IAWA Bulletin n.s., Vol. 4 (2-3), 1983 131

DEVELOPMENTAL CHANGES IN THE WOOD OF BOCCONIA VULCANICA DONN. SMITH

by

Billy G. Cumbie Division of Biological Sciences, University of Missouri, Columbia, Missouri 65211, U. S. A.

Summary Developmental changes in the xylem were xylem differs in a number of respects from studied in a stem of Bocconia vulcanica Donn. the wood in primary woody plants (Carlquist, Smith with a xylem radius of 3.0-4.5 cm. 1974). Growth rings are absent. The vascular cambium One of the families of dicotyledons that is nonstoried with fusiform initials averaging shows the transformation of herbaceous plants 282 J.Lill long. The specialised vessel members into woody plants is the Papaveraceae (Takhta­ are short, with oblique to transverse end walls, jan, 1980). This family consists predominantly simple perforations, and alternate intervascular of herbs, although species of Dendromecon pitting. Vessels are relatively uniform in diame­ tend to be shrubby, while those of Bocconia ter and arrangement throughout the wood. are shrubs, small trees, or woody herbs (Law­ Fibres have moderately thin walls and do not rence, 1951). Bocconia consists of ten species increase in length from the primary xylem to in tropical America (Standley & Steyermark, the cambium. Axial parenchyma is paratracheal, 1946). Halle et al. (1978) list Bocconia as an scanty to vasicentric. Rays are exclusively mul­ example of Corner's model, a monocaulous tiseriate, tall, and heterocellular with a predo­ tree with a single aerial meristem producing an minance of erect and square cells. Sheath cells unbranched axis with lateral inflorescences. occur along the sides. There are no fibres in the However, Standley and Steyermark (1946) de­ secondary phloem and a periderm is not pres­ scribe the inflorescence in Bocconia as being ent. The xylem and bark are similar in many terminal. A plant of Bocconia frutescens L., respects to that formed in some groups of dico­ growing in a greenhouse, is monocaulous, con­ tyledons that are basically herbaceous with sisting of a linear sympodium. After producing evolution toward woodiness. a stem with several internodes, a terminal inflo­ Key words: Secondary xylem, woodiness, rays, rescence forms. A new lateral branch forms paedomorphosis. at a node at the base of the inflorescence, and pushes the inflorescence to one side as it grows Introduction upward. Thus, Bocconia is an example of the One of the major trends of evolution in the Chamberlain model (Halle et aI., 1978). dicotyledons has been the derivation of herba­ This paper describes the developmental ceous species from woody forms. However, this changes occurring in the wood of Bocconia vul­ evolutionary trend is not irreversible. Carlquist canica Donn. Smith, a small tree 3-8 m tall, (1966, 1974), Cronquist (1968), Stebbins from the mesic mountain forests of Guatemala. (1974), and Takhtajan (1980) point out that Some observations are also made of the bark. trends in growth habit have been reversed in The main objective is to compare the structure several unrelated families of dicotyledons that of the xylem in Bocconia with that in second­ are basically herbaceous, including, for exam­ arily woody species of dicotyledons to deter­ ple, the Compositae, Campanulaceae, Cheno­ mine if it is likely that this woody form has podiaceae, and Phytolaccaceae. The evolution evolved from herbaceous ancestors. of woody forms from herbaceous ancestors has been in response to highly specific ecological Materials and Methods conditions, such as adaptation to mesic, rela­ Material of Bocconia vulcanica was collected tively uniform conditions on islands and conti­ by W.E. Harmon, 3592 (University of Missouri nents (Carlquist, 1974). Herbarium, 92676), August 6, 1970, from a Woody species that are derived secondarily mesic mountain forest on the north slopes of from herbaceous forms have limited cambial Volcano Agua, Sacatepequez, Guatemala. From activity and are typically smaller than primary the material, a dried oval stem 8.5 x 10 cm in woody plants. Although they may be similar in diameter was used for this study. The stem had growth form to truly woody plants, they in­ a massive pith, 1.4 cm in diameter, and a xylem clude unusual forms such as rosette trees and radius of 3.0 x 4.5 cm. A relatively thick bark shrubs, and 'woody' herbs. The secondary covered the stem.

Downloaded from Brill.com09/28/2021 08:13:39AM via free access 132 IAWA Bulletin n.s., Vol. 4 (2-3),1983

Along the 4.5 cm xylem radius, two rectan­ Usually there are no vessels in these strands. gular blocks 7 mm wide and about 10 mm deep Lateral to these strands, and separating them were cut. One block was divided into five from the large strands containing vessels, are pieces, each with a radial thickness of 9 mm, interfascicular areas or medullary rays 2-3 cells and prepared for sectioning using standard par­ wide (Fig. I). These extend the entire height of affin techniques (Sass, 1958). Several trans­ the sections of about 10 mm, and consist of verse, tangential, and radial sections were cut thin-walled, extremely erect cells. from each piece and stained with hematoxylin Growth rings are absent in the secondary xy­ and safranin. The other block was used to pre­ lem. Vessel members I mm out from the pri­ pare macerations at 5 mm intervals along the mary xylem are markedly shorter than those in xylem radius using Jeffrey's method (Sass, the outer primary xylem (Fig. 13). During the 1958). Similar blocks were cut along the 3 mm production of the first 20 mm of secondary xy­ radius. After dividing one block into six pieces lem, vessel member length shows a slight but (five wood and one bark) and embedding in continuous decrease. Subsequently, vessel mem­ paraffin, several transverse sections were cut ber length remains relatively constant at about off each piece. Each piece was then remounted 290 pm (Fig. 13). The vessel members are spe­ and completely cut into serial tangential sec­ cialised, with oblique to transverse end walls, tions 10 pm thick (about 3000 wood and 720 simple perforations, and rather large, alternate bark sections). The other block along the 3 cm intervascular pitting. During the formation of radius was used to prepare macerations at 5 mm the first mm of wood, vessel diameter increases intervals. Some additional transverse sections to an average of 120 pm, and remains uniform were cut from small pieces of wood embedded throughout secondary growth (Fig. 13). The in glycol methacrylate (JB-4, Polysciences Inc., distribution of vessels is about the same in all Warrington, PA). These sections, 3-5 J.1.m thick, the secondary xylem. Solitary vessels, multiples were stained with toluidine blue. and small clusters occur in about equal num­ Average values for secondary xylem vessel bers. In the inner 5 mm of wood, there are ap­ member length and fibre length are based on proximately 12-13 vessels/mm2, whereas in 50 measurements for each maceration sample. the remainder of the wood the frequency is on­ The average length of primary xylem vessel ly about 5-6/mm2 (Figs. 2, 3). members and fibres are based on 10-25 mea­ Moderately thin-walled fibres occur through­ surements from serial tangential sections. Ves­ out the wood (Figs. 1-3). They possess sparse, sel diameter and frequency were determined minute, simple pits. Although there is consider­ from the entire transverse sections, and the de­ able variation in fibre length (Fig. 13), the mean velopmental changes in ray structure were de­ length is rather uniform throughout secondary termined from the serial tangential sections. growth, averaging 625 J.1.m. Axial parenchyma With one exception, terminology used fol­ is para tracheal, scanty to vasicentric through­ lows the recommendations of the Committee out the secondary xylem (Figs. 2, 3). The thin­ on Nomenclature, lAW A (1964). In the case walled parenchyma consists mainly of fusiform of rays, initials of individual cells within a ray cells with a few strands of two cells. are called ray cell initials, whereas the entire At the beginning of secondary growth, pri­ group of such initials is called ray initial. This mary rays (Barghoorn, 1940) originate opposite modification of accepted terminology, first the medullary rays (Fig. I). These are similar to proposed by Cheadle and Esau (1964), is neces­ the medullary rays, being 2-3 cells wide and sary when describing the transformation of fu­ most extending the entire height of the sections. siform initials into ray initials. All cells are erect. Primary rays are absent in the fascicular areas of the xylem (Fig. I). During Results the formation of the first few mm of wood, the Primary xylem vessels occur in widely spaced primary rays increase in width by the conver­ strands surrounding the massive pith. These sion of fusiform cambial initials into ray cell strands are typically large and contain numer­ initials along the sides of the ray initials (Fig. 4). ous metaxylem vessels (Fig. I). The vessel mem­ These erect cells along the sides of the rays are bers show a fairly abrupt decrease in length, sheath cells, and are quite similar to fibres. As a from 700 to about 540 pm, progressing from result, the limits of the rays are difficult to de­ the inner to the outer metaxylem, and an in­ termine, particUlarly in transverse sections (Figs. crease in diameter from 47 to 90 pm (Fig. 13). 1, 2). At the same time that the rays are in­ In the central region between these strands are creasing in width, transverse divisions in ray additional strands consisting of wide, fibrous cell initials decreases the height of the ray cells. cells (Fig. I) with transverse or abruptly tapered While the primary rays are undergoing devel­ end walls. These cells average 800 pm long. opmental changes, some secondary rays appear

Downloaded from Brill.com09/28/2021 08:13:39AM via free access IAWA Bulletin n.s. , Vol. 4 (2-3), 1983 133

Fig. 1-4. Bocconia vulcanica Donn. Smith. Bar equals 250 J.Lm . - 1-3: Transverse sections of xylem. - I: Primary xylem and inner secondary xylem. Primary xylem consists of strands contain­ ing numerous metaxylem vessels alternating with strands consisting of fibres only. - 2: Secondary xylem 0.6 mm from primary xylem. - 3: Secondary xylem 25 mm from primary xylem. - 4: Tan­ gential section 2.0 mm out from primary xylem showing three primary rays and one secondary ray (second from left). Arrows indicate where rays will be broken apart into smaller rays.

Downloaded from Brill.com09/28/2021 08:13:39AM via free access 134 IAWA Bulletin n.s., Vol. 4 (2-3),1983

Fig. 5-8. Bocconia vulcanica Donn. Smith. Bar equals 250 pm. - 5-7: Tangential sections of secondary xylem. - 5: Xylem 6.0 mm out from primary xylem. Rayon left broken apart by elon­ gation of ray cell initials. One new secondary beginning to form from two cells. - 6: Xylem 10 mm out from primary xylem. Rays markedly heterocellular. - 7: Xylem 28 mm out from primary xylem. - 8: Transverse section of nonfunctional phloem about 2 mm out from cambium showing intact axial parenchyma cells and tangentially expanded ray cells.

Downloaded from Brill.com09/28/2021 08:13:39AM via free access lAW A Bulletin n.s., Vol. 4 (2-3), 1983 135

Fig. 9-10. Bocconia vulcanica Donn. Smith. Bar equals 250 JJ.ITI. - 9: Tangential section of non­ functional phloem showing intact axial parenchyma and rays. - 10: Transverse section of outer bark showing primary phloem fibres and outer parenchyma.

during the formation of approximately the in­ initials do not shorten by transverse divisions ner 7 mm of wood (Fig. 4). These rays regular­ (Fig. 4, arrows). These erect cells elongate to ly originate about midway between the primary become fusiform initials. By the time the inner rays. Secondary rays are formed by ray initials 10 mm of xylem is formed, most rays are about which arise by subdivision of fusiform initials 1.3 mm tall (Fig. 6). Subsequently, ray height in the cambium. Generally, a ray first appears is fairly constant, but the rays continue to in­ as a short ray 1-2 cells wide by the conversion crease in width throughout secondary growth. of a small group of tangentially and axially At a radial distance of 20 mm from the primary contiguous fusiform initials into ray cell initials xylem, the rays are 4-16 (ave. II) cells wide, by transverse divisions (Fig. II a-d). The aver­ and at a distance of 28 mm the rays are 6-17 age length of the fusiform cambial cells giving (ave. 12) cells wide (Fig. 7). Thus, the rays are rise to new ray cell initials during the forma­ multiseriate, fusiform in shape, and markedly tion of the inner secondary xylem is 290 JJ.ITI, heterocellular throughout the wood. The cen­ as reflected in the length of the cells when they tral portion of a ray consists of groups of small first appear in the xylem (Fig. 11 a). The sec­ procumbent cells surrounded by square and ondary rays increase in height by the addition erect cells .. Sheath cells occur along the sides of new ray cell initials at one or both margins throughout the wood and, because they are of the ray initial (Fig. II e, f, lower margin of very similar to the fibres, make it difficult to ray). After the height of the ray is established, recognise the limits of the rays, even in the the secondary rays increase in width in the same outermost wood (Fig. 3). The ray cells possess manner as the primary rays, and in a short ra­ thin secondary walls and are angular in outline dial distance become very similar to the pri­ throughout the secondary xylem (Figs. 4-7). mary rays. At the beginning of secondary growth, pri­ As the rays increase in width, they are broken mary rays occupy about 19% of the secondary up into shorter rays. In some regions of the ex­ xylem. This increases to 27% during the pro­ tremeIy tall ray initials, a few of the ray cell duction of the inner 10 mm of wood by the in-

Downloaded from Brill.com09/28/2021 08:13:39AM via free access 136 IAWA Bulletin n.s., Vol. 4 (2-3),1983

d d ® @)

Fig. 11-12. Camera lucid a drawings from tangential sections of xylem illustrating the formation of secondary rays. The first drawing in each series shows the precursors of the rays when first observed in the xylem. x 50. - 11: Ray originating 2.6 mm from primary xylem. Radial distance between a & f is 2.5 mm. - 12: Ray originating 22 mm from primary xylem. Radial distance between a & e is I mm.

crease in ray width and the formation of new by little or no elongation would also result in a secondary rays. Subsequently, the proportion gradual but steady decrease in length of the fu­ of the wood occupied by rays remains relative­ siform initials. ly constant throughout secondary growth. Dur­ A rather unusual bark is present. On the sur­ ing this latter period, very few additional sec­ face there are deep fissures about I cm apart. ondary rays appear. A ray originating during The region between the fissures is up to 8 mm this period arises from a small group of fusi­ thick while in the bottom of a fissure the bark form initials (Fig. 12 a) that average 280 )J.ffi is only 2-3 mm thick. In the thick areas the long, as reflected in the length of the cells bark consists offunctional phloem, about 1 mm when they are first observed in the wood. Dur­ thick, a region of nonfunctional phloem 5-6 ing the formation of a very small amount of mm thick (Fig. 8), and an outer region of ho­ wood, a multiseriate ray is formed that closely mogeneous cortical parenchyma (Fig. 10). At resem bles other rays (Fig. 12 b-e). In trans­ the outer edge of the nonfunctional phloem verse sections, such a ray appears to arise very there are some small strands of fibres with suddenly. moderately thick walls (Fig. 10). These are Although the stem material used was dried, very likely primary phloem fibres. The outer­ sections of the most recently formed xylem most mm or so of cortical parenchyma consists and phloem revealed the general organisation of collapsing cells (Fig. 10). The remaining pa­ of the vascular cambium. The fusiform initials renchyma cells throughout the bark appear in­ are short, averaging 282 )J.ffi (range 160-387 tact. The nonfunctional phloem consists of }.tm), and are nonstoried or locally storied. axial parenchyma as fusiform cells and strands Some walls in the recently formed xylem and of two, rarely three cells, and ray parenchyma phloem cells appear as if they resulted from cells which have expanded tangentially (Figs. recent anticlinal divisions in the fusiform ini­ 8,9). tials. Most of these are relatively long oblique anticlinal divisions, while a few are radial anti­ Discussion clinal or relatively long lateral divisions. With The results reported here are very similar to little or no elongation of the daughter cells, a the general description of Bocconia given by nonstoried cambium would result with locally Metcalfe and Chalk (1950). A high level of spe­ storied areas formed by radial anticlinal divi­ cialisation for the secondary xylem is shown by sions. The oblique anticlinal divisions, followed the very short vessel members with alternate

Downloaded from Brill.com09/28/2021 08:13:39AM via free access IAWA Bulletin n.s., Vol. 4 (2-3),1983 137

pitting and simple, oblique to transverse per­ foration plates, short libriform fibres with very fiber length small pits, para tracheal axial parenchyma, and specialised, exclusively multiseria te, heterocel­ lular rays (Barghoorn, 1941). The wood in Bocconia is similar in a number of respects to some species that are considered to be secondarily woody from an herbaceous ancestry. The most notable features are the E 400 length-on-age curves for vessel members and :t vessel member length fibres, and the form and composition of the 300 rays. The fact that there is little or no develop­

mental change in diameter and frequency of 200 vessels, fibre wall thickness, and amount and arrangement of axial parenchyma may also in­ vessel diameter dicate an herbaceous ancestry for Bocconia. Al­ 100~ so, the rather unusual bark structure is different from that in truly woody dicotyledons. The length-on-age curve for vessel member px 0 5 10 15 20 25 30 length in Bocconia vulcanica resembles curves distance from primary xylem in mm for many herbaceous dicotyledons, species sec­ ondarily woody from an herbaceous ancestry, Fig. 13. Length-on-age curves for vessel mem­ stem succulents, and other specialised types ber and fibre length, and diameter-on-age curve (Carlquist, 1962, 1975). While there is a de­ for vessel diameter in Bocconia vulcanica Donn. crease in vessel member length in the metaxy­ Smith. lem, there is an even greater decrease during the production of the first secondary xylem, followed by a slight but steady decrease during most of cambial activity. The curve is very sim­ dicotyledons in families that are predominantly ilar to that for Lobelia gibberoa and Delissea herbaceous. These characteristics may remain undulata (Carlquist, 1975). permanently juvenile or they may ultimately The length-on-age curve for vessel member be lost during secondary growth. Although there length in Bocconia vulcanica can be explained is some decrease in vessel member length in the in the following way. Decrease in length of primary xylem of Bocconia vulcanica, this de­ metaxylem cells is due to transverse divisions, crease is extended into the secondary xylem. so characteristic of the procambium. Transverse Mabberley (1974, 1982), who rejects the divisions are extended into the vascular cam­ paedomorphosis concept, offers a different ex­ bium during early secondary growth, judging planation for the type of length-on-age curve from the decrease in length, 540 to 337 pm, of found in Bocconia, by referring to the paper of vessel members in the first formed secondary Philipson and Butterfield (1976). According to xylem. A slight but steady decrease in vessel Mabberley, the rate of increase in the circum­ member length throughout most of cambial ac­ ference of the cambium in plants with a narrow tivity is due to relatively long oblique anticlinal pith is much higher in young stems than in older divisions in the fusiform initials with little or ones. In plants in which the anticlinal divisions no elongation of the daughter cells. Extension in the fusiform initials are oblique, the fre­ of transverse divisions from the procambium quency of division is not high enough to accom­ into the vascular cambium and the occurrence modate the rapidly expanding cambium. As a of relatively long oblique anticlinal divisions in result, the daughter cells formed in oblique an­ fusiform initials with little or no elongation of ticlinal divisions elongate to become longer the daughter cells is common in herbaceous than the dividing initial. Thus, the average length dicotyledons (Cumbie, 1963, 1969a, 1969b). of cambial cells increases during secondary The length-on-age curve for vessel member growth, and the length-on-age curve for vessel length in many specialised dicotyledons with member length parallels the curve for fusiform limited cambial activity is a form of juvenilism initial length. On the other hand, in plants with or paedomorphosis (Carlquist, 1962, 1975, a wide pith the first oblique anticlinal divisions 1980). According to this concept, juvenile take place at a greater distance from the centre characteristics of the primary xylem may be of the pith so that there is a relatively lower in­ extended into the secondary xylem in herbace­ crease in circumference, per unit of radial ous dicotyledons and in many specialised woody growth. Accordingly, extensive elongation of

Downloaded from Brill.com09/28/2021 08:13:39AM via free access 138 IAWA Bulletin n.s., Vol. 4 (2-3),1983

cells following oblique anticlinal division is not fibre length outwards from the pith. In Robinia required, and there will be no increase in length pseudoacacia (Hejnowicz & Hejnowicz, 1959) of fusiform initials or vessel members during and Balanites aegyptiaca (Parameswaran & Con­ secondary growth. rad, 1982), fibre length increases for several Mabberley's interpretation does not explain years. In Pterocarpus angolensis fibre length in­ the unusuallength-on-age curve for many plants, creases significantly during the formation of however. Baas found a 'juvenilistic' length-on­ the first two growth rings and then remains age curve in stems of Leptospermum crassipes constant (Chalk et aI., 1955). However, fibre (1977) and Vaccinium lucidum (1979), both length does not increase outwards from the pith species with a very narrow pith. He also observed in Aeschynomene elaphroxylon and Nesogor­ that in the woody family Araliaceae, some spe­ donia papavifera (Chalk et aI., 1955). It is signi­ cies with a very wide pith show a greater in­ ficant, however, that the fibres in Aeschyno­ crease in vessel member length than species mene elaphroxylon are unusual and undergo with a narrow pith (Baas, 1976). This is just little or no elongation during differentiation the opposite of what one would expect based (Beijer, 1927). on Mabberley's interpretation. Hibiscus lasio­ Some groups of dicotyledons that are basical­ carpus is a perennial herbaceous species which ly herbaceous with evolution toward woodiness forms stems with a narrow pith at the base, and are characterised by conspicuously heterocellu­ considerable secondary xylem. Yet the fusiform lar, multiseriate rays. The rays, which extend initials in this species decrease in length during from the interfascicular regions, are usually tall, early secondary growth and then remain rela­ and often show little or no breakup during sec­ tively constant (Cumbie, 1963). In Hibiscus, as ondary growth. They consist of erect and well as in other species with a 'juveniJistic' square cells with few or no procumbent cells. length-on-age curve, the frequency of anticlinal Erect sheath cells almost always occur along division in the cambium is just adequate to the sides of the rays, and are very similar to the keep pace with the increase in circumference of wood fibres. As a result, the limits of the rays the cambium throughout secondary growth. are often difficult to determine, particularly in Carlquist (1962, 1975) illustrated length-on­ transverse sections. The ray cells are usually age curves only for vessel members in explain­ thin-walled and angular in form in tangential ing paedomorphosis, although he assumed that sections. Species with this ray organisation are fibre length parallels vessel member length. The common in some Compo sitae (Cariquist, 1966, length-on-age curve for fibres in Bocconia vul­ 1974), Lobelioideae-Campanulaceae (Carlquist, canica does parallel that for vessel members. 1969a), Goodeniaceae (Carl quist, 1969b), the The variation that occurs is the result of differ­ genus Echium in the Boraginaceae (Cariquist, ences in intrusive growth of fibres during sec­ 1970), and in the Brassicaceae (Cariquist, 1971). ondary growth. Similar length-on-age curves for Barghoorn (1941) recognised that this type of fibres occur in Rubia fruticosa (Koek-Noorman, ray organisation often characterises highly spe­ 1976) and in Vaccinium lucidum (Baas, 1979). calised forms of dicotyledons with limited cam­ Horak (1981) found that vessel member length bial activity. As an example of such ray struc­ and fibre length is constant across xylem incre­ ture, he illustrates Bocconia frutescens. ments in Stegnosperma, which shows anomalous Ray organisation in Bocconia vulcanica re­ secondary growth by means of successive cam­ sembles in many respects that commonly found bia. I have also found that fibre length parallels in the woody species of predominantly herba­ vessel member length in several annual and per­ ceous groups. The rays are exclusively multi­ ennial herbs which accumulate considerable seriate, tall, and heterocellular with a predomi­ secondary xylem. nance of erect and square cells. The cells have In contrast to many plants with limited cam­ thin secondary walls and are angular in outline. bial activity, length-on-age curves for vessel Erect sheath cells occur along the sides of the members and fibres are different in truly woody rays throughout the secondary xylem. dicotyledons. In species with a nonstoried At first glance, the magnitude of the develop­ cambium there is an increase in vessel member men tal changes in the rays in Bocconia vulcanica and fibre length outwards from the pith due to seems very extensive. New secondary rays are the increase in average length of the fusiform formed, the very tall rays are broken into smal­ initials as the cambium increases in circumfer­ ler ones, and the rays increase in width through­ ence. In plants with a storied cambium, there is out secondary growth. However, the changes no increase in vessel member length outwards are very regular in occurrence and do not seem from the pith because the length of the fusi­ as complex as those that occur in truly woody form initials remains constant. In the few spe­ dicotyledons. Almost all secondary rays are cies studied, however, most show an increase in formed during the production of the first few

Downloaded from Brill.com09/28/2021 08:13:39AM via free access IAWA Bulletin n.s., Vol. 4 (2-3),1983 139

mm of secondary xylem, and in a very regular decrease in length of vessel members, the ab­ pattern. They originate midway between the sence of any increase in fibre length, the con­ primary rays. Subsequently, very few new rays stant vessel diameter and frequency, fibre wall are formed even though considerable secondary thickness, and amount and arrangement of axial xylem is produced. Judging from the length of parenchyma, the presence of specialised multi­ the cells when they first appear in the secondary seriate, heterocellular rays, and the absence of xylem, fusiform initials are converted into ray secondary phloem fibres and a periderm are all cell initials by simple subdivision without de­ features that may be interpreted as indicating crease in length or loss of any of the products. Bocconia may be derived from an herbaceous Subdivision of the cells is rather slow as reflect­ ancestry. Certainly these features are character­ ed in the tall sheath cells. This process is simi­ istic of many plants with reduced size and me­ lar to that in herbaceous dicotyledons (Cumbie, chanical strength. 1963, 1969a, 1969b). In truly woody dicotyle­ The idea that Bocconia may have evolved dons, rays are always changing by a number of from herbaceous ancestors also seems plausible complex phenomena: initiation of new rays, in­ in view of what can be inferred about the evolu­ crease or decrease in size, fusion, splitting, and tionary position of the genus (Curtis Clark, per­ disappearance (Philipson et aI., 1971). sonal communication). Bocconia has a micro­ Some additional features of the wood in petalous, wind-pollinated flower, and a single­ Bocconia vulcanica undergo little or no change seeded fleshy two-valved drupe, both of which during secondary growth. Fibre wall thickness give it an unusual position in a family of large and the amount and arrangement of axial pa­ entomophilous flowers and dry capsules. The renchyma is the same throughout the wood. only other clearly woody member isDendrome­ Although there is a slight increase in vessel dia­ con, which does not appear to be closely related meter and a decrease in vessel frequency, these to Bocconia, so it seems likely that Bocconia changes occur during the formation of the first has been derived from herbaceous ancestors. secondary xylem and remain constant through­ out the remaining secondary xylem. These fea­ Acknowledgements tures, established in the innermost secondary I should like to thank W. E. Harmon for pro­ xylem and extending throughout secondary viding the stem material used in this study. This growth, may also be regarded as a form of ju­ research was supported by a grant from the Re­ venilism. search Council of the Graduate School, Univer­ There is little information available on bark sity of Missouri-Columbia. structure in plants that may be secondarily woody. In his studies of herbaceous families References and genera that are woody on islands, Carlquist Baas, P. 1976. Some functional and adaptive (1974) found that many have few or no extra­ aspects of vessel member morphology. In: xylary fibres. The only extraxylary fibres in Wood structure in biological and technolo­ Bocconia vulcanica are the small strands of pri­ gical research (eds. P. Baas, A.J. Bolton & mary phloem fibres. D. M. Catling): 157-181. Leiden Bot. Ser. Considering the diameter of the stem, the 3. Leiden Univ. Press, The Hague. thickness of the bark, and the fact that the bark - 1977. The peculiar wood structure of Lep­ is deeply fissured, it is surprising that a peri­ tospermum crassipes Lehm. (Myrtaceae). derm is not present in Bocconia vulcanica. The IAWA Bull. 1977/2: 25-30. outer bark consists of cortical and phloem pa­ - 1979. The peculiar wood structure of Vac­ renchyma only. There is little tangential expan­ cinium lucidum (BI.) Miq. (Ericaceae). IAWA sion of the bark, by lateral enlargement of ray Bull. 1979/1: 11-16. parenchyma cells. In some Labiatae there is a Barghoorn, E.S. 1940. The ontogenetic devel­ special method of cork formation by the gra­ opment and phylogenetic specialization of dual suberisation of parenchyma cells in the rays in the xylem of dicotyledons. I. The bark without development of a phellogen primitive ray structure. Amer. J. Bot. 27: (Solereder, 1908). I was not able to determine 918-928. if there are any cells with suberised walls in the - 1941. The ontogenetic development and bark of Bocconia vulcanica. Nevertheless, the phylogenetic specialization of rays in the absence of extraxylary fibres and a true peri­ xylem of dicotyledons. II. Modifications of derm may be additional indicators of an herba­ the multiseriate and uniseriate rays. Amer. ceous ancestry for Bocconia. J. Bot. 28: 273-282. Thus, it seems likely that Bocconia evolved Beijer, J.J. 1927. Die Vermehrung derradialen from herbaceous ancestors based on the organi­ Reihen im Cambium. Rec. Trav. Bot. Neerl. sation of the secondary xylem and bark. The 24: 631-786.

Downloaded from Brill.com09/28/2021 08:13:39AM via free access 140 IAWA Bulletin n.s., Vol. 4 (2-3), 1983

Carlquist, S. 1962. A theory of paedomorphosis Horak, K. E. 1981 . Anomalous secondary thick­ in dicotyledonous woods. Phytomorphol­ ening in Stegnosperma (Phytolaccaceae). ogy 12: 30-45. Bull. Torrey Bot. Club 108: 189-197. -1966. Wood anatomy of Compositae: a IAWA, International Association of Wood Anat­ summary, with comments on factors con­ omists. Committee on Nomenclature. 1964. trolling wood evolution. Aliso 6: 25-44. Multilingual glossary of terms used in wood - 1969a. Wood anatomy of Lobelioideae anatomy. Konkordia, Winterthur. (Campanulaceae). Biotropica I: 47-72. Koek-Noorman, J. 1976. Juvenile characters in - 1969b. Wood anatomy of Goodeniaceae the wood of certain Rubiaceae with special and the problem of insular woodiness. Ann. reference to Rubia fruticosa Ait. IAWA Bull. Missouri Bot. Gard. 56 : 358-390. 1976/3: 38- 42. - 1970. Wood anatomy of Echium (Boragina­ Lawrence, G.H.M. 1951 . Taxonomy of vascular ceae). Aliso 7: 183- 199. plants. MacMillan, New York. - 1971. Wood anatomy of Macaronesian and Mabberley, D. J. 1974. Pachycauly, vessel-ele­ other species of Brassicaceae. Aliso 7: 365- ments, islands and the evolu tion of arbores­ 384. cence in 'herbaceous' families. New Phytol. -1974. Island biology. Columbia Univ. Press, 73 : 977-984. New York, London. - 1982. On Dr. Carlquist's defence of paedo­ - 1975. Ecological strategies of xylem evolu­ morphosis. New Phytol. 90: 751 - 755. tion. Univ. Calif. Press, Berkeley. Metcalfe, C. R. & L. Chalk. 1950. Anatomy of - 1980. Further concepts in ecological wood the dicotyledons. Clarendon Press, Oxford. anatomy, with comments on recent work Parameswaran, N. & H.Conrad. 1982. Wood and in wood anatomy and evolution. Aliso 9: bark anatomy of Balanites aegyptiaca in re­ 499-553. lation to ecology and taxonomy. lAW A Bull. Chalk, L. , E. B. Marstrand & J. P. de C. Walsh. n.s. 3: 75-88. 1955. Fibre length in storeyed hardwoods. Philipson, W. R. & B.G. Butterfield. 1967. A Acta Bot. Neerl. 4 : 339-347. theory on the causes of size variation in Cheadle, V. I. & K. Esau. 1964. Secondary wood elements. Phytomorphology 17 : 155 - phloem of Liriodendron tulipifera. Univ. 159. Calif. Publ. Bot. 36: 143-252. -, J. M. Ward & B.G. Butterfield. 1971. The Cronquist, A. 1968. The evolution and classifica­ vascular cambium: Its ontogeny and devel­ tion of flowering plants. Houghton Mifflin, opment. Chapman & Hall, London. Boston, Mass. Sass, J.E. 1958. Botanical microtechnique. 3rd Cumbie, B.G. 1963. The vascular cambium and Ed. Iowa State Univ. Press, Ames, Iowa. xylem development in Hibiscus lasiocarpus. Solereder, H. 1908. Systematic anatomy of the Amer. J. Bot. 50: 944-951. dicotyledons. Clarendon Press, Oxford. - 1969a . Developmental changes in the vas­ Standley, P.C. & J.A. Steyermark. 1946. Flora cular cambium of Polyganum lapathifolium. of Guatemala. Chicago Nat. Hist. Museum ; Amer. J. Bot. 56 : 139-146. Fieldiana: Bot. 24 (Pt IV): 347- 354. - 1969b. Developmental changes in the xylem Stebbins, G.L. 1974. Flowering Plants. Evolu­ and vascular cambium of Apocynum sibiri­ tion above the species level. Harvard Univ. cum. Bull. Torrey Bot. Club 96: 629-640. Press, Cam bridge, Mass. Halle, F., R. A. A. Oldeman & P. B. Tomlinson. Takhtajan, A.L. 1980. Outline of the classifica­ 1978. Tropical trees and forests. An archi­ tion of flowering plants (Magnoliophyta). tectural analysis. Springer-Verlag, Berlin. Bot. Rev. 46: 225-359. Hejnowicz, A. & Z. Hejnowicz. 1959. Variations of length of vessel members and fibres in the trunk of Robinia pseudoacacia. Acta Soc. Bot. Polon. 28: 453-460.

Downloaded from Brill.com09/28/2021 08:13:39AM via free access