without reservation.

The only variability in the patterns of or-ganogenesis in the eight species investigatedin this study is found in the formation of the

inner androecium. The main differences dealwith the amount and position of growth thatoccurs in the flanks of the common primordiaafter the initiation of the androecial and petalgrowth centers. Ischnosiphon elegans, Pleio-stachya pruinosa, Marantochloa purpurea, andCanna indica are the only species which showa significant difference in the initiation of atleast one androecial member. In I. elegans andP. pruinosa the stamen is initiated in the ven-tral flank of the common primordium belowthe initiation site of the petal. M. purpurea and

Canna indica show a similar pattern of for-mation of the callose staminode and petal: thestaminode is initiated in the flank of the com-mon primordium, below the petal.

Perhaps the most surprising observationconcerning organogenesis is that the formation

of the inner androecium is not uniform withinthe genus Calathea. In many cases, the patternof formation of a specific member of the innerandroecium is more similar between a speciesof Calathea and a species of some other genusthan it is among the species of Calathea. Forexample in the separation of the hooded stam-inode-peta l primordium a truncate , twohumped, primordium is formed. In Calatheavinosa, Ischnosiphon, Pleiostachya, Maranto-chloa, and Canna these outgrowths are of ap-proximately the same size, while in Calathealeopardinia, Calathea lancifolia, and Mono-tagma the inner outgrowth is distinctly larger.This diversity of developmental patterns inCalathea reflects the variability of floral formsin this large genus (approximately 150 species).

The similarities among the developmental pat-terns of the inner androecium, of species ofCalathea and species of other genera are dueto parallelisms.

Based on these results, it is possible to con-clude that the amount of variability in organ-ogenesis among closely related genera may beapproximately the same as that found amongdistantly related species of a large genus. Thereis no way to predict how similar the patternsof organogenesis will be simply by knowingthat the species which possess them belong to

the same (or different) genera. LITER ATUR E C ITED

ANDERSSoN, L. 1976. The synflorescense of the Mar-antaceae. organization and descriptive terminology.

Bot. Not. 129: 39-48.______ . 1977. The genus Ischnosiphon (Marantaceae).

opera Bot. 43: 1-113.______ . 1981 . The n eotropical genera of Marantaceae.

Circumscription and relationships. Nord. J. Bot. 1:

218-245.BERLYN, G. P., AND J. P. MIKSCHE. 1976. Botanical Mi-

crotechnique and Cytochemistry. Iowa State Univer-

sity Press, Ames.CoSTERUS, J. C. 1916. A fresh investigation into th estructure of the flower ofCanna. Ann. Jard. Bot. Bui-

tenzorg. 29, Deuxime Ser., Vol. 14: 165-184.

______. 1917. Die Uebereninstimmung und der Unter-schied in dem Bau der Blumen von Canna und der-

jenigen der Marantaceen. Ann. Jard. Bot. Buitenzorg.30, Deuxime Ser., Vol. 15: 59-93.

EICHLER, A. W. 1873. Ueber den Blthenbau von Canna.Bot. Zeitung 31: 177-189,193-198,208-218,225

232,241-247.______. 1875. Blthendiagramme construirt und erla-

tert. Vol. 1. W. Engelmann, Leipzig.______ . 1883. Beitrge zur Morphologie und Systematic

der Marantaceen. Akad. Wiss. Berlin, Physik. Math.Kl., Ab. I: 1-99.

Hourrum, R. E. 1951. The Marantaceae of Malaya. Gard.Bull Sing. 13: 254-296.

KAPLAN, D. R. 1967. Floral morphology, organogenesis

and interpretation of the inferior ovary in Downingia

bacigalupii. Amer. J. Bot. 54: 1274-1290.


17/17

Apri l, 1983] KIRCHOFFFLORAL ORGANOGENESIS 523

KENNEDY, H. 1977. Systematics and pollination of the"closed-flowered" species of Calathea (Marantaceae).Univ. Calif. Publ. Bot. 71: 1-90.

KRANZLIN, F. 1912. Cannaceae.In A. Engler [ed.], DasPflanzenreich IV 47: 1-77.

PAI, R. M. 1963. The floral anatomy ofCanna indica L.Bull. Bot. Soc. Coll. Sci., Nagpur 4(2): 45-53.

. 1965. Morphology of the flower in the Canna-ceae. J. Biol. Sci. 8: 4-8.

PAYER, J. B. 1857. Trait& d'organognie compare de lafleur. Texte et Atlas. Librairie de Victor Masson, Paris.

PoSLUSZNY, U., M. G. SCoTT, AND R. SATTLER. 1980.Revisions in the technique of epi-illumination lightmicroscopy for the study of floral and vegetativeapices. Can. J. Bot. 58: 2491-2495.

RAo, V. S., AND N. DoNDE. 1955. The floral anatomy ofCanna flaccida. J. Univ. Bombay 24(3): 1-10.

SATTLER, R. 1968. A t echnique for the study of floraldevelopment. Can. J. Bot. 46: 720-722.

SCHUMANN, K. 1902. Marantaceae. In A. Engler [ed.],Das Pflanzenreich IV 48: 1-176.

THoMPSoN, J. M. 1933. Studies in advancing sterility.Pt. VI. The theory of Scitaminean flowering. Publ.Hartley Bot. Lab. 11: 1-111.

TILAK, V. D., AND R. M. PAI. 1966. Studies in the floralmorphology of the Marantaceae I. Vascular anatomyof the flower of Schumannianthus virgatus Rolfe, withspecial reference to the labellum. Can. J. Bot. 44:

1365-1370., A N D . 1 9 6 8 . S t u d i es i n t h e f l o r a l m o r - phology of the Marantaceae. II. Vascular anatomy ofthe flower in two species of the genus Phrynium Willd.Proc. Indian Acad. Sci. Sec. B. 68(5): 240-249.

, A N D . 1 9 7 0 . S t u d i e s i n th e f l o r a l m o r -phology of the Marantaceae III. Vascular anatomy ofthe flower in some species of the genus Calathea.Marathwada Univ. J. Sci. 9 Sci. 2: 31-41.

WEBERLING, F. 1965. Typology of inflorescences. Bot.J. Linn. Soc. London 59: 215-221.

Documents

Floral Org a No Genesis in Five Genera of the Marantaceae and in Canna