BLUMEA 22 (1975) 311—407
Vegetative anatomy and the affinities of Aquifoliaceae,
Sphenostemon, Phelline, and Oncotheca
P. Baas
Rijksherbarium, Leiden, Netherlands
Contents
Summary 312
Introduction
Aims of this study 3 12
Historical 31 2
Synopsis of morphological characters 3*5
> Techniques and materials 3 I(
Descriptive part
Leaf J anatomy 3 7
2 Nodal anatomy 3 5
2 Twig anatomy 3 7
2 Bark anatomy 3 9
Wood 1 anatomy 33
Discussion of some anatomical characters
Indumentum 334
The unspecialized cells of the leaf epidermis 334 The stomatal complex 335
Cork warts 335
Leaf hypodermis 335 Crystals, sclereids, and marginalsclerenchyma 33d Nodal and vascularization anatomy petiole 337
Twig and bark anatomy 33"
Wood 338 anatomy
Affinities and taxonomic rank
Introduction 339
Sphenostemon 339
Phelline 34 2
Oncotheca 355
Nemopanthus 355 and the Celastrales. The position of Icacinaceae, Aquifoliaceae, Phellinaceae, Sphenostemonaceaein 356
General considerations and conclusions 35b
The leaf anatomical in Ilex Special part: range
Introduction 357
Loesener's for Ilex 357 system Specific descriptions and taxonomic notes 359 anatomical Lists ofDex species with certain leaf characters 387
Comparison of leaf anatomical data with Loesener's system 390 Subgenus Rybonia 390 Subgenus Byronia 390 Subgenus Euilex 391 Subgenus Prinus 399
Taxonomic conclusions 399
Infiraspecific variabilityin 15 species of Ilex 400 Latitudinal and altitudinal trends; ecological and functional anatomy 402 Acknowledgements 404
References BLUMEA — VOL. No. 312 22, 3, 1975
Summary
The ofleaf, node, twig, and bark of Phelline and Sphenostemonis anatomy Ilex, Nemopanthus, Oncotheca, , with wood of the latter Several characters recorded the described, together the anatomy 4 genera. are for
first time.
value of anatomical characters the wider affinities ofthe involved The systematic some for judging genera is discussed. Considering these characters together with macromorphological and palynological data, it is
concluded that Phelline and Sphenostemon each merit family status, and have affinities with Aquifoliaceae and
the Icacinaceae of the Celastrales. Oncotheca shows no affinities with families of Celastrales, and may possibly closest relatives but this has been studied detail. have its in Theales, problem not in great
The leaf of Ilex and the leaf anatomical of Ilex is discussed anatomy 95 species is described, great diversity reference classification. with to Loesener’s infrageneric In general,leaf anatomy cannot be used to support
his elaborate subdivision of the genus. Infraspecificleaf anatomical variability is reported and discussed for and functional 15 species of Ilex. Attention is paid to some geographical, ecological, aspects of the leaf
anatomical of variation in Ilex. This has for leaf range range a general significance systematic anatomy, it because shows the restricted value ofseveral characters for the discussion ofaffinities above the genus level.
Introduction
Aims of this study
The affinities of Sphenostemon from New Guinea, New Caledonia, and Queensland, and
of Oncotheca from New Caledonia have been discussed and disputed over the last decades
The inclusion of Phelline from Caledonia the in many publications. New in Aquifoliaceae
has also been challenged by several authors. The three genera have in common that they have been mentionedrepeatedly in connection with the Aquifoliaceae. For the development
of views the further collecting of more anatomical data on the two undisputed mem- bers of the Aquifoliaceae: Ilex and Nemopanthus, therefore seemed essential. With this aim
anatomical Ilex in mind a detailed and elaborate survey of the wood range in has been
in published (Baas, 1973); the present paper the leaf anatomy of Ilex will be treated in a similar the leaf and wood of way. Besides, anatomy Nemopanthus, Oncotheca, Phelline, and
Sphenostemon will be comprehensively described, together with observations on nodal,
bark It is considered data sounder twig, and anatomy. that these should provide a basis for anatomical comparison than the rather haphazard choice of miscellaneous species
which the contributed to such diverse conclusions several authors. in past by Comparisons
with other putative relatives will be based mainly on data from the literature, and will the with consequently suffer more from incompleteness than comparisons Aquifoliaceae
sensu stricto.
anatomical ofIlex should also be considered aim in itself. The leaf survey as an Although
it will be attempted to find correlations with the infrageneric classification proposed by
this aimed the Loesener (1901, 1908, and 1942), survey is mainly at determining leaf
anatomical range in a large and widely distributed genus, and at indicating possibly useful
characters for taxonomic which needed for this a future revision is badly genus.
Historical
views affinities As a starting point for a historical survey of on the ofIlex, Nemopanthus, and Loesener's and of the be Phelline, treatments (1901, 1908 1942) Aquifoliaceae can conveniently used; older literature is comprehensively listed by him. and differences The close affinities of Ilex Nemopanthus are beyond doubt; the only in P. Baas : Anatomy of Aquifoliaceae etc. 313
reduced and the free of external morphology are in the calyx in petals Nemopanthus as
contrasted by the well-developed calyx and the at their base almost always more or less fused, imbricate petals of Ilex. treated Phelline characterized valvate Loesener in a separate tribe, by apiculate petals.
Recently some doubts have been expressed about the inclusion ofPhelline in Aquifoliaceae. stated be On the grounds of pollen morphology Erdtman (1952) that arguments might advanced in favour of the assumption that Phelline has not yet been given its correct place referred in the system. Airy Shaw (1966, in Willis) doubtfully Phelline to Araliaceae (~ Aquifoliaceae) without giving further Takhtajan (1966) erected the family ’, arguments.
Phellineaceae 'related to Aquifoliaceae, from which it differs, however, by having valvate
petals, by the character of the inflorescence, the hemitropous or weakly campylotropous the wood the different and the leaf ovules, anatomy, quite sporoderm morphology, from Shaw venation' (quoted Takhtajan, 1969). His suggestion was followed by Airy in of corrected the latter the latest edition Willis' Dictionary (1973); the family name was by
into Phellinaceae.
When Baillon first he indicated links (1891) described the genus Oncotheca, with
Phelline and Sphenostemon from the 'Ilicinees' and at the same time he suggested a relation-
he the in in ship withEbenaceae. Later (1892) formally included genus Ilicaceae, his concept also including Cyrilla and Cliftonia, which are now in Cyrillaceae. Kronfeld (1892) in-
of but cluded Oncotheca in his treatment the Aquifoliaceae, Loesener (1897, 1901, and 1942)
in this and it Ebenaceae and never accepted the genus family assigned to (1897, 1942) also affinities with suggested Sapotaceae (1901). Guillaumin (1938) most strongly advocated
the Ebenaceous nature of Oncotheca and formally erected the tribe Oncotheceae within
Ebenaceae on the basis of both macromorphological and anatomical characters. More recently Airy Shaw (1965) erected the family Oncothecaceae, following unpublished Kobuski. An and suggestions by Ebenaceous affinity was suggested to be erroneous, with of affinities Theaceae were considered to be rather close. The pollen Oncotheca has not yet been properly described; Erdtman (1952) could only dispose of an inadequate slide ilicoid and be of of mixed smooth grains. Both types would, in any case, not suggestive Ebenaceous affinities.
The exceedingly chaotic taxonomic history of the genus Sphenostemon (including and Bernardi Idenburgia Nouhuysia ) has been very comprehensively reviewed by (1964).
of views With reference to his paper a survey the held before 1964 can be very brief. The reduction of Idenburgia Gibbs and Nouhuysia Lauterb. to Sphenostemon Baillon (Van
affinities with Steenis, 1952, 1955) implies presumed Aquifoliaceae (Baillon, 1875; Van
Steenis, 1955), with Trimeniaceae (Gibbs, 1917; Gilg & Schlechter, 1923), and with
Guttiferae (Lauterbach, 1912; Van Steenis, 1952). Older views on the affinities of Spheno-
the stemon sensu stricto had also differed: Loesener (1901, 1942) expelled genus from
and Theaceous Ochnaceous Aquifoliaceae suggested a or affinity, leaving possibilities open for a Anatomical and contributions have far added separate family. palynological so only to the confusion: Money c.s. (1950), Bailey & Swamy (1953), and Bailey (1956) excluded
Sphenostemon (Idenburgia and Nouhuysia) from Monimiaceae, Trimeniaceae, and Guttiferae,
reluctant and Bailey (1956) was also to accept an Aquifoliaceous affinity. Previously had Money c.s. (1950) excluded Idenburgia from Monimiaceae on morphological grounds.
Metcalfe's anatomical contribution(1956) to the problem also led him to conclude that it is that of is related evident none the genera ( Sphenostemon, Idenburgia, Nouhuysia) closely to either the Monimiaceae, Aquifoliaceae, or Guttiferae. The occurrence of styloid crystals
mention in prompted him to a resemblance this respect with Escalloniaceae, without, however, implying and Escalloniaceous affinity for Sphenostemon s.l. Ingle & Dadswell 314 BLUMEA VOL. 22, No. 3, 1975
(1961) in comparing the wood of Nouhuysia (New Guineanand Australian Sphetiostemon)
with that of Ilex, Platea (Icacinaceae), and Polyosma (Escalloniaceae) concluded that of these
Ilex bears resemblance to that at genera least Nouhuysia, and Icacinaceae have least as big a
claim for affinity as the Escalloniaceae. Erdtman (1952, 1954) drewattention to the 'strange
± unique' pollen grains of Sphenostemon sensu lato, stressing their lack of resemblance with
pollen of Aquifoliaceae. Lobreau (1969) in studying the pollen of Sphenostemon, found it
similar to that of Dichapetalaceae rather than to other putative relatives. Hutchinson
(1959, 1964, 1969, and 1974) meanwhile maintained Sphenostemon in Trimeniaceae, and
Airy Shaw, when formally publishing a family diagnosis for Sphenostemonaceae in 1972
also advocated Trimeniaceous the and a affinity, although anatomy pollen morphology in 'forbid' its inclusion that family. Airy Shaw moreover repeated objections against
inclusion in Aquifoliaceae, Guttiferae, Monimiaceae, and Escalloniaceae. Bernardi (1964) the observations scattered the literaturewith extending numerous in an original observa- concluded tion on the unitegmic nature of the ovules, from many sources ofevidence that should after all be included Sphenostemon in Aquifoliaceae as suggested by Baillon who first
described the genus.
The extensive amount of publications, listed above, seems to make further attempts to
find taxonomic of the three a proper position enigmatic genera Oncotheca, Phelline, and rather useless without Sphenostemon or prospect. However, most previous anatomical
publications have in common that they do not take the entire anatomical range of Aqui- and that the with this and families foliaceae sensu stricto into account comparisons other of the are too superficial. By adding to our knowledge anatomical range in Aquifoliaceae
sensu stricto and by considering the anatomical and macro-morphological characters in
it is that this settle the other putative relatives, hoped study can at least finally argument with affinities between and respect to possible Aquifoliaceae these genera.
considerable amount of data has accumulated on the A descriptive vegetative anatomy of Older the 5 genera studied. literature has been comprehensively summarized by
Solereder (1899, 1908), and Metcalfe & Chalk (1950), though special mention should be
made of Thévenard's (1906) anatomical study of Sphenostemon, Phelline, Ilex, and Nemo-
his leave features panthus. However, accurate descriptions appear to out many now con-
sidered as important diagnostic or taxonomic characters. Anatomical data on Ilex are furthermore provided by Senglet (1928), Zahur (1955), Ashworth (1963), Copeland
(1963), Gasanov (1969), Fretz & Dunham (1972), Korn & Frederick (1973), and by Baas
(1973), who also listed the more recent wood anatomical literature.
The of has received in Thevenard anatomy Sphenostemon attention papers by (1906), Bailey & Swamy (1953), Metcalfe (1956), Jutte (1958), and Ingle & Dadswell (1961).
Pennington (1953) also gave an account of the wood of Sphenostemon, but a comparison with the material vouched I could study leaves no doubt that the wood specimen she
studied does not belong to Sphenostemon.
of Phelline is included in Thevenard's survey Aquifoliaceae anatomy, and Takhtajan included data (1966) anatomical when he raised the genus to family rank. He recorded
anomocytic stomata, narrow vessels with many-barred scalariform perforation plates,
scalariform to alternate vessel wall pitting, scanty paratracheal parenchyma, mixed homo- with geneous rays with short tails, and fibres distinctly bordered pits. Pennington's of the wood of Phelline mislabelled misnamed material. (1953) account again suggests or Oncotheca has received little anatomical with and comparatively attention, scanty very superficial observations by Guillaumin (1938) and unpublished studies by Kobuski (see
Airy Shaw, 1965). Recently Miss C. S. Bissett (Chapel Hill, North Carolina, U.S.A.) anatomical her independently embarked on an study of Oncotheca, and conclusions P. Baas: Anatomy of Aquifoliaceae etc. 315
emphasizing floral morphological features will be the subject of a separate publication. less simultaneous with this Madame Danielle Lobreau-Callen has More or study
of has carried out a pollen morphological survey all genera involved. She shared her
that I could include from results prior to publication, so arguments pollen morphology in
discussion of the affinities and status of the The results my enigmatic genera. were pub- lished in a thesis when these investigations had been completed (Lobreau-Callen, 1975).
Her conclusions on Phelline and Sphenostemon are in general agreement with those presented here. Oncotheca is said to have affinities with Salvadoraceae because of more or less identical pollen.
SYNOPSIS OF MORPHOLOGICAL CHARACTERS
The descriptions below are mainly taken from the literature (Airy Shaw in Willis,
1973; Loesener, 1942). For Sphenostemon, Phelline, and Oncotheca most characters have been checked in the material present in the Rijksherbarium.
ILEX L.
Flowers unisexual, in axillary cymes or racemes, or fasciculate, rarely solitary, 4(5—9)- often with merous, isomerous, rarely heteromerous, pleiomerous ovary. Sepals connate at base, lobes imbricate. Petals mostly connate at the base, rarely free, lobes rounded, imbricate. Stamens alternipetalous, adnate to the base of the corolla, rarely free; filaments thin, anthers basidorsifixed, introrse, connective not prolonged. Pollen tricolporate,
as ofthe tectate- ectoapertures long, endoapcrtures developed constrictions ectoapertures, discoid clavate. Ovary superior, with sessile capitate or stigma, (2—)4—g(—22)-locular, with locule. Funicle 1 (rarely 2) axile, pendulous, anatropous, unitegmic ovule(s) per site Fruit frequently thickened ('obturator' or of repressed second ovule, cf. Herr, 1959). a
2—22-pyrened drupe. Seeds with copious endosperm.
Trees or shrubs, rarely climbers, evergreen or deciduous, with alternate, rarely opposite or whorled, stipulate leaves.
About 400 species; cosmopolitan.
NEMOPANTHUS Raf.
reduced with small Petals free. other characters Calyx to a narrow seam 0—5 tips. In as
Ilex, subgenus Prinus.
Two species in N. America.
ONCOTHECA Baill.
Flowers bisexual, in ± narrow thyrses with angular rhachis, 5-merous, isomerous,
Sepals connate, lobes orbiculate, strongly imbricate, persistent. Petals connate, lobes
Stamens rounded, imbricate. alternipetalous, inserted on the corolla; filaments thin, very anthers with thick short, basifixed, extrorse, connective, produced into a subulate, sharply indexed, appendage. Pollen tricolporate, ectoapertures long, endoapertures subcircular, with tectate-fmely perforate. Ovary superior, with 5 free recurved styles, 5-locular, 1
(rarely 2) pendulous, anatropous, unitegmic ovule(s) per locule. Funicle without obturator. Seed Fruit a 5-locular, ± woody drupe. unknown.
Evergreen shrubs with alternate, exstipulate leaves.
One species in New Caledonia. 13LUMEA VOL. 316 22, No. 3. 1975
PHELLINE Labill.
Flowers unisexual, in spicate or paniculate inflorescences, 4—6-merous, with a 2—5- small. merous ovary. Sepals connate at base, Petals free, valvate, with a small inflexed
apiculus. Stamens alternipetalous, free; filaments thin, anthers basidorsifixed, introrse, Pollen devel- connective not prolonged. tricolporate, ectoapertures long, endoapertures the tectate-echinate oped as constrictions of ectoapertures, or tectate-smooth. Ovary
superior, with sessile lobed stigma, and I axile, pendulous, hemitropous or slightly Funicle campylotropous, unitegmic ovule per locule. without obturator. Fruit a drupe
with 2—J stones. Seeds with copious endosperm. Evergreen shrubs with alternate, exstipulate leaves.
About 10 species in New Caledonia.
SPHENOSTEMON Baill.
Flowers unisexual in subterminal heteromerous. or bisexual, racemes, Sepals 4, free, caducous. decussate, early Petals 4, free, decussate, or rudimentary, or absent, ± fleshy,
caducous. with short obsolete early Stamens 4—12, i-seriate, free; thick, very or filaments,
thecae laterally on thick connective, or ventrally on thick petaloid stamen. Pollen tri- or
tetraporate, rarely inaperturate, tectate-reticulate or finely verrucate. Ovary superior, with sessile 2-locular with capitate stigma, I (rarely 2*) pendulous, anatropous, unitegmic Seeds ovule per locule. Funicle with obturator. Fruit a drupe with I or 2 bony pyrenes.
with copious endosperm, ruminate or not. shrubs with alternate Evergreen trees or (sometimes subopposite or pseudoverticillate) exstipulate leaves.
Seven species, in New Caledonia (series Sphenostemon), and New Guinea and Queens- land (series Apetala).
TECHNIQUES AND MATERIALS
Microtechnical procedures have been largely the same as reported elsewhere (Baas, leaf additional observations made cuticular 1970, 1972b, 1973). For anatomy were using
macerations (cf. Van Staveren & Baas, 1973), because these provide characters which are
not or hardly visible in sections. Stomata were measured in euparal mounts of free hand
paradermal sections; average values are based on 10 measurements. Almost all material studied is from the Rijksherbarium, but special mention should be made ofleaf material of some South American species from the Utrecht Herbarium (U), and some material of Oncotheca, Sphenostemon, and Phelline especially collected for this
study by Dr. H. S. McKee in New Caledonia. Part of the wood samples of Sphenostemon,
Nemopanthus, and Phelline were furthermore received from Madison (MADw & SJRw), and Sydney (SFCw) (abbreviations according to Stern, 1967). of leaf taken Although virtually all the studied material of Ilex was from flowering
misidentification that this specimens, the possibility of cannot be excluded, so in respect
this bears the limitations the wood anatomical This is study same as survey (Baas, 1973). due of of mainly to the lack an up to date world revision the genus.
*) Smith recorded the of 2 ovules locule in S. (1957) sporadic occurrence per lobosporus. P. Baas: Anatomy of Aquifoliaceac etc. 317
DESCRIPTIVE PART
Leafanatomy
ILEX
(see Plate I—V; fig. i—8, 17, 18) of In surface view: Glabrous or with indumentum varying density. Hairs
usually short, thick-walled, and unicellular, often confined to petiole and adaxial side of also abaxial side of and abaxial midrib, sometimes present on midrib, rarely present over the whole lamina. hair small number surface of Other types confined to a of species and slender 2-cellular hairs with cuticle similar hairs with thick including a warty (fig. 5); a basal and smooth cuticle hairs with number ofthin and sept a (fig. 4); long a varying septs, short hairs with broad base thick-walled a very subdivided by numerous closely spaced
Plate Hairs with recurved and with or septs (fig. 8; I, 6). straight or tips fairly narrow
broad bases. Cuticle of adaxial and/or abaxial surface smooth, coarsely to finely granular
and/or faintly to strongly striated. Epidermal cells of abaxial and/or adaxial surface poly-
gonal with straight to undulated anticlinal walls; undulations often confined to outer
adaxial cells in with anticlinal division walls. part; epidermal some species thin, secondary, cuticular Anticlinal flanges and cell walls pitted or not. Cells overlying midrib flattened to
elongated abaxially, isodiametric (not modified) or flattened to elongated adaxially.
and Stomata confined abaxial Primary, secondary, tertiary veins prominent or not. to with of epidermis, randomly orientated, mainly cyclocytic, sometimes 2 rings subsidiary cells but of (bicyclic), anomocytic, anisocytic, and complex cyclocytic in a fair number with combination of of these different species (Plate II; fig. 17); many species a 2 or more
stomatal types or with intermediate stomatal types. Peristomal rim usually present; outer
stomatal inner stomatal ledges well developed; ledges absent or very inconspicuous; polar
T-pieces to absent, guard cell on 18 long, 16 well-developed pairs average —42 /mi —34 fim wide. Giant stomata usually present, but varying in frequency. Cork warts frequent and of circular outline and of regular, or angular (Plate IV, 3 & 4), or infrequent irregular outline (probably traumatic) to absent.
In transverse thick. Adaxial section: Lamina dorsiventral, 90—700 fim
cuticle 1 thick, abaxial cuticle I—- thick. Unspecialized epidermal cells erect —28 fim 16fim
to flattened; adaxial cells much bigger to slightly smaller than abaxial cells. Adaxial
epidermis, and rarely also abaxial epidermis, containing a varying number of bulging,
usually tall, mucilage cells, often with periclinal division walls, sometimes resulting in a
complete 2—3-layered multiple epidermis (Plate IV, 2, 5, 6; fig. 10—15). Unspecialized abaxial epidermal cells with hgnified pitted thickenings to the innerpericlinal walls and
the of the anticlinal walls Adaxial inner part in some species (Plate III, 1—3; IV, 1). of cell of hypodermis I layer present or absent, composed unlignified parenchymatous to
collenchymatous cells or of lignified cells; rarely of more than one cell layer, and than
mostly composed ofbulging mucilagenous cells. Mesophyll composed of I—3(—4) layers of short tall cells and loose which be to palisade to compact spongy tissue may partly
(abaxially) or wholly lignified in a number of species. Midrib with a grooved, flattened, or raised raised adaxial surface and prominently abaxial surface, supplied with a simple to vascular sheathed of complex system partly or wholly by varying amounts sclerenchyma
fibres. Ground tissue of midrib collenchymatous to parenchymatous. Veins mostly
embedded, with a lignified or unlignified parenchymatous, rarely sclerenchymatous, bundle bundle sheath amounts of & sheath; enclosing various sclerenchyma (Plate V, 5 6) ; No. 318 BLUMEA VOL. 22, 3, 1975
Camera — lucida ofhair — Ilex Ilex micrococci Fig. 1 —9. drawings types. 1. pernyi (Boom 27299). —2.
3. Ilex serrata(Boom 6165). — 4. Ilex brasiliensis. — 5. Ilex amelanchier. — 6. Ilex cissoidea. — 7. Ilex zygophylla
= — (l lignified hypodermis). — 8. Ilex hypoglauca (Hallier B 2875). — 9. Phelline lucida (Schlechler 15340).
One construct the series: —2 —6, and —8. and may following morphological 1 —5—4, 1—3 1 —7 Fig. 4a
5a x 110; other figs. x 470. P. Baas: Anatomy of Aquifoliaceae etc. 319
veins occasionally vertically transcurrent through collenchymatous to parenchymatous or
sclerenchymatous bundle sheath extensions. Leaf margin with hardly modified ± compact
with solid strand fibres with chlorenchyma or a of sclerenchyma (Plate V, 4). Petiole a ±
similar vascular as midrib or almost pattern more complex (Fig. 18). Crystals exclusively
represented by small to large druses throughout lamina, rarely crystals minute and
solitary (prismatic to needle-shaped) in adaxial epidermis, or crystals solitary, prismatic
and accompanying major vascular bundles; very rarely absent. Idioblastic sclereids °f various in of shapes, but mostly irregularly fusiform, the mesophyll a restricted number of of species (Plate V, I—3); rarely regular elongated shape and confined to ground tissue of midriband petiole.
Note: The leafanatomy of the individualspecies and its implications for systematics
in Ilex will be treated in a separate section (see p. 359).
NEMOPANTHUS
(see Plate VI, 3)
In surface view: Glabrous. Cuticle of abaxial epidermis with conspicuous, adaxial rather widely spaced ridges (Plate VI, 3), cuticle faintly to conspicuously striated.
Epidermal cells of adaxial surface polygonal with curved walls; walls of abaxial surface
curved to undulated. Cuticular flanges weakly developed, not pitted. Cells overlying
also modified midrib elongate; cells over primary, secondary, and also many tertiary veins
(usually elongate) in both adaxial and abaxial epidermis. Stomata confined to abaxial
epidermis, randomly orientated, irregularly cyclocytic (tending to complex anisocytic or
—6 anomocytic) with 4 subsidiary cells; peristomal rims present; outer stomatal ledges not
cell very conspicuous, polar T-pieces absent; guard pairs on average 21—23 jim long,
wide. Giant stomata radier frequent. Cork warts absent or infrequent and 18—21 fim very ofirregular shape.
— thick. Adaxial and In transverse section: Lamina dorsiventral, 90
abaxial cuticle thick thinner. cells to those of adaxial 1 / epidermis often with convex inner walls and periclinal subdivision walls, possibly con- taining mucilage; proportion of subdivided cells variable with the specimen. Abaxial epidermal cells smaller than adaxial cells. Hypodermis absent. Mesophyll composed of i layer of palisade cells and unlignified spongy tissue. Midrib with flat to grooved adaxial surface and raised abaxial surface, supplied with crescentiform main vascular strand and small dorsal Ground of two bundles, accompanied by little or no sclerenchyma. tissue with midrib collenchymatous. Veins parenchymatous unlignified bundle sheaths, ex- in smallest veins in tending to lower and upper epidermis most cases; only embedded mesophyll. Bundle sheath enclosing small adaxial and abaxial fibre strands in major veins. Petiole druses with a similar system as midrib. Crystals infrequent to frequent as in meso- phyll, but mainly in vicinity of veins. Material studied. Nemopanthus mucronata (L.) Trel. (N. fascicularis Raf.); CANADA, Ontario: Kirk 308, Rolland Germain — Herb. — Quebec: 1444. U.S.A., Pennsylvania: J. R. Lowrie s—8 (1852). BELGIUM, cult. Calmpthout: Boom 27763. ONCOTHECA Plate (see VI, 4 & 5; fig. 19—21) In surface view: Glabrous. Cuticle smooth. Epidermal cells of both adaxial and abaxial surface with straight to slightly curved anticlinal walls; anticlinal cuticular flanges midrib well-developed, not pitted. Cells overlying hardly modified adaxially, square to elongate abaxially; cells over vein's not modified. Stomata confined to abaxial epidermis, 320 BLUMEA VOL. 22, No. 3, 1975 Fig. 10—15. Camera lucida drawings of adaxial epidermis in transverse section, x 300. — 10. Ilex amelan- chier idioblastic mucilage cell dotted. — 11. Ilex theezans (Reitz & Klein 3222), frequent periclinal subdi- visions and infrequentbulging mucilagecells. — 12.Ilex theezans (Dusén 10608),infrequentbulging mucilage — cells with periclinal subdivisions. 13. Ilex archboldiana, frequent anticlinal and periclinal subdivisions. — into — 14. Ilex umbellata,strongly bulging mucilage cells deeply penetrating palisade tissue. 15. Ilex argentina, ibid., with frequent subdivisions. Fig. 16. Abaxial epidermis of Ilex guianensis ( Broadway 4052), mixture of anisocytic and cyclocytic stomata, x 300. ofunusual be and randomly orientated, type: possibly to interpreted as paracytic anisocytic with subdivided subsidiary cells (Plate VI, 5), peristomal rims absent, outer stomatal ledges well-developed, inner ledges small, polar T-pieces absent, guard cell pairs on stomata m long, 20—22 m wide. Giant occasionally Cork warts average 29—32 [i [i present. absent, but necrotic patches of epidermal tissue rather frequent. In section: Lamina dorsiventral thick. Adaxial transverse 470—510fim cuticle abaxial cuticle — thick. cells 2—5 /im,1 1 2/um_ Epidermal, square to flattened, adaxially , f x than Adaxial of of larger abaxially. hypodermis well-developed, composed (i—2) 3 layers thick-walled continuous midrib. ± square parenchyma cells, over Mesophyll composed of of tall to short cells and rather 2—3 layers palisade compact unlignified spongy tissue differentiated thick- which is more or less into a thin-walled adaxial portion and a more walled abaxial portion. Midrib with slightly raised adaxial and prominently raised abaxial with vascular Ground of midrib surface, supplied system as in fig. 21. tissue parenchyma- P. Baas: Anatomy ofAquifoliaceae etc. 321 cells and tous tocollenchymatous, buton adaxial side composed of I layer of short palisade 3-layered hypodermis. Veins embedded, with parenchymatous unlignified bundle sheaths, enclosing well-developed adaxial sclerenchyma caps; abaxial sclerenchyma caps only Petiole with vascular & devoid present in major veins. complex open system (fig. 19 20), in tissue of of accompanying sclerenchyma. Crystals present as ordinary druses ground abaxial and adaxial lamina, and as minute star- or needle-shaped crystals in unspecialized epidermal cells. Thick-walled brachysclereids infrequent in ground tissue of midrib and petiole. Material studied. Oncotheca balansae NEW McKee Baill.; CALEDONIA: 3439, 3597, 4S49. PHELLINE & & (see Plate VIII IX; fig. 9 30) In surface view: Glabrous. Cuticle of adaxial epidermis mostly smooth, some- times finely granular or faintly striated; abaxial cuticle more frequently striated or granular (if striated, striae sometimes confined to neighbouring cells around stomata). Adaxial epidermal cells polygonal with straight to curved anticlinal walls; abaxial unspecialized due cells with straight to undulated walls, not pitted or with granular appearance to closely spaced irregular pits in anticlinal cuticular flanges. Cells overlying midrib usually sometimes square to elongate, rarely flattenedabaxially. Veins mostly not prominent, only also confined abaxial primary or even more rarely secondary veins prominent. Stomata to with epidermis, randomly orientated, anomocytic, 4—7 (mostly 4—5) neighbouring with inner stomatal cells, well-developed peristomal rims, very conspicuous and outer ledges, without polar guard cell long, wide. Giant T-pieces; pairs 45—88/ transverse section: thick. Adaxial In Lamina dorsiventral, 240—480[im abaxial thick. cells cuticle 1— thick, cuticle Epidermal erect to flattened, 2611m 1—14/Lm abaxial cells slightly or much smaller and more strongly flattened than adaxial cells. Hypodermis absent. Mesophyll not distinctly differentiated into palisade and spongy tissue but composed of c. 2 adaxial layers of very short cylindrical to almost isodiametric cells and loose tissue. Midrib with raised adaxial and fairly unlignified spongy prominently abaxial number of vascular bundles surface, containing a varying centric (amphicribral) of the whole (4—12) and a varying number collateral 'split-off' vascular bundles (o—6), fused and built of centric bundles in of P. lucida system not clearly up one specimen centric (McKee 5337), bundles with or without thin sclerenchyma sheaths, segments of Ground bundles separated by sclerenchymatous or parenchymatous 'rays' (Plate VIII, 4). tissue of midrib parenchymatous to collenchymatous. Veins embedded in mesophyll, with unlignified parenchymatous bundle sheaths enclosing no or little (mainly abaxial) with similar of vascular sclerenchyma. Petiole a or slightly more simple arrangement bundles as midrib with bundles distal end of absent (Plate IX), 4—15 in petiole. Crystals Sclereids or infrequent as minute druses in petiole, more rarely also in midrib and lamina. but sclerified cells in dumbeënsis and sclerified absent, some mesophyll present P. some ground tissue cells in midrib and petiole ofP. comosa and P. dumbeënsis. Material studied. variation anatomical See table I, which also gives the species-to-species of some leaf features. On the whole the leaf of Phelline is anatomy very constant. BLUMEA 322 VOL. 22, No. 3, 1975 in — Fig. 17. Stomatal types Ilex. a=anomocytic; b=cyclocytic; c=anisocytic; d=bicyclic; e=complex. Broken lines indicate presumed possibilities for phylogenetic derivation. SPHENOSTEMON Plate (see XI; fig. 22—24) In surface view: but unicellular hairs Usually glabrous, adpressed infrequent on abaxial surface of S. arfakensis; larger unicellular hairs with a narrow base and a much thicker in of and undulating body infrequent i specimen S.■ papuanum (BW 12160) abundant in S. pauciflorum. Adaxial cuticle smooth to coarsely granular; abaxial cuticle Adaxial smooth, faintly striated or granular. epidermal cells with straight to slightly un- dulated anticlinal walls, unspecialized abaxial cells with straight to strongly undulating ('zig-zag') walls. Cuticular flanges well-developed, rarely pitted. Cells overlying midrib cells usually elongate, over primary and secondary veins (rarely also over tertiary veins) usually modified in abaxial epidermis, rarely modified in adaxial epidermis. Stomata confined abaxial to epidermis, randomly orientated or with pores tending to be parallel with —6 absent primary veins, anomocytic with 4 neighbouring cells; peristomal rims or poorly developed, outer and inner stomatal ledges well-developed, polar T-pieces absent or only partly distinct (only cross-piece), guard cell on long, pairs average 30—35 /im wide. Cork 25—33/rm Giant stomata rarely present. warts usually absent but present and of outline adaxial surface of of S. irregular on 1 specimen papuanumi (NGF 19932). P. Baas: Anatomy of Aquifoliaceae etc. 323 over different — leaf anatomical featuresfeatures different Table I Phelline. Distribution of some species and specimens. of cell pitted petiole end in adaxial or cells (p) (centric) distal flanges (m) striated granular epidermal size of undulated thickness in midrib of Material studied sclerified mesophyll cuticle cuticle abaxial walls cuticular stomatal lamina thickness cuticle and number bundles petiole (all from New Caledonia) P. biHardieri Panch. - + - - Franc s.n. (=Herb. Bonati 439) + 58 x 52 360 4—6 15 P. brachyphylla Baill. - McKee 10020 - - + - 45 x 44 300 1 —2 8 P. comosa Labill. McKee 4552 a + - - - 66 x 59 400 3—4 P 6 McKee 4552 b* + - - - 300 6—8 P 9 P. confertifolia Baill. - Schlechter 15215 - + - + 72 x 69 320 10—12 4 P. dumbeënsis Guill. McKee 4540 + + - - 74 x 69 480 8—10 m+p 13 P. erubescens Baill. McKee 5367 - - - - 49 x 46 360 3—4 - 4 P. lucida Vieill. - - - McKee 4856 + + 80 x 72 390 18—22 7 McKee 5337 - + - + 88 x 77 460 24—26 - 5 - + - x —14 - Franc s.n. + 72 67 350 12 7 - - - Schlechter 15340 + + 64 x 57 240 4 —5 5 P. macrophylla Baill. + - - - Franc 2489 + 58 x 55 370 7—8 17 - - + - - McKee 2686 68 x 62 400 9 —11 7 - - - - Balansa 583 + 47 x 46 240 2 —4 15 y same of McKee 4552 two leaves were studied of the same herbarium sheet; stomata were not measured in b. thick. Adaxial In transverse section: Lamina dorsiventral, 250—670/tm cuticle — abaxial cuticle thick. cells to flatten- 1 6/um thick, 1 —3 /urn Epidermal (erect) square ed, adaxially much bigger than abaxially. Hypodermis absent or 2—3-layered of ± collenchymatous translucent cells on adaxial side of lamina, continuous but less regular midrib. of tall cells and variable over Mesophyll composed of 2—5 layers short to palisade a of Midrib with raised to amount lignified or unlignified spongy tissue. slightly grooved adaxial surface and prominently raised abaxial surface, supplied with open ± V- to incurved or with closed arch-shaped vascular system with strongly margins a ± triangular both sheathed amounts of fibres. system (fig. 22—24), in cases by varying sclerenchyma 324 BLUMEA VOL. 22, No. 3, 1975 Ground tissue of midrib collenchymatous to parenchymatous; lignified in 'pith'. Veins embedded in mesophyll, with lignified or unlignified parenchymatous bundle sheaths varying amounts of abaxial Petiole distal with enclosing sclerenchyma. through part open closed with additional latero-dorsal which be or system as in midrib, wing bundles, may centric collateral or (amphicribral). Crystals present as medium-sized to very large styloids of (up to 350X 30/tm) in mesophyll and phloem major vascular bundles. Material studied. See table II. Note. Within Sphenostemon there are two distinct leaf anatomical groups: the New Caledonian which characterized species, are by a multilayered epidermis and a closed vascular system in petiole and midrib, and the remaining species. This coincides with the infrageneric classification proposed by Van Steenis (1955). Table II — Sphenostemon.Sphenostemon Distribution of some leaf anatomical featuresfeatures over differ- ent species andand specimens. rH «M P 0) X M X ■H C •H Id N M co a A 3 p 01 p id ■P 5 •H O >1 CO id •o p p C CP CO ■H CO "O CO CO 0) <0 c to CP id h c x c •H •H -O X -0 NEW CALEDONIAN SPECIES S. balansae Baill. - - - + - Balansa 1330* 360 + + S. comptonii Baker Compton 1693* ~ 670 + + - - - - - + - + + - + Bernard! 10164 31 X 28 570 S. pachycladum Baill. McKee 23070 - - - + - 31 X 25 580 + + - - + + - McKee 5428 - - - + - 30 X 27 480 - - - - + - + + - - Franc s.n. 31 X 27 610 NEW GUINEAN SPECIES S. arfakensis (Gibbs) Steen. + + Kostermans 2217 + " + + + 35 X 32 380 - - S. papuanum (Lauterb.) Steen. - " - + - - - - + NGF 19932 31 X 29 370 BW 12160 (transverse + + 250 + sections only) S. pauciflorum (Smith) Steen. ++ - + + - - + + Brass 30858 31 X 30 290 QUEENSLAND SPECIES S. lobosporus (F.Mue 11.) Smith + Rodd 269 - - - + + 32 X 30 400 - - + - - Clemens (1947) - " + + + 34 X 33 310 + + V* specimens of which only slides were available borrowed from the Jcdrell Laboratory, and which could not bebe used for certain characters.characters P. Baas: Anatomy of Aquifoliaceae etc. 325 Nodal anatomy ILEX (see fig. 25—28) of Ilex is illustrated in —28. The nodes be The range in nodal anatomy fig. 25 may I. or trilacunar with three unilacunar with a single trace (I. asprella, I. longipes, and serrata) other ofwhich be of traces (the species studied), the median one may composed 3 —7 more less or separate units. node with vascular strand the The unilacunar corresponds a single throughout petiole of in in the species studied (fig. 25). The rudimentary stipules are devoid vascular supply this case. The into several trilacunar nodes may be subdivided types: and from to the stele without taking in 1. Lateral traces very small running stipules part vascular supply of the petiole (fig. 26; I. goshiensis and I. verticillata). off side branches the of size, sometimes centric, 2. Lateral traces giving to stipules, varying in the I. and assuming the position of latero-dorsal wing bundles petiole (fig. 27; anomala, I. aquifolium, I. clemensiae, I. coriacea, I. grandifolia, I. hypoglauca, I. oppositifolia I. I. I. and I. p.p., I. pedunculosa, spicata, stapfiana, theezans, zygophylla). but with the median vascular strand in the petiole 3- Lateral traces as above, fusing and Latero-dorsal bundles then split offat a (I. opaca, I. oppositifolia p.p., I. versteeghii). higher level in the petiole (fig. 28). the The median vascular strand always enters the stele at the level of leafinsertion on but the laterals enter the stele at a much lower level, even below the stem, may sometimes and next lower leaf insertion (I. clemensiae I. pedunculosa p.p.). Unusual deviations from the above variants have been encountered in some specimens of I. oppositifolia, and in I. hypoglauca. In I. oppositifolia (Clemens 31375) one side of the the side opposite-leaved node conforms to the normal trilacunar condition, but opposite and the has with lateral traces side of the median trace 1 on other side. 4 gaps 2 on one of three Clemens the node is trilacu- The median traces are each composed units. In 31895 1 but median trace of three units, and the other of seven. nar on both sides, one is composed the of wound tissue insect and the The latter trace is in vicinity (probably attack) splitting the have traumatic In Clemens 2557 the situation is up of median trace may a cause. 3 _ again different with single median strands and lateral traces which are fused with the the median traces are of median trace over some distance. In I. hypoglauca composed 3 with three independent traces at the more or less distinct units, probably corresponding The different nodal conditions be linked with a diversity of primary growth stage. may the distal end ofthe but besides the few correlations vascular tissue arrangement in petiole, mentioned above, it should be noted that in Ilex complex vascular bundle arrangements nodes. are always associated with trilacunar anatomical descriptions, unless stated otherwise). Material studied (for localities see material cited under leaf I. De Wilde & I. PNH I. I. anomala, Heller 2735; aquifolium, Dorgelo 2528; asprella, 21584; clemensiae. I. , Voogd I. hypoglauca, Clemens 50259; I. goshiensis. Beyrick 926, Moran 5131 (Japan); grandifolia, De 1134; Biltmore Herb. I. Lawrence & Dress I. oppositifolia, Anderson 8536, Hallier 2875; I. longipes, 4063; opaca,, 246; Wilson Moran I. serrata, Boom 61651 I- Clemens 31375, 31895, 32557; I. pedunculosa, s.n., 5329 (Japan); Haviland Hallier I. Dusén spicata, VanOostsîroom 13113, I. stapfiana, 3357,,Jacobs 5057, 650; theezans, 10608; SAN 1. Brass 11772; I. verticillata. Rolland-Germain 6216; I. zygophylla, 39419. versleeghii,, BLUMEA VOL. No. 326 22, 3, 1975 Vascular bundle in the distal end of the in Ilex. — Fig. 18. arrangements petiole Types 1—4: simple open; type 5: abaxially open; types 6 & 7: simple closed; type 8: complex closed; type 9: complex open; type 10: ‘double’. — Xylem hatched, phloem dotted. NEMOPANTHUS Node unilacunar with a single trace. Material studied. N. mucronata, Rolland-Germaitt 1444. ONCOTHECA (see fig. 29) with Node pentalacunar 5 traces, the median one somewhat larger than the laterals. The individual bundles become incurved to the basal of the and strongly centric in part petiole show some fusion and splitting up higher up in the petiole. Material studied. O. McKee balansae, 3439, 4859. PHELLINE (see fig. 30) three the the Nodes trilacunar with traces, all of which are ± centric in cortex and of petiole, through incurving the collateral stelar traces. Three types may be distinguished: 1. Lateral traces attached to the stele at the same level as the leaf andmedian trace. Splitting the the of bundles occurs from basal part of petiole upwards (J(P. brachyphylla and P. lucida). 2. Lateral traces attached to the stele several nodes below the innervated leaf, giving the impression of cortical bundles throughout the stem. Splitting of bundles as in I. (P. erubescens). 3. As above, but splitting of both median and lateral traces occurs in cortex prior to the of entering petiole so that 6(—7) traces enter the base the petiole (.P. billardieri). P. Baas: Anatomy ofAquifoliaceae etc. 327 The final numberof vascular bundles in the distalend ofthe petiole results from varying of numbers of splits or 'branching' median and lateral traces in the petiole. In the simplest bundle cases only the median splits into an adaxial and abaxial bundle so that the final number ofbundles is 4 (see table I). Material studied P. billardieri,Herb. Bonati 439; P. brachyphylla, McKee 7944; P. erubescens,, McKee 3367; P. lucida. Schlechter 13345■ SPHENOSTEMON Nodes trilacunar with three traces. The lateral traces fuse with the median trace in the base ofthe petiole, and always are connected with the stele at the same level as the median in S. McKee the with trace (as fig. 28). pachycladum, 5428, presents only exceptional case a unilacunar node with a trace composed of three units; large amounts ofnecrotic tissue in the pith of this specimen are suggestive of traumatic causes underlying this aberrant nodal anatomy. Material studied. S. arfakensis, Kostermans 2217; S. cf. comptonii, Bernardi 10164; S. pachycladum, McKee Franc S. NGF S. Rodd 5428, s.n.; papuanum, 19932; lobosporus, 269. Twig anatomy Based on limited observations. ILEX absent leaves but often Cork Hairs or present, as on more frequent. arising in subepider- flattened cells with mis, mostly composed of cells, in some species containing unilateral of which adaxial lignified wall thickenings, in I. zygophylla palisade-like cells are mostly heavily sclerified. Cortex homogeneous and parenchymatous to collenchymatous, or with scattered brachysclereids. Perivascular sclerenchyma in a more or less continuous of fibres with absent from cylinder, composed interspersed brachysclereids, young twig in nodal of I. coriacea. Pith parenchymatous, lignified or not, with very large air spaces I. coriacea smaller in the inter- region of (aerenchyma; the intercellular spaces are much nodes of this species). Crystals mainly as clusters in cortex and pith, but also as solitary confined the prismatic ones in some species (this type mainly to pith), rarely as styloid- like size mixed with clustered in nodal crystals (maximum 30x8 /mi) ones region (I. and abundant nodal than zygophylla I. hypoglauca). Crystals always more in region in internodes. Material studied. See under Nodal anatomy. NEMOPANTHUS Glabrous. Cork arising in subepidermis, composed offlattenedcells. Cortex parenchyma- Pith tous to collenchymatous. Perivascular sclerenchyma ring as in Ilex. not lignified or clusters less lignified in peripharal part only. Crystals present as in cortex and frequently in pith. Material studied. under Nodal and old studied. See anatomy; a very young a 3 year twig were ONCOTHECA Glabrous. Cork arising in subepidermis, composed of flattened cells. Cortex parenchym- atous, containing idioblastic branched sclereids. Perivascular sclerenchyma composed of individual fibre linked each other sclerified Pith groups to by parenchyma tissue. lignified No. 328 BLUMEA VOL. 22, 3, 1975 with clusters of thick-walled brachysclereids. Crystals as irregular clusters in cortex and pith parenchyma. Material studied. under Nodal See anatomy. PHELLINE (see fig. 9) Hairs confined to leafaxils, thin-walled uniseriatewith subglobular top-cells (see fig. 9; glandular?). Cork arising in subepidermis, composed of square to flattened cells, containing Fig. 19—24. Camera lucida drawings of midrib and petiole in transverse section, x 17. — 19—21. Oncotheca balansae (McKee 4849), basal and distal part of petiole, and midrib respectively. — 22 & 23. — Sphenostemon lobosporus (Rodd 269), basal and distal end of petiole. 24. Sphenostemon pachycladum (Franc s.n.), distal end ofpetiole. — Xylem hatched, phloem dotted, sclerenchyma black. P. Baas: Anatomy ofAquifoliaceae etc. 329 a varying number of cells with unilateral adaxial U-shaped lignified wall thickenings. Cortex parenchymatous to collenchymatous, in some species containing brachysclereids. Perivascular absent of individual fibre Pith sclerenchyma or composed groups only. paren- minute druses in chymatous, lignified or not. Crystals present as very rare to frequent cortex and pith. Material studied. under Leaf See anatomy (table I). SPHENOSTEMON Indumentum usually absent, in hairy species early caducous because of cork formation (S. arfakensis). Cork arising in subepidermis, composed offlattened cells, containing some cells with unilateral adaxial lignified wall thickenings. Cortex parenchymatous, containing variable of sclerotised cells. Perivascular scleren- a amount (usually very high) lignified to of small fibre linked sclerified Pith chyma composed groups by lignified to parenchyma. mostly entirely lignified, rarely of mixed lignified and unlignified cells (S. pachycladum). size in but in Crystals as styloids of varying (140x8 —200 X 14/mi) mainly cortex, also pith. under Leaf Material studied. Sec anatomy (table II). Bark anatomy Based on limited observations. ILEX Tissue outside secondary phloem usually split off during microtechnical procedures and of and not studied (cf. Twig anatomy). Preserved in slides I. marquesii composed ofbroad homogeneous cork zone of flattened cells, cortical tissue containing sclerified cells and of and mixed groups cells, a perivascular sclerenchyma ring. Ground tissue ofsecondary phloem withhomogeneous appearance in transverse sections, and composed of sieve tubes, companion cells phloem parenchyma; at some distance from the with and distributed of thick-walled cambium infrequent irregularly large groups brachysclereids, which are often crystalliferous. Fibres absent. Rays broad and narrow, but the former triangular in transverse section due to dilatation, mostly unlignified, parts adjacent to sclereid-groups in ground tissue sometimes sclerified; entirely lignified in and confined I. sp. (Jacobs 7928, a climbing species). Crystals solitary prismatic to sclereids. Material studied. Miscellaneous specimens of the following species studied for wood anatomy with adhering bark fragments (for collection numbers see Baas, 1973): I. anomala, I. aquifolium, I. caroliniana, I. cassine, I. curranii, I. decidua, I. fargesii, I. glabra, I. montana, I. sideroxyloides, I. subrugosa, I. theezans, and I. sp. (Jacobs 7928, climber from the Philippines). NEMOPANTHUS all Broad Similar to Ilex bark in characters. rays not lignified. under Material studied. N. mucronata, see Wood anatomy. ONCOTHECA Cork layers also deep-seated in secondary phloem, composed of layers of thin-walled flattened cells, some of the cell-layers being crystalliferous. Ground tissue of sieve tubes, companion cells, and parenchyma alternating with ± of of which are Fibres layered patches axially elongate sclereids, many crystalliferous. 330 BLUMEA VOL. 22, No. 3, 1975 — Fig. 25—28. Reconstructions of nodal types in Ilex. 25. Unilacunar with 1 trace. — 26. Trilacunar; lateral traces supply stipules only. — Trilacunar; lateral continuous 27. traces as wing bundles throughout petiole; stipules supplied by side-branches of lateral traces. — 28. Trilacunar; lateral traces fuse with median trace. — 29. Node of Oncotheca; pentalacunar. absent. Rays broad and narrow, the former with very little dilatationand usually sclerified. Crystals of irregular shapes but mostly solitary in special cork layers; solitary prismatic in phloem sclereids. Material studied. O. balansae, McKee 22947. PHELLINE Material bark of mature not available. Secondary phloem of sieve tubes, companion cells and traversed and broad in phloem parenchyma by narrow unlignified rays twigs. absent. Crystals See also under Twig anatomy. SPHENOSTEMON Cork superficial, of thin-walled flattened cells. Cortical tissue almost entirely sclerified. Perivascular of with sclerenchyma mainly brachysclereids, interspersed small fibre groups. of and sclerified Secondary phloem sieve tubes, companion cells, partly phloem paren- with chyma interspersed secondary phloem fibres. The latter more or less tangentially banded diffuse in broad and or aggregates. Rays narrow, the former triangular and mostly sclerified. Crystals abundant as styloids, mostly accompanied by small irregular 'satellite' crystals in phloem parenchyma cells. P. Baas: Anatomy of Aquifotiaceae etc. 331 wood anatomical studies ofthe Material studied. Bark fragmentsadhering to specimens used for following species: S. lobosporus, S. pachycladum, and S. papuanum. Wood anatomy ILEX See Baas, 1973 NEMOPANTHUS (see Plate VI, 1 & 2) General features: Wood light and fairly soft. Colour white. Growth rings distinct. Texture fine. Heartwoodnot differentiated. Microscopic features: Wood semi-ring porous. Growth rings distinct. 2 —156/mm solitary and in radial multiples of 2—4_( —7), 41 —46% solitary, Vessels 98 , . member — vessel length angular to oval in T.S., tangential diameter 21—34—35 —610— Inter-vessel rarely tending to alternate; pits 380—570 780jum. pits opposite, enclosed within the — with to pit rounded rectangular 4 6/um, oval slit-like apertures borders. Vessel—ray and vessel—parenchyma pits similar but half-bordered. Perforations scalariform end walls with —18 bars. Helical thickenings and in oblique 13 —20—27 first-formed vessel contents absent. Thin-walled tyloses in MADw 19208only occurring in secondary xvlem. Ground tissue composed of thin-walled fibre-tracheids with bordered ______the with slit-like pits mainly confined to radial walls; pits c. 4/tm apertures extending fibre-tracheid Helical beyond the pit borders; length 580—750—850—980/im. thickenings strands of to 6 cells. Rays absent. Parenchyma scanty diffuse, mostly 4, occasionally up uniseriate multiseriate cells heterogeneous II, rays 9—11/mm, rays up to 5 wide, 3 —4/tnm, sheath and tallest to high, infrequent cells Crystals, silica, pith rays up 1.3 mm present. flecks absent. Material studied. U.S.A. N. mucronata: MADw 783, MADw 19208, SJRw 40391 (material of stems c. 3—4 cm diam.; low altitude). ONCOTHECA (see Plate VII) General features: Wood heavy and hard. Colour brown. Growth rings absent. Texture rather fine. Heartwood not differentiated. Growth absent. Vessels Microscopic features: Wood diffuse porous. rings 2 & 46/mm 88 & solitary, rarely in radial pairs, round to oval in T.S., tangential 29 , 96% & — diameter — vessel member Inter- 34—55 length 460—990 1030 round with slit-like vessel pits alternate to diffuse, to oval, 4—5 /urn, slightly oblique confined within the borders. and ± apertures pit Vessel—ray vessel—parenchyma pits similar but half-bordered andsometimes elongate and transitional. Perforations scalariform sometimesalmost Helical in oblique end walls with 8— 15—24 bars, reticulate. thickenings Ground and tyloses absent. Dark solid contents present in vessels outside sap wood region. bordered tissue composed ofvery thick-walled fibre-tracheids with very numerous pits on both radial and with slit-like confined within the tangential walls, pits 4—5 [im apertures Helical thickenings pit-borders; fibre-tracheid length 820—1150 & 1550—23 70/tin. scattered diffuse absent. Parenchyma mainly very scanty paratracheal, rarely as cells, cells. uniseriate strands of 3—7, mostly of 4—5 Rays heterogeneous II, rays ii/mm, cells tallest to 1.2 mm high; infrequent multiseriate rays 2—3(—5) wide, 3/mm, rays up rhomboidal sheath cells present. Central cells of broad rays square to erect. Solitary cells. Silica and absent. crystals frequent in ray pith flecks 332 BLUMEA VOL. 22, No. 3, 1975 of vascular bundle from node distal end of Fig. 30. Diagrams arrangement to petiole in Phelline billardieri. — Lateral traces and their derivatives black. Broken lines indicate derival from the same bundle. McKee Material studied. NEW CALEDONIA. O. balansae: 22947, Schmid s.n. (material of stems 3—4 cm diam.; altitude 600—1200 m). PHELLINE (see Plate X) General features: Wood fairly light and soft. Colour yellowish. Growth Texture rings indistinct. coarse through very conspicuous broad rays. Heartwood not differentiated. Microscopic features:. Wood diffuse porous. Growth rings very faint to absent. 2 Vessels 5i/mm mainly solitary (c. 90%), occasionally in radial angular to , pairs, roundedin T.S., diameter vessel member tangential 29—43—63 fi m, length 1530—2130— Perforations scalariform almost vertical 2900/rm. in end-walls, resulting in strong overlap of with bars. Inter-vessel a vertical adjacent vessel-members, 91 —134—204 pits in single oval row, and variously flattened, 7 —10/im wide. Vessel—parenchyma pits similar but half-bordered. Vessel—ray pits transitional to scalariform, half-bordered, occasionally unilaterally compound. Tyloses, vessel contents, and helical thickenings absent. Ground tissue of with the composed fairly thick-walled fibre-tracheids numerous bordered pits on radial and with sht-like the tangential walls; pits 6—■7 Mm apertures extending to pit „ , x , r border fibre-tracheid Helical absent. margins; length 1760—2420—3030/401. thickenings and Parenchyma diffuse, often touching on the vessels easily to be mistaken for vessels in because of similar cell strands cells T.S., ± diameter, 4—7 long. Rays heterogeneous II, uniseriate rays 4/mm, broad rays extremely large, occupying about 40% vol. of the P. Baas: Anatomy ofAquifoliaceae etc. 333 cells often of and wood, up to 14 wide, over 2 cm high, composed large cells, with fre- quent sheath cells, 1.3/mm. Crystals, silica and pith flecks absent. studied. Phelline NEW CALEDONIA: 600—1200 Material lucida Vieillard; SJRW 14399 (altitude m). SPHENOSTEMON (see Plate XII) General features: Wood of medium density and hardness. Colour light brown to reddish brown. Growth rings absent or very faint. Texture rather fine, but broad rays usually visible to the naked eye, especially in S. lobosporus. Heartwood not differentiated. Wood diffuse Growth absent Microscopic features: porous. rings or ex- 2 remainder radial tremely faint. Vessels 12—24/mm —84% solitary, in multiples of , 37 2 oval to vessel diameter vessel —4, slightly angular, tangential 45—71 —90—143 fxm, member Inter-vessel length 950—1690—2730—3500/tm. pits typically scalariform, covering the whole tangential wall with a single row of pits, infrequently transitional to opposite. Vessel—parenchyma and vessel—ray pits mostly opposite to transitional, also and diffuse to alternate in S. papuanum, sometimes tending to scalariform in S. lobosporus, in half-bordered, 6/um or wider, rarely unilaterally compound. Perforations scalariform with bars. of several very oblique end-walls, 32—67—118—185 Bars perforation plates repeatedly forked to form a honey-comb-like pattern (Plate XII, 5) in part of the per- forations. Tyloses, vessel contents, and helical thickenings absent. Ground with tissue composed offibre-tracheids bordered pits on both radial and tangen- tial walls; those on the tangential walls only where in contact with parenchyma; pits c. with slit-like the fibre-tracheid 5 fi m oblique apertures extending beyond pit borders; Helical absent. diffuse in length 920—1840—3100—3920/rm. thickenings Parenchyma in aggregates (mainly short tangential lines), rather frequently touching on the vessels, —8 cells. broad strands of 3 Rays heterogeneous II, uniseriate rays 5—8/mm, rays usually cells in 2 —4(—6) wide; S. lobosporus up to 10 cells wide, 1—3/mm, up to 3 mm high and (in S. pachycladum only up to 1.2 mm), with fairly frequent sheath cells. Silica pith flecks absent. Large styloid crystals accompanied by smaller irregular crystals in axial parenchyma of S. pachycladum only. Material studied. QUEENSLAND. S. lobosporus: Rodd 26g, CSIRO 13692, 14216,33123 (alt. c. 1200 m). — S. — NEW NEW CALEDONIA. pachycladum: McKee 23070 (ait. c. 1000 m). GUINEA. S. papuanum: Pulle 266, Pullen 5422, Schodde 1328, Vink 17086, 17492 (alt. 1200—3000 m). Notes on individual The differences wood between species. in anatomy the studied 3 species are very small and have mostly been recorded in the generic de- S. the whole in vessel of scription. papuanum covers range frequency, percentage solitary and shortest in vessels, number of bars per perforation of the genus. Vessel members are S. and in S. but the is continuous. The clear pachycladum longest papuanum, range ± only is qualitative distinctive species character the presence of styloid crystals in S. pachycladum. Judging from the twig anatomy studied of the other Sphenostemon species, the presence the of styloids in secondary xylem parenchyma characterizes the New Caledonian group of species only. 334 BLUMEA VOL. 22, No. 3, 1975 DISCUSSION OF SOME ANATOMICAL CHARACTERS Indumentum found the leaves of the studied the Hair types on genera play only a minor part in because taxonomic discussions, in most species the indumentum is absent or of an 'un- is is in the specialized' nature, that of a type which widely distributed dicots as a whole. The hair of Ilex well types are arranged in morphological series in fig. I—8. One may be imagine that these series reflect phylogenetic trends. However, no evidence can produced as to the directions such developments might have had. The — to my knowledge — hair of and be derived from the unique type I. hypoglauca is probably specialized can unicellular common type through an intermediate stage as present in I. zygophylla, which other also close species is in respects to I. hypoglauca. The unspecialized cells of the leaf epidermis is There a considerable variational range in outline of the unspecialized cells, due to varying degrees of anticlinal wall undulation in both Ilex, Phelline, and Sphenostemon. As discussed this feature is level within Ilex. An on p. 400, not even diagnostic at the species exception should probably be made for the strongly zig-zagging walls of I. verticillata, also present (but to a lesser extent) in I. longipes and I. serrata, which produce highly characteristic cell Presence absence of of anticlinal cuticular patterns. or pitting the flanges and of granular structures in the cuticular layers are variable features in Ilex, even at the level These characters have also been tested species (see p. 400). in Icacinaceae (Van Staveren & and Winteraceae in Baas, 1973), (Bongers, 1973), and were found to vary their diagnostic value in the different families and genera. The lignification of (part of) the abaxial epidermal cells in several species of Ilex is noteworthy. Though not absolutely constant at the species level (cf. I. canariensis, I. crenata, I. triflora), the distribution of this character over the different sections recognized by Loesener indicates some taxonomic importance. The variability at the species level may be dueto age of the persistent leaves involved (cf. Napp-Zinn, 1973 & 1974). There are threevariants to the feature 'abaxial epidermis lignified': (1) unspecialized cells lignified Plate I. only (I. argentina, I. boliviano, I. canariensis p.p., I. cassine; III, 1, I. crenata p.p., and cells dumosa, I. ovalifolia; Plate III, 3; (2) unspecialized cells and guard lignified (I. I. opaca, retusa, and I. triflora); (3) all cells, including subsidiary cells, lignified (I. amara; Plate IV, 1, I. martiniana; cf. similar situation in aberrant stoma of I. cassine: Plate III, 2). Cador(i90o) reported sclerified abaxial epidermal cells in I. affinis, I. amara, I. cassine p.p., I. chamaedrifolia, I. glazioviana, I. syniplociformis, I. vitis-idaea. Thcvenard(1906) and several other authors mentioned thick-walled pitted abaxial epidermal cells in some species, now known to possess lignified cells (I. cassine, and I. opaca). For I. triflora, however, Thevenard did not details about the abaxial sclerified give any epidermis. He did, however, report epidermal cells for I. chamaedrifolia. Lignified abaxial epidermal cells occur in the — according to Loesener — closely related sections Polyphyllae, Brachythyrsae, Thyrsiflorae, and well several Symplociformes (Cador, 1900) as as in sections ( Crassifoliae, Dasyneurae, and Cassinoides) which Loesener (1908) all thought to be derived from a common stock. The of this in I. isolated of section occurrence character I. triflora and argentina as species Microdontae indicate the himself. may erroneous placement in system by Loesener The range ofepidermal organization as seen in transverse section and with reference to the of also occurrence anticlinal and periclinal subdivisions is given in fig. 10—15 (see Plate & Below there is in the extent to IV, 5 6). the species level a very strong variability P. Baas : Anatomy ofAqnifotiaceae etc. 335 which the adaxial epidermis has thin periclinal or anticlinal division walls, or in the num- ber cells of differentiated into bulging mucilage cells. The absolute presence or absence of these features is, however, of rather good diagnostic value. The character complex of the cells all unspecialized epidermal in genera studied is not as such that it can be used in discussions of the wider affinities of the taxa concerned. The stomatal complex The stomatal the range in type in genus Ilex is quite considerable. Comprehensive studies in other genera have also provided examples which are heterogeneous with regard stomatal Webster for & to type, e.g. (1956) Phyllanthus; Jansen Baas (1973) for Lophope- Baas for Citronella. This should induce in the stomatal talum; (1974) great care using type in discussions of wider affinities of Ilex the enigmatic taxa. Within occurrence of several stomatal of types or intermediatetypes in one leaf certain species even limits the diagnostic value at the species level. The of stomatal in Ilex is illustrated in most from range types fig. 17. It seems plausable a of the the least and morphological point view, to regard anomocytic type as specialized, the and as derived Ilex. There anisocytic, cyclocytic, bicyclic types in is, however, no evidence this it to prove hypothesis, and seems even likely that anomocytic stomata may have from of arisen other types in certain groups of Ilex, because the species subgenus — Prinus generally regarded as specialized Ilex representatives — are often characterized by anomocytic stomata. Ilex also Anomocytic stomata as occur in some species, are present in Phelline and is Sphenostemon, so that the anatomical evidence neutral with respect to presence or ab- affinities. The sence ofmutual stomatal size range in Phelline, considerable as it is, however, is the — considerable — in and beyond equally range Ilex, Nemopanthus, Oncotheca, the of Phelline Sphenostemon. In fact, stomata belong to the largest in the Dicotyledons as a & whole (cf. Napp-Zinn, 1973 1974). This quantitative difference may be used as a weak additional argument in favour of the isolated position of Phelline with respect to the other The stomatal of Oncotheca is difficult of the genera. complex to assign to any types pro- in the & Plate different posed past (cf. Fryns-Claessens Van Cotthem, 1973; see VI, 5). It is from in the other and be in favour of any type occurring genera can used to argue an isolated position of Oncotheca. Cork warts warts the leaf Cork of regular shape and regular frequency on lower surface, are of value which them. Ilex good diagnostic in those Ilex species are characterized by Within they are widely distributed over most sections recognized by Loesener, but in combination with other characters they may be profitably used in discussions of infrageneric affinities. Controversial about opinions their diagnostic value are probably due to the fact that taxonomists, probably including Loesener, have not always been able to distinguish at low magnifications the regular cork warts from irregular traumatic cork patches, which are ofvery variable frequency in several Ilex species studied by me. Leaf hypodermis The term hypodermis is here applied to (an) adaxial subepidermal layer(s) deviating the from chlorenchyma tissue, and in mature leaves not obviously derived from the epidermis. I3LUMEA VOL. 336 22, No. 3, 1975 Ilex restricted Hypodermal development in is to some species only (see p. 389); yet considerable in these few species show a structural diversity hypodermal anatomy. The of 2—3-layered mucilaginous hypodermis I. anomala, and the lignified hypodermis of I. zygophylla constitute examples of features, which are rare in Dicotyledons as a whole Within (Napp-Zinn, 1973 & 1974). Sphenostemon the presence of a multi-layered hypo- dermis characterizes one series only, so that the character here coincides with taxonomic in absence of subdivision. The great variability presence or and/or type hypodermis in Ilex and Sphenostemon renders this character useless for discussing the wider affinities of the studied. genera Crystals, sclereids, and marginal sclerenchyma Thevenard (1906) already attached great diagnostic value to the various crystal types of Ilex, Nemopanthus, Phelline, and Sphenostemon. Considering the entire crystal complement from and Ilex shows stem leaf, a wide range with solitary prismatic, clustered, minute restricted rod-like, and styloid-like crystals. The latter types are ofrare occurrence and to leaf epidermis and nodal of few upper region respectively species only. The and of have conspicuous frequent styloids Sphenostemon always played animportant part in the discussions of possible affinities of Sphenostemon. Bernardi (1964) has rightly criticized the overweighting of this feature. In most families for which styloids have been the literature Metcalfe & are restricted a few reported in (cf. Chalk, 1950) they to rep- resentatives only. The present observations of styloid-like crystals in the nodes of two Ilex be used for of species (p. 327), might even as supporting evidence mutual affinities be Ilex and Sphenostemon. This would, however, equally unjustified as using styloids too for of and strongly negative evidence, since the crystals Ilex are much smaller have a distribution different from those found pattern in Sphenostemon. Incidentally I recently of the the styloids exactly same type as in Sphenostemon in enigmatic genus Paracryphia, also from Caledonia. This has of and nodal anatomical New genus many its twig, leaf, features in common with Sphenostemon, but its floral morphology (cf. Baker, 1921) is entirely different from that of Sphenostemon. With more detailed studies of the overall of the Dicotyledons, probably much with vegetative anatomy more groups styloid crystals would be found than reported thus far in the literature. The situation may well be com- for cells & parable to that the so-called rare cristarque (cf. Baas, 1972, a, b; Jansen Baas, in remains 1973). Yet, the presence of styloids Sphenostemon a distinctive feature, to be taken into account in the overall comparisons with other groups. The Ilex of foliar sclereids of some species, mainly belonging to subsection Repandae section — — the first time. Microdontae are to my knowledge reported here for Their distribution over very similar species make them useful indicators ofinfrageneric affinities from the the within Ilex. They are probably derived sclerenchyma tissue supporting vascular bundles. This is most clear in species like I. confertiflora and I. cornuta, where only few sclereids which of the vascular bundle are present are always, over some length, part These the leaf anatomical links Ilex with caps. 2 species moreover constitute to species section marginal sclerenchyma strands of Aquifolioides (see p. 367). In other species the less fusiform sclereids network the more or constitute an irregular throughout mesophyll which of the (Plate V, 1—3) of only a part is directly in contact with the mechanical tissue veins. mentioned also The marginal sclerenchyma strands, briefly above, are restricted to a small number of closely related Ilex species. In some species (I. confertiflora, I. insignis, I. integra, and I. latifolia) a more or less intermediate condition is present, with a rudimentary P. Baas: Anatomy ofAquifoliaceae etc. 337 marginal vascular bundle, mostly associated with massive abaxial to lateral sclerenchyma these — rather strands — are detached from the vascular bundle and caps; locally caps or then form true marginal sclerenchyma strands. It is therefore probable that the marginal sclerenchyma strands in Ilex are derived from marginal vascular bundles, and should be regarded as specialized structures. This would be in agreement with the generally special- floral of ized structure, inflorescence morphology, and stomatal type (frequently bicyclic) the the species exhibiting this feature. Likewise occurrence of foliar sclereids may be regarded as a specialization. The above discussed sclerenchymatous structures in some Ilex species are absent from the other genera, and cannot be used in the discussion of the wider affinities of these taxa. Nodalanatomy and petiole vascularization the of Sinnott's classical on nodal Ilex has been From time paper anatomy (1914), of with both trilacunar and unilacunar whilst quoted as an example a genus nodes, species intermediatein with very small lateral traces were regarded as this respect. The present confirms this but also extends the known in Ilex with the in study view, range type as I. goshiensis and I. verticillata, where the lateral traces supply the stipules only (a condition similar to that in someEuphorbiaceae, Singh, 1972), or with lateral traces entering the stele at much lower levels than the median trace, and finally cases where several traces depart from constitutes derived condition in the median gap. Probably the unilacunar node the Ilex. This is in with the distribution of unilacunar nodes with agreement over species a specialized inflorescence morphology and gynoecium (Loesener, 1908). The of in distal vascularization the petiole, particularly as seen the part, covers an even wider Ilex than nodal This also range in its anatomy (fig. 18, 25—28). range considerably limits the use of this character complex in the discussion of the wider affinities of the taxa involved. Even within Ilex its taxonomic value be because the different may disputed within types (simple and complex) may occur a single species (notably I. hypoglauca, where the peculiar hairs virtually exclude the possibility of misidentification of the material of similar Ilex been for studied). Ranges a magnitude as in have reported in instance der the genus Commiphora (Burseraceae, Van Walt, 1971) and for the tribe Tilieae of the Tiliaceae vascular (Herlemont, 19s 1). In Ilex the complex type of arrangement is mainly confined to species belonging to the subgenera Rybonia and Byronia, with the exception of two species of section Indico-Malaicae, which Loesener (1908) supposed to be ancestral stock derived from the same as subgenus Byronia. These subgeneric taxa have and advance favour of retained several primitive characters (Loesener), one may this in the vascularization the and the regarding complex patterns as primitive, simple pattern as the derived condition in Ilex. The nodes and less and petiole anatomy of Nemopanthus Sphenostemon are more or within the of but Phelline and Oncotheca be isolated range Ilex, appear to in this respect. The pentalacunar node and the associated petiolar vascular pattern in Oncotheca, and the centric bundles throughout the petiole in Phelline in particular, constitute unusual con- ditions for the whole Metcalfe Dicotyledons as a (Sinnott, 1914; & Chalk, 1950; Napp- Zinn, 1973 & 1974). most similar Petioles to those ofPhelline occur as far as I know, in some members of Winteraceae; species of Exospermum and Zygogynum also having a high proportion of bundles These centric (Bailey & Nast, 1944). species also have some collateral bundles in addition, however, and the whole vegetative anatomy and floral morphology of the Winteraceae is far from of a cry that Phelline. Many Dicotyledons, including some species 338 BLUMEA VOL. 22, No. 3, 1975 of have less latero-dorsal bundles. Ilex, Icacinaceae, etc., centric or more or centric wing It is not known whether these are comparable to the centric bundles occurring throughout the petiole in Phelline. Ontogenetic studies wouldbe needed to check this. Twig and bark anatomy The has much for histology ofyoung twigs not yielded useful information taxonomic the discussion. It is difficult to evaluate the taxonomic significance of absence of brachy- sclereids in the perivascular region of Phelline, because it is not impossible that these are in The if formed older material. pith and cortex of Oncotheca and Phelline are very broad with those of other has beenrecorded in the compared the three genera; this character not useful which descriptions, because it is impossible to give quantitative values are independ- of ent the developmental stage. Bark anatomical studies have not been carried out on a sufficiently wide scale in Dicots as a whole to evaluate the significance of the diversity reported here (Esau, 1969). For the moment the differential bark characters can only be listed and used in addition to other distinguishing characters and similarities in the systematic discussion. of that sclerification broad is Within Ilex, it is interest to note complete of the rays confined to the only climber used for this study. Although such rays are known to occur in erect trees as well (e.g. Fagus and Platanus), it is tempting to me to give a functional for interpretation Ilex: the sclerified rays together with the perivascular sclerenchyma the which of functional provide a mechanical protection of phloem, may be more signif- icance in climbers than in erect trees or shrubs. Woodanatomy The wood of all is of with its scalariform vessel 5 genera an unspecialized type per- of forations, fibre-tracheids, and heterogeneous rays. This complex characters occurs in a fairly large number of miscellaneous dicot families (cf. Metcalfe & Chalk, 1950) without affinities of the necessarily indicating mutual groups concerned. Apparently primitive has wood structure been retained in most major groups of the Dicotyledons. Ilex discussed in The wide wood anatomical range in has been elaborately a previous in Ilex have paper (Baas, 1973), and most of the latitudinal and altitudinal trends found been established trends wider of der Graaff & since as general in a sample genera (Van the Baas, 1974). Because 'ecological' or 'geographical' trends are very significant for wood features of anatomical which are generally regarded as phylogenetic importance, the other we must take the distribution of 4 genera into account in comparing theirwoods with that of Ilex. and Nemopanthus conforms in its wood anatomy to Ilex species of high latitudes, close affinities with witnesses species of subgenus Prinus through some special features (cf. Baas, 1973). Sphenostemon shows some overlap in vessel member length with Ilex, but its number of bars exceeds the values Ilex. The differences per perforation highest in quantitative gain in importance if one considers the altitudinal distribution of Sphenostemon (1000—3000 m; cf. Pronounced scalariform do Baas, 1973). inter-vessel pits as in Sphenostemon not occur in Ilex, although some species, notably I. anomala, show tendencies in this direction. As willbe discussed elsewhere closer similarities with the wood of Sphenostemon are found in Icacinaceae Yet there is overall wood anatomical between some (Platea). a great similarity some tropical Ilex species and Sphenostemon. The wood anatomy is thus inconclusive with regard to affinities and taxonomic status of Sphenostemon. P. Baas : Anatomy ofAquifoliaceae etc. 339 The wood ofPhelline differs strongly from that ofIlex in general and from species from the same latitudes and altitudes in particular in vessel diameter, vessel member length, ofbars uniseriate short number per perforation, inter-vessel pits, very large rays, and to a of vessels. vessels of a more lesser extent in percentage solitary Particularly its are primitive with bars which differ from the elaborate scalariform type its numerous closely spaced only in dicotyledons by the pit areas the tracheids of some vesselless as e.g. Trochodendron, absence ofpit membranes. Amongst the dicotyledons with vessels, the wood anatomy of Phelline belongs to the most primitive. The wood of Oncotheca also differs rather strongly from that ofIlex and the other three here the in genera, but differences are, apart fromsome quantitative ones, mainiy qualita- low tive features. Vessel member length and number of bars per perforation are in com- the and The parison to Ilex species of same altitude latitude. type of fibre-tracheid: very with radial and thick-walled very numerous large bordered pits on the tangential walls, encountered ofthe other The fact rather several is not in any genera. type is in common in families of the Theales and several other groups. The mainly paratracheal parenchyma is also different from that in the other genera. thus be of Woodanatomy may used to argue in support of the isolated position Phelline, Oncotheca, and to a much lesser extent also of Sphenostemon. However, much larger wood anatomical ranges in presumably natural families as Icacinaceae, Myrtaceae, etc., make the evidence decision about the and wood anatomical inconclusive with respect to a status affinities of those genera. One might equally well use the unspecialized wood anatomical shared all mutual affinities. would characters, by genera, to support This, however, imply also that one should support relationships between Symplocaceae, Aquifoliaceae, Cornaceae, similarlow Theaceae, Eucryphiaceae, Styracaceae, Ericaceae, etc. etc., on the basis of speciali- zation levels in their woods. AFFINITIES AND TAXONOMIC RANK Introduction In the affinities of Hua and I have summarized the my paper on Afrostyrax (1972b) results of most of with families from different orders. my comparisons widely plant I refrain from this practice here, because it would mean summing up differential characters from external morphology, vegetative anatomy, palynology, etc., etc., for numerous families. Such information is readily available in the literature to those who would like to all recorded the The discussion of the affinities explore possibilities not in following pages. has been limited and status of the different genera largely to positive conclusions only, families restricted whilst the negating of affinities with certain is to those families seriously considered relatives. been as putative The actual comparative work has carried out along the same lines as in my Hua and Afrostyrax study, and embraced all families of Celastrales and Theales, taken in the widest possible sense, together with numerous other families the for scattered throughout system, which invited more detailed comparisons one reason another. The anatomical characters used for the summarized or comparisons are in table III. SPHENOSTEMON In comparing the anatomy of Sphenostemon with that of Aquifoliaceae sensu stricto (Ilex and that few characters are Nemopanthus), it appears only differentiating present: crystal and vessel One with complement, bark histology, some characters. may contrast this a 340 BLUMEA VOL. 22, No. 3. 1975 ??) ex with clusters types to prismatic compl 60 ICACINACEAE fibres or observations. 0 300—2200 c. in other 3 polymorphous 1 impie— stratified to scalariform solitary polymorphous 00 phloem alternate styloids 1— (+ druses r-i s c. 0 original Icacinaceae. with fibres with SPHENOSTEMON 2400 for 35 118 anomocytic 30— simple stratified phloem 1600— 67— scalariform styloids poratereticulate- verrucate together (1942) 3 smooth (1973), Sleumer PHELLINE druses 88 centric Baas andànd anomocytic uniseriate colporate echinate- 45— bundles 2130 134 minute 3 of ? & Staveren characters, with stone prismatic ONCOTHECA type perforate Van and complex unusual 29—32 stratified elongated cells alternate druses solitary colporatefinely 1030 15 (1941), Callen-Lobreau. morphological 5 + characters. Howard of prismatic to & NEMOPANTHUS a fibres Madame and Synopsis stratified ite .cyclocytic 23 Bailey by alternate colporate see 21— simple without 600 oppos druses solitary clavate 1 not 18 microscopical + from provided data pris- are rarely clusters Baas (1973), & Van Staveren to important kindly ILEX complex fibres stratified or solitary rarely 42 —1920 58 scalariform some Icacinaceae anomocytic anisocytic cyclocytic 3 impie— opposite, styloid-like pollen bicyclic 18— 1— without 600 13— alternate druses matic very colporateclavate s not and of macromorphological + for on Survey Data Data length Data For of pits in — — phloem bars type guard traces) vasculari- complement III (number member of of length and Table Inter-vessel Stomatal Range cell Nodes gaps Petiole zation Secondary Vessel Number Crystal Pollen P. Baas : Anatomy of Aquifoliaceae etc. 341 overall in also in leaf good similarity wood anatomy, and to a lesser extent anatomy. Considering differential characters from the floral region: arrangement and morphology of the perianth, morphology of stamens, pollen, and fruit (Bailey & Swamy, 1953; Bailey, 1956; Lobreau, 1969), there are, however, enough reasons to look for relatives of Sphenostemon outside Aquifoliaceae. Recently Airy Shaw (1972) again advocated affinitiesofSphenostemon with Trimeniaceae, and although he was aware of the anatomical differences 'forbidding' inclusion in Tri- meniaceae it seems useful to repeat once more the anatomical differences between the stomatal oil and nodal and 2 taxa: type, mucilage cells, crystal complement, anatomy, fibres the wood of a whole of septate in Trimeniaceae provide complex differences, although in other wood anatomical features Sphenostemon and Trimeniaceae are much could alike. From the reproductive sphere one add several more important differences the absence in (floral whorl, merousness of ovary, of an obturator Trimeniaceae, morpholo- gy of the style). All these differences make a close mutual affinity very unlikely. Moni- also offer miaceae, in some interpretations embracing Trimeniaceae, an equal or even larger number of differences in anatomy and external morphology (Money c.s., 1950). Escalloniaceae, Dichapetalaceae, Guttiferae, Ochnaceae, Theaceae, and Icacinaceae are the in in remaining families mentioned the literature (see p. 313), because of resemblances one or more characters. Of these the Escalloniaceae, Dichapetalaceae, Guttiferae, and Ochnaceae need not to be discussed because of obvious dissimilarities, most of which have been extensively discussed in the literature. For anatomical differences the reader may be referred to Metcalfe & Chalk (1950). Theaceae, suggested by Loesener as possible relatives, share a number ofprimitive wood value characters with Sphenostemon (for their systematic see p. 338); however, nodal anatomy, number of integuments, pollen morphology, and many other characters (see exclude the close relative. Keng, 1962) family as a the discussion Icacinaceae, like Aquifoliaceae a member of the Celastrales, first entered when Ingle & Dadswell (1961) compared the wood of Platea with that of Sphenostemon, similar this anatomical and concluded that the genera were very in respect. The wood because of the shared inter- resemblance is very striking indeed, particularly scalariform leaf of Platea different vessel pits. The anatomy is, however, quite (Van Staveren & Baas, Platea 1973). If one extends the comparisons beyond there are, however, many points of similarity between Sphenostemon and Icacinaceae. This is mainly due to the enormous of anatomical diversity this family (Bailey & Howard, 1941; Heintzclmann & Howard, Staveren & 1948; Van Baas, 1973; Baas, 1974), but some similarities with genera of with scalariform vessel and trilacunar cf. groupI (Icacinaceae perforations nodes, Bailey & that Both Howard, 1941) are so striking they may be used to advocate natural affinity. and of rather similar that of and leaf- wood anatomy e.g. Apodytes are to Sphenostemon, the with same holds for the other genera anomocytic or anomocytic to cyclocytic stomata of bars of in also of group I. Elaborate forking the perforation plates as Sphenostemon occurs in some Icacinaceae (Metcalfe & Chalk, 1950). The few bark samples ofIcacinaceae studied by me contain phloem fibres in more or less regular bands, also including some sclerified short cells, as in Sphenostemon. The presence ofstyloids in the bark of Gonocaryum calleryanum (belonging to Bailey & Howard's group II) reported by Ingle & Dadswell (1961) I could confirm. On the whole Sphenostemon would fit into Icacinaceae quite well on the basis of The differences other characters of about the vegetative anatomy. in are order those Ilex: and several same as with perianth, stamens, ovary differing in respects. the ovules with thickened funicle shared However, pendulous unitegmic anatropous a are by Ilex, Sphenostemon, and Icacinaceae. The unusual stamens of Sphenostemon would not 342 BLUMEA VOL. 22, No. 3, 1975 be much in since wide very out of place Icacinaceae, this family shows a very range in The wide of stamen morphology (Sleumer, 1942, 1971). pollen morphological range could also accomodate the so-called unusual of Icacinaceae (Lobreau, 1969, 1973) type Sphenostemon. evidence from and external The complete vegetative anatomy morphology therefore for affinities. settle leaves possibilities open either Aquifoliaceous or Icacinaceous To this realize is question one should that choosing between the two alternatives a rather futile exercise, because Icacinaceae and Aquifoliaceae are both anatomically and also macro- also close. If the morphologically (see Valeton, 1886; Sleumer, 1942) very we compare for differential macromorphological and anatomical features Ilex, Icacinaceae, and Spheno- and be similar stemon (table III, synopsis, Sleumer, 1942), Sphenostemon appears to more to Icacinaceae than to Aquifoliaceae; in passing it should be remarked that the differences and Icacinaceae few. Bernardi's for between Aquifoliaceae are very strong plea maintaining Sphenostemon in Aquifoliaceae must therefore be declined. Accepting his solution would also and imply that one should merge the Aquifoliaceae with Icacinaceae, although the between the families such solution would be and affinities two are close, a unpractical, would not reflect the (admittedly few) constant differential characters separating the two in Icacinaceae taxa. Including Sphenostemon the would meet similar, though less strong and I therefore the for also objections, prefer to support family status Sphenostemon as the doneby Erdtman (1954) and Airy Shaw (1972). This status clearly recognizes unusual oddment floral morphology, supported by only one anatomical (styloids) and avoids of the of Celastrales well-established families. problems upsetting system by lumping I the of disagree with Erdtman's and Airy Shaw's opinions on natural affinity Spheno- stemonaceae. This monogeneric family is according to the whole character complex of related both and macromorphology and anatomy to Icacinaceae Aquifoliaceae, and should accordingly be treated as a member of the Celastrales. Bailey (1956) and Metcalfe (1956) have left the question unanswered whether Nouhuysia and Baillon should be sensu Van Steenis Sphenostemon regarded as one genus. The anatomi- cal differences reported in the descriptive part and table II are so much outweighed by I Van anatomicalsimilarities that fully support Steenis' decision (1955) to merge Nouhuysia with Sphenostemon. PHELLINE The fact that Loesener always has accepted Phelline as a member of the Aquifoliaceae for it in its taxonomic probably accounts that only recent years position has been question- anatomical and ed. Yet if we consider the morphological differences with Ilex, these are those and Ilex. The at least of the same order as between Sphenostemon very unusual petiole anatomy with concentric vascular bundles, the extremely large guard cell pairs, the vessels and with valvate very primitive narrow huge rays, together petals, hemitropous to weakiy campylotropous ovules, and different pollen ectexine make the genus rather outstanding from Aquifoliaceae sensu stricto as well as from Icacinaceae and from Spheno- stemonaceae. In the with these differences similarities comparison Ilex, onemay contrast by in fruit anatomy (Thevenard, 1906), floral plan, and an unspecialized wood anatomy. with less and the Comparisons Icacinaceae yield more or the same overall similarities wood ofPhelline is more compatible with the huge range in that family thanwith the wood of Ilex; rays similar to those in Phelline occur for instance in Citronella. Within or outside the Celastrales I claim for have not been able to find genera or plant families with a bigger affinity than Aquifoliaceae or Icacinaceae. Araliaceae, vaguely suggested by Airy Shaw in P. Baas: Anatomy of Aquifoliaceae etc. 343\I I. thick x — I. Plate I. Ilex. Hairs. — 1. brasiliensis, x 90. — 2. ibid., showing basal septs, 360. 3. cissoidea, I. base division — hairs, x 90. — 4. ibid., showing broad and anticlinal wall, x 360. 5. pubescens, long septate B base with numerous broken tips broken off. — 6. I. hypoglauca(Hallier 2875), broad septs, tip off, x 360. 13LUMEA VOL. No. 344\II 22, 3, 1975 stomata I. stomata — 2. I. Plate II. Ilex. Stomata, all x 360. — 1. amara, anomocytic. verticillata, I. stomata — I. anomocytic. — 3. I. integerrima, stomata anisocytic. — 4. clemensiae, ‘complex’. 5. pulogensis, I. —6: cuticular stomata cyclocytic. — 6. altaclarensis, stomata bicyclic. — 1 & 2: paradermal sections; 3 macerations. P. Baas: Anatomy of Aquifoliaceaeetc. 345\III III. — I. cassine Boom Plate Ilex. Abaxial epidermis. 1. ( 38139), unspecialized cells lignified, stomata cyclocytic with thin-walled subsidiary cells, x 360. — 2. ibid., aberrant stomatal complex with lignified — I. cells with subsidiary cells at low focus, x 800. 3. ovalifolia, unspecialized lignified, stomata anisocytic thin-walled subsidiary cells, x 360. — 4. I. paraguariensis, unspecialized cells thick-walled (not lignified), stomata cyclocytic with thin-walled subsidiary cells, x 360. — 5. I. aquifolium var. heterophylla, cuticle smooth, peristomal rims (arrow), SEM, x 350. — 6. I. hypoglauca(Hallier D 2875), cuticle with conspicuous striations, peristomal rims (arrow), SEM, x 730. BLUMEA VOL. 346\IV 22, No. 3, 1975 Plate IV. Ilex. — 1. I. amara, transverse section of abaxial epidermis showing thickened adaxial lignified stomata 800. I. section cell walls (arrows) and (s), x — 2. amelanchier, paradermal of adaxial epidermis with scattered mucilage cells, x 360. — 3. I. amplifolia, regular cork wart, maceration, x 360. — 4. I. coriacea, wart of I. anti- regular cork peculiar outline, maceration, x 360. — 5. archboldiana, adaxial epidermis with clinal and periclinal division walls, x 220. — 6. I. argetititia, bulging mucilage cells and division walls of adaxial epidermis, x 220. — 7. I. perado (Lems 2722), locally 2-layered adaxial hypodermis + epidermis, — 8. I. anomala of — x 220. (Van Balgooy 1689), 2-layered hypodermis bulging mucilage cells, x 90. 9. I. zygophylla, lignified adaxial hypodermis (arrow), note very tall epidermis cells, x 90. P. Baas: Anatomy of Aqtiifoliaceae etc. 347\V V. I. 1. transverse section Plate Ilex. Sclerenchyma. — 1—3. buergeri, foliar sclereids, x 220; showing elongated branched sclercids (arrow) mainly in palisade tissue; 2. ibid., note also fibrous bundle sheath section sclereids in I. (arrow); 3. paradermal showing between palisade cells. — 4. pernyi (cult. hort. Leiden), — transverse section with marginal sclerenchyma strand, x 90. 5. I. anomala (Van Balgooy 1689), lignified — 6. I. hort. bundle sheath enclosing very little abaxial sclerenchyma (arrow), x 360. bioritsensis (cult. bundle sheath massive abaxial Leiden), unlignified, enclosing sclerenchyma cap, x 220. 348\VI BLUMEA VOL. 22, No. 3, 1975 Plate VI. mucronata. 1. MADw section 1—3. Nemopanthus 19208, transverse of wood, x 90; 2. ibid., tangentialsection ofwood, x 90; 3. Kirk 308, abaxial epidermis, maceration showing characteristic cuticular and — & section ridges peristomal rims, x 360. 4 5. Oncotheca balansae (McKee 3439). 4. transverse of lamina, with 2—3-layered hypodermis, x 90; 5. abaxial epidermis, paradermal section with characteristic of arrangement subsidiary cells, x 360. P. Baas: Anatomy of Aquifoliaceaeetc 349\VII — — section, Plate VII. Oncotheca balansae. Wood structure ( McKee 22947). 1. transverse bleached, x 90. — — radial unbleached = 3. section, bleached, 4. 2. ibid., (p parenchyma cell), x 360. tangential x 90. section, note crystals (arrow), bleached, x 90. — 5. radially split surface, perforation plate, SEM, x 350. BLUMEA VOL. No. 350\VIII 22, 3, 1975 Plate Phelline. Leaf macerations. — stomatal VIII. anatomy, 1 —3: 1. P. comosa, anomocytic complex, anticlinal flanges not granular, x 360. — 2. P. confertifolia, ibid., anticlinal flanges granular, x 360. — 3. P. macrophylla (Balansa 583), anticlinal flanges undulated, x 220. — 4. P. lucida (McKee 4856), transverse section concentric of midrib bundle; p: phloem, x: xylem, s: sclerenchyma, r: ‘ray’. P. Baas: Anatomy of Aquifoliaceae etc 351\IX lucida 2. — basal of distal Plate IX. Phelline. Leaf anatomy. 1—3. P. (McKee 5337), x 36; 1. part petiole; — P. 2686 distal of of midrib. — part petiole; 3. 4. P. confertifolia, midrib, x 36. 5. macrophylla(McKee ), part petiole, x 22. 352\X BLUMEA VOL. 22, No. 3, 1975 Plate X. Phelline lucida. Wood — structure. 1. transverse section, x 36. — 2. ibid. (p=parenchyma cell), x 220. — 3. tangential section, x 36. — 4. radially split surface, scalariform perforation plates, SEM, x 370. P. Baas: Anatomy ofAquifoliaceae etc. 353\XI cuticular ano- Plate XI. Sphenostemon. Leaf anatomy. — 1. S. pachycladum (McKee 23070), maceration, — S. mocytic x 360. — 2. S. ibid., x 360. pachycladum (McKee 23070), transverse stomata, arfakensis , 3. section of lamina with styloids (arrow) and hypodermis (h), x 90. — 4. S. pauciflorum, transverse section of lamina with and of hair — section of abaxial styloids (arrow) part undulating (u), x 90. 5. ibid., paradermal epidermis showing indumentum of undulating hairs, x 90. — 6. S. arfakensis, maceration of abaxial surface with adpressed hair, x 220. 354\XII BLUMEA VOL. 22, No. 3, 1975 transverse Plate XII. Sphenostemon. Wood structure. — 1—4. S. lobosporus (CSIRO 14216). 1. section, bleached, x 36; 2. tangentialsection, bleached, x 36; 3. transverse section, unbleached, showing parenchyma — S. distribution (arrow), x 90; 4. tangential section showing scalariform inter-vessel pits, x 360. 5. of in — papuanum (Van Royen 3692), radial section, detail branching scalariform perforation plate, x 360. 6. ibid. (Pullen 266), tangential section, bleached, x 36. — 7. S. pachycladum (McKee 23070), ibid., x 36. P. Baas: Anatomy of Aquifoliaceae etc. 355 do be related to Phelline if their anatomies Willis (1966), certainly not appear to are compared. Phelline to rank, in order to do I therefore support Takhtajan's proposal to raise family anatomical features which make it distinct from justice to its macromorphological and Aquifoliaceae and Icacinaceae. Affinitiesto both these families are present, but ifall characters less close than those of with these two are considered these are probably Sphenostemon families. ONCOTHECA Ebenaceous for Oncotheca The early and often repeated suggestions ofan affinity cannot This of Guillaumin for this be supported on anatomical grounds. in spite s (1938) support and additional idea, using superficial anatomical observations comparisons as arguments. different: in of vessel fibre The woods of the two taxa are entirely type perforations, & Chalk, 1950). pitting, and parenchyma distribution (cf. Metcalfe Within Aquifoliaceae Oncotheca is out of place because of its pollen, pentalacunar node, of distribution in the wood, and peculiar stomatal type, type fibre-tracheids, parenchyma also its crystal complement. The floral morphological differences between Ilex and etc.) provide Oncotheca (insertion of stamens, stamen morphology, number of styles, for Oncotheca from The rather heterogeneous further arguments excluding Aquifoliaceae. with Oncotheca than Icacinaceae provide no further characters in common Aquifoliaceae, so Affinities that affinities with this family are also not indicated. with Sapotaceae (Loesener, if considers the anatomical differences (Metcalfe & Chalk, 1901) are most unlikely one 1950). Lobreau-Callen's suggestion (1975) of affinities with Salvadoraceae is not supported leaf of Salvadoraceae differ from Oncotheca. by anatomy. Wood and structure widely Shaw and Theaceous affinities, as recently suggested by Airy (1965) Takhtajan (1969), also to Theales but with find support in wood anatomy. Caryocaraceae, belonging a very w w nodes It seems likely that different woodanatomy, possess pentalacunar (Schofield, 1968). of the will remain out of on the grounds of the combined characters Oncotheca, genus B with see place in any of the existing Thealeanfamilies (for a comparison Theaceae, Keng, has 1962). However, since Miss Carol Bissett (U.S.A., Chapel Hill) independently em- Thealean — Oncotheca with — mainly barked on a thorough comparison of putative relatives, I refrain from further elaborate discussions here. NEMOPANTHUS The affinitiesof Nemopanthus with Ilex have never been challenged and recently (Clark, been raised whether should be with 1974) the question has even Nemopanthus not merged would from anat- Ilex. Such a taxonomic decision not meet any opposition vegetative identical that of certain Ilex omists: wood, bark, and leaf anatomy are more or less to the abaxial leaf species of subgenus Prinus. Only the peculiar cuticular markings on surface of Nemopanthus mucronata make it stand out from all Ilex species so far studied anatomically. The overall differences between Nemopanthus and Ilex are in fact smaller than those between Sphenostemon series Sphenostemon from New Caledonia and series Apetala from New Guinea and Queensland. BLUMEA No. 356 VOL. 22, 3, 1975 The position of Icacinaceae, Aquifoliaceae, Phellinaceae, and Sphenostemonaceae in the Celastrales a of for classification that From survey recent systems Angiosperm it appears no agree- ment exists about the delimitation of the order Celastrales (Thorne, 1968; Melchior in Engler's Syllabus, 1964; Hutchinson, 1974; Takhtajan, 1969; Cronquist, 1968). Of these that ofThome untraditional suborder Theineae systems is most in including Aquifoliaceae in of his Theales, whilst Icacinaceae and Celastraceae, together with Stackhousiaceae, are in of suborder Celastrineae his Santalales. However, I cannot agree with the separation of such anatomically and morphologically similar families as Icacinaceae and Aquifoliaceae into widely different orders. A discussion of the affinities ofall families of the Celastrales as based on their anatomies is it is that beyond the scope of this paper, but interesting to note Aquifoliaceae, Phellinaceae, wood of Sphenostemonaceae, and part of the Icacinaceae have the more primitive anatomy the whole order, only encountered in few other representatives of which only Cyrillaceae and Celastraceae merit further attention. Celastraceae are often regarded as very close to but there in few Aquifoliaceae, anatomically are not many points common. Although of Celastraceae their members possess scalariform perforations in woods, these are yet very different in overall wood anatomy from Aquifoliaceae (Dr. A. M. W. Mennega, Utrecht, rather similar private communication). Cyrillaceae are anatomically to Aquifoliaceae (both in wood and leaf), but characters from floral and pollen morphology have induced several students of Angiosperm classification to rank them amongst the Ericales, with also share wood anatomical features. One wonder which family they many may only whether the similarities of Cyrillaceae with Ericales as well as with some Celastrales are the indicative of affinities between result of convergent evolution, or whether they are truly the two plant orders. Within and Celastrales the group of Icacinaceae, Aquifoliaceae, Phellinaceae, Spheno- which have retained stemonaceae probably constitutes a natural assemblage of families, features. Within Icacinaceae lines of many primitive woodanatomical many specialization of show have occurred, so that some its genera now a very derived type ofwoodanatomy. and well be isolated Aquifoliaceae, Phellinaceae, Sphenostemonaceae may regarded as groups derived from the same ancestors as the primitive Icacinaceae. GENERAL CONSIDERATIONS AND CONCLUSIONS The taxonomic decisions to raise Phelline and Sphenostemon to family rank, as supported remain criticism here by anatomical evidence, open to by those who want to incorporate of & each of these odd genera into one the existing families. Carlquist (1959), Swamy Bailey (1949), and Philipson (1967) have discussed the problems of treating aberrant remain since universal rules for genera, but the matter will always controversial, no family delimitation exist. Palynological and anatomical contributions usually do not alter this problem. Maintaining Phelline in Aquifoliaceae would render the family heterogeneous for certain floral morphological, wood anatomical, and leaf anatomical features. In itself the Ilex also floral and anatomical genus covers a wide morphological range (Loesener, 1942; Baas, 1973, and this paper), but here the heterogeneity can be arranged in morphological series; in other words, the structural variation is more or less continuous. The characters in Phelline that serve to distinguish it from Ilex and Nemopanthus do not fit at all into these be used the of Phelline in distinct morphological series; this may to support treatment a P. Baas: Anatomy of Aquifoliaceae etc. 357 family. Moreover, the inclusion of Phelline in Aquifoliaceae would do injustice to the similarities this has with of the — equally strong genus some representatives structurally Oncotheca the would — For and very diverse Icacinaceae. Sphenostemon same reasoning apply in comparisons with either Ilex or any other families. the in and of the However, considering tremendous range macromorphology anatomy the of the Icacinaceae, a family generally accepted as natural, one may question validity is mention character reasoning adopted above. In this family it hard to any constant the and the structure which except anatropous pendulous unitegmic ovules, ovary varies limits of the — in within narrow (Sleumer, 1942). Most variation especially vegetative be — in continuous anatomy and pollen morphology can, however, arranged morpholog- ical series, which makes it impossible to split the family into separate groups which are distinct from each other in a combinationofseveral characters. The lack ofsuch continuous series of character complexes in the group of taxa: Phelline, Sphenostemon, Aquifoliaceae has sensu stricto, and Icacinaceae as a whole, induced me to support the recognition of Phellinaceae and Sphenostemonaceae. like Whatever the merits of Bailey's (1956) proposal to treat genera Sphenostemon as for ‘incertae sedis’, there is no reason now to maintain this position Sphenostemon and information has become available on the and Phelline, since much more anatomy mor- oftheir relatives. The phology possible proposal to treat them as members of the Celastrales is one of the major conclusions of this study. The significance of the structural range in Ilex, analysed in a separate section below, be far classification or affinities are con- cannot as yet fully established, as as infrageneric the for cerned. A new revision is required first. However, general significance systematic of the wide wood- and leaf anatomical Ilex should not be underesti- anatomy range in within that mated. The great variability of several anatomical characters a genus implies the systematic value of these characters is very restricted in discussions of affinities above the level. genus SPECIAL PART: THE LEAF ANATOMICAL RANGE IN ILEX Introduction this the leaf of of Ilex described and discussed. In section anatomy 95 species is Mostly but of the infraspecific anatomical only one specimen per species was studied, 15 species number of variation was studied in a restricted specimens. Shaw the total number of Ilex According to (1973) extant species may be estimated at The aimed c. 400. selection of species for this study was at covering most infrageneric taxa also include of recognized by Loesener, and to representatives from all regions the large of a few did distribution area ofIlex. Of the monotypic subgenus Yrbonia and sections I not study any material. Loesener’s system for Ilex of Loesener's classification (1901, 1908, 1942) Ilex remains the only comprehensive one. for from Some proposals adjustments based on material certain geographic areas (e.g. Hu, and these be indicative of 1949—1950, 1967;Edwin, 1965) have been made, although may Loesener's have followed them because this would make the shortcomings in system, I not incoherent. comparisons of the leaf anatomical data with systematic treatment reference of Loesener's from which For easy a summary system (1942), infrageneric below. taxa of which no material was used for this study are omitted, is given BLUMEA 358 VOL. 22, No. 3, 1975 P. Baas: Anatomy of Aquifoliaceae etc. 359 Subgenus Rybonia Subgenus Byronia Series A. Eubyronia Series B. Micrococca Subgenus Euilex Series A. Lioprinus Section Excelsae Subsection Umbelliformes Subsection Laxae Section Cassinoides Section Dasyneurae Section Crassifoliae Series B. Paltoria Section Polyphyllae Section Vacciniifoliae Series C. Aquifolium Section Lemurenses Section Aquifolioides Subsection Oxyodontae Subsection Insignis Section Microdontae Subsection Eumicrodontae Subsection Repandae Subsection Vomitoriae Subsection Stigmatophorae Subsection Sideroxyloides Section Prinifoliae Section Megalae Section Micranthae Section Rugosae Series D. Thyrsoprinus Section Indico-Malaicae Section Thyrsiflorae Section Brachythyrsae Subgenus Prinus Series A. Euprinus Series B. Prinoides For a better understanding of Loesener's ideas of infrageneric relationships and special- of his ization level a copy phylogenetic tree (1908) is given in fig. 31. At that time Loesener did know the which him the not yet opposite-leaved species, prompted to erect new subgenus Rybonia in 1942. It is interesting to note that Loesener's visual representation of his ideas about infrageneric classification is less rigid than his formal recognition of sub- and intercalated that genera, sections, 'series'. It shows for instance some sections from different series are in fact closer to each other than to sections from the same series in subgenus Euilex, and that section Indico-Malaicae, though treated in subgenus Euilex was thought to be derived from the same ancestors as subgenus Byronia. Specific descriptions and taxonomic notes Explanatory note In the specific descriptions only positive characters have been recorded. Absence ofhairs, ofsubdivisions in the epidermal cells, of mucilage cells, of a continuous hypodermis, of idioblastic sclereids, of marginal sclerenchyma strands, and the unlignified condition of the abaxial epidermis, the bundle sheaths and the Fig. 31. Loesener’s phylogenetic tree for Ilex (copied from Loesener, 1908: 94). Infrageneric taxa above the broken line with All derived from the trunk’ and correspond extantgroups. sections ‘main (Proto-Ilex) section Indico-Malaicae belong to subgenus Euilex. BLUMEA 360 VOL. 22, No. 3, 1975 recorded the of spongy tissue are not in descriptions. Likewise the lignification or mucilaginousnature the is recorded if in the consists of translucent hypodermis only positively present; case hypodermis ordinary this is in if is continuous whole adaxial cells not specified the descriptions. Only a hypodermis over the side of the lamina, this character is included in the descriptions. Many species of Ilex show a local differ- of the recorded. Cork recorded entiation hypodermis over midrib,but this condition is not warts are only if they are of regular outline and of frequent occurrence. Occasional deviations from the above-mentioned For stomatal size will adopted procedure are thought to be self-explanatory. average values per specimen be the full The with given only; not range. petiole types correspond fig. 18. Occasionally data have been recorded about the outline of the walls of unspecialized epidermal cells; this has only been done for strongly undulatingwalls and straight walls. In the remaininginstances of curved undulated walls the of undulation value in of to degree has been proved to have no diagnostic a number in view species (p. 400), and therefore no data on cell wall outline as seen surface has been included for those Other which species. characters, appeared to be very variable below the species level in a number ofcases (see p. 400) are also left out of the descriptions. The individual should in I notes on the species most cases be regarded as very tentative suggestions. am that in in- very well aware leaf anatomical resemblance need not be indicative of natural affinity several I stances.Yet, it is hoped that the notes will be ofinterest for future taxonomic work on Ilex. Since have not considered I am that several of will be macromorphological characters, sure my suggestions not supported, but if ofthe leads to better in the the ofthe only part suggestion a insight infrageneric relationships, purpose taxonomic notes will have been served. is is Ifreference made to suggestions by Loesener, his monograph of 1901 quoted, unless stated otherwise. Species not known to Loesener in 1908 are cited with full reference. Ilex altaclarensis = (Loud.) Dall. 'camelliaefolia' I. aquifolium x perado (subg. Euilex, series C, sect. Aquifolioides, subsect. Oxyodontae) — Plate II, 6 Adaxial and abaxial cuticle unicellular hairs midrib Lamina 410/im. 10/im. Sparse on and petiole. Stomata bicyclic, 38 x 36/rm. Adaxial epidermal cells with frequent anticlinal subdivisions. Hypodermis of 1 cell layer. Marginal sclerenchyma strand. Petiole with vascular simple open system (type 3). Material studied. NETHERLANDS: cult. hort. Leiden. Note. This specimen is very similar to I. aquifolium, I. bioritsensis, I. cornuta, I. dipyrena, I. integra, I. perado, and I. pernyii with which it shares the marginal sclerenchyma strands and bicyclic stomata together with numerous less outstanding anatomical features. Ilex amara (Veil.) Loes. var. longifolia Reiss. (subg. Euilex, series D, sect. Brachythyrsae) — Plate II, 1; IV, 1 hairs Lamina Adaxial cuticle abaxial cuticle unicellular on 270fim. 5 fim; 3 fim. Sparse midrib and petiole. Stomata mainly anomocytic, but also tending to cyclocytic or rarely 26x21 abaxial anisocytic, fim. Lignified epidermis (including neighbouring/subsidiary cells and Cork Bundle of guard cells). warts. sheaths and part spongy tissue strongly Petiole with vascular biréfringent. simple open system (type 4). Material studied. BRAZIL. Santa Catarina: Reitz & Klein 3932. Note. Anatomically there are similarities with the American species I. boliviana, I. dumosa, I. ovalifolia, I. martiniana, and I. retusa, mainly through the shared lignified epidermis and regular cork warts. I. triflora from Malesia also resembles I. amara in this respect. Ilex amelanchier Curt. series — Plate (subg. Prinus, B) IV, 2 ; fig. 5, 10 Lamina Adaxial and abaxial cuticle Indumentum of slender 2-cellular 90 fim. c. 1 fim. hairs with cuticle warty on lower leafsurface. Stomata anomocytic to cyclocytic, 20X 19 P. Baas : Anatomy of Aquifoliaceae etc. 361 Adaxial with cells. Petiole with [xm. epidermis slightly bulging mucilage simple open vascular system (type i—2). Material studied. U.S.A.: cult. Arn. Arbor. 14740 ( = herb. Boom 40305). Note. The only outstanding character of I. amelanchier is the indumentum of long, rather uniseriate, warty hairs. The thin, anatomically featureless mesomorphic leaves are also characteristic for the other species of subgenus Prinus and some other, miscellaneous species. Ilex amplifolia Rusby (subg. Euilex, series A, sect. Excelsae, subsect. Laxae) — Plate IV, 3 Adaxial cuticle abaxial cuticle Lamina 230/tm. c. 4//m; c. 2 jum. Stomata cyclocytic, cells with walls. Adaxial with 34X 29 fim. Epidermal strongly undulating epidermis anticlinal subdivisions and containing bulging mucilage cells. Cork warts. Petiole with simple open vascular system (type 4). Material studied. BOLIVIA: Krukoff 11198(U). Note. Leaf anatomy does not provide clues for affinities of this dubiously placed species (cf. Loesener, 1901: 129). Theother species of sect. Excelsae I studiedall lack regular cork warts. Ilex anomala Hook. & Am., inch f. sandwicensis (Endl.) Loes. and f. taitensis (Gray) Nadaud (subg. Byronia, series A) — Plate IV, 8; V, 5 abaxial cuticle Lamina 300 —500/im. Adaxial cuticle 6—9/mi; 3 —7/um. Stomata intermediate anomocytic to cyclocytic, 29x26—41 x 35 jim (several sizes present!). anticlinal of Adaxial epidermis with subdivisions. Hypodermis 1—3 layers of bulging mucilaginous cells. Bundle sheaths lignified. Irregular crystalline bodies of unidentified Petiole with composition in hypodermis of OAHU 198. simple open vascular system (type 4). In midrib simple closed system (type 6) present in Van Balgooy 1689 only. HAWAII — TAHITI Van Material studied. (f. sandwicensiss): OAHU 198; Heller 2735. (f. taitensis): Balgooy 1689; Vesco 1847. Note. The mucilaginous hypodermis in I. anomala is the most striking feature of this species. A slightly similar but less well-developed hypodermis occurs in I. brasiliensis and tuerckheimii differ several other Stomatal vasculari- I. which, however, in respects. size, zation of and and outline of midrib petiole, midrib in transverse section vary strongly within this correlation species without, however, showing a with the two geographical forms. Anatomically (also according to the wood, Baas, 1973) this species seems rather isolated. Ilex aquifolium L., inch f. bacciflava (West.) Rehder, f. ferox (Ait.) Schneider, and f. heterophylla (Ait.) Loes. (subg. Euilex, series C, sect. Aquifolioides, subsect. Oxyodontae) — Plate III, 5 Adaxial cuticle abaxial cuticle Lamina 430—470/tm. 5 —8/mi; 7—10/im. Sparse unicellular hairs on midrib and Adaxial petiole. Stomata bicyclic, 34x33—36 x 36/tm. epidermis with anticlinal subdivisions. Hypodermis of 1 cell layer. Marginal sclerenchyma strand. Petiole with vascular simple open system (type 3—4). Material studied. MOROCCO: De Wilde & Dorgelo 2528. — NETHERLANDS ((f. bacciflava, f. ferox, and f. heterophylla): cult. hort. Leiden. 362 BLUMEA VOL. 22, No. 3, 1975 Note. I. aquifolium is anatomically very close to I. altaclarensis, I. bioritsensis, I. cornuta, I. dipyrena, I. integra, I. perado, and I. pernyi (see note under I. altaclarensis). In spite of the whole great diversity of external leaf morphology, the anatomical characters are on the very constant. Ilex archboldiana Merr. & Perry, J. Am. Arb. 22 (1941) 259 (? Subg. Euilex, series C, — Plate sect. Rugosae) IV, 5; fig. 13 Adaxial cuticle 6 abaxial cuticle Lamina 400/tm. —7 //m; 3—5 /im. Stomata cyclocytic and to anomocytic, occasionally anisocytic, 32 x 277ml. Adaxial epidermis with anticlinal cells. Bundle sheaths and of periclinal subdivisions, containing mucilage part spongy tissue Petiole with vascular lignified. simple open system (type 3). Material studied. NEW GUINEA. Mt. Giluwe: Schodde 1904. anatomical Note. This species does not show unusual features. Anatomy supports Merrill & Perry's (I.e.) statement that this species is close to I. versteeghii, with which it the features of thick leaves and shares many characters, e.g. xeromorphic partly lignified tissue and lignified bundle sheaths. spongy Cient. Ilex argentina Lillo, An. Soc. Argent. 72 (1910) 171 (subg. Euilex, series C, sect. subsect. — Plate Microdontae, Repandae) IV, 6; fig. 15 Lamina Adaxial cuticle abaxial cuticle Stomata 220//m. 3 /mi; 2[im. mainly cyclocytic, 28 X Abaxial with cells and occasionally bicyclic, 27/;m. epidermis lignified unspecialized strongly undulating anticlinal walls. Adaxial epidermis with anticlinal and periclinal sub- divisions, with large bulging mucilage cells. Bundle sheaths sometimes strongly bire- Petiole with vascular fringent. simple open system (type 4). Material studied. ARGENTINA. Tucuman: Vervoorst 6834. Note. The lignified abaxial epidermis, strongly sinuous anticlinal walls of the abaxial epidermis, and partly bicyclic stomata make this species rather outstanding. Lillo (I.e.) suggested a close affinity with I. paraguariensis from which it differs, however, in several leaf anatomical characters. Of the other species with a lignified epidermis, I. canadensis resembles I. argentina most closely. This may, however, be entirely due to parallel developments in rather unrelated species. Ilex asprella (Hook, et Arn.) Champ, (subg. Prinus, series B) Adaxial cuticle 2 abaxial cuticle 1 Lamina 1307ml. —3 /um; —27011. Stomataanomocytic cells with walls. Adaxial to cyclocytic, 37 x297ml. Epidermal strongly undulating epi- dermis with periclinal subdivisions, and containing bulging mucilage cells. Petiole with simple open vascular system (type 1). Material studied. PHILIPPINES: PNH 21584. Loesener's Note. This species agrees well with otherspecies studied in subgenus Prinus. also suggestion of a close affinity with I. longipes Chapm. from N. America is therefore supported by leafanatomy. Ilex bioritsensis Coll. sect. Hayata, J. Sc. Tokyo 30 (1911) 53 (subg. Euilex, series C, Aquifolioides, subsect. Oxyodontae) — Plate V, 6 P. Baas: Anatomy of Aquifoliaceae etc. 363 Adaxial cuticle abaxial cuticle Indumentum of Lamina 360/tm. 9/im; 671m. sparse unicellular hairs confined to petiole in cult. Leiden, abundant onadaxial side of midrib and lateral veins in Boom 13379. Stomata cyclocytic to bicyclic, 27x27711x1. Marginal scleren- strand. Petiole with vascular chyma simple open system (type 3). Material studied. NETHERLANDS: cult. Dedcmsvaart (=Herb. Boom 13379) and cult. hort. Leiden. that bioritsensis Note. Hayata's remark (I.e.) I. is very close to I. pernyi is fully sup- ported by the leaf anatomical characters. See also notes under I. aquifolium and I. alta- clarensis. I. bioritsensis is native in Formosa. Ilex boliviana Britton (subg. Euilex, series A, sect. Dasyneurae) Lamina 200 Adaxial cuticle c. abaxial cuticle 6 Sparse unicellular 71m. 2573m; —771m. hairs on petiole and midrib. Stomata cyclocytic, 26 x 26 «m. Abaxial epidermis with lignified unspecialized cells. Adaxial epidermis with straight walls. Cork warts. Bundle sheaths Petiole with vascular lignified. Many veins vertically transcurrent. simple open system (type 4). Material studied. BOLIVIA: Britton & Rusby 450. this I. Note. Leaf anatomically species resembles I. amara, martiniana, I. ovalifolia, cork I. retusa, and I. triflora which share the frequent warts and lignified epidermis. I. boliviana stands with out, however, by its extremely scleromorphic anatomy massive the bundles and lower sclerenchyma girders linking many of vascular with upper thick epidermis, and by a very cuticle. Ilex brasiliensis (Spreng) Loes. var. pubiflora (Reiss.) Loes. (subg. Euilex, series C, sect. — Plate and Megalae, subsect. Pedicellatae) I, I 2 ; fig. 4 Lamina Adaxial cuticle abaxial cuticle — slender hairs with 270 fim. 4/tm; 1 271m. Long thick basal abundant both leaf surfaces. Stomata septs on cyclocytic to anomocytic, 26x21 Adaxial with anticlinal subdivisions. of of fim. epidermis Hypodermis I layer cells. bulging, probably mucilaginous, Petiole with simple open vascular system (type 4). Material studied. BRAZIL : Lindeman & De Haas 3367 (U). the adaxial cell either of Note. It is difficult to classify subepidermal layer as part a The of multiple epidermis or as a 'true' hypodermis. alignment the bulging subepidermal cells is most suggestive of a multiple epidermis. The uniseriate hairs, with several to many of constitute distinc- septa which the basal sept is usually much thicker than the others, a this tive featureof species. Loesener's suggestion of an affinity with I. theezans agrees more less with or leaf anatomy. Ilex brassii & Arb. Merr. Perry, J. Am. 20 (1939) 334 (? subg. Byronia, series A) Lamina abaxial cuticle — Stomata 22071m. Adaxial cuticle 371m; 1 271m. anisocytic, Adaxial with anticlinal and 27 x2071m. epidermis periclinal subdivisions, containing cells. Bundle sheaths and of fusiform bulging mucilage part spongy tissue lignified. Some sclereids of midrib. Petiole with vascular in ground tissue simple open system (type 4). Material studied. NEW GUINEA : Brass 7686. Note. Merrill and Perry's remark (I.e.) that I. brassii resembles I. ledermannii is leaf both do show distinc- supported by anatomy, although species not very outstanding BLUMEA VOL. No. 364 22, 3, 1975 I. I. tive characters. Similar leaf histologies occur in e.g. I. cymosa, laurifolia, macrophylla, also similarities with the I. sclerophylloides, and I. wallichii from Malesia. There are New All these World species I. integerrima and I. sideroxyloides. species have predominantly and bundle sheaths. anisocytic stomata and mostly lignified spongy tissue Strongest the resemblance is with I. laurifolia which shares with I. brassii occurrence of occasional thick-walled fibres in the peripheral abaxial ground tissue of the midrib. I. laurifolia was, however, placed in subgenus Euilex by Loesener. This leaves the problem of where to place I. brassii unsettled. Ilex brevicuspis Reiss. (subg. Euilex, series C, sect. Microdontae, subsect. Eumicrodontae) Adaxial cuticle abaxial cuticle 1—2 irregularly Lamina 130/rm. 1—3 /im; fim. Sparse midrib. curved, unicellular hairs on Stomatacyclocytic, sometimes irregular, 22x20/mi. vascular Petiole with simple open system (type 4). Material studied. BRAZIL. Santa Catarina: Smith & Klein 11440. Note. This species does not show outstanding features. Through its mesomorphic be taxonomically leaves it resembles species from subgenus Prinus, but this may insignif- icant. series Indico- Ilex brunnea Merr., Philip. J. Sc. Bot. 10 (1915) 318 (subg. Euilex, D, sect. Malaicae) Adaxial cuticle 6 abaxial cuticle Stomata mainly anisocytic Lamina 300 fim. /cm; 9 fim. Adaxial with subdivisions and also tending to cyclocytic, X 21 epidermis periclinal o t j „ 32 rfim. cells. Petiole with vascular 4). bulging mucilage simple open system (type Material studied. PHILIPPINES: Am. Arbor. 34404 (=Edaho s.n). brunnea Note. None of the leaf anatomical characters of I. are very outstanding. Merrill's placement (I.e.) in section Indico-Malaicae cannot be supported or challenged leaf there resemblance with I. cissoidea and I. using anatomy, though is some stapfiana, sections also from this section. However, other species from different or even subgenera I. I. I. I. I. I. sebertii, and I. theezans like cymosa, krugiana, laurifolia, macrophylla, rotunda, similar in leaf with I. brunnea. show or stronger resemblance anatomy Ilex buergeri Miq. f. rolfei (Elm.) Loes. (subg. Euilex, series C, sect. Microdontae, subsect. — Plate Repandae) V, 1—3 Adaxial cuticle 6 abaxial cuticle Sparse unicellular hairs Lamina 190/im. /im\ 4—5/im. Bundle on midrib and petiole. Stomata cyclocytic, occasionally bicyclic, 28X28/tm. sheaths partly fibrous and lignified. Sclereids abundant throughout mesophyll. Petiole with vascular simple open system (type 2). Material studied. PHILIPPINES: BS 40633. Note. The abundant fusiform sclereids in the leaf mesophyll constitute the most with striking feature of this species. Similarities other species showing sclerenchyma Of idioblasts (I. curranii, I. ficoidea, I. formosana, I. glomerata, and I. mitis) are very striking. these species, Loesener included I. ficoidea, I. formosana, and I. glomerata in the same sub- from other section. There also similarities with section. I. mitis is, however, an are some and species lacking sclereids, notably I. dipyrena, I. fargesii, I. goshiensis, I. integra, I. P. Baas: Anatomy of Aqttifoliaceae etc. 365 I. the tolucana. Of these I. goshiensis and I. tolucana are in the same section as buergeri, from sections but all from series C. Aquifolium of subgenus others are different are yet Euilex. Ilex canariensis Poir. (subg. Euilex, series A, sect. Cassinoides) Adaxial cuticle abaxial cuticle 6 Stomata cyclocytic, Lamina 320fim. 4 —5 /4m; fim. Abaxial anticlinal walls of epidermal cells strongly undulating. 39 x 35 /-«ni, unspecialized Adaxial epidermis with lignified unspecialized cells in Bornmueller 414 only. epidermis vascular with periclinal subdivisions and bulging mucilage cells. Petiole with simple open system (type 4). 681 — TENERIFFE: — PALMA: Cool & Den Tex Material studied. MADEIRA: Bornmueller 414. (sections only). Pitard s.n. (26—1—1905). Bornmueller of the Note. In some of the lamina of 414 _a portion unspecialized partsr __ _ Cool and Pitard once more epidermal cells is lignified, this is not so in & Den Tex 681 s.n., the exemplifying the absence of absolutely diagnostic anatomical characters in majority resemble each other closely. of Ilex species. In other respects the three specimens very resemblance with and cassine. The is Anatomically there is some I. argentina I. species very also in distinct from I. perado in both leaf and wood anatomy (see Baas, 1973: 228), spite Thevenard multicellular hairs and small of the same biotope they inhabit. (1906) reported find cork warts for I. canariensis, but I did not these. Ilex caroliniana(Lam.) Loes.; I. vomitoria Ait. (subg. Euilex, series C, sect. Microdontae, subsect. Vomitoriae) hairs 260 Adaxial cuticle abaxial cuticle I—2 Unicellular on Lamina fim.. 3 —5 /im; /xm. mainly cyclocytic, lamina and petiole fairly abundant adaxially, sparse abaxially. Stomata and occasionally anisocytic, Adaxial epidermis with periclinal subdivisions t t # 25x23 fim. (type 2). bulging mucilage cells. Cork warts. Petiole with simple open vascular system U.S.A. Material studied. Oregon: Suhsdorfs.n. (4—1909)- rather small cork there Note. Except for the numerous warts, are no outstanding leaf anatomical characters in this species. I. coriacea, I. divaricata, I. glabra, and I. pedunculosa share with I. divaricata is of the section. many characters I. caroliniana. Of these only same and Ilex cassine L. inch var. angustifolia Ait., var. myrtifolia (Walt.) Chapman (subg. Euilex, series A, sect. Cassinoides) — Plate III, I and 2 Lamina Adaxial cuticle abaxial cuticle — Fairly long, occasionally 400_ 671m; 3_ 4 fim. r fim. r ti on midrib and wholeabaxial surface. Stomata septate hairs abundant petiole, sparse over Anticlinal walls of cells mainly cyclocytic, occasionally anisocytic, 31 x 26 fim. epidermal , / , , r strongly undulating. Abaxial epidermis with lignified unspecialized cells. Subsidiary cells rarely lignified (Plate III, 2). Adaxial epidermis with anticlinal and periclinal subdivisions, vascular 4). and bulging mucilage cells. Petiole with simple open system (type Schallert Material studied. U.S.A. cult. Florida: Boom 38139 (var. angustifolia); N. Carolina: s.tt. (var. myrtifolia, maceration only). section have abaxial Note. I. canariensis p.p. and I. opaca of the same also a lignified leaf of these two resembles I. cassine most closely in overall epidermis; species I. opaca 366 BLUMEA VOL. 22 No. 3, 1975 Thevenard recorded for this but could anatomy. (1906) solitary crystals species, I only find clustered ones. Ilex celebensis Capit., Bull. Soc. Bot. France 57 (1910) 236 (subg. Byronia, series A) 180 Adaxial cuticle abaxial cuticle unicellular hairs Lamina fim. 6—7jum; 3 /mi. Sparse midrib and Stomata occasionally bicyclic, X on petiole. mainly cyclocytic, 25 23 /mi. bire- Bundle sheaths partly fibrous and lignified. Part of spongy tissue rather strongly fringent but probably not lignified. Minute crystals in adaxial epidermis. Petiole with simple open vascular system (type 2). Material studied. CELEBES: Teysmann 14119. Note. Capitaine's suggestion (I.e.) of affinities with both I. sandwicensis (= I. anomala) and I. is leaf because of in cymosa not supported by anatomy differences hypodermal and stomatal alternative development type respectively. It is, however, impossible to give suggestions on the grounds of leaf anatomy. Ilex chinensis Sims; I. purpurea Hassk. (subg. Euilex, series A, sect. Excelsae, subsect. Umbelliformes) Adaxial and abaxial cuticle unicellular hairs midrib. Lamina 240/im. 2/im. Sparse on Adaxial Stomatacyclocytic, rarely bicyclic, 22x 21/im. epidermis containing somebulging In mucilage cells. Petiole with simple open (type 4) to almost closed vascular system. midrib completely closed (type 6). CHINA. Fan & Li Material studied. Hunan: 549. Note. Loesener's suggestion of an affinity with I. pedunculosa is not fully supported leaf because of and several other characters. Other affinities by anatomy petiole anatomy are hard to indicate, but affinities with I. kwantungensis, suggested by Hu (1950) can be supported. Ilex cissoidea sect. — Plate and Loes. (subg. Euilex, series D, Indico-Malaicae) I, 3 4 fig. 6 Adaxial cuticle abaxial cuticle Broad-based hairs abundant Lamina 1007ml. 3 fim; 2[im. hairs wall on both surfaces; frequently with anticlinal division in basal part (fig. 6). Stomata anisocytic to 26x21 Adaxial with anticlinal subdivi- cyclocytic, /mi. epidermis Bundle sheaths and rather Petiolewith sions. mesophyll strongly biréfringent. simple open vascular system (type 4). Material studied. BORNEO : Endert 1732. Note. The indumentum of broad-based hairs is the only outstanding feature in the leaf this in anatomy of species. Similar hairs occur I. oppositifolia and I. revoluta, also from These of leafanatomical Borneo. species differ, however, in a considerable number other features from I. cissoidea. The other species of section Indico-Malaicae studied (I. spicata, I. clemensiae, and I. stapfiana) are also quite different from I. cissoidea. I. brunnea treated in this Merrill number the section by shares, however, quite a of characters, but on the whole other leaf anatomical evidence is inconclusive with regard to relationships with species. Ilex clemensiae Heine, Mitt. Bot. Staatssaml. MUnchen I, 6 (1953) 208 (subg. Euilex, — Plate scries D, sect. Indico-Malaicae) II, 4 P. Baas: Anatomy of Aquifoliaceae etc. 367 Adaxial and abaxial cuticle unicellular hairs on Lamina 220/an. 9—12/rm. Sparse midrib and petiole. Stomata anisocytic to cyclocytic, frequently tending to complex, Adaxial with subdivisions and mucilage cells. 38 X 29/mi. epidermis infrequent periclinal Petiole with vascular Cork warts. Bundle sheaths lignified. interrupted open system (type 5). Material studied. N. BORNEO. Mt. Kinabalu: Clemens 30259. cork and vascular in petiole and Note. Frequent warts an abaxially open system of which in section Indico- midrib are the more striking characters this species was put with leaf Malaicaeby Heine(I.e.). However, within this section no species agreat anatomi- clemensiae other candidates for close cal resemblance to I. was encountered, nor can affinity be indicated on the basis of leafanatomical characters. Ilex confertiflora Merr., Lingnan Sc. J. 13 (1934) 35 (subg. Euilex, series C, sect. Micro- dontae, subsect. Repandae) to Lamina Adaxial and abaxial cuticle — _ Stomataanisocytic cyclocytic, 300fim. 5_ 6/um. x tissue partly strongly biréfringent. Few irregu- occasionally complex, 23 23 fim. Spongy from of lar sclereids associated with but tending to be detached sclerenchyma caps veins. with vascular Very small marginal sclerenchyma strand, occasionally associated elements. vascular Petiole with simple open system (type 3). Material studied. CHINA. Kwantung: Tsang 20347. Note. Merrill's suggestion (I.e.) of affinities with I. ficoidea, as supported by Hu evidence because of the of (1950), are in agreement with the leaf anatomical presence link with foliar sclereids. I. confertiflora at the same time provides a leaf anatomical sub- of the of marginal sclerenchyma. section Oxyodontae section Aquifolioides through presence Ilex coriacea (Pursh) Chapm. (subg. Euilex, series A, sect. Cassinoides) I — Plate IV, 4 unicellular hairs Lamina 260/mi. Adaxial cuticle 16/mi; abaxial cuticle 12/mi. Sparse Stomata X 26 Adaxial on midrib and petiole. cyclocytic to anomocytic, 27 /mi. epidermis with anticlinal and periclinal subdivisions, containing idioblastic bulging mucilage cells. Petiole with 'double' vascular Cork warts of characteristic shape. system (type 10). Material studied. U.S.A. Georgia: Beyrick 926. Note. Affinities with I. canadensis, I. cassine, and I. glabra as suggested by Loesener shares the vascularization are supported by leaf anatomy only for I. glabra, which petiole cork Cork similar those type and frequent (though slightly different) warts. warts to in in minor I. coriacea occur in I. macfadyenii from Jamaica, which differs, however, several leaf anatomical characters. almost Thevenard (1906) reported an continuous hypodermis for I. coriacea (under the & subdivisions synonym I. lucida Torr Gray), but in Beyrick 926 there are only sporadic with mucilage cells in the epidermis. Ilex corauta Lindl. & Paxt. 'Burfordii' (subg. Euilex, series C, sect. Aquifolioides, subsect. Oxyodontae) Adaxial cuticle abaxial cuticle — unicellular Lamina 400/rm. 12/tm; 5 6/um. Sparse and and hairs on midrib petiole. Stomata cyclocytic bicyclic, 28x27/mi. Few irregular sclereids associated with, but tending to be detached from sclerenchyma caps of veins. Petiole with vascular Marginal sclerenchyma strand. simple open system (type 3). VOL. No. 368 BLUMEA 22, 3, 1975 Material studied. EGYPT. Cairo, cult. Wight nurseries: Boom 38771. one I. altaclarensis, I. aquifolium, I. Note. This species clearly belongs to group with from bioritsensis, I. dipyrena, I. integra, I. perado, and I. pernyi, all subsection Oxyodontae. from There are also similarities with I. fargesii section Lemurensis, and with species having foliar sclereids. Ilex crenata Thunb. (subg. Euilex, series B, sect. Polyphyllae) hairs Adaxial cuticle 8 abaxial cuticle Sparse unicellular on Lamina 230/im. //m; 7//m. Adaxial with midrib and petiole. Stomata cyclocytic, 27 x277ml. epidermis periclinal subdivisions and bulging mucilage cells. Cork warts. Petiole with simple open vascular system (type 3). Material studied. NETHERLANDS : cult. hort. Leiden; see also below. others studied Note. Besides the cultivated specimen cited above, 10 were (of 9 specimens macerations only), because one deviating specimen of I. crenata var. luzonica cells in Loes. (Moluccas, Ceram, Eyma 2044) showed lignified epidermal and also differed _ leaf characters. the material epidermal cells a number of other In remaining no lignified which indeed casts doubt on the identification ofEyma All were found, 2044. specimens _ The were constant for the occurrence of cork warts. stomatal type ranged from almost exclusively cyclocytic, to anomo- to cyclocytic, or aniso- to cyclocytic. Remarkably section enough the aberrant Eyma collection resembles the other species studied of abaxial Polyphyllae: I. dumosa from Brazil, mainly through the shared lignified cells in the epidermis. series C, sect. Microdontae, Ilex curranii Merr., Philip. J. Sc. 17 (1920) 273 (subg. Euilex, subsect. Repandae) Adaxial cuticle abaxial cuticle unicellular hairs on Lamina 230/1111. 6/im; 4/um. Sparse midrib and petiole. Stomata cyclocytic, occasionally bicyclic, 38 x 28/mi. Bundle sheaths partly fibrous and lignified. Sclereids fusiform and moderately branched, partly associated with of veins, abundant in adaxial of Petiole with sclerenchyma caps part mesophyll. simple open vascular system (type 3). Material studied. PHILIPPINES. Luzon: PNH 7364. but Note. Merrill (I.e.) originally placed I. curranii in section Rugosae, Loesener (Tn,oA ricrbtlv treated it member ofsubsection where as a Repandae, this species obviously ' r \-^*l —/ --O / * * foliar other features affinity. belongs with its abundant sclereids and many witnessing I. Loes. series Ilex cymosa Bl.; pleiobrachiata (subg. Byronia, A) Indumentum Adaxial cuticle abaxial cuticle 1 Lamina 150—270fim. 1—5,«m; —\fim. unicellular hairs whole abaxial and absent or of sparse to fairly numerous over surface, but a variable on adaxial side of midrib and on petiole. Stomata mainly anisocytic, pro- X to Adaxial epidermis with or without portion tending to cyclocytic, 21 18 29x23 //m. cells. periclinal subdivisions, usually but not in all specimens with mucilage Bundle of in Kep. Field No. 71108. sheaths lignified. Part spongy tissue usually lignified, except Petiole with vascular (type 4), but with complex closed usually simple open system system in Wirawan 194 (type 8). No. Material studied. MALAYA: Kep. Field No. 71108; Sinclair 10813. ■— SINGAPORE: Sing. Field 38887 — — SUMATRA: Dumas — SAN Wirawan 1630. (coll. Sinclair). BORNEO: Kostermans 9235; 48318. JAVA: 194. P. Baas: Anatomy of Aquifoliaceae etc. 369 Note. The variation encountered in these seven specimens is very large. Hardly any of indicate of the characters is constant. Because this strong variability it is possible to high degree to I. but the distribution many species with a of similarity cymosa specimens, different Loesener is random, of those species over the subgenera recognized by virtually This is what would suggesting that these similarities have no taxonomic significance. one of the leaf anatomical characters in I. expect, since none predominantly occurring cymosa Within there similarities with I. are unusual within Ilex. subgenus Byronia, are good ledermannii, I. macrophylla, I. sclerophylloides, and I. wallichii. Steenis. The leaf The identity of the material studied was kindly checked by Prof. Van with of to anatomical variation is not correlated the grouping specimens as belonging Loesener's delimitation. The most I. cymosa or I. pleiobrachiata according to specific striking example of variability is offered by the vascularization of the petiole, which is in of a complex type one specimen only. (1955) (subg. Euilex, series Ilex danielis Killip & Cuatrec., Trop. Woods 101 14 C, sect. Micranthae, subsect. Epunctatae) unicellularhairs 600 Adaxial cuticle abaxial cuticle on Lamina fim. 14/tm; 4 /mi. sparse adaxial side ofmidrib and major veins. Stomata cyclocytic, sometimes tending to anomo- Adaxial epidermis with occasionalpericlinal subdivisions and idioblastic cytic, 25 x 22/jm. with 4). bulging mucilage cells. Petiole simple open vascular system (type Material studied. COLOMBIA : Hatheway 1602. shows features in and the anatomical evidence Note. This species no outstanding Ilex, is therefore rather neutral with regard to thesystematic placement adopted here, following Killip and Cuatrecasas' (I.e.) suggestion of close affinities with I. guianensis. From the latter of subsection in species and other species Epunctatae studied, I. danielis differs, however, stomatal type. Ilex decidua Walt. (subg. Prinus, series B) 2-cellular Adaxial cuticle 2 abaxial cuticle 1—2 Sparse Lamina 140/tm. —3 / Heller Material studied. U.S.A. Texas: Heller & 2414. does show Note. As most species of subgenus Prinus, this mesomorphic species not outstanding anatomical features. Ilex dipyrena Wall. (subg. Euilex, series C, sect. Aquifolioides, subsect. Oxyodontae) Adaxial cuticle abaxial cuticle —8 unicellular Lamina 8—9 /im; 7 //m. Sparse hairs Stomata and Epidermal cells on petiole and midrib. cyclocytic bicyclic, 31 X29/tm. with straight anticlinal walls. Marginal sclerenchyma strand. Petiole with simple open vascular system (type 3). Material studied. HIMALAYA: Heybroek 1 75. Note. With its partly bicyclic stomata and well-developed marginal sclerenchyma 370 BLUMEA VOL. 22, No. 3, 1975 strand, I. dipyrena clearly belongs to the same group of species as I. aquifolium, I. bioritsensis, I. cornuta, I. integra, I. perado, and I. pernyi. divaricata Ilex Mart. (subg. Euilex, series C, sect. Microdontae, subsect. Sideroxyloides) Adaxial cuticle abaxial cuticle unicellular hairs Lamina 300/Mn. 9 /mi; 4—5 fim. Sparse and midrib. on petiole Stomata mainly cyclocytic, occasionally tending to anisocytic or X cells with anticlinal walls. Adaxial anomocytic, 29 23 /cm. Epidermal straight epidermis with periclinal division walls and bulging mucilage cells. Cork warts. Petiolewith simple vascular open system (type 4). Material studied. VENEZUELA: Maguire, Rowan & Wurdack 30735 (U). Note. According to Loesener this species is close to I. sideroxyloides. This is not leaf because of differences in stomatal and supported by anatomy, mainly type occurrence of regular cork warts. Similarities with other species (e.g. I. coriacea, I. danielis, I. theezans, I. based and do and tuerckheimii) are not on any unusual characters therefore not provide evidence of mutual strong affinity. Ilex dubia Trel. mollis (Don) var. (Gray) Loes.; I. montana Torr. & Gray (subg. Prinus, series B) Adaxial and abaxial cuticle Indumentum of Lamina no/ U.S.A. N. Material studied. Carolina: Stauffer c.s. 5958. Note. As in most other species of subgenus Prinus the leaves are mesomorphic and without outstanding anatomical features. Ilex dumosa Reiss. (subg. Euilex, series B, sect. Polyphyllae) Adaxial cuticle abaxial cuticle unicellular hairs Lamina 190/mi. 6 /mi; 3 fim. Sparse on anomocytic to 28 x Abaxial with petiole. Stomata cyclocytic, 23 fim. epidermis lignified cells. Bundle sheaths unspecialized strongly biréfringent. Petiole with simple open vascular system (type 3). Material studied. BRAZIL. Parana: Dustn 11076. Note. Through its lignified abaxial epidermis, this species resembles the aberrant of var. luzonica. Similarities also with and specimen I. crenata are present I. amara I. boliviano of different sections and series. Franch. Ilex fargesii (subg. Euilex, series C, sect. Lemurenses) Lamina Adaxial cuticle abaxial cuticle 280fim. 6fim; 4—5 fim. Stomata cyclocytic, occasionally bicyclic, Petiole with simple vascular i in 29x28 fim. open system (type basal in distal part; type 3 part). Material studied. NETHERLANDS, cult. Wageningen: Boom 30300. this lacks foliar leaf anatomical Note. Although species sclereids, there are good similarities with I. mitis, which Loesener regarded to be a related species. P. Baas: Anatomy ofAquifoliaceae etc. 371 flex ficoidea Hemsl. (subg. Euilex, series C, sect. Microdontae, subsect. Repandae) & Adaxial cuticle — abaxial cuticle 2 Indumentum Lamina 210 230/rm. 4 6/ x Sclereids fusiform and bicyclic, 29x27 & 31 30/im. moderately branched, partly of abundant in adaxial of associated with sclerenchyma caps veins, part mesophyll. Petiole with vascular simple open system (type 3). Material studied. CHINA. Kwantung: Tsang 20343, 20824. Loesener's and Hu's of close with Note. (1950) suggestions a affinity I. buergeri are leaf and fully supported by anatomy. I. cornuta, I. curranii, I. formosana, I. glomerata, I. also similar to ficoidea the of, mitis are very I. through common occurrence amongst and lack feature but others, foliar sclereids. I. goshiensis I. integra this are similar in many other respects. flex formosana Maxim, (subg. Euilex, series C, sect. Microdontae, subsect. Repandae) Adaxial cuticle — abaxial cuticle unicellular Lamina 250/im. 5 6/im; 3—4/tm. Sparse hairs midrib. Stomata Adaxial on petiole and cyclocytic, tending to bicyclic, 31 X 2971m. subdivisions. epidermis with periclinal Sclereids fusiform or moderately branched, partly associated with of abundant in adaxial of mesophyll. sclerenchyma caps veins, part of fusiform sclereids leaf Petiole with vascular Concentration in margin. simple open system (type 3). Material studied. TAIWAN: Taiwan Herb. 11132. this has Note. Almost identical as it is to I. ficoidea, species the same anatomical affinities. affinities I. thus Hu's (1950) remarks about close with ficoidea are fully supported her of I. and I. in the series. by leaf anatomy, as well as treatment glomerata buergeri same flex glabra (L.) Gray (subg. Euilex, series A, sect. Cassinoides) Adaxial cuticle abaxial cuticle unicellular Lamina 310/tm. 3 —5 junv, 6—7/tm. Sparse hairs and midrib. X Adaxial with on petiole Stomata cyclocytic, 24 23 /um. epidermis infrequent anticlinal subdivisions. Cork warts. Petiole with 'double' vascular system (type 10). Material studied. U.S.A. New Yersey: Lawrence & Dress 269. of vascular the rather unusual for Ilex Note. The arrangement tissue in petiole is and other is shared by I. coriacea. This, together with some resemblances, strongly supports Loesener's suggestion of close affinities between the two species. Other overall similarities based character with some species are not on special complexes, and are probably not relevant for judging natural affinities. Thevenard (1906) reported a locally mucilaginous of could epidermis for this species, but no trace this be found in Lawrence & Dress 269. flex glomerata King (subg. Euilex, series C, sect. Microdontae, subsect. Repandae) Adaxial cuticle abaxial cuticle and Lamina 180/tm. 2 —3 /um; 2/um. Stomata cyclocytic cells with anticlinal walls. Sclereids bicyclic, 32 x 29/im. Epidermal strongly undulating fusiform and moderately branched, partly associated with sclerenchyma caps of veins, abundant adaxial of Petiole with vascular in part mesophyll. simple open system (type 3). Material studied. BORNEO. Mt. Kinabalu: Clemens 28843. BLUMEA VOL. 22, No. 3, 1975 372 indicate affinities with I. Note. Bicyclic stomata and fusiform foliar sclereids buergeri, and I. mitis. I. I. I. I. I. curranii, I. ficoidea, I. formosana, fargesii, integra, bioritsensis, dipyrena, also share characters with I. and I. latifolia many glomerata. seriesC, sect. Microdontae, Ilex goshiensis Hayata,Mat. Fl. Formosa (1911) 54 (subg. Euilex, subsect. Sideroxyloides) hairs l Adaxial cuticle abaxial cuticle Sparse unicellular amina 310/zm. 7 /xm; 3—4/an. midriband Stomata and 28 x with fairly broad bases on petiole. cyclocytic bicyclic, 25 /tm. Adaxial epidermis with infrequent anticlinal subdivisions. Solitary crystals present in midrib. Petiole with simple open vascular system (type i). Moran Material studied. JAPAN. Ryukyu Islands, Okinawa: 5131. follows s Note. The treatment of I. goshiensis in subsection Sideroxyloides Hayata There (I.e.) suggestions of affinities with I. championii of that subsection. are, however, celebensis of several anatomical differences between I. sideroxyloides and I. divaricata. I. leaf but since no subgenus Byronia resembles I. goshiensis anatomically in many respects, taken of close mutual special features are involved this cannot be as proof affinity. Mich. (subg. series A) Ilex grandifolia Merr., Pap. Ac. Sc. 19 (1933) 163 Byronia, Adaxial and abaxial cuticle Stomatacyclocytic to anomocytic, Lamina400» um. 3—4/rm. Adaxial with anticlinal subdivisions and bulging mucilage cells. 32 x 26 /xm. epidermis Petiole with complex vascular Bundle sheaths and spongy tissue lignified. open system (type 9). Material studied. SUMATRA: De Voogd 1134. I. Note. According to Merrill (I.e.) this species belongs in the same group as cymosa, hence its in and is probably most closely relatedto I. macrophylla Wall.; treatment subgenus Byronia here. The main difference between these species and I. grandifolia is the stomatal other characters roughly similar. The complex vascularization of the petiole type, being the cuticular striations of the abaxial recall I. cymosa p.p. and very conspicuous epidermis There also similarities with and I. hypoglauca from the same subgenus. are some species from section Indico-Malaicae of subgenus Euilex, series D. Standi, (subg. Ilex guianensis (Aubl.) O. Ktze, inch var. arimensis Loes.; I. panamensis Euilex, series C, sect. Micranthae, subsect. Epunctatae) — Fig. 16 Stomata Adaxial cuticle abaxial cuticle 1 Lamina 240—300fxm. 3—5 /im; —3 fxm. predominantly anisocytic, sometimes cyclocytic in Broadway 4052, predominantly Allen in cyclocytic, sometimes anomocytic or anisocytic in 3570, cyclocytic to anisocytic Lindeman Adaxial epidermis with periclinal Lanjouw & 1495, 22x19—26x21/101. vascular (type 4). subdivisions and bulging mucilage cells. Petiole with simple open system — — SURINAM: & TRINIDAD: Material studied. PANAMA: Allen 3370 (U). Lanjouw Lindeman 1495 (U). Broadway 4052. with Note. Leafanatomically this species is not very unusual, and bears resemblance of Ilex of different sections Of all these I. a great number species or even subgenera. of the subsection is closest leaf I. inundata and I. umbellata are jenmanii same anatomically. rather also quite similar, suggesting that subsection Epunctatae constitutes an anatomically homogeneous assemblage. P. Baas: Anatomy of Aquifoliaceae etc. 373 Ilex hanceana Maxim, (subg. Euilex, series B, sect. Vacciniifoliae) Adaxial cuticle 8 abaxial cuticle unicellular Lamina 300/cm. —9/cm; 3—4/cm. Sparse hairs midrib and petiole. Stomata cyclocytic, occasionally X on anisocytic, 25 25 /im. Adaxial epidermis with occasional anticlinal subdivisions. Solitary crystals in ground tissue of petiole and midrib. Petiole with simple open (type 2) to almost closed vascular midrib with closed system; in completely system (type 6). Material studied. JAPAN. Iriomote I.: Gressitt 600 (U). Note. The solitary crystals in addition to clustered ones constitute the only distinctive Affinities other clear. character of this species. with species studied are not Ilex hypoglauca (Miq.) Loes. (subg. Byronia, series A) — Plate I, 6; III, 6; fig. 8 _ ... _ Adaxial cuticle — abaxial cuticle Indumentum Lamina 210—400//IT1. 4 15 /um; 3—7/tm. withnumerous basal both surfaces ofbroad-based hairs septs present on (Plate I, 6; fig. 8). Anderson Hallier and Havilatid Stomata mainly cyclocytic (. 8536, B 2875, 2872) or cyclocytic — Adaxial with to anisocytic (other 2 specimens), 26x22 31 X 27/im. epidermis periclinal subdivisions and cells Bundle sheaths and bulging mucilage in varying frequencies. part of tissue Petiole with closed vascular spongy lignified. usually complex system (type 8), but with simple open system (type 4) in Kostermans 4704. BORNEO: B Kostermans — SUMATRA. Material studied. Anderson 8536, Hallier 2875, Haviland 2872, 4704. Banka: isotype ('P.P. 2t\ ?HorsJield). Peculiar hairs with broad basal subdivided thin characterize Note. a part, by septs this resemble species. I. zygophylla has an indumentum of hairs which slightly those of but the hairs of that and I. hypoglauca, septs occur only sporadically in species, are always hair if The the vascular in few per present. variability of complex the petiole in I. hypo- glauca is remarkable. Besides with I. zygophylla there are anatomical affinities with several Malesian species of subgenus Byronia, and also with several other Malesian species of subgenus Euilex: I. brunnea, I. versteeghii, I. archboldiana, I. laurifolia, and also with I. share features oppositifolia of subgenus Rybonia. All these species some scleromorphic (thick cuticle, lignified bundle sheaths and partly lignified spongy tissue). Ilex insignis Hook. /. (subg. Euilex, series C, sect. Aquifolioides, subsect. Insignis) Adaxial cuticle abaxial cuticle Stomata Lamina 380/tm. 7—8/cm; 4 —6/tm. cyclocytic, Adaxial with partly bicyclic, 27 x 28/mi. epidermis infrequent periclinal subdivisions, containing mucilage ceils. Bundle sheaths strongly birefringent. Tendency for marginal sclerenchyma strand through reduction of vascular tissue in marginal vein, or through off thin of splitting sclerenchyma strand from massive sclerenchyma cap this marginal Petiole with vascular vein. simple open system (type 4). Sikkim: Material studied. INDIA. J. D. Hooker s.n. (Herb. Ind. Or. Hook./. & Thomson). o ofthe same subsection the of N t e. As in I. latifolia appearance marginal sclerenchyma from subsection is variable in this species. Affinities are also with species Oxyodontae. Ilex series integerrima (Veil.) Reiss. var. ebenacea (Reiss.) Loes. (subg. Euilex, C, subsect. — Plate sect. Megalae, Pedicellatae) II, 3 Adaxial cuticle —8 abaxial cuticle I Stomata Lamina 320/cm. 7 //m; —2/cm. mainly cyclocytic, 26 X 22 Adaxial with anticlinal and anisocytic, occasionally /mi. epidermis BLUMEA VOL. No. 374 22, 3, 1975 periclinal subdivisions, probably containing mucilage cells. Petiole with simple open vascular system (type 4). Material studied. BRAZIL. Parana: Hatschbach 7550. Note. Close affinities with I. theezans as advocated by Loesener, are supported by these anatomical resemblances in most characters, though are not very distinctive, so that affinities with othertaxa could also be advocated. I. brasiliensis of the same subsection differs in a number of leaf anatomical features. Ilex integra Thunb. (subg. Euilex, series C, sect. Aquifolioides, subsect. Oxyodontae) Adaxial cuticle abaxial cuticle Stomata cyclocytic and Lamina 330/cm. 7/cm; 5/cm. for strand reduction of bicyclic, 33x29/cm. Tendency marginal sclerenchyma through vascular tissue in marginal vein, and through lateral extension of sclerenchyma cap of marginal vein. Petiole with simple open vascular system (type 3). Material studied. JAPAN. Hondo: N.S.M. 527. resembles those of Note. This species closely the same subsection (.I. aquifolium etc.). of subsection foliar sclereids and There are also similarities with species Repandae having with species from subsection Insignis. Ilex inundataPoepp. (subg. Euilex, series C, sect. Micranthae, subsect. Epunctatae) Adaxial cuticle abaxial cuticle unicellular hairs Lamina200/cm. 3—4/cm; 1 /cm. Sparse Adaxial with anticlinal on midrib. Stomata anisocytic to cyclocytic, 23 x 20/cm. epidermis ceils. and periclinal subdivisions, containing mucilage Petiole with simple open vascular system (type 4). Material studied. BRAZIL: Kappler 182. Note. According to Loesener this species is close to I. umbellata. This can be supported leaf anatomical by a good overall leaf anatomical similarity. None of the characters in affinities be I. inundata is, however, very distinctive, so that with many other species could supported. Ilex jenmanii Loes. (subg. Euilex, series C, sect. Micranthae, subsect. Epunctatae) Adaxial cuticle —6 abaxial cuticle Stomata to Lamina 400 /cm. 4 /cm. 3 /cm. anisocytic Anticlinal walls of undulated. Abaxial and cyclocytic, 27x23/cm. epidermis strongly adaxial epidermis containing bulging mucilage cells. Spongy tissue partly birefringent, Petiole with vascular probably unlignified. simple open system (type 4). Material studied. SURINAM: Lindeman 704 (U). of the subsection without N o t e. As the previous species same I. jenmanii is outstanding several characters and, although similar to species of this subsection, similarities with other well. miscellaneous species are present as Ilex kleinii Edwin in Reitz, Fl. ilustr. Catar. I, Aquifol. (1967) 41 (not assigned to a subgenus or section) Lamina Adaxial cuticle abaxial cuticle 1—2 Stomata mainly nojum. 3 /cm; /cm. Adaxial with cyclocytic, occasionally anomocytic or anisocytic, 22x19/cm. epidermis infrequent anticlinal subdivisions. Petiole with simple open vascular system (type 4). P. Baas: Anatomy of Aquifotiaceae etc. 375 Material studied. BRAZIL. Santa Catarina: Reitz & Klein 3966 (U). to afford a Note. There are no distinctive anatomical characters in this species sugges- tion about systematic affinities. Ilex krugiana Loes. (subg. Euilex, series C, sect. Microdontae, subsect. Eumicrodontae) Adaxial cuticle abaxial cuticle Stomata cyclocytic, Lamina i8o//m. 4—6/tm; 3 //m. Adaxial with frequent anticlinal occasionally bicyclic or complex, 30x25 //m. epidermis with and periclinal subdivisions, containing mucilage cells. Petiole simple open vascular system (type 4). Material studied. BAHAMAS: Eggers 4462. features, the leaf microstructure N o t e. In the absence ofoutstanding leaf anatomical of with from which is neutral with respect to Loesener's suggestion an affinity I. brevicuspis, few anatomical I. krugiana differs in a minor aspects. series Ilex kwantungensis Merr., J. Arn. Arb. 8 (1927) 8 (subg. Euilex, A, sect. Excelsae) Adaxial cuticle abaxial cuticle Indumentum of rather Lamina 190fim. 3—5 /tin; 2/Lin. sides of lamina. long unicellular hairs with thick walls and recurved tips present on both x 22 Adaxial epidermis with periclinal subdivisions, containing Stomata cyclocytic, 24 [im. closed (type 6) mucilage cells. Petiole with simple open vascular system (type 4); simple in midrib. Fan & Li Material studied. CHINA. Hunan: 259. Note. Merrill (I.e.) doubtfully referred I. kwantungensis to section Excelsae. Hu (1950) chinensis her which contains treated the species together with I. in series Chinensis, many known I. and I. lonicerifolia species not to Loesener, except I. chinensis (syn. purpurea) section Excelsae. To which Loesener (1942) both referred to subsection Umbelliformes of this the therefore treat I. in section, which conform to Loesener system I kwantungensis finds leaf chinensis and I. kwantungensis are leaf anatomically good support in anatomy; I. alike, particularly in the midrib vascularization. Ilex latifoliaThunb. (subg. Euilex, series C, sect. Aquifolioides, subsect. Insignis) 600 Adaxial cuticle —6 abaxial cuticle Stomata cyclocytic Lamina fim.. 4 //m; 3—\fim. and cells with straight anticlinal walls. Tendency for bicyclic, 37X34/tm. Epidermal marginal sclerenchyma strand through reduction of vascular tissue in marginal vein, or of of the latter. Petiole with simple vascular lateral extension sclerenchyma cap open intermediate between and 4). system (± type 3 Material studied. FRANCE, cult. Orleans: Boom 15484. subsection (I. aquifolium, Note. This species is anatomically very close to Oxyodontae the I. bioritsensis, I. cornuta, I. dipyrena, I. integra, I. perado, and I. pernyi) although marginal vascular Similarities with sclerenchyma is still associated with a bundle in I. latifolia. of also well with I. and I. glomerata I. insignis the same subsection are present, as as fargesii of other subsections. Thévenard(1906) recorded short hairs for this species, but these were absent the of this character below from the specimen studied by me, indicating variability the species level. I. latifolia is native in Japan. BLUMEA VOL. 376 22, No. 3, 1975 subsect. Ilex laurifolia Zipp. ex Loes. (subg. Euilex, series A, sect. Excelsae, Laxae) Adaxial cuticle abaxial cuticle Lamina 440 fim. 7—8/tm; 4—5 /.im. Stomata mainly with and anisocytic, rarely cyclocytic, 30 x 23 [im. Adaxial epidermis anticlinal periclinal cells. Bundle sheaths and of subdivisions, containing mucilage part spongy tissue lignified. Some fusiform sclereids in ground tissue of midrib. Minute solitary crystals in adaxial vascular epidermis. Petiole with simple open system (type 4). NEW GUINEA: Van Material studied. Royen 5415. Note. With mainly anisocytic stomata and leaf scleromorphy expressed in partly lignified mesophyll and a lignified bundle sheath, this species resembles several Malesian those of and of Euilex. species, notably subgenus Byronia, series A, some species subgenus I. amplifolia, from the same subsection but native in S. America differs in several characters. Loesener's remark about similarities with I. in the anatomical character cymosa finds support There also from New complex. is a very striking similarity with I. brassii also Guinea, with which species it shares the, for Ilex unusual, sclerified cells in the abaxial ground tissue of the midrib. Ilex ledermannii Loes., Bot. Jahrb. 59 (1925) 81 (subg. Byronia, series A) Abaxial and adaxial cuticle Lamina i8o//m. c. I fim. Stomata anisocytic to cyclocytic, Adaxial with cells. Bundle sheaths and of 28 x 21 fim. epidermis bulging mucilage part Petiole with vascular spongy tissue lignified. simple open system (type 4). Material studied. NEW GUINEA: Koster BW 7135. Note. Like several species from subgenus Byronia, I. ledermannii has scleromorphic leaves with lignified mesophyll and bundle sheaths. Several other Malesian species from subgenus Euilex (I. archboldiana, I. laurifolia, and I. stapfiana) are also rather similar. Ilex longipes Champ, (subg. Prinus, series B) Lamina Adaxial cuticle abaxial cuticle Indumentum of i6o//m. 1—2/im; c. 1 /mi. rather broad-based midrib short but hairs on and petiole. Stomata predominantly ano- abaxial mocytic, sometimes anisocytic or cyclocytic, 26 X 22fim. Unspecialized epidermal cells with strongly undulated anticlinal walls. Adaxial epidermis with periclinal sub- divisions. Both epidermides containing bulging mucilage cells. Petiole with simple open vascular system (type i). Material studied. U.S.A. Tennessee: Biltmore Herb. 4063. of leaf N o t e. As most species subgenus Prinus the is typically mesomorphic , anatomy with distinctive for the rather broad-based hairs and the fact no characters, except sparse that both abaxial and adaxial epidermides are mucilaginous and contain bulging cells. Ilex Arn. montana Griseb. macfadyenii (Walp.) Rehd., J. Arb. 3 (1922) 215; I. (Sw.) (subg. Euilex, series A, sect. Excelsae, subsect. Laxae) Adaxial cuticle abaxial cuticle unicellular hairs Lamina 190 /um. 2—3 /mi; 3 fim. Sparse onmidrib and 26 x Adaxial petiole. Stomata cyclocytic, rarely complex, 22 /mi. epidermis with anticlinal subdivisions and mucilage cells. Cork warts with characteristic outlines. Petiole with vascular simple open system (type 4). studied. Material JAMAICA: Krug & Urban 5813. P. Baas: Anatomy of Aquifoliaeeae etc. 377 Note. The sunken cork warts with angular outlines together with other similarities of The recall the situationin I. coriacea of section Cassinoides the same series. leaf anatomy of I. amplifolia of the same subsection also resembles I. macfadyenii in many characters. Ilex macrophylla Wall, ex Hook. f. (subg. Byronia, series A) Adaxial cuticle abaxial cuticle Stomata Lamina 210/rm. 2—3/tm; 1fim. anisocytic, Adaxial with rarely cyclocytic or complex, 23 X epidermis strongly undulating anticlinal walls, with periclinal subdivisions, and containing mucilage cells. Bundle of Petiole with vascular sheaths and part spongy tissue lignified. simple open system (type 4). Material studied. SUMATRA: Achmad 991. Note. Affinities with I. pleiobrachiata (=I. cymosa) as suggested by Loesener are also simi- supported by the leaf anatomy of this rather scleromorphic species. There are larities with several other tropical species sharing the scleromorphic character complex, but lacking distinctive anatomical features. Ilex martinianaD. Don (subg. Euilex, series D, sect. Thyrsiflorae) 260 Adaxial and abaxial cuticle unicellular hairs abaxial Lamina fim. 5 fim. Sparse on surface and adaxially of midrib. Stomata mainly anisocytic, occasionally cyclocytic, cells with anticlinal walls. Abaxial 26 X 23 fim. Epidermal strongly undulating epidermal cells (including subsidiary cells and guard cells) with lignified walls. Adaxial epidermis with infrequent anticlinal subdivisions. Cork warts. Locally developed hypodermis over midrib lignified. Bundle sheaths lignified. Spongy tissue strongly birefringent. Petiole with simple open vascular system (type 4). Material studied. BRITISH GUIANA : Schomburgk 385. Note. Frequent cork warts and a lignified abaxial epidermis characterize this species, which shares these features with I. ovalifolia of the same section. I. retusa from S. America and I. triflora from Malesia, each from different series, also resemble I. martinianarather closely. Ilex micrococca Maxim, (subg. Byronia, series B) — Fig. 2 Adaxial cuticle — abaxial cuticle Indumentum of Lamina 3 4fim; 2 —3 fim. and lateral veins. fairly long, partly septate, hairs on petiole, and on adaxial side ofmidrib Stomata mainly anomocytic, often tending to cyclocytic, occasionally anisocytic, 31 X21 urn. Adaxial epidermis composed of bulging mucilage cells; abaxial epidermis also con- cells. taining some bulging Petiole with simple open vascular system (type 4). CHINA. & Material studied. Kweichow: Steward, Chiao Cheo 713. Note. With its mesomorphic leaves, mainly anomocytic stomata, and partly septate hairs this species recalls subgenus Prinus rather than other species of subgenus Byronia. Ilex microdonta Reiss. (subg. Euilex, series C, sect. Microdontae, subsect. Eumicrodontae) Lamina Adaxial and abaxial cuticle unicellular hairs midrib. 3001am. c. 2 fim. Sparse on Stomata or X Petiole with mainly cyclocytic, rarely anisocytic complex, 26 21 fim. simple closed vascular system (type 7). Material studied. BRAZIL. Santa Caterina: Reitz & Klein 8400. 378 BLUMEA VOL. 22, No. 3, 1975 be Note. Loesener's suggestion of an affinity with I. brevicuspis cannot supported by in In the leaf anatomy. The two species differ mainly petiole anatomy. absence of very distinctive features, the existing overall leaf anatomical similarities with I. hanceana from and with China Japan I. chinensis from may be systematically quite irrelevant. Ilex mitis (L.) Radlk. (subg. Euilex, series C, sect. Lemurenses) Adaxial cuticle abaxial cuticle Lamina 360fim. 9 fim; 4—5 fim. Stomata cyclocytic, Slender unbranched tending to bicyclic or complex, 28 x 26 fim. undulating sclereids, adaxial Petiole with sparse throughout mesophyll. Minute solitary crystals in epidermis. simple open vascular system (type 4). Maas Material studied. KENYA: Geesleranus 5977. foliar combination with a number of other less N o t e. By its infrequent sclereids, in distinctive this the anatomical features, species approaches group of I. buergeri, I. curranii, I. ficoidea, I. formosana, and I. glomerata, all of subsection Repandae from Asia. I. mitis stands the of rod- the out, however, by possession minute to needle-shaped crystals in I. section Lemurensis adaxial epidermis. Similarities with fargesii from are also present. Ilex opaca Ait. (subg. Euilex, series A, sect. Cassinoides) Adaxial cuticle abaxial cuticle Lamina 310/im. 9 fim; 4 —5jum. Stomata bicyclic, sometimes cyclocytic, x Epidermal cells with strongly undulating anticlinal 37 35 fim. walls. Abaxial epidermis with lignified unspecialized cells and guard cells. Bundle sheaths and of Petiole with closed vascular part spongy tissue lignified. simple system (type 6). Lawrence Material studied. U.S.A. New Jersey: & Dress 246-, some other specimens for testinglignification of epidermal cells. abaxial and Note. Lignified unspecialized epidermal cells, bicyclic stomata, a simple closed vascular make this rather This combinationof characters system species outstanding. been Ilex and has not encountered in any other Ilex species. canariensis I. cassine from the section but have same share the lignified epidermis simply monocyclic or cyclocytic to and vascular Ilex also of anisocytic stomata an open system. glabra and I. coriacea section of crucial characters Cassinoides share none the of I. opaca. Loesener's remark that I. opaca similar several the absence is in respects to I. cassine can be supported by leafanatomy, in of other species resembling I. opaca more closely. Ilex oppositifolia Merr., J. Arn. Arb. 20 (1939) 222 (subg. Rybonia) Adaxial cuticle abaxial cuticle Unicellular rather Lamina 660 /im. 18—28 fim; 4—5 /um. broad-based, thick-walled hairs on both sides of lamina. Stomata irregularly cyclocytic, Adaxial with anticlinal and subdivisions. 37X31 fim. epidermis periclinal Hypodermis differentiated in several of lamina. Bundle sheaths and of adaxially parts part spongy Midrib of sessile leaves with closed vascular tissue lignified. ± complex system (type 8). Material studied. BORNEO: Clemens 31895. Note. As the single representative of Loesener's (1942) subgenus, this species shares anatomical characters with I. hypoglauca of subgenus Byronia and with I. archboldiana and I. versteeghii of subgenus Euilex, section Rugosae. The complex closed arrangement of the vascular of the midrib distinctive feature of this tissue (petioles are ± absent) is a species, which has a leaf very scleromorphic anatomy. P. Baas: Anatomy of Aquifoliaceae etc. 379 — Plate Ilex ovalifolia Mey. (subg. Euilex, series D, sect. Thyrsiflorae) III, 3 Adaxial cuticle abaxial cuticle Stomata mainly Lamina 320/rm. 4—6/tm; 3 fim. x 28 cells with strongly undulating anisocytic, occasionally cyclocytic, 33 /irn. Epidermal anticlinal walls. Abaxial epidermis with lignified unspecialized cells. Adaxial epidermis with periclinal subdivisions, containing mucilage cells. Cork warts. Bundle sheaths and of Petiole with simple vascular (type 4). part spongy tissue lignified. open system Material studied. SURINAM: B.B.S. 179 (U). Note. Lignified abaxial unspecialized epidermal cells, mainly anisocytic stomata, and regular, frequent cork warts together with scleromorphic features such as partly tissue and bundle sheaths of this recall the situationin I. lignified spongy lignified species each of different sections. martinianaof the same section but also of I. triflora and I. retusa, Ilex Hil. series paraguariensis St. var. sincorensisLoes. (subg. Euilex, C, sect. Microdontae, subsect. Repandae) — Plate III, 4 Lamina Adaxial and abaxial cuticle Unicellular hairs on whole 440/um. 7—9 jum. midrib and adaxial surface, onabaxial surface confined to major veins. Stomatacyclocytic, Abaxial unspecialized epidermal cells with straight thick anticlinal walls; 32x21 fim. very those of adaxial epidermis less markedly so. Cork warts. Minute solitary crystals in closed vascular adaxial epidermis. Petiole with simple system (type 6). Material studied. BRAZIL. Bahia: Ule 7082. Note. Thick-walled unlignified abaxial epidermal cells, a simple closed vascular distal end of the and cork warts are remarkable features system in the petiole, frequent of is of the studied this species. Within subsection Repandae, I. repanda the only one species is which shares the petiole vasculature with I. paraguariensis. I. paraguariensis furthermore the which characterized by minute prismatic to needle-like crystals in adaxial epidermis, character is also encountered in several unrelated species. I. find it impossible to indicate of leaf rather aberrant close affinities I. paraguariensis, and anatomically it is anyway whole. within subsection Repandae and also within section Microdontae as a Ilex peduncularis F. v. Muell. (subg. Byronia, series A) Adaxial cuticle abaxial cuticle irregularly Lamina 270 fim. 3—4/im; 2/jm; Stomata Adaxial cyclocytic, anisocytic, ± anomocytic or tending to complex, 26 x 23 /tm. epider- mis with frequent anticlinal and periclinal subdivisions and containing some mucilage cells of and bundle sheaths lignified. Petiole with (multiple epidermis). Part spongy tissue vascular 4). simple open system (type AUSTRALIA: Brass Material studied. 20237. The Note. This species is without outstanding anatomical features. anatomy agrees fairly well with several other species from the same series (e.g. I. cymosa). Ilex pedunculosa Miq. (subg. Euilex, series A, sect. Excelsae, subsect. Umbelliformes) Adaxial cuticle abaxial cuticle unicellular Lamina 190/nn. 3—5/im; 2 —3 /im. Sparse hairs on midrib and petiole. Stomata cyclocytic, occasionally anisocytic, 30x27/1111. Petiole with simple open vascular system (type 2). P. Material studied. JAPAN: A. Wilson s.n. 380 BLUMEA VOL. 22, No. 3. 1975 Note. Lack oflatero-dorsal wing bundles in the petiole is the only noteworthy leaf of There several anatomical feature this species. are points of difference with both I. chinensis and I. rotunda from the same subsection. Similarities with some other species used mutual affinities. cannot be positively to support Ilex perado Ait.; I. platyphylla Webb. (subg. Euilex, series C, sect. Aquifolioides, subsect. — Plate Oxyodontae) IV, 7 Adaxial cuticle — abaxial cuticle than Lamina 410—470 /J,m. 4 3—12 /im(thicker in unicellular hairs ofsheet adaxialcuticle Hunt 45 only). Sparse onpetiole go9.67—700 only. and Adaxial with Stomata cyclocytic bicyclic, 34x33—36x35/im. epidermis frequent anticlinal subdivisions. Hypodermis of 1—2 cell layers. Marginal sclerenchyma strand. Petiole with vascular (type 4). simple open system 3 or AZORES: Hunt — CANARY ISLANDS. Lems sine Material studied. 43 (U). Teneriffe: 2722; nomine, Leiden herb, sheet 909.67—700. Note. Marginal sclerenchyma strands, bicyclic stomata together with almost all affinities with of other less distinctive anatomical features indicate the species group I. aquifolium, I. bioritsensis, I. cornuta, I. dipyrena, I. integra, and I. pernyi. There are also similarities with I. fargesii. The species is also more or less anatomically identical to I. altaclarensis, which is the hybrid between I. perado and I. aquifolium. Ilex pernyi Franch. (subg. Euilex, series C, sect. Aquifolioides, subsect. Oxyodontae) — Plate V, 4; fig. 1 Adaxial cuticle abaxial cuticle Lamina 510—530/tm. 9—16/mi; 5 —8/mi. Sparse, fairly broad-based, unicellular hairs on midrib and petiole. Stomatacyclocytic and bicyclic, x & x 28 Adaxial with infrequent anticlinal subdivisions. Marginal 27 29 30 fim. epidermis strand. Petiole with vascular sclerenchyma simple open system (type 3). Material studied. GERMANY, cult. hort. Munich: Boom 27299. — NETHERLANDS: cult. hort. Leiden. with Note. I. pernyi belongs the previous species to the same group which is charac- terized by marginal sclerenchyma strands, bicyclic stomata, and several other leaf ana- tomical characters. Ilex fargesii and I. integra also resemble I. pernyi to a considerable extent. I. pernyi is native in China. Ilex pubescens Hook. & Arn. (subg. Euilex, series C, sect. Prinifoliae) — Plate I, 5 Adaxial and abaxial cuticle Indumentum of hairs Lamina 150/mi. c. I/mi. long septate both leafsurfaces. Adaxial with on Stomata anisocytic, 21 X 19/cm. epidermis periclinal subdivisions. Both epidermides containing bulging mucilage cells. Petiole with simple vascular open system (type 3). CHINA. S. Y. Lan Material studied. Kwantung: 20137. Note. Abundant long uniseriate hairs characterize this mesomorphic species, which shows I. some anatomical affinities with species from subgenus Prinus, particularly amelanchier. The exclusively anisocytic stomata of I. pubescens plead, however, against inclusion in subgenus Prinus. Loesener's remark that this species is intermediate between subgenus Euilex and subgenus Prinus can be fully confirmed by the anatomical evidence. P. Baas: Anatomy ofAquifoliaceae etc. 381 series Ilex pulogensis Merr., Philip. J. Sc. Bot. 5 (1910) 358 (subg. Euilex, D, sect. — Plate Indico-Malaicae) II, 5 Adaxial cuticle abaxial cuticle — Stomata cyclocytic Lamina 300 //m. 3 //m; 4 5/tm. Adaxial with subdivisions, and bicyclic, 32x23 epidermis periclinal containing __ /im. strongly bulging mucilage cells. Cork warts. Bundle sheaths lignified. Petiole with vascular simple open system (type 4). Material studied. PHILIPPINES : Conklin & Buwaya PNH 79634. Note. Frequent cork warts constitute the most outstanding leaf anatomical feature of this soecies. which Merrill (I.e.) assigned to section Indico-Malaicae. In the species resemblance with studied and belonging to that section it is impossible to find sufficient In view of the rather leaf I. pulogensis to support Merrill's treatment. 'unspecialized' Loesener of it is, however, impossible to indicate alternatives. anatomy I. pulogensis, but the (1942) listed I. pulogensis as a species 'ohne Punkte' (without cork warts), 2 speci- PNH BS both have mens present in the Rijksherbarium ( 79634; 31835) frequent regular cork warts. Ilex repanda Griseb. (subg. Euilex, series C, sect. Microdontae, subsect. Repandae) Lamina Adaxial cuticle —6 abaxial cuticle Stomata cyclocytic, 290 /im. 4 /tm; 3 —5 /um. Adaxial with subdivisions and bulging mucilage cells. 29 x27/urn. epidermis periclinal Bundle sheaths birefringent, possibly lignified. Petiole with simple closed vascular system (type 7). Material studied. BAHAMAS: Curtiss 88. somewhat resembles I. microdonta of the Note. Through its petiole type I. repanda characters other same section from Santa Catarina. In the absence of other distinctive no its affinities. suggestions can be made about Ilex retusa Klotsch (subg. Euilex, series A, sect. Crassifoliae) abaxial cuticle unicellular Adaxial cuticle — c. Sparse Lamina 480/tm. 15 19/tm; 15 //m. and Abaxial hairs on midrib and petiole. Stomata cyclocytic anisocytic, 34 x 32/tm. epidermis with lignificd unspecialized cells and guard cells. Adaxial epidermis with periclinal subdivisions and bulging mucilage cells. Cork warts. Bundle sheaths lignified. Petiole with vascular simple open system (type 4). Material studied. BRITISH GUIANA. Maguire & Fanshawe 32483 (U). cork with abaxial Note. Well-developed and frequent warts together a lignified and other features affinities of this with I. ovalifolia and I. epidermis suggest species triflora of different sections. Ilex revoluta Stapf (subg. Euilex, series C, sect. Rugosae) cuticle Lamina 330/tm. Adaxial cuticle 10—i2//m; abaxial 6—8 /mi. Moderately long, whole abaxial curved, unicellular hairs on midrib and petiole, also infrequent over Adaxial surface. Stomata cyclocytic, sometimes tending to anomocytic, 33 x 29/nn. epidermis with infrequent anticlinal and periclinal subdivisions, possibly containing mucilage cells. Cork warts. Petiole with simple open vascular system (type 3). Material studied. : BORNEO Meyer SAN 29253. BLUMEA VOL. No. 382 22, 3, 1975 others section. Note. The species differs from the studied of the same Affinities with species of other sections cannot be suggested because of the lack of enough distinctive characters. Ilex rotunda Thunb. (subg. Euilex, series A, sect. Excelsae, subsect. Umbelliformes) Adaxial and abaxial cuticle and Lamina 220 fim. 2 fim. Stomata cyclocytic anisocytic, to X28 Adaxial with subdi- occasionally tending complex, 33 /urn. epidermis periclinal visions and bulging mucilage cells. Very minute solitary crystals in adaxial epidermis. with Petiole simple open vascular system (type 4). Material studied. CHINA. Kwantung: Lan 20237. Note. This species shows no unusual anatomical features. There are several points of difference with I. chinensis and I. pedunculosa of the same subsection. Ilex sclerophylloides Loes. (subg. Byronia, series A) Lamina Adaxial cuticle abaxial cuticle 380/tm. 4—SEm; 2 —3 /. system (type 4). Material studied. BORNEO: Herb. For. Dept. Sarawak 4381. for of features such Note. Except a strong expression scleromorphic as lignifi cation of features spongy tissue and bundle sheaths, there are no unusual in this species. There of are good overall similarities with several other species subgenus Byronia (I. brassii, also I. cymosa, I. macrophylla) but with some species of subgenus Euilex. Ilex scrabidula Merr. & Perry, J. Arn. Arb. 22 (1941) 258 (? subg. Euilex, series D, sect. Indico-Malaicae) Lamina Adaxial cuticle 2 abaxial cuticle —2 Stomata 320fim. —3 /j,m; 1 fim. mainly Cork Bundle sheaths cyclocytic, occasionally anisocytic, 28x24fim. warts. lignified. Petiole with vascular simple open system (type 3). Material studied. NEW GUINEA: Brass 10302. Note. Merrill & Perry (l.c.) remarked that this species somewhat resembles I. spicata; hence its treatment in section Indico-Malaicae here. Leaf anatomically there are no very cork Similar anatomies in remarkable characters, except abundant warts. can be found some species, notably I. spicata, supporting Merril & Perry's suggestions. I. spicata differs, however, in its complex vascular supply of petiole and midrib. sebertii Ilex Panch. & Seb. (subg. Euilex, series A, sect. Excelsae, subsect. Laxae) Adaxial cuticle 8 — abaxial cuticle Lamina 330 fim. 9fim\ 6—8 fim. Stomata mainly cyclocytic, occasionally Adaxial epidermis with anticlinal and anisocytic, 36x31 fim. periclinal subdivisions, containing mucilage cells. Part of spongy tissue biréfringent, Petiole with vascular possibly weakly lignified. simple open system (type 4). Material studied. NEW CALEDONIA : McKee 2362. There Note. are no distinctive anatomical features in this species, and the anatomical evidence is rather neutral with regard to Loesener's suggestion of an affinity with I. laurifolia. P. Baas: Anatomy ofAquifoliaceae etc. 383 series A) — Ilex serrata Thunb., incl. var. sieboldii (Miq.) Loes. (subg. Prinus, Fig. 3 Adaxial and abaxial cuticle c. 1 Unicellular broad-based Lamina no—130/im. //in. hairs surfaces. Stomata X Abaxial epidermis numerous on both leaf anomocytic, 24 17/mi. Petiole with with strongly undulating anticlinal walls forming an interlocking pattern. simple open vascular system (type 2). Material studied. NETHERLANDS. cult.Wagcningen: Rijksarborelum 946 (=Boom 616}). —JAPAN: Leiden herb, sheet 904.134—327. Note. Strongly undulating anticlinal epidermal walls, anomocytic stomata, together in line the characters of most of with a mesomorphic leaf anatomy, are well with species indumentum subgenus Prinus, notably I. verticillata of the same series. The difference in had been because between these two species would not hold if more material studied, Loesener recorded both glabrous and puberulous forms or varieties for the two species. Ilex sideroxyloides (Sw.) Griseb. (subg. Euilex, series C, sect. Microdontae, subsect. Sideroxyloides) Adaxial cuticle abaxial cuticle 1 Stomata mainly Lamina 300/tm. 2 —3 /mi; —2/mi. cells with anisocytic, occasionally cyclocytic, 28 x 20/mi. Epidermal strongly undulating anticlinal walls. Adaxial epidermis with periclinal subdivisions; both epidermides con- vascular taining bulging muciiage cells. Petiole with simple open system (type 4). Material studied. WEST INDIES: Lesser Antilles. Eggers 6234. in this to justify Note. There are no distinctive leaf anatomical features species diva- meaningful systematic suggestions. There are rather important differences with I. ricata from the same subsection. Ilex sikkimensis Kurz (subg. Euilex, series C, sect. Aquifolioides, subsect. Insignis) Lamina Adaxial cuticle —8 abaxial cuticle Stomata 170/tm. 7 /mi; 3—5 /mi. irregularly Petiole with vascular cyclocytic, sometimes anomocytic, 29x27/1111. simple open system (type 4). Material studied. SIKKIM. Sinchal: Thomson s.n. anatomical characters. Note. This species is without distinctive leaf There are some with and the differences I. insignis I. latifolia of same subsection. Ilex spicata Bl. (subg. Euilex, series D, sect. Indico-Malaicae) Adaxial and abaxial cuticle 2 Stomata cyclocytic, x Lamina 350/tm. —3 /mi. 29 24/mi. sheaths Adaxial epidermis containing some bulging mucilage cells. Cork warts. Bundle of lignified. Petiole with complex lignified. Part spongy tissue birefringent, possibly closed vascular system (type 8). Material studied. JAVA : Van Ooslstroom 13113. Note. I. stapfiana of section Indico-Malaicae resembles I. spicata in many respects but differs from lacks regular cork warts. I. cissoidea of the same section more strongly I. spicata. Ilex stapfiana Loes. (subg. Euilex, series D, sect. Indico-Malaicae) Adaxial cuticle abaxial cuticle Lamina400/tm. 5 —7/«n; 4—6/tm. Stomata anomocytic BLUMEA VOL. No. 384 22, 3, 1975 Adaxial with subdivisions and strongly to cyclocytic, 29 x 20jum. epidermis periclinal cell bulging mucilage cells. Bundle sheaths lignified. Spongy tissue with mucilaginous walls. Petiole with complex closed vascular system (type 8). Material studied. BORNEO. Sarawak: Haviland 3557. with by Note. Loesener's suggestion about affinities I. spicata are largely supported with Besides a adaxial leaf anatomy (see above). showing strongly mucilaginous epidermis subdivided of the tissue in I. stapfiana has mucilaginous large bulging cells, part spongy other sub- cell walls. The species shows also similarities with some Malesian species from and with I. from subgenus Rybonia. genus Byronia oppositifolia Ilex theezans Mart. (subg. Euilex, series C, sect. Megalae, subsect. Pedicellatae) — Fig. 11 and 12 Adaxial cuticle —6 abaxial cuticle 2 m. Indumentum Lamina 300—320/urn. 4 /jm; —3 [i of unicellular hairs midrib. Stomata and in absent or sparse on anisocytic cyclocytic X & X Adaxial epidermis with different ratios in the 2 specimens studied, 25 21 29 23 /um. cells. or bulging mucilage anticlinal and periclinal subdivisions, containing few many cells. Part of tissue Abaxial epidermis also with some bulging spongy biréfringent, possibly vascular lignified. Petiole with simple open system (type 4). Material studied. BRAZIL: Dttseti io6o8\ Santa Catarina: Reitz & Klein 3222 (U). Because Note. The two specimens studied differ from each other in some characters. features this shares leaf of this, and in the absence of distinctive leaf anatomical species its with a number of from sections and different subgenera. anatomy large species many of subsection to these similar I. brasiliensis and I. integerrima the same belong species. for theezans Thevenard (1906) recorded several mucilaginous hypodermal layers I. (as vascularization different from that in in I. anomala), and furthermore described a midrib the material studied here. Possibly his material was misidentified. Ilex tolucana Hemsl. (subg. Euilex, series C, sect. Microdontae, subsect. Repandae) Adaxial cuticle abaxial cuticle Indumentum of thick- Lamina 380/nm. 6/im; 2/tm. hairs Stomata walled, slender, fairly long, unicellular over whole lamina. cyclocytic, Guard cells lignifted (?). Adaxial with anticlinal tending to complex, 27 x 27fim. epidermis vascular Midrib with simple subdivisions. Petiole with simple open system (type 4). closed system (type 6). Material studied. MEXICO: Meyer & Rogers 2942 (U). and distinctive in this Note. Indumentum, guard cells, midrib vascularization are with similar leaf species. Within subsection Repandae I have not seen species a anatomy. ofsection Microdontae. There are some similarities with I. microdonta Ilex triflora Bl. (subg. Euilex, series C, sect. Microdontae, subsect. Stigmatophorae) — uni- Adaxial cuticle abaxial cuticle 2 Sparse Lamina 130—210/mi. 3—8/tm; 8/tm. also of Stomata cyclo- cellular hairs on midrib and petiole, rarely on abaxial side lamina. in cytic and anisocytic in different proportions different specimens, 25 X24—32x29/mi. few Guard cells lignified. Unspecialized cells of abaxial epidermis mostly lignified; very Adaxial with subdivisions and cells lignified only in Forbes 2goi. epidermis periclinal o y _ Cork warts. tissue partly bulging mucilage cells in very variable frequencies. Spongy lignified. Petiole with simple open vascular system (type 3). P. Baas: Anatomy ofAquifoliaceae etc. 385 Van Sleenis — Material studied. BORNEO: Hallier B 2471. —JAVA: 17146. SUMATRA. Forbes 2goi,Jacobson 536, Maradjo 214, Meijer 6722. with Note. In spite of its variability the species remains outstanding its lignified (at least abaxial and cork warts. This in combinationwith the pro parte) epidermis frequent World partly anisocytic stomata would suggest affinities with the New species I. martiniana and I. ovalifolia from section Thyrsiflorae of series D. Thyrsoprinus. 268 series Ilex tuerckheimii Loes.; Symb. Antill. (Ed. Urban) 7 (1912) (subg. Euilex, A, sect. Excelsae, subsect. Laxae) hairs midrib Adaxial and abaxial cuticle m. Sparse unicellular on Lamina 280/tm. 2—3 /2 of and petiole. Stomata anisocytic and cyclocytic, 24x21/201. Hypodermis 1—2 layers of partly bulging cells, probably containing mucilage. Spongy tissue with abaxial ligni- cells. Petiole fied layer, remainder composed of thick-walled ± collenchymatous with vascular 4). simple open system (type Material studied. DOMINICA: Fuerles 1683. in I. anomala and Note. A mucilaginous 'true' hypodermis, comparable to that some the to die abaxial epidermis strongly lignified cells of the spongy tissue in layer adjacent characters this Within subsection Laxae none of the are the only distinctive of species. of species studied resembles I. tuerckheimii very closely, nor have I found evidence closer affinities with species from other sections or subsections. series Micranthae, Ilex umbellataKl. var. humirioides (Reiss.) Loes. (subg. Euilex, C, sect. subsect. Epunctatae ) — Fig. 14 abaxial Lamina Adaxial and cuticle 1—2 unicellular hairs onpetiole, 310/2m. /2m. Sparse Adaxial midrib, and major veins. Stomata anisocytic and cyclocytic, 18 x i6/2m. epidermis cells. with periclinal subdivisions and strongly bulging mucilage Petiole with simple open between and vascular system (intermediate type 3 type 4). Material studied. BRITISH GUIANA: Fanshaw For. Dept. 3693 (U). anatomical features. Affinities with I. Note. This species shows no outstanding could be inundata, I. guianensis, and I. sideroxyloides as suggested by Loesener supported by from different leaf well as affinities with several other infrageneric anatomy, as species taxa recognized by Loesener. Arb. series Ilex versteeghii Merr. & Perry, J. Arn. 22 (1941) 259 (subg. Euilex, C, sect. Rugosae) Lamina Adaxial and abaxial cuticle Stomata cyclocytic, 530/2m. 6/2m. 38x30/2m. cells with thick anticlinal walls. Adaxial epidermis Unspecialized abaxial epidermal very with anticlinal and periclinal subdivisions, containing bulging mucilage cells. Bundle of tissue Minute in adaxial sheaths and part spongy lignified. solitary crystals epidermis. vascular Petiole with simple open system (type 4). Material studied. NEW GUINEA: Brass 11772. cells and Note. Except for thick-walled unspecialized epidermal a scleromorphic leaf tissue, lignified bundle sheaths), this species lacks anatomy (partly lignified spongy close I. I. distinctive features. Of section Rugosae, I. archboldiana is very to versteeghii. BLUMEA VOL. No. 386 22, 3, 1975 cork and in other char- revoluta of the same section differs in having regular warts some several of and of other acters as well. Overall similarities with species subgenus Byronia series and sections of subgenus Euilex are also present. Ilex verticillata (L.) Gray (subg. Prinus, series A) — Plate II, 2 Adaxial and abaxial cuticle X 21 Lamina i6o/tm. c. I fim. Stomata anomocytic, 35 /mi. Unspecialized abaxial epidermal cells with strongly undulating anticlinal walls forming Adaxial with subdivisions and an interlocking pattern. epidermis infrequent periclinal cells. vascular slightly bulging, possibly mucilaginous Petiole with simple open system (type 4). Material studied. CANADA. Quebec: Roland Germain 6216. Note. With its mesomorphic leaf anatomy, anomocytic stomata, and strongly undulating epidermal abaxial anticlinal cell walls this species resembles I. serrata and several other species of subgenus Prinus. Ilex vitiensis Gray (subg. Euilex, series A, sect. Excelsae, subsect. Laxae) Adaxial cuticle abaxial cuticle Unicellular hairs on Lamina 270 / tissue strongly birefringent. Petiole with simple open vascular system (type 4), but median strand composed of several discrete bundles. Material studied. FIJI: Degener 14592. Note. A small-celled hypodermis is the only outstanding feature of this species, resembles others the latter which according to Loesener amongst I. laurifolia. However, characters. The of species differs from I. vitiensis in several leaf anatomical only species subsection Laxae with hypodermal development is I. tuerckheimii, but the type of hypo- dermis is different. The anatomical evidence for indicating close affinities is therefore inconclusive. Ilex wallichii Hook. f. (subg. Byronia, series A) Adaxial cuticle abaxial cuticle Stomata mainly Lamina 320/mi. 5 —6/im; 3—4/mi. Adaxial with bulging anisocytic, rarely cyclocytic, 19X 18/mi. epidermis infrequent with simple mucilage cells. Bundle sheaths and part of spongy tissue lignified. Petiole open vascular system (type 4). Material studied. MALAYA: Md. Nur SFN 33996. a close I. can be supported by Note. Loesener's suggestion of affinity with cymosa less within the anatomical of There is leaf anatomy: I. wallichii is more or range I. cymosa. also resemblance with other from I. a some species subgenus Byronia (I. brassii, ledermannii, and I. sclerophylloides) and with several species from subgenus Euilex (I. guianensis, I. inundata, I. jenmannii, and I. laurifolia). Ilex zygophylla Merr., J. Arn. Arb. 20 (1939) 222 (subg. Euilex, series B, sect. Vaccinii- series — Plate foliae or subg. Byronia, A) IV, 9; fig. 7 — Indumentum Lamina Adaxial cuticle 18— abaxial cuticle 700/on. 22/im; 14 16/tm. broad-based and with several basal of thick-walled hairs, sometimes rather septs (as in P. Baas: Anatomy of Aquifoliaceae etc. 387 often whole lamina. Stomata I. hypoglauca), unicellular, present over mainly cyclocytic, Adaxial with anticlinal with rarely anisocytic, 42 X 34/tm. epidermis straight walls, in- frequent anticlinal subdivisions. Hypodermis of 1 cell layer, lignified. Petiole with simple vascular open system (type 4). Material studied. N. BORNEO: Meijer SAN38459. Note. WhenMerrill (I.e.) described this species as opposite-leaved he mentioned that series both I. zygophylla and I. oppositifolia might belong to either subgenus Byronia A, or subgenus Euilex, section Excelsae. Loesener (1942), however, corrected Merrill's obser- vation on I. zygophylla and recorded leaves alternate or in whorls of three. He referred the species to section Vacciniifoliae. Anatomically the species is rather outstanding with its — for Ilex unique — lignified adaxial hypodermis, its indumentum of broad-based hairs which are partly septate as in I. hypoglauca, and the unusually large-sized stomata. The leaves are moreover highly scleromorphic with strongly lignified spongy tissue and sheaths. the affinities bundle Anatomically are with subgenus Byronia (I. hypoglauca) of Euilex. the rather than with section Excelsae or Vacciniifoliae subgenus However, species remains anatomically rather isolated. Lists of Ilex species with certain leafanatomical characters reference and identification lists of the distribution of the For easy purposes are given characters more important leaf anatomical over the different species studied. Because of considerable overlapping between species it will be impossible to key out individual Ilex combinations ofcharacters from the lists can be species on leaf anatomical characters, but used to narrow down the possibilities and find indications for the identification. doubtful the followed If than In intermediate or cases species name is by ±. more one been the specimen has studied, species name may be followed by p.p. (pro parte) indicating that the feature involved is not constant for that species. Hairs longand/or uniseriate I. amelanchier I. decidua I. kwantungensis(±) I. pubescens I. brasiliensis I. dubia I. micrococca (±) I. revoluta (±) I. cassine Hairs unicellular with broad septate base I. hypoglauca I. zygophylla (±) Abaxial epidermis (at least partly) lignified I. I. canadensis I. dumosa I. ovalifolia amara p.p. I. argentina I. cassine I. martiniana I. retusa I. I. I. I. boliviana crenata p.p.? opaca triflora Adaxial epidermis with anticlinal division walls I. altaclarensis I. cassine I. inundata I. pernyi I. amplifolia I. cissoidea I. krugiana I. revoluta I. anomala I. coriacea I. laurifolia I. sebertii I. aquifolium I. formosana I. macfadyenii I. theezans I. archboldiana I. goshiensis I. martiniana I. tolucana I. argentina I. grandifolia I. oppositifolia I. versteeghii I. brasiliensis 1. hanceana I. peduncularis I. zygophylla I. brassii I. integerrima I. perado 388 BLUMEA VOL. 22, No. 3, 1975 Adaxial epidermis containingmucilage cells I. amelanchier I. crenata I. kwantungensis I. revoluta (±) I. I. I. I. amplifolia cymosa p.p. laurifolia rotunda I. archboldiana I. danielis I. ledermannii I. sebertii I. argentina I. decidua (±) I. longipes I. sideroxyloides I. asprella I. divaricata I. macfadyenii I. spicata I. brassii I. dubia (±) I. macrophylla I. stapfiana I. brunnea I. grandifolia I. micrococca I. theezans I. canadensis I. guianensis I. ovalifolia I. triflora I. caroliniana I. hypoglauca I. peduncularis I. umbellata I. cassine I. insignis I. pulogensis I. versteeghii I. chinensis I. integerrima (±) I. repanda I. verticillata (±) I. I. clemensiae I. inundata retusa I. wallichii I. coriacea I. krugiana At least part ofthe stomata bicyclic I. altaclarensis I. chinensis I. formosana I. krugiana I. aquifolium I. cornuta I. glomerata I. latifolia I. curranii I. I. I. argentina goshiensis opaca I. bioritsensis I. dipyrena I. insignis I. perado I. buergeri I. fargesii I. integra I. pemyi I. celebensis I. ficoidea Stomata predominantly anomocytic I. I. verticillata amara longipes I. serrata I. I. decidua Stomata mainly anomocytic to cyclocytic I. amelanchier I. asprella I. dubia I. micrococca I. anomala I. brasiliensis I. dumosa I. stapfina I. archboldiana I. coriacea I. grandifolia Stomata predominantly anisocytic I. brassii I. integerrima I. martiniana I. sclerophylloides I. brunnea I. laurifolia I. ovalifolia I. sideroxyloides 1. I. I. I. wallichii cymosa p.p. macrophylla pubescens Stomata mainly anisocytic to cyclocytic I. I. triflora I. cissoidea I. hypoglaucap.p. peduncularis p.p. I. clemensiae I. inundata I. retusa 1. tuerckheimii I. confertiflora I. jenmannii I. rotunda I. umbellata I. I. ledermanii I. theezans I. vitiensis cymosa p.p. p.p. I. guianensisp.p. Stomata predominantly cyclocytic I. amplifolia 1. curranii I. kleinii I. revoluta I. argentina I. danielis I. krugiana I. scrabidula I. boliviana I. divaricata I. kwantungensis I. sebertii I. brevicuspis I. fargesii I. macfadyenii I. sikkimensis I. buergeri I. ficoidea I. microdonta I. spicata I. mitis I. I. canadensis I. formosana theezans p.p. I. caroliniana I. glabra I. oppositifolia I. tolucana I. cassine I. hanceana I. I. triflora paraguariensis p.p. I. celebensis I. hypoglauca p.p. I. pedunculosa I. versteeghii I. chinensis I. insignis I. pulogensis I. zygophylla I. crenata I. integra I. repanda P. Baas: Anatomy ofAquifoliaceae etc. 389 Guard cell 20 shorter pairs µm long or I. amelanchier I. umbellata I. wallichii Guard cell pairs 35 µm long or longer I. altaclarensis I. clemensiae I. oppositifolia I. versteeghii I. 1. I. I. verticillata aquifoliump.p. latifolia perado p.p. I. I. I. I. asprella opaca sebertii zygophylla 1. canariensis Regular cork warts martiniana I. revoluta I. amara I. coriacea I. I. amplifolia I. crenata I. ovalifolia I. scrabidula I. boliviana I. divaricata I. paraguaricnsis I. spicata I. caroliniana I. glabra I. pulogensis I. triflora I. clemensiae I. macfadyenii I. retusa Continuous hypodermis present I. altaclarensis I. aquifolium I. perado I. vitiensis I. anomala I. brasiliensis I. tuerckheimii I. zygophylla (mucilaginous) (mucilaginous) (mucilaginous) (lignified) At least tissue part of spongy lignified or strongly birefringent I. I. I. I. tuerckheimii amara hypoglauca oppositifolia I. archboldiana I. laurifolia I. ovalifolia I. versteeghii I. brassii I. ledermannii I. peduncularis I. vitiensis I. confertiflora I. macrophylla I. sclerophylloides I. wallichii I. martiniana sebertii I. I. cymosap.p. I. zygophylla I. I. I. grandifolia opaca triflora Bundle sheaths lignified or at least strongly birefringent I. I. amara I. clemensiae I. martiniana sclerophylloides anomala I. I. I. scrabidula I. cymosa opaca I. archboldiana I. dumosa I. oppositifolia I. spicata I. argentina (±) I. grandifolia I. ovalifolia I. stapfiana I. boliviana I. hypoglauca I. peduncularis I. versteeghii I. brassii I. insignis I. pulogensis I. vitiensis I. buergeri I. laurifolia I. repanda (±) I. wallichii I. celebensis I. ledermannii I. retusa I. zygophylla I. cissoidea (±) I. macrophylla Foliar sclereids present throughout lamina I. buergeri I. cornuta(±) I. ficoidea I. glomerata I. confertiflora (i) I. curranii I. formosana I. mitis Marginal sclerenchyma strands I. altaclarensis I. confertiflora (±) I. insignis (±) I. perado I. I. aquifolium I. cornuta I. integra (±) pernyi I. bioritsensis I. dipyrena I. latifolia (±) Minute solitary crystals in adaxial epidermis I. celebensis I. mitis I. rotunda I. versteeghii I. laurifolia I. paraguariensis 390 BLUMEA VOL. 22, No. 3, 1975 bundles Vascular system of petiole simple open without wing (types 1 & 2) I. amelanchier I. caroliniana I. longipes I. serrata I. asprella I. celebensis I. pedunculosa Vascular system ofpetiole simple closed (types 6 & 7) I. chinensis ( ) I. microdonta I. I. :j= opaca repanda closed in midrib Vascular system only I. I. hanceana I. I. tolucana anomala p.p. kwantungensis I. chinensis of 8 Vascular system petiole complex open to closed (types & 9) I. I. I. I. cymosa p.p. hypoglauca p.p. spicata stapfiana I. grandifolia I. oppositifolia Vascular of system petiole interrupted open (type 5) I. clemensiae Vascular system of petiole’double’ (type 10) I. coriacea I. glabra Comparison of leafanatomical data with Loesener’s system the anatomical the In this section leaf ranges of variation of individual infrageneric taxa recognized by Loesener will be discussed. Suggestions will be made about the affinities of these taxa in cases where the leaf anatomical evidence has obvious taxonomic implications. The number of species ofwhich each individual section is composed is given between brackets, and is taken from Loesener (1942). Subgenus RYBONIA (1 species in Borneo) i species studied: I. oppositifolia. this Loesener established subgenus in 1942, without discussing its systematic position From of the of list within the genus Ilex. his placing subgenus Rybonia at beginning the of 5 subgenera, close to subgenus Byronia, one may conclude that he regarded I. oppositifolia of Ilex. The leaf of this indeed recalls that of as a primitive representative anatomy species vascular some species of subgenus Byronia, and the complex system in the midrib of the leaves be used the idea that is derived from the can to support subgenus Rybonia same ancestral stock as section Indico-Malaicae of subgenus Euilex, and as subgenus Byronia. There are also leaf anatomical similarities with some species treated as members of section Rugosae of subgenus Euilex. Subgenus BYRONIA Series A. EUBYRONIA (c. 20 species in tropical Asia, Australia, and Polynesia) studied: I. 11 species anomala, I. brassii, I. celebensis, I. cymosa, I. grandifolia, I. hypoglauca, I. ledermannii, I. macrophylla, I. peduncularis, I. sclerophylloides, I. wallichii. The stomatal type may be predominantly anomocytic, anisocytic, or cyclocytic within P. Baas: Anatomy of Aquifoliaceae etc. 391 celebensis. Stomatal size from this series, whilst partly bicyclic stomata occur in I. ranges 18 varies from the unusual in I. to 19 x to The indumentum type hypoglauca consisting ofshort unicellular hairs, or it is absent. Cork warts are absent or at least never The adaxial be of regular outline and frequent occurrence. epidermis may entirely un- modified, or contain mucilage cells, or show periclinal and/or anticlinal subdivisions with or without mucilaginous daughter cells. A 'true' (mucilaginous) hypodermis occurs in I. in all The vascular in the anomala only. The bundle sheaths are lignified species. pattern distal end of the the from to the more petiole covers ranges simple open (1 —4) complex strands absent from all types (7—9). Sclereids and marginal sclerenchyma are species. show number of features: all anomala and Most species a scleromorphic species (except I. I. celebensis) have partly lignified spongy tissue. Series with a Eubyronia thus appears to be anatomically very heterogeneous great diversity in most leaf anatomical features relevant for the characterization of individual Ilex species. The heterogeneity would be somewhat reduced by expelling I. anomala and I. celebensis from this series, so that series Eubyronia would lack hypodermal differentiation and bicyclic stomata. I. anomala is also wood anatomically rather aberrant (Baas, 1973). The series ofLoesener's other only way to characterize Eubyronia, and most infrageneric taxa leaf anatomically, wouldbe to record the most frequent occurrence of certain features like: sheaths often bundle lignified; spongy tissue mostly partly lignified; with anisocytic additional characterization characters like: cork stomata, etc. As one might use negative warts, sclereids, and marginal sclerenchyma strands absent, abaxial epidermis not lignified, etc. Series B. MICROCOCCA (1 species in East Asia) 1 species studied: I. micrococca. leaf With its septate hairs, anomocytic to cyclocytic stomata, and simple mesomorphic is from of series Similarities with anatomy this species very different those Eubyronia. much subgenus Prinus are stronger. Subgenus Byronia as a whole The with subgenus is leaf anatomically very heterogeneous. Similarities of some species thoseofother subgenera (Rybonia, Euilex, section Indico-Malaicae, and Prinus) may indicate true affinities be due evolution. Theanatomical or to parallel or convergent heterogeneity does not necessarily challenge Loesener's grouping, because he based this subgenus on the characters of inflorescence flower. common occurrence of supposedly primitive and One could that within such of which these characters have accept a group species in primitive been other characters like those of leaf have become retained, anatomy specialized in different ways. Subgenus EUILEX Series A. LIOPRINUS Excelsae and Section (c. 3 1 species in East Asia, Polynesia, Tropical America) 10 species studied: subsection Umbelliformes: I. chinensis, I. kwantungensis, I. pedunculosa, I. rotunda; subsection Laxae: I. amplifolia, I. laurifolia, I. macfadyenii, I. sebertii, I. tuerckheimii, I. vitiensis. An indumentum is consists in I. kwan- absent or ofshort unicellular hairs only, except has and tungensis which numerous long hairs. The stomata are usually aniso- cyclocytic, of the is sometimes has one types predominant or exclusive. Only I. chinensis some partly 392 BLUMEA VOL. 22, No. 3, 1975 Stomatal from bicyclic to complex cyclocytic stomata. size ranges 22x21 to 36x31 fjim. Regular frequent cork warts occur in I. amplifolia and I. macfadyenii only. The adaxial be contain epidermis may entirely unmodified, mucilage cells, or show periclinal sub- divisions with mucilaginous daughter cells, or anticlinal subdivisions. A small-celled in hypodermis occurs I. vitiensis, whilst a mucilaginous hypodermis of partly subdivided tuerckheimii. bulging cells occurs in I. The bundle sheaths and mesophyll are entirely almost all and The vascular unlignified or so in species except I. laurifolia I. vitiensis. system is in of the petiole mainly of the common type 4, but a single strand (type 1) occurs and a closed 6 the midrib I. chinensis and I. pedunculosa simple type system in (at least of) I. kwantungensis. Foliar sclereids and marginal sclerenchyma strands are absent from all species. Section Excelsae is therefore also rather heterogeneous leaf anatomically, albeit less so and than e.g. series Eubyronia. The range in epidermal hypodermal differentiation in Laxae is rather well the variation in in subsection great, as as vascular patterns petiole and because midrib in subsection Umbelliformes. I. amplifolia seems out of place of its frequent cork warts. Some species, notably I. laurifolia, are similar to species of subgenus Eubyronia, and section Indico-Malaicae of subgenus Euilex. Leaf does not characters subsection anatomy provide good diagnostic to separate Umbelliformes and subsection Laxae. Cassinoides 16 and Section (c. species in Macaronesia, America, East Asia) studied: 3 species I. canadensis, I. cassine, I. coriacea, I. glabra, I. opaca. In section Cassinoides the indumentum be consist of short unicellular may absent, hairs, hairs The or of long septate (I. cassine). stomata are mainly cyclocytic but occasionally stomata or to stomata also whilst has anisocytic transitions anomocytic occur, I. opaca predominantly bicyclic stomata. Stomatal size from x to x Regular ranges 24 23 39 35 /cm. cork are and The abaxial of warts present in I. coriacea I. glabra only. epidermis I. canadensis is p.p., I. cassine, and I. opaca has lignified unspecialized cells. The adaxial epidermis entirely unmodified or is mucilaginous with periclinal subdivisions. The bundle sheaths I. are lignified in I. opaca only. The mesophyll is generally unlignified though in opaca it is be partly lignified. The vascular pattern of the petiole may oftype 4, 6, or 10 (I. coriacea and I. glabra). Sclereids and marginal sclerenchyma strands are absent from all species studied. The of be anatomical heterogeneity this section thus also appears to considerable, with absence ofcork and of variables. presence or warts a lignified epidermis as the most striking I. canadensis from Macaronesia resembles I. cassine from North America. Ilex coriacea and I. America glabra, both from N. America, also seem mutually related. I. opaca from N. appears rather isolated anatomically. Section Dasyneurae (c. 3 species in S. America) 1 species studied: I. boliviana. to abaxial with Anisocytic cyclocytic stomata, long hairs, a lignified epidermis frequent regular cork warts, and lignified bundle sheaths characterize this scleromorphic species. Apparently, cork warts are not a constant feature of this section, because Loesener (1942) the recorded that other 2 species have 'punktlose Blatter'. Anatomical resemblance with section I. retusa of the related Crassifoliae seems systematically significant (cf. Loesener, 1908: 47—49). Affinities with I. martinianaand I. ovalifolia ofseries D, section Thyrsiflorae andwith I. triflora of section Microdontae can also be advocated on the strength of anatomi- cal similarities in rather unusual character complexes. P. Baas: Anatomy of Aquifoliaceae etc. 393 Section Crassifoliae (c. 11 species in tropical S. America) i species studied: I. retusa. is similar I. boliviana and the for the I. retusa very to same characterization, except indumentum, would apply. Affinities would therefore also be indicated with section Microdontae, section Thyrsiflorae, and section Dasyneurae. Series Lioprinus as a whole The series is characterized the anatomically very heterogeneous, and can only be by absence of some — elsewhere in Ilex unusual — features such as broad-based septate hairs, sclereids, and marginal sclerenchyma strands. Series B. PALTORIA from Section Polyphyllae (c. 20 species mainly in tropical S. America, I. crenata East Asia and North Malesia) 2 species studied: I. dumosa, I. crenata. itself considerable of the from Although I. crenata shows a degree variation, two species different geographical regions resemble each other. The material studied is, however, too all character limited to conclude whether species of Polyphyllae are characterized by the combination shared by I. crenata and I. dumosa. At least the occurrence of cork warts (cf. of Loesener's accounts species with 'unpunktierte Blatter', 1942) and of a lignified abaxial epidermis (cf. variability of I. crenata) is not constant within this section. Other features and for which the section appears to be heterogeneous are subdivision mucilaginous the adaxial and the bundle sheaths. Loesener's nature of epidermis lignification of sug- gestion (1908) about possible links with section Crassifoliae of series Lioprinus could be supported because a lignified epidermis and frequent cork warts have also been found in I. retusa of that section. Section Vacciniifoliae (c. 26 species in S. America, Malesia, and Ceylon) studied: 2 species I. hanceana, I. zygophylla. The rather leaf of hanceana does allow dis- 'unspecialized' anatomy I. not systematic cussions of relevance for this widely distributed section. I. zygophylla is entirely different and be closer series A than this section and series. may to subgenus Byronia to Series Paltoria as a whole Only 4 species from two sections have been studied of this series. Inclusion of I. zygo- Paltoria phylla makes series anatomically very heterogeneous. Series c. AQUIFOLIUM Asia Section Lemurenses (c. 12 species; 1 in Africa, the remainder in East and Indo- Malesia) 2 species studied: I. fargesii, I. mitis. According to Loesener (1908: 57) the two species used for study belong to the relatively advanced of this which said intermediate fase between species section, is to represent an series series section Lioprinus and Aquifolium. Anatomically links are indicated with Aquifolioides and Microdontae subsection Repandae. Within the small sample of species the anatomical the restricted to studied, range is quite narrow; diversity is presnece or absence of foliar sclereids and of and the of of the bicyclic stomata, to degree curving vascular bundle in the petiole. 394 BLUMEA VOL. 22, No. 3, 1975 Section Aquifolioides (c. 21 species, for distribution see under subsections) Subsection and Oxyodontae (c. 14 species in Macaronesia, N. Africa, Eurasia) 7 species studied + 1 hybrid: I. altaclarensis (hybrid), I. aquifolium, I. bioritsensis, I. cornuta, I. dipyrena, I. integra, I. perado, I. pernyi. Subsection I n s i g n i s (c. 7 species in Indo-Malesia and East Asia) studied: 3 species I. insignis, I. latifolia, I. sikkimensis. Most of the species of section Aquifolioides share a highly characteristic complex ofleaf anatomical features. Except I. sikkimensis all species share at least partly bicyclic stomata, feature marginal sclerenchyma strands or a tendency towards this (I. insignis, I. integra, and I. latifolia), large stomata (27x27 to 38 x indumentum absent or composed of short unicellular hairs only, a simple open vascular system of the petiole of type 2 or 4, and the absence oflignification in the abaxial epidermis and mesophyll, and offrequent cork warts. the with bundle sheath. sikkimensis is I. insignis is only species a lignified I. aberrant with its cyclocytic to anomocytic stomata, and absence of marginal sclerenchyma strands. The anatomical homogeneity of section Aquifolioides (excluding I. sikkimensis) is very for of There exceptional Ilex, and strongly supports the naturalness of this group species. links and with are leaf anatomical with section Lemurenses through I. fargesii a number of section Microdontae This species from subsection Repandae. agrees remarkably well with Loesener's ideas about mutual relationships of those sections. His statement (1908: 59) that section level within sub- Aquifolioides represents a highly specialized evolutionary could be used to the rather beliefthat within Ilex characters genus Euilex, support intuitive like bicyclic stomata, marginal sclerenchyma strands and hypodermal development anatomical I. sikkimensis link with less represent specializations. might represent a special- ized the leaf not distinctive to indicate groups, although anatomy is enough a particular section. Section Microdontae for distribution under (c. 50 species, see subsections) Subsection Eumicrodontae (c. 6 species in tropical America) studied: microdonta. 3 species I. krugiana, I. brevicuspis, I. Subsection and and Asia) Repandae (c. 35 species in tropical subtropical America 10 species studied: I. argentina, I. buergeri, I. confertiflora, I. curranii, I. ficoidea, I. formosana, I. glomerata, I. paraguariensis, I. repanda, I. tolucana. Subsection Vomitoriae (1 species in N. America) 1 species studied: I. caroliniana. Subsection Stigmatophorae (c. 2 species in tropical and subtropical Asia) 1 species studied: I. triflora. Subsection Sideroxyloides (c. 6 species in East Asia and tropical America) studied: 3 species I. divaricata, I. goshiensis, I. sideroxyloides. The anatomical in section Microdontae is in the range very considerable, not only whole section as a but also in the individual subsections of which more than one species was studied. consist In subsection Repandae an indumentum may be absent, of short unicellular tolucana be broad-based hairs. hairs, or as in I. represented by longer unicellular Stomata least in are always at partly cyclocytic but in some species bicyclic stomata are present various in others in cell amounts, anisocytic stomata are present addition. The guard in size from to x 28 Regular cork warts are in one pairs range 23 X23 38 //m. present species only (I. paraguariensis), the adaxial epidermis is usually unmodified but has pericli- nal subdivisions with mucilaginous daughter cells in two species. A hypodermis is never P. Baas: Anatomy ofAqiiifoliaceae etc. 395 midrib. fully developed; only some species show local hypodermal differentiation in the I. argentina has lignified abaxial epidermal cells, and in I. paraguariensis these cells have thick walls. The bundle sheaths and tissue of the very unlignified spongy mesophyll may be not. The widi a vascular lignified or petiole is usually supplied simple open system closed in I. I. (type 3 or 4) but a cylinder occurs repanda and paraguariensis. Marginal but five sclerenchyma strands are usually absent, species are characterized by idioblastic foliar sclereids. The species of subsection Eumicrodontae are more or less within the leaf anatomical have range of subsection Repandae. Cork warts, a lignified epidermis, or foliar sclereids not been encountered in the species studied. Here too the stomata are always at least also partly cyclocytic, but in one species (I. krugiana) bicyclic stomata occur, and in I. in size microdonta anisocytic stomata are present addition. Stomatal ranges from 26x21 to The latter also shows local hypodermal development in the midrib 30X 25 //m. species region. I. caroliniana of subsection Vomitoriae with its frequent cork warts and lack of wing bundles in the petiole is not very close to any particular species of subsection Repandae, of characters somewhere subsection. although most its are present in that I. triflora of subsection Stigmatophorae with its combination oflignified abaxial epidermis and regular cork warts is unique amongst the species studied of this section. small of Subsection Sideroxyloides is anatomically heterogeneous, even in the sample studied. Stomatal absence of cork and hairs are the 3 species type, presence or warts main here. section is variables Through I. sideroxyloides the anatomical range of Microdontae again somewhat extended because stomata are predominantly anisocytic in this species. The leaf anatomical heterogeneity ofsection Microdontae makes it impossible to indicate affinities with other sections the for whole group of species studied. Leaf anatomy only a of 6 from subsection provides strong arguments to regard group species Repandae as related. This characterized the of foliar mutually closely group is by presence sclereids, cyclocytic or cyclocytic to bicyclic stomata andnumerous other less outstanding anatomical features. I. buergeri, I. confertiflora, I. curranii,I. ficoidea, I. formosana, and I. glomerata belong to this These with I. mitis section Lemurensis and group. species have anatomical links from with the the latter species from section Aquifolioides. It is noteworthy that species which are mainly provided with marginal sclerenchyma strands should share many features with species having another device to increase the scleromorphic nature of their leaves: foliar sclereids. I. cornuta and I. confertiflora are anatomically intermediate between these two Affinities of section section Lemurensis and section groups. Microdontae with Aquifolioides are well in line with Loesener's ideas about the derivation of this group. The anatomical For heterogeneity also implies that links between several other sections can be suggested. I. microdonta of subsection Eumicrodontae this would be section Vacciniifoliae and section Excelsae. For I. argentina of subsection Repandae affinities with section Cassinoides could be suggested because of resemblance with I. canariensis of that section. Anatomies similar to that of I. caroliniana of subsection Vomitoriae occur in section Cassinoides and section of with certain Excelsae. Ilex triflora subsection Stigmatophorae shares many leaf characters species of section Thyrsiflorae. Finally affinities could be suggested of I. divaricata of sub- section Sideroxyloides with sections Cassinoides, Micranthae, Megalae, and Excelsae. How- ever, since many of these similarities are based on rather indistinctive leaf anatomies, the value of these similarities for judging the natural affinities is probably highly restricted. BLUMEA VOL. No. 396 22, 3, 1975 Section Prinifoliae (1 species in East Asia) I. i species studied: pubescens. of this of of The anatomy mesomorphic species is suggestive that subgenus Prinus for difference stomatal except a in type. Section Megalae (c. 13 species in tropical S. America) studied: theezans. 3 species I. brasiliensis, I. integerrima, I. in leaf of in and The only variables the anatomy of this group species are stomatal type indumentum. I. brasiliensis stands both these characters with hairs and out in long septate cyclocytic to anomocytic stomata, instead of cyclocytic to anisocytic stomata occurring in the other two species. The characters shared by all three species are of very common within that affinities with could be advocated without occurrence Ilex, so many groups affinities section. really contributing much to our understanding of the wider of this Micranthae Section (c. 27 species in tropical America) studied: umbellata. 5 species I. danielis, I. guianensis, I. inundata, I. jenmanii, I. this if The variation within section is quite small. An indumentum, present, always with consists of short unicellular hairs; the stomata are usually cyclocytic to anisocytic, I. danielis as the only exception with cyclocytic to anomocytic stomata; stomatal size from cork 18x16 (I. umbellata) to x regular warts are absent; the ranges 27 23 pm; epi- dermis is usually subdivided and is at least always partly mucilaginous. The abaxial and bundle sheaths The vasculature is epidermis are never lignified. petiole simple open, and sclereids and marginal sclerenchyma strands are always absent. As in section Megalae this anatomical character complex is rather unspecialized and of common occurrence throughout Ilex. Section Rugosae (c. 5 species in Indo-Malesia and East Asia) 3 species studied: I. archboldiana, I. revoluta, I. versteeghii. Of three revoluta known and these species, only I. was by Loesener. I. versteeghii I. archboldiana here of Merrill are treated because suggestions by & Perry (1941). The two latter but revoluta deviates in cork species are very similar, I. having frequent warts, an indumentum of moderately long hairs, and non-lignified spongy tissue. affinities For this species no suggestions about can be given. For the other two species there links with and sections Euilex. These links are subgenus Byronia several of subgenus without characteristic and therefore rest on a scleromorphic anatomy very features, may be of little systematic significance. Series Aquifolium as a whole leaf This large infrageneric taxon with over 100 species shows a tremendous anatomical range in the sample of 44 species studied. Foliar sclereids and marginal sclerenchyma have been strands only found in species of this series, though they are absent from the them. Not covered majority of the entire anatomical range of Ilex is in series Aquifolium. Complex vascular systems and a mucilaginous hypodermis never occur. Bicyclic stomata show this are almost entirely confined to this series, but only three species outside it whilst of to Microdontaeand section feature, 14 species series Aquifolium (belonging section Aquifolioides) have at least partly bicyclic stomata. Within the series two clearly defined with anatomical groups can be distinguished: one marginal sclerenchyma strands and another with foliar sclereids. These two groups are in other characters anatomically time rather similar, and are at the same linked through intermediates. They probably P. Baas: Anatomy of Aquifoliaceae etc. 397 the levels ofleaf anatomical specialization in Ilex. The majority of species represent highest showing distinctive anatomi- presents a rather heterogeneous assemblage, only infrequently cal features. Series D. THYRSOPRINUS Section Indico-Malaicae (c. 18 species in Malesia, Ceylon, and Papuasia) I. I. cissoidea, I. clemensiae, I. pulogensis, I. scrabidula, I. spicata, 7 species studied: brunnea, I. stapfiana. and I. stapfiana are in- In Loesener's system only I. cissoidea, I. pulogensis, I. spicata, here because of of their eluded; the other three species are treated suggestions affinity by section authors. 'Punktierte Blatter undBlatter ohne Punkte' both occur within this according reflected the absence of frequent to Loesener (1942); this is in variability in presence or regular cork warts recorded here. The former is follows: indumentum absent or anatomical range of the four species as stomata cyclocytic to anisocytic, consisting of broad-based hairs (I. cissoidea), cyclocytic, stomatal size from cyclocyticto anomocytic, or cyclocytic to bicyclic (I. pulogensis), ranges cork warts frequent or absent, adaxialepidermis unmodified, 26 x 21 to 32 X 23 fim,regular with cells. Petiole with containing mucilage cells, or subdivided mucilaginous daughter and or with (I. complex closed vascular system (I. spicata I. stapfiana) simple open system but sometimes cissoidea and I. pulogensis).. The bundle sheaths are usually lignified, very as a or marginal sclerenchyma faintly so. Unusual features such hypodermis, sclereids, other I. clemensiae, would strands are absent. The inclusion of the three species, especially of short unicellular hairs is extend the anatomical range to some extent: an indumentum and the stomatal would to X (I. clemensiae), present in I. clemensiae, size range up 38 29fim also include the clemensiae). vascularization of the petiole would unusual type 5 (I. Regardless of the inclusion of the latter species in this section, Indico-Malaicae remains leaf with stomatal cork anatomically a rather heterogeneous group, indumentum, type, vascularization the main variables. Loesener's that this warts, and petiole as hypothesis be supported section, together with subgenus Byronia, is derived from a common stock can of vascularization of 8 are entirely restricted by leaf anatomy. Complex types petiole type and of with to subgenus Byronia section Indico-Malaicae. I. oppositifolia subgenus Rybonia, from the ancestors. Moreover, a similar midrib vascularization, could be derived same Indico-Malaicae share their overall leaf with several some species of section anatomy also exist with species of subgenus Byronia. It should, however, be stressed that similarities affinities with several species of subgenus Euilex, opening possibilities to advocate numer- ous sections. Section Thyrsiflorae (c. 10 species in tropical S. America) 2 species studied: I. martiniana, I. ovalifolia. similar with the characters shared: These two species are anatomically very following indumentum absent, stomata aniso- to cyclocytic, regular cork warts frequent, abaxial bundle epidermis and bundle sheaths lignified, vascular pattern in petiole simple open, from and unusual features absent. The remaining diversity: stomatal size ranging 26 x 23 adaxial unmodified mucilaginous and with sub- to 33 X 28/im, epidermis or periclinal divisions, is quite small in comparison with the overall similarities. this section and the section Loesener (1908) suggested a different ancestry for following (Brachythyrsae) than for section Indico-Malaicae of the same series. Support for this sugges- tion be advanced leaf anatomical character: abaxial cells. may using a lignified epidermal 'at the base' of section This feature also occurs in section Crassifoliae, which was put 398 BLUMEA VOL. 22, No. 3, 1975 Thyrsiflorae by Loesener. In other sections of series A. Lioprinus, which according to Loesener's ‘Stammbaum’ are even more primitive, the bgnified epidermis also occurs sporadically (in section Dasyneurae and section Cassinoides). Outside series D and A, ligni- Microdontae fied abaxial epidermides are only found in two species of section (series C), section one species of Rugosae (series C), and in section Polyphyllae (series B); the latter section is be sections series D Loesener. regarded to very close to some of by Within series D a lignified epidermis characterizes also the only species studied of section Brachy- thyrsae. The fact that within this of the unusual anatomical feature like the part genus Ilex, an occurrence of a lignified epidermis is strongly correlated with Loesener's ideas of classi- fication as based on macromorphology is suggestive of a considerable systematic value of this anatomical character. This is also supported by Cador's observations on some other of species the same or closely related sections showing a lignified abaxial epidermis. Whether this has also implications for the systematic position of those species from series until of the C having a lignified epidermis, remains uncertain a new taxonomic revision whole is carried out. genus Section Brachythyrsae (c. 6 species in Brazil) studied: i species I. amara. With its frequent cork warts and lignified epidermis this species recalls those of the previous section. The main difference is the predominance of anomocytic stomata in The resemblance I. amara. anatomical of I. amara with species from section Thyrsiflorae and of sections and in full with Loesener's Polyphyllae Crassifoliae are agreement suggestions about affinities (see above). Series Thyrsoprinus as a whole From Loesener's derivation of the inflorescences in this series and his phylogenetic tree (fig. 31) follows a polyphyletic origin for the different sections of this series. Only a limited number of of of the has been studied representatives part sections anatomically, but the leaf be used this Indico-Malaicae anatomy may to support hypothesis. Section shows links with subgenus Byronia, and the other sections witness affinities with sections Crassifoliae, Polyphyllae, Cassinoides, and Dasyneurae of series Lioprinus and series Paltoria of subgenus Euilex. As with most series treated before, the anatomical diversity of Thyrso- is prinus very great. Subgenus Euilex as a whole number This is Loesener's largest subgenus, numbering about230 species out of the total of Ilex therefore that the 271 species recognized in 1908. It is not surprising to note ana- this tomical range of subgenus virtually covers the whole range of the genus. The only character with base not occurring within this subgenus is the unusual hair type septate as occurring in I. hypoglauca of subgenus Byronia. If, however, Loesener's placing of I. zygophylla in section Vacciniifoliae would be correct, hairs resembling the 'hypoglauca' in The of Euilex type would also occur subgenus Euilex. anatomical diversity subgenus cannot be used to verify Loesener's classification and ideas of affinities. According to his derivations of several character complexes, the different series and sections present many diverse external and this well be specializations in morphology, may brought in agreement with anatomical The anatomical of several a great diversity. heterogeneity sections is, and raises however, more problematic, the question whether their arrangement and delimitation really reflects natural affinities. P. Baas: Anatomy of Aquifoliaceae etc. 399 Subgenus PRINUS Series A. EUPRINUS (C. 6 species in N. America and East Asia) 2 species studied: I. serrata, I. verticillata. Thin mesomorphic leaves and strongly undulating 'interlocking' anticlinal walls in the abaxial epidermis with anomocytic stomata, but without other outstanding anatomical studied. there anatomical features characterize the two species Yet is some heterogeneity stomatal absence of in indumentum, size (24x17 & 35x21 /iin), presence or mucilage cells the in the epidermis and presence or absence of wing bundles in petiole. There are with several of and also with obvious links species series B. Prinoides, some species with a leaf from other from mesomorphic anatomy subgenera, notably I. micrococca subgenus Byrottia. Series B. PRINOIDES (c. 10 species in N. America, East Asia, and India) studied: I. 5 species I. amelanchier, I. asprella, I. decidua, I. dubia, longipes. The of this share characters with those of The species series many series Euprinus. ana- is indumentum be absent is tomical heterogeneity quite restricted: an may but usually well-developed and consists of long septate hairs or non-septate broad-based hairs; the stomata or to cyclocytic with are usually anomocytic anomocytic very rarely anisocytic addition stomatal from the stomata in (I. decidua); size ranges 20x19 to 37x29/101; adaxial epidermis always contains mucilage cells, and frequently shows periclinal sub- such cork divisions; petiole vasculature is always simple open. Features as warts, sclereids, lignified epidermis, marginal sclerenchyma strands, lignified spongy tissue, and lignified from all these The bundle sheaths are absent mesomorphic species. interlocking undula- tions of the abaxial anticlinal epidermis walls in I. longipes recall the situation in series Besides similaritieswith with Euprinus. the latter series, there are anatomical resemblances I. micrococca (of subgenus Byronia) and I. pubescens and I. brevicuspis (of subgenus Euilex). All also characterized a leaf that the these species are by mesomorphic anatomy, so anatom- of ical similarities may reflect similar autecologies the species involved (in casu deciduous habit ?) rather than true affinity. whole Subgenus Prinus as a Through its deciduous mesomorphic leaves, this subgenus constitutes anatomically a There is for fairly homogeneous assemblage. a strong tendency anomocytic stomata to be and all other characters indumentum the predominant type, except the tend to be rather similar 'unspecialized'. Although there are some species in the other subgenera with a leaf would be this favour of mutual anatomy, it unjustified to use as a strong argument in affinity since most of these characters are a direct consequence of the mesomorphic habitof the foliage in these species. As Loesener's suggestions of affinities with other species also largely rest on this mesomorphic appearence as a macromorphological feature, the use of for additional of the leaf anatomical characters arguments wouldimply an overweighting the same character complex. Taxonomicconclusions The leaf anatomical heterogeneity of most infrageneric taxa recognized by Loesener raises die question about the naturalness of his classification. All subgenera show some or much in leaf this also overlapping their anatomical range with each other, and holds for most series, sections, and subsections. As similar situation in wood anatomy (Baas, 1973) could not be used to challenge BLUMEA VOL. No. 400 22, 3, 1975 because latitudinal and altitudinal trends strongly of Loesener's system, were suggestive factors in of wood anatomical ecological as directing agents the evolution diversity. also be correlated with but since Ilex Leaf anatomy in general may ecological factors, the shows different possibilities to arrive at xeromorphic or scleromorphic structures, still be witness ofnatural affinities (see 404). However, the actual structures present may p. correlation between leaf and Loesener's number of examples in which there is a anatomy to make alternative of Ilex based on system is very limited. Attempts groupings species, combinations ofanatomical characters have failed, with the few exceptions ofsmall species with characterized rather distinctive leaf anatomies the species marginal groups by e.g. sclerenchyma strands and/or foliar sclereids. distinctive anatomical characters do not The numerous Ilex species without provide with suggestions about natural affinities, and the distribution of the remaining species features of distinctive characters (special petiole types, hypodermal development, special of the different the epidermis, stomatal types infrequent occurrence etc.) over infrageneric of much more of than in taxa is suggestive a complex pattern relationships suggested fact also what would when Loesener's formal subdivision. This is in one suspect analyzing characterizations of his taxa. Loesener's rather vague infrageneric Infraspecific variability in 15 species of Ilex than studied. The number of Of the following 15 species, more one specimen was bioritsen- specimens studied is indicated betweenbrackets: I. anomala (4); I. aquifolium (4); I. I. sis (2); I. canariensis (3); I. crenata (11); I. cymosa (7); I. ficoidea (a); I. guianensis (3); hypo- I. theezans (2); I. triflora (6). glauca (5); I. opaca (3); I. perado (3); I. pernyi (2); I. serrata (2); studied for certain anatomical In some cases infraspecific variability was only characters, leaf of macerations were without considering the complete anatomy: e.g. I. crenata only of abaxial studied to check the lignification of the cell walls the epidermis. the Although the number of specimens studied for testing the diagnostic value of the results indicate of variability anatomical characters was very limited, a high degree characters. below the species level of most limits be Lamina thickness varies between wide or narrow as can seen in the descriptions of the species tested. The absence of indumentum is variable in four tested. This presence or an species hairs. with a distinctive always concerns species with only sparse unicellular Species more hair constant in this type are respect. smooth of the cuticular The texture of the cuticle, caused by a granular or structure layer (cf. Van Staveren & Baas, 1973) is variable in two of the species tested. Cuticular in five striations appeared to be variablein distinctness or even presence or absence species. is recorded in the Cuticular thickness always varies to a small or large extent, and specific descriptions. walls Theoutline of the unspecialized epidermal cells, as affected by the anticlinal being undulated variable of the tested. The variation straight, curved, or is in seven species entire in Ilex: viz. curved to usually covers only part of the range straight to curved, or undulated. Whether the anticlinal cuticular flanges are pitted or not was found to be variable in five species. The shape of the epidermal cells overlying the midrib (flattened, The venation of the leaves square, or elongate) is variable in eight species. pattern usually to which this causes little or no modification of the epidermal cell pattern; yet the degree four of the tested. The of thickened lignified cell walls occurs varies in species occurrence fewer in the abaxial epidermis is limited to few species of Ilex only, and of even species P. Baas: Anatomy of Aquifoliaceae etc. 401 than studied: in I. and I. cells more one specimen was opaca triflora lignified epidermal are of constant occurrence, although in onespecimen ofI. triflora (HallierB 2471) only a minor of cells found proportion the unspecialized is lignified; I. canadensis and I. crenata were to be variable for this character. is in The stomatal type also variable seven of the species tested, but this only concerns a the of different the leaf. for variability in proportion types occurring in same In instance show and and I. triflora some specimens predominantly anisocytic infrequent cyclocytic intermediate stomata, and other specimens have predominantly cyclocytic stomata with infrequent anisocytic ones. In several other species the ratio of anisocytic and cyclocytic stomata is also variable. Species with cyclocytic and/or bicyclic stomata are usually in stomatal See also the Stomatal size variations be constant type. specific descriptions. may anomala considerable, notably in I. in which it amounts to 31% of the average maximum value. ratios of the cell also albeit within Length/width guard pairs vary, usually fairly limits. an with about from narrow I. perado is exception 13% variation: 0.93 to 1.07. in and if Polar T-pieces (see Stace, 1965) are never very conspicuous Ilex, sufficiently distinct their value is limited because four found in this diagnostic species were to vary Peristomal also somewhat respect. rims, although virtually always present in Ilex, vary in distinctness in four species. Cork warts, if fairly frequent and of regular outline, are diagnostic for the few species absence and of (3) tested. Presence or frequency irregularly shaped, probably traumatic, cork warts (cf. Herridge, 1963) is always variable. The degree ofsubdivision of the adaxial epidermis in those species which tend to form variable of the a multiple, largely mucilaginous, epidermis is highly in six species tested. The the cells also these The degree of bulging of mucilaginous varies in species. presence of and type hypodermis is constant in the species tested. all tested the bundle sheaths with their cell walls In species are constant respect to being obvious lignified or not. However, in some species it is not in safranin/haematoxylin stained sections whether the bundle sheaths which limits the are lignified or not, applica- this character The bility of for diagnostic purposes. same applies to lignification of(part of) the the of lignified tissue varies in In I. spongy tissue. Here proportion some species. cymosa the cell walls of the spongy tissue may be lignified or not. In doubtful cases the and recorded be bundle sheaths spongy tissue are to strongly birefringent. It should be realized that unlignified cell walls also show birefringence, but secondary lignified wall thickenings are usually much more strongly birefringent. It is therefore likely that has occurred in the the lignification more strongly birefringent walls, although staining did not reveal this clearly. Wiesner's reaction (phloroglucinol and hydrochloric acid) has been used for the absence of only sporadically testing presence or lignin. The midrib in Ilex, as seen in transverse sections, is usually grooved but in a fair number of species it is flattened to raised. However, in three of the species tested this is variable. Vascularization of the midrib, but especially of the petiole varies in four of the species tested. This and whereboth is particularly striking in I. hypoglauca I. cymosa, simple open the This and highly complex types occur in same species. variability is not to be confused with the differences between the basal of the has existing part petiole (which always a simple system, if sectioned close enough to the place of attachment to the stem) and the distal ofthe part petiole. Foliar sclereids are ofrare occurrence in Ilex as a whole, and their diagnostic value has been 'tested' in of which both studied them. only one species (I. ficoidea), specimens possess Marginal sclerenchyma strands were also found to be of constant occurrence in the four species tested. However, in the 'intermediate' species (I. confertiflora, I. insignis, and I. their the leaf of leaf. latifolia) appearance varies along margin a single 402 BLUMEA VOL. 22, No. 3, 1975 Frequency of crystals is variable in most species tested. In only one specimen of I. all. the latter need theezans no crystals were found at Probably observation would cor- rection if a complete series of sections from leaf tip to petiole were made. The considerable variation below the species level in many leafanatomical characters is than the variations of and leaf even stronger infraspecific reported epidermal histological characters (Jansen & Baas, 1973; Bongers, 1973; Van Staveren & Baas, 1973; Van Vliet & of and Baas, 1975) in species Celastraceae, Winteraceae, Icacinaceae, Crypterotiiaceae respec- tively, although most of the characters discussed above appeared to show some degree of variation below the species level in these families as well. It should be stressed that in of Ilex probably an even higher degree leaf anatomical infraspecific variability exists than reported above. The numberof specimens studied was very limited for most species, and the number of variable characters for with more specimens will no doubt increase each species. be due The infraspecific variability may to genetical variability or ecological factors. Yet it is likely that within Ilex species two other factors will also have contributed to the first the of the variational range. In the place this may be age persistent leaves. Although leaves been selected for this full-grown have study, no data onthe actual age wereavailable, it is known that several leaf anatomical characters the life of and may change during span and persistent leaves (see Napp-Zinn, 1974, literature cited by him). Cuticular thickness, of the of degree ofsubdivision epidermis, lignification epidermis and mesophyll may have this. the lack of Ilex been affected by In second place, the a recent taxonomic revision for implies the possibility of doubtful specific delimitation and erroneous identifications, well of the variation recorded below which may underly part the species level. However, of of in a number cases the anatomical evidence almost excludes the possibility mis- characteristic identifications, because the species involved show a highly anatomical combination of characters: feature or I. anomala, I. aquifolium, I. bioritsensis, I. ficoidea, and small of I. hypoglauca, I. opaca, I. perado, I. pernyi, I. triflora. Even in this group nine of several anatomical species the infraspecific variability characters is very considerable. for In those species studied which infraspecific taxa are recognized no correlation has been found between anatomical variation and infraspecific classification. The above results made it hardly useful to document all leaf anatomical characters in I therefore chosen form of elaborate specific descriptions. have the very condensed which description in data on cuticular granulation and striations, outline of the epidermal cells, etc. have usually been left out altogether. Latitudinal and altitudinal trends; ecological and functionalanatomy Stimulated the results of studies wood anatomical by previous on variation (Baas, 1973; Van der Graaff & Baas, 1974) I have also searched for correlations between several leaf anatomical characters and latitude and altitude of provenance. For 76 of the 95 species collection data sufficient do this. studied, were to Four groups were distinguished: tropical mountain 6 from high species (altitude over 2500 m; species); tropical mountain species altitudes between 1000 and m lowland (altitude below 2500 (12 species); tropical species and and 1000 m; 18 species), subtropical temperate species (40 species). All four to be rather with to most groups appeared heterogeneous respect quantitative and qualitative leaf anatomical characters. Exceptions to this lack of altitudinal and latitudinal trends are the following: High mountain species in the tropics always have thick leaves, and mostly also thick P. Baas: Anatomy ofAquifoliateae etc. 403 cuticles. Their anticlinal epidermal cell walls are usually straight to curved, and rarely undulated. The percentage of species with undulated walls increases with decreasing for altitudes in the tropics: c. 17% high mountain habitats; c. 45% for mountain habitats between and and for the lowland 1000 2500 m altitude; c. 60% tropical (cf. Bongers, 1973, for results the and comparable in Drimyspiperita). In temperate subtropical regions c. 60% of the species have undulated anticlinal epidermal cell walls. bundle sheaths and ofmuch more Lignified lignified spongy tissue are frequent occur- less of and rence in the tropics (more or independent altitude) thanin temperate subtropical there difference between the World and Old regions. Remarkably enough is a New World this About of Ilex from and the tropics in respect. 65% 33 species tropical Asia have bundle in America of Pacific lignified sheaths; tropical only c. 27% 18 species show this character. and the For temperate subtropical species percentage is c. 10. Comparable for percentages lignified spongy tissue are c. 50, 11, and 9 respectively. of correlations. Of The distribution stomatal types also shows some peculiar the 22 the these species withbicyclic stomataonly 3 are from tropics, and are restricted to Malesian mountain The other from and with regions. 19 are mainly subtropical temperateEurasia, and the other the exception ofI. argentina, I. krugiana, I. opaca. On hand, anisocytic stomata are much more in evidence in the tropics: of the tropical species studied 12 show predomi- If also nantly anisocytic stomata, against 2 in subtropical and temperateregions. we include the with to stomata that of all species anisocytic cyclocytic it appears c. 50% tropical have ofother species studied at least partly anisocytic stomata (independent geographical for the and and factors) against c. 10% subtropical temperate species. For cyclocytic anomocytic stomata there are no latitudinal trends. Other characters like stomatal size and many qualitative characters which are limited to of Ilex correlated with a minority species only, are not clearly latitude or altitude of provenance. of the trends the It is not possible to give a satisfactory explanation in percentages recorded above. In those cases where the latitudinal trends are due to the frequent occur- rence of the characters in a special geographical area only (lignified bundle sheaths and I that these the of restricted mesophyll, bicyclic stomata), suggest are only consequence distributions of groups of closely related species; this implies that these trends are of no considerations but the the importance in ecological only stress significance of feature Ilex. trends involved as a systematic character within For the remaining any explanation is completely lacking. Latitude and altitude are of course very poor and rough indicators of the ecological and labels conditions of the species involved. Data from the literature on herbarium were insufficient lack information for unfortunately to overcome this of ecological most species studied. Nevertheless the in leaf in Ilex has induced diversity anatomy me to give some considerations about and functional general ecological aspects. Ilex anatomical which For many leaf characters could be listed are traditionally be- lieved to be of importance for drought resistance of the plants involved. With the ex- ception of the deciduous species, most Ilex species show such structural details in their leaves. Three main be in this categories can distinguished respect: in 1. Structures to prevent the loss of water. Examples Ilex are: thick cuticle andpericlinal cell stomatal outer epidermal walls, and conspicuous outer ledges providing a sheltered stomatal front cavity. 2. Water storage tissue. Examples are: hypodermal development (different types in Ilex), and specialization of certain epidermal cells to large mucilage cells. 404 BLUMEA VOL. 22, No. 3, 1975 in 3. Mechanical reinforcement (scleromorphy) order to prevent damage to protoplasts in case of wilting. Examples of these are manyfold: thick cuticle; thick epidermal and/or hypodermal cell walls; extra development of sclerenchyma adjacent to vascular bundles; foliar sclereids; marginal sclerenchyma strands; lignified epidermal cells; lignified bundle sheaths; and finally lignified spongy tissue. combination of all characters Not a single Ilex species shows a mentioned above. the have been Usually only few of characters realized in a single species. It should be stressed that the functions attributed to several anatomical features listed above, have only been assumed, and although generally accepted and logical (cf. Haber- landt, 1918, and Burstrom & Odhnoff, 1963), no factual evidence has been presented that survive. it is these structures are necessary for the species to In this respect of interest to Malesian note that many highly scleromorphic features occur in everwet lowland rain forest species or in species from montane mossy forests, hi temperate evergreen species features have obvious function: of loss xeromorphic seem to a more prevention water in times of reduced water uptake at low temperatures (Windslow & Havis, 1967). It is well-known that several xeromorphic (including scleromorphic) features are induced by conditions which from ecological are different drought, such as light intensity (Martin & Juniper, 1970) and mineral deficiency of the soil (Loveless, 1961). characters realized the Ilex In view of such facts the diverse xeromorphic in genus should not be interpreted as adaptive characters. For taxonomic studies these characters may have the same value as any other leaf anatomical characters. ACKNOWLEDGEMENTS I wish to record to Prof. Dr. C. G. G. who the my appreciation my supervisor J. van Steenis, suggested of this and has subject study, who always been a stimulating influence. The very initial stages of this often in Kew in I interrupted study were carried out the Jodrell Laboratory at 1969, and feel much indebted to Dr. C. R. Metcalfe and Mr. F. Richardson who have always generously shared their vast anatomical Miss M. knowledge, indispensable for this type ofwork. Gregory ofthe same department kindly provided references. me with bibliographical institutions in H. S. The curators ofthe mentioned the text, and Dr. McKee are gratefully acknowledged for providing valuable materials for this study. thanks are due Ger van who did of the work nodal Special to Mr. Vliet, most on anatomy reported in this Miss A. M. Kuenen and Mr. W. Star remembered here for their share in the microtechnical paper. are work. Mr. in Madame D. Callen-Lobreau (Arboretum de Chevreloup) and J. Muller were helpful discussing pollen morphological data. The scanning electron micrographs were mostly taken at the Leiden Institute for Geology and Mineralogy Mr. H. those of the wood of which executed Dr. M. by Kammeraat, except Phelline, were by Miss S. Jutte time Mr. C. L. Marks Mr. B. N. Mr. H. at the at Imperial College (London). and Kieft prepared the plates. J. executed of the Zoelen van Os en Mr. E. Vijsma most drawings. Miss M. van diligently typed the manu- script. Prof. Van Steenis and Prof. Dr. W. K. H. Karstens critically read the manuscript. Their help and Dr. W. in Vink's patient and careful editorial work have greatly reduced the number of mistakes this paper. REFERENCES (not including those cited under 'specific descriptions') ASH R. P. into mid vein and of certain of WORTH, 1963. Investigations anatomy ontogeny species the genus Ilex. J. Elisha Mitchell Sc. Soc. 79: 126—138. BAAS, P. 1970. Anatomical contributions to plant taxonomy I. Floral and vegetative anatomy of Eliaea from Madagascar and Cratoxylum from Indo-Malesia (Guttiferae). Blumea 18: 369—391. — Kostermansia 8: 1972a. The vegetative anatomy of malayana Soegeng. Reinwardtia 335—344. P. Baas: Anatomy of Aquifoliaceae etc. 405 to II. of and — 1972b. Anatomical contributions plant taxonomy The affinities Hua Pierre Afrostyrax Perkins et Gilg. Blumea 20: 161—192. in and and —1973. The wood anatomical range Ilex (Aquifoliaceae) its ecological phylogenetic significance. Blumea 21: 193—258. in Icacinaceae. outside Malesia. Acta Bot. — 1974. Stomatal types Additional observations on genera Neerl. 23: 193—200. of Caledonia and of New BAILEY, I. W. 1956. The relationship between Sphenostemon New Nouhuysia Guinea. J. Am. Arb. 37: 360—365. of Icacinaceae I—IV. Arn. Arb. 22: — & R. A. HOWARD. 1941. The comparative morphology the J. 125—132; 171—187; 432—442; 556—568. — of the Winteraceae IV. Anatomy of the node and & C. G. NAST. 1944. The comparative morphology vascularization of the leaf. J. Arn. Arb. 25: 215 —221. — and Nouhuysia. Arn. Arb. & B. G. L. SWAMY. 1953. The morphology and relationships of Idenburgia J. 34: 77—87- H. BAILLON, 1875. Sin: le nouveau genre Spenostemon. Bull. Soc. Linn. Paris 1: 53. Paris — Bull. Soc. Linn. 1891. Sur le nouveau genre Oncotheca. 2: 931 —932. — Histoire 1892. des Plantes 11: 219. Pines. Linn. (Bot.) BAKER, E. G. 1921. Plants from New Caledonia and the Isle of Eucryphiaceae. J. Soc. 45: 306. Baillon Steenis. Candollea BERNARDI, L. 1964. La position systeinatique du genre Sphenostemon sensu van 19: 199—205. Blumea BONGERS, J. M. 1973. Epidermal leaf characters of the Winteraceae. 21: 381—411. of BURSTROM, H. G., & C. ODHNOFF. 1963. Vegetative anatomy plants. ihres Gehaltes CADOR, L. 1900. Anatomische Untersuchungen der Mateblatter unter Beriicksichtigung an Thein. Bot. Centralbl. 84: 241—251, 275—283, 309—315, 340—345, 369—374. In: in (Lectures and CARLQUIST, S. 1961. Handling the anomalous genus. Recent Advances Botany Symposia IX Bot. —168. of International Congress, 1959) 1: 165 Southern Appalachians. J. CLARK, R. C. 1974. Ilex collina. A second species of Nemopanthus in the Am. Arb. 55: 435—441. COPELAND, H. F. 1963. Structural notes on Hollies (Ilex aquifolium and Ilex comuta, family Aquifoliaceae). Phytomorph. 13: 455—464. CRONQUIST, A. 1968. The evolution and classification offloweringplants. New Bot. EDWIN, G. 1965. Aquifoliaceae. In: Botany of the Guyana Highlands VI. Mem. York Gard. 12: 124—150. ENGLER, A. (ed. H. MELCHIOR). 1964. Syllabus der Pflanzenfamilien. Ed. 7. ERDTMAN, G. 1952. Pollen morphology and plant taxonomy. Angiosperms. — and Bot. Notis. 1954. Pollen morphology plant taxonomy. 1954: 65. ESAU, K. 1969. The phloem. In: Handbuch der Pflanzenanatomie V, 2. and FRETZ, T. A., & C. W. DUNHAM. 1972. Influence of 3 levels of light intensity on leaf structure, area color difference in American Ilex Ait. Miss Helen. Holly, opaca cv. Phyton 30: 135—140. FRYNS-CLAESSENS, E., & W. VAN COTTHEM. 1973. A new classification of the ontogenetic types of stomata. Bot. Rev. 39: 71—138. GASANOV, A. M. 1969. Anatomie des organes vegitatifs d'llex hyrcana Pojark. Izv. Akad. Nauk azerbajdzh. —28. SSR, biol Nauk 2: 23 mountains: GIBBS, L. S. 1917. A contribution to the phytogeography and flora ofthe Arfak 136. R. SCHLECHTER. Die GILG, E., & 1923. Monimiaceen-GattungIdenburgia. Bot. Jahrb. 58: 244. in GRAAFF, N. A. VAN DER, & P. BAAS. 1974. Wood anatomical variation relation to latitude and altitude. Blumea 22: 101—121. GUILLAUMIN, A. 1938. Observations morphologiques et anatomiques sur le genre Oncotheca. Rev. G6n. Bot. 50: 629—635. HABERLANDT, G. 1918. Physiologische Pflanzenanatomie. Ed. 5. HEINTZELMANN, C. E., & R. A. HOWARD. 1948. The comparativemorphology ofthe Icacinaceae. V. The Amer. Bot. pubescence and the crystals. J. 35: 42—52. Les Etude d'anatomie et de la HERLEMONT, R. H. J. 1951. Tiliacees d'Indochine. d'histologie compares feuille. Thesis Pharmacy 201, Lille. HERR, J. M. 1959. The development of the ovule and megagametophyte in the genus Ilex L. J. Elisha Mitchell Sc. SOC. 75: 107—128. of leaf of HERRIDGE, E. A. 1963. Pathological anatomy spots Holly. Phytopathol. 53: 481—487. S. The Ilex in China. Am. Arb. 348 —387, and —80; Hu, 1949—1950. genus I—V.J. 30: 233 —344; 31: 39 214—263. in the Pacific area Bot. — 1967. The evolution and distribution ofthe species ofAquifoliaceae (1). J. Jap. 42: 13—27- 406 BLUMEA VOL. 22, No. 3, 1975 & HUTCHINSON, J. 1959 & 1974. The families offlowering plants. Ed. 2 3. — of I. 1964. The genera flowering plants. — 1969. Evolution and phylogeny offlowering plants. of the of South Pacific INGLE, H. D., & H. E. DADSWELL. 1961. The anatomy secondary xylem West tree aid their of Tenth Pacific Science species as an to taxonomy. Paper Congress. leaf of Kokoona and JANSEN, W. T., &P. BAAS. 1973. Comparative anatomy Lophopetalum(Celastraceae). Blumea 21: 153—178. JUTTE, S. M. 1958. Hardwoods of Netherlands New Guinea. I. Nova Guinea, new ser. 9: 347—367. H. in Univ. Bot. KENG, 1962. Comparative morphological studies Theaceae. Calif. Publ. 33: 269—384. KORN, R. W., & G. W. FREDERICK. 1973. Development ofD-type stomata in the leaves ofIlex crenata var. Ann. Bot. convexa. 37: 647—656. KRONFELD, M. 1892. Aquifoliaceae. InEngler & Prantl, Natiirliche Pflanzenfamilien ed. 1, III, 5 : 183 —189. C. Nova LAUTERBACH, 1912. Guttiferae. Guinea 8: 843. LOBREAU, D. 1969. Les limites de l'ordre Celastrales d'apres le pollen. Pollen et Spores 11: 499—555. LOBREAU-CALLEN, D. 1973. Le pollen des Icacinaceae. II. Observations en microscopie electronique, Pollen Correlations, Conclusions. et Spores 15: 47—89. — des Celastrales Thesis 1975. Les pollens et groupes apparentes. Montpellier. LOESENER, TH. 1897. Aquifoliaceae. InEngler & Prantl, Natiirliche Pflanzenfamilien ed. 1, Nachtrage zu III, 5: 217—221. I. Kais. Deutsch. Akad. Naturf. — 1901. MonographiaAquifoliacearum Nova Acta Abh. Leop.-Carol. 78. — 1908. Ibid. II. Ibid. 89 (1). In Pflanzenfamilien ed. —86. — 1942. Aquifoliaceae. Engler & Prantl, Natiirliche 2, 20b: 36 LOVELESS, A. R. 1961. A nutritional interpretation of sclerophylly based on differences in the chemical composition of sclerophyllous and mesophytic leaves. Ann. Bot. 25: 168—174. MARTIN, J. T., & B. E. JUNIPER. 1970. The cuticles of plants. MERRILL, E. D., & L. M. PERRY, 1941. Plantae Papuanae Archboldianae VI. J. Am. Arb. 22: 253 —270. the ofthe of stem and METCALFE, C. R. 1956. The taxonomic affinities of Sphenostemonin fight anatomy its leaf. Kew Bull. 1956: 249—253. &L. CHALK. — 1950. Anatomy of the Dicotyledons. Monimiaceae. MONEY,L. L., I. W. BAILEY & B. G. L. SWAMY. 1950. The morphology and relationshipsofthe Am. J. Arb. 31: 372—404. II A In: VIII. NAPP-ZINN, K. 1973 & 1974. Anatomie des Blattes 1 & 2. Handbuch der Pflanzenanatomie. M. T. A of of the PENNINGTON, 1953. comparative study of wood anatomy fifty-four species family Aquifoliaceae. Thesis Univ. Virginia. Doct. Diss. Ser. Publ. No. 7981. W. R. Forst. New PHILIPSON, 1967. Grisefinia fil. Anomaly or fink. Zeal. J. Bot. 5: 134—165. Petiole of the Guttiferae and related families. New York Bot. Gard. SCHOFIELD, E. K. 1968. anatomy Mem. 18: 1—55. SENGLET, A. 1928. La melanogenese chez quelques plantes d'un interet pharmaceutique. Bull. Soc. bot. Geneve, ser. 2, 20: 385 —457. 8. SHAW, H. K. A. 1966 & 1973. In: J. C. Willis, A dictionary of flowering plants and ferns. Ed. 7 & Kew —1965. Diagnoses ofnew families,new names etc. for the seventh edition of Willis's 'Dictionary'. Bull. 18: 249—273. — 1972. Validation of the family name Sphenostemonaceae.Kew Bull. 27: 325 —326. A. Studies I. Trends of nodal vasculature. Ind. Bot. SINGH, 1972. in Euphorbiaceae specialization in J. 51: 350—355- E. of the node aid classification of the Amer. SINNOTT, W. 1914. The anatomy as an in Angiosperms. J. Bot. 1: 303—322. SLEUMER, H. 1942. Icacinaceae. In Engler & Prantl, Natiirliche Pflanzenfamilien ed. 2, 20b: 322 —396. — 1971. Icacinaceae. In Flora Malesiana I, 7: 1—87. notes II. Queensl. SMITH, L. S. 1957. New species of and on Queenslandplants Proc. Roy. Soc. 68: 43—50. SOLEREDER, H. 1899 & 1908. Systematische Anatomie der Dicotyledonen (& Erganzungsband). Mus. STACE, C. A. 1965. Cuticular studies as an aid to plant taxonomy. Bull. Brit. (Nat. Hist.) Bot. 4, I. Bot. STAVEREN, M. G. C. VAN, & P. BAAS. 1973. Epidermal leaf characters of the Malesian Icacinaceae. Acta Neerl. 22: 329—359. STEENIS, C. G. G. J. VAN. 1952. Reduction of two endemic monotypic Papuan genera. Acta Bot. Neerl. I : 94—98. — 1955. Some notes on the flora ofNew Caledonia and reduction of Nouhuysia to Sphenostemon. Svensk Bot. Tidskr. 49: 19—23. STERN, W. L. 1967. Index Xylariorum. Regnum Vegetabile 49. Am. SWAMY, B. G. L., & I. W. BAILEY. 1949. The morphology and relationships of Cercidiphyllum. J. Arb. 30: 187 —211. P. Baas: Anatomy ofAquifoliaceae etc. 407 TAKHTAJAN, A. 1966. Systema et Phylogenia Magnoliophytorum(in Russian). — 1969. Flowering Plants. Origin and dispersal. TH£VENARD, M. 1906. Recherches histologiques sur les Ilicacees. Thesis Paris. Aliso THORNE, R. F. 1968. Synopsis of a putatively phylogenetic classification of the flowering plants. 6: 57—66. VALETON, TH. 1886. Critisch overzicht der Olacineae B. et H. Thesis Groningen. VLIET, G.J. C.M. VAN, &P. BAAS. 1975. Comparative anatomy ofCrypteroniaceaes.l. Blumea22: 173—195. in WALT, J. J. A. VAN DER. 1971. 'n Taksonomies-morfologiese studie van die genus Commiphora Jacq. Suid-Afrika. Thesis Pretoria. WEBSTER, G. L. 1956. A monographic study of the West Indian species of Phyllanthus. J. Am. Arb. 37: 91 —122; 217—268. WINSLOW, C. C., & J. R. HAVIS. 1967. Water movement in stems ofAmerican Holly at low temperature. Hort. Sc. 2: 24—25. M. ZAHUR, S. 1955. Comparative study ofsecondary phloem of423 species ofwoody Dicotyledonsbelong- ing to 85 families. Cornell Univ. Agric. Exp. Stn. Mem. 358.