Skip to main content
NCBI home page
Annals of Botany logoLink to Annals of Botany
. 2017 Jan 8;119(4):533–543. doi: 10.1093/aob/mcw236

Unveiling the osmophores of Philodendron adamantinum (Araceae) as a means to understanding interactions with pollinators

Patrícia Gonçalves-Souza 1, Clemens Schlindwein 1, Stefan Dötterl 2, Elder Antônio Sousa Paiva 1,*
PMCID: PMC5458670  PMID: 28065928

Abstract

Background and Aims Araceae species pollinated by nocturnal Cyclocephalini beetles attract their pollinators by inflorescence scents. In Philodendron, despite the intense odour, the osmophores exhibit no definite morphological identity, making them difficult to locate. This may explain why structural studies of the scent-releasing tissue are not available so far.

Methods Several approaches were employed for locating and understanding the osmophores of Philodendron adamantinum. A sensory test allowed other analyses to be restricted to fertile and sterile stamens as odour production sites. Stamens were studied under light and electron microscopy. Dynamic headspace and gas chromatography–mass spectrometry were used to collect and analyse scents from different zones of the inflorescence.

Key Results The epidermal cells of the distal portion of fertile stamens and staminodes are papillose and, similar to the parenchyma cells of this region, have dense cytoplasm and large nuclei. In these cells, the composition of organelles is compatible with secretory activity, especially the great number of mitochondria and plastids. In this portion, lipid droplets that are consumed concomitantly with the release of odour were observed. Quantitative scent analyses revealed that the scent, with a predominance of dihydro-β-ionone, is mainly emitted by the fertile and sterile staminate zones of the spadix. An amorphous substance in the stomata pores indicates that the components are secreted and volatilized outside of the osmophore under thermogenic heat.

Conclusions Despite the difficulty in locating osmophores in the absence of morphological identity and inefficiency of neutral red staining, the osmophores of P. adamantinum have some features expected for these structures. The results indicate a functional link between thermogenesis and volatilization of osmophore secretions to produce olfactory signals for attracting specialized beetle pollinators. These first experimental data about the precise location of osmophores in Philodendron will stimulate studies in related species that will allow future comparison and the establishment of patterns of functional morphology.

Keywords: Araceae, cellular ultrastructure, insect–plant interaction, osmophores, Philodendron, pollination

INTRODUCTION

Odours emitted by osmophores attract pollinators, seed dispersers and predators of herbivores (Gang, 2005), and repel potential herbivores (Chapman et al., 1981). According to Vogel (1963), osmophores differ from other secretory glands by their location, the released volatile compounds, the duration of activity, and their anatomical structure.

While several chemical and ecological studies have focused on products of osmophores (Chapman et al., 1981; Huber et al., 2005; Vega et al., 2014), there is little information regarding their anatomy, mainly due to the difficulty in locating these secretory structures. The few detailed studies of osmophores are mostly dedicated to the Orchidaceae (Curry, 1987; Wiemer et al., 2009; Melo et al., 2010). Among the most strongly scented plants are Araceae (e.g. Dötterl et al., 2012). However, the cellular ultrastructure of osmophores remains poorly studied in this group of plants, except for Sauromatum guttatum (Skubatz et al., 1993, 1995; Skubatz and Kunkel, 1999; Hadacek and Webber, 2002).

Araceae comprises 106 genera and 2823 species (Govaerts and Frodin, 2002), with Philodendron being the second largest genus with approx. 700 species (Croat, 1997). The inflorescence in this genus has a short peduncle and is surrounded by a persistent spathe that opens from the apex to the base and closes soon after the release of pollen. The spadix has achlamydeous pistillate flowers at the base and staminate flowers at the apex, with the staminate flowers often being divided into an apical fertile and a basal sterile region (Mayo, 1991).

Many Araceae, including species pollinated by beetles (Gibernau, 2003), produce strong inflorescence scents (Vogel, 1963; Knudsen et al., 2006), and several of these cantharophilous species have osmophores, which are a hallmark of this syndrome (Gottsberger, 1977).

Despite receiving floral visitors of various insect groups, pollination of Philodendron seems to be accomplished exclusively by large scarab beetles of the tribe Cyclocephalini (Scarabaeidae, Dynastinae) (Croat, 1997). Although odour emissions have been reported for several species (Gibernau et al., 1999, 2000; Gibernau and Barabe, 2002; Maia et al., 2010) and their importance in attracting nocturnal beetle pollinators was recently shown (Dötterl et al., 2012; Gottsberger et al., 2013; Maia et al., 2013; Pereira et al., 2014), there are no morphological studies that permit location with any specificity of the volatile-producing structures in Philodendron.

Philodendron adamantinum Mart. ex Schott, a rupicolous species endemic to the Serra do Espinhaço mountain range of Minas Gerais, Brazil (Mayo, 1991), exhibits the typical set of features of cantarophily: the release of odour to attract beetle pollinators, inflorescence heating during odour production (thermogenesis), nutritive tissues, protogyny, and the presence of a pollination chamber at the basal part of the inflorescence, which accommodates the pollinators (Pereira et al., 2014). According to these authors, pollination in P. adamantinum is performed exclusively by beetles of Erioscelis emarginata (Mannerheim, 1829) (Scarabaeidae, Cyclocephalini), which are attracted by the release of volatile substances during the pistillate stage (at dusk) of the 36 h anthesis. Therefore, the release of odour through osmophores is of vital importance for the reproductive success of this species. Since the osmophores in P. adamantinum are inconspicuous structures, as they are in general in Philodendron and other Araceae, anatomical approaches together with chemical analytical techniques are needed to study their location and secretory tissue. Thus, in order to better understand the interaction between P. adamantinum and its pollinators, this study aimed to investigate the operation of its osmophores by determining their location and performing structural, ultrastructural as well as scent analyses.

MATERIALS AND METHODS

Plant material

Samples of Philodendron adamantinum Mart. ex Schott were collected in the Parque Estadual do Rio Preto, in the municipality of São Gonçalo do Rio Preto, Minas Gerais, Brazil. Flowers in various stages of development, according to the requirements of each analysis, were located in different populations throughout the park.

Collecting was performed during the reproductive period, from November to January, in 2012 to 2014. Voucher material was deposited in the BHCB herbarium of Universidade Federal de Minas Gerais (UFMG) under the number 161786.

Sensory test

Two newly opened inflorescences were collected immediately before odour release, and taken to the laboratory where they were dissected into the following parts: spathe, inflorescence axis, fertile staminate flowers, sterile staminate flowers and pistillate flowers. These parts were further cut into pieces of 1 cm in length (only the central portion of each region) and placed in different plastic containers (50 mL Falcon tubes). Five different people performed a blind evaluation of which container (i.e. part of the inflorescence) produced odour resembling that of the inflorescence (ripe fruit) or another odour.

Structural analysis and nature of secretion

The results of the sensory test determined the source of odour to be fertile and sterile stamens. Thus, the following tests were performed only in fertile and sterile staminate regions.

Three inflorescences were collected and, in accordance with the release of odour, five stages of development were defined: (1) approx. 5 d prior to odour release (initial opening of the spathe); (2) 8 h prior to odour release (morning of total opening of the spathe); (3) during release of odour (during anthesis of pistillate flowers); (4) 24 h after odour release during anthesis of staminate flowers when the spathe starts to close (after anther dehiscence); and (5) 36 h after odour release (with a closed spathe). Samples of fertile stamens, staminodes and the tissue of the inflorescence axis were processed for study by light microscopy and histochemical tests.

The samples were subjected to a vacuum in Karnovsky solution of pH 7·2 in 0·1 m phosphate buffer (Karnovsky, 1965) and fixed for 24 h. Subsequently, the material was dehydrated in an ethanol series (Johansen, 1940) and subjected to pre-infiltration and infiltration of synthetic resin (2-hydroxyethyl methacrylate) (Leica®). A rotary microtome was used to produce 6 μm thick sections which were stained with Toluidine blue, pH 7·4 (O’Brien et al., 1964), counterstained with ruthenium red solution, placed on slides and mounted in Entellan® for photo-documentation and study by light microscopy.

Samples of inflorescences from fertile and sterile staminate regions were analysed using scanning electron microscopy (SEM) during the release of odour. Samples were fixed in 2·5 % gluteraldehyde in 0·1 m phosphate buffer, dehydrated in an ascending ethanol series to the critical point in liquid CO2, metallized with gold (Robards, 1978) and observed using a Quanta 200 scanning electron microscope (FEI Company, Eindhoven, The Netherlands) at 12–20 kV.

Histochemical tests employed both samples fixed in Karnovsky solution and sectioned by freehand, and material included in resin and sectioned with a rotary microtome. The following tests were performed: Lugol to identify starch (Johansen, 1940); NADI for terpenes (David and Carde, 1964); 10 % aqueous ferric chloride solution for phenolic compounds (Johansen, 1940); aqueous solution of 0·02 % ruthenium red for detection of pectic compounds (Jensen, 1962); and Sudan red B for lipids in general (Brundrett et al., 1991).

For locating metabolically active tissues, two inflorescences, including the spathes were collected during the stage corresponding to the release of odour. Both were totally immersed in an aqueous solution of 0·01 % neutral red for 24 h (Vogel, 1963). After this period, the inflorescences were removed and examined with the naked eye. The spadices were sectioned to observe the internal tissue.

Ultrastructural analysis

Considering the results obtained in the odour assay, the analysis using transmission electron microscopy (TEM) focused on samples of fertile and sterile stamens. Samples of the apical portions of the stamens were collected during the release of odour. Sampled fragments were fixed in Karnovsky solution, pH 7·2 in 0·1 m phosphate buffer (Karnovsky, 1965) for 24 h, post-fixed in 1 % osmium tetroxide (0·1 m phosphate buffer, pH 7·2) for 2 h, washed in phosphate buffer (0·1 m, pH 7·2), dehydrated in an ethanol series and infiltrated with Araldite® resin (Roland, 1978). The resultant ultrathin sections of 50 nm were stained with uranyl acetate and lead citrate, and examined using a Tecnai G2-Spirit Transmission Electron Microscope (Philips/FEI Company, Eindhoven, The Netherlands) at 80 kV.

Volatile collection and analysis

The volatile compounds emitted by P. adamantinum were obtained from three first day inflorescences corresponding to the first peak of thermogenesis. The inflorescences were cut from the plants 1 h before odour release and taken to the laboratory. The spathe was removed and, 1 h after cutting, the whole spadix was placed in a polyester oven bag (10 × 20 cm, Toppits®, Germany) wherein the released fragrance was allowed to accumulate for 10 min. The air contained in each bag was removed by a suction pump (ASF Thomas, Inc., Germany) for 2 min at a rate of 200 mL min–1. The accumulated volatile compounds were captured in adsorbent tubes, which contained 1·5 mg of Tenax-TA (mesh 60–80) and 1·5 mg of Carbotrap B (mesh 20–40; both Supelco). After capturing the compounds released by the entire spadix, the fertile staminate, the sterile staminate and the pistillate regions of each inflorescence were cut (excluding the transition zones of approx. 5 mm), and volatiles were collected from these separate parts as described previously.

The volatiles trapped were analysed by gas chromatography–mass spectrometry (GC/MS) using an automatic thermal desorption (TD) system (TD-20, Shimadzu, Japan) coupled to a Shimadzu GCMS-QP2010 Ultra equipped with a ZB-5 fused silica column (5 % phenyl polysiloxane; 60 m, i.d. 0·25 mm, film thickness 0·25 μm, Phenomenex) as described before (Heiduk et al., 2015; Zito et al., 2015). GC/MS data were processed using the GCMSolution package, Version 2·72 (Shimadzu Corporation 2012). As cutting of the spathe/spadix might have induced the emission of some volatiles not released by undamaged inflorescences, only compounds identified by Pereira et al. (2014), which used plants from the same population, were considered.

For each substance, the total peak area in each sample was determined, its percentage (= relative amount with respect to total peak area) calculated and square-root transformed, and used for further statistical analyses. We did not use absolute amounts of compounds for analyses because the total absolute amounts strongly varied among individuals (see the Results). Semi-quantitative similarities in scent patterns among samples were calculated using the Bray–Curtis similarity index in the statistical software PRIMER 7.0.10 (Clarke and Gorley, 2015). Non-metric multidimensional scaling (NMDS) was used to depict variation in scent among the samples (Clarke and Gorley, 2015). To evaluate how well (or poorly) the ordination produced the observed similarity matrix, the stress value is given. The smaller the stress value, the better is the fit of the reproduced ordination to the observed similarity matrix, with values <0·2 being reasonable (Clarke and Gorley, 2015). Bray–Curtis similarities were also used to test for semi-quantitative differences in scent among different inflorescence parts using PERMANOVA (10 000 permutations) in Primer (Clarke and Gorley, 2015). PERMANOVA is a technique for testing the simultaneous response of one or more variables to one or more factors in an analysis of variation (ANOVA) experimental design on the basis of a (dis)similarity (distance) matrix with permutation methods (Anderson et al., 2008). Two different tests were performed. In the first test, we only included the spadix and the fertile and sterile staminate zones, as they were categorized in the sensory evaluations as smelling similar (see the Results). The second test considered, additionally, the samples collected from the pistillate zone. In both tests, inflorescence part was the categorical factor; we also included individual as a factor since several parts were used from a single inflorescence.

RESULTS

Location, structural organization and nature of secretion

The sensory evaluation detected a strong odour of ripe fruit, resembling the scent of a complete inflorescence only from the flowers that make up the fertile and the sterile staminate region. No odours were detected in the other regions of the inflorescence (i.e. the axis, the pistillate flowers and the spathe) in the sensory evaluation.

The fertile staminate region, located at the apex of the spadix (Fig. 1A), possesses numerous juxtaposed pyramidal to obpyramidal stamens (Fig. 1A, B). Each stamen is covered with a uniseriate epidermis and contains a tetrasporangiate anther with weakly delimited thecae (Fig. 1B, C) and longitudinal dehiscence. These stamens have only one vascular bundle that originates in the axis of the inflorescence. Its distal portion coincides with the apex of the sporangia, delimiting a sterile apical portion, at the top of connective (Fig. 1B, D). In this portion, the epidermal cells are papillose and form the exposed surface of the spadix (Fig. 1C, D).

Fig. 1.

Fig. 1

Open in a new tab

Fertile stamens of Philodendron adamantinum. (A) Diagram showing the distribution of flower types on the spadix; in the apical region are fertile staminate flowers, in the middle region are staminodes, and in the basal region pistillate flowers. (B) Longitudinal section of a staminate flower highlighting the sterile apical portion where phenolic substances accumulate in some of the parenchyma cells and the microsporangia; the arrow indicates the single central vascular bundle. Letters d and e indicate the position of the transverse sections shown in (D) and (E), respectively. (C) Apical surface of the stamens showing the homogeneously papillose epidermis. The inset shows a stoma located on this surface; note a residue in a stomata pore. (D) Transverse section of the sterile, apical portion; note the papillose epidermis and the absence of vascular tissues; here cells with dense cytoplasm occur among phenolic-accumulating cells. (E) Transverse section of the anther region showing the microsporangia and the vascular bundle; note the absence of papillae on the epidermal cells. (F and G) Detail of secretory cells of the apical portion of a fertile stamen. (F) At 8 h prior to release of odour; note the cells with dense protoplast. (G) At 24 h after the release of odour; note the intense vacuolization of the cells and phenolic epidermis. (ep, epidermis; st, secretory tissue; vb, vascular bundle).

Stomata are found throughout the surface of the stamens. When odours are released, the stomata at the apex are clogged by an amorphous substance that spreads among the papillae (Fig. 1C). On the sides of the stamens, papillae are absent on epidermal cells and show a flat external periclinal surface. These non-papillose epidermis cells are provided with large vacuoles in which phenolic substances accumulate (Fig. 1E). Prior to thermogenesis, the apical portion of the stamens have regions of cells with secretory characteristics, such as dense cytoplasm, conspicuous nucleus and poorly developed vacuome, interspersed by vacuolated cells that store phenolic substances (Fig. 1D, F). After thermogenesis, the secretory cells become vacuolated, but devoid of phenolic substances (Fig. 1G), but the epidermis becomes full of phenolics too.

The sterile staminate region is found in the middle portion of the inflorescence and is almost totally immersed in the chamber formed by the constriction of the spathe. This region consists of prismatic, juxtaposed, sterile stamens. The inner and outer structures of these staminodes are quite similar to that of fertile stamens, differing only in their larger size and by the absence of microsporangia.

Starch was observed in the ground tissue of the fertile stamens (Fig. 2A) and in some epidermal cells only before the first peak of thermogenesis. During the first peak of thermogenesis this starch is consumed (Fig. 2B). Just a few starch grains were observed in the staminodes, almost undetectable by light microscopy.

Fig. 2.

Fig. 2

Open in a new tab

Histochemical tests performed on the sterile, apical portion of fertile stamens of Philodendron adamantinum. (A, B) Lugol test for starch, shown in transverse section of parenchyma. (A) At 8 h prior to the intense emission of odour. (B) During the release of odour. (C, F) Sudan red test for lipids. (C) Longitudinal section of the apex of the stamen showing the accumulation of lipid substances in the sterile, apical portion. (D–F) Sequence of transverse sections of the region near the epidermis showing the consumption of lipid reserves. (D) At 8 h prior to the release of odour; note the droplets of lipid reserves dispersed in the cytoplasm. (E) During the release of odour; note the partial consumption. (F) At 24 h after the release of odour showing the exhaustion of the reserves. (G, H) NADI test for terpenes on the apical portion of fertile stamens. (G) At 4 h prior to the release of odour with a massive presence of terpenes from essential oils (blue). (H) During release of odour showing the absence of terpenes. (I) Inflorescence subjected to immersion in neutral red; note the positive result only in the spathe and the pistillate flowers indicated by the black arrow.

Lipid droplets dispersed in cytoplasm occurred in the cells of the non-vascularized apical portion of both fertile stamens and staminodes (Fig. 2C, D). During the first peak of thermogenesis these lipids are consumed (Fig. 2E) and they were absent in later stages (Fig. 2F). This lipid fraction is composed, at least in part, of terpenes, as demonstrated by the NADI test in fertile stamens (Fig. 2G, H). The terpene fraction showed a similar pattern to that described for lipids, with a noticeable decrease after thermogenesis (Fig. 2G, H).

After 24 h of immersion of the inflorescence in neutral red solution, the tips of pistillate flowers and the upper side of the spathe were stained intensely. The fertile and sterile staminate regions did not show color change (Fig. 2I), nor did the axis of the inflorescence. The spadix was sectioned longitudinally in order to inspect the internal tissues that were also not stained.

Ultrastructural organization

The ground tissue of the sterile apical portion of the fertile stamens has both cells with dense cytoplasm rich in organelles and cells vacuolated with phenolic substances (Fig. 3A). In the cells with phenolic substances, the vacuole occupies a large portion of the cell lumen, leaving the cytoplasm, poor in organelles, limited to a narrow band compressed by the cell wall (Fig. 3A).

Fig. 3.

Fig. 3

Open in a new tab

Secretory cells while emitting odour in fertile stamens and staminodes of Philodendron adamantinum. (A–G) Fertile stamens. (A) Secretory cells showing dense cytoplasm. (B) Cell showing a large population of organelles, especially amyloplasts and mitochondria; arrow highlights plastoglobules; note plastids lacking a thylakoid system. (C) Connection between two cells by field pits indicated by the black arrow; white arrows highlight numerous plastoglobules. (D) Population of mitochondria showing well-developed cristae. (E–G) Sequence of the process transforming a plastid into a vacuole. (E) Normal plastid at the top and at the bottom, the other with content already modified; in detail, double membranes of both structures. (F) Intermediate stage of plastid transformation into vacuoles. (G) Formed vacuole; the presence of plastoglobules (arrows) indicates the original plastid. (H, I) Staminodes. (H) Vesicles fused to the vacuole (white arrows) and vacuole fusion to the plasma membrane (black arrows), releasing contents into the periplasmic space. (I) Epidermal cell showing a population of organelles similar to that of other secretory cells. di, dictyosome; mi, mitochondria; nu, nucleus; pl, plastid; pl**, modified plastid; er, endoplasmic reticulum; va, vacuole.

The cells with dense cytoplasm have a high population of organelles and the vacuome is formed by many small vacuoles, inside of which is an amorphous material that is probably comprised of lipids (Fig. 3A). In these cells, the nucleus is evident with a conspicuous nucleolus (Fig. 3B); the wall is very thin and plasmodesmata communicate with adjacent cells (Fig. 3C). In addition to vacuoles, plastids, mitochondria, smooth endoplasmic reticulum and dictyosomes are among the organelles most commonly represented in the cytoplasm (Fig. 3B, C). The plastids have a poorly developed system of membranes, dense stroma, and numerous plastoglobules; starch grains can be seen in some of these organelles (Fig. 3B, C). The dictyosomes are few in number, have few cisternae, and show no indication of the production of vesicles. Also noteworthy is the high number of mitochondria with well-developed cristae scattered throughout the cytoplasm (Fig. 3D).

During the release of odour, some plastids indicate hydrolysis of starch and a considerable reduction in the density of the stroma (Fig. 3E). At this stage, these organelles lose their double membrane, the system of internal membranes dissolves and the organelle undergoes a contour alteration, expanding into an amoeboid shape (Fig. 3E, F). In the wake of these events, plastid transformation can still be recognized, especially by the presence of plastoglobules; inner membranes disappear and the stroma becomes even less dense (Fig. 3F) and, in some cells, it is possible to note plastoglobules within the vacuoles (Fig. 3G). Throughout the cytoplasm, fusions of small vacuoles can be observed and, not rarely, the fusion of a vacuole with the plasma membrane, when there occurs the extravasation of vacuolar content into the periplasmic and intercellular spaces (Fig. 3H).

The papillose epidermal cells arranged at the apex of the stamens have a composition of organelles similar to that observed in cells of the ground tissue; an elevated population of mitochondria, plastids and small vacuoles (Fig. 3I). In these cells the outer periclinal cell wall is slightly thick and lamellar; the thick and complete cuticle, along with the outer layers of the wall, form ripples.

Odour analyses

Dynamic headspace and GC/MS analyses of the complete spadix, and the different zones thereof, revealed that the scent of P. adamantinum is mainly released by the fertile and sterile staminate zones. Despite the high variation in total amount of scent trapped from the staminate zones (and the complete spadix), the pistillate zone released a 6- to 10-fold smaller amount of total scent compared with the sterile and fertile staminate zones, respectively (Table 1).

Table 1.

Absolute  (ng; median, min–max) amount of scent trapped  (10 min bagging + 2 min collecting) from three complete spadices as well as from the fertile staminate, sterile staminate, and pistillate zones thereof

Spadix Fertile staminate zone Sterile staminate zone Pistillate zone
Total amount of scent trapped  (ng)  (median, min–max) 337·9   (199·6–7300·0) 403·4  (127·8–1383·4) 241·6  (114·1–1302·0) 45·4  (11·2–46·0)
Aromatics
Methyl salicylate 0·2  (tr-124·7) 4·0  (0·3–6·7) 0·6  (0·0–2·3) 0·2  (0·1–0·4)
Miscellaneous cyclic compounds
2,2,6-Trimethy-6-vinyldihydro-2H-pyran-3(4H)-one 0·3  (0·2–104·1) 0·2  (0·0–7·3) 0·4  (0·4–2·7) 1·1  (0·7–2·2)
(E)-Methyl jasmonate 0·3  (0·0–0·3) – 0  (0–5) –
(Z)-Methyl jasmonate 8·0  (0·3–78·2) 0·5  (0·5–10·2) 1·8  (0·2–270·0) –
Terpenoids
α-Pinene 105·5  (19·1–401·5) 17·6  (7·5–21·5) 4·5  (4·2–32·8) 23·1  (tr–25·8)
Sabinene 3·4  (3·3–265·6) 8·3  (8·3–77·0) 1·3  (0·1–1·7) tr  (0·0–3·3)
β-Pinene 20·2  (3·4–130·0) 11·7  (4·5–67·0) 0·7  (0·1–5·0) 5·0  (0·1–5·0)
6-Methyl-5-hepten-2-one tr  (0·0–33·3) 1·3  (tr–3·3) tr  (tr–0·7) tr  (tr–tr)
β-Myrcene 12·2  (8·1–255·5) 5·0  (3·3–43·2) 13·3  (0·7–13·4) 0·1  (0·0–3·3)
Limonene 21·2  (15·0–333·3) 66·7  (6·3–76·2) 9·6  (4·2–23·4) 2·7  (0·0–6·7)
1,8-Cineole 23·5  (12·8–1000·0) 59·3  (57·1–300·0) 9·6  (3·5–14·5) 5·0  (tr–6·7)
γ-Terpinene 2·4  (1·5–108·8) 0·1  (tr–10·0) 0·8  (0·1–3·6) tr  (0·0–tr)
(Z)-Linalool oxide furanoid 0·1  (0·1–33·3) 0·1  (tr–3·7) 0·2  (tr–1·2) 0·3  (tr–0·4)
(E)-Linalool oxide furanoid 0·3  (0·1–166·7) 1·3  (tr–30·0) 0·2  (0·1–4·0) 0·5  (0·2–1·4)
Linalool 1·2  (0·6–587·8) 3·3  (3·3–33·3) 0·8  (0·3–21·7) 0·3  (0·1–0·8)
(E)-4,8-Dimethyl-1,3,7-nonatriene 1·2  (0·5–1·4) 0·6  (0·2–0·7) 0·3  (0·1–0·5) 0·4  (0·3–2·4)
(Z) + (E)-Linalool oxide pyranoid 2·4  (1·7–99·1) 0·7  (0·7–9·3) 0·5  (0·2–3·3) 0·2  (0·1–0·3)
4-Terpineol tr  (0·0–10·0) 0·1  (0·0–2·3) tr  (0·0–0·3) 0·2  (tr–0·5)
α-Terpineol 0·2  (0·1–133·1) 1·3  (0·3–8·5) 0·1  (0·1–1·5) tr  (tr–0·1)
Methyl citronellate tr  (tr–42·1) 0·0  (0–0·4) tr  (0·0–2·5) –
Theaspirane isomer I 0·3  (0·1–23·3) 0·7  (0·1–2·0) 0·3  (0·2–1·0) 0·1  (tr–0·1)
Methyl geranate 0·3  (tr–213·0) tr  (tr–4·1) 0·3  (0·0–30·3) tr  (0·0–tr)
Theaspirane isomer II 0·2  (0·1–22·1) 0·7  (0·2–2·5) 0·2  (0·2–0·8) 0·1  (0·0–0·1)
Cyclo-β-ionone tr  (0·0–0·3) 0·0  (0·0–0·3) – –
Dihydro-β-ionone 123·9  (99·2–2160·7) 133·5  (28·9–605·5) 136·4  (57·6–602·3) 1·7  (1·2–2·0)
Dihydro-β-ionol 1·3  (0·4–266·7) 0·7  (0·3–5·0) 0·3  (0·3–23·9) tr  (tr–tr)
(E)-8(9)-Dehydro-4(5)-dihydrotheaspirone 21·9  (20·6–464·7) 18·5  (2·0–100·8) 30·7  (17·2–264·2) tr  (tr–0·1)
β-Ionone 0·1  (0·1–30·0) 0·1  (tr–2·8) 0·5  (tr–5·0) 0·0  (0·0–tr)
Unknowns
0·6  (0·6–97·2)7* 4·4  (0·6–8·5)7 0·3  (0·1–17·5)7 tr  (0·0–tr)2
Open in a new tab

Values >30·0 are printed in bold.

*

The unknowns were pooled and the number of pooled compounds is indicated by superscript digits.

Consistent with the sensory evaluations, both staminate zones did not differ in scent composition (i.e. the relative amount of compounds) from the complete spadix (Pseudo-F2,8 = 2·15, P = 0·13). Indeed, with the exception of one sample (see also below), samples collected from the complete spadix and the staminate zones group in similar regions in our ordination plot (Fig. 4). However, a strong difference among inflorescence parts was found when including samples of the pistillate zone in the PERMANOVA model (Pseudo-F3,11 = 3·27, P = 0·001). Samples collected from the pistillate zone differed from the other parts, and had a much smaller loading on dimension 1 of our ordination plot (Fig. 4). Dihydro-β-ionone was most abundant in the samples collected from the spadix and the staminate zones, but was almost absent in the samples collected from the pistillate zone, which was dominated by α-pinene (Table 1). We observed that the terpenoid (E)-8(9)-dehydro-4(5)-dihydrotheaspirone was another compound released in high (relative) amounts by staminate zones, but in small (relative) amounts by pistillate zone. One sample collected from a sterile staminate zone contained additionally high amounts of (Z)-methyl jasmonate (the sample with the highest loading on dimension 1 in our ordination; Fig. 4).

Fig. 4.

Fig. 4

Open in a new tab

Plot of relative scent composition using non-metric multidimensional scaling of samples collected from complete spadices and different zones thereof. Samples collected from the complete spadix and both fertile and sterile staminate zones had similar scent compositions and group in the centre or right-hand side of the ordination, whereas samples collected from the pistillate zone group on the left-hand side of the plot. These findings were supported by PERMANOVA analyses: there were non-significant effects among samples collected from the complete spadix and both staminate zones, whereas a significant test outcome was obtained when additionally adding samples from the pistillate zone (see text for more details).

DISCUSSION

Location of osmophores

In agreement with sensory evaluations, scent analyses by dynamic headspace and GC/MS revealed that the staminate zones of the spadix produce a scent similar to that of the complete spadix. However, in contrast to the sensory evaluations, which did not detect a scent from the pistillate zone, the quantitative analyses detected the release of a small amount of scent with a different composition from this zone. Anatomical and ultrastructural analyses demonstrated that osmophores are located in the apical non-vascularized portion of both fertile and sterile staminate flowers. Papillose epidermal cells are characteristic of this region. The occurrence of such papillae in tissues that secrete volatile compounds has been reported for various other species of different families (Endress, 1984; Sazima et al., 1993; Vogel and Hadacek, 2004; Melo et al., 2010) and is a recurrent feature that can be used to indicate the presence of osmophores. In the aroid Lagenandra ovata, the osmophores are also located in the stamens (Buzgo, 1998), but structural data were not provided by this study, making further comparisons difficult. Using a staining approach, Leguet et al. (2014) showed the intracellular location of vesicles in the epidemical cells of staminate (fertile and sterile) flowers in Arum italicum.

Structure and ultrastructure of osmophores

Even in small quantities, the stomata of the apex of the stamens appear to function as a means of release of volatile substances. At the odour-releasing stage, the presence of amorphous substance in the stomata pore and the absence of cuticular pores in other epidermal cells are evidence that support this hypothesis. Stomata as a means of releasing secretion from osmophores have been observed in other species (Vogel, 1963; Melo et al., 2010). Another possible way for secretion release, according to Paiva (2016), is the diffusion of compounds through the cuticle without the necessity of rupture. Since the epidermal cells also contain terpenes, this possibility must also be considered.

It is surprising that the tissues that secrete volatile compounds in the spadix of P. adamantinum did not stain with neutral red solution (Fig. 2G) as is usual for tissues with characteristics of elevated metabolism, including osmophores (Vogel, 1963; and numerous subsequent publications of other authors). It is possible that the dye does not reach the internal secreting tissue due to the very small number of stomata and the integrity of the cuticle, particularly in the apical portion of the stamens. The result of the neutral red test in the inflorescences of P. adamantinum would lead to the erroneous conclusion that osmophores would be located in the spathe and the corresponding region of pistillate flowers. Our results thus indicate that non-staining with neutral red may not necessarily be conclusive.

The cells of the apex of the stamens possess dense cytoplasm, elevated populations of organelles, especially mitochondria, and a poorly developed vacuome in the period prior to release of odour. These traits, according to Fahn (1988), are characteristics of secretory tissue. As with P. adamantinum, S. guttatum (Skubatz et al., 1993, 1995) and several species of Orchidaceae have numerous mitochondria with well-developed cristae in the tissue of osmophores (Pridgeon and Stern, 1983; Curry, 1987; Melo et al., 2010). Melo et al. (2010) attributed this characteristic to the high consumption of energy during the secretory process.

The high density of mitochondria is also related to the process of thermogenesis arising from cellular respiration. According to Seymour et al. (1983), thermogenesis in Philodendron selloum occurs in staminate flowers, with approx. 70 % of the heat being emitted by sterile flowers. According to Walker et al. (1983), the heat generation in P. selloum flowers employs lipids as substrate. Thus, considering the close phylogenetic relationship between P. adamantinum and P. selloum, both species being members of the subgenus Meconostigma (Mayo, 1991), we suggest that P. adamantinum would share some of these characters.

The presence of plastoglobules is seen by many authors as evidence of the production of terpene compounds (Pridgeon and Stern, 1983; Curry, 1987). Although recent studies have revealed other functions for these structures, these studies do not rule out the role of metabolism and storage of lipid substances (Austin et al., 2006; Bréhélin et al., 2007; Bréhélin and Kessler, 2008). In Arum italicum, Leguet et al. (2014) used the presence of terpene vesicles to locate the osmophore.

Although the accumulation of starch in osmophores is reported in several studies (García et al., 2007; Wiemer et al., 2009; Melo et al., 2010), including being used as a characteristic indicative of the location of osmophores, the presence of starch is not universal; Vogel and Martens (2000) reported that osmophores in species of Arisaema (Araceae) do not exhibit starch reserves, though they maintain the emission of odour as well as constant elevated temperature throughout anthesis. The presence of starch in fertile stamens of P. adamantinum is restricted to the phase immediately prior to thermogenesis and release of odour. This starch coexists with terpenes, substances that are possibly part of the released bouquet. Therefore, we can infer that the starch does not participate in the synthesis of these terpenes, although starch could be consumed in other phases of the secretory process as a potential energy source for synthesis of other compounds.

The abundant vascularization in the region of the osmophores of Ceropegia elegans was related to the need for water and phloem-derived sugars during the production and emission of volatile compounds (Vogel, 1963). The absence of vascularization in the region of osmophores in P. adamantinum indicates that the resources used in the production and release of volatile compounds are most probably provided by reserves, which accumulate in the secretory tissues during their development.

The transport of the plastid content to the interior of vacuoles was also reported in nectaries of Cucumis sativus (Cucurbitaceae) (Peng et al., 2004) and Hymenaea stigonocarpa (Fabaceae) (Paiva and Machado, 2008). However, in both cases, the plastid does not become converted into a vacuole as observed in cells of osmophores of P. adamantinum in the present study. In C. sativus, the plastid is engulfed by the vacuole, and in H. stigonocarpa, the envelope is dissolved and subsequently forms part of the vacuole membrane. Peng et al. (2004) suggest that the nectar is modified within the plastids in association with vacuoles. In P. adamantinum, the synthesis of at least part of the released volatile compounds occurs within the plastids, and the transformation of these organelles in vacuoles can be related to the transport of the secretion to the periplasmatic space. The conversion of plastids within vacuoles is clearly demonstrated by the presence of plastoglobules, which were previously restricted to plastids within the vacuoles during the phase of secretion release.

The secretion of volatile compounds is, at least in part, granulocrine, since numerous vesicles are incorporated into the plasma membrane, releasing their contents to the periplasmatic space. This mode of secretion was observed in the osmophores of species of Orchidaceae and Passifloraceae (Pridgeon and Stern, 1983; García et al., 2007) and in the appendix of S. guttatum, Araceae (Skubatz et al., 1995). We found no expressive accumulation of substances in the periplasmatic space, which, according to Vogel (1983), is common in osmophores with volatile substances that are synthesized during the release of odour and not stored previously.

The appearance of an amorphous substance in stomata pores indicates that the secretion is not released in gaseous form, but instead is volatilized just outside the osmophore on the apical surface of staminate flowers. This may explain the precisely timed and extraordinarily intense floral scent production, as well as the coincidence with thermogenesis, which are common for aroid pollination systems involving scarab beetles. Furthermore, the observation that the material accumulated in the periplasmic space is later exposed on the osmophore suggests the release of secretion products according to the model proposed by Paiva (2016).

Odour analysis and pollinator attraction

Terpenoids are the main class of substances of the scent of P. adamantinum, with a predominance of dihydro-β-ionone, which is the primary volatile component of its floral bouquet (Table 1; Pereira et al., 2014), and this class of substances were found in some other Philodendron species as well as in other Araceae genera (Leguet et al., 2014). This substance alone was shown to attract the scarab beetle Erioscelis emarginata, the sole pollinator of this plant species. Dihydro-β-ionone is released from the staminate zones (see Table 1), and thus the staminate flowers are responsible not only for production of heat and pollen (only fertile ones), but also for scent production and release, and pollinator attraction. In addition, the high production of the (E)-8(9)-dehydro-4(5)-dihydrotheaspirone, found mainly in staminate flowers, may be related to attraction of P. adamantinum pollinators as suggested by Dötterl et al. (2012) for another plant species. Thus, the inflorescence of P. adamantinum, which consists of a spathe, and a spadix with fertile and sterile male flowers and pistillate flowers, functions as a unit. During receptivity of pistillate stage flowers, Erioscelis emarginata beetles are attracted in the early night to the inflorescence by scents released from the osmophores of both types of staminate flowers. Attracted beetles then eventually pollinate the weakly scented pistillate flowers in the chamber formed by the spathe at the base of the spadix, in which they stay for approx. 24 h. During the next evening, when the fertile staminate flowers release their pollen, beetles feed on the pollen, and thereby get covered with pollen, before leaving the inflorescence and being potentially attracted by a pistillate stage inflorescence.

As the P. adamantinum pollinators are nocturnal insects, the shelter formed by the floral chamber allows them to spend the daytime in the dark. Thus, this shelter is an additional floral reward, which adds to the food provided by the sterile stamens. So, in order to attract the pollinators to the floral chamber, scents emitted by the sterile staminate portion (the basal one) is likely important once this portion is within this chamber. Additionally, considering the structural similarity between the fertile and sterile stamens, we should not be surprised at the odour production in both.

The main compounds released from the pistillate zone, such as α- and β-pinene, have not yet been tested for their attractiveness to Cyclocephalini pollinators (Pereira et al., 2014). However, their presence mainly during non-attractive stages in other Philodendron species (Gottsberger et al., 2013) suggests that they are not involved in pollinator attraction. The massive presence of terpenoids as constituents of the floral bouquet of P. adamantinum is consistent with our NADI histochemical test, which revealed the presence of terpenes in the apical region of the staminate flowers.

Some of the compounds, such as limonene and 1,8-cineole, were released by some regions of the spadix in greater amounts than when compared with the entire spadix. This was unexpected and may be the result of injury caused by dissecting the spadix into the different zones. Overall, however, such effects potentially induced through cutting did not bias our conclusions, as shown, for example, in our ordination. With one exception, all samples collected from spadices and fertile and staminate zones group close together, showing that effects of cutting were not major and much smaller than the effects found between scent of the pistillate zone and the staminate zones/the spadix (Fig. 4).

Generally, with the exception of the samples collected from the pistillate zone, which all emitted small amounts of scent, there was considerable variation in absolute amount of scent among samples/inflorescences (see Table 1). As shown by Dötterl et al. (2012), the scent released by Philodendron is highly dynamic and varies by a factor of 40–50 within 1 h. Thus, the differences observed in our study can be explained by differences in the time when we cut the inflorescences and collected scent samples (variation of 1 h). Interestingly, our results also show that, despite this variation in absolute amount, the relative scent pattern is similar among the inflorescences/zones (see Supplementary Data Table S1).

Conclusions

Scent release by the inflorescences of P. adamantinum is a co-ordinated event with two main actors: the osmophores and thermogenesis (see Pereira et al., 2014). The osmophores are discrete and composed of secretory cells at the apex of fertile and sterile staminate flowers, where the scent, with a predominance of dihydro-β-ionone, is emitted. Chemicals released by osmophores are suggested to reach the external environment through stomata located in the apical surface of staminate flowers. There is evidence that these substances are volatilized at the epidermal surface, and the co-ordination with thermogenesis, which occurs at the staminate zones of inflorescence and at the exact same time as secretion release, seems to be essential to spread odour in order to attract specialized beetle pollinators.

SUPPLEMENTARY DATA

Supplementary data are available online at www.aob.oxfordjournals.org and consist of the following. Table S1: total absolute and relative amount (median, min–max) of scent trapped (10 min bagging + 2 min collecting) from three complete spadices as well as from the fertile staminate, sterile staminate and pistillate zones thereof.

Supplementary Material

Supplementary Data
Click here for additional data file. (16.4KB, docx)

ACKNOWLEDGEMENTS

We thank the Center of Microscopy (UFMG) for providing the equipment and technical support for experiments involving electron microscopy. We also thank two anonymous reviewers for their valuable comments on the manuscript. This work was supported through a research grant from the Conselho Nacional de Desenvolvimento Científico e Tecnológico–CNPq (308589/2011-4) for E.A.S.P and (312831/2013-7) for C.S.

LITERATURE CITED

  1. Anderson MJ, Gorley RN, Clarke KR.. 2008. Permanova+ for Primer: guide to software and statistical methods. Plymouth, UK: Primer-E. [Google Scholar]
  2. Austin II JR, Frost E, Vidi PA, Kessler F, Staehelin LA.. 2006. Plastoglobules are lipoprotein subcompartments of the chloroplast that are permanently coupled to thylakoid membranes and contain biosynthetic enzymes. The Plant Cell 18: 1693–1703. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Bréhélin C, Kessler F.. 2008. The plastoglobule: a bag full of lipid biochemistry tricks. Photochemistry and Photobiology 84: 1388–1394. [DOI] [PubMed] [Google Scholar]
  4. Bréhélin C, Kessler F, Wijk KJ.. 2007. Plastoglobules: versatile lipoprotein particles in plastids. Trends in Plant Science 12: 260–266. [DOI] [PubMed] [Google Scholar]
  5. Brundrett MC, Kendrick B, Peterson CA.. 1991. Efficient lipid staining in plant material with Sudan Red 7B or fluorol yellow 088 in polyethylene glycol–glycerol. Biotechnic and Histochemistry 66: 111–116. [DOI] [PubMed] [Google Scholar]
  6. Buzgo M. 1998. Odor differentiation in Lagenandra ovata (Araceae). Acta Botanica Yunnanica X: 76–88. [Google Scholar]
  7. Chapman RF, Bernays EA, Simpson SJ.. 1981. Attraction and repulsion of the aphid, Cavariella aegopodii, by plant odors. Journal of Chemical Ecology 7: 881–888. [DOI] [PubMed] [Google Scholar]
  8. Clarke KR, Gorley RN.. 2015. Primer v7: user manual/tutorial. Plymouth, UK: Primer-E. [Google Scholar]
  9. Croat TB. 1997. A revision of Philodendron subgenus Philodendron (Araceae) for Mexico and Central America. Annals of the Missouri Botanical Garden 84: 311–704. [Google Scholar]
  10. Curry KJ. 1987. Initiation of terpenoid synthesis in osmophores of Stanhopea anfracta (Orchidaceae): a cytochemical study. American Journal of Botany 74: 1332–1338. [Google Scholar]
  11. David R, Carde JP.. 1964. Coloration différentielle des inclusions lipidique et terpéniques des pseudophylles du pin maritime au moyen du réactif Nadi. Comptes Rendus Hebdomadaires dês Séances de l’ Academie dês Sciences Paris 258: 1338–1340. [Google Scholar]
  12. Dötterl S, David A, Boland W, Silberbauer-Gottserger I, Gottsberger G.. 2012. Evidence for behavioral attractiveness of methoxylated aromatics in a dynastid scarab beetle-pollinated Araceae. Journal of Chemical Ecology 38: 1539–1543. [DOI] [PubMed] [Google Scholar]
  13. Endress PK. 1984. The role of inner staminodes in the floral display of some relic Magnoliales. Plant Systematics and Evolution 146: 269–282. [Google Scholar]
  14. Fahn A. 1988. Secretory tissues in vascular plants. New Phytologist 108: 229–257. [DOI] [PubMed] [Google Scholar]
  15. Gang DR. 2005. Evolution of flavors and scents. Annual Review of Plant Biology 56: 301–325. [DOI] [PubMed] [Google Scholar]
  16. García MTA, Galati BG, Hoc PS.. 2007. Ultrastructure of the corona of scented and scentless flowers of Passiflora spp. (Passifloraceae). Flora 202: 302–315. [Google Scholar]
  17. Gibernau M. 2003. Pollinators and visitors of aroid inflorescences. Aroideana 26: 73–91. [Google Scholar]
  18. Gibernau M, Barabé D.. 2002. Pollination ecology of Philodendron squamiferum (Araceae). Canadian Journal of Botany 80: 316–320. [Google Scholar]
  19. Gibernau M, Barabé D, Cerdan P, Dejean A.. 1999. Beetle pollination of Philodendron solimoesense (Araceae) in French Guiana. International Journal of Plant Sciences 160: 1135–1143. [DOI] [PubMed] [Google Scholar]
  20. Gibernau M, Barabé D, Labat D.. 2000. Flowering and pollination of Philodendron melinonii (Araceae) in French Guiana. Plant Biology 2: 331–334. [Google Scholar]
  21. Gottsberger G. 1977. Some aspects of beetle pollination in the evolution of flowering plants. Plant Systematics and Evolution 1: 211–226. [Google Scholar]
  22. Gottsberger G, Silberbauer-Gottsberger I, Seymour RS, Dötterl S.. 2013. Pollination and floral scent differentiation in species of the Philodendron bipinnatifidum complex (Araceae). Plant Systematics and Evolution 299: 793–809. [Google Scholar]
  23. Govaerts R, Frodin DG.. 2002. World checklist and bibliography of Araceae (and Acoraceae). Kew: Royal Botanic Gardens. [Google Scholar]
  24. Hadacek F, Weber M.. 2002. Club-shaped organs as additional osmophores within the Sauromatum inflorescence – odour analysis, ultrastructural changes and pollination aspects. Plant Biology 4: 367–383. [Google Scholar]
  25. Heiduk A, Kong H, Brake I, et al. 2015. Deceptive Ceropegia dolichophylla fools its kleptoparasitic fly pollinators with exceptional floral scent. Frontiers in Ecology and Evolution 3: 66. doi:10.3389/fevo.2015.00066. [Google Scholar]
  26. Huber FK, Kaiser R, Sauter W, Schiestl FP.. 2005. Floral scent emission and pollinator attraction in two species of Gymnadenia (Orchidaceae). Oecologia 142: 564–575. [DOI] [PubMed] [Google Scholar]
  27. Jensen WA. 1962. Botanical histochemistry: principles and practice. San Francisco: Freeman. [Google Scholar]
  28. Johansen DA. 1940. Plant microtechnique. New York: McGraw-Hill. [Google Scholar]
  29. Karnovsky MJ. 1965. A formaldehyde–glutaraldehyde fixative of high osmolality for use in electron microscopy. Journal of Cell Biology 27: 137–138. [Google Scholar]
  30. Knudsen JT, Eriksson R, Gershenzon J, Ståhl B.. 2006. Diversity and distribution of floral scent. Botanical Review 72: 1–120. [Google Scholar]
  31. Leguet A, Gibernau M, Shintu L, et al. 2014. Evidence for early intracellular accumulation of volatile compounds during spadix development in Arum italicum L. and preliminary data on some tropical Aroids. Naturwissenschaften 101: 623–635. [DOI] [PubMed] [Google Scholar]
  32. Maia ACD, Schlindwein C, Navarro DMAF, Gibernau M.. 2010. Pollination of Philodendron acutatum (Araceae) in the Atlantic forest of Northeastern Brazil: a single scarab beetle species guarantees high fruit set. International Journal of Plant Sciences 171: 740–748. [Google Scholar]
  33. Maia ACD, Gibernau M, Dötterl S, et al. 2013. The floral scent of Taccarum ulei (Araceae): attraction of scarab beetle pollinators to an unusual aliphatic acyloin. Phytochemistry 93: 71–78. [DOI] [PubMed] [Google Scholar]
  34. Mayo SJ. 1991. A revision of Philodendron subgenus Meconostigma (Araceae). Kew Bulletin 46: 601–681. [Google Scholar]
  35. Melo MC, Borba EL, Paiva EAS.. 2010. Morphological and histological characterization of the osmophores and nectaries of four species of Acianthera (Orchidaceae: Pleurothallidinae). Plant Systematics and Evolution 286: 141–151. [Google Scholar]
  36. O’Brien TP, Feder N, McCully ME.. 1964. Polychromatic staining of plant cell walls by toluidine blue O. Protoplasma 59:368–373. [Google Scholar]
  37. Paiva EAS. 2016. How do secretory products cross the plant cell wall to be released? A new hypothesis involving cyclic mechanical actions of the protoplast. Annals of Botany 117: 533–540. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Paiva EAS, Machado SR.. 2008. The floral nectary of Hymenaea stigonocarpa (Fabaceae, Caesalpinioideae): structural aspects during floral development. Annals of Botany 101: 125–133. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Peng YB, Li YQ, Hao YJ, Xu ZH, Bai SN.. 2004. Nectar production and transportation in the nectaries of the female Cucumis sativus L. flower during anthesis. Protoplasma 224: 71–78. [DOI] [PubMed] [Google Scholar]
  40. Pereira J, Schlindwein C, Antonini Y, et al. 2014. Philodendron adamantinum (Araceae) lures its single cyclocephaline scarab pollinator with specific dominant floral scent volatiles. Biological Journal of the Linnean Society 111: 679–691. [Google Scholar]
  41. Pridgeon AM, Stern WL.. 1983. Ultrastructure of osmophores in Restrepia (Orchidaceae). American Journal of Botany 70: 1233–1243. [Google Scholar]
  42. Robards AW. 1978. An introduction to techniques for scanning electron microscopy of plant cells In: JL Hall, ed. Electron microscopy and cytochemistry of plant cells. New York: Elsevier, 343–403. [Google Scholar]
  43. Roland AM. 1978. General preparations and staining of thin sections In: JL Hall, ed. Electron microscopy and cytochemistry of plant cells. New York: Elsevier, 1–62. [Google Scholar]
  44. Sazima M, Vogel S, Cocucci A, Hausner G.. 1993. The perfume flowers of Cyphomandra (Solanaceae): pollination by euglossine bees, bellows mechanism, osmophores, and volatiles. Plant Systematics and Evolution 187: 51–88. [Google Scholar]
  45. Seymour RS, Bartholomew GA, Barnhart MC.. 1983. Respiration and heat production by the inflorescence of Philodendron selloum Koch. Planta 157: 336–343. [DOI] [PubMed] [Google Scholar]
  46. Skubatz H, Kunkel DD, Meeuse BJD.. 1993. The ultrastructural changes in the appendix of the Sauromatum guttatum inflorescence. Sexual Plant Reproduction 6: 153–170. [Google Scholar]
  47. Skubatz H, Kunkel DD, Patt J, Howald W, Rothman T, Meeuse BJD.. 1995. Pathway of terpene excretion by the appendix of Sauromatum guttatum. Proceedings of the National Academy of Sciences, USA 92: 10084–1088. [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. Skubatz H, Kunkel DD.. 1999. Further studies of the glandular tissue of the Sauromatum guttatum (Araceae) appendix. American Journal of Botany 86: 841–854. [PubMed] [Google Scholar]
  49. Vega C, Herrera CM, Dötterl S.. 2014. Floral volatiles play a key role in specialized ant pollination. Perspectives in Plant Ecology, Evolution and Systematics 16: 32–42. [Google Scholar]
  50. Vogel S. 1963. The role of scent glands in pollination. On the structure and function of osmophores (transl. by Bhatti JS.). Washington, DC: Smithsonian Institution Libraries. [Google Scholar]
  51. Vogel S. 1983. Ecophysiology of zoophilic pollination In: OL Lange, Nobel PS, CB Osmond, H Ziegler, eds. Physiological plant ecology III. Responses to the chemical and biological environment. New York: Springer, 559–624. [Google Scholar]
  52. Vogel S, Hadacek F.. 2004. Contributions to the functional anatomy and biology of Nelumbo nucifera (Nelumbonaceae) III. An ecological reappraisal of floral organs. Plant Systematics and Evolution 249: 173–189. [Google Scholar]
  53. Vogel S, Martens J.. 2000. A survey of the function of the lethal kettle traps of Arisaema (Araceae), with records of pollinating fungus gnats from Nepal. Botanical Journal of the Linnean Society 133: 61–100. [Google Scholar]
  54. Walker DB, Gysi J, Sternberg L, DeNiro MJ.. 1983. Direct respiration of lipids during heat production in the inflorescence of Philodendron selloum. Science 220: 419–421. [DOI] [PubMed] [Google Scholar]
  55. Wiemer AP, Moré M, Benitez-Vieyra S, Cocucci AA, Raguso RA, Sérsic AN.. 2009. A simple floral fragrance and unusual osmophore structure in Cyclopogon elatus (Orchidaceae). Plant Biology 11: 506–514. [DOI] [PubMed] [Google Scholar]
  56. Zito P, Dötterl S, Sajeva M.. 2015. Floral volatiles in a sapromyiophilous plant and their importance in attracting house fly pollinators. Journal of Chemical Ecology 41: 340–349. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary Data
Click here for additional data file. (16.4KB, docx)

Articles from Annals of Botany are provided here courtesy of Oxford University Press

RESOURCES