r984l N. G. AND R. E. DENGLER:PLICATION FORMATION

Principes,23(1),1984, PP. 31 48

Formation of Plications in the Pinnate Leaves of r11 r'I t Lnrysalldocarpus Iutescens and the PalmateLeaves of Rhapisexcelsa

NaNcv G. DrxcrsR ANDRoNALD E. DEncrnn Department oJ'Botany, finiuersity ofToronto, Toronto, Ontario MSS lAl, Canada and Department of Botany, Scarborough College, Uniuersity oJ'Toronto, Vest HilI, Ontario MIC lA4, Canada

For over 100 years morphologists palmately compound leaves first appear as have known that the large, dissectedleaves hood-shaped protuberances which encircle of palms exhibit a developmental pathway the shoot apical meristem. A series of which is distinctly different from that of transverse to oblique folds or plications the compound leaves of all other flowering develop near the margins of the primor- . ln those dicotyledonsand mono- dium (Fig. lA); these indicate the position cotyledons having dissected leaf blades, of the future leaf blade. A distinctive {ea- leaflets arise as free lobes on the margin ture of these plications is that they do not of the primordial leaf. In contrast, leaflet extend to the leaf margin, leaving an inception in the palms occurs by folding unplicate marginal strip of tissue. During of the submarginal part of the lamina, fol- the growth of most palm leaves a localized lowed in most groups by the secondary separation of tissue results in the splitting splitting of the blade into individual leaf- oT-the lamina into individual leaflets. In lets. Since this mode of leaflet origin seems some groups of palms the splitting occurs to have no counterpart in other flowering in the ridges nearest the shoot apex (Fig. plant and is markedly complex in lB,C). This is the adaxial side of the leaf character, it has received a considerable and because the leaf bends away from the amount of attention from morphologists in shoot tip as it expands from the crown, the last century (von Mohl 1845, Hof- the mature leaflets are trough-shaped in meister 1868, Goebel 1884, 1926, Nau- cross section or induplicate. In other mann 1887, Deinega l89B) and more groups of palms, the splitting occurs in recently (Yampolsky 1922, E{mes 1953, the ridges farthest from the shoot apex Venkatanaray ana 1957 , Periasamy 1962, (abaxial ridges); this results in leaflets that Padmanabhan1963, 1967, Corner 1966, are ridge-shapedin crossseclion or redu- Padmanabhan and VeerasamY 1973). plicate (Fig. lD). In some palms, the loca- Despite the length of time that the prob- tion of the zones of separation does not lem has been studied and the large num- correspond to adjacent ridges so that indi- ber of papers devoted to this subject, con- vidual leaflets consist of a number of the troversies on the mechanism of the original original plications (Fig. lE). The tips of folding or plication remain unresolved and the leaflets also separate from the unpli- misinterpretations Persist. cate marginal strip of tissue which is In the past there has been general ephemeral in many groups, but which may agreement amongst these investigators on form a pair of conspicuous reins attached the general pattern of palm leaf morpho- to the two basal leaflets(Eames 1953). genesis. Both pinnately compound and The source of controversy in the liter- PRINCIPES lVoL. 28

X A

Aboxiol Adoxiol

B Adoxiol ---x ^---

Aboxiol UU (A) and the aPpearanceot a I. Diagramsillustrating an adaxialview of a hood-shapedpalm leaf primordium by the line (x x)' Abscissionthrough section(B) made through the ""u*"tgi""r plications in the plane indicated (C); through the abaxial ridges (arrow) will *" "J".irf ridges (arrJw) will result in iniuplicate leaflets abscission regions may result in leaflets consisting ,".Ji"."arpi"utl l"ufl"t" (D); or zonesoiabscission in thg intercostal of more than one Plication(E).

such a way that ature has been over the mechanism of nal separation of cells in break through inception of the original folds or plications internal schizogenous slits on alternate sides in a previously smooth lamina. The point the epidermal surface appearance of disagreement is whether tissue separa- of the leaf giving the pleated the differences tion is involved in the initiatiqn of the pli- (Fig. 2A,D,E). At first might appear cations themselves. For example, Hof- between these alternatives quite dif- meister (1868), Goebel (1884, 1926), trivial; however, each involves If tissue Deinega (lB9B) and more recently Per- ferent morphogenetic processes. internal cells of the iasamy (1962) and Corner (1966) con- separation did occur, schizogenous clude that the plications are formed by a ground tissue lining the to become process of differential growth followed by spaces and normally destined cells would now r foldi.te of the meristematic lamina as piotosynthetic mesophyll of the leaf and shown in Figure 2A-C. On the other hand, L" o.r ih" external surface as the protective cells von Mohl (1845), Naumann (I887), would differentiate contrast, the process Yampolsky (1922), Eames (1953) and of the epidermis. In only involves local- Padmanabhan(1963. I967) argue that of differential growth to give rise to the the process of pleat formation involves not ized zones of growth of tissue with- only differential growth, but also the inter- adaxial and abaxial ridges l 9841 N. G. AND R. E. DENGLER:PLICATION FORMATION 33

A Adoxiol

A boxiol

)

I C EV

2. Diagrams illustrating the formation of alternating adaxial and abaxial ridges through diferential growth (A,B,C) or through the extension of internal schizcgenous slits to the surface (A,D,E).

out the complicated redifferentiation of lutescensH. A. Wendl. (arecoid major cells. group, Moore I973) having pinnateredu- The purpose of this study is to resolve plicate leaves and excelsa this long standing controversy over the (Thunb.) Henry (coryphoidmajor group, mechanism of plication inception and, Moore 1973) having palmateinduplicate particularly, to test the hypothe;is of the leaves.Shoots of both specieswere col- schizogenous origin of the plications and lectedat FairchildTropical Garden,Coral subsequent redifferentiation of tissue. Since Gables,Florida and were shippedto the this mechanism has not been reported to University of California, Berkeley. This occur in any other groups, investigation was carried out in conjunc- this would be a further unique feature of tion with Dr. D. R. Kaplan, University of palm leaf development that might be of California, Berkeley, and the observations potential significance in evaluating the which are summarizedbelow were reported systematic relationships of palms, partic- in greater detail in three papers in the ularly with other having Canad.ianJournal of Botany (Kaplan et plicate leaves. al. I982a,6, Dengleret al. 1982). The Species having the two basic types of specific methods used to study the prob- palm leaf morphology were selected for lem of plicationinception are alsodescribed this investigation: Chrysalidocarpus in thesepapers. 28 34 PRINCIPES lVoL. Results

Morphogenesis iz Chrysalidocarpus lu- tescens. The mature leaf of Chrysali- docarpus measures about 2 meters in length and is differentiated into a pin- nately compound blade, a short and a smooth, tubular base (Fig. 3). Each of the approximately 80 leaflets has a single major vein or rib and is reduplicate. The leaves are arrapged in either a clockwise or counterclockwise phyllotaxis (Kaplan el al. I9B2a). The developing leaves are tightly packed together within the apical bud and their appearance in scanning electron micrographs and in sectioned material viewed in the light microscope suggest that considerable mutual pressure is exerted between adjacent leaves (e.9., Fig. 6). Figure 4 shows a young leaf about 0.5 mm in length in which there is no external indication of plication inception but which clearly shows the hood-shaped lamina, and a sheathing leaf base which encircles the shoot apex and younger leaf primordia. The upper limit of the sheath of the next oldest leaf can be determined by the loca- tion of the constriction near the base of the leaf lamina (Fig. 4, arrow). Developing 3 leaves of Chrysalidocarpus are charac- One margin of teristically asymmetrical. 3. Mature leaf oI Chrysalidocarpus lutescens the lamina protrudes through the sheath showing the pinnate lamina, petiole and tubular of the next oldest leaf; this is the anodic sheathing base. Scale line equals 50 cm. (*) margin. The other margin is appressed against the rachis; this is the tathodic (-) margin. The leaf illustrated in Figure 4 is from a shoot with counterclockwise phyl- ridee formation on the abaxial side of the lotaxis and therefore the right hand mar- anodic ma.gin; the folding of the cathodic ein as viewed from the adaxial side is margin against the rachis obscures the anodic (*). Leaves from shoots with ridges (Fig. 5A,B). As the leaf develops, clockwise phyllotaxis have the anodic (*) the abaxial surface appears flat in surface marein on the left. view and the furrows between the abaxial Plications first appear in leaves 0.6 to ridees resemble slits incised in the surface 1.0 mm in length as a series of slight of ihe lamina (Fig. 6). Plications do not ridges and furrows having a vertical ori- extend to the margin or apex of the leaf, entation in the distal part of the lamina leaving an unplicate margin and a hood- (Fig. 5A). Formation of the ridges on the shaped leaf apex (Fig. 6). While the first adaxial side of the leaf is seen to precede formed apical plications have a vertical 35 N. G. AND R. E. DENGLER:PLICATION FORMATION

leaves of chrysalitlocarpus Lute'scens'Fig' 4' Leaf pri- 4,6. Scanning electron mrcrographs of developing (B), hood shaped lamina-(L) with anodic (*) and mordium 0.5 mm in length. NJte'th" sheathing l""uf ba." 'l'he formed by the sheath of the.next oldest leaf' Scale cathodic (-) margins. arrow indicates the constriction lineequals200prr.l.ig.5A.Leafprimordim-aboutlmminlength.Notetheverticalorientationofabaxial ritlges (ArlR) of the anodic margin' Scale line equals r-idges'(AbR) on the catirodic l*f -argir and the adaxial leaf a" shown in Fig- 5A and with the plications used lor t06-d. Fig. 58. The anodic -u.gi."of the same Scale Iine equals 100 ir'm' Fig' 6' Leaf 2 mm in histoiogical Inalysis, 2, 3, 4 above the constriction, labelled. below the constriction (arrow)' }lote the unplicate t"rrgtrrir which new plications are being {ormed basipetally (I)' Scale line (RP) impression of younger leaf on adaxial surface of sheath *lisi" tUNal, rachis protusion ani equals 200 trrm. 28 36 PRINCIPES [VoL.

(Fig. orientation, later formed plications are ori- layers thick and has a smooth surface ented in the horizontal plane. Based on B). The appearance of slight ridges on the length and surface appearance, new pli- adaxial side of the lamina is associated (parallel cations appear to be added both acrope- with localized areas of periclinal tally and basipetally until the leaf reaches to the surface) and oblique divisions in a leneth of about 5.5 mm with about 40- middle layers of the lamina in addition to 42 plications on each lamina half (Fig. 7). slight cell enlargement in the adaxial pro- (Fig. During the period of plication initiation, todermal and subprotodermal layers thickening growth of the rachis axis results 9). Further growth of the adaxial ridges in the formation of a rachis protrusion is accompanied by the development of which covers the base of the adaxial ridges zones of mostly anticlinal (perpendicular and furrows (Figs. 5A.b). Elongation of to the surface)divisions alternating with the leaf, particularly above the point of the ridges (Fig. l0). This pattern of growth confluence of the margins, marked by an results in an increase in surface area of arrow in Figure 7, results in a gradual the portion of the lamina undergoing pli- reorientationof the apical plicationsto the cation. Since the plicated part of the lam- horizontal plane. Later in development, ina is bordered by the unplicate leaf mar- elongating plications come to lie in the gin, the leaf apex, the rachis, and the vertical plane. This vertical packing of petiole (Fig. 6), and since the enlargement leaflets against the rachis can be seen in of these parts of the leaf does not keep a new sword leaf as it emerges from the pace with the plicate lamina, the compar- crown. Prior to emergence, splitting of the atively rapid extension of the confined blade into leaflets occurs. When the leaf lamina is accommodated by buckling of (Fig' is fully elongated, expansion of the pulvi- the lamina between adaxial ridges nus at the base of each leaflet causes it to l0). This growth pattern brings about the open and bend away from the rachis (Cor- eppearance of abaxial ridges and furrows, ner I 966). and further growth in surface area results in the compression of adjacent ridges on lu- Hi.stogenesis in Chrysalidocarpus both sides of the leaf, giving the slit-like tescens. Documentation of the early appearance of the adaxial and abaxial fur- stages of leaf development using scanning rows (Fig. li). The plications deepen in electron microscopy allowed us to under- the radial plane by further intercalary stand the complex three dimensional rela- growth in the intercostal sector, causing tionships of parts of the leaf during the the adaxial and abaxial ridges to be dis- and critical stages of plication formation placed from one another (Fig. I2). to orient young leaf primordid for section- All developmental stages examined in half of ine. Serial sections of the anodic this study lacked any indication of internal just above the point of con- the lamina slits opening to the surface to form alter- (ridges 2, 3, 4 in Fig. 58) were striction nating ridges and furrows, although slightly made from the unplicate margin through oblique sections taken near the rachis the plications to the rachis in a plane per- occasionally (and incorrectly) gave the pendicular to the axis of plication exten- impression of internal schizogenousspaces sion. These series of sections were ana- (Dengler et aI. 1982). Analysis of serial lyzed to determine the pattern of cell sections also showed that the protoderm growth described below; outline drawings as a continuous layer of cells on of sections from the midregion of the pli- appears and abaxial surfaces of cations only are illustrated in Figures 8- both the adaxial 12. the leaf throughout the developmental observa- Prior to the formation of plications the stases observed. Both of these lamina of the leaf is five to seven cell tions support the alternative of differential ,)Bll N. G. AND R. E. DE\GLER: PLICATION FORN'IATION

7 nLm in lengtb' Arrorv indicates Scanning-*" electron micrograph of a leaf of Chrysalidocarpus lut.escens apex. Scale line equals 500 pm' p"iir of confluence of ih" urpli"ut" rnargins with the hoodlike PRINCIPES [VoL. 2B

O.5mm

2.5mm

10 12 older leavesof Chrysalidocarpus- B 12. Outline drawingsillustrating the midregion of plications of successively Scaleline equals50 pm' Fig' B' Sectionof lutescens.Sections are orientedwith the adaxialsurface upPermost. the margin' Fig' 9' Sectionof the plicate ,rr" ,"fii""r" lamina of a leaf 0.5 mm in length taken Sd'pm from The adaxialridge of plication 2 abo,vethe lamina of a leaf 0.25 mm in length taken Tdpm from the margin. leaf about l'0 mm in length taken 64 trrm constriction is labelled. l-ig. I0."Section of a plicate lamina of a marks the position of the abaxial furrow i-- ,rr" margin. The adax"ialridge of plication 2 is labelled. The arrow in length taken 84 prmfrom the margin' oitii. pti""tiJ.. Fig. 11. Section"o{plicate Iamlna jr^o1 a leaf 1.5 mm taken lB2 pm from the margin' Arrow flg. i2'. S"",i". fr"om the plicate lanina of a leaf 2.5 mm in length showstlirection of growth in intercostalsector (IS)'

cell layers were growth, but it can be argued that the most Counts of the number of of the adaxial ridge iigoro,.r test of the mode of plication for- made at the position abaxial ridge (adaxial mation would be an analysis of the num- (abaxial furrow); the intercostal sector in leaves ber of cell layers across the lamina. If the furrow) and the as shown in Figure 13' formation of internal schizogenous slits of increasing ages the number of cell layers is results in ridge and furrow inception, a A decrease in at any stage. These compar- decrease in the number of cell layers across not observed a dramatic the lamina at the locus of the furrows isons clearly demonstrate number of cell layers (from would be expected; conversely, if differ- increase in the position of the adaxial ridge ential growth alone accounts for plication 5 to l6) in the of these cells are asso- formation, one would exPecl onlY an (Fig. l3A); many of the pro- increase in number of cell laYers. ciaied with the differentiation r9841 N. G. AND R. E. DENGLER:PLICATION FORMATION

t6 Adoxiol ridge t4 a ao t2 aa a aa aa a a aaaa aarta lta ro a aaa aaaa oa I a a aaa IrO

6 4 2

o (l} t2 o Aboxiolridge ro

(l, (J I

o 6

qt 4 o E 2 :t z t2 lntercoslol secfor ro 8

6

4

2

2 3 4 6 Leof lengfh, mm (A), abaxial ridge 13. Scatter plots illustrating the changes in the number of cell layers in the adaxial ridge length. Arrows (B) and intercostal section (Cfin the leaves of Chrysalidocarpus lutescens 0'5 mm to 7.0 mm in indicate the stage of plication incePtion. 40 PRINCIPES [Vor. 2B cambial strand of the future leaflet mid- vein. An increase in the number of cell layers also occurs in the abaxial ridge (Fig. l38) and the intercostalsector (Fig. l3C) although the increase is less dramatic (from 5 to 7 layers). Tf splitting were occurring durine furrow formation and were com- pensai"d for by localized growth in the associatedridge so that the number of cell layers was not altered, one would expect to see cell wall patterns in the ridges indic- ative of this localized compensatory srowth. Since this was not observed we conclude that all of our qualitative and quantitative histological observations are consistent with the concept of differential growth-alternating zones of the leaf lam- ina undergoing different growth rates resulling in plication formation. 14. Leal of Rhapisercelsa showing differentiation into a palmatelamina, petiole and tubular sheathing Morphogenesls in Rhapis excelsa. The base.Arrow indicates adaxial hastula. Scale line equals 50 cm. mature foliage leaf of Rhapis is about 0.8 meters in length and is differentiated into a palmately compound blade, a long pet- number of monocotyledous species (Kap- iole and a fibrous tubular leaf base (Fig. lan 1973). The concave adaxial surface l4). Although the blade is divided into l0 oT the lamina is covered by the adaxial to 12 segments, segment number does not hastula and the convex abaxial side bears correspond to plication number and each the abaxial hastula. Plications are. not segment has two or more ribs along its externally visible on the abaxial side of the length. A pair of small crestlike append- lamina until leaves are about I mm in ages called hastulae are attached at the length (Fig. 16). The first abaxial ridges junction of the blade and the petiole on appear as subtle mounds on the narrow the adaxial and abaxial sides of the leaf horizontal ledge of tissue that represents (Fig. 15). As in Chrysalidocarpus, the the free margin of the leaf (Fig. 16). The leaves are arranged in either a clockwise adaxial ridges and furrows are obscured or counterclockwise spiral in the bud which by the adaxial hastula at this stage, imposes an asymmetry on the tightly althoueh sectioned leaf material reveals packed developing leaf primordia (Kaplan that the development of the adaxial ridges et al. I982b). is actually advanced over the abaxial ridges Prior to the inception of plications, the (Kaplan et al. I982b). As the leaf expands, young leaf primordium of Rhapis consists the lamina becomes more vertical in ori- of a tubular leaf base and a hood-shaped entation and the abaxial ridges become lamina in which an anodic and a cathodic sharply demarcated from each other by side can be distinguished (Fig. l5). The slit-like grooves (Fig. l7). The first plica- lamina differs from that of Chrysalido- tions to be initiated are at the apex of the calpus in that the apex is prolonged into leaf and subsequent plication initiation "Vorliufer- a conical protuberance or occurs in a basipetal direction along both soitze" characteristic of the leaves of a the anodic and cathodic margins. The 19B1l N, G. AND R. E. DENGLER:PLICATION FORMATION 41

E AdH tir,l s ."1 I *' q! FJ 'a '..+f. t7-. 15;

l5.Terminalbudsholving l5 IB. scanningelectronmicrographsof clevelopingleavesof Rhapisexcelsa.Fig- hastula (AdH), abaxial leaves 150 p-, +bO pm a.d 65b 1im nr length. The oldest leaf exhibits an adaxial "Vorl5rferspitze" (V). Scale line equals hastula (AbH), anodic (*) and cathodic (-) margins and a prominent between the leaf margin and 200 pm. Fig. 16. Leaf I mm in length. Slight uba"ial ridgei (arrow) are evident (AdH). Scale line equals 200 the ataxial f,astula (AbH). Adaxial riilges a.e cotered by the adaxial hastula trrm' adaxial hastula (AdH) Fig. lT. Leaf 2 mm in length. Abaxiallidges are numbered. Note unplicate rnargin_(UM), line equals 500 pm' Fig' lB' aritl abaxial hastula (AbH). Arrow indicates impression of next youngest leaf. Scale (AbR) evident' Scale line equals Leaf 3 mrn in length. Both adaxial ridg"" iedR; and abaxial ridges are 500 gm PRINCIPES [VoL.2B ridges and furrows of each plication clearly Histogenesis in Rhapis excelsa. Since do nol extend to lhe lamina margin and the orientation of the developing plications the crescent shaped zone of attachment of is vertical in the leaves of Rhapis excelsa, the plications to the rachis is obscured by a histological analysis of plication incep- the hastulae. tion in the leaves was based on cross sec- As the leaf elongates, the adaxial ridges tions of developing leaves. Also because and furrows begin to extend beyond the of the differences in stage of development adaxial hastula margin (Fig. i8). The of olications of one leaf and differences abaxial hastula elongates along with the along the length of each individual plica- given here lamina; so that .while the horizontal part tion, the histological description of the plications adjacent to the margin is will be based on cross sections from the exposed, the elongate, vertical portion of midregion of plication 2 (Fig. I7). As in the abaxial ridges is protected by the has- Chrysalidocarpus, rhe anodic margin of tula. The leaf illustrated in Figure 17 also the leaf protrudes from the sheath of the shows how mutual pressures within the next oldest leaf (Fig. 15) while the ca- apical bud leave the impression of the next thodic margin is somewhat compressed youngest leaf on the adaxial side of the inside the sheath. Although this results in leaf sheath (arrow). The final number of differences in the degrees of compression 20-24 plications is reached in leaves that of the developing margin at early stages, are uborlt 4.5 mm in length. Trichomes plications 2 from both the anodic and develop on the edges of the plication ridges cathodic margins will be used to describe beginning with the unplicate margin and histological development. proceeding basipetally; growth of the tri- Figure 20 is a section through the chomes and margin eventually obscures anodic margin of a leaf at the stage when "VorlSuferspitze." the Trichomes also plication 2 first appears as a slightly con- develop on the margins of both hastulae fex mound which is located between the in alignment with the plication furrows adaxially curved free margin and the (Fig. l8). Eventually, expansion of the adaxial ridge of plication l. The formation hastulae does not keep pace with the elon- of these ridges is associatedwith periclinal gation of the plications so that both adax- and oblique divisions in the internal cell ial and abaxial ridges are exposed (Fig. layers (Fig. 20). At this stage the adaxial l9). In the 6 mm leaf shown in Figure 19 furrow between the ridges of plications 1 the line of demarcation indicated by an and 2 appears as a broad depressionrather arrow on the abaxial ridges delimits the than a narrow furrow. As the leaf grows, original horizontal portion of the plication compression of the lamina brings the ridges from the more vertically oriented part of adjacent plications into closer proxim- which was appressed to and covered by ity, narrowing the adaxial furrow (Fig. 21). the abaxial hastula through much of early This occurs at earlier stages in plications development. Separation of the lamina into of the cathodic margin as compared with 10 to l2 sesments occurs within the inter- those of the less compressed anodic mar- costal zone of tissue lying between adja- ein. Internal divisions result in an increase cent adaxial and abaxial ridges as indi- in tissue layers at the locus of the adaxial cated in Figure lE. Elongation of the leaf ridee while anticlinal divisions and asso- and its segments continues until the new ciaied cell growth in the intercostal sector leaf emerges from the crown as a sword between the ridges extends the surface leaf. As in Chrysalidocarpus. expansion area of the lamina. Although extension of of the pulvinar region at the base of each the plications in this palmate leaf does not segment brings about the opening of the occur in such a confined space as observed folded seements. in Chrysalidocarpus, buckling of the 43 PLICATION FORMATION rs+l \. G. .\\D R. E. DE\GLER:

,,, * 't I AbH

the horizontalportion Ig.scanningelectronmicrographofaleafoffrhapisexcelsrl6mminlengthwhentheplicationshaveurd.ubu*iulil.*i" teunr' A'o*-i"tdicut"t wellbeyond ,r." "a""i?iT".,rr'"iaJni stages of development' Note elongated by,h";;;i;astulu at all 500 pm' of ti]e abaxial plication,idg"r;hi;";.;'*"ot"."a of hu"t,lu"' Scale line equals margin, plication ,idg". ""i;;;;i.l trichome development on leaf 44 PRINCIPES IVor.2B x't3l AdR K /*f ;{: t ; *i: 2 tr K I

r ,24

,t ,-l .*.{

! 3 t

$ _-.?r {!' :l :{

:!, la .

j'*1{ ?. 23 .4r r .' .1 --l' L *.1

20 23. Cross sections of developing leaves of Rhapis excelsa showing the midregion of plication 2. Fig. 2O. Section from the anodic side of a leaf 0.8 mm in length taken 187 plm from the leaf margin. The adaxial ridges of plications I and 2 are associated with localized periclinal and oblique divisions. Scale line equals 50 pm. Fig. 21. Section frorn the cathodic side of a leaf 1.0 mm in length taken 265 pm from the leaf tip. Adaxial ridges I and 2 are compressed so that the adaxial furrow (AdF) between them appears slit-like. Scale line equals I00 irm. Fig. 22. Section from the cathodic side of a leaf about I.2 mm in length taken 266 prn from the tip. The adaxial ridge (AdR) and associated abaxial furrow (AbF) of plication 2 are labelled. Scale line equals 100 pm. Fig. 23. Section from the cathodic side of a leaf 2 mm in length taken 619 ;lm from the leaf tip. Adaxial ridge (AdR) and intercostal sector (IS) of plication 2 are labelled. Scale line equals 100 g.m. ,rt5 ,lB+l N. G. AND R. E. DE\GLER: PLICATION FORN,IATION

the plications. In addition, iaurina also occurs and results in the for- nous origin of the numbers of cell lay- :nation of broadly rounded abaxial ridges careful counts of lamina. made along the .rncl narrow abaxial furrows (Ftg' 22)' ers acrosb the in leavesof a range lnlercalarygronth of lhe intercostalsee- lengthol the plications an increase in the tLrr results in the displacement of the of ages, only showed layers. One rvould expect ,,larialand abaxialridge' ar*a; from each number of cell if split- ,rlr"r(Fig. 23). All of thesesections shorv a tlecrease in number of cell layers the observed increase .r complete continuity of the protodermal ting were occurring; r-elllayers suPPorts lhe con- ,.Lrerat all stagesof development. in number oI growth' Perhaps the \lthough this histological analysis of cept of differential argument for our conclu- ;'lrcation origin in Rhapis gave no sug- most convincing similarity in the histology o1 -,..tionof .plitting.the samequanlitalii e sion is the in a palm with pin- analvsis of plication growth used in CArl'- developing plications ,,Lliilot'arprtswas carried out as an addi- nately compound leaves, Chrysalidocar- and a palm with palmately tional test of the mechanism of morpho- ptLs iutescens, excelsrt. Despite gene"i.. This analysis showed thal the compound leaves, Rhapis in the morphology of the greatest increase in number of cell layers the differences leaves, the orientation of the pli- ifro- 5 to 26) occurred in the adaxial young and the physical contraints on the ridge (Fig. 24A); as in Chrysalidoc(rrpus, cations eloping lamina. our cross sectionsof this is associated with the differentiation der plications of both Rhapis ar'd of the procambial stand of a major vas- developing are strikingly similar "ular bundle.Although the inereasein the ChrysaLidocarpus (cl'.. number of cell layers in the abaxial ridges Figs. 10"22\. separationdoe' (Fie. 24B) and the intercostalsectors (Fig' Our asserlionthal tissue in the initial formation of 24C) was less (from about 4 to about 8 not play a role plications is based on an understand- cell layers), no evidence of a decrease in the morphology of the young the number of cell layers was observed at ing of the gross of the plane of sec- any developmental stage. leaves, careful control tioning, use of serial sections so that posi- tion within a particularplication is known. Discussionand Conclusions recognition of the age of the leaf under of a large We conclude that the process of differ- obseivation and examination leaves. Although the ential growth at the locus of the adaxial number of individuai handicapped by ridges followed by inlercalary grou th earliest investigators were available to them, between ridges accounts for the original the limited techniques of this aspect formalion of plication. in a submarginal the fact that the mechanism ha' remained position (Dengler et aL. 1982, Kaplan e/ of palm leal morphogenesi' over a century is largely oL. I982b). This conclusion is based on uniesolved for one or more observations of oriented serial sections the result of misunderstanding above (Kaplan et a/' through developing plications in a closely of the factors listed graded series of leaf lengths; our obser- I9B2ct). Naumann's (iBB7) vations of sections of over 140 leaves of For instance,despite of the problem in ChrysaLidocarpus and 63 of Rhaprs were early clear statement initiation, his own .oniirt"nt with this interpretation. No sec- interpreting plication of Phoenix were tions revealed the presence of internal slits anatomical observations of the whole leaf, developingtoward the:urface or any based on cross sections be oblique to the breaks in protodermal continuity which an angle which would surprisingly, he would support the hypothesis of schizoge- developing plications. Not PRINCIPES [VoL.28 Adoxiol ridge A 28 24 lnru 20 t5 ao o aa l? ao I 4 vl (u o Aboxio I ridge t6 (u (J t2 mffiB o I tu lt 4 E t z Intercoslolgeclor

t2 I

4

o.4 0.8 1.2 t.6 2.O 2.4 Leof length, mm

24. Scatter plots illustrating the changes in the number of cell layers in the adaxial ridge (A), abaxial ridge (B) and intercostal sector (C) in leaves of Rhapis excelsafrom 0.5 to 2'5 mm in length. 47 DENGLER:PLICATION FORI\IATION 19841 N' G' AND R' E

sion (Fig. 20); only as additional plications young plications as poucnes described all the plica' they are initi"ated basipetally and having internal slits a"d deduced that "the in depth do the furrows tissue separation' As tions extend uere result of (Fig' become narrow and slit-like 22)' example, although Goebel(1926) enother have missed the stages the leaves of E.laeis Some investigators iook care in orienting inter- to with roundeJridges and have based -o that sections were made tangentlal on th" more comPressed of his critical stages oretations rl,c rnargin. some Fo1 too "duun""d stages of plication growl!, -,r'r)ear io be oblique sections made instancePadianabhan (1963, 1967) and ' i,l-" ,o the unplicate strip of marginal and VeerasamY (1973) he concluded that differen- Padmanabhan ::--,r,'. \ hile groo\es in have argued that the V-shaped ' : ir,,',"th las the process resulting sections of the leaves ol Cocos' ': . of plications, he failed ob.er,reJitt r,,rrttation -to as the zig- the Borassas and Phoenix, as well : : 'i r,l. convincing evidence for zag appearanceo[ the walls of the [ur- ,,r", hartism' Our own observations indi- ,oi*t. it the result of separalionbelween r'ate that plication inception occurs at a the internal ground meristem cells which relatively early state which is easily missed have an irregular alignment' Agarn, com- ibl a leaf length of 5'5 mm in Chrysali' of their illustrationstith our sec- a length of 4'5 mm in oarison rloccLrpus anJ of pli- ( tions indicates that the early stages R hapis).\ enkatanarayanaI957)'^who cation formation were probably missed and atrributed plication development in Lclcos irregular appearance of some tur- splitting.aPpears to hare that the to schizogenous"his of distor- have row surfaces may be the result oriented sections properly but to tion of susceptiblemerislematjc tissues missed the actual stage of plication incep- caused bv fixation and dehydration' a figures are of a relatively late tion. His many o[' our likely problem ihut ul.o plagued stage of pliiation growth and most adja- sections. ,"p"r"."nt u ,on" ofuppt"ssion of two ' The appearance of internal slits in some cent plications similar to that shown in sectionaf views has also provided a major Figure I I o[ this PaPer' source of evidence in supporl o[ lissue ih" *alot lines of evidence cited.in seoaration (Naumann 1887, Yampolsky support of the process of tissue seParatlon 1d22, Eu*". 1953, Padmanabhan 1963' uti, t) the slit-like appearance of the fur- 1967j. In our material of Rhapis and' rows in surface views of whole leaves and any indication of inter- views of growing plications' ChrysaLidocarpus in sectional result in .rul ,lit, couid be shown to be the and 2) the occurrence of internal slits- oblique section taken adja- A primary reason that of a somewhat some sections. inter- the cent to the rachis and the apparent ChrysaLidocarp.r's was select6d to test nal pocketscould alwals be shown lo be hypothesis of tissue separation.is the nar- the exlernal >urlace ol of the furrows continuousnith row slit-like appearance (Den- the leaf by following serial sections shortly after inception (Fig' 6)' However' rtl" 1982, Kaplan et al' I980b)' yottttg"t plications dem- "gler et sections through that the at Orr. observations indicate onstrale lhat the plication ridges are mechanism of differential growth alone can first rounded with the [urrows only sec- account for the original formation of pli- ondarily assuming a slit-like appearance cations in the leaves of palms' In this as the plications expand within a confined respect palm leaf development is similar (Figs.9-ll). This is more dramat- space of compound leaves of in Rhapis Whe.n to deveiopment ically illuitrated 1wo differential thel other flowlring plants in which adiacenlplications are frrst initiated' qrowth (in the iense of localized areas of are separatedby a broad shallowdepres- 48 PRINCIPES [VoL. 28

meristematic activity) also results in the 1926. Dei Gestaltungsverhdltnisseder formation of leaflets from free mareinal Palmenbl5tter.Ann. Jard. Bot. Buitenzorg.36: 16t 185. lobes.The unique aspectof palm leaI mor- HoFMEISTER,W. 1868. AllgemeineMorphologie phogenesisis the location of this localized der Gewdchse.Verlag von Wilhelm Engelmann, growth in a submarginal position and the Leipzig. subsequent secondary separation of the Keplex, D. R. 1973. The problem of leaf mor- phology and evolution in the monocotyledons. plications to form individual leaflets. While Quart. Rev. Biol. 48: 437 457. our investigation led to a rejection of the N. G. DnNcrrn, ,qNo R. E. DoNcrnn. hypothesis of splitting as the mechanism 1982a. The mechanismof plication inception of plication inception, it is clear that some in palm leaves: problem and developmental sort of tissue separation must occur in the morphology.Can. J. Bot. 60:2939 2975. AND -. 1982b. The abscissionof leaflets from one another and mechanismof plication inception in palm leaves: from the unplicate strip of marginal tissue. histogeneticobservations on the palmate leaf of Future research on palm leaf morphogen- Rhapis excelsa.Can. J. Bot. 6O: 2999 3O16. esis must test our conclusions in a wider MonL H. vov. 1845. Yermischte Schriften BotanischesInhaltes. Miinchen. range of palm taxa and must also examine MooRE, H. E., JR. 1973. The major groups of these more complicated aspects of leaflet palms and their distribution. Gentes Herbarium separation which are among the most dis- Ll:27 I4I. tinctive features of morphogenesis in this NAUMANN,A. 1887. Beitriige zur Entwicklungs- 193- unique plant group. geschichteder Palmenbldtter.Flora 70: 202, 209 2r8, 227-242, 250 257. PeouaNanHeN, D. 1963. Leaf developmentin LrrnReruneCrrno palms.Curr. Sci. 32: 537-539. 1967. Direct evidencefor schizogenous CoRNER,E. J. H. 1966. The Natural History of splitting in palm-leaflamina. Curr. Sci. 36: 467 Palms.Univ. of CaliforniaPress, Berkeley. 468. "split- DsrNtce, V. I898. Beitragezur Kenntnisder Ent- aNo S. Vren.lseuv. 1973. Late wicklungsgeschichtedes Blattes und der Anlage . ting" in the juvenile leaf of Phoenix dactylifera der Gefdssbiindel.Flora 85: 439 498. L. Curr. Sci. 42: 470-472. DENGLER,N. G., R. E. DeNcrrn, eNo D. R. Kep- PrRreseuv, K. I962. Morphologicaland ontoge- LAN. 1982. The mechanism of plication netic studiesin palms. I. Developmentof the inceptionin palm leaves:histogenetic observa- plicate condition in the palm leaf. Phytomor- tions on the pinnate leaf of Chrysalidocarpus phologyl2: 54-64. lutescens.Can. J. Bot. 6O: 2976 2998. VtNreteNenevANA, G. 1957. On certain aspects Eeurs, A. J. 1953. Neglectedmorphology of the of the developmentof the leaf of Cocos nucifera palm leaf. Phytomorphology3: I72 IB9. L. Phytomorphology7 : 297 -3O5. CoEBEL,K. 1884. VergleichendeEntwicklungs- YAMPoLSKY,C. 1922. A contributionto the study geschichteder Pflanzenorgane.Schenk's Hand- of the oil palm Elaeis guineensis Jacq. Bull. buch der Botanik (Breslau)3:99;432. Jard. Bot. Buitenzorg.Ser. 3, 5: 107 I74.