Kew Royal Botanic Gardens.

Kew bulletin online

. (page 42 of 69)
Online LibraryKew Royal Botanic GardensKew bulletin → online text (page 42 of 69)
Font size
QR-code for this ebook

least, from these specimens, Tashri»L.b .xxuuiv confused the taxonomy
of these grasses had h^vi^J^lirs time become, I may mention that there
are three sheetp ?'^^>ttfnls herbarium under the name. Sheet 1 contains
a panicle ar p^ leaves of the true * Lemon-grass ' or Sereh, a young
panicle ^^HMith some of the upper leaves of the officinal * Schoen-
atitlir'^y^^ (Camel's hay) and a small inflorescence of C. coloratus.
2 and 3 are the B^amatci grass. Sheet 1 is initialled by
illdenow, and Sheet 3 is accompanied by a label with the
ame * Atidropogon Schomumthus ' in his handwriting. Under
the circumstances it is not surprising that the Indian botanists of
the time, who depended on a few books and relied for the com-
parison of their species with -those of extra-Indian floras on the
support of their European colleagues, formed equally confused
ideas concerning these fragrant grasses. Thus Ainslie, in his
* Materia Medica't (1S13), refers *Comachee pilloo' to ^ Andro-
pogon SchoenanthuSy and adds to it as synonyms vernaculars
which in reality belong to the * Lemon-grass ' and to the * Camel's
hay.' Wight, who collected the grass repeatedly, distributed some
' of his specimens (No. 1806) under the same name. Others he

f submitted to Nees, who was then planning a monogra|/h of the

Indian Olumaceae^ which, however, was never completed. Nees
i named Wight's grasses, which were subsequently distributed with

t his determinations, and described them as opportunity oflEered.


♦ WiUdenow, Spec. Plant, vol. iv., part ii, (1806), p. 9Ur.
t AinsUe, Mat. Med. (1818), p. 75.

Digitized by



Foundation op the Species. — Among the grassee thus dis
tributed was the Kamatci pillu (No. 17006). It was named by
Nees * Andropogon caesit^, /?.' Unfortunately, Nos. 1700, 1700c,
and 1700d were also distributed under that name. No. 1700 was
Andropogon pufnilus, Roxb., No. 1700c Cymhopogon coloratiM^
and No. 1700d a diseased state of C. coloratvs. The distribution
of Andropogon pumilus as Andropogon caesiua was obviously a
mere accident, as it is evident from the original specimen in
Wight's own herbarium that Nees really meant to apply the
name Andropogon caesius to No. 1700a, which is the same as
No. 17006, but is a very weak, (shade ?) form. The description
of Andropogon caesius appeared in Hooker k Amott's ^ Botany
of Beeohey's Voyage '• a few years later. Nos. 1700 (recte 1700a),
17006, and 1700c of Nees' distribution are quoted, and it is
obvious that the description was drawn up from all three
indiscriminately. To make matters worse, Nees referred to
this composite species specimens collected by Millett and Yachell
near Macao, which are neither identical with Nos. 1700a and
17006, nor with 1700c, but represent what is generally accepted
as Andropogon hamatulus or Andropogon Nardus var. hamatvlus.
Nor was this all. In 1813, Nees revised his determinations of
those grasses in Meyen's * Beitr^ge zur Botanik ' (p. 190), and re-
duced Andropogon caesius to Andropogon Martini^ quoting Wight,
No. 1700 and No. 1806 (the latter = Kamatci), under Andropogon
Martini; Nos. 1700a and 17006 (both = Kamatci) under Andro-
pogon Martini^ a and /3 respectively ; and No. 1700c (C coloratus)
under Andropogon Martini^ y. He further referred Millett's and
Vachell's Chinese specimens to thelatter,of which at least Vachell's —
I have not seen the other — is Andropogon hamatulus^ and he cited
'does not perai^ x^nfeif^-hfid ikasJling, No. 1095, which evidently
present the question as to what the ' 6'6^^s8 therefore remains
Oolconda was, there can be little doubt that the tragot>o^on caesiu^^
grass, of which there were whole fields to traverNos. 1700a
KamatXi-pillu of the Tamils. Of this name we hear fdS> di^nbt
time in * Samuel Browne's Seventh Book of East Indian i?^ the
edited and commented on by Petivert (1702). The plants "^the
form the subject of the paper were collected " between the 1;,
and 20th June, A.D. 1696, in the ways between Fort St. George
and Trippetee, which is about 70 miles off." One of them was
' Comachee pillee,' and of it Browne says : " This is Schoenanth,
which the natives here have not in great Esteem ; sometimes in the
Moors' Camps, the Horses, Camels, and Oxen which carry burthens
eat nothing else ; it is generally 2 or 3 feet high here about (but
near Color in reech soyl, I have seen it 8 feet high) [this gigantic
grass is no doubt C, Martini'] and thick as a Quill or small Reed ;
It's sometimes by the natives put into their Decoctions for Fevers,
and with us is deservedly of more esteem." PetiverJ identified
the * Comachee pillee ' with Plukenet's * Oramen Dactylon
Maderaspatense'* figured on plate 119, fig. 2 of his Almagesta
(1691), the type of which is in Plukenet's herbarium — it is the

* Herbert de Jaiirer in Yalentini, Hist. Simpl. (1732), p. 392.
t In Petiver, Mr. Samnel Browne, his Seventh Book of East India Plants in
Phil. Trans, xxiil. (1702), p. 1252.
X Petiver, l.o., p. 1251,

Digitized by



in April in a thoroughly dry condition yielded 0'711 per cent, of
oil, the diflEerences in the yield being attributed to the first lot
being fresh, whilst the other was dry. No analysis of the oil has
yet been made.

10. Cymbopogon polyneuros, Stapf.

{Andropogon polyneuros^ Steud.)

Just as C. Martini is replaced in the south-east of the Deccan
Peninsula by C. caesiuSy so another species takes its place in the
south-west. This species, C. polyneuroSy is, however, much better
defined than G. caesius. It is a moderately robust grass with a
tendency to copious branching from the collar so as to form dense
tufts of culms, with somewhat persistent, narrow, basal sheaths,
rather fat, smooth blades with a rounded base, more or less glaucous
beneath and often suffused with purple along the margin, and
with short, contracted, variegated panicles, the herbaceous sheaths
being usually deep brown-green with a narrow scarious margin,
the spikelets being green in the lower part, and more or less
blackish-purple in the upper. It was first distributed by Wight
(No. 1705) under the name Andropogon versicolor^ N.E., a name
chosen no doubt in allusion to the variegation of the inflorescence.
Nees never published a description of it. On the other hand,
Steudel has, in his * Synopsis Plantarum Oraminearum ' (1855),
(p. 388), an * Andropogon versicolor^ Nees MSS.,' under which
he quotes M. Schoenanihm, Wall. Cat. n. 8794L.' Wallich's
* Cat. n. 8794L.' is in Wallich's own herbarium identical with
n. 8794K., which Steudel (I.e.) cites under Andropogon clan -
deatintiSy Nees. Steudel's description of Andropogon versicolor
agrees neither with Wight's No. 1705 issued as ^Andropogon
versicolor^ N.E.,' nor with Wallich's n. 8794L. It is not clear
what the plant, which Steudel had in mind, was; it cannot
well have been Wight's * Andropogon versicolor^ N.E.' Wight
does not indicate the locality where his No. 1705 was collected
beyond the general note " Peninsula Ind. Orientalis." It agrees
absolutely, however, with a grass which has frequently been
collected in the Nilgiris, among others by Hohenacker who dis-
tributed it as " 933, Andropogon (Cymbopogon) nardoides /J. minor
N. ah Ey ; this was made by Steudel* the type of his Andro-
pogon polyiieuros. That name being perfectly unambiguous, its
specific component will have to be retained for the Nilgiri grass
in question in preference to versicolor ^ although the latter has very
generally been applied to it. Outside the Nilgiris, G. polyneuros
has so fsur only been observed in Ceylon where it is, particularly
at higher elevations (up to 1,500 m.), a locally common plant.
Thwaitest ^^ already called attention to the ** rather agreeable
aromatic odour" of the infiorescences of this species, adding
'^ tiiat the essential oil appears to be situated principally at the base
of the spikelets." According to a note in the * Tropical Agricul-
turist ' for 1901 (p. 873), the odour of the crushed leaves resembles
that of fennel or anise. There it is also stated that the grass

* Stendel, Syn. PI. Glam. pare. L (1855), p. 885.
t Thwaitee, Bnum. PI. ZeyL (1864), p. 367.

Digitized by



(which was identified by Trimen as A. ScJMenanthns^ var. verai-'
color) is particularly common in the Island of Delft in Adam's
Strait^ and has, under the name ' Delft ffrass,* the reputation of
being a good fodder for horses. I have seen no specimens from
that locality.

In 1902, a volatile oil was prepared from a sample of the grass
collected on the hills about Ootacamund by Mr. Proudlock. The
average yield is given as 0*25 per cent. ; but so far no analysis
seems to have been made of it.

11. Vetiveria sdzanioides, Stapf.

{Andropogon muricatuSy Retz.)

Khas Khas (Hind. ?)— Vetiver (TamU).

Early history. — If we admit certain deductions of the
Sanscritists — and there is no objection to them from the botanist's
point of view — this grass, best known as ' Khas Khas ' or * Vetiver,'
must have been popular with the peoples of Northern India for a
very long time. W. Jones,* as long ago as 1795, identified the
Usira of Ealidasa with ^ Khas Khas,' and Hesslert did the same
in his translation of the Ayurvedas, whilst among the more recent
interpreters of Sanscrit plant-names Dutt| has come to the same
conclusion. Other Sanscrit names which have been interpreted in
the same sense are Virana^ Ldrnajjaka (or Lamaja) and Bald.
According to the ^ Pharmacographia Indica' (vol. iii., p. 571),
"In Vedic times the ancient Hindus were instructed to build
their houses in a place where the Virana and KiMa (Desnuh
stachya bipinnata^ Stapf) were abundant." Ldmajjaka is, in the
same work (I.e., p. 562), referred to * Camel's Hay ' (C Schoenanth%is\
but the synonyms ' Dirgha-mulaka ' (long-rooted) and ^ Jalasdya'
(aquatic) with which Ldmajjaka is connected in the Nighandas, are
much more descriptive of * Khas Khas,' and Heyne's§ and Elliot'sl
interpretation of the term as connoting the latter is therefore
more plausible. Hessler also renders the Bald of SuSruta with
Andropogon muricutus. According to Duttf it stands for
Pavonia odorata, another plant whose aromatic roots are
frequently used in Hindu medicine. But the fact that Bald in
Hindi actually also denotes the roots of ' Khas Khas ' and that the
Bengali, Gujerati and Mahrati synonyms Vdld and Val6 are applied
in the same sense, supports Hessler's identification. In proof of
the assumption that 'Khas Khas' was an article of some importance
long ago, the authors of the 'Pharmacographia Indica' (vol. iii.,
p. 572) also refer to the discovery of some copper plates in the
village of Basahi in the district of Etawah, south-east of A^ra, it
being stated that on these copper plates, which are dated A.D. IIOJ
and 1174, the grass is mentioned among the articles subject to
royalties. The actual term used is * turushka-danda,' which

• Jonee in Asiat. Research, vol. iv. ( 1 795), p. 306.

t Hessler, SuS^ita's Ayuryedas, vol. iii. f 1850), p. 174.

t Dutt, Mat. Med. Hind. (1900), p. .S21.

§ Heyne, Tracti Hist. Stat. India (1814), p. 130,

II BlUot, FL Andh. (1859), p. 106.

if Dntt, 1.0., p. 293.

Digitized by



Babn Rajendrala*la Mitra* interprets as meaning ^'aromatic
reed " (tomshka = aromatic substance, danda = stick), and hence
also * KhBB Khas/ The latter term, now so commonly nsed, is
supposed to be of Persian origin, but this appears to me very
doubtful. It is mentioned in the Makhzan-el-Adwiyahf as a
kind of * Izkhir ' used in India, also known as * Izkhir-i-Jami,'
(Izkhir-i-Ajami, foreign Izkhir), and called by the Persians
* bikh-i-w&ljBk • (wdla root).

The *Khas Elhas' was long asfo equally well known to the
Dravidic peoples of the South. Rheede| described and figured the
grass under the Malayalim name Ramacciam, which is still in use in
Travancore (Ramach-cJiamy Moodeen Sherif ; BatnaSSam^ Stolz).
He states that the roots (but not the leaves) are fragrant and sold
in the bazars for medicinal purposes to prepare lotions, infusions,
and decoctions. It is, he remarks, very common throughout
Malabar and diligently cultivated by the natives, who propagate it
by dividing the tufts and planting them in loose soil. He further
observes that the best Ramacciam grows near Tutocorim, the
port which in our own day is still the principal place of export of
the roots of * Khas Khas ' or ' Vetiver.' Rheede's figure represents a
leaf -tuft with the leaf -tops cut off. Although somewhat crude, it is
perfectly characteristic, and it is difficult to understand how the
'Ramacciam^ of the *Hortus Malabaricus' could ever have passed —
as it so frequently has done — for the Memon-grass.'. Hermann §
(1672-1677) also found the roots in similar use at Colombo in
Ceylon, where they were known as ^ LumbutschUveru {radix
odorata)* and the grass itself as ^ Saewaendaray'' which name
has survived to the present day. About 25 years later (in 1700)
Dr. Bulkley sent it to Ch. Du Bois from Madras under the Tamil
name * Vettyveer ' (= Vetiver), the vernacular name by which the
grass is best known in Europe. Petiver | also received specimens of
it from Samuel Browne of Madras at about the same time and
announced them in his ' Museum * as ^' Oramen MadraspcUanum
majus citjus locustae spinulis eleganter annatae suntJ*' Some of
them he sent to ScheuchzerT, who from them drew up one of
those classic descriptions which for completeness and accuracy
remained long unequalled in agrostological literature.

Foundation op thb Species. Synonymy.— No notice was
taken of Scheachzer's description or of Petiver's and Du Bois's
specimens, and when Linnaeus, about 1770,** received the grass
from Koenig he described it as something new under the name
Phalaris zizanioides. Koenig, however, also sent specimens of
the grass to Retzius, who published it as Androftogon mnncatustt
in 1783. This name, which was suggested by Koenig himself, was

• Babo R&jendrala*la Mitra in Joarn. As. Soo. Benjr. (Kist. k Lit.), vol. xlii.
(1873), p. 320 ; Proo. As. Soc. Beng. (1^73), p. 161.

t Stttf Dymook, Warden, and Hooper, Phurinaoogi. In ale i (IbUa), vol. lii.,
p. 672.

t Eheede, Hort. Malab., vol. xii (170.^, tob. 72.

§ Hermann, Mns. Zeyl. (1726), p. 51.

i Petiver. Mas. Petiv. (1699), p. 53, no. 559.
Soheuchzer, Agroetogr. (1719), p. 103.
* Linnaens, Mant Alt. (1771), p. 183.
tt Ret*. Obwrv., vol. iU. (1783). p. 48.

26295 D 2

Digitized by



flubseqaenUy adopted by Roxburgh and most other botanists. M ore
recently,* however, it has been replaced by AndropogonsqtuirrosuSj
a name adopted by the younger Linnaeusf for a plant, also com-
mnnicated by Koenig, who found it " circa Zeylonam natans supra
etagna profundiora," and entirely distinct from Andropogon
muricatus. The specimen is still in Linnaeus' herbarium and was
correctly identified by R. BrownJ with his Panicum ahortivum^
that is Chamaeraphis spinescens^ a characteristic floating grass of
the Indo-Malayan region. Retzius§ himself is responsible for the
erroneous reduction of Andropogon sqicarrosus to Andropogon
muricatua^ which recently has been revived, although Roxburgh!
long ago drew attention to the confusion. ' Zizanioides ' being the
earliest specific epithet, it will have to be adopted for the *• Ehas
^las,^ so that its name under Vetiveria must be F. zizanioides.

Uses op THK Roots. — Koenig, in a note reproduced by Retzius,
remarks: **Tamulis Woetiwaer. Radices ab indigenis usitat-
issimae ob gratum odorem quem aqua irroratae spargunt. Ex his
Fldbella praecipue parantur quae pennis Pavonum ornantur."
This property of the roots of *Kbas Khas' of emitting a
pleasant odour as often as they are wetted and as long as
they are wet was also mentioned by JonesH in 1795. It
has led from early times to their being woven into screens and
mats (tatties), which are hung over doors or set in windows ; in
hot weather, when frequently sprinkled with water, they cool and
perfume the air. The fans (Tamil, visri) mentioned by Koenig
act in the same way. The root, in the powdered state, enters into
the composition of an Abir^^* or perfumed powder used by the
Hindus at the Holi festival. Such an Abir, Abir Izkhir^ is already
mentioned in the * Ain-i-Akbari,'tt the Annals of the Emperor
Akbar,the appellation *Izkhir' standing here for * Izkhir-i-Ajami,'
that is *Khas Khas.' The ' Schoenanthus ' powder which Herbert
de Jager|| found in use at Golconda in the second half of the
17th century was also most likely *KhasKhas' powder. For what
he says is this : " In Golkonda, this Schoenanthus is used in
powder-form for washing the hands on account of the very pleasant
odour it imparts very quickly to the water; hut the odour ceases
as soon as the hands are dry^

While, however, the use of the roots of Vetiveria zizanioides for
medicinal purposes and in perfumery has been universal in India
for a very long period, 1 have failed to find, among the earlier
writers, any definite and indisputable reference to the extraction
of an oil from them. It is true that Hessler, in his translation
of Susruta, mentions (vol. i., p. 160) ^^Andropogi muricati
spiritus distillatus "; but the word which he renders as " Andro'
pogi (sic) muricati " is ' Mrinala,' which by others, as for instance

^ Hackel, Andropog. (in DO. Monogr. Phaner., toL vi, 1889), p. 642.
t Linn^ fil., Snppl. (1781), p. 433.
t R. Brown, Prodr. Fl. Nov. Holl. (1810), p. 193.
§ Eetz., I.C., vol. V. (1789), p. 21.

f Eoxburgh, Fl. Ind., ed. Carey k Wall., vol. i. (1820), p. 270.
^ Jones in Aeiat. Research, vol. iv. (1795), p. S06.
♦• Diet. Boon. Prod. India, vol. i. (1885), p. 7.
ft Hooper, in Calcntta Review, Oct. 1904.

tt Herb, de Jager, in Valentini, Hist. SimpL (1782), p. 392. See dUo p. of
this paper.

Digitized by



by Dntt,* is interpreted as meaning the leaf -stalk of the Lotus, So
that for this reason alone the passage quoted cannot be adduced
as proof of an early knowledge of the distillation of oil from
^Khas Khas' or *Vetiver.' Indeed, the distillation of Vetiver
oil in India seems to be of very limitedf extent and there is
hardly any export, the oil being mainly produced in European
distilleries from the imported root; but even the import of the
roots as a regular article of commerce appears to be of com-
paratively recent date.

Natural Area and Cultivation. — The natural area of Andro-
pogon muricatus in India and Ceylon includes practically the "whole
country,5n the north up to altitudes of 600 m. Although common
in many parts of the country, particularly on the banks of rivers
and in rich, marshy soil, it is also at present, as in Rheede's time,
occasionally cultivated, as for instance in Rajputana and in Chutia
Nagpur. Eastwards the area extends into Burma. Throughout
the Malayan region, however, it occurs only in the cultivated state
or as an escape from gardens. It has also been introduced into
the Mascarenes, the West Indies, and Brazil ; but it seems that in
these countries oil is not distilled to any appreciable extent, except
perhaps in Reunion, where the grass must have been in cultivation
for at least 100 years, as the first sample of Yetiver-oil that was
chemically examined^ (in 1809) came from there.

12. Andropogon (Sect. Amphilophis) odoratos, LisK


This is a little-known grass which was discovered by Dymock
at Thana in 1875 and mentioned on account of its strong odour of
ginger under its vernacular name, Usadhana, in the first edition
of his "Materia Medica of Western India" (p. 693). In the
second edition of that work (p. 853) it was referred to Andropogon
Nardtis. Subsequently it was, however, recognised as a new
species by Mrs. J. C. Lisboa,§ and described as A. odoratv^. This
very aromatic grass is used by the peasantry of the Thana district
for medicinal purposes. An essential oil of a golden- yellow to a
deep sherry colour, with a distinctive odour, was obtained from it
by distillation, but it has not yet become an article of commerce.
The odour is, according to the "Pharmacographia Indica," vol. iii.,
p. 570, at first that of cassia and rosemary, but afterwards that of
oil of cassia or, according to Oildemeister and Hoffmann,! ^^^^ ^^
pine-needle oiK


(Oildemeister and Hoffmann, Volatile Oils, p. 285.)

Oildemeister and Hoffmann mention in their work on volatile
oils a "ginger-grass oil," of which they say that it is "an inferior
quality of palmarosa oil, or a mixture of the latter with much (up

• Dutt, Mat. Med. Hind., ed. 2. ri900), p. 109.

t Dnthie (Fodd. Grass. N. India, p. 37) mentions that at Bhira, in Oudh, a
perfume called itae is extracted from the roots of Vetiveria zUanioides^ and nsed
medicinaUy under the name of urama,

X Vauquelin, in Ann. Mus. Hist. Nat. Par., vol. xiv. (1809), pp. 28-31.

§ LisboA in Joum. Bomb. Nat. Hist. Soc, vol. iv. (1889), p, 123.

II Oildemeister k, Hofbnann, Vol Oils, p. 299.

Digitized by


to 90 per cent.) torpentiiie oil or mineral oils/', and further, that
'* occasionally other grasses are also nsed in the distillation
{Andropogon laniger ?), for some ginger-grass oils possess a
phelandrene-like odour, which is entirely wanting in the palma-
rosaoU/* In the Semi- Annual Reports of Schimmel & Co. for
October-November, 1902, Ginger-grass oil of good quality is, how-
ever, spoken of as available from a new source of production on
the Madras coast. In the April-May number of the same Reports
for 1904, it is stated that phellandrene was detected m the portions
boiling up to 80°, geraniol was also obtained, and a new alcohol
(CioHijO). Investigations into the composition of this oil were
continued, and in the October-November number of the Semi-
Annual Reports for 1904, Schimmel & Co. further indicate (p. 46)
the presence of two turpenes (rf-limonene and dipentene). The
aldehyd has a peculiar odour, which suggests both cenanthic
aldehyde and citronellal. In the complete absence of herbarium
specimens it is, of course, impossible to trace the origin of this oil.
It may, however, be useful to mention that the name ^ ginger-oil '
or its equivalents in various Indian languages has been in use for
a considerable time. * Ginger-grass ' is, for instance, mentiontjd
by Ainslie in his * Materia Medica ' (1813), p. 115, together with
its Tamil equivalent, * Shukkunari-pillu.'* In this case the grass
was Cymbofjogon flexuosus {see p. ^{19). Since then the name has
been more frequently used for Cgmbopogon Martini. Among
the vernacular names, those derived from the Sanscrit name Sont
(dry ginger), and therefore equivalent to the English name
* Ginger-grass,' have been variously applied. Stolz has, for
instance, punihi-hulla for C. Martini. Edge worth, according to
Duthie, has *Sent (sentTia)^ for Vetiveria zizanioideSy whilst
*'Sondhi ' stands as one of the Indian synonyms of Izkhir in the
Makhzan el Adwiyah.


Key to the Grasses.

Ctmbopoook. — Racemes of spikelets paired on a common
peduncle which is supported by and often enclosed in a more or
less boat-shaped bract (spathe) ; all the sessile spikelets alike, with
the exception of one (or more) at the base of the racemes (at least
of one of each pair).

Tardily flowering perennials ; innovations intravaginal, forming
dense tufts ; culms from dense bunches of firm, persistent leaf-
sheaths, more or less widened below ; blades long, hard, rough-
edged throughout, filiform to linear; the first (outer) glume of
the sessile spikelet flat or concave between the keels : —

Panicle narrow, of short, dense fascicles of raceme-pairs ;
raceme-joints villose all over, hairs long, more or less con-
cealing the sessile spikelets ; awn usually a straight, very
short bristle (Series Schoenanthi) : —

Basal leaf-sheaths in dense tufts, tightly clasping, thick-
ened below ; blades more or less filiform and flexuous,
except when very short ; raceme-fascicles more or less
simple 1. C. SchoenarUhus.

• Shokkn, Tamil for the dried yoot of ginger.

Digitized by VjOOQIC


Basal leaf -sheaths ultimately loosened and curled ; blades
flat ; raceme-fascicles compound • . 2. C7. Jwarancusa.
Panicles more often large and very compound ; raceme-
joints glabrous or pubescent on the back, bearded along the
sides, hairs increasing in length upwards, but not concealing
the sessile spikelets; awns, if present (they are normally
absent in the cultivated forms), distinctly geniculate with the
knee exseHed (Series Citrati) : —

Online LibraryKew Royal Botanic GardensKew bulletin → online text (page 42 of 69)