Botany Publication and Papers Botany

1972 Raphide Crystal Cell Development in Leaves of () Harry T. Horner Iowa State University, [email protected]

R. E. Whitmoyer Ohio Agricultural Research and Development Center

Follow this and additional works at: http://lib.dr.iastate.edu/bot_pubs Part of the Botany Commons, Other Sciences Commons, and the Plant Biology Commons

Recommended Citation Horner, Harry T. and Whitmoyer, R. E., "Raphide Crystal Cell Development in Leaves of Psychotria Punctata (Rubiaceae)" (1972). Botany Publication and Papers. 52. http://lib.dr.iastate.edu/bot_pubs/52

This Article is brought to you for free and open access by the Botany at Iowa State University Digital Repository. It has been accepted for inclusion in Botany Publication and Papers by an authorized administrator of Iowa State University Digital Repository. For more information, please contact [email protected]. Raphide Crystal Cell Development in Leaves of Psychotria Punctata (Rubiaceae)

Abstract The distribution and development of raphide crystal cells in nodulated leaves of Psychotria punctatawere studied by light and electron microscopy. Crystal cells in the leaf are oriented in various ways depending on whether they occur in the spongy or palisade parenchyma. Crystals are never found within the bacterial nodules and are not concentrated around them. Developing leaf crystal cells become larger than surrounding cells and have larger nuclei and nucleoli. Raphides develop within membrane chambers in the large central vacuole in association with membrane complexes, vesicles and tubules, the latter measuring 10-13 nm in diameter. Certain cytoplasmic organelles, the plasmalemma, and a cytoplasmic vacuolar channel complex also appear to be associated with crystal development. These results are compared with other recent investigations dealing with calcium oxalate crystals in higher .

Disciplines Botany | Other Plant Sciences | Plant Biology

Comments This article is from Journal of Cell Science 11 (1972): 339. Posted with permission.

This article is available at Iowa State University Digital Repository: http://lib.dr.iastate.edu/bot_pubs/52 J. Cell Sci. ii, 339-355 (1972) 339 Printed in Great Britain

RAPHIDE CRYSTAL CELL DEVELOPMENT IN LEAVES OF PSYCHOTRIA PUNCTATA (RUBIACEAE)

H. T. HORNER, JR Department of Botany and Plant Pathology, Iowa State University, Ames, Iowa 50010, U.S.A. AND R. E. WHITMOYER Electron Microscope Laboratory, Ohio Agricultural Research and Development Center, Wooster, Ohio 46691, U.S.A.

SUMMARY The distribution and development of raphide cry8tal cells in nodulated leaves of Psychotria punctata were studied by light and electron microscopy. Crystal cells in the leaf are oriented in various ways depending on whether they occur in the spongy or palisade parenchyma. Crystals are never found within the bacterial nodules and are not concentrated around them. Develop- ing leaf crystal cells become larger than surrounding cells and have larger nuclei and nucleoli. Raphides develop within membrane chambers in the large central vacuole in association with membrane complexes, vesicles and tubules, the latter measuring 10-13 nra ln diameter. Certain cytoplasmic organelles, the plasmalemma, and a cytoplasmic vacuolar channel complex also appear to be associated with crystal development. These results are compared with other recent investigations dealing with calcium oxalate crystals in higher plants.

INTRODUCTION The distribution and biological significance of calcium and inorganic calcium salt crystals within plants have been recently reviewed by Arnott (1972) and Arnott & Pautard (1970). Other recent studies specifically dealing with various aspects of cal- cium oxalate crystals in flowering plants represent an increased interest in this area of plant calcification (Arnott, 1966; Arnott & Pautard, 1965; Frank, 1967, 1969a, b; Frank & Jensen, 1970; Mollenhauer & Larson, 1966; Price, 1970; Schotz, Diers & Bathelt, 1970; Wattendorf, 1969). A number of investigators have used the characteristic appearance and distribution of calcium salt crystals for taxonomic purposes (see Arnott & Pautard, 1970; Metcalfe & Chalk, 1950). Solereder (1908) presented a diagnosis of genera and tribes of the dicotyledonous Rubiaceae which contain various types of calcium oxalate crystals. Verdcourt (1958) and Bremekamp (1966) divide the Rubiaceae into sub- families on the presence or absence of raphide crystals. Even though the number of subfamilies and the taxonomic delineation differ between their 2 schemes, both authors agree that the subfamily is characterized by the presence of rap- hide crystals. This paper deals with the distribution and development of raphide crystal cells in 340 H. T. Horner, Jr and R. E. Whitmoyer leaves of Psychotria punctata, a member of the subfamily Rubioideae. The study is an outgrowth of several earlier investigations on bacterial leaf nodulation in this species (Horner, 1972; Horner & Lersten, 1968, 1972; Lersten & Horner, 1967; Whitmoyer & Horner, 1970). A comparison of the present results with previous works on plant crystal ultrastructure shows certain anatomical and developmental differences not previously reported. Their significance and possible relevance to calcium oxalate crystal formation in general and leaves with bacterial nodules is presented.

MATERIALS AND METHODS Very young leaves just expanding from terminal and lateral buds and older leaves from a greenhouse-grown clone of Psychotria punctata Vatke (= P. bacteriophila Val.) were used. Leaves of various ages were removed and cleared by the method of Horner & Arnott (1963) to show the orientation and frequency of raphide crystal cells. Leaves from a number of herbarium collections of other species of Psychotria and Pavetta and Ardisia (2 other genera containing leaf-nodulated species) were also cleared to verify the presence of crystals. Some leaves were fixed in CRAF III or FAA, dehydrated in an ethanol-xylene series, and embedded in Tissue- mat (56-5 °C m.p.; Sass, 1958). Sections were cut at 7-10/tm and stained with safranin- chlorazol black E-fast green for general purposes. Plastic-embedded sections (see below) 1 /tm thick were stained with Paragon Multiple Stain (Spurlock, Skinner & Katrine, 1966) as an adjunct to electron microscopy. Photographs were taken with an Orthomat camera mounted on a Leitz Ortholux microscope fitted with bright-field, phase-contrast and polarization optics. Very young leaves were fixed in 3-5% glutaraldehyde in 01 or 005 M phosphate buffer (pH 72-y4) for 2-24 h at 4 °C, washed 3 times, 20 min each, in cold buffer, and then post- fixed in 1 % osmium tetroxide (same buffer and temperature) for 1-2 h. Fixed tissue was dehydrated in an ethanol-propylene oxide series, embedded in Araldite-Epon (Whitmoyer & Horner, 1970) and sectioned at 1 /tm or 50-80 nm on a Reichert ultramicrotome. Thin sections were placed either on 400-mesh unsupported grids or 150-mesh Formvar-supported grids and stained for 10—20 min at 25 °C with uranyl acetate dissolved in methanol or with lead citrate (same time and temperature). Observations were made on an Hitachi HU-i iC, RCA EMU-3F or a Siemens Elmiskop IA electron microscope.

OBSERVATIONS Light microscopy Location of crystals within the leaf. Raphide crystals typically occur in parenchyma cells of the various aerial organs, especially the leaves, and they are confined to the mesophyll. Generally, they are evenly distributed throughout the lamina except where nodules are present (Fig. 1). A through-focus sequence from adaxial to abaxial leaf surface shows that crystal cells in the palisade parenchyma are usually oriented per- pendicular to the leaf surfaces (Fig. 2) whereas they are randomly oriented to each other and parallel to the leaf surfaces in the spongy parenchyma (Figs. 3-5). Raphide crystals appear differentially birefringent when viewed with polarization optics because of their random orientation in the lamina. Some crystal cells in older leaves and other plant parts show a loss of birefringence due possibly to a reabsorption of the crystal material. Similar observations were made in Coprosma (Rubiaceae) (Stevenson, 1953). Development of crystal cells. As leaf area increases, so does the number of raphide crystal cells. Young leaves just expanding out of the bud were studied for crystal cell Psychotria raphide crystals 341 development since these leaves are very small and contain the various ontogenic stages in a relatively small area. Paradermal sections through the spongy lamina of a young leaf show the distribution of developing crystal cells around young leaf nodules containing bacteria (Fig. 5). The earliest indications that a parenchyma cell will become a crystal cell are 2-fold: the cell becomes larger than its neighbouring mesophyll cells, and its nucleus and nucleolus appear larger (Fig. 6), possibly denoting that endopolyploidy has occurred (the latter possibility is being checked by quantitative cytophometric methods). A young crystal cell begins to elongate and develops a large central vacuole with a dense peripheral cytoplasm. The large, conspicuous chloroplasts observed in sur- rounding normal mesophyll cells are not seen in crystal cells. Parallel crystals appear in the vacuole and increase in number and length in the direction of cell elongation (Fig. 7). Numerous dense-staining bodies are found next to the developing crystals in the vacuole; these bodies become more peripheral as the number of crystals in- creases (Figs. 8-10). A fully elongated crystal cell is completely packed with crystals while the nucleus and cytoplasm are peripheral. It is presently not known whether crystal cells in older leaves and those leaves near abscission contain intact cytoplasm or not.

Electron microscopy Young crystal cell before crystal formation. The most noticeable difference between young crystal cells and neighbouring mesophyll cells is that the latter have vacuoles which are large and relatively free of any internal structures (Fig. 11). In contrast, young crystal cell vacuoles are filled with a profusion of membrane structures and tubular elements (Fig. 11). The membrane structures comprise vesicles, reticulated membrane complexes, and looser membrane complexes. Scattered among these mem- brane structures are many arrays of tubules (Figs. 11-13). The peripheral cytoplasm contains a normal complement of organelles but with smaller and less organized plastids than surrounding cells (Fig. 14). Dictyosomes are numerous and appear to be producing many vesicles. The plasmalemma is quite irregular in outline as compared with adjacent mesophyll cells and, at places, invagi- nates into the crystal cell cytoplasm (Figs. 16, 17). These so-called plasmalemmasomes (Marchant & Robards, 1968) are quite common, each usually containing an internal membrane complex consisting of tubular membrane elements folded in upon each other. Aside from this internal membrane complex, a granular body sometimes is evident which is not membrane bound. At certain places, the plasmalemmasomes can be seen protruding from the cytoplasm into the vacuole, still surrounded by the tonoplast (Fig. 17). No comparable structures with both enclosed membrane tubules and granular body have been observed unattached in the vacuole proper. However, there are numerous membrane complexes in the vacuole proper that display a loose membrane system (Fig. 18). The tonoplast is also irregular, producing vacuolar channels which extend deeply into the peripheral cytoplasm (Fig. 14). These channels are filled with vesicles and, at certain places, exhibit reticulate membrane complexes that appear to be extensions 342 H. T. Homer, Jr and R. E. Whitmoyer of the channel tonoplast (Fig. 15). Some of the complexes have dense regions asso- ciated with the membrane reticulum. It is not known whether the plasmalemmasomes and these latter reticulated membrane complexes are spatially associated with each other. However, they appear in the vacuole in increasing numbers indicating that their movement must be inward (Figs. 16, 18). The tubules which are present in the vacuole proper were never observed in the vacuolar channels or in the cytoplasm, even though microtubules occur in the cytoplasm near the cell wall. Crystal formation. Individual crystals appear first in the centre of the vacuoles and are usually associated with the membrane complexes and vesicles (Fig. 18). The complexes are still similar to the 2 forms described previously except that the reticu- lated membrane complexes typically display dense, compact central membrane bodies from which extend the reticulate membranes (Figs. 16, 18). These latter structures are comparable to the dense-staining bodies observed with light microscopy (Figs. 8-10). Along with looser membrane complexes, they are associated with developing crystal chambers which rapidly increase in number and size (Fig. 18). Each chamber is surrounded by a membrane and all develop in one direction, parallel to the long axis of the elongating cell. The vesicles are scattered among the membrane complexes and crystal chambers, but there is no evidence that the vesicles fuse with them. Both plasmalemmasomes and vacuolar channel membrane complexes are observed in various cytoplasmic forms during this time. The vacuolar tubules, measuring 10-13 nra m diameter, are also found throughout the vacuole (Figs. 16, 18). They usually, but not always, parallel the long axis of the crystal chambers and occasionally may be in close association with all of the vacuolar membrane structures. At high magnification (Figs. 12, 13), the tubules display a hollow core but are much smaller in diameter than published reports for cytoplasmic microtubules (25 nm; Porter, 1966) and Pi-protein tubules (18-24 nm> Cronshaw & Esau, 1967; Northcote & Wooding, 1966). They do fall into the range of P2-protein striated fibrils (9-15 nm; Cronshaw & Esau, 1967; Northcote & Wooding, 1966). The walls of the vacuole tubules do show substructure. Rotational analyses (unpub- lished) have been carried out with n = 4 being the more common reinforcement number. A comparison of the vacuolar tubules with cytoplasmic tubules in surrounding cells was made in order to help document their presence and relative size. Aside from microtubules found within the crystal cell, young sieve tube elements in nearby leaf veins contained both cytoplasmic P-protein tubules and microtubules. Measurements of the diameter of these tubules showed dimensions of about 20 nm and 25 nm, res- pectively. The cytoplasm of the developing crystal cell is dense and contains many organelles. Plastids are smaller than in adjacent mesophyll cells and have simplified lamellar systems with little or no stacking of membranes. Starch is absent. Mitochondria with well developed cristae are numerous. Dictyosomes are also abundant. The many vesicles attached to the cisternae indicate much activity while throughout the cyto- plasm there is endoplasmic reticulum (ER), both smooth and rough. Cytoplasmic ribosomes are also plentiful. Sometimes microbodies are partially encircled by the ER, the former being no more numerous than in adjacent mesophyll cells. Psychotria raphide crystals 343 Older crystal cells. Eventually the crystal cell reaches a maximum size and its vacuole fills with crystals (Fig. 19). At least in young leaves where this occurs, the peripheral cytoplasm is still present. The vacuolar channels and vacuole proper have a paucity of vesicles and membrane complexes while the vacuolar tubules are no longer evident. The dictyosomes appear less active and the plasmalemma is more regular.

DISCUSSION Solereder (1908) used the shape of calcium oxalate crystals to group the various tribes in the Rubiaceae. The types of crystals which he designated were: crystal-sand, clustered crystals, styloids (acicular crystals) and raphides. Cystoliths (composed of calcium carbonate) are apparently absent in the family. Verdcourt (1958) and Bremekamp (1966) taxonomically redefined the Rubiaceae but only used the presence or absence of raphide crystals to separate the subfamilies initially. They both agree that all members of the subfamily Rubioideae have raphides. These results were supported from leaf clearings, bud longisections of over 100 species, and many gross visual observations of herbarium specimens of nodulated and un- nodulated Psychotria species (N. R. Lersten, personal communication). Bremekamp (1966) states, 'That the presence of raphides is to be regarded as a taxonomically very important feature, follows from the fact that it is in the tribes in which it occurs, a general character, whereas this kind of crystals is completely absent in the other tribes of the Rubiaceae and also in all the families which are to be regarded as their nearest allies'. It was of interest to determine whether other leaf-nodulated species, aside from those in Psychotria, contained raphide crystals. The 2 other genera which we recognize as having true bacterial leaf-nodules (Homer & Lersten, 1972) are Pavetta (Rubiaceae, subfamily Ixoroideae) and Ardisia (Myrsinaceae). Pavetta, by its position in another subfamily of the Rubiaceae (and from clearings) does not contain raphide crystals but has crystal-sand (crystal-sand alone or together with clustered crystals; Solereder, 1908). Likewise, Ardisia does not have raphide crystals but contains druse crystals which were observed in leaves from herbarium collections. From these results, it is apparent that leaves of all leaf-nodulated taxa observed contain crystals of some type. It should be pointed out that all unnodulated species observed in these 3 genera also have their respective crystals. Such evidence would indicate that the presence and type of crystal are independent of the bacterial nodules. However, this does not affect their use as a taxonomic character. In Psychotria punctata the raphide crystal cells are present in a rather uniform manner throughout the leaf lamina. They are restricted to mesophyll cells and are not found within the leaf nodules. The increase in size of crystal cell, nucleus and nucleolus as compared to adjacent non-crystal forming cells denotes the first visible changes. Hurel-Py (1938, 1942) also observed enlarged nuclei in raphide-forming cells in some orchids in which the crystals form in the vacuole. Even though it has not been estab- lished, the enlarged nuclei in these cells may indicate that DNA replication has occurred. 344 H- T. Homer, Jr and R. E. Whitmoyer The absence of crystal cells within developing bacterial nodule mesophyll cells is worth noting since these cells originate from the lamina parenchyma cells and even- tually become a part of a special nodule microenvironment within the leaf. If there is any influence of the presence of the nodules on crystal formation within the adjacent lamina, it is not apparent since the crystals are evenly distributed throughout and not absent or clustered around the nodules. The orientation of an individual crystal cell within a leaf appears to be governed only by its position in the spongy or palisade parenchyma. It is clear at both the light and electron-microscope levels that the Psychotria raphide crystals form within the vacuole proper. This agrees with the origin and location of other calcium oxalate crystals recently reported by other investigators, except for Wattendorf (1969). The appearance and contents of the vacuoles in young crystal cells of Psychotria are strikingly different from vacuoles of adjacent mesophyll cells. Early in develop- ment the vacuoles become filled with the various membrane structures and tubules which apparently are associated with crystal chamber and crystal formation. Most investigators agree that there is a membrane surrounding each crystal, while in several instances, some kind of membrane complex or vesicle is shown in the vacuole asso- ciated with the crystal chambers (Arnott, 1966, 1972; Arnott & Pautard, 1965, 1970; Frank & Jensen, 1970; Mollenhauer & Larson, 1966; Price, 1970; Schotz et al. 1970). In Psychotria, 3 types of membrane structures are apparent: reticulate membrane complex, looser membrane complex, and vesicle. The origin of these membrane struc- tures and their roles in crystal formation are not completely understood at the present time. However, from our results it appears that these structures originate from the cytoplasm, plasmalemma, and tonoplast. The persistent and evident vacuolar channel network extending throughout the cytoplasm contains vesicles and reticulate mem- brane complexes during crystal formation. Even though there were many dictyosomes producing vesicles, no vesicles were seen fused with the tonoplast of the channels or vacuole proper. Schotz et al. (1970) feel that the ER and dictyosomes may be responsible for the vesicles, crystal chambers and calcium oxalate crystals in Oenothera. This was also suggested for Spirodela, with the contribution of surface-derived pinocytotic vesicles (Ledbetter & Porter, 1970). In Psychotria, the ER, even though it was quite evident, did not appear to produce any vesicles. The invaginations of the plasmalemma producing the so-called plasmalemmasomes (Marchant & Robards, 1968) were a consistent feature during crystal cell development. We feel they are real and represent membrane structures that are transporting some- thing into the cells and, more specifically, into the vacuoles. Dengg (1971), in an ultrastructural study of plant galls on Urtica, observed cells with druse crystals con- taining membrane structures similar to the plasmalemmasomes in Psychotria. She called them lomasomes. Our use of the former term follows the definition by Marchant & Robards (1968). It is important to emphasize that the crystal cells are undergoing a much higher rate of activity and enlargement than adjacent cells. Therefore, it seems most reasonable Psychotria raphide crystals 345 that membrane structures be produced for the rapid influx of materials necessary for the various metabolic activities associated with crystal formation. Several recent investigators support the concept of an endocytic phenomenon (Mahlberg, Olson & Walkinshaw, 1970, 1971) in which a membrane complex is formed from the plasma- lemma and invaginates into the cytoplasm and vacuole. We feel that our results support this concept. However, whether the plasmalemmasomes in Psychotria are different from the reticulate membrane complexes observed in the vacuolar channels and vacuole proper is not known at this time. The appearance of what may be 2 types of membrane complexes in the vacuole proper suggests that this is so. The presence of many tubules (10-13 nm m diameter) adds another structure to the complex crystal vacuole system. Frank & Jensen (1970) show tubules or fibrils forming a boundary associated with the open end of the crystal chamber in Canavalia. They do not give any further information about the tubules but suggest they '. . . are in- volved in the transportation of material from the cytoplasm to the site of crystal forma- tion in the vacuole'. The Psychotria crystal vacuole tubules definitely have a hollow centre and are smaller in diameter than reports for Pi-protein tubules or microtubules. They do fall within the range of cytoplasmic P2-protein striated fibrils (9—15 nm; Palevitz & Newcomb, 1971) but any indication of similarities at this time would be premature. The presence and general orientation of the tubules suggest that they are assembled in the vacuole proper and may aid both in the orientation of the developing crystal chambers and in movement of the various vacuolar structures. The plastids are smaller than normal mesophyll chloroplasts and contain a simplified lamellar system and no starch similar to that reported for some other crystal cell plastids (Arnott & Pautard, 1970; Price, 1970). The plastids do not contain specialized regions as reported by Arnott (1966) in Yucca and by Mollenhauer & Larson (1966) in Vanilla and Monstera roots. Microbodies are apparent within the cytoplasm and often associated with the ER but are no more frequent than in adjacent mesophyll cells (Frank & Jensen, 1970; Mollenhauer & Larson, 1966). In summary, the Psychotria crystal cells appear to contain various structures not all previously described in any one crystal cell. Both the vacuole and cytoplasm, by the very presence and number of structures, appear quite active and organized. However, it is not understood at this time what the relationships of the various structures are to each other or how these structures and materials arrive in the vacuole for crystal formation. Further work is being carried out to clarify these points. We wish to thank Mrs Sandy Blake for typing the manuscript.

REFERENCES ARNOTT, H. J. (1966). Studies of calcification in plants. In Hlrd Eur. Symp. on Calcified Tissues (ed. H. Fleisch, H. J. Blackwood & M. Owen), pp. 152-157. Berlin: Springer-Verlag. ARNOTT, H. J. (1972). Plant calcification. In Biological Mineralization (ed. I. Zipkin). New York: John Wiley & Sons. (In the Press.) ARNOTT, H. J. & PAUTARD, F. G. E. (1965). Development of raphide idioblasts in Lemna. Am. J. Bot. 53, 618 (Abstr.). ARNOTT, H. J. & PAUTARD, F. G. E. (1970). Calcification in plants. In Biological Calcification (ed. H. Schraer), pp. 375-446. New York: Appleton-Century-Crofts. 346 H. T. Homer, Jr and R. E. Whitmoyer BREMEKAMP, C. E. G. (1966). Remarks on the position, the delimitation and the subdivision of the Rubiaceae. Acta bot. neerl. 15, 1-33. CRONSHAW, J.&ESAU, K. (1967). Tubular and fibrillar components of mature and differentiating sieve elements, jf. Cell Biol. 34, 801-815. DENGG, E. (1971). Die Ultrastruktur der Blattgalle von Dasyneura urticae auf Urtica dioica. Protoplasma 73, 367-379. FRANK, E. (1967). Zur Bildung des Kristallidioblastenmusters bei Canavalia ensiformis DC. I. Z. Pflanzenphysiol. 58, 33-48. FRANK, E. (1969a). Zur Bildung des Kristallidioblastenmusters bei Canavalia ensiformis DC. II. Zur Zellteilung in der Epidermis. Z. Pflanzenphysiol. 60, 403-413. FRANK, E. (19696). Zur Bildung des Kristallidioblastenmusters bei Canavalia ensiformis DC. III. Gehalt an Oxalat, Stickstoff und Trockengewicht im Verlauf der Blattentwicldung. Z. Pflanzenphysiol. 61, 114-121. FRANK, E. & JENSEN, W. A. (1970). On the formation of the pattern of crystal idioblasts in Canavalia ensiformis DC. IV. The fine structure of the crystal cells. Planta 95, 202-217. HORNER, H. T., JR. (1972). Leaf. Yearbook of Science and Technology. New York: McGraw-Hill. HORNER, H. T., JR. & ARNOTT, H. J. (1963). Sporangial arrangement in North American species of Selaginella. Bot. Gaz. 124, 371-383. HORNER, H. T., JR. & LERSTEN, N. R. (1968). Development, structure and function of secretory trichomes in Psychotria bacteriophila (Rubiaceae). Am. J. Bot. 55, 1089-1099. HORNER, H. T., JR. & LERSTEN, N. R. (1972). Nomenclature of bacteria in leaf nodules of the Myrsinaceae and Rubiaceae. Int. J. syst. Bad. 22, 117-122. HUREL-PY, G. (1938). litude des moyaux vegetatifs de Vanilla planifolia. Revue Cytol. Cyto- physiol. veg. 3, 129-133. HUREL-PY, G. (1942). Sur les vacuoles des cellules a raphides. C. r. Seanc. Soc. Biol. 215, 31-33. LEDBETTER, M. & PORTER, K. (1970). Introduction to the Fine Structure of Plant Cells. Berlin: Springer-Verlag. LERSTEN, N. R. & HORNER, H. T., JR. (1967). Development and structure of bacterial leaf nodules in Psychotria bacteriophila Val. (Rubiaceae). J. Bact. 94, 2027-2036. MAHLBERG, P. G., OLSON, K. & WALKINSHAW, C. (1970). Development of peripheral vacuoles in plant cells. Am. J. Bot. 57, 962-968. MAHLBERG, P., OLSON, K. & WALKINSHAW, C. (1971). Origin and development of plasma mem- brane derived invaginations in Vinca rosea L. Am. J. Bot. 58, 407-416. MARCHANT, R. & ROBARDS, A. W. (1968). Membrane systems associated with the plasmalemma of plant cells. Ann. Bot. 32, 457-471. METCALFE, C. R. & CHALK, L. (1950). Anatomy of the Dicotyledons, vol. 1, 2. Oxford: Clarendon Press. MOLLENHAUER, H. H. & LARSON, D. A. (1966). Developmental change in raphide-forming cells of Vanilla planifolia and Monstera delidosa. J. Ultrastruct. Res. 16, 55-70. NORTHCOTE, D. H. & WOODING, F. B. P. (1966). Development of sieve tubes in Acer pseudo- platanus. Proc. R. Soc. B 163, 524-537. PALEVITZ, B. A. & NEWCOMB, E. H. (1971). The ultrastructure and development of tubular and crystalline p-protein in the sieve elements of certain Papilionaceous legumes. Protoplasma 72, 399-426. PORTER, K. R. (1966). Cytoplasmic microtubules and their functions. In Principles of Bio- molecular Organization, Ciba Fdn Symp. (ed. G. E. W. Wolstenholme & M. O'Connor), pp. 308-345. Boston: Little, Brown & Co. PRICE, J. L. (1970). Ultrastructure of druse crystal idioblasts in leaves of Cercidium floridum. Am.J. Bot. 57, 1004-1009. SASS, J. E. (1958). Botanical Microtechnique, 3rd edn. Ames: Iowa State Univ. Press. SCHOTZ, F., DIERS, L. & BATHELT, H. (1970). Zur Feinstruktur der Raphidenzellen. I. Die Entwicklung der Vakuolen und der Raphiden. Z. Pflanzenphysiol. 63, 91-113. SOLEREDER, H. (1908). Systematic Anatomy of the Dicotyledons. Oxford: Clarendon Press. SPURLOCK, B. O., SKINNER, M. S. & KATTINE, A. A. (1966). A simple rapid method for staining epoxy-embedded specimens for light microscopy with the polychromatic stain Paragon-1301. Am. J. din. Path. 46, 252-258. Psychotria raphide crystals 347 STEVENSON, G. B. (1953). Bacterial symbiosis in some New Zealand plants. Ann. Bot. 17, 343-345- VERDCOURT, B. (1958). Remarks on the classification of the Rubiaceae. Bull. yard. bot. Etat Brux. 28, 209-281. WATTENDORF, J. (1969). Feinbau und Entwicklung der verkorkten Calciumoxalatkristallzellen in der Rinde von Larix decidiia Mill. Z. Pflanzenphysiol. 60, 307-347. WHITMOYER, R. E. & HORNER, H. T., JR. (1970). Developmental aspects of bacterial leaf nodules in Psychotria bacteriophila Val. (Rubiaceae). Bot. Gaz. 131, 193-200.

{Received 16 February 1972) 348 H. T. Homer, Jr and R. E. Whitmoyer

Fig. i. Leaf clearing showing the distribution of raphide crystal cells within the lamina and in relation to a bacterial nodule (n) and leaf veins, x 71. Fig. 2. Leaf clearing. Plane of focus at level of palisade parenchyma. Crystal cells are perpendicular to leaf surfaces, x 164. Fig. 3. Leaf clearing. Plane of focus at level of spongy parenchyma and veins. Crystal cells are randomly oriented to each other and parallel to leaf surfaces, x 164. Fig. 4. Portion of cross-section of mature leaf lamina displaying a crystal cell in spongy parenchyma and a part of one crystal cell in palisade parenchyma (arrow), x 425. Fig. 5. Paradermal plastic section through young leaf lamina with several crystal cells at various stages of development. Two young leaf nodules (arrows) with their enclosed intercellular bacteria are evident, x 643. Fig. 6. Very young crystal cell which is larger than surrounding mesophyll cells. Large vacuole does not contain any crystals, x 1500. Psychotria raphide crystals 350 H. T. Horner, JrandR.E. Whitmoyer

Fig. 7. Young crystal cell with enlarged nucleus and nucleolus. Crystals are present in central vacuole. x 1500. Fig. 8. Numerous dense bodies (arrows) are evident in vacuole around developing crystals, x 1500. Fig. 9. Crystal cell vacuole is filled with crystals and cytoplasm is appressed to cell wall. Nucleus is flattened, x 1500. Fig. 10. Older crystal cell is many times larger than adjacent mesophyll cells. Dense bodies are still present in vacuole. x 1500. Fig. 11. Young crystal cell prior to crystal formation. Vacuole stains differently from adjacent mesophyll cells (right) and contains various membrane structures and tubules (arrows). Peripheral cytoplasm lacks chloroplasts containing starch as shown in lower right adjacent cell, x 8900. Fig. 12. Cross-sections of crystal vacuole tubules showing hollow centre. Tubules measure between 10-13 nm m diameter, x 315000. Fig. 13. Longitudinal section of crystal vacuole tubules, x 315000. Psychotria raphide crystals 352 H. T. Homer, Jr and R. E. Whitmoyer

Fig. 14. Portion of crystal cell cytoplasm displaying ribosomes, mitochondria (>»), endo- plasmic reticulum (er), plastid, vesicles and vacuolar channel complex (vc) with included vesicles. Vacuole proper (v) is at lower right with several crystal chambers (cc) evident. Arrow depicts one place where central vacuole is connected with vacuolar channel complex, x 41 500. Fig. 15. Reticulate membrane complex within a portion of vacuolar channel complex. Dense region denotes where some of the membranes are compacted to form a dense body, x 52800. Fig. 16. Periphery of crystal cell showing wall, cytoplasm with plasmalemmasome (arrow) and vacuole containing tubules, membrane complexes and a few vesicles, x 36000. Fig. 17. Plasmalemmasome extending into the vacuole proper. Cytoplasm and cell wall are evident at the bottom. Tonoplast surrounds outer portion of plasmalemmasome. x 32600. Psychotria raphide crystals

r

23-2 354 H. T. Horner, Jr and R. E. Whitmoyer

Fig. 18. Centre of crystal-cell vacuole where the first crystals appear. The developing crystal chambers are in close association with the membrane structures, loose com- plexes (Z), reticulate complexes (r), and vesicles. Numerous tubules are visible in various orientations, x 37800. Fig. 19. Older crystal cell. Vacuole nearly filled with crystals. Peripheral cytoplasm still displays dictyosomes. Vacuolar channels and vacuole do not contain as many vesicles. Membrane complexes are not as numerous, x 13400. Psychotria raphide crystals 355

i