BLUMEA 24 (1978) 123—127

Morphology of the pistil in Malvaceae-Ureneae

W.A. van Heel

Rijksherbarium, Leiden, Netherlands

Summary

The puzzling fact that the pistils of Malvaceae-Ureneae have five locules, but ten, instead of five, stylar

branches, is made clear by a study of their development.

Introduction

Those Malvaceae that have stylar branches twice the number of carpids are grouped

in the together by Schumann (1893) Ureneae. Normally, there are five locules in the

and there is which divides into instead of ovary, only one style, ten, five, stigmatic all of branches. According to Schumann, in Pavonia species ten carpids develop, which five retarded in their assumed that are development. Moreover, Schumann the changing of the position carpids (antesepalous in some species, antepetalous in others) depends on the the found whether antepetalous or antesepalous carpids are retarded. However, I that the location of the Malvaceae the (1966) carpel primordia in depends on degree of of the the of of the contortion petal-androecium system at time origin carpel primordia. Probably, Schumann took his knowledge of the development of the carpids from

Payer (1857). Payer described and figured the primitive composition of the Pavonia

pistil. According to him ten protuberances arise and soon become elevated on a common Whereas the base five of and membrane. at of them, on their inside, a locule is formed, the

other five remain sterile. Payer convincingly dismissed the assertion ofDuchartre (1845)

that five carpels have two styles each by ' dédoublement’.

to Saunders the formation of According (1936) ten separate styles is proof beyond of her of this the question theory carpel polymorphism. According to theory syncarpous of five 'fertile and of the classical and pistil consists carpels' (the septs placentae theory) five 'sterile carpels' (the locules and their walls of the classical theory). Whereas normally the 'sterile have the of 'fertile only carpels' styles, in case the Ureneae also the carpels' have. the of based the However, theory carpel polymorphism is exclusively on pattern of the vascular and bundles, not at all on developmental facts (Van Heel, 1969). The fact that locules but do further the bases of the of are formed, not develop, at second whorl in lobata and invalidates her carpel primordia Urena Pavonia praemorsa, argument. As it is that ten carpel primordia are initiated, evident ten stylar components and flower. stigmatic branches result in the open However, the question remains how in the flower is of open the ovary built up of five carpels, seemingly without any trace the other five. this To answer question, a study of the development was made, which is here re-

ported, by means of direct photography of early developing stages combined with This combination enables of microtomy. a good visualization the development of flower

structure. BLUMEA 124 VOL. 24, No. I, 1978

— of I 8: Pavonia spinifex, series c.s. ofa young pistil, X 30. 9—14: Ibidem, younger, x 45. W. A. van Heel: Morphology ofthe pistil in Mahaceae-Ureneae 125

MATERIALS AND METHODS

Cav. Botanic The species studied are i. Pavonia praemorsa (cult. Gardens, Leiden); hastata 2. Pavonia spinifex Cav. (cult. Bot. Garden, Groningen ); 3. Pavonia Cav. (ibid.); Malvaviscus arboreus lobata the field 4. Cav. (ibid.); 5. Urena (collected in near Bogor, ofall West Java*). Fixation material was either in FAPA, or in CRAF.

in Microtome sections were stained as follows, x. Safranin (CHROMA), 1% sol. 50% Rinse in Astra Blau FM alcohol, 1.5—2 hrs; 2. water, 1 min.; 3. Alcohol 30%, 1 min.; 4.

(CHROMA), 0.5 g. in 2% tartaric acid, I min.; 5. Rinse in water, 1 min.; 6. Aceton, 1 This schedule from and min., 3 X; 7. Xylol, 3 X; Depex (GURR). was adapted Maacz Vdgis (1961) and used for thin (6 /um) sections of the primordial stages**. Its result,

especially for the cell walls, is superior to the conventional safranin-fast green staining. The of the follows: the with photography primordial stages was as 1. Prepare objects

the aid of fragments of razor blades mounted on needle holders, leaving enough of the

stalks make solution Prior to handling easy; 2. Stain in a of JKJ in 50% alcohol; 3. to

photography, mount the objects under water on insect needles. The needles are put in

portions ofparaffin, fixed along the inside edges of petri-dishes; the whole is painted dull- black reflections. The allow to prevent stain keeps long enough (1 —2 hrs.) to photography; the with mounted 4. For photography LEITZ Ultropak system immersion caps was used,

on an Ortholux microscope equiped withan automatic camera. A KODAK PanatomicX film Ultrafin was used with a yellow-green filter, valued at 18 Din, and developed in think with Sattler that this is (TETENAL) 1 /20 for 8 minutes at 20°C. I (1968 and 1973) a quick, relatively cheap method, which especially brings well forward the cell walls.

However, the method is of limited application because of the low depth of focus.

THE DEVELOPMENT OF THE PISTIL

The the adult structure: development is marked by two factors, which are decisive for

whorl 1. the ten primordia originate as two successive whorls of five, 2. in the second the is ovules formed. development of an ovary limited according to the species, and are not In the following the primordia will be called the carpel primordia in accordance with the discussed carpel theory, although possibly the Ureneae do not fit the theory, as will be later. By this terminology a whorl of five fertile carpel primordia is followed on the five apex by a whorl of sterile carpel primordia (it is difficult to distinguish an outer and an inner whorl). The origin of the first whorl is shown for Malvaviscus in photo 1; the diameter this the pistil primordia are 1/5—1/4 mm in at stage. Prior to photography, and the young petal-stamen tubes, surrounding partially covering pistil primordia, were removed. the of the with to Along periphery apex five carpel primordia, a triangular trapeziform outline, arise, extending laterally. They remain separated by less active radial

which about cell wide. It is after the less active ofthe strips are two rows only strips apex have become and broader radial cell that to higher (3 —4 rows), they begin bulge forward, marking the origin of the sterile carpel primordia. The result is that ten less active strips between ten carpel primordia remain along the periphery of the apex (photo 2). Con- reach sequently, the fertile carpel primordia have a lead in development. They deep into

*) My sincere thanks are due to the Director of the Kebun Raya, Bogor, Indonesia,who gave me the oppor- subsidized tunity to work there for a period in 1969. This visit was by the 'Stichtingvoor Wetenschappelijk

Onderzoek van de Tropen' (WOTRO), the 'Maatschappij voor Wetenschappelijk Onderzoek in de Tropen (Treub Maatschappij)', the 'Greshoff's Rumphius Fonds', and the 'Van Leersumfonds'.

**) My thanks are due to Judith Kramer-Wiltink for taking care of this part of my work. No. 126 BLUMEA VOL. 24, 1, 1978

Pavonia series of of 15—27: praemorsa. c.s. a young pistil, X 45. 28—35: Ibidem, slightly older, X 45. W. A. van Heel: Morphology of the pistil in Malvaceae-Ureneae 127

the apex at the onset of the sterile carpel primordia, and may already show the first signs

of locule formation. This is the reason why (in adult stages) the proximal pistil part ap-

be formed the fertile albeit which pears to by carpels only, separated by apex tissue passes upwards into the tissue of the sterile carpels. The growth of the fertile and sterile carpel primordia proceeds towards the central

of the while at the time the itself becomes broader and parts apex, same apex slightly how the higher (photo 3, figs. 9—14. Sterile carpels at x). It is clearly shown in photo 4

bases of the sterile situated on the than those of the carpel primordia are higher apex

fertile carpel primordia. The latter become much wider at the base, occupying the corners

of a now slightly pentagonal pistillary primordium. Radial regions of the apex which are the situated medianly in front of the fertile carpel primordia, do not take part in inward development of these primordia. Therefore, the inward growth takes the shape of two of radial lateral flanks a carpel primordium enclosing a locule (photos 4 and 5). The

about in diameter this The of the pistil primordia are 1/3 mm at stage. top part primordia

overarches the locule and forms its roof. Growth in width forms the back of the locule.

sterile without locule however. In The carpel primordia are prominent formation, almost Pavonia spinifex, P. hastata, and Malvaviscus arboreus locule formationis absent or so,

lobata and Pavonia small locules evident in Urena it is slight (photo 11), in praemorsa are but do much further. (photo 12, fig. 37 at arrows), they not develop ofthe between The activation ten radial strips of the apex, which remain the ten carpel for that onwards primordia, is very important the final structure of the pistil. From phase the their and inward primordia appear as laterally connected, except by top parts growing

parts (photos 3 and 4). In this way the ovary wall originates. The centripetal growth of the flanks of the fertile carpels needs special attention. Grow- the fertile ing inwards, flanks of adjacent carpels in most cases become connected by the

development of a 'bridging' tissue in front of the sterile carpel primordia (photo 6).

However, in Pavonia praemorsa and in Urena lobata adjacent flanks remain separated by

'commissural' time fusion slits for a long (figs. 51 —58 at x), until the overall secondary at

the end ofthe these slits into the ventral slits of the sterile development. Upwards pass and carpels, which have a limited locule formationin these two species (photo 11, 12, 14). Also flanks all in Malvaviscus adjacent were found not to reach one another in cases of the of flank between fertile (figs. 42—50). As a consequence growth a bridging carpel

primordia, the centripetal growth of the sterile carpel primordia is limited at this level of the young pistil, their position remains more to the periphery; they are seemingly carried Also of upwards by the further growth of the fertile carpels. as a consequence the

bridging flank, the of the of the most formed the regions upper part ovary in cases are by fertile The valid for carpels only. same is the proximal parts of the style. This also means that the the tissue which passes upwards into sterile carpels (stylar components) can, at be of the anthesis, recognized only ifthe development is well known, the more so because late extension of the locules. It is situated slightly central to the sinus between two locules,

Urena minor as indicated for (photo 14 at x). In the other species, which have a sterile locule formation, this zone is still less evident, as shown for Malvaviscus (photo 17). later In stages the flanks ofeach carpel touch each other and fuse secondarily. However, before that the ovule of which long stage, originates exactly on the strip the apex was not occupied by the growth of the flanks of the fertile carpel. Photo 8 shows halfof a fertile carpel primordium of Pavonia spinifex, as photographed after a median longitudinal

madeof a such shown of the flanks of the section was pistil primordium as in photo 7. One fertile carpel is to have far inwards the front of seen grown as as region in an adjacent terile carpel. The ovule primordium originates more outwards, however, unconnected BLUMEA No. 128 VOL. 24, i, 1978

of Malvaviscus series of of 36—41: Pavoniapraemorsa, continuation figs. 28—35. 42—50: arboreus, c.s. a

Urena series of c.s. of young pistil, X 30. 51 —59: lobala, a young pistil, X 45. W. A. van Heel: Morphology of thepistil in Malvaceae-Ureneae 129

with a ‘Querzone’, and most probably unconnected with the carpel flanks. The ovule bottom the locule the fertile originates as it were at the of of carpel, where a locus on the cell apex remained free. Its development starts by division in the second and third layer

In the becomes of cells. this way ovule axillary to an enveloping bract. During later

of ovular microtome sections reveal 'cellular stages growth, never any regular con- the ovule cells tinuity' with the flanks of carpel oncross-section; the invariably 'come from

low in in the of the down'. Also later stages the ovules always appear exactly median radii ifthe locules and fertile carpels, ovary are of equal size form. It is only when the locules that are unequal some ovules appear contiguous to one of the flanks, suggesting a 'sub-

marginal' placentation on cross-section. The ofthe result roof ovary and the style are formed as a ofthe continuing centripetalous and overarching growth of the sterile and fertile carpel primordia, followed by their

adherence and ultimate fusion in the middle. Photo 8 makes clear how the closure and

style formation takes place when the inward growing rims of the fertile and sterile carpels

will contact one another in the middle. The of the roofof the the composition ovary or of the proximal parts of the style depends on the measure centripetal growth of fertile

this is described carpel parts. Usually, growth quite considerable, as above. In that case the of the of five fertile the middle and proximal parts style consist carpel parts in five

regions of tissue leading to the sterile carpel parts in the periphery. The distal stylar parts, of five sterile which have however, consist carpel parts grown in centripetal direction,

and fertile 6 and Sterile carpel parts in the periphery (photo 15; figs. 7. parts at x). Photo 13

and fig. 23 show the upward transition to the well developed tissue of the sterile carpels different in sectors of a young pistil of Pavonia praemorsa. Thus, a style consisting of ten radial the is components begins to be formed when pistil primordium about J mm in diameterand The the height (photo 9). extreme growth in length of style, as compared the is is with ovary, shown when the primordium about J mm high (in photo 9). Exter-

the is this nally, composition often parts evident, and five shorter stylar parts are(in case) the on top of fertile carpels. Internally, however, fusion has already started, in a way as

described by Baum (1949) and by Boeke (1973). This fusion is marked on cross-section by of cell walls the and a zigzag system at sutures (photo 16; figs. 18, 19, 32). Distally, the constituent remain parts free as stigmatic branches.

In the lower of the strands of tube tissue part style five pollen transmitting are formed

by division ofthe epidermal cells of the flanks of the fertile carpels at those places where they touch and fuse in the median radii ofthe fertile carpels. The pollen tube transmitting tissue extends upon the funicles of the ovules. Up in the style ten transmitting strands into the leading stigmatic branches are always present (photo 18). All strands fuse into one cylinder of pollen tube transmitting tissue in the middle of the style during the final of five stages development. In Pavonia praemorsa and Urena lobata one to alternating strands can be formed which terminate blindly downwards, in addition to the usual five strands

of due the pollen tube transmitting tissue. No doubt this is to initial formation of locules

of the sterile carpels in these species. In Malvaviscus this was also observed occassionally.

THE VASCULAR BUNDLE PATTERN IN THE PISTIL

Saunders (1936) presented descriptions of this pattern. Unfortunately, however, she created in connection with the her own terminology, theory of carpel polytnorphism, which makes her descriptions rather unintelligible. As intensive anatomical investigation would first knowl- be needed, a detailed description lies outside the scope of this paper. A edge of, and a comparison with, the more simple cases in the Malveae, in which all BLUMEA 130 VOL. 24, No. I, 1978

is carpels are fertile, would be necessary. Therefore a concise description given here. Immediately above the level of divergence of vascular bundles for the hypocalyx, the

base marked later vascular flower is by ten, on cross-section U-shaped or concentric, large

bundles, which are mostly arranged as an inner antesepalous and an outer alternisepalous of the bundles the median whorl. The outer parts antesepalous main diverge as sepal

bundles. The outer parts ofthe antepetalous mainbundles diverge as the sepal commissural

marginal bundles and the combined petal-stamen bundles. The innermost flank parts of which the antepetalous bundles move to the centre as ten pistillary traces, are located in

between sepal and petal radii. It is also possible that, above the level of divergence ofthe

and closed stele is which cross-section is sepal petal-stamen traces, a reconstituted, on

five-angled in the petal radii. Presumably, the one or the other possibility depends on the

amount of apical growth intervening between the origin of the sepal and petal-stamen other. If primordia on the one hand, and the carpel primordia on the there is a lag between

a stele will be from which also the two, reconstituted, ten gynoecial traces originate. into Each pair of gynoecial vascular traces, if taken as antepetalous pairs, divides two

carpellary dorsal bundles to the outside, two placental-stylar bundles to the inside, and,

not always, two lateral carpellary bundles in the middle. The dorsal and lateral bundles

the Still ramify upwards; the more so as ovary wall develops. more upwards they con-

and the bundles. Each ovule receives a vascular bundle verge again contact stylar single

which is connected by two traces with the pair of placental bundles.

From the above descriptions it seems that the stylar bundles are located as two lateral

for each a fertile evident bundles stylar component belonging to carpel. However, it is shifted the that these bundles are slightly sidewards, and are nearer to botmdaries between which the fertile carpels and the tissue upwards passes into that of the sterile carpels form (photos 13 and 17). Upwards, these bundles commissural bundles between the

constituentparts of the style, which belong to the sterile and fertile carpels resp. In thestyle fusion these bundles develop only after the lateral of the components has eliminated the

morphological boundaries between the two sets. It is observed that exactly in those

where cell in the regions the fusion starts, a division process dermaland subdermal layers

gives rise to commissural stylar bundles (photo 16). Upwards, each of these divides into

two lateral bundles for two adjoining stigmatic branches (photo 18). In this way a supply of vascular bundles to the sterile carpel components is effectuated, which the sterile carpels

could not receive direcdy from the baseofthe flower, by lack ofwell developed proximal parts. the In Uretta, ten more gynoecial traces may arise from antesepalous main bundles in base noted take location of the flower (also by Saunders). They up a placental-stylar

traces of the sterile carpels. They occur only if the sterile carpels are developed to some

extent.

DISCUSSION

This in in paper presents a case floral morphology which the study of development

proves essential. With the discovery that ten carpel primordia originate in two successive

whorls, the one fertile, the other sterile, the main cause for the aberrant final structure

and in Pavonia becomes evident. In Urena lobata (Endress, pers. comm.) praemorsa a carpel of the second whorl occasionally canbe fertile. This is in accordance with the fact that the

carpels ofthe second whorl clearly have initial locules in these two species. A development of a second whorl of carpels, which together with the first forms one pistilary structure, is for known for Punica, for Siphonodon (Croizat, 1947), and of course the Navel orange. If, W. A. van Heel: Morphology ofthe pistil in Malvaceae-Ureneae 131

the fruits with seeds, in the Hibisceae and in the Malvales in evolution, capsular many as

are considered and fruits with as a whole, primary, many monospermic parts advanced,

then the case of the Ureneae may represent a transitional state. During the first phase of development the carpel primordia become trapezoid bodies by apical growth. Subsequent growth in thickness, which is centripetal and lateral, takes

the overall formationof cell The developing carpels surround a free place by rows. apex. of free front of fertile Regions the apex, situated laterally in the carpel primordia, in centripetal direction join in with the apical growth, forming the flanks of the carpels.

As this occurs over a free apex, which at the same time becomes broader and higher itself, a locule results medianly in front of the fertile carpels. This description is different from such 'the of the is descriptions as margins carpels grow (bend) inwards', which and false. typological language with a developmental tongue therefore The manner of of the development carpel primordia resembles most that of enveloping basal parts of leaves. In the ovule the of the free that fact, primordia originate on regions apex are enveloped by the carpels, not on the flanks of the carpels themselves, as described above. the classical that the ovules the of As a consequence theory, holding are placed on margins carpels in all Angiosperms, cannot stand here; there is no room for heterotopy in the the theory. Also carpels should be named by a neutral term, such as gynoecial appendage,

bract It is that the in which the ovules or (cf. Sattler, 1974, 1975). possible ways are placed in a commonpistillary structure are more plastic thanadmitted hitherto. Inthis connection the other of the such the development of the pistil in groups Malvaceae, as in Hibisceae, in which the gynoecial appendages have more than a single ovule, should be studied.

of the in in Because variety gynoecial structure obviously allied groups, the Malvaceae are a promising family for comparative developmental study. sterile The joint development of the fertile and carpel primordia to form one pistillary wall, is called congenital fusion in the field of non-developmental morphology. The

has developmental analysis of these processes started only recently. An account was Sattler the presented by on Fourth Symposium of Plant Morphology at Strasbourg in

In there real fusion of that the March, 1977. no case is organs. It is evident meristematic different for various processes are cases of 'congenital fusion'. In our case it may be surmised that mantle of the floral is but that first the a peripheral apex potentially active, limited this has been activity is to two successive whorls of appendages. However, not the of The meristem described above makes subject our present study. 'bridging' the flanks of because of two adjacent ovule-enclosing gynoecial appendages grow up together the meeting in alternateradii of the meristems that give rise to the flanks. Meristematicactivity just extends to that region, resulting in an undulating (sinusoid) wall. On the other hand, median flanks in the radii the fusionof the is a real fusion process between two contiguous of Also the formationof the takes parts organs (called 'postgenital' fusion). style place by fusion of the ofboth the and secondary stylar components ovule-enclosing vacant gynoeci- The al appendages. problem of 'fusions' is intimately related with the problem of con- of vascular servatism the bundle pattern in the changing structure. In our case a new

is five structure produced in which pollen can reach ovules by way of ten stigmatic branches of the the vascular style. Correspondingly, pattern changes to give a new feature in the form often commissural bundles which, by simple division, supply both thenormal and the branches. five extra five stigmatic Thus a new structure induces a new vascular bundle pattern, which is against the idea of conservatism.

is It necessary to find techniques for visualizing the complex patterns of development which take the floral The is of the The place on apex. Ultropak photography one ways. results should be complemented andchecked by microtomesections. By combining both better be reached. techniques a insight may 132 BLUMEA VOL. 24, No. 1, 1978

REFERENCES

in Ost. Bot. BAUM, H. 1949. Uber die postgenitale Verwachsung Karpellen. Zs. 95: 86—94.

BOEKE, J. H. 1973. The use oflight microscopy versus electron microscopy for the location of postgenital

fusions in Proc. Kon. Ned. Akad. Wetensch. C. plants. 76: 528—535. L. A in CROIZAT, 1947. study the Celastraceae. Lilloa 13: 31—43.

DUCHARTHE, M. P. 1845. Observations sur l'organogenie de la fleur dans les plantes de la famille des Mal-

Nat. Serie vacees. Ann,. Sci. 3, 4: 123—150.

HEEL, W. A. VAN 1966. Morphology of the androecium in Malvales. Blumea 13: 177—394.

1969. The synangial natureof pollen sacs on the strength of 'congenital fusion' and 'conservatism of

the vascular bundle system', with special reference to some Malvales. Proc. Kon. Ned. Akad.Wetensch.

C. —206. 72: 172

MAACZ, G. J., & E. VAGAS, 1961. Triple stain Astrablau, Auramin, Safranin. Mikroskopie 16: 40—43.

Traite PAYER, J.-B. 1857. d'organogeniccomparee de la fleur.

SATTLER, R. M. 1968. A technique for the study of floral development. Canad. J. Bot. 46: 720—722.

Press. 1973. Organogenesis offlowers. A photographic Text-Atlas. Univ. ofToronto

1974. A new approach to gynoecial morphology. Phytomorphology 24: 22—34.

1974- A new conception of the shoot ofhigher plants. J. Theor. Biol. 47: 367—382.

Bot. —266. 1975- Organverschiebungen und Heterotopienbei Bliitenpflanzen. Jahrb. Syst. 95: 256

1977- Kronrohrenentstehung bei Solanum dulcamara L. und "kongenitale Verwachsung". Ber.

Bot. Deutsch. Ges. 90: 29—38. E. R. On features of floral construction and in the Malvaceae. SAUNDERS, 1936. certain arrangement Ann. Bot. 50: 247—282.

SCHUMANN, K. 1893. Elaeocarpaceae. Tiliaceae. Malvaceae. Bombacaceae. Sterculiaceae. In: E. & P., Nat.

Pfl. fam. 3, 6: 1 —99.

EXPLANATION OF PHOTOGRAPHS

1—3: Malvaviscus arboreus, pistil primordium, external view, X 150. 4: Pavonia spinifex, ibidem, X 150;

6: Pavonia Pavonia 5: ibidem, X 180. hastata, ibidem, X 180. 7: spinifex, young pistil, external view, X 150;

8: ibidem, median l.s. of fertile X ibidem, X 110. 10: Malvaviscus exter- carpel, 120; 9: arboreus, young pistil, Pavonia nal view, X 90. 11: Urena lobata, pistil primordium, external view, X 180. 12: praemorsa, pistil

external Pavonia primordium, view, X 150. 13: praemorsa, cf. fig. 23, X 250. 14: Urena lobata, c.s. of centre Pavonia ofthe pistil at anthesis, X 200. 15: spinifex, cf. fig. 7, x 200. 16: Pavoniapraemorsa, cf. fig. 32, X 400.

Malvaviscus arboreus, of centre at X 18: 17: c.s. ofpistil anthesis, 150; ditto, subdistal c.s. of style at anthesis,

x 150. Malraceae-Ureneae W. A. van Heel: Morphology of the pistil in 133 BLUMEA VOL. 134 24, No. I, 1978 W. A. van Heel: Morphology of the pistil in Malvaccae-Ureneae 135 BLUMEA No. 136 VOL. 24, I, 1978 W. A. Heel: the in Malvaceae-Ureneae van Morphology of pistil 137 BLUMEA VOL. No. 138\I 24, 1, 1978

of Plate I A. Memecylanthus neo-ealedonicus (Deplanche 413, K), transverse section flower, showing fleshy

B. hypanthium(base of calyx, corolla tube, and stamens), x 20.— Periomphale pancheri(Schlechter 15426, L), of section axile — C. balansae transverse ovary, showing placentation (ovules outlined), X 75. Periomphale section (Balansa 2776, K), tangential longitudinal of stem, X 260. Casparian strips (darkly stained with safranin) of the endodermis form a net-like pattern. — D. Memecylanthus neo-caledonicus, tangential longi-

— E. tudinal section of secondary wood, showing tall rays, x 120. Periomphale balansae, tangential longi-

section tudinal ofsecondary wood, showing tall rays, x 120.