Plant Ecology and Evolution 144 (1): 44–63, 2011 doi:10.5091/plecevo.2011.436 REGULAR PAPER Morphology and development of spikelets and lowers in Cyperus and Pycreus (Cyperaceae) Alexander Vrijdaghs1*, Marc Reynders2, A. Muthama Muasya3, Isabel Larridon2, Paul Goetghebeur2 & Erik F. Smets1,4 Laboratory of Plant Systematics, Institute of Botany and Microbiology, K.U. Leuven, Kasteelpark Arenberg 31, P.O.Box 02437, BE-3001 Leuven, Belgium 2 Research Group Spermatophytes, Department of Biology, Ghent University, K. L. Ledeganckstraat 35, BE-9000 Ghent, Belgium 3 University of Cape Town, Department of Botany, Private Bag, 7700 Rondebosch, South Africa 4 Netherlands Centre for Biodiversity Naturalis (section NHN), Leiden University, P.O.Box 9514, NL-2300 RA Leiden, the Netherlands *Author for correspondence: alexander.vrijdaghs@bio.kuleuven.be 1 Background – Pycreus, Kyllinga, and Queenslandiella cluster together with Cyperus within the Cyperus s. lat. clade, one of the two large clades in Cypereae. However, in contrast with Cyperus, they have laterally lattened pistils/nutlets. Pycreus, Kyllinga and Queenslandiella form morphologically well circumscribed independent genera. In the context of a broader systematic project to work out a well supported, evolution based taxonomy for Cyperus s. lat., we present in this paper general morphological and developmental data of species of Pycreus in comparison with three species of Cyperus, including C. laevigatus with dorsiventrally lattened nutlets. Approach – Freshly collected material was investigated using scanning electron microscopy (SEM) and light microscopy (LM). Special attention was given to spikelet and gynoecial development. Results – SE micrographs of all species studied show an indeterminate rachilla with distichously arranged glumes, each subtending a bisexual lower. In spikelets of C. capitatus and P. pumilus, the proximal glume sometimes subtends a lateral spikelet instead of a lower. In the species of Pycreus studied, each lower sits in a cavity formed by the growth of the rachilla, which is congenitally fused with the wings of the glume of the higher, alternate lower. Glumes appear successively, each soon forming a lower primordium in its axil, which develops according to a general cyperoid ontogenetic pattern. In Pycreus, the stigma branches grow out from dorsiventrally positioned primordia. During gynoecium development, a hypogynous stalklet (gynophore) appears in all species studied. Conclusion – In spikelets of Pycreus, the rachilla and wings of the glumes are congenitally fused and consequently develop with epicaulescent displacements of the glumes resulting in typical spikelets with lowers in cavities. In spikelets of Cyperus, a similar though less pronounced development results in spikelets with zigzagging rachilla. The particular positions of the stigma branches in C. laevigatus and Pycreus are explained by the development of the gynoecium from an annular primordium, which facilitates shifts in localisation of the stigma primordia. Though we consider the combination of the typical spikelet ontogeny and the independently originated laterally lattened nutlets to be strong arguments in favour of a genus Pycreus, a phylogenetic conirmation that the taxon is monophyletic is an absolute, until now unfulilled, condition. Moreover, the consequences for the giant genus Cyperus must be taken in consideration. Key words – Cypereae, Cyperus, Cyperus s. lat., laterally lattened dimerous gynoecium, loral ontogeny, Pycreus, scanning electron microscopy, spikelet. INTRODUCTION Taxonomical data of Cyperus s. lat. According to molecular phylogenetic studies in Cyperaceae (Muasya et al. 2009a), the subfamily Cyperoideae comprises most of the cyperaceous genera, including the derived Cypereae clade (corresponding to Cypereae sensu Goetghebeur 1998). Within this clade, Cyperus and allied genera, called Cyperus s. lat., form a subclade that is sister to a Hellmu­ thia­Scirpoides­Isolepis­Ficinia clade (Muasya et al. 1998, 2001b, 2009a, Simpson et al. 2007, ig. 1). Based on the em- All rights reserved. © 2011 National Botanic Garden of Belgium and Royal Botanical Society of Belgium – ISSN 2032-3921 Vrijdaghs et al., Floral ontogeny in Cyperus and Pycreus (Cyperaceae) bryological study of Van der Veken (1964) and corroborated by more recent molecular phylogenetic studies (Muasya et al. 2002, 2009a, 2009b, Simpson et al. 2007), several smaller satellite genera appear to be nested within the Cyperus s. lat. clade, such as among others, Kyllinga, Queenslandiella and Pycreus. Each of these is characterised by specialised inlorescence and lower morphologies. Kyllinga can be distinguished by its reduced spikelets and lowers with laterally lattened ovaries, Pycreus by lattened spikelets and lowers with laterally lattened ovaries, and Queenslandiella by dehiscent spikelets (formerly placed in Mariscus) and lowers with laterally lattened ovaries (Goetghebeur 1986). In 1998, Goetghebeur wrote: “Pycreus and Kyllinga, plus some highly specialized smaller taxa are often excluded [from Cyperus s. lat.] and recognized at the generic level. Authors who include these taxa into Cyperus s. lat. mostly maintain them on the subgeneric level.” Moreover, the more derived part of Cype­ rus s. lat., including C. capitatus, C. laevigatus, and Pycreus, consists of genera with C4 photosynthesis and Kranz anatomy (ig. 1). Since Kyllinga, Pycreus and Queenslandiella are not sister taxa, we hypothesize independent and multiple origin of the laterally lattened pistil. Inlorescence morphology in Cyperus s. lat. and Pycreus The inlorescence in Cyperoideae is a compound inlorescence, essentially a panicle of spikelets with the main axis called a culm. The ultimate branch in a cyperoid inforescence is always a lateral spikelet, consisting of a rachilla and spirally to distichously placed glumes, each subtending (or not) a bisexual (most Cyperoideae) or unisexual (Cariceae) lower. Lateral spikelets are subtended by a bract and have a prophyll (Goetghebeur 1998). Terminal spikelets end the culm or a branch of it as a (co)lorescence sensu Troll (1964; see Weberling 1992), and as a consequence it is separated from its prophyll by the length of the culm/lateral branch, which Figure 1 – Simpliied cladogram of Cypereae based on Muasya et al. (2009a). In dark grey, taxa of which species were used in this study. Cyperus luzulae is a C3 species, whereas C. capitatus and C. laevigatus are C4 species. 45 Pl. Ecol. Evol. 144 (1), 2011 E A B C D F H G L 46 J I M N K Vrijdaghs et al., Floral ontogeny in Cyperus and Pycreus (Cyperaceae) constitutes also the axis of the terminal spikelet (Haines & Lye 1983, Goetghebeur 1998). In Cyperus s. lat., spikelets are distichously organised, which can be considered as a synapomorphy though a number of reversals to terete spikelets are recorded (Muasya et al. 2001b). In Pycreus, the inlorescence is antheloid with as well the culm as lateral branches ending with a terminal spike of lattened spikelets. In Pycreus, the inlorescence is antheloid with as well the culm as lateral branches ending with a terminal spike of lattened spikelets. Vrijdaghs et al. (2010) showed that cyperoid spikelets, including several, mostly distichously organised controversial ones that by some authors were interpreted as sympodial (e.g. Celakovsky 1887, Kern 1962, Zhang et al. 2004), have an indeterminate rachilla and can be considered to be an open spike as cited by Weberling (1992). Guarise & Vegetti (2008: 41) reported that in Cyperus, section Luzuloidei, fascicles of spikelets occur, “which can be serial, prophyllar, or mixed”. Serial fascicles of spikelets are mainly found in the distal part of the lorescence and paraclades, the latter being a repetition of the main inlorescence’s structure. A serial fascicle of spikelets is subtended by a single bract. Guarise & Vegetti (2008: 55) also mentioned a ‘torsion’ within the spikelets: “some spikelets appear with the glumes in the same plane as the pherophyll and prophyll, or in an intermediate position”. Several species in Cyperus s. lat., formerly grouped together in Mariscus, have dehiscent spikelets. Haines (1967: 57) reported a ‘pulvinus’ or swelling body at the base of lateral spikelets in Cyperus tenuis Sw., stating “But at the attachment of the prophyll, and probably a part of the prophyll, is a pulvinus which adjusts the position of both the branch and the umbel bract that subtends it”. Haines & Lye (1983: 17) mentioned “a callus is developed at the prophyll base, swelling of this callus causing divergence of the shoot”. Floral morphology and development in Cyperus s. lat. and Pycreus Flowers in Cyperoideae either have a perianth (3 + 3 parts or less) of varying size and shape or lack a perianth as observed in most species of Cypereae (e.g. Goetghebeur 1998, Muasya et al. 2009b). The androecium in most Cyperoideae is haplostemonous with usually three stamens with basiix and introrse anthers (Bruhl 1991, Vrijdaghs et al. 2005a), resulting from the reduction of the inner staminal whorl (Takh- tajan 1997). However, particularly in Cyperus s. lat., the number of stamens can be reduced to two or one (Haines & Lye 1983). In the irst developmental stages in cyperoid lowers, the stamens grow faster than the gynoecium (Vrijdaghs et al. 2005a), but at maturity of the lower, the stigma branches usually are functionally active before the pollen grains are released (Goetghebeur 1998). In many species of Cyperus, an apiculus or connective crest is formed on the top of the anthers (Haines & Lye 1983). The pistils in lowers within the Cyperus s. lat. clade vary from triangular with three stigma branches to dorsiventrally or laterally compressed with only two stigma branches. Raynal (1966) studied some African Cyperus species (e.g. C. meeboldii Kük., C. clavinux C.B.Clarke, C. lateriticus Raynal) with triangular nutlets and a single stigma branch. Most species with a dorsiventrally lattened pistil were often classiied in a separate taxon (Juncellus) by several authors (e.g. Clarke 1893, Kükenthal 1936, Podlech 1960). Already Clarke expressed some doubt: “This species [Juncellus pustulatus] has differentiated itself into Juncellus, but has not broken its connection with Cyperus entirely yet.” (Clarke 1901: 308). In some former Juncellus species, even within single specimens, the lowers can have both trigonous and dorsiventrally compressed nutlets (e.g. Cyperus alopecuroides Rottb., C. pustu­ latus Vahl, C. pygmaeus Rottb.). The polyphyletic dispersion of the Juncellus species was conirmed by many other authors (e.g. Goetghebeur 1986, Muasya et al. 2002). Therefore, a separate genus Juncellus is no longer recognised. Moreover, dorsiventrally lattened pistils can also be found in diverse other cyperoid genera such as Dulichium arundinaceum (L.) Britton, Eleocharis, Fimbristylis, Nemum, and Carex. On the other hand, laterally compressed pistils are restricted to three genera, Pycreus, Kyllinga and Queenslandiella. Blaser (1941) showed that the laterally lattened pistil in Pycreus concurs with new vascular patterns. Several authors based the subdivisional classiication within Pycreus among others on the morphology of the fruit wall epiderm cells (e.g. Clarke 1897, Chermezon 1919, Kükenthal 1936). Clarke (1897: 155) described the epiderm cells of nutlets in his “Zonatae” as follows: “Supericial cells of the nut longitudinally oblong; nut often appearing zonate by reason of the narrow ends of the cells running into an undulating or broken horizontal line.” In several species, these cells contain silica bodies, though ◄ Figure 2 – Cyperus luzulae, SE micrographs of loral ontogeny. A, lateral view of the rachilla apex, with six distichously placed glumes at successive developmental stages (numbered 1–6 from young to older; ‘1’, ‘3’ and ‘5’ show the wings of alternately positioned glumes); B, lateral view of glume 2 and a lower primordium in its axil; C, apical view of developing ovary wall surrounding a central ovule primordium with two adaxial and one abaxial stigma primordium; D, lateral view of a developing bractless spikelet belonging to a spikelet fascicle subtended by a common bract (not visible here) with lowers at successive developmental stages (encircled) and numbered from 1 (distal lower) to 7 (proximal lower). Arrows shows the wings of glume 4, which is also visible as the glume protecting the rachilla apex. The main axis, indicated as rachis, actually belongs to another, older spikelet in the fascicle; E, lateral view of a developing gynoecium. A single style appears (arrowed); F, lateral view of a developing ovary, with three stigma branches becoming papillose (encircled); G, lateral view of a part of a spikelet. Proximally, a developing lower with elongating stamen, and a stigma branch protruding above it (arrowed). At right hand side a glume with a wing enveloping the rachilla and a part of the stamen of the alternate lower; H, apical view of a part of a spikelet with removed glumes and stamens (arrowed); I, apical view of the distal part of a spikelet, with some glumes removed (arrows indicate the wings of the glumes); J, adaxial view of a developing lower; K, detail of developing style and stigma branches; L, nutlet with gynophore (arrowed); M, detail of apical part of a glume, with numerous stomata (encircled) and prickles (arrowed); N, detail of prickles. Abbreviations: a, anther; F, lower primordium; f, ilament; fa, loral apex; G, glume; nu, nutlet; o, ovule primordium; ov, ovary wall (primordium); R1, rachilla; Ra, rachis; s, stamen primordium; sg, stigma (primordium); st, style; *, rachilla apex. 47 Pl. Ecol. Evol. 144 (1), 2011 A B D E H C F J I K M 48 G L N O P Vrijdaghs et al., Floral ontogeny in Cyperus and Pycreus (Cyperaceae) according to Metcalfe (1971), they have little taxonomical value. Aims This study represents the irst results in a broader project in which the Cyperus s. lat. clade is investigated in analogy with our earlier study of the Hellmuthia­Scirpoides­Isolepis­ Ficinia clade, which resulted in several publications (Muasya et al. 2009a, 2009b, Vrijdaghs et al. 2005b, 2006a, 2006b, 2009). By combining molecular phylogenetic data, anatomy, morphology and spikelet/loral ontogeny, our goal is to clarify the evolution of Cyperus s. lat. and the position of the so-called ‘satellite genera’ within it. In this paper, we present and discuss original SEM and LM images of the morphology and spikelet/loral development in species of Pycreus, which were selected based on our preliminary phylogenetic data and compared with two Cyperus C4 and one Cyperus C3 species (respectively C. laevigatus, C. capitatus, and C. luzulae), starting from the hypothesis that Pycreus can be considered to be a genus of its own. Of the three genera with laterally compressed pistils, the mainly African genus Pycreus (±120) was chosen to be examined irst, because it is the largest one. MATERIALS AND METHODS Plant material Inlorescences of the species studied were collected in the ield and at the Ghent University botanical garden (table 1) and subsequently ixed in FAA (70% ethanol, acetic acid, 40% formaldehyde, 90/5/5). Spikelets and loral buds were dissected in 70% ethanol under a Wild M3 (Leica Microsystems AG, Wetzlar, Germany) stereo microscope equipped with a cold-light source (Schott KL1500; Schott-Fostec LLC, Auburn, NY, USA). Since in Cyperus s. lat. most spikelets have many and a variable amount of lowers, and consequently in order to avoid the use of abstract numbers, (lower subtending) glumes are numbered from young (1) to old (x). Scanning electron microscopy (SEM) Table 1 – Species of Cypereae (Cyperaceae) studied and voucher data. taxa Cyperus capitatus Poir. Cyperus laevigatus L. Cyperus luzulae Rottb. Pycreus bipartitus C.B.Clarke idem (ig. 12) Pycreus lavescens (L.) P.Beauv. ex Rchb. idem (ig. 12) Pycreus pelophilus (Ridl.) C.B.Clarke idem (ig. 12) Pycreus poly­ stachyos subsp. holocericeus (Rottb.) P.Beauv. idem (ig. 12) Pycreus pumilus (L.) Nees idem (ig. 12) Pycreus sanguinolentus (Vahl) Nees collector and origin Goetghebeur, Sep. 2004, HBUG 2003-1782 (w) Goetghebeur, Sep. 2004, HBUG1997-1237 Reynders, Nov. 2007, HBUG2003-1192 Vrijdaghs, HBUG1900-3306 Reynders, Nov. 2004, HBUG 2003-0327 (s) Laegadr, Ecuador Reynders, Jul. 2007, HBUG2005-0401 (s) voucher PG10744 PG10202 AV05 GENT101015 Muasya, 2005, Kenya Musili, 2005, Kenya AM2585 PM029 Reekmans, Burundi Reynders, Jul. 2007, HBUG 2006-1258 (w) GENT2547 Lewalle, Burundi Muasya, 2005, Kenya GENT6290 AM2150 Reekmans, Burundi Reynders, Jul. 2007, HBUG2006-1753 (w) GENT5795 The prepared material was washed twice with 70% ethanol for 5 minutes and then placed in a mixture (1/1) of 70% ethanol and DMM (dimethoxymethane) for 5 minutes. Subsequently, the material was transferred to 100% DMM for 20 min, before it was CO2 critical point dried using a CPD 030 critical point dryer (BAL-TEC AG, Balzers, Liechtenstein). The dried samples were mounted on aluminium stubs using ◄ Figure 3 – Cyperus capitatus, SE micrographs of loral ontogeny. A, apical view of the rachilla apex, and two lower primordia (arrowed) at early stages of development; B, differentiating lower primordium with three stamen primordia and a loral apex; C, annular ovary primordium surrounding a central ovule primordium (encircled), and three stamen primordia; D, idem as in ‘C’, with the ovary wall growing up from the base; E, position of a lower at early developmental stage with respect to the rachilla; F, apical view of developing lower, with the ovary wall enveloping the ovule (two adaxial stigma primordia and an abaxial one appear); G, lateral-adaxial view of developing lower (lateral stamen is removed); H, adaxial view of a developing lower, with one lateral stamen removed. The four (!) stigma primordia are growing out (encircled); I, adaxial view of a developing lower (arrow indicates single style); J, developing gynoecium and a single stamen, with stigma branches protruding high above the stamen (encircled); K, developing stamen before the elongation starts, with apiculus (left upper corner inset) and papillose cells at the bases of the pollen sacs (right hand side inset); L, elongated, withered stamen, with spiralised anther; M, ovule with obturator hairs covering the micropyle (arrowed); N, nutlet, with withered style still present; O, distal part of a culm, in the transition zone between lorescence and lateral branches. These are spiro-tristichously positioned and each subtended by a bract, whereas in the terminal spikelet (lorescence), the glumes are distichously arranged. This explains the position of the proximal glume-like bracts subtending a rudimentary spikelet; P, middle-apical part of spikelet with two developing lowers (encircled) and the wings of the glumes of the higher, opposite lower (arrows). Abbreviations: a, anther; f, ilament; fa, loral apex; G, glume; nu, nutlet; o, ovule (primordium); ov, ovary wall (primordium); Rl, rachilla; s, stamen primordium; sg, stigma (primordium); st, style; *, rachilla apex. 49 Pl. Ecol. Evol. 144 (1), 2011 A F D C E G I H J L 50 B K M N O Vrijdaghs et al., Floral ontogeny in Cyperus and Pycreus (Cyperaceae) Leit-C and coated with gold with a SPI-ModuleTM Sputter Coater (SPI Supplies, West-Chester, PA, USA). Images were obtained on a Jeol JSM-6360 (Jeol, Tokyo) at the Laboratory of Plant Systematics (K.U. Leuven). Light microscopy (LM) Samples were prepared in ethanol 70% and subsequently gradually transferred to ethanol 100%. Then, the samples were transferred to LR White Resin, hard grade (London Resin Company Ltd, Reading, England) in a graded LR White Resin/ethanol series using solutions of 25/75, 50/50, 75/25, 100/0 resin/ethanol 100% for at least 5 h each. Next, the samples were placed in a closed capsule illed with fresh resin, and hardened at 60°C during 48 h. Sections of 2 µm were made with a rotation microtome (Microm HM360 Waldorf, Germany) and subsequently stained with 0,1% toluidine blue. The stained sections were ixed on microscopy slides using Eukitt© quick hardening mounting medium (Fluka Chemie GmbH, Switzerland). Observations were done with a light microscope (Leitz Dialux 20, Van Hopplynus, Brussels, Belgium) equipped with a camera (PixeLINK PL-B622CF, Ottawa, Canada) with specially developed software (Microscopica v1.3, Orbicule, Leuven, Belgium). RESULTS The development and morphology of spikelet and lower in Cyperus and Pycreus are described below. Cyperus: spikelet structure In all species studied, the spikelet consists of a open axis (rachilla) and many distichously arranged glumes, each subtending a bisexual lower (igs 2A, 3A, 4A & B). Glumes develop fast, the older glumes not only protecting the lower they subtend, but also the apical part of the spikelet (igs 2A, B & D, 3A, 4A, B & M). Mature glumes have lateral wings, which partially envelop the rachilla and alternate, lower lower (igs 2D & I, 3P, 4A, C & L, ig. 13). The basal part of glume and wings is congenitally fused with the rachilla (ig. 13B–E). In C. luzulae, mature glumes have conspicuous prickles at the distal side, as well as high numbers of stomata (ig. 2M & N). Cyperus: loral ontogeny A new glume originates below the rachilla apex, forming a rim-like primordium (igs 2A & D, 3A, 4A–C). Soon, a lower primordium appears in the axil of the glume. The lower primordium expands laterally, forming a stamen primordium at each side, followed by a third abaxial one (igs 2A, B & D, 3A & B, 4B–D). In C. luzulae, usually there is a single, lateral stamen primordium (ig. 2A, B & D). Simultaneously, the loral apex becomes convex (igs 2B & D, 3B, 4D) and starts differentiating into an annular ovary primordium surrounding a central ovule primordium (igs 3C–E, 4E). Subsequently, the ovary wall grows up from the base, enveloping the ovule (igs 2C & D, 3C–E, 4E). On its top, one abaxial and two adaxial stigma primordia appear (igs 2C & D, 3F & G). The stigma primordia grow out into three papillose stigma branches (igs 2D–F & K, 3G–J). In C. laevigatus, only two laterally positioned stigma primordia appear, which results in a dorsiventrally lattened ovary (ig. 4F–L & O). In C. capitatus, samples with four stigma branches occur (ig. 3H). Meanwhile, the ovary wall continues its growth, forming a single style (igs 2E & F, 3I & J, 4I–K). Simultaneously with the development of the ovary, the stamen primordia differentiate into ilament and anther (igs 2G, 3F–J, 3P, 4F–H). Until this stage, the development of the stamens is as fast as or faster than the development of the pistil (igs 2G, 3G–I, 3P, 4E–H). However, at the later loral developmental stages style and stigma branches elongate faster, so that eventually they protrude above the stamens and even the glume (igs 2J, 3J, 4J–M). Meanwhile, the base of each pollen sac becomes papillose (igs 2J, 3K, 4K), and on the top of the anther an apiculus is formed (igs 3K, 4K). The ovule primordium develops into an anatropous bitegmic ovule, and within the locule, in a zone around the micropyle, hairs appear (igs 3M, 4N). In C. capitatus, the anther of the mature stamen becomes spiralised (ig. 3L). The nutlets of C. laevigatus and C. luzulae have a hypogynous stalklet, also called gynophore (igs 2L, 4O). The nutlet in C. laevigatus is dorsiventrally lattened (igs 4O, 12I). Pycreus: spikelet structure The spikelet in all species studied consists of an indeterminate rachilla and many distichously arranged glumes, each subtending a bisexual lower (igs 5A & B, 7A, 8A & B, 9A & ◄ Figure 4 – Cyperus laevigatus, SE micrographs of loral ontogeny. A, lateral view of a spikelet apex with lower subtending glumes at successive stages of development, numbered ‘1’ (youngest) to ‘6’ (oldest). The wings of each glume envelop partially the alternate, lower lower (arrowed); B, apical-abaxial view of spikelet apex with developing glumes; C, detail of very young glume subtending a lower primordium, and a wing of the alternate, higher glume (arrowed); D, differentiating lower primordium with three stamen primordia, and a part of the loral apex; E, developing lower with ovary wall growing up, and three stamen primordia beginning to differentiate; F, apicalabaxial view of a developing lower. Two laterally positioned stigma primordia are growing out on the top of the ovary wall, which envelops the ovule. Filaments and anthers are well developed; G, apical view of developing lower. The two stigma primordia are growing out; H, apical view of a transversally cut spikelet, with two alternating lowers at intermediary developmental stages (encircled); I, abaxial view of a developing lower; J, adaxial view of a developing lower; K, detail of a developing stamen, with apiculus (arrowed); L, abaxial view of a developing lower (encircled) in a tranversely cut spikelet. The subtending glume is removed. The wings of the higher, opposite lower can be seen (arrowed); M, apical part of a spikelet, with several, distichously placed glumes, and protruding style branches; N, lateral view of an ovule, with funiculus (black line) and obturator hairs covering the micropyle (arrowed); O, dorsiventrally lattened nutlet with a hypogynous stalklet or gynophore (arrowed). Abbreviations: a, anther; F, lower primordium; f, ilament; G, glume; nu, nutlet; o, ovule primordium; ov, ovary wall (primordium); Rl, rachilla; s, stamen primordium; sg, stigma (primordium); st, style; *, rachilla apex. 51 Pl. Ecol. Evol. 144 (1), 2011 C A B D F G H K 52 E I L J M N Vrijdaghs et al., Floral ontogeny in Cyperus and Pycreus (Cyperaceae) B, 10A & B). The basal part of glume and wings is fused with the rachilla (ig. 13F–I). At later stages, each lower stands in an alcove-like cavity (igs 5B, 5J, K & M, 6A, C & E, 7B & C, 8A & B, 9B & C, 10B & C, ig.13). In P. pumilus, a glumelike bract in proximal position alternating with the prophyll subtends a lateral axis, which develops into a spikelet (ig. 5A & J), instead of a lower. In P. pelophilus, P. polystachyos and P. sanguinolentus, an adaxial swelling body can be seen at the base of the spikelet (igs 6G, 7H, 9L). The spikelet of P. polystachyos has a long irst internode or epipodium, which is enveloped by the tubular sheath of the spikelet prophyll (ig. 7G & H). The developing glumes at the apical part of the spikelet envelop the apex of the rachilla with a bonnet-like mucro (igs 5B, 7A, 9A). Pycreus: loral ontogeny Glumes originate below the indeterminate spikelet apex (igs 5A & B, 7A, 8A & B, 9A & B, 10A & B), forming a rim-like primordium, of which the edges partially envelop the alternate, lower lower primordium (igs 5C, 5H & J, 7B, 8B). Soon after the formation of a new glume primordium, a lower primordium appears in its axil (igs 5B & C, 7B, 8B, 9B & C, 10B). With the glume developing, the lower primordium expands laterally, forming two lateral stamen primordia, followed with some delay by a third abaxial one (P. lavescens, P. sanguinolentus; igs 9C–E, 10B & C). In species with lowers with two stamens, no abaxial stamen primordium is formed (P. bipartitus, P. pelophilus, P. polystachyos; igs 7B, 8B–D). In lowers with only one stamen, the lower primordium expands laterally, forming only one stamen primordium (P. pumilus; ig. 5B–D). Simultaneously with the formation of the stamen primordia, a loral apex appears (igs 8C, 9D). Next, the loral apex differentiates into an annular ovary primordium, surrounding a central ovule primordium (igs 5E, 6A, 8D, 9E & F, 10C & D). The ovary wall primordium grows up from the base, gradually enveloping the central ovule (igs 5F & H, 6A & B, 8D & E, 9F, 10E). At this stage, on the top of the ovary wall two dorsiventrally positioned stigma primordia appear (igs 5G, I & J, 6C, 7C, 8E, 9G, 10F). The development of the adaxial stigma primordium is sometimes slightly delayed with respect to the abaxial one (igs 8E & F, 9G–I, 10F–H, 11A & B). In P. bipartitus, the early adaxial stigma primordium sometimes splits (igs 8G & H). Subsequently, the growing ovary wall develops a single style without distinct style base, while the stigma primordia grow out into two papillose stigma branches (igs 5G & K, 6D–F, 7D & E, 8G & H, 9I, 10G & H, 11A & B). In P. sanguinolentus, at this stage, an annular constriction appears in the apical part of the ovary (ig. 9K). Meanwhile, the stamen primordia have developed into introrse stamens with basiixed anthers with longitudinal slits (igs 5G, H, I & K, 6C–F, 7C–F, 8E–G, H & J, 9G & H–J, 10E & F, 11A & B). In semi-mature lowers of P. lavescens, and P. sanguinolentus, the anthers are as long as or longer than the ilaments (igs 9J, 11A & B), whereas in P. bipartitus, P. pelophilus, P. polystachyos and P. pumilus the anthers are relatively short with respect to the ilaments (igs 5K–N, 6D–G, 7E & F, 8J & K). In P. pelophilus and P. pumilus, a short connective stalklet appears between ilament and anther (igs 5M, 6G & H). In all species studied, the cells at the base of the pollen sacs in developing anthers become more or less papillose (igs 5M & N, 6H, 7F, 8G & K, 9J, 10A & B). An apiculus is absent or remains rudimentary, with the apical cells becoming papillose (e.g. in P. bipartitus. ig. 8H). Maturing gynoecia and nutlets have a hypogynous stalklet or gynophore (e.g. igs 7I, 8K, 11C & D, 12A, C, E & G–I). In P. polystachyos and P. pumilus, the cells of the nutlet wall each contain a conspicuous tabular silica body (igs 7I & J, 12A–F), in P. polystachyos often with microsatellites around its top. In P. pelophilus, similar cells only occur in the center of each lateral side (ig. 12C & D). In P. lavescens, the epidermal cells of the mature nutlet become longitudinally elongated (zonate cells) pushing up the transverse cell walls, which gives the nutlet its typical wrinkled appearance (ig. 11D & E, 12H). ◄ Figure 5 – Pycreus pumilus, SE micrographs of loral ontogeny. A, lateral view of a branched spikelet with proximally a glume-like bract subtending a secondary spikelet (encircled). The prophylls (P) of the main and secondary spikelet are parallel to each other. All visible lowers have a single stamen; B, detail of a spikelet apex with 11 glumes, each subtending a lower (primordium), numbered 1–11 from young ‘1’ to older ‘11’. In lower 11, the wings (arrowed) of the opposite, higher glume (number 10, only partially visible) form the walls of an alcove-like cavity in which the lower develops; C, detail of a young glume with lower primordium. At the right hand side, the wing of the alternate, superior glume is visible (arrowed); D, differentiating lower with primordia of stamen and ovary wall; E, lower with developing stamen and early gynoecium. The ring primordium of the ovary wall surrounds the central ovule primordium; F, developing lower with ovary wall growing up from the base, and stamen with distinct ilament and anther; G, lateral-abaxial view of part of a spikelet with three lowers at different developmental stages. In the middle lower, two dorsiventrally oriented stigma primordia appear on the top of the ovary wall (encircled). In the lower lower, the ovary wall entirely envelops the ovule, a single style appears, and the stigma primordia are growing (encircled); H & I, detail of the development of the ovary and appearance of the dorsiventrally positioned stigma primordia, and simultaneously the development of the stamen; J, transversely cut proximal part of a main spikelet with proximally a glume-like bract, subtending a lateral spikelet. Alternately of it, the second glume can be seen, subtending a lower of which only the developing gynoecium is visible. This is partially enveloped by the wing (arrowed) of the third glume (removed together with the lower it subtends). This wing is fused with the rachilla of the main spikelet; K, lateral-abaxial view of a part of a spikelet with two lowers at developmental stages following on the developmental stage at ‘I’. In the lowest lower, a single style appears (arrowed); L, adaxial view of a developing gynoecium and a glume with a conspicuous mucro (encircled) subtending a lower; M, lateral view of a part of a spikelet. In the lowest lower, consisting of a gynoecium and two stamens, a ‘connective stalklet’ can be observed between ilament and anther (arrowed); N, lateral view of semi-mature lower with two stamens, protected by the wings of the alternate, higher glume. Abbreviations: a, anther; B, bract; F, lower primordium; f, ilament; G, glume; o, ovule primordium; ov, ovary wall (primordium); P, prophyll; Rl, rachilla; s, stamen primordium; sg or white dot, stigma (primordium); st, style; W, wing of glume; *, rachilla apex. 53 Pl. Ecol. Evol. 144 (1), 2011 B A C E F D G H Figure 6 – Pycreus pelophilus, SE micrographs of loral ontogeny. A, transverse section in the distal part of a spikelet, with two lowers at early developmental stages. In the lower below, two stamen primordia and an annular ovary primordium surrounding a central ovule primordium are visible. In the upper lower, the stamen primordia start differentiating into anther and ilament (not visible), and the annular ovary primordium grows up from the base; B, growing ovary wall enclosing the central ovule; C, apical-abaxial view of a developing lower and part of a tranverse section through the distal part of the rachilla. Two dorsiventrally positioned stigma primordia appear on the top of the ovary wall. In between the glume and the lower it subtends, two wings of the alternate, higher lower (removed) partially envelop the lower (arrowed). The wings are fused with the rachilla; D, abaxial view of a developing lower. A single style appears. The anthers are shorter than the ilaments; E, lateral view of the middle part of a spikelet with removed glumes. Two developing lowers are visible, each partially envelopped by the wings of the higher, opposite glume (arrowed); F, abaxial view of a semi-mature lower; G, lateral view of the proximal part of a spikelet, with spikelet subtending bract, prophyll of the spikelet, proximal glume and proximal lower (encircled), partially hidden by the wing of the next glume. At the base of the prophyll, a swelling body or pulvinus is visible (arrowed); H, detail of the connective stalklet (arrowed) in between ilament and anther. Abbreviations: a, anther; B, bract; co, connective; f, ilament; Fp, proximal lower primordium; G, glume; Gp, proximal glume; o, ovule primordium; ov, ovary wall (primordium); P, prophyll; ps, pollen sac; Ra, rachis; Rl, rachilla; s, stamen primordium; sg or white dot, stigma primordium; st, style; W, wing; *, rachilla apex. Anatomical data Cross sections were made at different levels through developing spikelets of Cyperus laevigatus (ig. 13A–E) and Pycreus lavescens (ig. 13F–I). Figure 13A serves as a key to symbols 54 for 13B–I. Cross sections at the basal part of a lower in C. laevigatus (ig. 13B–D) and P. lavescens (ig. 13F–H) reveal that glume and rachilla are fused below the level where the ilaments are clearly distinguishable. Cross sections at anther Vrijdaghs et al., Floral ontogeny in Cyperus and Pycreus (Cyperaceae) A B C D E H I G F J Figure 7 – Pycreus polystachyos, SE micrographs of loral ontogeny. A, rachilla apex and irst glume (encircled); B, transverse section in the apical part of the spikelet, showing a newly formed glume with wings (arrowed) partially enveloping the alternate, lower lower primordium. The glume subtending this lower primordium is removed; C, lateral view of a developing lower. The wings (arrowed) of the alternate, higher lower contribute to its protection. The ovary wall is enveloping the ovule, and two dorsiventrally positioned stigma primordia appear. The two stamen primordia are differentiating into ilament and anther; D, developing lower. A single style is formed, and the stigma primordia grow out into stigma branches; E, semi-mature lower. The anther becomes shorter than the ilament; F, lateral view of a part of a spikelet with two semi-mature lowers. The cells at the bases of the pollen sacs become papillose; G, entire spikelet, with a long irst internode enveloped by a sheath-like prophyll. At the base of the prophyll, a swelling body is present (encircled). Stigma branches protrude above the glumes (arrowed); H, detail of the irst internode (white bar) and spikelet prophyll, with a conspicuous swelling body. The spikelet subtending bract is removed. (I) Nutlet with hypogynous stalklet or gynophore; J, detail of the surface of the nutlet, with tabular silica-bodies. Abbreviations: a, anther; F, lower primordium; f, ilament; G, glume; Gp, proximal glume; nu, nutlet; ov, ovary wall (primordium); P, prophyll; Ra, rachis; Rl, rachilla; st, style; white dot, stigma (primordium); *, rachilla apex. level show a separate glume and rachilla (ig. 13E & I). In both species, a cross section through the rachilla at internode level is butterly-shaped (ig. 13E & I). DISCUSSION Spikelet development and morphology In all our observations presented here on Cyperus and Py­ creus, the spikelet consists of an indeterminate rachilla and numerous, acropetally developing glumes, each subtending a single lower. This concurs with our earlier observa- tions in a wide range of cyperoid species (Vrijdaghs et al. 2006a, 2007, 2010). In Cypereae, a lateral spikelet (which is deined as ‘ultimate branch’ and hence should not have any ramiication within it) is not always clearly distinguishable from a branched partial inlorescence; in some species, a secondary spikelet instead of a lower is formed in the axil of a glume (e.g. in Hellmuthia; Vrijdaghs et al. 2006b). This was also observed in Ficinia (Muasya, unpubl. res.), Cyperus (igs 2D, 3E), and Pycreus (ig. 5A & J). Therefore, in strict sense, in such cases a glume subtending a secondary spikelet should be called ‘glume-like bract’, and the rachilla of the 55 Pl. Ecol. Evol. 144 (1), 2011 C A E H D B F I G J K Figure 8 – Pycreus bipartitus, SE micrographs of loral ontogeny. A, apical view of a spikelet apex with glumes/lowers at different developmental stages, numbered from young ‘1’ to older ‘5’. Encircled is a proximal developing lower, with the ovary wall enclosing the ovule, and two developing stamens; B, spikelet apex with very young glume subtending a yet undifferentiated lower primordium. Arrows indicate wings of two superposed glumes at the other side of the spikelet; C, differentiating lower primordium with two lateral stamen primordia and a conspicuous loral apex; D, developing lower with the two stamens removed. The ovary wall is enveloping the central ovule; E, apical view of a developing lower. Two dorsiventrally oriented stigma primordia originate on the top of the ovary wall; F, lateral view of a developing lower. On the top of the anthers, an apiculus appears (arrowed); G, lateral view of a developing lower. A single style appears, with the stigma primordia growing out into stigma branches (encircled). The adaxial stigma primordium is split into two (arrowed); H, detail of stigma primordia with splitted adaxial one (encircled), and apiculus (arrowed); I, developing ovule with the micropyle nearly bent back over 180° (arrowed). The funiculus is indicated with a black line; J, mature lower. Stigma branches are encircled; K, lateral view of mature lower. The gynoecium/nutlet has a hypogynous stalklet or gynophore (arrowed). Abbreviations: a, anther; F, lower primordium; f, ilament; G, glume; o, ovule primordium; ov, ovary wall (primordium); s, stamen primordium; sg, stigma (primordium); st, style; te, outer tegument; ti, inner tegument; *, rachilla apex. 56 Vrijdaghs et al., Floral ontogeny in Cyperus and Pycreus (Cyperaceae) main spikelet ‘rachis’. In C. luzulae, spikelets belong to a serial fascicle of spikelets (ig. 2D), in which several spikelets originate in the axil of a common bract, the one above the other, as described by Guarise & Vegetti (2008: ig. 8). We also observed similar spikelet clusters in C. eragros­ tis Lam. (both belonging to the section Luzuloidei; Denton 1978). In all C3 species studied, the position of the spikelet prophyll of rather distally on the rachis positioned spikelets is shifted in comparison with the plane determined by the distichous arrangement of the spikelet’s other glumes. In C4 species, such torsion has not been observed. Similar observations are reported by Guarise & Vegetti (2008). In C. capitatus, the culm ends in a terminal spikelet (lorescence) in which the glumes are distichously placed. Below the lorescence, lateral branches, each subtended by a bract, are spiro-tristichously positioned (ig. 3O). This allows us to interpret the proximal glume-like bracts as bracts subtending a lateral spikelet, positioned out of the plane determined by the higher distichously placed glumes of the terminal spikelet. In the transition zone between lorescence and the lower part of the culm with lateral branches, primordia in the axil of a bract have a high lexibility to become lower or lateral axis. This lexibility to determine a given, yet undetermined primordium in the axil of a glume(-like bract) also explains the presence of secondary spikelets in spikelets of e.g. Pycreus pumilus (ig. 5A & J; Vrijdaghs et al. 2010). The glumes in all species studied are winged, with the wings of one glume partially enveloping the opposite, lower lower. In both Cyperus and Pycreus, the basal part of the glume including (part of) the wings is congenitally fused with the rachilla (igs 6C, 7B, 13) and grows up with it. This is most obvious in P. pumilus, whereas in P. pelophilus (ig. 6) and P. lavescens (ig. 10), a large part of the wings grows free from the rachilla. Consequently, the main part of the glume and the lower primordium it subtends are epicaulescently displaced to a more apical position (actually, until the next node) on the rachilla. As a result, the rachilla itself is winged along the common growth zone (Vrijdaghs et al. 2010). In Pycreus, this epicaulescent metatopic displacement is more pronounced than in species of Cyperus or other Cyperoideae, resulting in the typical alcove-like cavities along the rachilla, of which the lateral walls consist mainly of the wings of the opposite, higher glume (ig. 13). In Pycreus, the glumes often have a prolonged midvein or mucro, which becomes cap-shaped, protecting the rachilla apex (e.g. igs 5B, 7A, 9A). At the adaxial lower part of prophylls of both inlorescence branches and spikelets in P. pelophilus, P. polystachyos and P. sanguinolentus, an adaxial swelling body can be seen (igs 6G, 7H, 9L). We also observed it in other Cypereae, such as C. luzulae (Reynders, unpubl. res.) and Kyllinga Rottb. (Huygh, University of Ghent, Belgium, and Vrijdaghs, unpubl. res.). These observations allow conirming Haines’ (1967) suggestion that the swelling body or pulvinus is part of the prophyll. Floral ontogeny and fruit morphology The loral ontogenetic pattern in Cyperus and Pycreus is similar to the pattern observed by us in many other Cyperoideae (e.g. Vrijdaghs et al. 2005, 2009). However, there is no formation of perianth primordia, which is a common feature for Cyperus s. lat. and Cypereae. However, in the Ficinia­Iso­ lepis clade, two southern African species previously named as Scirpus (S. falsus and S. icinioides) were added, as well as the formerly mapanioid Hellmuthia. These three species are the only recorded Cypereae with remnants of a perianth (Simpson et al. 2003, Vrijdaghs et al. 2006, Muasya et al. 2009a, 2009b). In Pycreus, the number of stamens is highly variable, with basic number three as in most other Cyperoideae (e.g. igs 9D, E, G & H, 10C, E & F). Kükenthal (1936) reported that nearly half of the 72 species he recognized in Pycreus have a constant number of two stamens instead of three. In these cases it is the abaxial stamen that does not develop (e.g. igs 6, 8A–G). In some species, the number of stamens can also vary within the species, and even within a single plant (e.g. P. pumilus, ig. 5M). We observed a tendency to delay the formation of the abaxial stamen or to reduce it completely in various other cyperoid genera, such as Eriophorum, Scir­ poides (Vrijdaghs et al. 2005a), Fuirena (Vrijdaghs et al. 2004), Ficinia and Isolepis (Vrijdaghs et al. 2005b). From these observations, we may deduce that the reduction of the abaxial stamen occurred independently in different cyperoid clades. In all cases, this tendency can probably be explained by a limited spacial freedom to develop the three stamens. Pycreus pumilus, with its highly compacted spikelets and lowers with usually one, sometimes two stamens, clearly illustrates this. Moreover, in stamens of lowers of P. pelophilus and P. pumilus, a ‘connective stalklet’ appears in semi-mature stamens. Similar observations were made in other Cypereae (e.g. Kyllinga and Oxycaryum; Vrijdaghs, unpubl. res.). We admit that this ‘connective stalklet’ acts as an articulation allowing the anther to bend over for better pollen dispersal by the wind. As in all other Cyperoideae studied by us, the gynoecium in the species of Cyperus and Pycreus presented here are formed from an annular ovary primordium surrounding a central ovule primordium. Since the ovary wall in Cyperoideae is not resulting from a postgenital fusion of three distinct carpel primordia but growing up from an annular ovary primordium, new possibilities arise in organizing the vascularisation of the gynoecium and consequently also for its morphology, such as the positions and number of the stigmas. In Pycreus, only two stigma branches are formed, positioned dorsiventrally, which results in laterally lattened gynoecia/nutlets (igs 5–11). Similar pistils also occur in Kyllinga and Queenslandiella. However, molecular phylogenetic data (Muasya et al. 2009a) show that these genera form different clades within Cyperus s. lat., which suggests that this feature evolved independently in each of the three genera characterized by it. Also in C. laevigatus, only two, though laterally positioned stigma primordia appear, resulting in a dorsiventrally lattened pistil/ nutlet (ig. 4F–H). Goetghebeur (1986) suggested that such a pistil, wich also occurs in other Cyperus species and other cyperoid genera such as Blysmus, Dulichium, Eleocharis and Fimbristylis, results from the reduction of the abaxial carpel and a fusion of the two remaining adaxial carpels. However, each attempt to explain the Pycreus type pistil using the carpel concept fails. Moreover, in strict sense, carpels are not present in cyperoid Cyperaceae since the ovary originates 57 Pl. Ecol. Evol. 144 (1), 2011 B A D C E F G I H J 58 K L Vrijdaghs et al., Floral ontogeny in Cyperus and Pycreus (Cyperaceae) A C F B D E G H Figure 10 – Pycreus lavescens, SE micrographs of loral ontogeny. A, lateral view of a spikelet apex, with glumes/lowers at successive developmental stages numbered from young to older 1–3; B, rachilla apex with young glume subtending a lower primordium; C–D, differentiating lower primordium with one abaxial and two adaxial stamen primordia, and with the loral apex differentiating into an annular ovary primordium (arrowed) surrounding a central ovule primordium. In ‘D’, the stamen primordia start differentiating into ilament and anther; E, transverse section through the rachilla, with three lowers (1, youngest; 3, oldest) at different developmental stages. Flower ‘1’ is shown from an adaxial viewpoint, lowers ‘2’ and ‘3’ from an abaxial viewpoint, each with removed stamens. In lower ‘2’, the adhesion of the wings of the subtending glume of lower ‘1’ to the rachilla can be seen (encircled); F, apical view of a transverse section through the rachilla, with two lowers. The right hand one is less developed, with two dorsiventrally stigma primordia appearing on the top of the ovary wall (arrowed); G–H, apical view of a developing gyncoecium. A single style appears, and the stigma primordia grow out, the adaxial one (arrowed) delayed with respect to the abaxial stigma primordium. Abbreviations: a, anther; F, lower primordium; f, ilament; G, glume; o, ovule primordium; ov, ovary wall (primordium); s, stamen primordium; sg, stigma (primordium); W, wing; *, rachilla apex. ◄ Figure 9 – Pycreus sanguinolentus, SE micrographs of loral ontogeny. A, apical part of a spikelet, with glumes/lowers at successive developmental stages, numbered from young ‘1’ to older ‘6’; B, detail of rachilla apex with a young glume primordium with undifferentiated lower primordium. The wings of the alternate glume reach the underlaying lower (arrowed); C, detail of a glume and lower primordium, which is expanding laterally. The wings of the glume envelop partially the rachilla (arrowed); D, differentiating lower primordium, with two lateral and a slightly delayed abaxial stamen primordium, and a loral apex. E, developing lower. The loral apex is starting to form an annular ovary primordium (arrowed); F, developing ovary, with ovary wall enveloping the central ovule; G, apical view of a developing lower. Two dorsiventrally positioned stigma primordia are visible on the top of the ovary wall; H, apical-adaxial view of a developing lower. The wings (arrowed) of the opposite, higher lower (not in the image) envelop partially the lateral stamens; I, lateral view of a developing lower. A single style appears; J, lateral view of a developing lower. The stigma branches are growing out; K, lateral view of semi-mature lower, one lateral stamen is removed. An annular constriction around the apical part of the ovary is formed (arrowed); L, entire spikelet. At the base of the prophyll, a conspicuous swelling body or pulvinus is visible (arrowed). Abbreviations: a, anther; B, bract; f, ilament; fa, loral apex; G, glume; Gp, proximal glume; o, ovule primordium; ov, ovary wall (primordium); P, prophyll; Ra, rachis; Rl, rachilla; s, stamen primordium; sg, stigma (primordium); st, style; *, rachilla apex. 59 Pl. Ecol. Evol. 144 (1), 2011 A B D C E Figure 11 – Pycreus lavescens, SE micrographs of loral ontogeny. A–B, lateral-adaxial view of developing lower. In ‘B’, the delay of the development of the adaxial stigma branch diminishes; C, longitudinal section of a fruit wall with rests of the obturator hairs, and a hypogynous stalklet or gynophore (arrowed); D, nutlet; E, detail of the fruit wall. Abbreviations: a, anther; f, ilament; nu, nutlet; ov, ovary wall; sg, stigma (primordium); st, style. from an annular ovary primordium. We believe that the organizational freedom resulting from the congenital fusion of the carpels into an annular ovary primordium made laterally lattened nutlets like in Pycreus, as well as dorsiventrally lattened nutlets like in C. laevigatus, possible. accidents; the meristematic zones from which the stigma branches originate (we call them stigma primordia because they are not carpel tips, though we do not exclude that they are homologous with carpel tips) can be splitted (dédoublement). In P. bipartitus, at early developmental stages, two adaxial stigma branches can occur (ig. 8G). Haines & Lye (1983) also reported the presence of three stigma branches in some specimens of P. nigricans. It is tempting to interpret these observations as an argument to state that the adaxial stigma branch in Pycreus resulted from the fusion of the two ancestral lateral ones. However, how to explain the presence of four stigma branches in C. capitatus (ig. 3H)? Therefore, we consider these particular structures rather as developmental In P. lavescens, the development of the adaxial stigma branch at early developmental stages is slightly delayed with respect to the abaxial one (igs 10, 11A). This too might be explained by a temporary lack of space. In P. sanguinolentus, an apical constriction of the ovary appears at semi-mature stage. We observed a similar phenomenon in Fuirena ab­ normalis C.B.Clarke (Vrijdaghs et al. 2004). In maturing lowers of several Cyperus and Pycreus species, a rudimentary hypogynous stalklet or gynophore appears (e.g. igs 2E 60 Vrijdaghs et al., Floral ontogeny in Cyperus and Pycreus (Cyperaceae) A B C D E F G H I Figure 12 – SE micrographs of mature nutlets in Pycreus (A–H) and Cyperus (I). A, P. bipartitus, lateral view of a nutlet with a gynophore (arrowed); B, P. bipartitus. Detail of the fruit wall epidermis with cells with small conical silica bodies; C, P. pelophilus, lateral view of a nutlet with a gynophore (arrowed); D, P. pelophilus. Detail of the fruit wall epidermis with cells with each a tabular silica body; E, P. pumilus, lateral view of a nutlet with a gynophore (arrowed); F, P. pumilus. Detail of the fruit wall epidermis with small cells, each illed with a tabular silica body; G, P. sanguinolentus, lateral view of a nutlet with a hypogynous stalklet or gynophore (arrowed); H, P. lavescens, lateral view of a nutlet with a gynophore (arrowed). The epidermis consists of zonate (longitudinally elongated) cells; I, Cyperus laevigatus. Dorsiventral view of a nutlet with a gynophore (arrowed). Abbreviation: nu, nutlet. & L, 4O, 7I, 8K). This also occurs in other genera in Cypereae, such as Ficinia, Isolepis, and Scirpoides (Vrijdaghs et al. 2005a, 2006b). Fruit wall epiderm cells in Pycreus pelophilus, P. pumilus and P. polystachyos have (at least partially) similar, tabular silica-bodies (ig. 12C–F). The fruit wall epiderm cells in P. lavescens are zonate and do not have silica-bodies (ig. 11D & E, 12H). Pycreus bipartitus has fruit walls with isodiametric epiderm cells with small conical silica bodies (ig. 12A & B), which is also reported in P. sanguinolentus, though we did not observe this in nutlets from herbarium specimens (ig. 12G). According to Metcalfe (1971), only the few neither conical nor tabular silica-bodies found in some species might have systematic value. CONCLUSIONS The spikelet ontogeny and morphology in the Cyperus and Py­ creus species studied concurs with our observations in many other Cyperoideae that cyperoid spikelets consist of an indeterminate rachilla and many glumes which usually subtend (or not) a bisexual lower (Cariceae and sclerioid Cyperaceae not included). However, in Cypereae, proximal bracts of the spikelet may axillate a secondary spikelet. We consider this phenomenon to be a result of the lexibility plants have to activate different developmental patterns (to become a lower, a spikelet or a vegetative axis) in yet undetermined primordia. Spikelets in Cyperus s. lat. have a typical zigzagging morphology, resulting from a congenital fusion of the rachilla and the wings of the glumes, which causes epicaulescent growth 61 Pl. Ecol. Evol. 144 (1), 2011 A B C D E F G H I Figure 13 – A, Key for B–I; LM image of a cross section through a spikelet of Pycreus lavescens, at the height of the ilaments. The green coloured areas are the fusion zones between rachilla and wings of a glume. The section of this glume and the lower in its axil (three ilaments and centrally the gynoecium) are coloured in red. The wing tips are also coloured in red, and arrowed. The rachilla is coloured in yellow. In each section shown in igure B–I a similar glume with the lower it subtends can be observed, as well as a fusion zone of wings and rachilla; B–I, LM images of cross sections through the spikelet at different levels in Cyperus laevigatus (B–E) and Pycreus lavescens (F–I); B & F, cross sections through the basal level of a glume and its lower. In P. lavescens, the fusion of glume and rachilla (green arrow) is less complete than in C. laevigatus; C,D, G & H, cross sections at ilament level; E & I, cross sections at anther level (or internode). Here, the glume is free from the rachilla. Abbreviations: F, lower; G, glume; Rl, rachilla. of the glumes with the rachilla. The particular morphology of a spikelet in Pycreus results from a pronounced epicaulescent growth of the glumes with the rachilla. The loral ontogeny in all species studied occurs according to the general cyperoid loral ontogenetic pattern, though no perianth primordia are formed. The pistil, as it originates from an annular primordium, gets more organisational freedom, which is illustrated by the two dorsiventrally positioned stigma branches in Py­ creus, as well as the two laterally positioned stigma branches in species with dorsiventrally lattened nutlets, such as C. laevigatus. Only on condition that in cladistic analysis Py­ creus would appear as a monophyletic taxon, we think that the combination of 1) its particular spikelet ontogeny resulting in a ‘Pycreus-type’ spikelet, 2) the laterally lattened ovaries/nutlets which originated independently in the evolution from other taxa with similar ovaries, are strong arguments to consider this taxon to be a genus on its own. However, we also realise that this would make Cyperus paraphyletic. ACKNOWLEDGEMENTS We thank Nathalie Geerts and Anja Vandeperre for assistance with the LM work. This work was supported inancially by research grants of the K.U. Leuven (OT/05/35), Belgium, the Fund for Scientiic Research - Flanders (Belgium) (F.W.O., G.0268.04) and the Special Research Fund (BO5622, Ghent University), Belgium. We also thank the three anonymous reviewers of this paper. 62 REFERENCES Blaser H.W. (1941) Studies in the morphology of the Cyperaceae I. Morphology of lowers. A. Scirpoid genera. American Journal of Botany 28: 542–551. DOI: 10.2307/2437000 Bruhl J.J. (1991) Comparative development of some taxonomically critical loral/inlorescence features in Cyperaceae. Australian Journal of Botany 39: 119–127. DOI: 10.1071/BT9910119 Clarke C.B. (1893) In Hook F. Flora of British India 6(19): 594. Clarke C.B. (1897) Cyperaceae. In Thiselton-Dyer W.T. Flora Capensis 7,1: 1–192. Clarke C.B. (1901) Cyperaceae 2. In Thiselton-Dyer W.T. Flora of tropical Africa 8(2): 385–524. Reeve, London. Celakovsky L. (1887) Ueber die ährchenachtige Partial-inlorescenzen der Rhynchosporeen. Berichte der Deutschen Botanischen Gesellschaft 5(1): 148–152. Chermezon H. (1919) Un genre nouveau des Cyperacées. Bulletin du Musée National d’Histoire Naturelle (Paris) 25: 60–63. Denton M.F. (1978) A Taxonomic treatment of the Luzulae group of Cyperus. Contributions from the University of Michigan Herbarium: 11(4): 197–271. Goetghebeur P. (1986) Genera Cyperacearum. Een bijdrage tot de kennis van de morfologie, systematiek en fylogenese van de Cyperaceae-genera. PhD thesis, Groep Plantkunde, Rijksuniversiteit Gent, Ghent, Belgium. Goetghebeur P. (1998) Cyperaceae. In: Kubitzki K. (ed.) The families and genera of vascular plants. IV. Flowering plants – Monocotyledons: 141–190. Berlin, Springer-Verlag. Guarise N.J., Vegetti A.C. (2008) The inlorescence structure of Cyperus L. section Luzuloidei Kunth. (Cyperaceae). Plant Sys- Vrijdaghs et al., Floral ontogeny in Cyperus and Pycreus (Cyperaceae) tematics and Evolution 271: 41–63. DOI: 10.1007/s00606-0070590-6 Haines R.W. (1967) Prophylls and branching in Cyperaceae. Journal of the East African Natural History Society 26(1): 51–70. Haines R.W., Lye K.A. (1983) The sedges and rushes of East Africa. Nairobi, East African National History Society. Kern J.H. (1962) New look at some Cyperaceae mainly from the tropical standpoint. Advancement of Science 19: 141–148. Kükenthal G. (1936) Cyperaceae-Scirpoideae-Cypereae 2. In: Engler A. (ed.) Das Planzenreich 101: 161–671, ig. 19–65. Berlin, Engelmann. Metcalfe C.R. (1971) Anatomy of the monocotyledons. Volume 5. Cyperaceae. London, Oxford University Press. Muasya A.M., Simpson D.A., Chase M.W., Culham A. (1998) An assessment of suprageneric phylogeny in Cyperaceae using rbcL DNA sequences. Plant Systematics and Evolution 211: 257–271. DOI: 10.1007/BF00985363 Muasya A.M., Simpson D.A., Chase M.W., Culham A. (2001a) A phylogeny of Isolepis (Cyperaceae) inferred using plastid rbcL and trnL–F sequence data. Systematic Botany 26: 342–353. DOI: 10.1043/0363-6445-26.2.342 Muasya A.M., Simpson D.A., Chase M.W. (2001b) Generic relationships and character evolution in Cyperus s.l. (Cyperaceae). Systematics and Geography of Plants 71: 539–544. DOI: 10.2307/3668698 Muasya A.M., Simpson D.A., Chase M.W. (2002) Phylogenetic relationships in Cyperus s.l. (Cyperaceae) inferred from plastid DNA sequence data. Botanical Journal of the Linnean Society 138: 145–153. DOI: 10.1046/j.1095-8339.2002.138002145.x Muasya A.M., Simpson D.A., Verboom G.A., Goetghebeur P., Naczi R.F.C., Chase M.W., Smets E. (2009a) Phylogeny of Cyperaceae based on DNA sequence data: current progress and future prospects. Botanical Review 75: 2–21. DOI: 10.1007/s12229008-9019-3 Muasya A.M., Vrijdaghs A., Simpson D.A., Chase M.W., Goetghebeur P., Smets E. (2009b) What is a genus in Cypereae: phylogeny, character homology assessment and generic circumscription. Botanical Review 75: 52–66. DOI: 10.1007/ s12229-008-9018-4 Podlech D. (1960) Ueber einige Cyperaceen Südafrikas. Mitteilungen der Botanischen Staatssammlung. München 3: 521–530. [Biodiversity Heritage Library: http://www.biodiversitylibrary. org/item/52384#page/555/mode/1up] Raynal J. (1966) Notes cyperologiques: IV. Trois Cyperus africains a style indivis. Adansonia 6(2): 301–308. Simpson D.A., Muasya A.M., Alves M., Bruhl J.J., Dhooge S., Chase M.W., et al. (2007) Phylogeny of Cyperaceae based on DNA sequence data – a new rbcL analysis. In: Monocots III/ Grasses IV. Aliso 23: 72–83. [available at http://www.herbarium.lsa.umich.edu/Monocots%20III%20Cyperaceae%20phylogeny.pdf] Takhtajan A. (1997) Diversity and classiication of lowering plants. New York, Columbia University Press. Troll W. (1964) Die Inloreszenzen, Typologie und Stellung im Aufbau des Vegetationskörpers. Bd I, II, 1. Jena 1964/69, Germany. Van der Veken P. (1964) Bijdrage tot de systematische embryologie der Cyperaceae-Cyperoideae. PhD thesis, K.U. Leuven, Leuven, Belgium. Vrijdaghs A., Goetghebeur P., Smets E., Muasya A.M., Caris P. (2004) The nature of the perianth in Fuirena (Cyperaceae). South African Journal of Botany 70: 587–594. [available at http://bio.kuleuven.be/sys/site/Publicaties/2004/Alex%20Vrijdaghs/The%20nature%20of%20the%20perianth%20in%20 Fuirena.pdf] Vrijdaghs A., Caris P., Goetghebeur P., Smets E. (2005a) Floral ontogeny in Scirpus, Dulichium and Eriophorum (Cyperaceae), with special reference to the perianth. Annals of Botany 95: 1199–1209. DOI: 10.1093/aob/mci132 Vrijdaghs A., Goetghebeur P., Muasya A.M., Caris P., Smets E. (2005b) Floral ontogeny in Ficinia and Isolepis (Cyperaceae), with focus on the nature and origin of the gynophore. Annals of Botany 96: 1247–1264. DOI: 10.1093/aob/mci276 Vrijdaghs A. (2006a) A loral ontogenetic approach to homology questions in non-mapanioid Cyperaceae – Een bloemontogenetische benadering van homologie-vraagstukken bij nietmapanioïde Cyperaceae. PhD thesis, K.U.Leuven, Leuven, Belgium Vrijgahs A., Goetghebeur P., Smets E., Muasya A.M. (2006b) The loral scales in Hellmuthia (Cyperaceae, Cyperoideae) and Paramapania (Cyperaceae, Mapanioideae): an ontogenic study. Annals of Botany 98: 619–630. DOI: 10.1093/aob/mcl138 Vrijdaghs A., Goetghebeur P., Smets E., Caris P. (2007) The Schoenus spikelet: a rhipidium? A loral ontogenetic answer. In: Columbus J.T., Friar E.A., Porter J.M., Prince L.M., Simpson M.G.(eds) Monocots: comparative biology and evolution – Poales. Aliso 23: 204–209. [available at http://bio.kuleuven.be/ sys/site/Publicaties/2007/Alex%20Vrijdaghs/Aliso%20Vrijdaghs%20et%20al%202007.pdf] Vrijdaghs A., Muasya A.M., Goetghebeur P., Caris P., Nagels A., Smets E. (2009) A loral ontogenetic approach to homology questions within the Cyperoideae (Cyperaceae). Botanical Review 75: 30–51. DOI: 10.1007/s12229-008-9021-9 Vrijdaghs A., Reynders M., Larridon I., Muasya A.M., Smets E., Goetghebeur P. (2010) Spikelet structure and development in Cyperoideae (Cyperaceae): a monopodial general model based on ontogenetic evidence. Annals of Botany 105: 555–571. DOI: 10.1093/aob/mcq010 Weberling F. (1992) Morphology of lowers and inlorescences. Cambridge, Cambridge University Press. Zhang X., Wilson K.L., Bruhl J.J. (2004) Sympodial structure of spikelets in the tribe Schoeneae (Cyperaceae). American Journal of Botany 91: 24–36. DOI: 10.3732/ajb.91.1.24 Manuscript received 26 Mar. 2010; accepted in revised version 27 May 2010. Communicating Editor: Elmar Robbrecht. 63